Formation and stability of organic zwitterions - The carbon acid pK as of the trimethylsulfonium and tetramethylphosphonium cations in water

Article abstract of DOI:10.1139/v05-155

We report second-order rate constants of k_DO = 7.5 × 10^-4 and 9.9 × 10^-5 (mol/L)^-1 s ^-1 for exchange for deuterium of the first methyl proton of the trimethylsulfonium and tetramethylphosphonium cations, respectively, in D ₂O at 25°C and 1 = 1.0 (KCl). The data were analyzed to give the following carbon acidities for these cationic carbon acids in water: (CH ₃)₃S⁺, pK_a = 28.5; (CH ₃)₄P⁺, pK_a = 29.4. These acidities are close to those of the neutral carbon acids acetonitrile and dimethylacetamide. This provides evidence that a portion of the stabilization of the cyanomethyl carbanion is due to resonance delocalization of negative charge from carbon to cyano nitrogen.

Full text of DOI:10.1139/v05-155

1536
Formation and stability of organic zwitterions —
The carbon acid pKas of the trimethylsulfonium
and tetramethylphosphonium cations in water1
Ana Rios, AnnMarie C. O’Donoghue, Tina L. Amyes, and John P. Richard
Abstract: We report second-order rate constants of kDO = 7.5 × 10–4 and 9.9 × 10–5 (mol/L)–1 s–1 for exchange for
deuterium of the first methyl proton of the trimethylsulfonium and tetramethylphosphonium cations, respectively, in
D2O at 25 °C and I = 1.0 (KCl). The data were analyzed to give the following carbon acidities for these cationic car-
bon acids in water: (CH3)3S+, pKa = 28.5; (CH3)4P+, pKa = 29.4. These acidities are close to those of the neutral car-
bon acids acetonitrile and dimethylacetamide. This provides evidence that a portion of the stabilization of the
cyanomethyl carbanion is due to resonance delocalization of negative charge from carbon to cyano nitrogen.
Key words: carbon acids, carbanions, ylides, proton transfer.
Résumé : Opérant dans le D2O, à 25 °C et une valeur de I = 1,0 (KCl), on a déterminé que les constantes de vitesse
du deuxième ordre pour l’échange du deutérium du premier proton méthyle des cations triméthylsulfonium et tétramé-
thylphosphonium sont respectivement de kDO = 7,5 × 10–4 et 9,9 × 10–5 (mol/L)–1 s–1. Les données ont été analysées
pour en tirer les acidités carboniques suivantes pour ces acides carboniques cationiques dans l’eau: (CH3)3S+, pKa =
28,5 et (CH3)4P+, pKa = 29,4. Ces acidités se rapprochent de celles d’autres acides carboniques, tel l’acétonitrile et le
diméthylacétamide. Ces résultats montrent qu’une portion de la stabilization du carbanion cyanométhyle est due à une
délocalization de résonance de la charge négative du carbone vers l’azote du groupe cyano.
Mots clés : acides carboniques, carbanions, ylures, transfert de proton.
[Traduit par la Rédaction] Rios et al. 1542
Introduction
tetramethylphosphonium and trimethylsulfonium cations.
Thus, the ylides of phosphonium and sulfonium cations are
stabilized relative to the corresponding ylides of ammonium
cations by the greater polarizability of phosphorus and sulfur
than of nitrogen, and possibly by hyperconjugation (15, 16)
and back-bonding from carbon to the d-orbitals of phospho-
rus and sulfur (17).
We are interested in characterizing the effects of strongly
electron-withdrawing substituents on carbon acidity and
have reported the kinetic and thermodynamic acidities of a
variety of weak carbon acids in water (1–12). The very low
carbon acidity of methane (13) is increased by substituents
that provide polar and (or) resonance stabilization of nega-
tive charge at the methyl carbanion; and the most powerful
purely polar stabilization of a methyl carbanion is provided
by a cationic group at a neighboring atom. In 1955 Doering
and Hoffmann (14) reported rate constants for deprotonation
of the tetramethylammonium, tetramethylphosphonium, and
trimethylsulfonium cations by hydroxide ion at several tem-
peratures. Their data showed that the tetramethylammonium
cation is a significantly weaker carbon acid than the
CH3
H3C N CH3
CH3
CH3
H3C P CH3
CH3
CH3
S
H3C
CH3
Me4N+
Me4P+
Me3S+
In recent years there have been several studies that docu-
ment the relatively strong carbon acidity of alkyl sulfonium
(18–26) and alkyl phosphonium (27, 28) cations. However,
the carbon acid pKas for the parent tetramethylphosphonium
and trimethylsulfonium cations in aqueous solution have not
been reported. To this end, we have determined the rate con-
stants for deprotonation of the tetramethylphosphonium cat-
ion (Me4P+) and trimethylsulfonium cation (Me3S+) in D2O
at 25 °C, using modern NMR methods to monitor transfer of
deuterium from the solvent to carbon acid. Our results agree
well with the results of Doering and Hoffmann (14); they ex-
tended earlier work by showing that there is no catalysis of
deprotonation of these cationic carbon acids by buffer bases.
