Carbanion Stabilization by Adjacent Sulfur: Polarizability, Resonance, or Negative Hyperconjugation? Experimental Distinction Based on Intrinsic Rate Constants of Proton Transfer from (Phenylthio)nitromethane and 1-Nitro-2-phenylethane

Article abstract of DOI:10.1021/jo9719463

(Phenylthio)nitromethane, PhSCH₂NO₂, is about as acidic as PhCH₂NO₂ and about 4 pK_a-units more acidic than CH₃NO₂ in water or aqueous DMSO, showing the well-known acidifying effect of thio substituents in the α-position of carbon acids. Over the years various interpretations have been offered for the acidifying effect of sulfur groups: d-p π-resonance, polarizability, and negative hyperconjugation. Assuming that the nature of the factors that potentially stabilize the transition state of the proton transfer from the carbon acid are the same as those that potentially stabilize the carbanion, we show that a distinction between these interpretations can be based on the effect of the phenylthio group on the intrinsic rate constants (k_o) of proton transfer. Such intrinsic rate constants were determined for the deprotonation of PhSCH₂NO₂ and PhCH₂CH₂NO₂ by amines in water and 90% DMSO-10% water; in both solvents k_o for PhSCH₂NO₂ was found to be substantially higher than for PhCH₂CH₂NO₂ as well as for other nitroalkanes reported previously. Based on a detailed analysis of how various factors such as resonance, inductive effects, polarizability, and positive and negative hyperconjugation affect the intrinsic rate constants for proton transfer, it is concluded that the high k_o values for PhSCH₂NO₂ result from a combination of the inductive and polarizability effect of the PhS group and that d-p π-resonance and negative hyperconjugation play a minor role if any.

Full text of DOI:10.1021/jo9719463

1944
J . Org. Chem. 1998, 63, 1944-1953
Ca r ba n ion Sta biliza tion by Ad ja cen t Su lfu r : P ola r iza bility,
Reson a n ce, or Nega tive Hyp er con ju ga tion ? Exp er im en ta l
Distin ction Ba sed on In tr in sic Ra te Con sta n ts of P r oton Tr a n sfer
fr om (P h en ylth io)n itr om eth a n e a n d 1-Nitr o-2-p h en yleth a n e
Claude F. Bernasconi* and Kevin W. Kittredge
Department of Chemistry and Biochemistry of the University of California, Santa Cruz, California 95064
Received October 22, 1997
(Phenylthio)nitromethane, PhSCH2NO2, is about as acidic as PhCH2NO2 and about 4 pKa-units
more acidic than CH3NO2 in water or aqueous DMSO, showing the well-known acidifying effect of
thio substituents in the R-position of carbon acids. Over the years various interpretations have
been offered for the acidifying effect of sulfur groups: d-p π-resonance, polarizability, and negative
hyperconjugation. Assuming that the nature of the factors that potentially stabilize the transition
state of the proton transfer from the carbon acid are the same as those that potentially stabilize
the carbanion, we show that a distinction between these interpretations can be based on the effect
of the phenylthio group on the intrinsic rate constants (ko) of proton transfer. Such intrinsic rate
constants were determined for the deprotonation of PhSCH2NO2 and PhCH2CH2NO2 by amines in
water and 90% DMSO-10% water; in both solvents ko for PhSCH2NO2 was found to be substantially
higher than for PhCH2CH2NO2 as well as for other nitroalkanes reported previously. Based on a
detailed analysis of how various factors such as resonance, inductive effects, polarizability, and
positive and negative hyperconjugation affect the intrinsic rate constants for proton transfer, it is
concluded that the high ko values for PhSCH2NO2 result from a combination of the inductive and
polarizability effect of the PhS group and that d-p π-resonance and negative hyperconjugation
play a minor role if any.
In tr od u ction
hyperconjugation, has also been invoked by several
authors.9,11-13 This mechanism involves double bond-
no bond resonance structures (e.g., eq 2).14
It is well-known that R-alkylthio or R-arylthio groups
increase the acidity of adjacent CH bonds. This increased
acidity has led to numerous synthetic applications.1 The
acidifying effect has been attributed to the stabilization
of the carbanion by sulfur.2 The mechanism by which
this stabilization occurs has generated much interest.
Originally this stabilization was explained in terms of a
resonance effect involving the d orbitals of sulfur, i.e.,
d-p π-bonding between the carbanion lone pair and
sulfur 3d orbitals,2-7 e.g., eq 1. However, theoretical
work has challenged this notion, suggesting that the
The d-p π-bonding mechanism is no longer strongly
advocated; regarding polarizability and negative hyper-
conjugation, they may well both be important.9,11-13
However, based on a recent high level ab initio study,
Wiberg et al.15 have concluded that negative hypercon-
jugation is the main factor in the stabilization of the
dimethyl sulfide anion by sulfur. Similar conclusions
were reached by Cuevas and J uaristi.16
In the present study we attempt to evaluate the
relative influence of the various interaction mechanisms
by using an approach that combines kinetic and ther-
modynamic data. It is based on the assumption that the
kinds of the factors that potentially stabilize the transi-
tion state of the deprotonation of a carbon acid are the
same as of those that potentially stabilize the carbanion,
i.e., polarizability, d-p π-bonding, negative hyperconju-
gation, and possibly others (see below). It exploits the
stabilization is mainly due to the polarizability of
sulfur.8-11 A third interaction mechanism, negative
(1) (a) Corey, E. J .; Seebach, D. Angew. Chem., Int. Ed. Engl. 1965,
4, 1075. (b) Seebach, D. Synthesis 1969, 1, 19. (c) Corey, E. J .;
Erickson, B. W. J . Org. Chem. 1971, 36, 3553.
(2) For reviews, see (a) Price, C. C.; Oae, S. Sulfur Bonding; Ronald
Press: New York, 1962. (b) Cram, D. J . Fundamentals of Carbanion
Chemistry; Academic Press: New York, 165; pp 71-84.
(3) Oae, S.; Tagaki, W.; Ohno, A. Tetrahedron 1964, 20, 417.
(4) Eliel, E. L.; Hartmann, A. A.; Abatjoglou, A. G. J . Am. Chem.
Soc. 1974, 96, 1807.
(5) (a) Wolfe, S.; LaJ ohn, L. A.; Bernardi, F.; Mangini, A.; Tonachini,
G. Tetrahedron Lett. 1983, 24, 3789. (b) Wolfe, S.; Stolow, A.; LaJ ohn,
L. A. Tetrahedron Lett. 1983, 24, 4071.
(6) Bernardi, F.; Mangini, A.; Tonachini, G.; Vivarelli, P. J . Chem.
Soc., Perkin Trans. 2 1985, 111.
(9) Lehn, J .-M.; Wipff, G. J . Am. Chem. Soc. 1976, 98, 7498.
(10) Bernardi, F.; Csizmadia, I. G.; Mangini, A.; Schlegel, H. B.;
Wangbo, M.-H.; Wolfe, S. J . Am. Chem. Soc. 1975, 97, 2209.
