Hammett analyses of halocarbene-halocarbanion equilibria

Article abstract of DOI:10.1021/ol400698y

Substituted arylchlorocarbenes (X = H, p-Cl, p-CF₃, p-F, m-Cl) reacted reversibly with Cl^- in dichloroethane to form the corresponding aryldichloromethide carbanions. Equilibrium constants and rate constants for the forward and rever

Full text of DOI:10.1021/ol400698y

ORGANIC
LETTERS
2013
Vol. 15, No. 8
2014–2017
Hammett Analyses of
HalocarbeneꢀHalocarbanion Equilibria
Lei Wang, Robert A. Moss,* and Karsten Krogh-Jespersen*
Department of Chemistry and Chemical Biology, Rutgers,
The State University of New Jersey, New Brunswick, New Jersey 08903, United States
moss@rutchem.rutgers.edu; krogh@rutchem.rutgers.edu
Received March 15, 2013
ABSTRACT
Substituted arylchlorocarbenes (X = H, p-Cl, p-CF3, p-F, m-Cl) reacted reversibly with Clꢀ in dichloroethane to form the corresponding
aryldichloromethide carbanions. Equilibrium constants and rate constants for the forward and reverse reactions were correlated by the Hammett
equation. DFT methods were used to compute equilibrium constants and electronic absorption spectra.
We reported a Hammett analysis of equilibrium con-
stants for the formation of π-complexes between p-sub-
stituted arylchlorocarbenes and 1,3,5-trimethoxybenzene
in pentane.1 More recently, we spectroscopically charac-
terized equilibria between phenylhalocarbenes, halide
ions, and phenyldihalomethide carbanions; eq 1 (X = Cl
or Br). At 298 K, equilibrium constants were 4.01 or 3.01
Carbenes 1aꢀ1e were generated by laser flash photolysis
(LFP)4 of the corresponding diazirines in 1,2-dichloroethane
(DCE) at temperatures ranging from 294 to 304 K. The
diazirines were prepared by standard Graham oxidations of
available5 or commercial amidines. In the presence of added
tetrabutylammonium chloride (TBACl), LFP of the diazir-
ines gave carbanions 2aꢀ2e, in addition to the carbenes.2 The
experimental absorption spectra were calibrated, and the
corresponding spectra appear in the Supporting Information
(SI).6 The experimental and principal computed absorption
maxima of the carbenes and the carbanions are included
in Table 1.7
M
ꢀ1 for X = Cl or Br, respectively.2
sK
PhCCX þ Xꢀ
PhCX2
ð1Þ
F
s
R
Herein we present a detailed Hammett study of equili-
bria between arylchlorocarbenes (1), chloride, and aryldi-
chloromethide carbanions (2), eq 2.
Equilibrium constants for eq 2 were determined in DCE
as described for 1a/2a in ref 2. Consider, for example, the
p-Cl case, 1b/2b. The ratio of the 1b absorbance at 324 nm
to the 2b absorbance at 388 nm varied with [TBACl]: A324/
A388 decreased with the increasing concentration of
TBACl.8
(4) We used a xenon fluoride excimer laser operating at 351 nm,
emitting 42ꢀ56 ns pulses with 60ꢀ70 mJ power.
(5) Moss, R. A.; Terpinski, J.; Cox, D. P.; Denney, D. Z.; Krogh-
Jespersen, K. J. Am. Chem. Soc. 1985, 107, 2743. Graham, W. H. J. Am.
Chem. Soc. 1965, 87, 4396.
(6) Calibration corrects the intensities of the observed UVꢀvis
absorptions for wavelength-dependent variations in sample absorptiv-
ity, xenon monitoring lamp emission, and detector sensitivity from 244
to 800 nm. Calibration curves appear in the SI.
(7) wB97XD/6-311þG(d) was used for calculation of ΔG and K in
simulated dichloroethane. Excited state calculations (TD-B3LYP/6-
311þG(d)//wB97XD/6-311þG(d)) in simulated dichloroethane pro-
duced absorption wavelengths and oscillator strengths (f). See SI for
computational details.
We report σ/F correlations for Keq, kf, and kb of eq 2 and
forcomputedvaluesofKeq. The resultsrevealthesenseand
magnitude of the substituent electronic effects that operate
in these equilibria. A previous study of kf for eq 2 gave
F = þ0.86 for the capture of Clꢀ by ArCCl.3
(1) Wang, L.; Moss, R. A.; Thompson, J.; Krogh-Jespersen, K. Org.
Lett. 2011, 13, 1198.
(2) Wang, L.; Moss, R. A.; Krogh-Jespersen, K. J. Am. Chem. Soc.
2012, 134, 17459.
(3) Moss, R. A.; Tian, J. Org. Lett. 2006, 8, 1245.
(8) Each ratio remained constant from 150 to 200 ns after the laser
flash (200ꢀ300 ns for 1c/2c).
r
10.1021/ol400698y
Published on Web 04/08/2013
2013 American Chemical Society

