The lewis base-catalyzed silylation of alcohols-a mechanistic analysis

Article abstract of DOI:10.1021/jo5016568

Reaction rates for the base-catalyzed silylation of primary, secondary, and tertiary alcohols depend strongly on the choice of solvent and catalyst. The reactions are significantly faster in Lewis basic solvents such as dimethylformamide (DMF) compared with those in chloroform or dichloromethane (DCM). In DMF as the solvent, the reaction half-lives for the conversion of structurally similar primary, secondary, and tertiary alcohols vary in the ratio 404345:20232:1. The effects of added Lewis base catalysts such as 4-N,N-dimethylaminopyridine (DMAP) or 4-pyrrolidinopyridine (PPY) are much larger in apolar solvents than in DMF. The presence of an auxiliary base such as triethylamine is required in order to drive the reaction to full conversion.

Full text of DOI:10.1021/jo5016568

Article
pubs.acs.org/joc
The Lewis Base-Catalyzed Silylation of AlcoholsA Mechanistic
Analysis
Pascal Patschinski, Cong Zhang, and Hendrik Zipse*
Department of Chemistry, Ludwig-Maximilians-Universitat, 81377 Munchen, Germany
̈
̈
S
* Supporting Information
ABSTRACT: Reaction rates for the base-catalyzed silylation
of primary, secondary, and tertiary alcohols depend strongly
on the choice of solvent and catalyst. The reactions are
significantly faster in Lewis basic solvents such as dimethyl-
formamide (DMF) compared with those in chloroform or
dichloromethane (DCM). In DMF as the solvent, the reaction
half-lives for the conversion of structurally similar primary,
secondary, and tertiary alcohols vary in the ratio
404345:20232:1. The effects of added Lewis base catalysts
such as 4-N,N-dimethylaminopyridine (DMAP) or 4-pyrroli-
dinopyridine (PPY) are much larger in apolar solvents than in
DMF. The presence of an auxiliary base such as triethylamine
is required in order to drive the reaction to full conversion.
A number of alternative protocols have also been explored
INTRODUCTION
■
using different combinations of bases and/or solvents, such as
those by Hernandez [triethylamine (Et3N, 2)/4-N,N-dimethy-
laminopyridine (DMAP, 3a)],5 Chang (Et3N/1,1,3,3-tetrame-
thylguanidine and Et3N/DBU),6 Weiffen (18-crown-6/
K2CO3),7 Lombardo [i-Pr2NEt (4)],8 and Fuchs (18-crown-
6/KH).9 The first two of these studies employ catalyst systems
closely related to that used in the Corey procedure, as both
employ a nitrogen heterocycle as the actual catalytic base.
Despite this apparent similarity, the reaction catalyzed by
DMAP in apolar organic solvents shows higher selectivities in
the transformation of polyol substrates carrying primary and
secondary hydroxyl groups (Scheme 1b). As shown in Scheme
2 for the example of DMAP, catalytic Lewis bases are believed
to react with silyl chlorides to form silylpyridinium ion pairs,
whose subsequent reaction with the alcohol substrate yields the
silylether products together with the protonated (and thus
deactivated) pyridine base.10 Reactivation of the catalyst then
requires the action of an auxiliary base such as 2.
The silylation of hydroxyl groups represents one of the most
important protecting-group strategies in the manipulation of
polyfunctional molecules.1−3 The usefulness of this reaction
was demonstrated by Venkateswarlu and Corey in 1972 using
tert-butyldimethylsilyl chloride (TBSCl, 1) in DMF as the
solvent with imidazole as the base and catalyst for the
protection of secondary alcohols (Scheme 1a).4 In the same
Scheme 1. (a) Protection of a Secondary Alcohol Using
Corey’s Method;4 (b) Selectivity in the Silylation of
Unsymmetric 1,2-Diols5
This mechanism is practically identical to the consensus
mechanism for the pyridine-catalyzed acylation of alcohols
employing anhydrides as acylating reagents.11−13 For these
acylation reactions, more electron-rich (and thus more active)
pyridines have recently14−16 been developed through extension
of the DMAP structure, and here we explore the usefulness of
these catalysts for the silylation of primary, secondary, and
tertiary alcohols.
report it was also demonstrated that TBS ethers can be cleaved
effectively under mild conditions using tetra-N-butylammonium
fluoride in THF. This “Corey method” has since been applied
to a multitude of substrates containing (mainly) primary and
secondary hydroxyl groups, thus documenting its general
usefulness.
Received: July 22, 2014
Published: July 31, 2014
© 2014 American Chemical Society
8348
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
of silyl chloride 1 (Figure 3) and the appearance of silylated
alcohol 6a could be fitted using an effective second-order rate
law. The corresponding effective rate constant keff could then be
used to characterize the reaction in terms of its effective
reaction half-life t1/2 (see the Supporting Information for
details).
Scheme 2. Mechanism of the DMAP-Catalyzed Silylation of
Alcohols5
It was recently shown for acylation reactions that the
auxiliary base plays an important role in maintaining the catalyst
activity over many turnover cycles.11 This point was therefore
tested in the silylation reaction (Figure 1) with 4 mol % DMAP
in the presence of different amounts of auxiliary base. As can
readily be seen from the turnover curves in Figure 4, the rates
of silylation were practically identical for the reactions involving
2.2, 1.7, and 1.2 equiv (relative to alcohol 5a) of Et3N as the
auxiliary base. This implies that the auxiliary base is not directly
involved in the catalytic cycle but merely needed to regenerate
the catalyst. When too little auxiliary base is present for this
latter task, the reaction slows down dramatically after a certain
percentage of turnover (as is visible from the turnover curve for
0.7 equiv of Et3N). In the absence of auxiliary base, the reaction
is extremely slow and not easily analyzed in terms of a second-
order rate law. Thus, we conclude that silylation reactions such
as the one described in Figure 1 are best run in the presence of
1.2 equiv of the auxiliary base.
The effectiveness of Et3N (2) as an auxiliary base for the
silylation was subsequently explored by rerunning the bench-
mark reaction using other auxiliary bases such as trioctylamine
(TOA, 8), N,N-diisopropylethylamine (DIPEA, 4), 1,4-
diazabicyclo[2.2.2]octane (DABCO, 9), 1,8-bis-
(dimethylamino)naphthalene (Proton Sponge, 10), and N-
methylmorpholine (NMM, 11). The results compiled in Table
1 show nearly no variation in the reaction half-life among bases
with similar pKa values, such as DIPEA (30.7 min), TOA (30.5
min), and Et3N (30.2 min).
