Nucleofugality and nucleophilicity of fluoride in protic solvents

Article abstract of DOI:10.1021/jo300141z

A series of p-substituted benzhydryl fluorides (diarylfluoromethanes) were prepared and subjected to solvolysis reactions, which were followed conductometrically. The observed first-order rate constants k₁(25 °C) were found to follow the correlation equation log k₁(25 °C) = s_f(N_f + E_f), which allowed us to determine the nucleofuge-specific parameters N_f and s_f for fluoride in different aqueous and alcoholic solvents. The rates of the reverse reactions were measured by generating benzhydrylium ions (diarylcarbenium ions) laser flash photolytically in various alcoholic and aqueous solvents in the presence of fluoride ions and monitoring the rate of consumption of the benzhydrylium ions by UV-vis spectroscopy. The resulting second-order rate constants k_-1(20 °C) were substituted into the correlation equation log k_-1 = s_N(N + E) to derive the nucleophilicity parameters N and s_N for fluoride in various protic solvents. Complete Gibbs energy profiles for the solvolysis reactions of benzhydryl fluorides are constructed.

Full text of DOI:10.1021/jo300141z

Article
pubs.acs.org/joc
Nucleofugality and Nucleophilicity of Fluoride in Protic Solvents
Christoph Nolte, Johannes Ammer, and Herbert Mayr*
Department Chemie, Ludwig-Maximilians-Universitat Munchen, Butenandtstrasse 5-13 (Haus F), 81377 Munchen, Germany
̈
̈
̈
S
* Supporting Information
ABSTRACT: A series of p-substituted benzhydryl fluorides
(diarylfluoromethanes) were prepared and subjected to
solvolysis reactions, which were followed conductometrically.
The observed first-order rate constants k1(25 °C) were found
to follow the correlation equation log k1(25 °C) = sf(Nf + Ef),
which allowed us to determine the nucleofuge-specific
parameters Nf and sf for fluoride in different aqueous and
alcoholic solvents. The rates of the reverse reactions were
measured by generating benzhydrylium ions (diarylcarbenium ions) laser flash photolytically in various alcoholic and aqueous
solvents in the presence of fluoride ions and monitoring the rate of consumption of the benzhydrylium ions by UV−vis
spectroscopy. The resulting second-order rate constants k−1(20 °C) were substituted into the correlation equation log k−1 = sN
(N + E) to derive the nucleophilicity parameters N and sN for fluoride in various protic solvents. Complete Gibbs energy profiles
for the solvolysis reactions of benzhydryl fluorides are constructed.
electrofugality parameter Ef, and combinations of leaving
groups and solvents are characterized by the nucleofuge-
specific parameters Nf and sf .
INTRODUCTION
■
Fluorine is a rare element in organic natural products. Only approxi-
mately one dozen organic fluorine compunds are doubtlessly
identified.1−4 More fluorinated compounds that are isolated
from natural material, e.g., sponges, probably originate from
human sources.5,6 In contrast, other halides have been detected
in several thousand natural products. On the other hand, fluoro-
substituted compounds have become highly important in
medicinal and agricultural chemistry,7−13 and 20−25% of the
drugs in the pharmaceutical pipeline contain at least one fluorine
atom.12,13 Since the van der Waals radius of fluorine (1.47 Å) is
between that of oxygen (1.52 Å) and hydrogen (1.20 Å),14
incorporation of fluorine into a biologically active substance
strongly affects the electronic properties without large changes
of its structure. Bioavailability, lipophilicity, blood−brain barrier
permeability, and metabolic stability of a pharmaceutically
active molecule can, therefore, be customized by incorporation
of fluorine. However, incorporation of fluorine into complex
organic molecules is often a challenging task for the synthetic
chemist.15−20 Nucleophilic substitutions with [18F]-fluoride are
the key steps in various syntheses of radiopharmaceuticals used
in positron emission tomography (PET).21−26 The low
reactivity of the C−F bond has already been recognized by
Ingold27 and Hughes.28 Consequently, only few quantitative
data on the leaving group abilities29−35 and the nucleophilic
reactivities of fluoride are available.36,37
log k (25°C) = s (N + E )
(1)
1
f
f
f
So far we had not yet incorporated fluoride into these scales,
which presently include parameters for more than 100 leaving
group/solvent pairs. In this work, we report on an efficient
synthesis of a series of substituted benzhydryl fluorides 2b−f,
the rate constants (k1) for their solvolyses in various solvents,
and the rates (k−1) of the reactions of fluoride ions with
benzhydrylium ions (Scheme 1, Table 1) in a variety of
solvents.
Scheme 1. Synthesis and Ionization of Benzhydryl Fluorides
In recent years, the Kronja group and we have introduced a
novel approach to analyze leaving group abilities in solvolysis
reactions.38−44 By using benzhydrylium ions of variable
stabilization as reference electrofuges, we have been able to
compare nucleofugalities of anions and neutral leaving groups
in different solvents over a wide range of nucleofugality. For the
correlation of the solvolysis rate constants k1 (s−1) we have
employed eq 1, where carbocations are characterized by the
RESULTS
■
Synthesis of Benzhydryl Fluorides. While Swain and co-
workers used hydrogen fluoride for the preparation of trityl
Received: January 18, 2012
Published: February 16, 2012
© 2012 American Chemical Society
3325
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
Table 1. Benzhydrylium Ions 3a−l Employed in This Study
and Their Electrofugalitites Ef and Electrophilicities E
Scheme 2. Simplified Solvolysis Scheme for SN1 Reactions of
Benzhydryl Fluorides
a solution of 80E20W (80% ethanol, 20% water) containing
0.08 M piperidine, showed a linear correlation between the
initial concentration of the benzhydryl fluoride 2e and the
conductance at the end of the reaction within the investigated
concentration range (see Figure S1 of the Supporting Information).
As a consequence, monoexponential increases of the conduc-
tance (G) were observed during the solvolysis reactions, and
the first-order rate constants k1 (Table 2) were obtained by
Table 2. Solvolysis Rate Constants k1 of the Benzhydryl
Fluorides 2a−f in Different Solvents (at 25 °C)
a
solvent
ArAr′CHF
k1 (s−1
)
methanol (100M)
2f
3.91 × 10−2
5.49 × 10−3
1.99 × 10−3
3.42 × 10−4
7.77 × 10−2
1.42 × 10−2
4.34 × 10−3
7.82 × 10−4
5.75 × 10−5
2.75 × 10−7
4.26 × 10−3
5.63 × 10−4
1.63 × 10−4
8.35 × 10−2
9.28 × 10−3
4.66 × 10−3
1.01 × 10−3
9.98 × 10−5
7.90 × 10−3
1.11 × 10−3
4.85 × 10−4
3.55 × 10−3
1.87 × 10−3
3.80 × 10−4
3.28 × 10−5
1.20 × 10−3
1.43 × 10−4
9.26 × 10−5
2e
2d
2c
2f
80E20W
2e
2d
2c
2b
2a
2f
b
c
ethanol (100E)
60AN40W
2e
2d
2f
a
b
From ref 42. From ref 45.
2e
2d
2c
2b
2f
fluoride and the parent benzhydryl fluoride (2a) from the
corresponding alcohols,35 Ando avoided the use of hydrofluoric
acid by treating benzhydryl bromide with silver fluoride
dispersed on calcium fluoride (Scheme 1).46 Following this
procedure, the symmetrical benzhydryl fluorides 2b and 2f
were synthesized, isolated, and characterized. The liquid 2b was
purified by distillation under high vacuum, and 2f was
recrystallized from dichloromethane/diethyl ether. The other
benzhydryl fluorides were not isolated because they decom-
posed during distillation and did not crystallize readily. They
were synthesized in acetonitrile solution (0.2 M) and used for
kinetic investigations without prior evaporation of the solvent.