Table 1. UV Absorptions of Carbenes 1 and Carbanions 2a
Table 2. Equilibrium and Kinetic Data for eq 2
b
b
c
c
c
X
λobsd
λcalcd
fb
λobsd
λcalcd
fc
X
Kexp (Mꢀ1 a
)
Kcalcd (Mꢀ1 a
)
10ꢀ8kfb
10ꢀ7kb
H
292
324
324
316
308
291
310
288
293
296
0.4770
0.5454
0.5254
0.4874
0.3880
404
388
388
388
388
393
396
397
388
393
0.1961
0.2421
0.2570
0.1776
0.2139
H
4.11
31.3
242
1.51 ꢁ 10ꢀ1
5.00
8.76 ꢁ 104
6.91 ꢁ 10ꢀ2
1.08 ꢁ 102
1.97
3.16
5.19
1.02
3.85
4.91
p-Cl
p-CF3
p-F
p-Cl
p-CF3
p-F
1.01
0.215
3.36
3.04
55.7
m-Cl
m-Cl
0.692
a Reported in nm in DCE. Experimental absorbances are calibrated.
See ref 7 for computational details. b Principal carbene absorptions.
Only the strong π f p absorption is listed; weak σ f p carbene
absorptions were observed at ∼600 nm. c Principal carbanion
absorptions.
a In DCE. b Rate constants in Mꢀ1 sꢀ1. See SI for temperatures and
errors. c In sꢀ1
.
of Table 2, the computed values for eq 2 with X = p-NO2,
p-Me, and p-OMe were included in the regression anal-
ysis.14 Hammett substituent constants were taken from
Smith and March.15
First consider Kexp: inclusion of K for X = F (1d/2d) is
deleterious to the correlation. Omitting Kexp for p-F
leads to an excellent four-point correlation of log K vs
σp, with F = þ3.26 (r = 0.995); cf. Figure S-41. Inclusion
of the p-F point causes significant decay in the fitting: log
K vs σp now gives F = þ3.67 and r is reduced to 0.937; cf.
Figure S-42. The positive values obtained for F imply
that carbanion formation is favored by electron with-
drawing substituents X, which destabilize carbene 1 but
Values of this ratio were determined over the 150ꢀ200
ns time interval as [TBACl] varied from 0.25 to 0.49 M; see
the SI for the resulting LFP spectra. BeerꢀLambert anal-
ysis of the data used computed oscillator strengths (f; see
Table 1)7 in place of the unknown extinction coefficients
of 1b and 2b.9,10 A plot of A324/A388 vs 1/[TBACl] gave a
linear correlation (Figure S-17 in the SI), whose slope
(0.072) led to K = 31.3 ((1.7) Mꢀ1 for eq 2 when
X = p-Cl.11
Values of K were similarly determined for the other
carbene/carbanion pairs; data for the unsubstituted 1a/2a
pair were taken from ref 2. Values of observed (Kexp) and
calculated (Kcalcd) equilibrium constants7 are collected in
Table 2. These values are specific to DCE solution and
ignore the possible aggregation of TBACl or the TBA salts
of the carbanions.12 We attempted to measure K for eq 2
with X = p-Me or p-NO2. However, with p-Me, K was too
small and the carbanion absorbance (at 404 nm) was too
weak for reliable measurement. With p-NO2, K was too
large and the carbene absorption (at 324 nm) could not be
accurately measured; the carbanion appeared at 500 nm.
The kinetics of the forward process of eq 2 were followed
by LFP, monitoring the apparent rate of formation of each
carbanion as a function of [TBACl]. Rate constants (kf)
were obtained from the slopes of these correlations; see the
SI.13 Values of kb for the reverse process were obtained
from the experimental data for K and kf, according to
kb = kf/K. The rate constants are collected in Table 2.
Hammett analyses were performed for Kexp, Kcalcd, kf,
and kb of Table 2. For Kcalcd, in addition to the five cases
stabilize carbanion 2. The p-F substituent has σp
=
þ0.15, indicating an electron withdrawing group that
should enhance carbanion formation relative to H (σp = 0).
However, the computed Kcalcd values in Table 2 indi-
cate that p-F suppresses carbanion formation relative
to H; Kcalcd(X = F)/Kcalcd(X = H) = 0.46. In other
words, p-F acts as a net donor of electron density in
eq 2; electron donation from a lone pair on F by
resonance outweighs electron withdrawal by induc-
tion. Given that p-F directly interacts with the carbene
or carbanion center via the phenyl ring, σpþ (ꢀ0.07) is a
possible alternative for σp (þ0.15) to represent p-F in a
Hammett analysis of eq 2.
Indeed, an excþellent correlation of log Kexp vs σ is
obtained using σp for F (and σm for m-Cl); cf. Figure 1
where F = þ3.18 (r = 0.996). Note, however, that using
þ
σp also for p-Cl leads to a decreased quality in the fit;
F = þ3.01, þbut r decreases to 0.959, Figure S-43.16
Using σp for p-F, rather than σp, is one possible
response to the computational results indicating that the
p-F substituent behaves as a net donor of electron density
(9) Moss, R. A.; Wang, L.; Odorisio, C. M.; Krogh-Jespersen, K.
J. Am. Chem. Soc. 2010, 132, 10677.
þ
in eq 2. The use of σp is similar to using σm for m-Cl.
(10) Formally, the computed oscillator strengths are proportional to
the frequency-integrated absorption coefficients. Although the widths of
the carbene and carbanion absorption bands differ, we did not correct
for this. We conservatively estimate that the fcarbene/fcarbanion ratio
accurately reproduces the desired εcarbene/εcarbanion extinction constant
ratio to within a factor of 2.
However, if a “pure” approach is preferred, one can simply
accept the correlation þwith σp for p-F. The difference in F
values (þ3.18 with σp vs þ3.67 with σp) is neither very
large nor mechanistically significant. In all the Hammett
(11) K = (slope)ꢀ1(f2/f1), where f2 is the computed oscillator strength
of the carbene and f1 is the analogous factor for the carbanion. For 1b,
f2 = 0.5454 at 324 nm and f1 = 0.2421 at 388 nm (cf. Table 1).
(12) The computed and experimental K values clearly fall in the same
order. Numerical differences may have arisen, because the calculations
did not treat the dissolved ions in sufficient detail, ignoꢀring inter aliþa
cationꢀanion interactions between both TBAþ and Cl , and TBA
and 2.
(14) Calculated7 values of K for eq 2 with X = p-NO2, p-Me, and
p-OMe are 1.46 ꢁ 1014, 1.16 ꢁ 10ꢀ3, and 4.24 ꢁ 10ꢀ6 Mꢀ1, respectively.
See Table S-1 in the SI.
(15) Smith, M. B.; March, J. March’s Advanced Organic Chemistry,
Reactions, Mechanisms, and Structure, 6th ed.; Wiley: New York, 2007;
p 404.
(16) The 3pꢀ2p overlap of Cl with the para phenyl carbon atom is less
efficient for transmission of electronic effects than the 2pꢀ2p FꢀC
overlap; hence, σp is more appropriate than σpþ for the p-Cl substituent.
(13) kf values were redetermined in DCE. The values in ref 2 were
measured in MeCN/THF or CCl4.
Org. Lett., Vol. 15, No. 8, 2013
2015