RESULTS AND DISCUSSION
■
Reaction with Primary Alcohol 5a. Initial experiments
were performed for the reaction of naphthalen-1-ylmethanol
(5a) with 1 (1.2 equiv) and 2 (1.2 equiv) as the auxiliary base
in CDCl3 using different nucleophilic catalysts, as depicted in
Figure 1. Dioxane was added as an internal standard. The
1
progress of the silylation reaction was monitored by H NMR
spectroscopy, and initial attempts were devoted to calibration
of the signals of alcohol 5a and its silyl ether 6a. The signal of
the CH2 group of the alcohol, however, was found to shift in an
unfavorable manner as the reaction progressed, making it
unsuitable for full kinetic analysis. The methyl group signals of
the silyl group were found to be more useful in this respect, and
a practical protocol was developed to monitor the reaction
progress to near completion (see the Supporting Information
1
The true advantage of TOA over the other two bases clearly
lies in the much better solubility of its ammonium chloride salt
in organic solvents. The reaction with NMM, an amine of lower
basicity, led to a significantly extended reaction half-life (127.7
min). Under these conditions, regeneration of DMAP (with
pKa = 9.7) is not effective anymore.18 Despite its low basicity,
DABCO (9) is quite efficient in its ability to promote the
silylation reaction, with t1/2 = 26.5 min. However, this latter
result more likely reflects its activity as a catalytic Lewis base
rather than its ability to regenerate protonated DMAP (in line
with earlier studies of the kinetic and thermodynamic basicities
for details). On closer inspection, all of the H NMR spectra
obtained under the reaction conditions indicated the presence
of small amounts of bis(silyl) ether 7 (Figure 2). In
experiments with defined amounts of added water, 7 was
rapidly generated from 2 equiv of silyl chloride 1 and 1 equiv of
water. These results show that the reactions in the following
mechanistic studies contain no more than 2% water. The
conversion of silyl chloride 1 follows an effective rate law
involving first-order behavior of all of the reactants and the
catalyst and zeroth-order behavior of the auxiliary base (see the
Supporting Information for full information). The conversion
Figure 1. Silylation of naphthalen-1-ylmethanol (5a) with TBSCl (1) and Et3N (2) in CDCl3 together with structures of all of the catalysts used
(3a−g).
8349
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
1
Figure 2. H NMR spectra of the benchmark reaction in CDCl3 using the primary alcohol 5a, 2 as the auxiliary base, and DMAP as the catalyst.
Table 1. Rate Data for the Silylation of Alcohol 5a Using
DMAP (4 mol %) as the Catalyst with Various Auxiliary
Bases (1.2 equiv) in CDCl3
Figure 3. Silylation of alcohol 5a with 1 (1.2 equiv) using 2 (1.2
equiv) as the base and DMAP (4 mol %) as the catalyst in CDCl3 as
1
monitored by H NMR spectroscopy.
a
b
c
keff in L·mol−1·s−1. Half-life in min. Only 7% conversion.
of DABCO).17 Why the reaction slows down somewhat with
Proton Sponge (10) as the auxiliary base (t1/2 = 37.4 min) is
not immediately evident from the aqueous-phase pKa values.
Reaction half-lives were determined for all of the catalysts
shown in Figure 1, including DMAP, PPY (3b), imidazole (3c),
and N-methylimidazole (3d) as well as the electron-rich
pyridines 3e, 3f, and 3g recently developed for acylation
reactions (Table 2).14,25 All of the benchmark experiments were
performed at a catalyst loading of 4 mol %. The least efficient
catalyst (t1/2 = 163.9 min) is imidazole (3c), which is used in
large excess in the original Corey procedure. This is
approximately 5 times slower than the reaction with DMAP
Figure 4. Turnover curves for the silylation of alcohol 5a catalyzed by
4 mol % DMAP in the presence of variable concentrations of Et3N (2)
in CDCl3 as monitored by H NMR spectroscopy.
1
8350
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
Table 2. Silyl Cation Affinities (SCAs) and Rate Data for the
Silylation of Alcohol 5a Using Different Lewis Base Catalysts
at 4 mol % Catalyst Loading
Figure 5. Correlations of effective rate constant keff vs catalyst
concentration for the silylation of alcohol 5a using DMAP, PPY, 3d,
3f, and 3g as catalysts.
Table 3. Silyl Cation Affinities (SCAs) and Effective Rate
Constants k′eff for the Silylations of Primary Alcohol 5a and
Secondary Alcohol 5b Using Different Catalysts
a
Silyl cation affinities at 298.15 K (in kJ/mol) were calculated at the
b
MP2(FC)/G3MP2large//MPW1K/6-31+G(d) level of theory. Sol-
vation energies were calculated at the PCM/UAHF/MPW1K/6-
c
d
31+G(d) level of theory for CHCl3. keff in L·mol−1·s−1. Half-life in
min.
(30.2 min) and about 18 times slower than the reaction with
the most efficient catalyst, 3g (t1/2 = 9.0 min).
For the reactions catalyzed by DMAP, PPY, 3d, 3f, and 3g,
rate measurements were repeated at different catalyst
concentrations, and a clear linear correlation between the rate
constant keff and the catalyst loading was observed in all cases
(Figure 5). As was previously found for other Lewis base-
catalyzed reactions such as the aza-Morita−Baylis−Hillman
reaction and in acylation reactions,13,26 this implies that only
one catalyst molecule participates in the rate-limiting step. The
slope of the correlation line (k′eff; see Table 3) reflects the
intrinsic catalytic efficiency of the catalyst, while the intercept
(b) represents the rate of the background reaction in the
absence of catalyst (eq 1).
a
[b]
Silyl cation affinities at 298.15 K (in kJ/mol) in CHCl3. k′eff in L·
mol−1·s−1.
Theoretical Evaluation of the Silyl Transfer Enthalpy.
In order to establish a possible link between the catalytic
efficiency and the Lewis basicity of the catalyst, the affinity of
each catalyst toward the tert-butyldimethylsilyl cation at 298.15
K was calculated at the MP2(FC)/G3MP2large//MPW1K/6-
31+G(d) level of theory (Table 2). Solvent effects in CHCl3
were computed at the PCM/UAHF/MPW1K/6-31+G(d) level
of theory. The silyl cation affinities (SCAs) were determined
relative to that for pyridine as the reference base using the
isodesmic silyl group transfer reaction shown in Figure 6. The
SCA of −24.7 kJ/mol for imidazole (3c) thus implies that the
tert-butyldimethylsilyl group attaches to imidazole 24.7 kJ/mol
more strongly than to pyridine. The calculations yielded the
lowest SCA values for the least effective catalysts 3c and 3d,
higher values for the more efficient catalysts DMAP and PPY,
′
keff = keff[cat] + b
(1)
Visual inspection of Figure 5 indicates that the rate of the
background reaction is quite small. This point was confirmed in
an experiment run without any catalyst, where a conversion
below 1% was determined after 36 h in CDCl3. This implies
that the auxiliary base Et3N (2) present in the reaction mixture
is not catalytically active.