An analysis of these data by the methods that we have devel-
oped to determine equilibrium carbon acidities in water (1–
Received 25 April 2005. Published on the NRC Research Press
Web site at http://canjchem.nrc.ca on 25 November 2005.
A. Rios. Departamento de Química Física, Facultad de
Química, Universidad de Santiago, 15706 Santiago de
Compostela, Spain.
A.C. O’Donoghue. Department of Chemistry, University
College Dublin, Belfield, Dublin 4, Ireland.
T.L. Amyes and J.P. Richard.2 Department of Chemistry,
University at Buffalo, SUNY, Buffalo NY 14260 USA.
1This article is part of a Special Issue dedicated to organic
reaction mechanisms.
2Corresponding author (e-mail: jrichard@chem.buffalo.edu).
Can. J. Chem. 83: 1536–1542 (2005)
doi: 10.1139/V05-155
© 2005 NRC Canada

Rios et al.
1537
12) provides the carbon acid pKas for deprotonation of
Me4P+ and Me3S+ in water.
the substrate and KOD at the same ionic strength (I = 1.0,
KCl) to give a final substrate concentration of 10 mmol/L.
At timed intervals an aliquot (600 µL) was removed from
the reaction mixture and the pD was adjusted to 7 to 8 by
the addition of concentrated DCl and 50 µL of 0.4 mol/L
phosphate buffer in D2O. The aliquots were either analyzed
immediately by H NMR spectroscopy or were frozen for
analysis at a later time.
The exchange for deuterium of the first proton of the
methyl groups of Me3S+ was followed by monitoring the
disappearance of the singlet at 2.911 ppm owing to the CH3
Experimental
Materials
1
All organic and inorganic chemicals were reagent grade or
better and were used without further purification. Trimethyl-
sulfonium methyl sulfate, tetramethylphosphonium chloride,
tetramethylammonium chloride, quinuclidine hydrochloride,
1,1,1,3,3,3-hexafluoro-2-propanol (HFIP), 2,2,2-trifluoro-
ethanol (TFE), and KOD (40% wt, +98% D) were from
Aldrich. Deuterium oxide (99.9% D) and deuterium chloride
(35% w/w, 99.5% D) were from Cambridge Isotope Labora-
tories.
groups and the appearance of the triplet at 2.899 ppm (JHD
=
1
2 Hz) owing to the CH2D groups, by H NMR. The ex-
change for deuterium of the first proton of the methyl groups
of Me4P+ was followed by monitoring the disappearance of
the doublet at 1.867 ppm (JHP = 16 Hz) owing to the CH3
groups and the appearance of the doublet of triplets at
1.852 ppm (JHP = 16 Hz, JHD = 2.5 Hz) owing to the CH2D
groups. The reactions were followed during the exchange for
deuterium of up to 25% and 5% of the first proton of the
methyl groups of Me3S+ and Me4P+, respectively.
Preparation of solutions
The acidic proton of quinuclidine hydrochloride was ex-
changed for deuterium before use as described previously
(1). TFE and HFIP were dissolved directly in D2O (99.9%
D), but at no time did this introduce >0.5 atom% of protium
into the solvent. Quinuclidine, trifluoroethanol, and HFIP
buffers were prepared by dissolving their acidic forms and
KCl in D2O followed by the addition of KOD to give the de-
sired acid:base ratio or pD at I = 1.0 (KCl).
A
CH
3
[2]
R =
A
CH D
2
A
+
CH
3
2
Solution pD was determined at 25 °C using an Orion
Model 720A pH meter equipped with a Radiometer
GK2321C or pHC4006-9 combination electrode. Values of
pD were obtained by adding 0.40 to the observed reading
of the pH meter (29). The concentration of deuterioxide
[3]
ln R = –kobsdt
ion at any pD was calculated from eq. [1], where Kw
=
Values of R, which is a measure of the progress of the
deuterium exchange reaction (4, 31), were calculated accord-
10–14.87 (mol/L)2 is the ion product of D2O at 25 °C (30) and
γOL = 0.79 or 0.75 is the apparent activity coefficient of
lyoxide ion under our experimental conditions determined
by the GK2321C or the pHC4006-9 electrode, respectively
(2).
ing to eq. [2], where A
and A
are the integrated areas
CH
CH D
3
2
of the signals owing to the CH3 and CH2D groups, respec-
tively. During the early stages of the reaction of Me3S+, it
was possible to obtain baseline separation of only the most
upfield peak of the triplet owing to the CH2D groups. In
these cases, the total integrated area of the triplet due to the
CH2D groups (ACH D) was obtained by multiplying the inte-
grated area of the r2esolved peak by three, and the area of the
10p D−pK
w
[1]
[DO−] =
γOL
1H NMR spectroscopy
singlet owing to the CH3 groups (A ) was obtained as the
CH
3
1H NMR spectra at 500 MHz were recorded in D2O at
25 °C on a Varian Unity Inova 500 spectrometer. The relax-
ation times for the methyl protons of Me3S+ and Me4P+,
determined for 10 mmol/L solutions of substrate at I =
1.0 (KCl), were found to be in the range of T1 = 6 to 7 s.