In general, great care has to be taken to exclude traces of water
and acid during synthesis and handling of these compounds,
since traces of acid lead to autocatalytic decomposition of the
benzhydryl fluorides.31
Kinetics of Benzhydryl Fluoride Solvolyses. When
compounds 2b−f were dissolved in aqueous or alcoholic
media, an increase of conductance was observed. As the
solvolyses (Scheme 2) of alkyl fluorides are prone to
autocatalysis,31 amines were added to trap the protons and to
deprotonate the released hydrofluoric acid quantitatively, thus
ensuring a linear dependence of the conductance on the
reaction progress. Calibration experiments, i.e., stepwise
addition of the rapidly solvolyzing benzhydryl fluoride 2e to
d
80AN20W
60A40W
2e
2d
2e
2d
2c
2b
2f
80A20W
90A10W
2e
2f
a
Mixtures of solvents are given as (v/v); solvents: A = acetone, AN =
b
acetonitrile, E = ethanol, M = methanol, W = water. Rate constant
was determined by Swain et al.35 and not used for the calculation of Nf
c
and sf parameters. Eyring activation parameters: ΔH⧧ = 62.5 kJ mol−1,
d
ΔS⧧ = −81.0 J mol−1 K−1. Eyring activation parameters: ΔH⧧ = 64.1
kJ mol−1, ΔS⧧ = −74.7 J mol−1 K−1.
fitting the monoexponential function (eq 2) to the time-
dependent conductance Gt.
−k t
1
G = G (1 − e
) + C
(2)
t
∞
3326
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
The majority of ionization rate constants (k1) were deter-
mined at least at two different concentrations of piperidine
(0.08 to 0.16 M, see Supporting Information); for some
systems (2c and 2e in 60AN40W and 2d in 100E) 2,6-lutidine
was additionally used as additive for the determination of k1. In
all cases the rate constants varied only within the typical experi-
mental error margin (<13%).
The observation of first-order kinetics and the fact that the
observed first-order rate constants are independent of the
nature and concentration of the added amines (Figure 1)
the abscissa (Ef axis) and the sf parameters as the slopes of the
correlation lines (Table 3).
Table 3. Nucleofuge-Specific Parameters Nf and sf for the
Fluoride Anion in Various Solvents
a
a
solvent
100M
Nf
sf
−1.44
−1.24
−2.08
−1.47
−2.13
−2.23
−2.72
0.99
0.92
1.13
0.83
0.99
0.79
1.07
80E20W
100E
60AN40W
80AN20W
60A40W
80A20W
a
Nf and sf are defined in eq 1.
calcd
The rate constant k1
= 2.05 × 10−7 s−1 for the solvolysis
of 2a calculated by using eq 1, the nucleofugality parameters for
fluoride in 80E20W (Table 3), and the electrofugality parameter
for the unsubstituted benzhydryl cation 3a (Table 1) is in
excellent agreement with the previously reported experimental
rate constant of 2.75 × 10−7 s−1,35 which demonstrates the
power of eq 1 and the practicability of these parameters.
Nucleophilicity of Fluoride Anions in Various Sol-
vents. It is well-known that nucleophilicity is not simply the
reverse of nucleofugality,42 and therefore, the nucleophilic
reactivity of the fluoride anion has been determined separately.
Previously, we determined the nucleophilicity parameters N and sN
of chloride and bromide in various solvents47 by measuring the
rate constants k−1 (M−1 s−1) of their reactions with benzhydrylium
ions and correlation of the data by eq 3. As demonstrated
previously, rate constants of the reactions of carbocations with
charged and neutral nucleophiles can be expressed by eq 3,
where k−1 is the second-order rate constant at 20 °C, sN is a
nucleophile-specific slope parameter, N is a nucleophilicity
parameter, and E is an electrophilicity parameter.45
Figure 1. Observed rate constants (k1) for the solvolysis of 2d in
60AN40W (60% acetonitrile, 40% water, v/v) at various concen-
■
▲
●
trations of amine ( , triethylamine; , 2,6-lutidine; , piperidine).
indicate that the observed first-order rate constants equal k1. If
common-ion return would occur (k−1[F−] ≈ ks), an increase of
the piperidine concentration would lead to an increase of the
overall rate because trapping of the carbocation by the amine
(kamine) would suppress the ion recombination. The second-
order rate law for an SN2 reaction requires a linear increase of
kobs with the concentration of the amine, which can be excluded
from the data shown in Figure 1.
Plots of log k1 for the solvolyses of 2b−f in various solvents
versus the electrofugality parameters Ef of the benzhydrylium
ions are linear (Figure 2), indicating the applicability of eq 1.
log k (20°C) = s (N + E)
(3)
−1
N
We now used laser flash photolysis and stopped-flow
techniques to characterize the nucleophilic reactivity of fluoride
ions by measuring the rates of their reactions with benzhydryl
cations 1b−l in a series of solvents.
The benzhydryltriphenylphosphonium tetrafluoroborates
4b−h (Scheme 3), which were employed as precursors for
Scheme 3. Precursors 4b−j for the Laser-Flash Photolytic
a
Generation of the Benzhydrylium Ions 3b−j
a
Substituents R1 and R2 are defined in Table 1.
Figure 2. Plots of log k1 for the solvolysis reactions of various
benzhydryl fluorides versus the electrofugalities Ef. The correlation
lines for 80E20W and 100M are shown in Figure S8 of the Supporting
Information. Mixtures of solvents are given as (v/v); solvents: A =
acetone, AN = acetonitrile, E = ethanol, M = methanol, W = water.
the laser flash-photolytic generation of benzhydryl cations, were
prepared by treating the corresponding benzhydrols with
equimolar amounts of triphenylphosphane and aqueous
tetrafluoroboric acid, followed by heating to 140−180 °C.48
The benzhydryltributylphosphonium tetrafluoroborates 4i and
4j were prepared in acetonitrile solution by adding tributylphos-
phane to the corresponding benzhydrylium tetrafluoroborates 3i,j
From these correlations, obtained for the fluoride leaving group
ability in different protic solvents, one can extract the
nucleofugality parameters, Nf, as the negative intercepts on
3327
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
a
Table 4. Rate Constants for the Reactions of Benzhydrylium Ions 3b−l with Fluoride (k−1) and Pure Solvent (ks)
solvent (N1, sN)
Ar2CH+
E
k−1 (M−1 s−1
)
ksexp (s−1
)
ks calcd by eq 1 (s−1
)
ksexp/kscalcd
100M (7.54, 0.92)
3h
3g
3f
−1.36
−0.83
0.00
1.56 × 106
4.12 × 106
1.14 × 107
c
7.39 × 105
2.44 × 106
9.91 × 106
1.97 × 107
4.87 × 105
1.66 × 106
1.79 × 105
1.02 × 106
1.40 × 105
3.15 × 105
1.04 × 106
1.14 × 104
3.16 × 104
9.61 × 104
2.81 × 105
1.03 × 106
3.35 × 107
9.38 × 103
3.08 × 104
9.49 × 104
2.52 × 105
7.05 × 103
2.12 × 104
6.78 × 104
k
4.85 × 105
1.49 × 106
8.65 × 106
3.15 × 107
9.39 × 104
4.76 × 105
4.25 × 104
8.81 × 105
1.5
1.6
1.1
0.63
5.2
3.5
4.2
1.2
2.8
1.8
0.94
5.8
5.4
2.9
2.4
1.5
0.55
5.2
5.7
3.2
2.5
6.9
6.6
3.5
b
3e
3g
3f
0.61
80E20W (6.68, 0.85)
100W (5.20, 0.89)
−0.83
0.00
3.83 × 106
1.06 × 107
1.09 × 105
1.02 × 106
1.37 × 105
3.28 × 105
1.20 × 106
1.66 × 105
3.94 × 105
1.82 × 106
4.36 × 106
9.04 × 106
d
d
e
3f
0.00
3d
3f
1.48
f
10AN90W (5.16, 0.91)
0.00
4.96 × 104
e
f
3e
3d
3h
3g
3f
0.61
1.78 × 105
e
f
1.48
1.10 × 106
g
60AN40W (5.02, 0.90)
80AN20W (5.02, 0.89)
−1.36
−0.83
0.00
1.97 × 103
g
5.90 × 103
g
3.30 × 104
g
3e
3d
3b
3h
3g
3f
0.61
1.17 × 105
g
1.48
7.08 × 105
h
g
3.63
2.0 × 108
6.72 × 105
1.72 × 106
5.10 × 106
9.76 × 106
2.89 × 106
6.03 × 106
1.85 × 107
1.97 × 103
2.95 × 104
4.17 × 105
2.31 × 106
1.12 × 108
1.40 × 108
2.69 × 108
2.40 × 109
8.63 × 109
6.10 × 107
−1.36
−0.83
0.00
1.81 × 103
5.36 × 103
2.94 × 104
1.02 × 105
1.02 × 103
3.21 × 103
1.93 × 104
i
j
j
3e
3h
3g
3f
0.61
90AN10W (4.56, 0.94)
98AN2W
−1.36
−0.83
0.00
3l
−7.02
−5.53
−3.85
−3.14
−1.36
−0.83
0.00
3k
3j
k
k
3i
k
3h
3g
3f
k
k
k
3c
3b
2.11
3.85 × 105
6.18 × 106
3.63
a
b
Mixtures of solvents are given as (v/v); solvents: AN = acetonitrile, E = ethanol, M = methanol, W = water. Calculated from the intercept of plots
c
of kobs vs [F−]. Previously ks = 8.4 × 106 s−1 has been reported in ref 52. The rate constant k−1 could not be determined because the slope in the plot
d
of [F−] versus kobs was too small. Calculated from the intercept of plots of kobs vs [F−]. In ref 53, ks = 1.0 × 105 and 7.8 × 105 s−1 have been
e
reported for 3f and 3d in 100W, respectively. Calculated from the intercept of plots of kobs vs [F−]. In ref 51, ks = 9.55 × 104 and 7.99 × 105 s−1 have
f
been reported for 3f and 3d in 9AN91W, respectively. The N1 and sN parameters for 9AN91W from ref 51 were used for the calculation of kscalcd
.