8351
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
Figure 7. Correlations of rate constant keff vs catalyst concentration for
the silylation of alcohol 5b in CDCl3 using DMAP, PPY, 3d, 3f, and 3g
as catalysts.
Figure 6. Plot of silyl cation affinity (relative to the reference base
pyridine) vs log(keff) for the silylation of alcohol 5a with various
catalysts in CHCl3.
and the highest values for the most active catalysts, pyridines
3e, 3f, and 3g. Despite the fact that the correlation between
SCA values and reaction rates is not perfect for the group of
highly active catalysts, the trend of higher catalytic reactivity for
the more Lewis basic catalysts is in line with expectation for the
mechanism shown in Scheme 2 involving pre-equilibrium
formation of the silylated catalyst and subsequent rate-limiting
transfer of the silyl group to the substrate alcohol. A similar
correlation has recently been observed for the Lewis base-
catalyzed acylation of alcohols.15
Reaction with Secondary Alcohol 5b. Reaction rates for
the secondary alcohol 1-(naphthalen-1-yl)ethanol (5b) were
determined under conditions nearly identical to those for
primary alcohol 5a. As expected, the reaction rate for the
silylation of 5b is much lower than that for primary alcohol 5a.
With DMAP at a catalyst loading of 4 mol %, for example, the
half-life for the silylation of primary alcohol 5a is 30.2 min,
while that for the silylation of secondary alcohol 5b is
3166 min. This decrease of approximately 2 orders of
magnitude in the absolute reaction rate was also observed for
all of the other Lewis base catalysts investigated here (Table 3).
Kinetic measurements were again repeated at various catalyst
loadings, and linear correlations between keff and the catalyst
concentration were also observed for secondary alcohol 5b
(Figure 7).
The effective rate constants k′eff for the silylation of secondary
alcohol 5b extracted from the correlations in Figure 7 with the
aid of eq 1 indicate that pyridines 3f and 3g are again the most
effective catalysts, followed by PPY, 3d, and DMAP. However,
the difference between the most and least efficient catalysts (3g
vs DMAP) amounts to a factor of only 3.1. This range is
significantly smaller than that observed for the silylation of
primary alcohol 5a, which implies that the ratio of rate
constants k′eff for primary and secondary alcohols increases
systematically with the calculated Lewis basicity of the catalyst
(Table 3).
The original Corey procedure employs a reaction temper-
ature of 35 °C in order to increase both the solubilities of the
reaction components and the reaction rate. The silylation
reaction of secondary alcohol 5b was therefore repeated using
catalyst 3f at 30 mol % catalyst loading at different reaction
temperatures in CDCl3 (see the Supporting Information for
more information). As expected, the reaction rate increased for
the benchmark reaction chosen here, providing reaction half-
lives of 162.7 min at 23 °C and 129.1 min at 45 °C. Formal
analysis of the moderately higher reaction rates at higher
temperatures using an Eyring plot implies an activation
enthalpy of ΔH⧧ = +5.8 kJ mol−1 and an activation entropy
of ΔS⧧ = −269.7 J K−1 mol−1. The value obtained for the
activation enthalpy is quite small for a reaction in solution,
whereas the negative value obtained for the activation entropy
is typical for an effective third-order reaction. Similar results of
ΔH⧧ = +12.8 kJ mol−1 and ΔS⧧ = −240 J mol−1 K−1 have
recently been determined for the PPY-catalyzed isobutyrilation
of secondary alcohol 5b.27 Even though the Eyring plot is based
on only three data points, the remarkable similarity of the
activation parameters for the Lewis base-catalyzed silylation and
acylation of alcohol 5b points to an associative reaction
mechanism in both cases.
Again with respect to the Corey procedure mentioned in the
Introduction, we note that silylation reactions are often
performed in polar aprotic solvents such as dimethylformamide
(DMF) or dimethyl sulfoxide (DMSO). The influence of
solvent polarity on the reaction rate was therefore studied for
the silylation of secondary alcohol 5b with TBSCl (1) in
CD2Cl2, CDCl3, and DMF-d7 (Figure 8). With catalyst 3g at
30 mol % loading and Et3N (2) as the auxiliary base (1.2
equiv), the reaction was first studied in CDCl3 and found to
proceed with a half-life of 176.3 min. Repeating the reaction in
CD2Cl2 under otherwise identical conditions yielded a slightly
faster reaction with a half-life of 115.4 min. In contrast, the
reaction in DMF-d7 was found to be so much faster that
accurate rate data could not be determined under these
conditions. Omission of the catalyst as well as the auxiliary base
eventually allowed accurate measurements but led to only 80%
conversion. With Et3N as the auxiliary base, full conversion was
observed, and a half-life of 6.7 min was determined for DMF-d7.
Other solvents such as THF, acetone, and acetonitrile were also
tested, but full analysis was impeded by formation of
inhomogeneous reaction mixtures (most likely due to
precipitation of Et3NH+Cl−). The significant increase in the
reaction rate in DMF-d7 compared with the two halogenated
solvents cannot be rationalized with common solvent
parameters such as ET30 values or Gutman donor numbers
8352
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
Figure 8. Reactions of secondary alcohol 5b with TBSCl (1) in
different solvents. The reactions in CDCl3 and CD2Cl2 involved 30
mol % catalyst 3g and Et3N (1.2 equiv) as the auxiliary base, while the
reaction in DMF-d7 was performed without any catalyst.