The relaxation delay between pulses was at least 10-fold
greater than the longest T1 of the protons of interest. Chemi-
cal shifts were reported relative to HOD at 4.67 ppm. Base-
lines were subjected to a first-order drift correction before
the determination of the integrated peak areas.
difference between the total peak area for both the singlet
and the triplet and A
(2). For the reaction of Me4P+,
CH D
2
only the most upfield peak of the doublet of triplets due to
the CH2D groups was cleanly resolved. The area of the dou-
blet of triplets due to the CH2D groups (ACH D) was obtained
by multiplying the integrated area of the r2esolved peak by
six, and the area of the doublet due to the CH3 groups (A
)
CH
3
was obtained as the difference between the total peak area
for all the signals and A
.
CH D
2
Semilogarithmic plots of reaction progress (R) against
time were linear with negative slopes equal to kobsd (s–1,
eq. [3]), which is the rate constant for exchange of a single
proton of the substrate.
A triplet at 3.164 ppm (J = 0.5 Hz) was observed for the
four equivalent methyl groups of the tetramethylammonium
cation. No signal for a CH2D group was observed after the
incubation of the substrate in the presence of 1.0 mol/L
KOD at 25 °C for 16 days. Several spurious signals ap-
peared during this prolonged reaction time, but no attempt
was made to assign these signals.
Kinetic measurements
All reactions were carried out in D2O at 25 °C and a con-
stant ionic strength of 1.0 maintained with potassium chlo-
ride. The deuterium exchange reactions of Me3S+ and
Me4P+ in buffered solutions were initiated by mixing solu-
tions of the substrate and the buffer at the same pD and ionic
strength (I = 1.0, KCl) to give a final substrate concentration
of 10 mmol/L. The reactions of Me4P+ and Me4N+ in alka-
line solutions of D2O were initiated by mixing solutions of
© 2005 NRC Canada

1538
Can. J. Chem. Vol. 83, 2005
Table 1. First-order rate constants for exchange for deuterium of the first proton of Me4P+ and Me3S+
in D2O at 25 °C and I = 1.0 (KCl).
[B]T (mol/L)b
pD
[DO–] (mol/L)c
kex (s–1)d
a
Carbon acid
(CH3)4P+
Brønsted base
DO–
fB
0.29
0.22
0.17
0.13
0.042
2.8× 10–5
2.2×10–5
1.6×10–5
1.2×10–5
3.9×10–6
CF3CH2O–
0.2
0.092
12.22
2.83×10–3
3.1×10–7
0.18
0.28
0.37
12.30
12.34
12.37
3.40×10–3
3.73×10–3
4.00×10–3
3.6×10–7
3.4×10–7
3.7×10–7
(CH3)3S+
(CF3)2CHO–
0.90
0.20
0.05
10.77
1.01×10–4
1.0×10–7
0.10
0.25
0.50
10.77
10.83
10.91
1.01×10–4
1.15×10–4
1.39×10–4
9.6×10–8
9.8×10–8
1.1×10–7
Quinuclidine
CF3CH2O–
0.05
11.50
5.40×10–4
3.7×10–7
0.20
0.05
0.20
0.05
11.47
11.92
11.89
12.04
5.04×10–4
1.42×10–3
1.32×10–3
1.87×10–3
3.3×10–7
9.0×10–7
7.6×10–7
1.42×10–6
0.40
0.10
0.20
12.12
2.25×10–3
1.42×10–6
aFraction of the Brønsted catalyst in the reactive basic form.
bTotal concentration of the Brønsted catalyst.
cConcentration of deuterioxide ion at the pD of the experiment calculated from eq. [1].
1
dObserved first-order rate constant for deuterium exchange determined by H NMR spectroscopy.
Results
Scheme 1.