g
h
The N and sN parameters for 67AN33W from ref 51 were used for the calculation of kscalcd. This value has to be considered approximate due to the
i
j
low slope of the kobs vs [F−] plot. In ref 51, ks = 9.82 × 104 s−1 has been reported for 3f in 80AN20W. In ref 51, ks = 7.11 × 102 and 9.87 × 104 s−1
have been reported for 3h and 3f in 90AN10W, respectively. Not determined.
k
until complete decolorization was achieved. The stabilized
benzhydrylium tetrafluoroborates 3i−l were prepared as
previously described.45
because we observed phase separation when trying to dissolve
potassium fluoride in these mixtures.
The benzhydrylium ions 3b−j were generated in these
solvents by irradiation of 4b−j with a 7 ns laser pulse from the
fourth harmonic (266 nm) of a Nd/YAG laser.48,50 In the
absence of added nucleophiles, we typically observed
monoexponential decays of the absorbances of 3b−j. Fitting
The choice of suitable sources of fluoride anions is not trivial.
For the kinetic investigations in methanol, water, and 10AN90W,
we used potassium and cesium fluoride, which were sufficiently
soluble in these solvents. Methanol was the only anhydrous
solvent, where both potassium fluoride and cesium fluoride could
be used as fluoride source. In other anhydrous solvents, such as
acetonitrile, the solubility of alkali fluorides is too low for kinetic
studies. Unlike tetrabutylammonium chloride and bromide, which
we used in the previous study,47 anhydrous tetrabutylammonium
fluoride is prone to elimination and is stable for only several hours
at room temperature.49 Therefore, we used the commercially
available tetrabutylammonium fluoride trihydrate, which did not
allow us to characterize fluoride in anhydrous solvents. Tetra-
butylammonium fluoride trihydrate was exclusively used as fluo-
ride source for the investigations in the aqueous solvent mixtures
(98AN2W, 90AN20W, 80AN20W, 60AN40W, 80E20W)
obst
of the single exponential function At = A0 e−k + C provided
the first-order rate constants for the reactions of 3b−j with the
solvents, ks, which are listed in Table 4. When the carbocations
were generated in the presence of added fluoride, the observed
rate constants increased linearly with the concentration of
added fluoride (Figure 3). As expressed by eq 4, the observed
pseudo-first-order rate constants kobs are the sum of a second-
order term for the reactions of the carbocations with fluoride
ions (k−1 in Scheme 2) and a first-order term for the reactions
of the carbocations with the solvents (ks).
−
k
= k [F ] + k
(4)
obs
−1
s
3328
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
Figure 3. Decay of the absorption of 3f in methanol, observed at
500 nm in the presence of 0.3 M CsF. Inset: Determination of
k−1 from a plot of kobs against the fluoride ion concentration. Alter-
○
●
natively, KF ( ) and CsF ( ) were used as the fluoride source.
Plots of kobs versus the fluoride concentrations resulted in
linear correlations according to eq 4 as exemplified in Figure 3.
The second-order rate constants k−1 for the reactions with
fluoride listed in Table 4 were obtained from the slopes of these
plots. The intercepts correspond to the background reaction
with the solvent (ks), which were also determined independ-
ently (Table 4). Table 4 demonstrates the good agreement of
the experimental data with the values calculated from eq 3 using
the previously published N1 and sN parameters of the solvents.51
As the fluoride anion is a weak base, we had to make sure
that fluoride and not hydroxide was the acting nucleophile in
aqueous solvents. For the reaction of the bisanisyl carbenium
ion (3f) with fluoride in water, we used fluoride concentrations
ranging from 0.06 to 0.91 M. From pKb for fluoride in water
(10.9), one calculates concentrations of hydroxide from 0.87 ×
10−6 to 3.4 × 10−6 M. With eq 3 and the published nucleo-
Figure 4. 1H NMR spectra (200 MHz) recorded for product analysis:
(A) (tol)2CHCl in CD3CN; (B) 1H NMR spectrum recorded 5 min after
combining (tol)2CHCl with 3.5 equiv of tetrabutylammonium fluoride
trihydrate in 60AN40W (deuterated solvents); (C) (tol)2CHF (2b) in
CD3CN; (D) (tol)2CHOH in 60AN40W (deuterated solvents).
philicity parameters for hydroxide in water (N = 10.47, sN
=
0.61)54 we can calculate hypothetical pseudo-first-order rate
constants k1Ψ(OH−) = 2.1 to 8.3 s−1 for the reaction of 3f with
hydroxide at the above-mentioned concentrations of fluoride.
As shown in the Supporting Information, rate constants kobs of
(1.84−2.80) × 105 s−1 were observed for the reaction of the
bisanisyl carbenium ion (3f) with fluoride. Therefore, the
nucleophilic reactivity of the hydroxide anion can be neglected
for the evaluation of the kinetic experiments.
the relative reactivities of the benzhydryl cation 3b toward F− and
H2O (OH− negligible as discussed above). This ratio shall be
compared with the ratio of the absolute rate constants for the re-
action of 3b with F− and H2O. The second-order rate constant
for the reaction of 3b with fluoride in 60AN40W has been deter-
mined (k−1 = 2.0 × 108 M−1 s−1).56 The rate constant for the
reaction of 3b with 60% aqueous acetonitrile is 3.35 × 107 s−1
Further evidence that fluoride, and not hydroxide, was the
1
active nucleophile was obtained by a H NMR spectroscopic
product analysis. A solution of 4,4′-dimethylbenzhydryl chloride
((tol)2CHCl) in deuterated acetonitrile was combined with a
solution of tetrabutylammonium fluoride trihydrate in aqueous
acetonitrile to yield a solution of 0.15 M 4,4′-dimethylbenz-
hydryl chloride and 0.54 M tetrabutylammonium fluoride in
60% aqueous acetonitrile (60AN40W). In line with a calculated
solvolysis rate constant of k1 ≈ 2.4 s−1 at 25 °C,55 (tol)2CHCl
1
(see Table 4). According to the H NMR spectrum, 70% of
(tol)2CHCl (c0 = 0.15 M) are converted into the benzhydryl
fluoride 2b. Thus, the average concentration of fluoride [F−]av
during the reaction is 0.49 M (eq 5).