(11) Schleyer, P. v. R.; Clark, T.; Kos, A. J .; Spitznagel, G. W.; Rohde,
C.; Arad, D.; Houk, K. N.; Rondan, N. G. J . Am. Chem. Soc. 1984, 106,
6467.
(12) Hopkinson, A. C.; Lien, M. H. J . Org. Chem. 1981, 46, 998.
(13) Schleyer, P. v. R.; Kos, A. J . Tetrahedron 1983, 39, 1141.
(14) Equation 2 can be visualized as donation of the carbanion lone
pair to the C-S σ* antibonding orbital.
(15) Wiberg, K. B.; Castejon, H. J . Am. Chem. Soc. 1994, 116, 10489.
(16) Cuevas, G.; J uaristi, E. J . Am. Chem. Soc. 1997, 119, 7545.
(7) Bordwell, F. G.; Bares, J . E.; Bartmess, J . E.; Drucker, G. F.;
Gerhold, J .; McCollum, G. J .; Van der Puy, M.; Vanier, N. R.;
Matthews, W. S. J . Org. Chem. 1977, 42, 326.
(8) Streitwieser, A., J r.; Williams, J . E. J . Am. Chem. Soc. 1975,
97, 191.
S0022-3263(97)01946-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/26/1998

Carbanion Stabilization by Adjacent Sulfur
J . Org. Chem., Vol. 63, No. 6, 1998 1945
fact that typically the relative importance of these factors
in stabilizing the transition state is not the same as in
stabilizing the carbanion.
The reasoning is as follows. As has been amply
demonstrated,17 the transition state of the deprotonation
of carbon acids activated by π-acceptors is imbalanced
in the sense that sp3 f sp2 rehybridization of the
R-carbon and charge delocalization into the π-acceptor
group (Y in eq 3) lag behind proton transfer, i.e., δY/δC <
pyramidal carbon. Hence, relative to the degree of proton
transfer, the transition state with its charge mainly
localized on the adjacent carbon benefits more from the
stabilizing influence of the sulfur than the carbanion.
(Phenylthio)nitromethane, PhSCH2NO2, appeared to
be a suitable candidate for the study of the effect of an
R-thio group along the lines described above because
proton transfers from nitroalkanes have strongly imbal-
anced transition states.17,18,20 Furthermore, intrinsic rate
constants for the deprotonation of several nitroalkanes
which would be suitable for purposes of comparison, are
already known. With the above reasoning in mind, we
studied the proton-transfer kinetics of the reaction of
PhSCH2NO2 and also of 1-nitro-2-phenylethane, PhCH2-
CH2NO2, with amines in water and 90% DMSO-10%
water. The latter substrate was not only included to
serve as additional reference nitro compound but because,
as will become apparent, it displays some interesting
features of its own.
ø/(1 - ø) in eq 3; δB is the amount of charge transferred
from B to the carbon acid at the transition state, -δC
and -δY are the partial negative charges on C and Y,
respectively, at the transition state, while -ø is the
negative charge on Y and -1 + ø the negative charge on
C of the product ion. In a nutshell, the reason for this
lag is that the degree of charge delocalization into the Y
group is coupled to the degree of C-Y bond formation
which in turn depends on the fraction of charge that has
been transferred from the base to the carbon acid. Since
neither C-Y π-bond formation nor charge transfer are
complete at the transition state, the charge delocalized
into the Y group is just a fraction of a fraction, and
therefore quite small.17,18 The consequence of this lag is
that resonance stabilization of the transition state is
disproportionately weak compared to that of the anion,
and hence the intrinsic19 rate constant of the reaction is
reduced compared to that of a reaction where Y is not a
π-acceptor, or a weaker π-acceptor.17
Similar reasoning should be applicable to resonance
effects that involve d-p π-bonding (eq 1) or negative
hyperconjugation (eq 2), i.e., their development at the
transition state is expected to lag behind proton transfer.
In other words, if d-p π-bonding or negative hypercon-
jugation are the dominant carbanion-stabilizing mecha-
nisms of R-thio groups, one expects the intrinsic rate
constant of the reaction to be reduced if X in eq 3 is a RS
group. This is because the charge delocalization symbol-
ized by the resonance structures 1 (d-p π resonance) or
2 (negative hyperconjugation) will have made minimal
Resu lts
Gen er a l F ea tu r es. Proton-transfer rates and pKa
values of PhSCH2NO2 and PhCH2CH2NO2 were deter-
mined in water at 25 °C and in 90% DMSO-10% water
(v/v) at 20 °C. The reason for choosing two different
temperatures was that in water there are more literature
data available for comparisons at 25 °C, while in 90%
DMSO, 20 °C has been the standard temperature used
in previous studies.
Most measurements were made in amine buffers and
KOH solutions. Pseudo-first-order conditions with the
nitroalkanes as the minor component were used through-
out. The ionic strength was kept constant with KCl at
0.5 M in water and at 0.06 M in 90% DMSO. The
pKCa H values for PhSCH2NO2 and PhCH2CH2NO2 as
carbon acids were determined by classic spectrophoto-
-
metric methodology according to eq 4 where A, AC and
ACH are the absorbances at pH ∼ pKCa H, pH . pKaCH and
AC- - A
pKCa H ) pH + log
(4)
A - ACH
pH , pKCa H, respectively. The pKNa OH values of the aci-
forms were only determined in 90% DMSO-10% water;
they were obtained kinetically as described below.
Kin etics in Wa ter . Proton-transfer rates were mea-
sured for the reactions of PhSCH2NO2 and PhCH2CH2-
NO2 with primary aliphatic and secondary alicyclic
amines. The reactions can be described by eq 5
progress at the transition state, just as is the case for
the resonance structure 3 (p-p π resonance). That
positive hyperconjugation follows this pattern has been
shown by Kresge.18
On the other hand, if carbanion stabilization with X
) RS in eq 3 is mainly a polarizability effect, the intrinsic
rate constants should be enhanced. This is because
polarizability effects do not depend on charge delocaliza-
tion/resonance development, i.e., they can operate on a
with
k1 ) kH1 O + kB1 [B] + k1OHaOH
(6)
(7)
2
-
(17) For recent reviews, see (a) Bernasconi, C. F. Acc. Chem. Res.
1987, 20, 301. (b) Bernasconi, C. F. Acc. Chem. Res. 1992, 25, 9. (c)
Bernasconi, C. F. Adv. Phys. Org. Chem. 1992, 27, 119.
(18) Kresge, A. J . Can. J . Chem. 1974, 52, 1897.
(19) The intrinsic rate constant, ko, for a reaction with a forward
rate constant k1 and a reverse rate constant k-1 is defined as ko ) k1
) k-1 when K1 ) 1 (∆G° ) 0); in proton transfers statistical factors
are usually included.
k-1 ) kH-1aH+ + k-BH1 [BH+] + kH2O
-1
where B is the amine. All measurements were carried
(20) Bordwell, F. G.; Boyle, W. J ., J r. J . Am. Chem. Soc. 1972, 94,
3907.