and may better be understood in terms of the direct
involvement of DMF-d7 as a Lewis base catalyst.28
In order to detect possible transient intermediates formed
between silyl chloride 1 and Lewis base catalysts in reactions
run in CDCl3 or adducts with DMF-d7 run in this latter solvent,
the reaction progress was monitored using 29Si NMR
spectroscopy. Together with the results of numerous control
experiments, the outcome of the 29Si NMR measurements can
be summarized as follows (Table 4 and Figure 9): (i) Transient
Table 4. 29Si NMR Shifts (in ppm) of Several Compounds in
CDCl3 and DMF-d7 in Comparison with Calculated 29Si
Shifts
a
entry
TBSCl (1)
CDCl3
36.10
DMF-d7
29Si calcd
34.97
37.11, 10.21
37.11, 10.21
37.14, 18.32
41.88, 10.21
10.21
TBSCl (1) + Et3N
TBSCl (1) + 5b
1 + AgSbF6 (1 equiv)
1 + AgSbF6 (0.5 equiv)
6a
20.58
18.43
12.18
9.91
20.30
20.94
20.22
14.41
11.31
18.56
17.19
13.95
42.43
38.39
34.89
49.29
Figure 9. Comparison of 29Si NMR results in CDCl3 and DMF-d7
with gas-phase calculations at the DF-LMP2/IGLO-III//MPW1K/6-
31+G(d) level of theory.32
6b
18.48
6c
12.06
7
10.22
12
20.34
20.39
14.00
see the Supporting Information), and bis(silyl) ether 7]. Even
mixtures of equimolar amounts of silyl chloride 1 and the Lewis
base DMAP or 3g did not lead to new signals in the 29Si NMR
spectrum. However, ion pair intermediates i1 and i2 were
generated in CDCl3 through the reaction of DMAP or 3g with
the more reactive silyl triflate 1b. The new resonances observed
at +33.25 ppm (for DMAP adduct i1) and +32.16 ppm (for 3g
adduct i2) can, together with results from 1H, 13C, and NOESY
measurements, be assigned to the ion pair structures shown in
Figure 9.29,30 These can also be located as true minima in gas-
phase geometry optimizations at the MPW1K/6-31+G(d)
level. However, the 29Si chemical shifts predicted for i1 and i2
using a recently developed32 theoretical protocol are system-
atically shifted downfield by several parts per million.
12b
12 + DMF (12c)
TBSOTf (1b)
TBSOTf + 3a (i1)
TBSOTf + 3g (i2)
TBSOTf + DMF (i3)
13.98, 10.25
41.97, 10.38
43.71, 9.9
33.25, 9.9
32.16, 9.8
44.06, 9.9
a
Geometries were obtained at the MPW1K/6-31+G(d) level of
theory. 29Si NMR shifts were calculated at the DF-LMP2/IGLO-III//
MPW1K/6-31+G(d) level of theory.
ion pair intermediates formed between a Lewis base catalyst
such as DMAP or pyridine 3g and silyl chloride 1 could not be
detected in CDCl3 under the conditions of the benchmark
reactions. Aside from silyl chloride 1 and silyl ether products
6a−c, the only other detectable species under these conditions
were small amounts of hydrolysis products [silanol 12, which
could be isolated and crystallized as hydrate 12b (for details,
(ii) Transient ion pair intermediates formed between DMF
and silyl chloride 1 cannot be detected in DMF-d7 under the
conditions of the benchmark reaction shown in Figure 8. Aside
from silyl chloride 1 and silyl ether product 6b, the only other
8353
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
detectable species were again the hydrolysis products 12 and 7.
Differences between 29Si chemical shifts in CDCl3 and DMF-d7
are generally quite small unless specific solute−solvent
interactions become important. This is the case for silanol 12,
whose 29Si resonance moved from +20.4 ppm in CDCl3 to
+14.0 ppm in DMF-d7 as a result of hydrogen-bonding
interactions. This shift of 6 ppm was readily reproduced by 29Si
NMR shift calculations for silanol 12 and its DMF-bound
complex 12c (Figure 9).
However, the 29Si NMR signal for silyl chloride 1 in DMF-d7
at +37.1 ppm changed substantially upon addition of AgSbF6 as
a reagent for the precipitation of AgCl. The new resonance at
+41.9 ppm obtained under these conditions was practically
identical to that obtained for the more reactive silyl triflate 1b
in DMF-d7 and can be attributed to ion pair i3 (with either
SbF6 or triflate as the counterion; Figure 9).29−31 Ion pair i3
(with triflate as the counterion) could again be located as a true
minimum in gas-phase geometry optimizations at the
MPW1K/6-31+G(d) level. As already observed for ion pairs
i1 and i2, the calculated 29Si NMR signal predicted for i3 is
shifted downfield by several parts per million relative to that
observed in DMF-d7 solution. Attempts to identify ion pair
intermediate i3 also in CDCl3 solution revealed an interesting
temperature effect: while the 29Si NMR signal for i3 in DMF-d7
solution remained effectively unchanged over the temperature
range from −50 to +70 °C, a mixture of silyl triflate 1b and
DMF (1.5:1 ratio) in CDCl3 showed signals for i3 (+45.2 ppm)
only at −50 °C. This signal vanished upon warming to room
temperature, in line with expectations for the reversible
formation of ion pair i3. Complementary observations can be
Figure 10. Kinetic measurements on 5b in DMF-d7 (top) with no
catalyst and no auxiliary base and (bottom) with the addition of Et3N
(2) after 60 min of reaction.
using identical conditions as shown before in DMF-d7 with 1.2
equiv of Et3N as the auxiliary base and no catalyst. However,
the reaction was very slow under these conditions: while full
conversion was reached after several minutes for primary and
secondary alcohols 5a and 5b, the first product signals for
tertiary silyl ether 6c were observed only after several hours of
reaction time. On the basis of the data collected up to 55%
conversion, one can predict a reaction half-life of t1/2 = 109 029
min (or approximately 75 days) for tertiary alcohol 5c (Figure
11). Repeating the reaction in the presence of catalyst 3g (30
1
made in the H NMR spectrum, where the broad signals
observed for DMF at room temperature at +8.55 and +3.46
ppm split into two sharp singlets at −50 °C.
(iii) Even though all of the 29Si NMR evidence points to
tetracoordinate silicon intermediates only, one should not
dismiss the possibility of pentacoordinate silicon species or true
silyl cation intermediates prematurely.31 A pentacoordinate
isomer of ion pair i1 (termed i4) could actually be identified as
a local minimum on the MPW1K-D2/6-31+G(d) potential
energy surface and was found to be located +42.8 kJ/mol
higher compared with i1.33 The calculated 29Si chemical shift
for i4 amounts to −32.24 ppm, which is in line with predictions
for other pentacoordinate silicon species.34 Signals in this range
of the 29Si NMR spectrum were not detected in any of the
measurements performed in this study, and we therefore
exclude the formation of pentacoordinate intermediates in the
reaction of DMAP with silyl triflate 1b.
The auxiliary base plays an important role during the reaction
in DMF. While the reaction reaches a plateau at 80%
conversion in pure DMF, addition of 1.2 equiv of Et3N
(relative to alcohol 5b) leads the reaction to full conversion.
For the experiment in pure DMF-d7, the 29Si NMR spectrum
shows signals of starting materials as well as product (Figure 10,
top). After addition of Et3N, only the product signal was
observed in 29Si NMR measurements.