1
was not observable in a H NMR spectrum taken immediately
−
−
after mixing the two solutions. As depicted in Figure 4, the
characteristic doublet at δ = 6.46 (d, 1 H, 3JHF = 46.0 Hz, CHF)
for 2b and singlet at δ = 5.69 indicated a 2.4/1 ratio of 2b and
4,4′-dimethylbenzhydrol. As the benzhydryl fluoride 2b will
solvolyze with a half-life of 1.9 h under these conditions (Table 2),
the observed 4,4′-dimethylbenzhydrol cannot arise from
hydrolysis of 2b. Hence, the observed product ratio reflects
[F ] = [F ] − 0.5[2b]
av
0
= (0.54 − 0.5 × 0.70 × 0.15) M = 0.49 M
(5)
Multiplication of the second-order rate constant (k−1) for the
reaction of 3b with [F−]av yields the pseudo-first-order rate
constant for the reaction of 3b with F−, which is divided by ks
3329
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
to yield an expected product ratio of 2.9/1 (eq 6), in fair
1
agreement with the product ratio observed by H NMR (2.4).
−
8
k
[F ]
[2b]
[(tol) CHOH]
2
2.0 × 10 × 0.49
−1 av
=
=
= 2.9
7
k
s
3.4 × 10
(6)
A more accurate comparison of the kinetic results with the
1H NMR product analysis is provided in the Supporting
Information.
When plotting log k−1 for the reactions of F− with benzhydryl
cations against the electrophilicities E of the corresponding
benzhydryl cations, linear correlations according to eq 3 were
obtained (Figure 5). In 98AN2W solution, we could determine
■
◆
Figure 6. Plot of log k−1 ( ) and log ks ( ) for the reactions of
benzhydrylium ions with fluoride ions and methanol versus their
electrophilicity parameters E.
in eq 3, see below), with the consequence that the accurate
determination of the small contribution of the k−1[F−] term
becomes difficult.
Another limitation of the laser flash photolytic technique is
the recombination of the benzhydrylium ions with the
phosphine photoleaving group. For that reason, we could not
determine rate constants kobs < 105 s−1 in 100W and in
10AN90W.
Plots of the second-order rate constants k−1 for the reactions
of 3 with fluoride (Table 4) against the electrophilicity para-
meters E of the benzhydrylium ions were linear (Figure 5) for
rate constants kobs up to 1 × 108 s−1 and yielded the sN
parameters and the nucleophilicity parameters N of fluoride in
methanol and aqueous solvents (Table 5). Rate constants log
Table 5. Nucleophilicity Parameters N and sN for Fluoride in
Various Solvents
Figure 5. Plot of log k−1 for the reactions of benzhydrylium ions with
fluoride ions versus their electrophilicity parameters E. Mixtures of
solvents are given as (v/v); solvents: AN = acetonitrile, M = methanol,
W = water.
a
solvent
N
sN
100M
11.31
0.63
(0.53)
(0.66)
0.64
b
b
b
80E20W
100W
(13.2)
(7.7)
b
absolute rate constants for the reactions of fluoride with a wide
variety of benzhydrylium ions. The rate constants for the re-
actions of fluoride with 3k and 3l, determined by the stopped-
flow method, lie on the same graph as the rate constants for the
reactions of 3b−j, which were determined by the laser flash
photolysis technique, demonstrating the consistency of the
results obtained by the different methods (Figure 5). The linear
correlation for the rate constants for 3 h−l bends downward as
k−1 exceeds 108 M−1 s−1 due to the proximity of the diffusion
limit (ca. 2 × 1010 M−1 s−1). An analogous behavior was
observed for numerous other nucleophiles.57
Only a narrow range of carbocations could be investigated in
solvent mixtures containing a higher percentage of water or
alcohols. The reactions of fluoride with the more stabilized
benzhydrylium ions (E < −2) do not proceed quantitatively,
and the fast reversible reactions with F− have to compete with
the slower but irreversible reactions with water or methanol
giving rise to kinetics that are difficult to evaluate. On the other
hand, the reactions of fluoride with the less stabilized
benzhydrylium ions (E > 2) are difficult to follow, because
for highly electrophilic carbocations the major fraction of kobs
(eq 4) is due to the reaction of the benzhydrylium ion with
the solvent. As shown for the example of methanol in Figure 6 the
rate constants for the reactions of 3 with the solvents are more
sensitive toward variation of the electrophiles than those of the
reactions of 3 with fluoride (fluoride has a lower sN parameter
10AN90W
60AN40W
80AN20W
90AN10W
98AN2W
8.05
9.75
0.63
11.40
12.27
10.88
0.59
0.59
0.83
a
Mixtures of solvents are given as (v/v); solvents: AN = acetonitrile,
b
E = ethanol, M = methanol, W = water. Correlation based on only
two data points.
kobs > 8.0 were not used for the determination of reactivity
parameters N and sN as these reactions approach the diffusion
limit (see above).
DISCUSSION
■
With the nucleofugality parameters (Nf, sf) listed in Table 3, it
now becomes possible to directly compare the leaving group
abilities of fluoride and those of other common leaving groups
in different solvents. As sf ≈ 1, a general overview can be
derived from the Nf parameters,42 some of which are compared
in Table 6. Depending on the solvent, fluoride is a slightly
weaker or better leaving group than 3,5-dinitrobenzoate.42
Replacing fluoride by chloride accelerates the solvolysis reac-
tions approximately by 4−5 orders of magnitude, while replac-
ing fluoride by bromide results in an acceleration of approxi-
mately 5−6 orders of magnitude. Thus the unsubstituted
3330
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
Table 6. Comparison of Nucleofugalities Nf of Important
Leaving Groups in Different Solvents and Solvolysis Half
Lives of Ph2CHX in 80E20W at 25°C
leaving group X
a
solvent
DNB
F
Cl
Br
80A20W
100E
−2.34
−2.05
−1.43
164 d
−2.72
−2.08
−1.24
39 d
2.03
3.01
2.93
4.36
27 s
1.82
80E20W
3.24
b
τ1/2
7 min
a
b
DNB = 3,5-dinitrobenzoate. Solvolysis half-life of Ph2CHX in 80%
aqueous ethanol (80E20W) at 25 °C calculated by eq 1 by using
Ef(3a) = −6.03 from Table 1, Nf from this table, and sf = 0.98 (for
DNB), 0.92 (for F), 0.99 (for Cl), and 0.95 (for Br); sf for DNB, Cl,
and Br from ref 42.
benzhydryl fluoride will solvolyze in 80E20W with a half-life
of 1 month, whereas the half-life is 7 min for benzhydryl
chloride, 27 s for benzhydryl bromide, and only 50 ms for the
parent benzhydryl tosylate (not shown in Table 6). These
findings are in agreement with previous results by Swain and
Scott who reported chloride/fluoride ratios of 106 and 105 for
the couples trityl chloride/trityl fluoride (85% aq acetone) and
tert-butyl chloride/tert-butyl fluoride (80% aq ethanol).19 The
poor nucleofugality of F− is commonly attributed to the high
C−F bond energy. Table 7 shows a further reason: The
Figure 7. Comparison of the solvolysis rates (at 25 °C) for the
reactions of the 4,4′-dimethyl-substituted benzhydryl derivative
tol2CH-X with different leaving groups X (DNB = 3,5-dinitroben-
zoate). Mixtures of solvents are given as (v/v); solvents: A = acetone,
AN = acetonitrile, E = ethanol, W = water.
in solvolyses of benzhydryl chlorides and bromides, where the
ion combination is diffusion-limited, i.e., where the transition
states correspond to the carbocations, other factors must
contribute.42
Fluoride is not only a poorer nucleofuge than the other
halide ions but also a poorer nucleophile in protic solvents.