1946 J . Org. Chem., Vol. 63, No. 6, 1998
Bernasconi and Kittredge
F igu r e 1. Plots of kobsd vs amine concentration for the reaction
of PhCH2CH2NO2 with morpholine in water at 25 °C. [: pH
8.18; 1: pH 8.30; b: pH 8.78; 9: pH 9.08; 2: pH 9.38. Data
at pH 8.48 and 9.26 omitted for clarity.
+
F igu r e 2. Plot of slope(B) vs aH according to eq 9 for the
reaction of PhCH2CH2NO2 with morpholine in water at 25 °C.
Note this plot includes points from data that were omitted from
Figure 1.
out at pH values considerably above the estimated
Ta ble 1. Su m m a r y of Ra te Con sta n ts a n d p KCa H Va lu es
for th e Rea ction s of P h SCH2NO2 a n d P h CH2CH2NO2 w ith
Am in es a n d OH- in Wa ter a t 25 °Ca
pKNa OH of the aci-form of XCH2NO2 (see Discussion) and
-
hence protonation of XCHdNO2 on the nitro group is
not included in eq 5.
The pseudo-first-order rate constant for equilibrium
approach is given by eq 8. Kinetic determinations were
BH
-1
b
b
B
pKBa H
kB1 (M-1 s-1
)
k
(M-1 s-1
)
PhSCH2NO2 (pKCH ) 6.67 ( 0.02)
a
piperidine
11.24
9.94
9.34
8.78
10.87
9.64
8.21
5.50
4.70
3.16 × 103
1.05 × 103
4.18 × 102
1.53 × 102
5.92 × 102
1.15 × 102
1.3 × 101
0.36
0.083
0.564
0.814
1.19
0.0374
0.123
0.387
5.33
piperazine
kobsd ) k1 + k-1
(8)
PZ-CH2CH2OHc
morpholine
made in amine buffers at a pH near pKBa H of the buffer.
Typically kobsd was measured at 5-7 buffer concentra-
tions for a given pH; the raw data are reported in Table
S1 of the Supporting Information.21 Plots of kobsd vs [B]
or [BH+] gave excellent straight lines; the rate constants
n-butylamine
2-methoxyethylamine
glycinamide
cyanomethylamine
methoxyamine
OH-
0.067
6.26
4.59 × 104
kB1 and kBH which are the main focus of this work were
PhCH2CH2NO2 (pKCH ) 8.55 ( 0.02)
-1
a
obtained from the slopes of these plots given by eqs 9
and 10, respectively, where KCa H is the acidity constant
of the carbon acid and KaBH the acidity constant of the
piperidine
11.24
9.94
9.34
8.78
10.87
9.64
5.50
4.70
3.24
0.854
0.292
0.129
6.61 × 10-3
3.48 × 10-2
4.73 × 10-2
7.62 × 10-2
2.25 × 10-3
5.64 × 10-3
3.90 × 10-1
4.23 × 10-1
piperazine
PZ-CH2CH2OHc
morpholine
n-butylamine
2-methoxyethylamine
cyanomethylamine
methoxyamine
OH-
4.70 × 10-1
6.94 × 10-2
3.47 × 10-4
5.97 × 10-5
aH
-1 KBa H
aH
KCa H
+
+
slope([B]) ) kB1 + kBH
) kB1 1 +
(9)
(
)
19.7
KBa H
1 aH
KCa H
slope([BH]) ) kB
+ k ) k 1 +
(10)
BH
BH
a
b
Ionic strength 0.5 M (KCl). Estimated experimental error
-1
-1
(
)
3-5%. c 1-(2-hydroxyethyl)piperazine.
aH
+
+
buffer acid.
Figure 1 shows plots of kobsd vs morpholine concentration
At pH J pKCa H the equilibrium was approached from
at various pH values while Figure 2 shows a plot of slope-
the carbon acid side, eq 9 was solved for kB1 using KaCH
(B) vs aH according to eq 9. The pKCa H value of 8.63
+
BH
-1
obtained from this plot is in excellent agreement with
the spectrophotometric value of 8.55.
determined spectrophotometrically, and k
was ob-
tained as kBH ) kB1 KaBH/KaCH. At pH j pKaCH the equilib-
-1
We also determined kO1 H by measuring kobsd as func-
tion of [KOH] in the absence of amine buffers where eq
rium was approached from the carbanion side, eq 10 was
BH
-1
solved for k
and kB1 was obtained as kB1 ) kBH KCa H
/
-1
8 simplifies to kobsd ) kO1 H aOH . All rate constants are
KBa H
. In the case of PhCH2CH2NO2 with morpholine,
-
pKCa H is close to pKBa H which allowed a kinetic determi-
nation of pKCa H from the pH-dependence of the slopes.
summarized in Table 1.
Kin etics in 90% DMSO-10% Wa ter . Rates were
measured for reactions with secondary alicyclic amines
only. The experimental methodology was the same as
in water. However, most of the plots of kobsd vs [B] or
(21) See paragraph concerning Supporting Information at the end
of this paper.

Carbanion Stabilization by Adjacent Sulfur
J . Org. Chem., Vol. 63, No. 6, 1998 1947
F igu r e 3. Plot of kobsd vs ammonium ion concentration for
the reaction of PhCH2CHdNO2 with morpholinium ion in
90% DMSO-10% water at 20 °C, pH 8.59.
F igu r e 4. Inversion plot according to eq 14 for the reaction
of PhCH2CHdNO2 with morpholinium ion in 90% DMSO-
10% water at 20 °C, pH 8.59.
-
-
Sch em e 1
(13.63) so that eq 12 simplifies to eq 13. Inversion plots
of 1/kobsd vs 1/[BH+] according to eq 14 afforded values
BH
-1
for Kassoc/kBH and (1+ aH /KNa OH))/k . A representative
+
-1
example is shown in Figure 4. In principle, a pH-
BH
-1
k
[BH+]
kobsd
)
(13)
(14)
aH
KNa OH
+
1 +
+ Kassoc[BH+]
[BH+] were curved; Figure 3 shows a representative
example. The curvature is reminiscent of similar obser-
vations reported for the reactions of nitromethane and
phenylnitromethane with amines in 90% DMSO-10%
water.22 It can be attributed to an association between
the nitronate ion and the protonated amine to form a
hydrogen bonded complex as shown in Scheme 1; note
that the scheme also includes the aci-form of the nitroal-
kanes because pKNa OH is much higher in 90% DMSO-
10% water than in water and played a significant role in
some of the kinetic runs with PhCH2CH2NO2.
aH
+
1 +
KNa OH Kassoc
1
)
+
BH
-1
BH
-1
kobsd
k
[BH+]
k
BH
-1
dependence of the (1 + aH /KNa OH))/k term would then
+
yield KNa OH and kBH separately from which Kassoc could
-1
finally be obtained. However, the relatively low value
of pKNa OH made this method unsuitable for an accurate
determination of this acidity constant and pKNa OH was
therefore obtained from kinetic data for the protonation
Based in Scheme 1, kobsd is given by eq 11; in this
solvent the intercepts of plots of kobsd vs [B] or [BH+] were
-
of PhCH2CHdNO2 in methoxyacetic acid buffers.