These results imply that the reaction is inhibited by hydrogen
chloride generated during the reaction process without auxiliary
base and can resume when the acid is removed from the
reaction mixture. The major role of imidazole in the Corey
procedure described in the Introduction therefore seems to be
more that of an auxiliary base (rather than that of a catalyst).
Reaction with Tertiary Alcohols. Finally, reaction rates
were determined for silylation of alcohol 5c with TBSCl (1)
Figure 11. Conversion vs time plots for alcohols 5a, 5b, and 5c in
DMF-d7 with Et3N (1.2 equiv) as the auxiliary base and without any
catalyst.
mol %) under otherwise identical conditions led to no
acceleration of turnover, which again supports the Lewis basic
solvent DMF-d7 as the only catalytically active species under
these conditions (see the Supporting Information for further
information). This implies that the silyl chloride reagent 1 used
here is intrinsically not reactive enough to turn over tertiary
substrates in a synthetically meaningful way. However, the
limited kinetic data available for alcohols 5a, 5b, and 5c can be
combined to extract relative reactivities of 404345:20232:1 for
these three substrates under otherwise identical conditions
(DMF-d7, rt, Et3N as the auxiliary base, no catalyst). This
implies that the reactivity difference between secondary alcohol
8354
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
5b and tertiary alcohol 5c (a factor of 20 232) is much larger
than the reactivity difference between primary alcohol 5a and
secondary alcohol 5b (a factor of 20.0).
amine bases such as Et3N, TOA, and DIPEA. A moderate
increase in the reaction temperature is helpful to solve solubility
issues for certain substrates but otherwise leads to only
moderate rate enhancements.
CONCLUSION
■
Among the various factors influencing the silylation of primary
and secondary alcohols, the choice of solvent and catalyst are
clearly the most important (and interdependent). In DMF as
the solvent there is actually no need for added catalysts because
of the high catalytic activity of this solvent alone. Despite the
fact that direct detection of silylated DMF intermediates by 29Si
NMR spectroscopy has not been successful, the formation of
these intermediates from silyl chlorides and DMF and their
subsequent reaction with substrate alcohols remains the best
mechanistic hypothesis. The reactions are much slower in
apolar organic solvents such as dichloromethane (DCM) and
chloroform, even in the presence of highly active catalysts. For
the purpose of preparing silyl ethers of primary and secondary
alcohols with the known silyl chloride reagents, the established
Corey procedure thus still provides the most rapid and
economical means. However, should the selective trans-
formation of primary over secondary alcohols be the goal, the
conclusions will be somewhat different, as already pointed out
previously.5 The reactivity difference between primary and
secondary alcohols amounts to 20.0 under DMF conditions,
compared with 51 for N-methylimidazole (3d) in CDCl3 and
120−145 for electron-rich pyridines such as 3g and 3f in
CDCl3. When these ratios are plotted against the silyl cation
affinities of the respective Lewis bases as in Figure 12, it
EXPERIMENTAL SECTION
■
General Methods. All air- and water-sensitive manipulations were
carried out under a nitrogen atmosphere using standard Schlenk
techniques. Calibrated flasks for kinetic measurements were dried in
the oven at 120 °C for at least 12 h prior to use and then assembled
quickly while still hot, cooled under a nitrogen stream, and sealed with
a rubber septum. Commercial chemicals were of reagent grade and
were used as received, unless otherwise noted. CDCl3 was refluxed for
at least 1 h over CaH2 and subsequently distilled. H and 13C NMR
1
spectra were recorded at room temperature. All 1H chemical shifts are
reported in parts per million (δ) relative to CDCl3 (δ 7.26); 13C
chemical shifts are reported in parts per million (δ) relative to CDCl3
(δ 77.16). 1H NMR kinetic data were measured on a 200 MHz
spectrometer at 23 °C. HRMS spectra (ESI-MS or EI-MS) were
obtained using an FT instrument. For all of the kinetic measurements
with reaction times longer than 24 h, the reaction mixtures were
mechanically shaken at room temperature. For each reaction, the
rotation speed was set at 480 turns/min. Analytical TLC was carried
out using aluminum sheets coated with silica gel 60 F254. All other
chemicals were purchased from commercial suppliers at the highest
available grade, stored over orange gel in a desiccator, and used
without any further purification.
General Procedure I for Benchmark Reactions of 5a with 2
mol %/3 mol %/4 mol % Catalyst. A 0.2 mL aliquot from 5.0 mL
of stock solution I [TBSCl (542 mg, 3.6 mmol), dioxane (0.088 g,
0.086 mL)], 0.2 mL from 5 mL of stock solution II [5a (475 mg, 3.0
mmol), Et3N (364 mg, 0.50 mL)], and 0.2 mL from 5 mL of stock
solution III (0.06/0.09/0.12 mmol of catalyst) were mixed in an NMR
tube and sealed.
General Procedure II for Benchmark Reactions of 5a with
0.25 mol %/0.5 mol %/1 mol % Catalyst. A 0.2 mL aliquot from
5.0 mL of stock solution I [TBSCl (542 mg, 3.6 mmol), dioxane
(0.088 g, 0.086 mL)], 0.2 mL from 5 mL of stock solution II [5a (475
mg, 3.0 mmol), Et3N (364 mg, 0.50 mL)], and 0.2 mL from 10 mL of
stock solution III (0.015/0.030/0.060 mmol of catalyst) were mixed in
an NMR tube and flame-sealed.
General Procedure III for Benchmark Reactions of 5b with 4
mol %/10 mol %/20 mol %/30 mol % Catalyst. A 0.2 mL aliquot
from 5.0 mL of stock solution I [TBSCl (1.54 mg, 3.6 mmol), dioxane
(0.088 g, 0.086 mL)], 0.2 mL from 5 mL of stock solution II [5b (517
mg, 3.0 mmol), Et3N (2.36 mg, 0.50 mL)], and 0.2 mL of 2 mL stock
solution III (0.048/0.12/0.24/0.36 mmol of catalyst) were mixed in an
NMR tube and flame-sealed.
General Procedure IV for Temperature-Dependent Reac-
tions of 5b with 30 mol % Catalyst 3f. : A 0.2 mL aliquot from 5.0
mL of stock solution I [TBSCl (1.54 mg, 3.6 mmol), dioxane (0.088 g,
0.086 mL)], 0.2 mL from 5 mL of stock solution II [5b (517 mg, 3.0
mmol), Et3N (2.36 mg, 0.50 mL)], and 0.2 mL from 2 mL of stock
solution III (0.36 mmol of catalyst 3f) were mixed in an NMR tube
and flame-sealed. During the NMR measurement the temperature was
set to the temperature of choice.