As shown in Figure 8, nucleophilicity increases in the series
Table 7. Eyring Activation Parameters for the Solvolyses of
Benzhydryl and tert-Butyl Halides in Different Solvents
ΔH⧧
ΔS⧧
a
entry
substrate
solvent
(kJ mol−1
)
(J mol−1 K−1
)
1
2
3
4
5
Ar2CHF (2d)
60AN40W
60AN40W
60AN40W
80A20W
64.1
83.0
86.5
90.3
85.1
−74.7
−34.6
0.55
b
b
(3-FC6H4)PhCH-Cl
(3-FC6H4)PhCH-Br
c
tert-butyl chloride
−51.6
−35.0
c
tert-butyl bromide
80A20W
a
Mixtures of solvents are given as (v/v); solvents: A = acetone, AN =
b
c
acetonitrile, W = water. This work (Supporting Information). Data
from ref 58.
activation entropies for the solvolysis reactions of benzhydryl
fluorides are considerably more negative than those of
benzhydryl chlorides and bromides. Thus, entries 1−3 show
that in 60% aqueous acetonitrile, ΔS⧧ becomes more negative
in the series Ar2CHBr to Ar2CHCl and Ar2CHF. The same
trend, more negative entropy of activation for R-Cl than for
R-Br, has also been observed in the tert-butyl series (entries 4 and 5).
The higher degree of solvent orientation needed for the solvation
of fluoride ions is in line with the relative hydration energies for
halide ions (−ΔHh°), which increase from I− (294 kJ mol−1),
Br− (335 kJ mol−1), Cl− (366 kJ mol−1) to F− (502 kJ mol−1).26
As illustrated for the 4,4′-dimethyl-substituted benzhydryl
derivatives tol2CH-X in Figure 7, the leaving group abilities
increase significantly in the series F− ≈ DNB ≪ Cl− < Br−, but
the exact ranking depends on the solvent.
Figure 8. Comparison of the nucleophilic reactivities of fluoride with
those of other halide anions in different solvents. Mixtures of solvents are
given as (v/v); solvents: AN = acetonitrile, M = methanol, W = water.
F− < Cl− < Br− in water, aqueous acetonitrile, and methanol.47 The
variable sensitivities sN for fluoride in different solvents (Table 5)
imply that the relative reactivities depend slightly on the nature
of the electrophile, which is also illustrated in Figure 5. The bis-
dihydrofuranyl annelated benzhydrylium ion 3h, for example,
reacts 104 times faster with F− in 98% acetonitrile/2% water
than in 100% water. One can extrapolate that in a solution
containing 15% water F− has a reactivity in between that of 98%
acetonitrile/2% water and 100% water. Because of the problems
obtaining anhydrous F−,49 we refrain to present a number for
As previously reported for several other leaving groups,42,59
the sensitivity parameters sf of fluoride decrease slightly with
increasing water content of the solvents. Since the carbocation
character is not fully developed in most of the benzhydryl
fluoride solvolyses investigated in this work (see below), the
trend in sf might be explained by a smaller degree of charge
separation in the transition states in solvents with a high
percentage of water. As analogous trends in sf are also observed
3331
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
anhydrous acetonitrile. A strong reduction of the nucleophilic
reactivity of fluoride ions by water molecules has also been
observed in SN2 reactions.36
The low nucleophilicity of fluoride in protic solvents
compared to chloride and bromide accounts for the fact that
common-ion return is rarely encountered in SN1 reactions of
alkyl fluorides. As a result, deviations from the first-order rate
law, due to reversible ionization toward the end of the kinetic
experiments, have not been observed in any of the benzhydryl
fluoride solvolyses described above.
A quantitative rationalization for this observation is given in
Figure 9, where the first-order rate constants of the reactions of
Figure 10. Free energy profiles (kJ mol−1) for the solyolyses of
differently substituted benzhydryl fluorides in 80AN20W at 25 °C
a
(20 °C for the reactions of 3a−h with nucleophiles). Calculated by
b
c
using eq 1. Direct measurement. Calculated by using eq 3.
Figure 9. Plot of log k1Ψ for the reactions of benzhydrylium ions with
aqueous acetonitrile 60AN40W (ks) and with fluoride in 60AN40W
(k1Ψ = [F−]k−1) at fluoride concentrations of 1.0 M (upper graph) and
at 1.7 mM (lower graph) versus their electrophilicity parameters E.
can, therefore, conclude that only alkyl fluorides that give much
less stabilized carbocations, i.e., substrates RF that require
heating to solvolyze within reasonable time periods, will ionize
via carbocation-like transition states.
various benzhydrylium ions with water in 60% aqueous acetonitrile
are compared with the corresponding pseudo-first-order rate
constants with fluoride (i.e., k−1[F−]) at different concen-
trations of fluoride. At substrate concentrations which are
typical for solvolysis experiments (1.7 mM), the reaction with
F− is approximately 102 times slower than the reaction with the
solvent. Only when high concentrations of fluoride are
employed, the reaction with F− is faster than the reaction
with water as demonstrated for the solvolysis of bis(p-tolyl)-
methyl chloride [(4-MeC6H4)2CHCl] in a 0.54 M solution of
nBu4N+ F− in 60AN40W (Figure 4).
EXPERIMENTAL SECTION
Materials. Silver Carbonate. Prepared according to a litera-
ture procedure by combining diluted aqueous solutions of silver
nitrate with potassium bicarbonate.60
■
Ando’s Fluorination Agent. Prepared according to the literature
procedure by Ando.46 Silver carbonate (10 g, 36 mmol) was mixed
with calcium fluoride (40 g, 0.51 mol) and ground thoroughly. The
mixture was transferred to a Teflon round-bottom flask equipped with
a Teflon stirring bar, before water (40 mL) was added. Then conc HF
(3.0 mL, 73 mmol) was slowly added to the stirred suspension, using a
plastic syringe equipped with a Teflon tube. The suspension was
stirred for 30 min, followed by evaporation of the water and drying for
several hours in the vacuum (1.3 × 10−3 mbar at 40−50 °C). The
fluorination agent (49 g, 98%) was obtained as a slightly yellowish
free-flowing granular powder (Ando reported a colorless powder). The
reagent can be stored in an opaque flask under air and can be used for
at least 1 month, but we recommend storage in a glovebox under
argon. The concentration of active fluorination agent (AgF) was
1.5 mmol/g in the resulting material.
CONCLUSION
■
The rate constants for the forward and backward reactions of
benzhydryl fluoride ionization can be combined to construct
quantitative energy profiles for the solvolysis reactions (Figure 10),
which differ significantly from those previously derived for
benzhydryl chlorides and bromides.
While SN1 solvolyses of benzhydryl chlorides and bromides,
which are commonly investigated at room temperature (i.e., 0.1 s
< τ1/2 < 1 d), have carbocation-like transition states (i.e., barrier-
free combinations of R+ with Cl− or Br−), solvolysis reactions of
benzhydryl fluorides, which proceed with similar rates, typically
do not have carbocation-like transition states. As shown in
Figure 10, the solvolysis of 2b, for which a solvolysis half-life of
approximately 63 h can be calculated, still yields a carbocation
that does not undergo barrier-free recombination with the
Benzhydryl Bromides. Used previously but some were only partially
characterized.61−66 For their synthesis, the corresponding benzhydrols
were refluxed with acetyl bromide (10 equiv) for 15 min, followed by
evaporation under vacuum. The remaining residue was crystallized
from CH2Cl2/n-pentane (ca. 1/10). In some cases good results were
also obtained with only 5 equiv of acetyl bromide. In general, better
results were achieved with the high excess of acetyl bromide, which
often yielded crystalline material after evaporation of the acetyl
bromide and acetic anhydride.