interactions between Bu4Nþ and 2 will reduce the effective
negative charge on the carbanion and mitigate the electro-
nic effects wielded by X; accordingly, F will decrease.
Possible aggregation of Bu4NþArCCl2 could cause an
ꢀ
additional dispersal of negative charge and further diminu-
tion of F.
The F values for the Hammett correlations of Kexp and
Kcalcd are positive, indicating that eq 2 shifts toward the
product carbanion (K increases) as X becomes more
electron withdrawing, stabilizing the carbanion and desta-
bilizing the carbene. An analogous F value (þ2.48) was
observed for the equilibrium between carbenes 1 and
1,3,5-trimethoxybenzene affording the corresponding
π-complexes.2
A previous determination of kf for eq 2 in MeCNꢀ
THF and CCl4 solvents gave F = þ0.86 (r = 0.985).3 A
redetermination in DCE yields the data in Table 2. Omit-
ting X = F leads to an excellent correlation vs σp (σm for
m-Cl) with F = þ0.79 (r = 0.999, Figure S-46). Inclusion
of the point for X = F causes a very significant decay in
fitting vs σp (F = þ1.06, r = 0.776, Figure S-47). However,
using σpþ for p-F and σm for m-Cl, we obtain a reasonablþe
correlation, Figure 3, where F = þ1.05 (r = 0.955). If σp
þ
Figure 1. Hammett correlation of log Kexp vs σp [σp for p-F,
σm for m-Cl] for the equilibria of eq 2. See Table 2 and text:
F = þ3.18 (r = 0.996).
Figure 2. Hammett correlation of log Kcalc vs σp [σpꢀ for p-NO2,
σpþ for p-F, σm for m-Cl] for the equilibria of eq 2. See Table 2
and text: F = þ12.0 (r = 0.993).
þ
Figure 3. Hammett correlation of log kf vs σp [σp for p-F, σm
for m-Cl] for the forward reaction of eq 2. See Table 2 and text:
F = þ1.05 (r = 0.955).
analyses, we present correlations that use σpþ or σp for p-F.
Again, the F values are comparable and the reader can
weigh the benefits of the two approaches.
We also obtained an excellent correlation of the com-
puted K values7 of Table 2 and ref 14; cf. Figure 2. Here we
used σpꢀ for p-NO2,3 σpþ for p-F, σm for m-Cl, and σp for
the other five substituents. We find F = þ12.0 (r = 0.993).
A comparably good fit was obtained using σp for p-F (F =
þ12.3, r = 0.993, Figure S-44), whereas using σpþ for p-Cl
and p-F led to some degradation in the fit (F = þ9.6, r =
0.977, Figure S-45).
is used for all X (except m-Cl), the fit is poor with
F = þ0.99, r = 0.906, Figure S-48.
The F value for the forward process of eq 2 is positive;
electron withdrawing groups that destabilize carbenes 1
and stabilize carbanions 2 accelerate the reaction. The
magnitude of F (∼0.8ꢀ1.0) is about a third of F for the
equilibrium of eq 2, presumably because only a fraction of
the unit negative charge imposed on the carbanion is
present at the transition state of the forward reaction.
An excellent correlation is obtained for the kb of the
reverse reaction of eq 2, if X = F is omitted: a plot of log kb
vs σp (σm for m-Cl) gives F = ꢀ2.49, r = 0.993, Figure S-49.
Inclusion of X = F affords a slightly poorer correla-
tion with F = ꢀ2.63, r = 0.979, Figure 4. Use of σpþ for
Why is F for Kcalcd (þ12.0) so much higher than F for
K
exp (þ3.18)? The calculated K values refer to eq 2 with a
full unit of negative charge on carbanion 2, which will thus
be very sensitive to electronic effects exerted by substituent
X. However, in the experimental systems, cation/anion
2016
Org. Lett., Vol. 15, No. 8, 2013