The exchange for deuterium of the first proton of the
methyl groups of Me3S+ and Me4P+ in D2O at 25 °C and I =
1.0 (KCl) was followed by 1H NMR spectroscopy at
500 MHz (Scheme 1). Deuterium exchange at Me3S+ results
in disappearance of the singlet at 2.911 ppm owing to the
CH3 groups and appearance of an upfield-shifted triplet at
2.899 ppm owing to the CH2D groups. Similarly, exchange
at Me4P+ results in disappearance of the doublet at
1.867 ppm and the appearance of a doublet of triplets at
1.852 ppm owing to the CH2D groups. The deuterium iso-
kex = 9kobsd
CH3
S
CH3
S
A
B
H3C
CH3
H3C
CH2D
kex = 12kobsd
CH3
CH3
H3C P CH3
CH3
H3C P CH2D
CH3
various concentrations of deuterioxide ion or added buffers
in D2O at 25 °C and I = 1.0 (KCl). There is no significant
change in the value of kex for the deuterium exchange reac-
tion of Me4P+ as the concentration of trifluoroethanol buffer
is increased from 0.092 to 0.37 mol/L at pD 12.3. Similarly,
the value of kex for the deuterium exchange reaction of Me3S+
is independent of the concentration of added hexafluoro-
isopropanol, quinuclidine, or trifluoroethanol buffers (Table 1).
The absence of buffer catalysis of deuterium exchange
shows that, in basic solution, these cationic carbon acids are
deprotonated exclusively by the solvent.
1
tope effect on the H NMR chemical shift and the H–D cou-
pling constant are similar to those observed in our previous
work (1, 2, 4–6, 9, 10, 32, 33).
First-order rate constants (kobsd (s–1)) for exchange for
deuterium of a single proton of the cationic carbon acids
were determined as the slopes of semilogarithmic plots of
reaction progress R (eq. [2]) against time according to
eq. [3] (not shown). The reaction of Me3S+ to give the
monodeuteriated product occurs nine times as fast as the ex-
change of a single proton of the reactant, so that kex = 9kobsd
where kex (s–1) is the first-order rate constant for exchange of
the first proton of the reactant to give the monodeuteriated
product (Scheme 1A) (4, 31). Similarly, the reaction of
Me4P+ to give the monodeuteriated product occurs 12 times
as fast as the exchange of a single proton of the reactant, so
that kex = 12kobsd (Scheme 1B) (4, 31).
Figure 1 shows the dependence of kex (s–1) on the concen-
tration of deuterioxide ion in D2O for the deuterium ex-
change reaction of Me4P+. The slope of this correlation
gives kDO = 9.9 × 10–5 (mol/L)–1 s–1 (eq. [4]) where kDO is
the second-order rate constant for deuterium exchange cata-
lyzed by the deuterioxide ion.
Table 1 gives the values of kex (s–1) for the deuterium ex-
change reactions of Me3S+ and Me4P+ in the presence of
[4]
kex = kDO[DO–]
© 2005 NRC Canada

Rios et al.
1539
Fig. 1. Dependence of kex (s–1) for exchange for deuterium of
the first proton of Me4P+ on the concentration of deuterioxide
ion in D2O at 25 °C and I = 1.0 (KCl). The slope gives the
second-order rate constant (kDO) for DO–-catalyzed deuterium
exchange.
Fig. 2. pD–rate profiles of kex for deuterioxide ion catalyzed ex-
change for deuterium of the first proton of Me3S+ (᭜, this
work), Me4P+ (᭹, this work), CH3CN (᭿, data from ref. 4), and
CH3C(O)NMe2 (᭡, data from ref. 10) in D2O at 25 °C and I =
1.0 (KCl).
these deuterium exchange reactions are catalyzed by buffer
bases. (3) We wanted to obtain pKas for these cationic car-
bon acids in water and to draw comparisons with the other
data that we have reported over the last decade or so.
In 1955 Doering and Hoffmann (14) monitored the deute-
rium exchange reactions of several cationic carbon acids in
D2O by combustion of the recovered carbon acid to a mix-
ture of D2O and H2O, whose deuterium enrichment was then
determined by the “method of the falling drop”. Table 2
shows that the second-order rate constants (kDO) for
deuterioxide-ion-catalyzed exchange of the first proton of
Me4P+ and Me3S+ determined here agree to within 33%
with the earlier values of Doering and Hoffmann (14). The
slightly higher values reported in the earlier work were ob-
tained at a temperature of 26.8 °C compared with 25 °C
used in this work.
[DO– ] (mol/L)
Figure 2 shows the pD–rate profiles for the deuterium ex-
change reactions of Me4P+ (᭹) and Me3S+ (᭜). In cases
where values of kex were determined over a range of buffer
concentrations (Table 1), the values of pD and kex for the re-
action at the lowest buffer concentration were used for
Fig. 2. The solid line through the data for Me4P+ was calcu-
lated from the value of kDO = 9.9 × 10–5 (mol/L)–1 s–1 using
eq. [5]. A value of kDO = 7.5 × 10–4 (mol/L)–1 s–1 for Me3S+
was determined from the fit of the data to eq. [5] (10). Fig-
ure 2 also shows data for the deuterium exchange reactions
of acetonitrile (᭿) (4) and dimethylacetamide (᭡) (10) taken
from our earlier work.