For these experiments eq 11 becomes eq 15 (no as-
sociation of the acid with the nitronate ion was observed)
k-1
kobsd ) k1 +
(11)
aH
KNa OH
+
1 +
+ Kassoc[BH+]
k-1
kobsd
)
(15)
aH
+
1 +
very small and usually negligible which means that kobsd
can be approximated by eq 12. In the case of PhCH2-
CH2NO2 all experiments were done at pH , pKaCH
KNa OH
and the reciprocal slopes of plots of kobsd vs methoxyacetic
acid concentration are given by eq 16. Such slopes were
determined at 10 different pH values; they afforded
BH
-1
k
[BH+]
kobsd ) kB1 [B] +
(12)
pKNa OH ) 8.55 ( 0.03 and kRCOOH ) (3.88 ( 0.10) × 103
aH
+
-1
1 +
+ Kassoc[BH+]
M-1 s-1
.
KNa OH
aH
RCOOH
+
1
1
)
+
(16)
(22) Bernasconi, C. F.; Kliner, D. A. V.; Mullin, A. S.; Ni, J .-X. J .
Org. Chem. 1988, 53, 3342.
RCOOH
slope[(RCOOH)]
k
k
-1
KNa OH
-1

1948 J . Org. Chem., Vol. 63, No. 6, 1998
Bernasconi and Kittredge
Ta ble 2. Su m m a r y of Ra te Con sta n ts, p KCa H, p KNa OH, a n d KAssoc Va lu es for th e Rea ction s of P h SCH2NO2 a n d
P h CH2CH2NO2 w ith Am in es in 90% DMSO-10% Wa ter a t 20 °Ca
BH
-1
pKBa H
kB1 (M-1 s-1
)
k
(M-1 s-1
)
Kassoc (M-1 s-1
)
b
b
c
B
PhSCH2NO2 (pKCH ) 10.65 ( 0.02; pKNa OH ) 6.79 ( 0.09)
10.74
10.23
9.53
a
piperidine
2.49 × 104
1.15 × 104
2.35 × 103
6.82 × 102
2.02 × 104
3.02 × 104
3.10 × 104
3.75 × 104
piperazine
PZ-CH2CH2OHd
morpholine
8.91
PhCH2CH2NO2 (pKCH ) 13.68 ( 0.02; pKNa OH ) 8.55 ( 0.03)
piperidine
10.74
10.23
9.53
a 2.88
1.84
0.447
0.117
2.52 × 104
5.17 × 103
6.31 × 103
6.86 × 103
50
64
82
62
piperazine
PZ-CH2CH2OHd
morpholine
8.91
a
b
Ionic strength 0.06 M (KCl). Estimated experimental error 8-12%; this error is significantly higher than in water because of the
curved buffer plots and additional uncertainty introduced by the KNa OH equilibrium in the reaction of PhCH2CH2NO2. c Estimated
d
experimental error 20 to 30%. 1-(2-hydroxyethyl)piperazine.
With PhSCH2NO2 all experiments with amine buffers
were performed at pH . pKNa OH so that eq 12 simplifies
to eq 17. Kassoc is relatively small and hence at low buffer
concentrations Kassoc[BH+] , 1. This allowed determi-
k
[BH+]
BH
-1
kobsd ) kB1 [B] +
(17)
1 + Kassoc[BH+]
nation of kB1 and kBH from the slopes according to eqs 9
-1
or 10. Kassoc was then found via eq 18 based on data at
high [BH+]. To determine pKNa OH, rates of protonation
Kassoc
1
1
)
+
(18)
BH
-1
BH
kobsd - kB1 [B]
k
[BH+]
k
-1
of PhSCHdNO2- in chloroacetate buffers were measured
and the data treated using eq 16; they yield
pKNa OH ) 6.79 ( 0.09 and kRCOOH ) (2.99 ( 0.19) × 103
-1
M-1 s-1
.
Discu ssion
F igu r e 5. Statistically corrected Brønsted plots for the
reactions in water at 25 °C. O: PhSCH2NO2 + R2NH; 4:
PhSCH2NO2 + RNH2; b: PhCH2CH2NO2 + R2NH; 2: PhCH2-
CH2NO2+ RNH2. The points of intersection between the
Brønsted lines and the dashed vertical line correspond to log
ko.
Rate constants and pKa values are summarized in
Table 1 for the reactions in water and in Table 2 for the
reactions in 90% DMSO-10% water; the latter includes
Kassoc values. Figures 5 and 6 show Brønsted plots while
the Brønsted âB values (dlog kB1 /dpKaBH)) and intrinsic
rate constants (ko ) kB/q ) k-BH1 /p at pKaBH - pKCa H + log-
1
9-PhS-fluorene vs 9-tert-butylfluorene although the effect
in these cases is less dramatic. Further attenuation of
the acidifying influence of the PhS group was found for
nitroalkanes but the effects are still substantial, e.g.,
PhSCH2NO2 in DMSO is 5.6 pK-units more acidic than
CH3NO2.
(p/q) ) 0)23 are reported in Table 3. Table 3 also includes
previously determined âB and log ko values for PhCH2-
NO2 and HOCH2CH2NO2.24
22
p Ka Va lu es. The acidifying effect of the R-phenylthio
group on carbon acids has been extensively documented
by Bordwell et al.12 For carbon acids whose conjugate
base has the negative charge substantially localized on
the R-carbon the effect of the phenylthio group is quite
dramatic. For example, in DMSO PhSCH2CN is 11.7 pK-
units more acidic that CH3CH2CN, and PhSCH2SO2Ph
is 10.5 pK-units more acidic than CH3CH2SO2Ph. Less
than half of the acidifying effect, which is larger than
that of an R-phenyl group, could be attributed to induc-
tive electron withdrawal. Similar conclusions were
reached with carbon acids that lead to more delocalized
anions such as PhSCH2COPh vs CH3CH2COPh and
Our own results show that in 90% DMSO-10% water
and in water the pKaCH of PhSCH2NO2 is about the same
as for PhCH2NO2, or 4.15 pK-units more acidic than CH3-
NO2 in 90% DMSO-10% water and 3.53 pK-units more
acidic than CH3NO2 in water (Table 3). The reduction
in the pK-difference upon making the solvent aqueous
is undoubtedly the result of stronger solvation of the
nitronate ion which diminishes the influence of the PhS
group.
The pKCa H of PhCH2CH2NO2 in water is about the
same as the pKa of CH3CH2NO2 (8.60),25 1.67 units lower
than that of CH3NO2 and 0.85 units lower than that of
(23) p and q are statistical factors: p is the number of equivalent
protons on BH+ while q is the number of equivalent basic sites on B.
(24) Bernasconi, C. F.; Panda, M.; Stronach, M. W. J . Am. Chem.
Soc. 1995, 117, 9206.
(25) Pearson, R. G.; Dillon, R. L. J . Am. Chem. Soc. 1953, 75, 2439.