4,4′-Dimethylbenzhydryl Bromide (1b). Obtained from 4,4′-
dimethylbenzhydrol (3.99 g, 18.8 mmol) and acetyl bromide
(23.1 g, 188 mmol): 4.15 g (80%), colorless crystals; mp 48.0−
fluoride ion in aqueous acetonitrile. For the more rapidly
calcd
ionizing substrates 2f (τ1/2 = 88 s) and 2h (τ1/2
= 8 s) the
62
1
48.6 °C. H NMR spectra agreed with literature data. 1H NMR
(300 MHz, CDCl3): δ 2.35 (s, 6 H), 6.28 (s, 1 H), 7.13−7.16 (m, 4 H),
7.34−7.37 (m, 4 H). 13C NMR (75.5 MHz, CDCl3): δ 21.2, 55.9, 128.5,
energy well for the resulting carbocations is much deeper, and
one can extrapolate that similar situations should be encountered
for other alkyl halides with comparable ionization rates. One
3332
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
129.3, 138.0, 138.5. MS (+EI): m/z (%) = 195.2 (100) [M+ − Br]. Anal.
Calcd for C15H15Br (275.2): C, 65.47; H, 5.49. Found: C, 65.70; H,
5.55.
Kinetics of Solvolysis Reactions. Preparation of Stock
Solutions of the Benzhydryl Fluorides 2c−e for Kinetic Measure-
ments. In a carefully dried Schlenk flask, the fluorinating agent
(1.00 g containing 1.5 mmol AgF) was suspended in MeCN
(3 mL). A solution of the corresponding benzhydryl bromide
(1.0 mmol) in MeCN (2 mL) was added to the stirred sus-
pension at 0 °C. After 30 min, the solution was filtered (1H NMR
spectroscopic analysis in CD3CN showed quantitative con-
version of 1c−e as well as formation of 2c−e indicated by the
characteristic doublet for CHF). The clear solutions were used
directly for solvolytic measurements and could be used for
approximately 2 days.
Calibration. Calibration experiments were performed by stepwise
addition of 50 μL portions of a 0.2 M solution of 2e in acetonitrile to
30 mL of 80E20W (80% ethanol, 20% water) containing 0.08 M
piperidine. After the addition of a portion of 2e the conductance after
at least 200 s (for 2e in 80E20W: τ1/2 = 49 s) was recorded before the
next portion of 2e was added. Plots of the conductance against the
initial concentration of added benzhydryl fluoride 2e were linear
(Figure S1 in Supporting Information), demonstrating that solvolytic
rate constants can be determined reliably by time-dependent
conductance measurements.
4-Methoxybenzhydryl Bromide (1c). Obtained from 4-methoxy-
benzhydrol (2.25 g, 10.5 mmol) and acetyl bromide (12.9 g, 105 mmol):
1
2.17 g (74%), colorless crystals; mp 49.5−50.9 °C. H NMR spectra
agreed with literature data.66 1H NMR (300 MHz, CDCl3): δ 3.79
(s, 3 H), 6.30 (s, 1 H), 6.84−6.87 (m, 2 H), 7.25−7.38 (m, 5 H),
7.45−7.48 (m, 2 H). 13C NMR (75.5 MHz, CDCl3): δ 55.5, 55.8,
114.1, 128.1, 128.5, 128.6, 129.9, 133.5, 141.4, 159.5. MS (+EI): m/z
(%) = 197.2 (100) [M+ − Br].
4-Methoxy-4′-methylbenzhydryl Bromide (1d). Obtained from 4-
methoxy-4′-methylbenzhydrol (2.33 g, 10.2 mmol) and acetyl bromide
(13.0 g, 102 mmol): 2.03 g (69%), colorless crystals; mp 39.1−
40.2 °C. 1H NMR (300 MHz, CDCl3): δ 2.35 (s, 3 H), 3.81 (s, 3 H),
6.30 (s, 1 H), 6.85−6.88 (m, 2 H), 7.14−7.16 (m, 2 H), 7.35−7.40
(m, 4 H). 13C NMR (75.5 MHz, CDCl3): δ 21.2, 55.5, 56.0, 114.0,
128.4, 129.3, 129.8, 133.7, 138.0, 138.5, 159.4. MS (+EI): m/z (%) =
211.2 (100) [M+ − Br].
4-Methoxy-4′-phenoxybenzhydryl Bromide (1e). Obtained from
4-methoxy-4′-phenoxybenzhydrol (1.98 g, 6.46 mmol) and acetyl
bromide (7.95 g, 64.7 mmol): 2.03 g (67%), slightly pink crystals; mp
Kinetics. The following solvents were commercially available and
used as received: acetone (99.8%), acetonitrile (extra dry, water content
<50 ppm), methanol (99.8%). Commercially available absolute
ethanol was freshly distilled from sodium/diethyl phthalate. Doubly
distilled water (impendance 18.2 Ω) was obtained from a water
purification system. Solvolysis rate constants of benzhydryl derivatives
(Table 2 and Table S2 in the Supporting Information) were
monitored by conventional conductometry. Freshly prepared solvents
(30 mL, with amine additive in case of benzhydryl fluorides 2) were
thermostatted ( 0.1 °C) at the given temperature for 5 min prior to
adding the substrate. Typically 0.25 mL of a 0.2 M stock solution of
the benzhydryl derivative in acetonitrile was injected into the solvent
upon which an increase of conductance was observed. This increase
was recorded at certain time intervals resulting in about 3000 data
points for each measurement. The first-order rate constants k1 (s−1)
were obtained by least-squares fitting of the conductance data to a
1
61.3−62.0 °C. H NMR (300 MHz, CDCl3): δ 3.81 (s, 3 H), 6.32
(s, 1 H), 6.86−7.11 (m, 7 H), 7.33−7.44 (m, 6 H). 13C NMR (75.5
MHz, CDCl3): δ 55.47, 55.52, 114.0, 118.4, 119.4, 123.8, 129.8, 129.9,
130.0, 133.5, 136.1, 156.8, 157.3, 159.5 MS (+EI): m/z (%) = 289.3
(100) [M+ − Br].
4,4′-Dimethoxybenzhydryl Bromide (1f). Obtained from 4,4′-
dimethoxybenzhydrol (1.31 g, 5.36 mmol) and acetyl bromide (6.61 g,
53.8 mmol): 1.27 g (77%), slightly pink crystals; mp 72.8−73.9 °C. 1H
NMR spectra agreed with literature data.65 1H NMR (400 MHz,
CDCl3): δ 3.80 (s, 6 H), 6.32 (s, 1 H), 6.85−6.87 (m, 4 H), 7.37−7.39
(m, 4 H). 13C NMR (100 MHz, CDCl3): δ 55.5, 56.1, 113.9, 129.8,
133.7, 159.3. MS (+EI): m/z (%) = 227.2 (100) [M+ − Br]. Anal.
Calcd for C15H15BrO2 (307.2): C, 58.65; H, 4. 92. Found: C, 58.67; H,
4.90.
4,4′-Dimethylbenzhydryl Fluoride (2b). In a carefully dried
Schlenk flask, the fluorination agent (4.85 g, containing 7.28 mmol
AgF) was suspended in acetonitrile (12 mL). A solution of 1b (1.00 g,
3.63 mmol) in acetonitrile (5 mL) was added dropwise at 0 °C. After
stirring for 30 min, the solution was filtered, and the solvent was
removed under reduced pressure at room temperature. The crude
product was distilled in the vacuum (bp 160 °C/5.0 × 10−3 mbar) to
give 2b as a colorless oil: 0.61 g (78%). 1H NMR (400 MHz,
single-exponential equation Gt = G∞(1 − e−k t) + C. Each rate constant
1
was typically averaged from at least three kinetic runs. Only in two
cases (2f in 90A10W and 2b in 80E20W) measurements were
performed only once as these reactions were very slow. All solvolyses
reported in Table 2 and Table S2 in the Supporting Information were
performed at 25 °C.
Nucleophilic Reactivity of Fluoride in Different Solvents. Sol-
vents and Fluoride Sources. KF of p.a. grade, CsF (99.9%) and
tetrabutylammonium fluoride trihydrate (98%) were used as
fluoride sources for the kinetic experiments. Acetonitrile (extra
dry, water content <50 ppm) and methanol (99.8%) were used
without further purification for laser-flash experiments.