⎛
⎜
⎜
⎝
⎞
⎟
⎟
⎠
kDOKw
Rate-determining step for hydron exchange
[5]
logkex = log
+ pD
The rate constant (kDO) for exchange of the first proton of
these cationic carbon acids in D2O may be limited either by
the chemical step of proton transfer from the substrate to
deuterioxide ion (kp, Scheme 2), or if this step is reversible
(k–p >> kreorg), by the rate constant for reorganization of sol-
vent to place a molecule of D2O in a position to deliver a la-
beled hydron to the ylide, kreorg ≈ 1011 s–1 (Scheme 2) (4, 11,
34). It has been shown in earlier work that reorganization of
solvent is largely rate-determining for the lyoxide ion cata-
lyzed deprotonation of cyanoalkanes (4). The following
comparisons between proton transfer at cyanoalkanes and at
Me4P+ and Me3S+ provide strong evidence that these reac-
tions proceed with the common rate-determining step of sol-
vent reorganization.
γOL
There was no detectable formation of a CH2D group after
the reaction of Me4N+ in 1.0 mol/L KOD in D2O at 25 °C
for 16 days. During this time, the appearance of several un-
identified NMR signals showed that there were other reac-
tion(s) of Me4N+, but these were not characterized. A peak
for a CH2D group that is 1% of the area of the peak due to
the CH3 groups could have been easily detected in these ex-
periments. This was used to calculate an upper limit of
kDO < 4 × 10–8 (mol/L)–1 s–1 for the deuterium exchange re-
action of Me4N+ catalyzed by deuterioxide ion.
Discussion
(1) There is a simple correlation between the decreasing
importance of buffer catalysis of carbon deprotonation
on moving from dihydroxyacetone phosphate (35) to
acetamide (10) (Scheme 3) and the increasing reactivity
and decreasing lifetime of the carbanion product of car-
bon deprotonation (10). No buffer catalysis is observed
for the limiting case where the reorganization of solvent
The present work was initiated for the following reasons:
(1) We wanted to determine second-order rate constants for
the deuterioxide ion catalyzed deuterium exchange reactions
of Me4P+ and Me3S+ using modern analytical methods and
under the same conditions used in our other studies of the
acidity of weak carbon acids. (2) We wanted to determine if
© 2005 NRC Canada

1540
Can. J. Chem. Vol. 83, 2005
Table 2. Kinetic and thermodynamic acidity of simple weak carbon acids in aqueous solution.
kDO
kHO
kHOH
(s–1)c
pKa
d
Carbon acid
(CH3)4N+
(CH3)4P+
((mol/L)–1 s–1)a
((mol/L)–1 s–1)b
(CH3)e
pKa
<4×10–8f
<1.7×10–8g
4.1×10–5g
1011
1011
>32.8
29.4
>33.4
30.0
9.9×10–5h
(1.2×10–4)i
7.5×10–4h
(CH3)3S+
3.1×10–4g
1011
28.5
29.0
(1.0×10–3)i
CH3CNj
2.7×10–4
1.1×10–4
2.9×10–6k
1011
7.3×109k
28.9
29.4
28.9
29.4
5.7×10–6
k
CH3C(O)NMe2
Note: At 25 °C and I = 1.0 (KCl).
aSecond-order rate constant for deprotonation of the carbon acid by deuterioxide ion in D2O.
bSecond-order rate constant for deprotonation of the carbon acid by hydroxide ion in H2O.
cFirst-order rate constant for protonation of the carbanion by solvent water, estimated as kHOH = kreorg = 1011 s–1,
where kreorg is the rate constant for solvent reorganization (see text).
dAcidity of the carbon acid in water, calculated according to eq. [6] of the text.
eStatistically corrected acidity of a single methyl group of the carbon acid.
fUpper limit calculated as the rate constant for a reaction that would have given a value of R = 1/1.005 (eq. [2]) af-
ter 16 days (see text).
gCalculated from the value of kDO using the relationship kHO = kDO/2.4, where 2.4 is the maximum secondary solvent
deuterium isotope effect on the reactivity of lyoxide ion (see text).
hData from Table 1.
iData from ref. 14.
jData for acetonitrile were taken from ref. 4.
kData for dimethylacetamide were taken from ref. 10.
Scheme 2.
kp
kreorg
Z
Z
Z
O
Z
O
D
=
H
H
H
D
D
H
k-p
H
D
H
H
S
H
O
H
H
O
D
D
CH3
Z
H3C
CH3
H3C P CH3
Scheme 3.