Carbanion Stabilization by Adjacent Sulfur
J . Org. Chem., Vol. 63, No. 6, 1998 1949
unfavorable. The fact that in 90% DMSO the pK differ-
ence between PhCH2CH2NO2 and CH3NO2 is reduced
from 1.67 to 1.12 units is consistent with the notion that
hyperconjugation is at least a contributing factor: be-
cause of the large charge separation, 4b is expected to
be less stable in the less polar DMSO than in water.
The pKNa OH values of the aci forms of PhSCH2NO2 and
PhCH2CH2NO2 were only determined in 90% DMSO.
J ust as is the case with CH3NO2 and PhCH2NO2,
pKNa OH is several pK-units lower than the corresponding
pKCa H values but the difference, pKaCH - pKNa OH varies
considerably from compound to compound (Table 3).
There is decreasing trend in pKCa H - pKaNOH with in-
creasing acidity which seems to be a general pattern for
nitro compounds;28 the particularly low pKaCH - pKNa OH
for PhCH2NO2 has been attributed to the possibility of
resonance stabilization of the aci form.22
Based on pKNa OH ) 3.25 for CH3NO2 and 3.90 for
PhCH2NO2 in water,28 the pKaNOH values for PhSCH2-
NO2 and PhCH2CH2NO2 can be expected to be below 4.0
in this solvent. This is confirmed by the fact that the
rate constants for deprotonation of these compounds by
methoxyamine (pKBa H ) 4.70) obtained via eq 10 lie on
the Brønsted line defined by the other primary aliphatic
amines; if equilibrium protonation on the nitro group had
been significant in the kinetic runs with methoxyamine
buffers, use of eq 10 would have led to a depressed rate
constant.
F igu r e 6. Statistically corrected Brønsted plots for the
reactions in 90% DMSO-10% water at 20 °C. O: PhSCH2-
NO2 + R2NH; b: PhCH2CH2NO2 + R2NH. The points of
intersection between the Brønsted lines and the dashed
vertical line correspond to log ko.
HOCH2CH2NO2 (Table 3). The reason for the low
pKCa H of PhCH2CH2NO2 is probably the same as the
reason for the high acidity of CH3CH2NO2 and HOCH2-
CH2NO2 compared to CH3NO2, i.e., stabilization of the
nitronate ion by the CH3, PhCH2, and HOCH2 groups,
respectively, attached to the sp2 R-carbon. The classical
interpretation of this stabilization is in terms of hyper-
conjugation18,26 (4b) although an alternative explanation
in terms of a strengthening of the bond between the RCH2
In tr in sic Ra te Con sta n ts. The log ko value for
PhSCH2NO2 is much larger than for all the other ni-
troalkanes listed in Table 3. Based on the reasoning
outlined in the Introduction, this result implies that the
-
extra stabilization of PhSCHdNO2 compared to say
-
-
CH2dNO2 or PhCH2CHdNO2 must be mainly the
result of a polarizability effect of the phenylthio group
rather than due to d-p π-bonding (eq 1) or negative
hyperconjugation (eq 2). This is the main qualitative
conclusion of our study. This conclusion is particularly
interesting in view of Wiberg and Castejon’s15 recent ab
initio study and of Cuevas and J uaristi’s DFT study16
which led them to regard negative hyperconjugation as
the main stabilizing factor exerted by an R-alkylthio
group.
carbon and the R-carbon upon rehybridization of the
R-carbon from sp3 in the acid to sp2 in the anion (4a ) is
possible. The bond strength argument appears now to
be the preferred explanation for the stabilization of
alkenes by alkyl groups27 but this does not necessarily
exclude hyperconjugation as a contributing factor in the
stabilization of nitronate ions. This is because the
hyperconjugative structure (4b) is more akin to that of a
carbocation (5b) where charge is merely delocalized
whereas the hyperconjugative structure of an alkyl-
It is not clear at this point why there is disagreement
between our conclusions and those reached by Wiberg
et al.15 and J uaristi et al.16 That negative hyperconju-
gation plays a significant role can be based on stereo-
electronic9,16 as well as bond lengths15 arguments but this
does not necessarily mean that it is the dominant factor.
In other words, even though the calculated geometric
parameters are consistent with negative hyperconjuga-
tion, the actual stabilization of the carbanions derived
from the hyperconjugation may be smaller than that
derived from the polarizability effect. Another view has
been offered by a referee. According to this referee,
negative hyperconjugation should only operate on pyra-
midal but not on planar carbons. This means that
negative hyperconjugation, just like polarizability, should
mainly stabilize the transition state and lead to an
increase rather than decrease in log ko. This view
appears to be based on the fact that some of the studies
showing the stabilizing effect of sulfur involved pyrami-
substituted alkene (6b) leads to the creation and separa-
tion of two opposite charges which is energetically
(26) Streitwieser, A., J r.; Hammons, J . H. Prog. Phys. Org. Chem.
1965, 3, 43.
(27) Streitwieser, A.; Heathcock, C. H.; Kosower, E. M. Introduction
to Organic Chemistry, 4th ed.; Macmillan: New York, 1992; p 258. (b)
McMurry, J . Organic Chemistry, 4th ed.; Brooks/Cole: Pacific Grove,
CA, 1996; p 197. (c) J ones, M., J r. Organic Chemistry; Norton and
Co.: New York, 1997; p 134.
(28) Nielsen, A. T. In The Chemistry of the Nitro and Nitroso Groups;
Feuer, H., Ed.; Wiley: New York, 1969; Part 1, p 349.

1950 J . Org. Chem., Vol. 63, No. 6, 1998
Bernasconi and Kittredge
Ta ble 3. Br øn sted âB Va lu es a n d In tr in sic Ra te Con sta n ts for th e Rea ction s of Nitr oa lk a n es w ith Am in es
a
b
b
PhSCH2NO2
6.67
PhCH2CH2NO2
HOCH2CH2NO2
9.40
PhCH2NO2
6.88
CH3NO2
Water, 25 °C
pKCa H
8.55
10.22 (10.40)e
âB(R2NH)c
0.52 ( 0.04
1.02 ( 0.15
0.63 ( 0.02
-0.13 ( 0.05
0.56 ( 0.01
-1.16 ( 0.02
0.61 ( 0.03
0.62
-0.59
0.57
0.48
-0.86
0.64
log ko(R2NH)c
âB(RNH2)d
-0.19 (-0.25)e
log ko(RNH2)d
-2.06 ( 0.07
-1.30
90% DMSO-10% Water, 20 °C
pKCa H
10.65
6.79
3.86
13.68
8.55
5.13
10.68
7.73
2.95
14.80 (14.98)e
8.65
pKNa OH
pKCa H - pKNa OH
6.15
âB(R2NH)c
log ko(R2NH)c
f,g
0.85 ( 0.03
4.08 ( 0.02
3.06
0.76 ( 0.06
2.51 ( 0.23
3.79
0.69
1.75
0.69
3.06 (3.01)e
log(kDo /kWo
W∆DpKCa H
)
f,h
2.61
3.80
3.26
4.58
3.89
5.13
a
b
d
Reference 24. Reference 22. c Secondary alicyclic amines. Primary aliphatic amines. e Statistically corrected to make CH3NO2
CH
comparable to nitroalkanes with only two R-protons. f D ) 90% DMSO, W ) water. Ratio refers to ko for R2NH. W∆DpKa
)
g
h
pKCa H(D) - pKCa H(W).