Commercially available absolute ethanol was freshly distilled
from sodium/diethyl phthalate. Doubly distilled water (Im-
pedance 18.2 Ω) was prepared with a water purification system.
Laser-Flash Photolysis. The measurements were performed in an
air-conditioned laboratory at 20 1 °C. The benzhydryl cations 3b−j
were generated by irradiating solutions (A266 nm ≈ 0.1−1.0) of the
precursor salts 4b−j with a 7 ns laser pulse (7 ns pulse width, 266 nm,
40−60 mJ/pulse). The benzhydryl cations 3b−h were generated from
the triphenylphosphonium salts 4b−h. For the photogeneration of 3i,j,
stock solutions of the tributylphosphonium salts 4i,j in acetonitrile
were prepared by mixing the appropriate amounts of benzhydrylium
tetrafluoroborates 3i,j and tributylphosphine. The system was
equipped with a fluorescence flow cell and a synchronized diaphragm
liquid pump that allowed complete replacing of the sample volume
between subsequent laser pulses. Kinetics were measured by following
the UV−vis absorbance decay of the benzhydryl cations at their
absorbance maxima. Averaged data obtained from ≥48 individual runs
were used for further evaluations. First-order rate constants (kobs) were
calculated by least-squares fitting of the absorbance data to a single
2
CD3CN): δ 2.33 (s, 6 H), 6.47 (d, JHF = 47.0 Hz, 1 H), 7.19−7.26
1
(m, 8 H). 13C NMR (100.6 MHz, CD3CN): δ 21.2, 95.2 (d, JCF
=
3
2
170.0 Hz), 127.2 (d, JCF = 6.0 Hz), 130.1, 138.5 (d, JCF = 22.1 Hz),
139.3 (d, 5JCF = 2.0 Hz). 19F NMR (376 MHz, CD3CN): δ −166.2 (d,
2JFH = 47 Hz). MS (+ EI): m/z (%) = 214.3 (42) [M+], 199.2 (100)
[C14H12F+]. Anal. Calcd for C15H15F (214.3): C, 84.08; H, 7.06.
Found: C, 84.27; H, 6.78.
4,4′-Dimethoxybenzhydryl Fluoride (2f). In a carefully dried
Schlenk flask, the fluorination agent (3.91 g, containing 5.86 mmol
AgF) was suspended in acetonitrile (5 mL). A solution of 1f (1.5 g,
4.88 mmol) in acetonitrile (10 mL) was added dropwise at 0 °C. After
stirring for 30 min, the solution was filtered, and the solvent was
removed under reduced pressure at room temperature. The crude
product was crystallized from CH2Cl2/n-pentane to give 2f as slightly
pink crystals: 0.60 g (50%); mp 72.8−73.9 °C. To avoid
decomposition, 2f was stored under argon at −35 °C. 1H NMR
2
(400 MHz, CD3CN): δ 3.78 (s, 6 H), 6.45 (d, JHF = 48.0 Hz, 1 H),
6.92−6.95 (m, 4 H), 7.26−7.30 (m, 4 H). 13C NMR (100 MHz,
1
3
CD3CN): δ 55.9, 94.9 (d, JCF = 169.0 Hz), 114.8, 128.9 (d, JCF
=
6.0 Hz), 133.5 (d, 2JCF = 23.0 Hz), 160.7 (d, 5JCF = 2.0 Hz). MS (+ EI):
+
m/z (%) = 246.1 (0.1) [M+], 228.1 (38) [C15H16O2 ], 227.1 (100)
[M+−F]. HRMS (+EI) Calcd. for C15H15FO2: 246.1056; Found:
246.1045.
Further benzhydryl fluorides were prepared in solution and directly
used for the solvolytic investigations without prior isolation (see below).
obst
exponential function At = A0 e−k + C. The second-order rate
3333
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
constants (k−1) were obtained using the slopes of the linear plots of
kobs against the fluoride ion concentration. First-order rate constants
(ks) were obtained from independent measurements with laser-
photolytically generated benzhydrylium ions in the absence of fluoride
ions.
(17) Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470−477.
(18) Furuya, T.; Strom, A. E.; Ritter, T. J. Am. Chem. Soc. 2009, 131,
1662−1663.
(19) Tang, P.; Furuya, T.; Ritter, T. J. Am. Chem. Soc. 2010, 132,
12150−12154.
Stopped-Flow Method. The kinetics of the reactions of F− with
3k,l were determined by using the stopped-flow method and isolated
benzhydrylium tetrafluoroborates as described previously.35 The decay
of their absorbance was monitored by UV−vis spectroscopy at the
absorption maxima of 3k,l. Pseudo-first-order rate constants kobs were
obtained from at least six runs at each fluoride concentration. The
absorbance-time curves were fitted to the single exponential function
(20) Watson, D. A.; Su, M.; Teverovskiy, G.; Zhang, Y.; García-
Fortanet, J.; Kinzel, T.; Buchwald, S. L. Science 2009, 325, 1661−1664.
(21) Lasne, M.-C.; Perrio, C.; Rouden, J.; Barre, L.; Roeda, D.; Dolle,
F.; Crouzel, C. Top. Curr. Chem. 2002, 222, 201−258.
(22) Miller, P. W.; Long, N. J.; Vilar, R.; Gee, A. D. Angew. Chem., Int.
Ed. 2008, 47, 8998−9033.
(23) Littich, R.; Scott, P. J. H. Angew. Chem., Int. Ed. 2012, 51, 1106−
1109.
(24) Schirrmacher, R.; Wangler, C.; Schirrmacher, E. Mini-Rev. Org.
Chem. 2007, 4, 317−329.
(25) Elander, N.; Jones, J. R.; Lu, S.-Y.; Stone-Elander, S. Chem. Soc.
Rev. 2000, 29, 239−249.
(26) Adams, D. J.; Clark, J. H. Chem. Soc. Rev. 1999, 28, 225−231.
(27) Ingold, C. K.; Ingold, E. H. J. Chem. Soc. 1928, 2249−2262.
(28) Cooper, K. A.; Hughes, E. D. J. Chem. Soc. 1937, 1183−1187.
(29) Swain, C. G.; Scott, C. B. J. Am. Chem. Soc. 1953, 75, 246−248.
(30) Swain, C. G.; Mosely, R. B. J. Am. Chem. Soc. 1955, 77, 3727−
3731.
obst
At = A0 e−k + C to yield the rate constants kobs (s−1). The second-
order rate constants k−1 (M−1 s−1) for the reactions of 3 with fluoride
ions were obtained from the slopes of linear plots of kobs versus the
fluoride concentrations.
ASSOCIATED CONTENT
* Supporting Information
■
S
Details for the kinetic procedures, product studies, and copies
of NMR spectra of 1b−f, 2b, and 2f. This material is available
free of charge via the Internet at http://pubs.acs.org.
(31) Coverdale, A. K.; Kohnstam, G. J. Chem. Soc. 1960, 3806−3811.
(32) Swain, C. G.; Knee, T. E. C.; MacLachlan, A. J. Am. Chem. Soc.
1960, 82, 6101−6104.
AUTHOR INFORMATION
Corresponding Author
*E-mail: herbert.mayr@cup.uni-muenchen.de.
Notes
■
(33) Toteva, M. M.; Richard, J. P. J. Am. Chem. Soc. 2002, 124,
9798−9805.
(34) Kevill, D. N.; D’Souza, M. J. J. Org. Chem. 2004, 69, 7044−7050.
(35) Swain, C. G.; Mosely, R. B.; Bown, D. E. J. Am. Chem. Soc. 1955,
77, 3731−3737.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(36) Landini, D.; Maia, A.; Rampoldi, A. J. Org. Chem. 1989, 54,
328−332.