Naphthalen-1-ylmethanol (5a). NaBH4 (0.567 g, 15 mmol, 0.5
equiv) was dissolved in 100 mL of THF, and the solution was cooled
to −10 °C. 1-Naphthaldehyde (4.68 g, 30 mmol, 4.07 mL, 1.0 equiv)
was dissolved in 50 mL of THF, and this solution was added dropwise
to the reaction mixture. The reaction mixture was allowed to stir for 30
min at rt, and the reaction process was monitored by TLC. The
reaction was quenched by addition of 2 M HCl until no H2 appeared.
The reaction mixture was extracted three times with DCM (20 mL)
and washed with brine (20 mL). The combined organic phases were
dried over MgSO4, and the solvent was removed under reduced
pressure. Column chromatography on silica (ihexane/EtOAc, 4:1) led
to a white solid product 5a in 95% yield (4.50 g). 1H NMR (300 MHz,
CDCl3): δ 2.67 (bs, 1H, OH), 5.05 (s, 2H, CH2), 7.40−7.61 (m, 4H),
Figure 12. Calculated silyl cation affinities vs selectivity ratios of
chosen catalysts. The ratio for DMF was determined from the
measurements in DMF as the solvent.
becomes apparent that the stability of the Lewis base−silyl
cation adducts are at least partially responsible for the observed
selectivities: the least stable intermediate (silylated DMF)
provides the lowest selectivities, while the much more stable
silylpyridinium intermediates all yield selectivities greater than
100. The choice of auxiliary base or reaction temperature
appears to be of minor importance in comparison.
The auxiliary base is needed in all of the protocols to
neutralize the generated acid and guide the reaction to full
conversion. In protocols employing pyridine catalysts, the
auxiliary base is also needed to avoid catalyst deactivation
through protonation. This role can be fulfilled by a variety of
8355
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
7.81−7.87 (m, 1H), 7.88−7.96 (m, 1H), 8.03−8.14 (m, 1H). 13C
NMR (75 MHz, CDCl3): δ 63.4, 123.7, 125.3, 125.9, 126.3, 128.5,
128.7, 131.3, 133.8, 136.3. MS (EI) m/z (%): 158.1 ([M + H]+, 83),
141.1 ([M − OH]+, 20), 129.2 ([M − CH2OH]+, 100). HRMS (EI):
C11H10O requires 158.0732 g/mol, found 158.0726 g/mol.
Si(CH2)2), 0.91 (s, 9H, SitBu), 1.89 (s, 6H, (CH3)2), 7.36−7.53 (m,
3H), 7.55 (dd, J = 7.3, 1.2 Hz, 1H), 7.74−7.77 (m, 1H), 7.84−7.87
(m, 1H), 8.82−8.85 (m, 1H). 13C NMR (75 MHz, CDCl3): δ −2.2,
18.6, 26.3, 32.7, 76.4, 122.5, 124.8, 124.9, 125.1, 128.4, 128.8, 128.9,
131.3, 134.9, 144.2. 29Si NMR (80 MHz, CDCl3): δ 12.18. MS (EI)
m/z (%): 285.17 ([M − CH3], 2), 185.08 ([M − TBS], 10), 169.09
([M − OTBS], 13), 127.06 ([M − Naph], 3), 75.09 ([C6H3], 100),
57.07 ([M − tBu], 5). HRMS (EI): C18H25OSi (6c −CH3) requires
285.1675, found 285.1662.
1,3-Di-tert-butyl-1,1,3,3-tetramethyldisiloxane (7). 1 (2.00 g,
13.4 mmol) was dissolved in 5 mL of acetonitrile and 1 mL of H2O.
After the addition of KI (2.49 g, 15.0 mmol), the mixture was stirred at
rt for 12 h. The upper layer was pipetted off and distilled (80 °C, 15
mbar; oil bath: 100 °C) affording 0.95 g of 7 (3.85 mmol, 28.7%) as a
colorless liquid. 1H NMR (300 MHz, CDCl3): δ 0.85 (s, 18H,
SiCH3tBu), 0.00 (s, 12H, (CH3)2SitBu). 13C NMR (75 MHz, CDCl3):
δ 25.7, 18.1, −3.0. 29Si NMR (80 MHz, CDCl3): δ 9.91. MS (EI) m/z
(%): 246.02 ([M+], 0.02), 189.11 ([M+ − tBu], 2.8), 147.05 (100),
73.04 ([M+ − tBu(CH3)2Si], 1.9). HRMS (EI): C12H30Si2O requires
246.1835, found 246.1828.
2-(Naphthalen-1-yl)propan-2-ol (5c). Magnesium (2.38 g, 100
mmol) and LiCl (2.08 g, 50 mmol) were dissolved in 150 mL of THF,
and a little bit of iodine was added. The reaction mixture was allowed
to stir for 15 min. 1-Bromonaphthalene (8.22 g, 40 mmol) was added
slowly, and the mixture was stirred for 30 min and refluxed for 1 h.
The reaction mixture was cooled to rt, and acetone (4.64 g, 80 mmol)
was added slowly. After 1 h of stirring at rt, the mixture was refluxed
for 3 h. The reaction was quenched by addition of saturated NH4Cl
under ice cooling. The reaction mixture was extracted three times with
DCM (50 mL) and washed with brine (20 mL). The combined
organic phases were dried over MgSO4, and the solvent was removed
under reduced pressure. Recrystallization from ihexane led to a white
solid product 5c in 82% yield (6.1 g). 1H NMR (300 MHz, CDCl3): δ
1.89 (s, 6H), 7.36−7.74 (m, 4H), 7.80 (d, J = 8.1 Hz, 1H), 7.85−7.98
(m, 1H), 8.76−9.01 (m, 1H). 13C NMR (75 MHz, CDCl3): δ 31.6,
74.1, 122.7, 124.8, 125.2, 125.3, 127.4, 128.6, 129.1, 131.0, 134.9,
143.4. MS (GC/EI) m/z (%): 186.23 ([M + H]+, 54), 171.20 ([M −
CH3]+, 100), 153.19 ([M − CH3 − CH3]+, 32), 128.16 ([M −
(CH3)2COH]+, 22). HRMS (GC/EI): C13H14O requires 186.1045 g/
mol, found 186.1036 g/mol.
tert-Butyldimethyl(naphthalen-1-ylmethoxy)silane (6a).