Ta ble 4. Acid ities a n d In tr in sic Ra te Con sta n ts of XCH2NO2 Rela tive to CH3NO2. Dissection of Rela tive In tr in sic Ra te
Con sta n ts in to Con tr ibu tion s by Reson a n ce, In d u ctive, Hyp er con ju ga tion , a n d P ola r iza bility Effects
PhSCH2NO2
PhCH2CH2NO2
HOCH2CH2NO2
PhCH2NO2
Water
1.85
∆log KXa CH
2NO2
3.73
1.27
1.00
3.52
-0.61
∆log kXo CH
δlog kRo
(obsd)a
2NO2
-0.91
-0.34
<-0.61
>0 (small)
δlog kIo
δlog kho
>0
≈0 (small)
≈-0.91
>0 (small)
>0 (small)
+
-
δlog kdoR + δlog kho
e0
>0
δlog kPo
90% DMSO
1.30
∆log KXa CH
2NO2
4.33
1.07
4.30
∆log kXo CH
δlog kRo
(obsd)
2NO2
-0.50
-1.26
-1.25
δlog kIo
δlog kho
>0
<0 (small)
>-0.50
>0 (small)
+
-
δlog kdoR + δlog kho
e0
>0
δlog kPo
Based on statistically corrected pKCa H and log ko for CH3NO2.
a
dal carbons in conformationally rigid systems.4,9,16 How-
ever, it contradicts the premise, elaborated upon in the
Introduction, that negative hyperconjugation is akin to
a resonance effect as described by eq 2 and, by its very
nature, leads to a partial planarization of the R-carbon.15
We therefore respectfully disagree with this referee.
from carbon acids in situations where the rates are not
significantly affected by steric effects.17c,29,30
An a lysis of Ma in F a ctor s Affectin g th e In tr in sic
Ra te Con sta n ts. We choose CH3NO2 as reference
compound. This allows us to define ∆log kXo CH NO by
2
2
eq 19 and ∆log KXa CH NO by eq 20. The observed
2
2
∆log kXo CH NO and ∆log KXa CH NO values are summarized
2
2
2
2
One question that we have not yet addressed is how
one may assess the magnitude of the polarizability effect
on log ko. This is not a trivial task because there is a
complex interplay of other factors that affect the intrinsic
rate constants of proton transfers, as can be seen from
the large differences in the log ko values among the four
nitroalkanes used in the comparison with PhSCH2NO2.
in Table 4; they are based on statistically corrected
∆log koXCH NO ) log kXo CH NO - log koCH NO (19)
2
2
2
2
3
2
∆log KXa CH NO ) log KaXCH NO - log KCa H NO
)
2
2
2
2
3
2
pKCa H NO - pKXa CH NO (20)
3
2
2
2
In the following section these factors will be analyzed
in some detail. This analysis will focus on the reactions
with secondary alicyclic amines because there is a more
complete set of data available. However comparable
conclusions would be reached on the basis of the reactions
with the primary amines. Incidentally, the intrinsic rate
constants for these latter reactions (log ko(RNH2)) are
substantially lower than log ko(R2NH) for the secondary
alicyclic amines, a common finding for proton transfers
pKCa H and log ko values for CH3NO2.32 The relevant
factors that can affect ∆log kXo CH NO include p-p π
resonance (R), inductive effects (I), positive hyperconju-
gation and/or bond strength effects elaborated upon in
the section on pKa values (h+),33 negative hyperconjuga-
tion (h-), d-p π resonance (dR), and polarizability (P).
This may be expressed by eq 21 but it should be noted
2
2

Carbanion Stabilization by Adjacent Sulfur
J . Org. Chem., Vol. 63, No. 6, 1998 1951
that typically only a few terms in eq 21 are simulta-
neously operative in any given case. Steric effects are
probably of minor importance in this study34 and have
not been included.
are the main factors24 so that eq 21 takes on the form of
eq 23. The inductive effect of the HOCH2 group is
slightly electron withdrawing, although less than that
+
∆log koHOCH CH NO ) δlog koI + δlog koh
(23)
2
2
2
+
∆log kXo CH NO ) δlog koR + δlog koI + δlog kho
+
2
2
h-
of the phenyl group in PhCH2NO2.37 Hence δlog koI
should be slightly positive. The net decrease in log ko
δlog k + δlog kdoR + δlog kPo (21)
o
compared to that for CH3NO2 (∆log kXCH NO ) -0.34 in
2
2
o
We now analyze the relevant contributions to
+
water) is thus the result of a negative δlog koh
. The
∆log kXo CH NO for the four compounds listed in Table 4.
2
2
+
negative value of δlog koh arises from the fact that the
transition state benefits very little from hyperconjugation
and/or the bond strength effect because sp3 f sp2
hybridization of the R-carbon has made little progress.40
P h CH2CH2NO2. The situation with this compound is
similar to that with HOCH2CH2NO2, i.e., eq 24 applies.
P h CH 2NO2. This is
a
familiar case20,22 but
∆log koPhCH NO has not been analyzed by means of eq 21.
The factors to consider here are p-p π resonance and
the electron-withdrawing inductive effect of the phenyl
group, i.e., eq 21 reduces to eq 22.
2
2
However, here δlog kI is very small because the PhCH2
∆log koPhCH NO ) δlog koR + δlog koI
(22)
2
2
o
+
∆log koPhCH CH NO ) δlog koI + δlog koh
(24)
2
2
2
As has been amply documented, resonance in the
product ion lowers the intrinsic rate constant.17 This is
a consequence of the imbalance which prevents the
resonance effect from stabilizing the transition state in
proportion to the degree of proton transfer. This means
δlog kRo < 0.35
group has a very small inductive effect; in fact, depending
on which substituent constant is taken as a measure of
the inductive effect, it may either be considered slightly
electron withdrawing or electron donating.37 We shall
assume that δlog kIo ≈ 0 and hence ∆log koPhCH CH NO
≈
2
2
2
The electron-withdrawing inductive effect of the phenyl
+
.
δlog kho
This is consistent with our finding that
group is expected to increase ko, i.e., δlog kI > 0 because,
o
δlog kXo CH NO for PhCH2CH2NO2 is more negative than
for HOCH2CH2NO2.
2
2
due to the imbalance, the negative charge at the transi-
tion state is mainly localized on the R-carbon (δC in eq
3), i.e., close to the phenyl group. This allows a particu-
larly effective stabilization of the transition state by this
group. In contrast, because in the nitronate ion the
charge is mainly delocalized into the nitro group and far
away from the phenyl group, stabilization of the nitronate
ion by this latter group is relatively ineffective.17,36
P h SCH2NO2. This compound represents the most
complex case because, apart from the inductive effect, it
is not clear a priori which of the potential factors
associated with the phenylthio group, such as d-p π
resonance, negative hyperconjugation, and polarizability,
are important. If all factors were to contribute, eq 25
-
Since the observed ∆log kXo CH NO for PhCH2NO2 is
negative (-0.61 in water and -1.26 in 90% DMSO) it is
clear that the resonance effect is dominant.