No detectable general
base catalysis
Strongest general
base catalysis
H
H
H
H
H
OH
H
H
H
H
CH3
H
CH3
S
NH3
O
H
H
H
~
~
~
~
H C P CH
3
3
O
>
>
>
H3C
~
~
CH3
O
O
N
CH3
H2N
EtO
MeO
2-
OPO3
(Ref. 33)
(Ref. 2)
(Ref. 6)
(Ref. 10)
(This work)
(Ref. 4)
to place a molecule of D2O in a position to deliver a la-
beled hydron to the carbanion is rate-determining, be-
cause it is not possible for the buffer catalyst to increase
the rate of solvent reorganization. Therefore, the obser-
vation that there is no buffer catalysis of carbon
deprotonation either of cyanoalkanes (4) or of Me4P+
and Me3S+ (this work) is consistent with a common
H3C
H
CH3
S
H
CN
rate-determining step of solvent reorganization (kreorg
Scheme 2) for these three proton transfer reactions.
,
1
2
© 2005 NRC Canada

Rios et al.
1541
Scheme 4.
(2) There are similar intrinsic rate constants at ∆pKa = 0 of
ko ≈ 104 (mol/L)–1 s–1 for deprotonation of the dimethyl-
9-fluorenylsulfonium ion (1) by primary amines in
water–DMSO (95:5) (20) and ko ≈ 103.6 (mol/L)–1 s–1 for
deprotonation of 9-cyanofluorene (2) in water–DMSO
kHO
Z
Z
+
HO
H
kHOH
H
H
H
H
+
(90:10) (18). This shows that the α-SMe2 and α-CN
substituents have similar effects on the Marcus intrinsic
barrier for proton transfer at carbon. The rate constants
kDO = 2.7 × 10–4 (mol/L)–1 s–1 for deprotonation of
acetonitrile (4) and kDO = 7.5 × 10–4 (mol/L)–1 s–1 for
deprotonation of Me3S+ (this work) are also similar. The
reactions, which proceed with similar rate constants and
Marcus intrinsic barriers, also have a similar thermody-
bond acceptor (42), and this solvent effect on pKa becomes
less with decreasing stabilization of the conjugate base by
hydrogen bonding to water. However, a substantially lower
pKa for Me3S+ in the organic solvent DMSO than in water
would require an unprecedented preferential solvation of the
products of proton transfer by DMSO. The pKa of 18.2 for
deprotonation of Me3S+ in DMSO is also inconsistent with
the pKa of 17.8 for deprotonation of the benzylic carbon of
Me2(PhCH2)S+ in DMSO (24) because the α-phenyl
substituent should result in a substantial increase in carbon
acidity. We note that the pKa of 18.2 for Me3S+ in DMSO
was reported in a Table in a review article (42), but to the
best of our knowledge, it has not been further documented.
namic drive force (12, 36, 37). Therefore, if k–p
>
kreorg ≈ 1011 s–1 for deprotonation of CH3CN, as pro-
posed in earlier work (4, 34), then the same relationship
must hold for deprotonation of Me3S+. We are not aware
of any data that would allow us to estimate the effect of
+
an α-PMe3 substituent on the intrinsic barrier to depro-
tonation of carbon.
Carbon acid pKas
Substituent effects on carbon acidity
Table 2 gives the carbon acid pKas for deprotonation of
Me4P+ and Me3S+ in water obtained in this work. These
pKas were calculated using eq. [6] derived from Scheme 4
where kHO is the second-order rate constant for deprotona-
tion of the carbon acid by hydroxide ion in H2O (Table 2),
kHOH = kreorg = 1011 s–1 is the first-order rate constant for
protonation of the carbanion zwitterion by solvent water (4,
10, 11, 34), and Kw = 10–14 (mol/L)2 is the ion product of
water.
Computational studies provide evidence for formal reso-
nance delocalization of negative charge onto nitrogen at the
cyanomethyl carbanion (43). However, it is not clear
whether this delocalization of charge results in a large stabi-
lization of this carbanion. For example, delocalization of
negative charge at the acetate anion provides a simple expla-
nation for the relatively high acidity of acetic acid. However,
it has been proposed that this high acidity is largely due to
inductive stabilization of negative charge by the electron-
withdrawing carbonyl group (44). The following observa-
tions provide evidence that there is significant stabilization
of the cyanomethyl carbanion by both polar and resonance
interactions between the negative charge and the α-CN
group.
(1) The polar effect of a cationic α-substituent on carbanion
stability must be larger than the polar effect of the neu-
tral α-cyano group, where the interaction is between the
formal negative charge of the carbanion and the partial
positive charge of the cyano dipole. Therefore, the es-
sentially identical values of pKa = 28.9 for acetonitrile
(4) and pKa = 29.0 for a single methyl group of Me3S+
(Table 2) show that there is sufficient resonance stabili-
zation of the cyanomethyl carbanion to balance the
greater stabilization of the sulfur ylide from polarization
and back-bonding effects.