Naphthalen-1-ylmethanol (0.32 g, 2 mmol) and 1 (0.36 g, 2.4
mmol) were dissolved in 15 mL of DCM, and Et3N (0.33 mL, 0.24 g,
2.4 mmol) was added. After addition of DMAP (0.01 g, 0.08 mmol),
the reaction mixture was stirred at rt for 12 h. The mixture was washed
with sat. aq. NH4Cl solution and extracted with DCM (3 × 5 mL), and
the combined organic phases were dried over MgSO4. The solvent was
removed under reduced pressure. Column chromatography (ihexane/
DCM, 4:1) yielded 0.44 g of 6a (1.62 mmol, 82%) as a colorless oil.
1H NMR (300 MHz, CDCl3): δ 0.14 (s, 6H, Si(CH2)2), 0.97 (s, 9H,
SitBu), 5.22 (s, 2H, CH2), 7.47−7.59 (m, 3H), 7.59−7.61 (m, 1H),
7.79 (d, J = 8.1 Hz, 1H), 7.86−7.93 (m, 1H), 8.00−8.05 (m, 1H). 13C
NMR (75 MHz, CDCl3): δ −5.2, 18.5, 26.0, 29.2, 63.4, 123.3, 123.8,
125.45, 125.51, 125.8, 127.5, 128.6. 29Si NMR (80 MHz, CDCl3): δ
20.58. MS (EI) m/z (%): 272.17 ([M], 0.6), 215.09 ([M − tBu]+, 72),
141.07 ([M − OTBS]+, 100) 115.05 ([TBS], 13). HRMS (EI):
C17H24OSi requires 272.1596, found 272.1590.
Di-tert-butyldimethylsilanol Hydrate (12b). Potassium hydrox-
ide (0.18 g, 2.4 mmol) was dissolved 1 mL of H2O and 0.25 mL of
methanol, and the solution was stirred at 0 °C. 1 (0.38 g, 2.5 mmol)
was dissolved in 2 mL of diethyl ether, and this solution was slowly
added to the reaction mixture. The reaction mixture was stirred for 10
min. The product was extracted three times with diethyl ether, and the
solvent was removed under reduced pressure. Slow removal of the
1
solvent led to the desired crystallized product. H NMR (300 MHz,
CDCl3): δ 1.52 (s, 2H, H2O), 0.89 (s, 18H, SiCH3tBu), 0.07 (s, 12H,
(CH3)2SitBu). 13C NMR (75 MHz, CDCl3): δ 25.8, 17.9, −3.6. 29Si
NMR (80 MHz, CDCl3): δ 20.39.
ASSOCIATED CONTENT
* Supporting Information
■
S
Explanation of methods, theoretical calculations, time con-
version plots, NMR spectra, and the CIF file for compound
12b. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: zipse@cup.uni-muenchen.de
■
tert-Butyldimethyl(1-(naphthalen-1-yl)ethoxy)silane (6b). 1-
(Naphthalen-1-yl)ethanol (0.35 g, 2 mmol) and 1 (0.36 g, 2.4 mmol)
were dissolved in 15 mL of DCM, and triethylamine (0.33 mL, 0.24 g,
2.4 mmol) was added. After addition of DMAP (0.01 g, 0.08 mmol),
the reaction mixture was stirred at rt for 48 h. The mixture was washed
with sat. aq. NH4Cl solution and extracted with DCM (3 × 5 mL), and
the combined organic phases were dried over MgSO4. The solvent was
removed under reduced pressure. Column chromatography (ihexane/
DCM, 4:1) yielded 0.46 g of 6b (1.52 mmol, 76%) as a colorless oil.
1H NMR (300 MHz, CDCl3): δ −0.01 (s, 3H, SiCH3tBu), 0.10 (s, 3H,
Notes
The authors declare no competing financial interest.
REFERENCES
■
(1) (a) Wuts, P. G. M.; Greene, T. W. Greene’s Protective Groups in
Organic Synthesis, 4th ed.; John Wiley & Sons: Hoboken, NJ, 2006.
(b) Kocienski, P. J. Protecting Groups, 3rd ed.; Thieme: Stuttgart,
Germany, 2005.
SiCH3tBu), 0.95 (s, 9H, SitBu), 1.60 (d, J = 6.4 Hz, 3H, CH3CHOR),
5.61 (q, J = 6.6 Hz, 1H, CH), 7.44−7.55 (m, 3H), 7.67−7.78 (m, 2H),
7.85−7.91 (m, 1H), 8.11 (d, J = 7.3 Hz, 1H). 13C NMR (75 MHz,
CDCl3): δ −4.8, −4.9, 18.3, 25.9, 26.6, 68.5, 122.7, 123.3, 125.2, 125.5,
125.6, 127.2, 128.8, 129.9. 29Si NMR (80 MHz, CDCl3): δ 18.43. MS
(EI) m/z (%): 215.16 (6), 155.17 ([M − C6H15OSi], 23), 141.15
(33), 115.14 ([C6H15Si], 13), 76.09 ([C6H4], 27), 75.09 ([C6H3],
100). HRMS (EI): C18H26OSi requires 286.1753, found 286.1744.
tert-Butyldimethyl((2-(naphthalen-1-yl)propan-2-yl)oxy)-
silane (6c). 2-(Naphthalen-1-yl)propan-2-ol (0.37 g, 2 mmol) and
imidazole (1.36 g, 20 mmol) were dissolved in 10 mL of DMF, and 1
(1.51 g, 10 mmol) was slowly added. The reaction mixture was stirred
at rt for 12 h and at 70 °C for 7 days. The mixture was washed with
sat. aq. NH4Cl solution, extracted with DCM (3 × 10 mL), and
washed once with brine, and the combined organic phases were dried
over MgSO4. The solvent was removed under reduced pressure.
Column chromatography (ihexane) yielded 0.3 g of 6c (1.00 mmol,
50%) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ −0.14 (s, 6H,
(2) White, J. D.; Carter, R. G. Silyl Ethers. In Science of Synthesis;
Fleming, I., Ed.; Thieme: Stuttgart, Germany, 2002.
(3) Crouch, D. Synth. Commun. 2013, 43, 2265−2279.
(4) Venkateswarlu, A.; Corey, E. J. J. Am. Chem. Soc. 1972, 94, 6190−
6191.
(5) Chaudhary, S. K.; Hernandez, O. Tetrahedron Lett. 1979, 20, 99−
102.
(6) Kim, S.; Chang, H. Synth. Commun. 1984, 14, 899−904. Kim, S.;
Chang, H. Bull. Chem. Soc. Jpn. 1985, 58, 3669−3670.
(7) Lissel, M.; Weiffen, J. Synth. Commun. 1981, 11, 545−549.
(8) Lombardo, L. Tetrahedron Lett. 1984, 25, 227−228.
(9) Braish, T. F.; Fuchs, P. L. Synth. Commun. 1986, 16, 111−115.