2
2
would apply. In eq 25 δlog kdoR and δlog koh are expected
to be negative while δlog kIo and δlog kPo are expected to
-
∆log koPhSCH NO ) δlog kIo + δlog kodR + δlog kho
+
HOCH2CH2NO2. For this compound the inductive
effect and positive hyperconjugation/bond strength effects
2
2
δlog kPo (25)
(29) (a) Bernasconi, C. F.; Hibdon, S. A. J . Am. Chem. Soc. 1983,
105, 4343. (b) Bernasconi, C. F.; Bunnell, R. D. Isr. J . Chem. 1985,
26, 420. (c) Bernasconi, C. F.; Paschalis, P. J . Am. Chem. Soc. 1986,
108, 2969. (d) Bernasconi, C. F.; Terrier, F. J . Am. Chem. Soc. 1987,
109, 7115. (e) Bernasconi, C. F.; Stronach, M. W. J . Am. Chem. Soc.
1990, 112, 8448.
(30) The higher ko values for secondary amines is a well-known
phenomenon caused by differences in the solvation energies of the
respective protonated amines and the fact that at the transition state
solvation of the incipient protonated amine lags behind proton
transfer.17c,31
be positive. Since the experimental ∆log kXo CH NO is
strongly positive (1.27 in water and 1.07 in 90% DMSO)
the latter two terms are clearly dominant. However, it
is difficult to evaluate the relative magnitude of
δlog kIo and δlog kPo . From the analysis of ∆log
2
2
kXo CH NO for PhCH2NO2, HOCH2NO2, and PhCH2CH2-
2
2
NO2, it appears that δlog kI is typically quite small, as
o
pointed out previously.17a,b For PhSCH2NO2 one expects
a somewhat larger δlog koI term because the inductive
effect of the PhS group is larger than that of the other X
(31) (a) Bell, R. P. The Proton in Chemistry, 2nd ed.; Cornell
University: Ithaca, New York, 1973; Chapter 10. (b) J encks, W. P.
Catalysis in Chemistry and Enzymology; McGraw-Hill: New York,
1968; p 178.
(32) Statistically corrected for the fact that CH3NO2 has three while
XCH2NO2 has only two acidic protons.
(37) For example, the σI values for Ph, HOCH2, PhCH2, and PhS
are 0.12, 0.05, 0.03, and 0.30, respectively,38 while the F values are
0.12, 0.03, -0.04, and 0.30, respectively.39
(33) h+ is symbolizing positive hyperconjugation; this symbol is used
for simplicity and is meant to include bond strength effects.
(34) The previously mentioned fact that log ko(R2NH) is significantly
larger than log ko(RNH2) is a clear indication that steric effects are
not important.29
(38) Exner, O. In Correlation Analysis in Chemistry; Chapman, N.
B., Shorter, J ., Eds.; Plenum Press: New York, 1978; p 439.
(39) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165.
(40) A referee has suggested that “important solvation changes
between neutral and anion are bound to complicate analysis”. In view
of the fact that the OH group does not get ionized, we do not believe
this to be an important issue. However, as has been discussed in ref
24, intramolecular hydrogen bonding between the OH and NO2 group
(35) It may be useful to remind ourselves that it is the intrinsic rate
constant, not the actual rate constant which is lowered; the latter is
actually somewhat enhanced.22 It is the fact that the experimental
rate constant is not enhanced as much as one would expect if the
resonance effect were developed in proportion to the degree of proton
transfers which leads to the reduction in the intrinsic rate constant.
We add this clarification in response to a referee’s comment.
(36) Incidentally, this is also the reason in the deprotonation of
substituted phenylnitromethanes by amines the RCH Brønsted value
(variation of the phenyl substituent) is much larger than the Brønsted
âB value (variation of the amine).20
-
in HOCH2CHdNO2 might possibly play a role in stabilizing the
nitronate ion. This effect would lag behind proton transfer at the
a similar way as hyperconjugation and bond
strength effects and hence would have a similar effect on log ko. It
transition state in
+
can therefore be regarded as included in the δlog kho term.

1952 J . Org. Chem., Vol. 63, No. 6, 1998
Bernasconi and Kittredge
groups.37 Based on Bordwell’s7 analysis of the relative
contribution of the inductive effect to the acidifying effect
proton transfer at the transition state coupled with the
fact that nitronate ions are much less solvated by DMSO
than by water. If one regards the W∆DpKaCH values
discussed above as an approximate measure of the
strength of solvation of the nitronate ion by water, there
is indeed a rough correlation between log(kDo /k0W) and
W∆DpKCa H, consistent with the above notion about the
of the R-phenylthio group in
a number of carbon
acids, we estimate that close to a third or half of the
∆log koXCH NO value could be due to the δlog kIo term.
This implies that the δlog kPo term is at least half to
2
2
two-thirds of the ∆log kXo CH NO value if δlog kodR and
2
2
-
δlog kho are negligible, and larger than that if the
origin of the solvent effect.
Associa tion betw een Nitr on a te Ion s a n d R 2NH2
-
+
δlog kdoR and/or δlog kho terms are significant.
.
In 90% DMSO-10% water there is measurable associa-
tion between the ammonium ions and the nitronate ions
(Scheme 1). The association is strong enough with
A comment regarding the circumstances under which
the polarizability effect enhances ko is in order. As
outlined in the Introduction this enhancement is a
consequence of the accumulation of negative charge on
the R-carbon of the imbalanced transition state which
allows a disproportionately strong stabilization of the
transition state by the adjacent sulfur. This is similar
to the explanation of why electron-withdrawing inductive
effects enhance ko. However, there are some important
differences between the two effects that result from the
fact that polarizability effects are proportional to the
square of the charge ((δC)2 in eq 3) and drop off very
steeply with the fourth power of distance, whereas
inductive effects are simply proportional to the charge
and drop off only with the square of the distance.41 With
respect to the inductive effect, this means that, as long
as the transition state is imbalanced in the sense
that charge delocalization lags behind proton transfer,
δlog kIo will always be positive. In contrast, as discussed
in more detail elsewhere,17b δlog kPo is positive only if the
imbalance is substantial and proton transfer at the
transition state has made significant progress. Both
conditions are amply met in the deprotonation of
nitroalkanes.17b
-
PhCH2CHdNO2 to allow a determination of Kassoc for
+
all R2NH2 although the Kassoc values suffer from large
experimental uncertainty (Table 2). In the case of
PhSCHdNO2- the association is weaker and only allows
crude Kassoc values to be obtained with morpholinium and
1-(2-hydroxyethyl)piperazinium ion.
The dependence of Kassoc on R2NH2+ and nitronate ion
appears to follow the expected trend for hydrogen bonded
complexes, i.e., Kassoc increases with increasing oxygen
basicity of the nitronate ion (pKNa OH) and increasing
acidity of R2NH2+ although the latter trend is quite weak,
perhaps on account of the large experimental uncertain-
ties. The same dependencies were observed for Kassoc in
-
- 22
the case of CH2dNO2 and PhCHdNO2
.