⎛
⎜
⎜
⎝
⎞
⎟
⎟
⎠
kHOH
[6]
pKa = pKw + log
kHO
The values of kHO (Table 2) were calculated from the cor-
responding rate constants for deprotonation of the carbon
acids by deuterioxide ion (Table 1) using the relationship
kHO = kDO /2.4 where 2.4 is the maximum secondary solvent
deuterium isotope effect on the reactivity of lyoxide ion
(38). This is appropriate because almost the entire difference
in the basicity of deuterioxide and hydroxide ions will be ex-
pressed in the transition state for carbanion formation when
solvent reorganization is rate-determining for exchange (4).
The principle of microscopic reversibility requires that the
reverse protonation of these carbanion zwitterions by solvent
water is also limited mainly by solvent reorganization, with
a rate constant on the order of that for the dielectric relax-
ation of water, kHOH = kreorg = 1011 s–1 (39–41).
(2) The Hammett substituent constants σ = 0.82 for p-
The greater carbon acidity of Me3S+ and Me4P+ than of
Me4N+ (Table 1) is well-known (14). It represents the
greater stabilization of sulfur and phosphorus than of nitro-
gen ylides from polarization of charge by the larger third-
row atoms and, possibly, by hyperconjugation (15, 16) and
back-bonding from carbon to the d orbitals of sulfur and
phosphorus (17).
We are not able to account for the difference between the
reported pKas of 28.5 (Table 2) and 18.2 (42) for depro-
tonation of Me3S+ in water and DMSO, respectively. The
change from DMSO to water causes a large decrease in pKa
when the conjugate base of the acid is a strong hydrogen
NMe3+ and σ = 0.66 for p-CN (45) show that a p-NMe3
+
substituent provides a larger stabilization of negative
charge than does a p-CN substituent when the interac-
tion occurs across a benzene ring. By comparison, the
observation that pKa = 28.9 for acetonitrile is much
lower than pKa > 33.4 for a single methyl group of
Me4N+ (Table 2) shows that an α-CN substituent pro-
vides a substantially larger stabilization of negative
+
charge at a neighboring carbon than does an α-NMe3
group. This shows that there is sufficient resonance sta-
bilization of the cyanomethyl carbanion to balance the
greater polar stabilization of the nitrogen ylide.
© 2005 NRC Canada

1542
Can. J. Chem. Vol. 83, 2005
14. W.v.E. Doering and A.K. Hoffmann. J. Am. Chem. Soc. 77,
521 (1955).
15. R.K. Bansal, N. Gupta, S. Singh, K. Karaghiosoff, P. Mayer,
and M. Vogt. Tetrahedron Lett. 45, 7771 (2004).
16. J.F. King, R. Rathore, Z. Guo, M. Li, and N. Payne. J. Am.
Chem. Soc. 122, 10 308 (2000).
It is interesting that no large intrinsic barrier to proton
transfer arises from the requirement for the movement of
negative charge from nitrogen to carbon on moving to the
transition state for protonation of α-cyano carbanions (4, 12,
46, 47). As discussed earlier (4), the proton transfer to α-
cyano carbanions resembles proton transfer to acetate anion
(48) in that the requirement for localization of negative
charge at the reacting basic atom (carbon or oxygen) pro-
ceeding to the transition state does not result in a large in-
trinsic barrier for proton transfer.
17. D.A. Dixon, T.H. Dunning, R.A. Eades, and P.G. Gassman. J.
Am. Chem. Soc. 105, 7011 (1983).
18. C.F. Bernasconi and F. Terrier. J. Am. Chem. Soc. 109, 7115
(1987).
19. C.J. Murray and W.P. Jencks. J. Am. Chem. Soc. 110, 7561
(1988).
20. C.J. Murray and W.P. Jencks. J. Am. Chem. Soc. 112, 1880
(1990).
21. N.H. Werstiuk and S. Banerjee. J. Org. Chem. 46, 470 (1981).
22. G. Barbarella, A. Garbesi, and A. Fava. Helv. Chim. Acta, 54,
253 (1971).
23. V. Cere, C. Paolucci, S. Pollicino, E. Sandri, and A. Fava. Tet-
rahedron Lett. 24, 839 (1983).
The α-fluorenyl group at the dimethyl-9-fluorenylsulfonium
ion (1, pKa = 13.7 in water–DMSO (95:5) (20)) causes a 16-
unit decrease in the statistically corrected pKa of 29.5 for a
single proton of Me3S+, while the α-fluorenyl group at 9-
cyanofluorene (2, pKa = 10.7 in water–DMSO (90:10) (18))
causes a larger 19-unit decrease in the statistically corrected
pKa of 29.4 for a single proton of acetonitrile. This shows
that the addition of an α-fluorenyl group results in a larger
+
attenuation of the α-SMe2 than of the α-CN substituent ef-
24. J.-P. Cheng, B. Liu, Y. Zhao, Y. Sun, X.-M. Zhang, and Y. Lu.
J. Org. Chem. 64, 604 (1999).
fect on carbon acidity. We suggest that this reflects the par-
ticularly sharp falloff (1/r4) in the stabilization of negative
charge by polarization onto a atom with increasing radius of
separation of the interacting centers (45).