We thank Dr. Markus Horn and Dr. Armin Ofial for assistance
during the preparation of the manuscript. We are grateful to the
Deutsche Forschungsgemeinschaft (Ma673/22-1) for financial
support.
(37) de Oliveira Baptista, M. J. V.; Widdowson, D. A. J. Chem. Soc.,
Perkin Trans. 1 1978, 295−298.
(38) Denegri, B.; Streiter, A.; Juric,
H. Chem.Eur. J. 2006, 12, 1648−1656.
(39) Denegri, B.; Ofial, A. R.; Juric, S.; Streiter, A.; Kronja, O.; Mayr,
́
S.; Ofial, A. R.; Kronja, O.; Mayr,
́
REFERENCES
■
H. Chem.Eur. J. 2006, 12, 1657−1666.
(40) Streidl, N.; Antipova, A.; Mayr, H. J. Org. Chem. 2009, 74,
7328−7334.
(1) Gribble, G. W. J. Chem. Educ. 2004, 81, 1441−1449.
(2) Gribble, G. W. The Handbook of Environmental Chemistry; Part N,
Organofluorines; Springer: Heidelberg, 2002.
(3) Vaillancourt, F. H.; Yeh, E.; Vosburg, D. A.; Garneau-Tsodikova,
S.; Walsh, C. T. Chem. Rev. 2006, 106, 3364−3378.
(4) O’Hagan, D.; B. Harper, D. J. Fluorine Chem. 1999, 100, 127−
133.
(41) Streidl, N.; Branzan, R.; Mayr, H. Eur. J. Org. Chem. 2010,
4205−4210.
(42) Streidl, N.; Denegri, B.; Kronja, O.; Mayr, H. Acc. Chem. Res.
2010, 43, 1537−1549.
(43) Matic,
6019−6024.
(44) Juric,
3854.
́
M.; Denegri, B.; Kronja, O. Eur. J. Org. Chem. 2010,
(5) Deng, H.; O’Hagan, D.; Schaffrath, C. Nat. Prod. Rep. 2004, 21,
773−784.
́
S.; Denegri, B.; Kronja, O. J. Org. Chem. 2010, 75, 3851−
(6) Xu, X.-H.; Yao, G.-M.; Li, Y.-M.; Lu, J.-H.; Lin, C.-J.; Wang, X.;
Kong, C.-H. J. Nat. Prod. 2003, 66, 285−288.
(45) Mayr, H.; Bug, T.; Gotta, M. F.; Hering, N.; Irrgang, B.; Janker,
B.; Kempf, B.; Loos, R.; Ofial, A. R.; Remennikov, G.; Schimmel, H.
J. Am. Chem. Soc. 2001, 123, 9500−9512.
́ ́
(7) Begue, J.-P.; Bonnet-Delpon, D. Bioorganic and Medicinal
Chemistry of Fluorine; John Wiley & Sons, Inc.: New York, 2007.
(8) Tressaud, A.; Haufe, G. Fluorine and Health: Molecular Imaging,
Biomedical Materials and Pharmaceuticals; Elsevier: Amsterdam, 2008.
(9) Maienfisch, P.; Hall, R. G. Chimia 2004, 58, 93−99.
(46) Ando, T.; Cork, D. G.; Fujita, M.; Kimura, T.; Tatsuno, T.
Chem. Lett. 1988, 1877−1878.
(47) Minegishi, S.; Loos, R.; Kobayashi, S.; Mayr, H. J. Am. Chem.
Soc. 2005, 127, 2641−2649.
(10) Ismail, F. M. D. J. Fluorine Chem. 2002, 118, 27−33.
(11) O’Hagan, D. J. Fluorine Chem. 2010, 131, 1071−1081.
(48) Ammer, J.; Baidya, M.; Kobayashi, S.; Mayr, H. J. Phys. Org.
Chem. 2010, 23, 1029−1035.
(12) Muller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881−
̈
1886.
(49) Sun, H.; DiMagno, S. G. J. Am. Chem. Soc. 2005, 127, 2050−
2051.
(13) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem. Soc.
Rev. 2008, 37, 320−330.
(50) For the photoheterolysis of quaternary phosphonium salts, see:
(a) Alonso, E. O.; Johnston, L. J.; Scaiano, J. C.; Toscano, V. G. J. Am.
Chem. Soc. 1990, 112, 1270−1271. (b) Alonso, E. O.; Johnston, L. J.;
Scaiano, J. C.; Toscano, V. G. Can. J. Chem. 1992, 70, 1784−1794.
(c) Imrie, C.; Modro, T. A.; Rohwer, E. R.; Wagener, C. C. P. J. Org.
Chem. 1993, 58, 5643−5649. (d) Imrie, C.; Modro, T. A.; Wagener,
C. C. P. J. Chem. Soc., Perkin Trans. 2 1994, 1379−1382.
(51) Minegishi, S.; Kobayashi, S.; Mayr, H. J. Am. Chem. Soc. 2004,
126, 5174−5181.
(14) Bondi, A. J. Phys. Org. Chem. 1964, 68, 441−451.
(15) Rozen, S. In Patai’s Chemistry of Functional Groups Supplement
D2, The chemistry of halides, pseudo-halides and azides Part 1; Patai, S.,
Rappoport, Z., Eds.; Wiley-Interscience, New York, 2009; pp
629−709.
́
(16) Carroll, M. A.; Dax, K.; Gouverneur, V.; Gree, R.; Haufe, G.;
Lequeux, T. P.; Miles, J. A. L.; Percy, J. M.; Roig, R.; Sai Krishna
Murthy, A.; Sandford, G.; Tardivel, R.; Tredwell, M.; Yamazaki, T. In
Science of Synthesis; Percy, J. M., Ed.; Thieme: Stuttgart, 2006; Vol. 34.
3334
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335

The Journal of Organic Chemistry
Article
(52) McClelland, R. A.; Kanagasabapathy, V. M.; Steenken, S. J. Am.
Chem. Soc. 1988, 110, 6913−6914.
(53) Pham, T. V.; McClelland, R. A. Can. J. Chem. 2001, 79, 1887−
1897.
(54) Minegishi, S.; Mayr, H. J. Am. Chem. Soc. 2003, 125, 286−295.
(55) Calculated by eq 1 using Ef = −3.44, Nf = 3.84, and sf = 0.96
from ref 42.
(56) Because of the low accuracy of this rate constant (see plot kobs vs
[F−] on p S22 of the Supporting Information) and the proximity of
the diffusion limit, this rate constant was not included in the
correlations of Figure 5.
(57) Mayr, H.; Ofial, A. R. Angew. Chem., Int. Ed. 2006, 45, 1844−
1854.
(58) Cowie, G. R.; Fitches, H. J. M.; Kohnstam, G. J. Chem. Soc.
1963, 1585−1593.
(59) Denegri, B.; Kronja, O. J. Phys. Org. Chem. 2009, 22, 495−503.
(60) Spencer, J. F.; Le Pla, M. Z. Anorg. Chem. 1910, 65, 10−15.
(61) Baidya, M.; Kobayashi, S.; Mayr, H. J. Am. Chem. Soc. 2010, 132,
4796−4805.
(62) Padmanaban, M.; Biju, A. T.; Glorius, F. Org. Lett. 2011, 13,
98−101.
(63) Smith, J.; Tedder, J. M. J. Chem. Res. Synop. 1988, 108−109.
(64) Shpanko, I. V.; Serebryakov, I. M.; Oleinik, N. M.; Panchenko,
B. V. Org. React. (Tartu) 1996, 30, 3−6.
(65) Carlier, P. R.; Zhao, H.; MacQuarrie-Hunter, S. L.; DeGuzman,
J. C.; Hsu, D. C. J. Am. Chem. Soc. 2006, 128, 15215−15220.
(66) Ohkawa, S.; Setoh, M.; Terashita, Z.-i. (Takeda Chemical
Industries, Ltd., Japan) Patent US6248766 B1, 2001.
3335
dx.doi.org/10.1021/jo300141z | J. Org. Chem. 2012, 77, 3325−3335