X = F yields a correlation of comparable quality with
F = ꢀ2.12, r = 0.973, Figure S-50. Finally, a correlation
with σpþ for all points (except m-Cl) gives an inferior result;
F = ꢀ2.03, r = 0.943, Figure S-51.
and the equilibrium. For the kb process, electron donating
groups destabilize the carbanion but stabilize the product
carbene, accelerating the reaction and generating a nega-
tive F. The magnitude of F for kb (∼2.0ꢀ2.6) is about twice
that of F for kf. Presumably, the slower reverse reaction
senses the transition state for eq 2 as being ‘later’ along the
reaction coordinate than the faster forward reaction; hence
the reverse reaction shows a greater sensitivity to the
electronic effects of X.
In summary, substituted arylchlorocarbenes react with
chloride reversibly to form the corresponding aryldichloro-
methide carbanions. Experimental and computed equili-
brium constants, as well as rate constants for the forward
and reverse reactions, are correlated by the Hammett
equation. The mechanistic implications of the resulting
F values are analyzed.
Acknowledgment. We are grateful to the National
Science Foundation for financial support.
Supporting Information Available. Figures S-1ꢀS-51;
computational details; computed thermodynamic param-
eters for eq 2 (Table S-1); computed geometries and
energetics for carbenes and carbanions appearing in
eq 2. This material is available free of charge via the
Internet at http://pubs.acs.org.
Figure 4. Hammett correlation of log kb vs σp (σm for m-Cl) for
the reverse reaction of eq 2. See Table 2 and text: F = ꢀ2.63
(r = 0.979).
The sign of F for the reverse reaction of eq 2 is negative,
in contrast to the positive F for both the forward reaction
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 8, 2013
2017