(10) (a) Akiba, K.-y.; Iseki, Y.; Wada, M. Bull. Chem. Soc. Jpn. 1884,
57, 1994−1999. (b) Akiba, K.-y.; Iseki, Y.; Wada, M. Tetrahedron Lett.
1982, 23, 3935−3936.
(11) (a) Held, I.; Larionov, E.; Bozler, C.; Wagner, F.; Zipse, H.
Synthesis 2009, 2267−2277. (b) Held, I.; Villinger, A.; Zipse, H.
8356
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357

The Journal of Organic Chemistry
Article
Synthesis 2005, 1425−1430. (c) Wei, Y.; Held, I.; Zipse, H. Org.
Biomol. Chem. 2006, 4, 4223−4230.
(12) (a) Muller, C. E.; Schreiner, P. R. Angew. Chem., Int. Ed. 2011,
̈
50, 6012−6042. (b) Vedejs, E.; Jure, M. Angew. Chem., Int. Ed. 2005,
44, 3974−4001. (c) Wurz, R. P. Chem. Rev. 2007, 107, 5570−5595.
(d) Spivey, A. C.; Arseniyadis, S. Top. Curr. Chem. 2009, 291, 233−
280.
(13) Xu, S.; Held, I.; Kempf, B.; Mayr, H.; Steglich, W.; Zipse, H.
Chem.Eur. J. 2005, 11, 4751−4757.
(14) Heinrich, M. R.; Klisa, H. S.; Mayr, H.; Steglich, W.; Zipse, H.
Angew. Chem., Int. Ed. 2003, 42, 4826−4828.
(15) Larionov, E.; Achrainer, F.; Humin, J.; Zipse, H. ChemCatChem
2012, 4, 559−566.
(16) De Rycke, N.; Couty, F.; David, O. R. P. Chem.Eur. J. 2011,
17, 12852−12871.
(17) Baidya, M.; Kobayashi, S.; Brotzel, F.; Schmidhammer, U.;
Riedle, E.; Mayr, H. Angew. Chem., Int. Ed. 2007, 46, 6176−6179.
(18) Soovaeli, L.; Rodima, T.; Kaljurand, I.; Kuett, A.; Koppel, I.;
Leito, I. Org. Biomol. Chem. 2006, 4, 2100−2105.
(19) Ibrahim, I. T.; Williams, A. J. Chem. Soc., Perkin Trans. 2 1982,
1459−1466.
(20) Mayer, B. J.; Spencer, T. A.; Onan, K. D. J. Am. Chem. Soc. 1984,
106, 6343−6348.
(21) Pawelka, Z.; Zeegers-Huyskens, T. Can. J. Chem. 2003, 81,
1012−1018.
(22) Fujii, T.; Nishida, H.; Abiru, Y.; Yamamoto, M.; Kise, M. Chem.
Pharm. Bull. 1995, 43, 1872−1877.
(23) Virtanen, N.; Polari, L.; Valila, M.; Mikkola, S. J. Phys. Org.
̈
̈
Chem. 2005, 18, 385−397.
(24) Hibbert, F.; Hunte, K. P. P. J. Chem. Soc., Perkin Trans. 2 1983,
1895−1899.
(25) (a) Singh, S.; Das, G.; Singh, O. V.; Han, H. Tetrahedron Lett.
2007, 48, 1983−1986. (b) Held, I.; Xu, S.; Zipse, H. Synthesis 2007,
1185−1196.
(26) (a) Fischer, C. B.; Xu, S.; Zipse, H. Chem.Eur. J. 2006, 12,
5779−5784. (b) Tandon, R.; Unzner, T.; Nigst, T. A.; De Rycke, N.;
Mayer, P.; Wendt, B.; David, O. R. P.; Zipse, H. Chem.Eur. J. 2013,
19, 6435−6442. (c) Lindner, C.; Tandon, R.; Liu, Y.; Maryasin, B.;
Zipse, H. Org. Biomol. Chem. 2012, 10, 3210−3218.
(27) (a) Larionov, E.; Mahesh, M.; Spivey, A. C.; Wei, Y.; Zipse, H. J.
Am. Chem. Soc. 2012, 134, 9390−9399. (b) Vedejs, E.; Daugulis, O.;
Harper, L. A.; MacKay, J. A.; Powell, D. R. J. Org. Chem. 2003, 68,
5020−5027.
(28) (a) Reichardt, C. Chem. Rev. 1994, 94, 2319−2358.
(b) Gutmann, V. Coord. Chem. Rev. 1976, 18, 225−255.
(29) Bassindale, A. R.; Stout, T. Tetrahedron Lett. 1985, 26, 3403−
3406.
(30) (a) Bassindale, A. R.; Stout, T. J. Chem. Soc., Perkin Trans. 2
1986, 221−225. (b) Bassindale, A. R.; Lau, J. C.-Y.; Stout, T.; Taylor,
P. G. J. Chem. Soc., Perkin Trans. 2 1986, 227−231.
(31) (a) Lambert, J.; Zhao, Y.; Zhang, S. M. J. Phys. Org. Chem. 2001,
14, 370−379. (b) Chult, C.; Corriu, R. J. P.; Reye, C.; Young, J. C.
Chem. Rev. 1993, 93, 1371−1448. (c) Dilman, A. D.; Ioffe, L. Chem.
Rev. 2003, 103, 733−772.
(32) The methodological analysis of different theoretical approaches
is described in: Zhang, C.; Patschinski, P.; Panisch, R.; Wender, J. H.;
Holthausen, M.; Zipse, H. Phys. Chem. Chem. Phys. 2014, 16, 16642−
16650.
(33) The pentacoordinate intermediate i4 exists as a local minimum
only on the MPW1K/6-31+G(d)+D2 and MP2(FC)/6-31+G(d)
potential energy surfaces. Reoptimization of this structure at the
MPW1K/6-31+G(d) level (i.e., without the added D2 dispersion
correction) leads back to ion pair intermediate i1.
(34) (a) Arshadi, M.; Johnels, D.; Edlund, U.; Ottosson, C.-H.;
Cremer, D. J. Am. Chem. Soc. 1996, 118, 5120−5131. (b) Reed, C. A.
Acc. Chem. Res. 1998, 31, 325−332. (c) Kinrade, S. D.; Deguns, E. W.;
Gillson, A.-M. E.; Knight, C. T. G. Dalton Trans. 2003, 3713−3716.
(d) Holmes, R. R. Chem. Rev. 1996, 96, 927−950.
8357
dx.doi.org/10.1021/jo5016568 | J. Org. Chem. 2014, 79, 8348−8357