Con clu sion s
The main objective of the present study has been to
identify the principal interaction mechanism that can
explain the strong acidifying effect of the thiophenoxy
group on PhSCH2NO2. Apart from an inductive effect,
the possible interaction mechanisms include d-p π-res-
onance, negative hyperconjugation, and polarizability. On
the basis of our findings that the intrinsic rate constant
(ko) for the deprotonation of PhSCH2NO2 by amines is
substantially higher than for the deprotonation of other
nitroalkanes, it was concluded that the dominant inter-
action mechanism is the polarizability of the sulfur group.
This conclusion follows from an analysis which shows
that, because of the imbalanced nature of the transition
state its stabilization by the thiophenoxy group is dis-
proportionately strong compared to the stabilization of
the nitronate ion; this is what causes the increase in ko.
This contrasts with the expectation that d-p π-resonance
or negative hyperconjugation should depress ko.
Solven t Effect on p KCa H a n d log ko. All nitro
compounds in Table 3 show the typical increase in
pKCa H as the solvent is changed from water to 90%
DMSO, reflecting mainly the loss of hydrogen bonding
solvation of the nitronate ion by water. The magnitude
of the solvent effect on pKCa H varies somewhat depend-
ing on the X group in XCH2NO2 (see W∆DpKaCH in Table
3). For PhSCH2NO2 and PhCH2NO2 W∆DpKaCH is
smaller than for CH3NO2; apparently part of the loss in
solvation of the respective nitronate ion in 90% DMSO
is offset by enhanced inductive and polarizability effects
in the former case, and enhanced inductive and resonance
effects in the latter. For PhCH2CH2NO2 W∆DpKaCH is
larger than for CH3NO2; in this case the extra stabiliza-
tion of the anion by positive hyperconjugation is less
effective in 90% DMSO than in water which, as men-
tioned earlier, must be the result of destabilization of the
charge separated structure 4b in DMSO.
Exp er im en ta l Section
1H NMR spectra were recorded on a Bruker 250-MHz
spectrometer. Kinetic experiments were carried out on a
Durrum-Gibson stopped-flow spectrophotometer equipped with
OLIS computerized software. UV-vis spectra were obtained
on a Perkin-Elmer Lambda 2 or Hewlett-Packard 8562 diode
array spectrophotometer. pH measurements were made on
an Orion 611 pH meter.
Ma ter ia ls. Water was taken from a Milli-Q water purifica-
tion system. Piperidine, 1-(2-hydroxyethyl)piperazine, pip-
erazine, morpholine, 2-methoxyethylamine, and n-butylamine
were refluxed over CaH2 and freshly distilled prior to use.
Methoxyacetic acid was distilled prior to use. Chloroacetic
acid, cyanomethylamine, glycinamide, and methoxyamine
were recrystallized from ethanol. KOH and HCl solutions
were prepared from “dilut-it” from Baker Analytical. DMSO
was distilled from CaH2 and stored over 4 Å molecular sieves.
1-Nitro-2-phenylethane is a yellow oil that was synthesized
by the procedure of Varma and Kabalka44 by reducing â-ni-
As to the intrinsic rate constants, they are all consider-
ably larger in 90% DMSO than in water, with log
(kDo /kWo )42 values varying between 2.61 and 3.79 (Table
3). As discussed in detail elsewhere,17c,43 there are
several factors that contribute to the solvent effect on ko
for deprotonation of carbon acids. The main one is the
lag in the solvation of the incipient nitronate ion behind
(41) Taft, R. W.; Topsom, R. D. Prog. Phys. Org. Chem. 1987, 16, 1.
(42) The log(kDo /kWo ) in Table 3 are for the reactions with secondary
alicyclic amines, with kDo referring to 90% DMSO. Note that if the
temperatures in the two solvents were the same, the log(kDo /koW
)
values would be slightly higher.
(43) Gandler, J . R.; Bernasconi, C. F. J . Am. Chem. Soc. 1992, 114,
631.

Carbanion Stabilization by Adjacent Sulfur
J . Org. Chem., Vol. 63, No. 6, 1998 1953
trostyrene with NaBH4. 1H NMR (250 MHz, CDCl3): δ3.31
(2H, t, phenyl-CH2), 4.60 (2H, t, CH2NO2), 7.35 (5H, m, ArH).
(Phenylthio)nitromethane is a clear oil that was synthesized
by the procedure of Barrett et al.45 1H NMR (250 MHz,
CDCl3): δ5.44 (2H, s, CH2), 7.35 (5H, m, ArH); lit.45 5.45 (2H,
s, CH2), 7.25-7.5 (5H, m, ArH).
electrode. The solutions of 90% Me2SO-10% water were
prepared by adding 10 mL of aqueous solution to a 100-mL
volumetric flask and topping off with Me2SO. The reaction
mixtures were adjusted to have an ionic strength of 0.06 M
by addition of KCl. pH measurements were made with an
Orion Ross electrode and an Orion 80-03 Sure-Flow reference
electrode. For 90% Me2SO-10% water, the pH meter was
calibrated according to Halle´ et al.46 The pKa values of the
carbon acids were determined in chloroacetate (for PhSCH2-
NO2) and methoxyacetate buffer (for PhCH2CH2NO2) by clas-
sical spectrophotometric procedures at the λmax of the anion
of the substrate.
Kin etic Ru n s a n d Sp ectr a . The kinetic experiments were
run under pseudo-first-order conditions with the substrate as
the minor component. The reactions were monitored spectro-
photometrically at the λmax of the anion of the substrate. For
reactions at pH < pKCa H, the carbon acid was placed in a 10-3
M KOH solution and then mixed in the stopped-flow apparatus
with the respective amine buffer containing 10-3 M HCl.
Rea ction Solu tion s: p H a n d p Ka Mea su r em en ts. The
pKBa H values of the amines and pKAa H of the acids were
determined by measuring the pH of various buffer ratios and
plotting log([B]/[BH+]) vs pH according to the Henderson-
Hasselbach equation pH ) pKa + log([B]/[BH+]) where the
intercept is the pKa and the slope is unity. The pH of reaction
solutions for stopped-flow runs was measured in mock-mixing
experiments that simulated the stopped-flow runs. Water
solutions were prepared at an ionic strength of 0.50 M by
addition of KCl, and pH measurements were performed with
an Orion Ross electrode and an Orion no. 800500 reference
Ack n ow led gm en t. This research has been sup-
ported by Grant No. CHE-9307659 from the National
Science Foundation.
Su p p or tin g In for m a tion Ava ila ble: Table S1-S29, ki-
netic data (16 pages). This material is contained in libraries
on microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
J O9719463
(44) Varma, R. S.; Kabalka, G. W. Synth. Commun. 1985, 15, 151.
(45) Barrett, A. G. M.; Dhanak, D.; Graboski, G. G.; Taylor, S. J .
Org. Synth. 1990, 68, 8.
(46) Halle´, J . C.; Gaboriaud, R.; Schaal, R. Bull. Soc. Chim. Fr. 1970,
2047.