25. O. Hofer and E.L. Eliel. J. Am. Chem. Soc. 95, 8045 (1973).
26. O.T. Magnusson and P.A. Frey. Bioorg. Chem. 30, 53 (2002).
27. X.-M. Zhang and F.C. Bordwell. J. Am. Chem. Soc. 116, 968
(1994).
28. A.-P. Bettencourt, A.M. Freitas, M.I. Montenegro, M.F. Niel-
sen, and J.H.P. Utley. J. Chem. Soc. Perkin Trans. 2, 515
(1998).
Acknowledgement
29. P.K. Glasoe and F.A. Long. J. Phys. Chem. 64, 188 (1960).
30. A.K. Covington, R.A. Robinson, and R.G. Bates. J. Phys.
Chem. 70, 3820 (1966).
We acknowledge the National Institutes of Health (GM
39754) for generous support of this work.
31. C.J. Halkides, P.A. Frey, and J.B. Tobin. J. Am. Chem. Soc.
115, 3332 (1993).
32. R.W. Nagorski and J.P. Richard. J. Am. Chem. Soc. 118, 7432
(1996).
33. J.P. Richard and R.W. Nagorski. J. Am. Chem. Soc. 121, 4763
(1999).
34. J.C. Fishbein and W.P. Jencks. J. Am. Chem. Soc. 110, 5087
(1988).
References
1. T.L. Amyes and J.P. Richard. J. Am. Chem. Soc. 114, 10297
(1992).
2. T.L. Amyes and J.P. Richard. J. Am. Chem. Soc. 118, 3129
(1996).
3. R.W. Nagorski, T. Mizerski, and J.P. Richard. J. Am. Chem.
Soc. 117, 4718 (1995).
4. J.P. Richard, G. Williams, and J. Gao. J. Am. Chem. Soc. 121,
715 (1999).
5. A. Rios and J.P. Richard. J. Am. Chem. Soc. 119, 8375 (1997).
6. A. Rios, T.L. Amyes, and J.P. Richard. J. Am. Chem. Soc.
122, 9373 (2000).
35. J.P. Richard. J. Am. Chem. Soc. 106, 4926 (1984).
36. R.A. Marcus. J. Phys. Chem. 72, 891 (1968).
37. A.J. Kresge. Chem. Soc. Rev. 2, 475 (1973).
38. A.J. Kresge, R.A. More O’Ferrall, and M.F. Powell. In Iso-
topes in organic chemistry. Vol. 7. Edited by E. Buncel and
C.C. Lee. Elsevier, New York. 1987.
7. Y. Chiang, A.G. Griesbeck, H. Heckroth, B. Hellrung, A.J.
Kresge, Q. Meng, A.C. O’Donoghue, J.P. Richard, and J.
Wirz. J. Am. Chem. Soc. 123, 8979 (2001).
8. A. Rios, J. Crugeiras, T.L. Amyes, and J.P. Richard. J. Am.
Chem. Soc. 123, 7949 (2001).
39. K. Giese, U. Kaatze, and R. Pottel. J. Phys. Chem. 74, 3718
(1970).
40. U. Kaatze. J. Chem. Eng. Data, 34, 371 (1989).
41. U. Kaatze, R. Pottel, and A. Schumacher. J. Phys. Chem. 96,
6017 (1992).
9. A. Rios, J.P. Richard, and T.L. Amyes. J. Am. Chem. Soc.
42. F.G. Bordwell. Acc. Chem. Res. 21, 456 (1988).
43. K.B. Wiberg and H. Castejon. J. Org. Chem. 60, 6327 (1995).
44. M.R.F. Siggel, A. Streitwieser, and T.D. Thomas. J. Am.
Chem. Soc. 110, 8022 (1988).
124, 8251 (2002).
10. J.P. Richard, G. Williams, A.C. O’Donoghue, and T.L. Amyes.
J. Am. Chem. Soc. 124, 2957 (2002).
11. T.L. Amyes, S.T. Diver, J.P. Richard, F.M. Rivas, and K. Toth.
45. J. Hine. Structural effects on equilibria in organic chemistry.
J. Am. Chem. Soc. 126, 4366 (2004).
Wiley, New York. 1975.
12. J.P. Richard, T.L. Amyes, and M.M. Toteva. Acc. Chem. Res.
34, 981 (2001).
13. A. Streitwieser, M.R. Granger, F. Mares, and R.A. Wolf. J.
46. C.F. Bernasconi. Acc. Chem. Res. 20, 301 (1987).
47. F. Hibbert, F.A. Long, and E.A. Walters. J. Am. Chem. Soc.
93, 2829 (1971).
Am. Chem. Soc. 95, 4257 (1973).
48. M. Eigen. Angew. Chem. Int. Ed. Engl. 3, 1 (1964).
© 2005 NRC Canada