Leaving Group Effects on the Selectivity of the Silylation of Alcohols: The Reactivity-Selectivity Principle Revisited

Article abstract of DOI:10.1021/acs.orglett.5b01536

TBS protection of primary alcohol naphthalen-1-ylmethanol (4a) and secondary alcohol 1-(naphthalen-1-yl)ethanol (4b) has been studied under various reaction conditions. The primary/secondary selectivity is largest in the comparatively slow Lewis base catalyzed silylation in apolar solvents and systematically lower in DMF. Lowest selectivities (and fastest reaction rates) are found for TBS triflate 1b, where only minor effects of solvent polarity or Lewis base catalysis can be observed. (Chemical Equation Presented).

Full text of DOI:10.1021/acs.orglett.5b01536

Letter
pubs.acs.org/OrgLett
Leaving Group Effects on the Selectivity of the Silylation of Alcohols:
The Reactivity−Selectivity Principle Revisited
Pascal Patschinski* and Hendrik Zipse*
̈
̈
Department of Chemistry, Ludwig-Maximilians-Universitat, Butenandtstrasse 5-13, 81377 Munchen, Germany
*
S Supporting Information
ABSTRACT: TBS protection of primary alcohol naphthalen-1-
ylmethanol (4a) and secondary alcohol 1-(naphthalen-1-yl)ethanol
(
4b) has been studied under various reaction conditions. The primary/
secondary selectivity is largest in the comparatively slow Lewis base
catalyzed silylation in apolar solvents and systematically lower in DMF.
Lowest selectivities (and fastest reaction rates) are found for TBS
triflate 1b, where only minor effects of solvent polarity or Lewis base
catalysis can be observed.
he silylation of alcohols is one of the most important
reactions in the chemo- and regioselective manipulation of
decisively on the catalysts and solvents used: while S = 20.0 using
DMF, 2 as solvent, and Lewis base catalyst, significantly higher
selectivities of 123 (DMAP, 6) and 130 (PPY, 7) have been
T
1
−3
complex organic molecules.
Its usefulness as a protecting
group strategy derives, in part, from the ability to differentiate
primary and secondary alcohols through the combined use of
silylation reagents of different size and reactivity, catalysts,
auxiliary bases, and solvents. The most frequently employed
reagents such as tert-butyldimethylsilyl chloride (TBSCl, 1a)
combine a silyl group of intermediate size with a leaving group of
moderate reactivity. The latter makes reagent 1a compatible with
a number of activation protocols, of which the “Corey
procedure” involving DMF (2) as the solvent in combination
with imidazole (3a) as base and the Lewis base catalyzed
obtained for electron-rich pyridines in CDCl . The enhanced
3
selectivity observed in these latter cases is accompanied by a
significant reduction in absolute reaction rates, which also implies
that combinations of 6 or 7 with DMF as solvent provide
7
effectively the same rate as obtained with DMF alone. This
result may be rationalized with the often invoked “reactivity−
8
selectivity” principle, and we therefore explore here the
influence of other factors responsible for the reaction rates in
silylation reactions. This particularly concerns the choice of the
leaving groups present in the silylation reagents.
The reactivity of TBSCl 1a toward alcohols 4a and 4b was
activation in apolar organic solvents with triethylamine (Et N,
3
4
−6
characterized already under a variety of reaction conditions
3
b) are the most common ones.
We have recently shown for the reaction of reagent 1a with
primary and secondary alcohols 4a and 4b that the selectivity
defined here as the ratio of reaction rates S = kprim/k ) depends
7
before. We reiterate here that the rate of reaction of 4b in DMF-
d7 does not depend on the type or amount of the auxiliary base
added, as long as there is sufficient base present to neutralize the
HCl byproduct. In the complete absence of auxiliary base, initial
(
sec
rates are practically identical to those using 1.2 equiv of Et N 3b,
but the reaction eventually comes to a halt at just below 80%
Scheme 1. Silylation of Primary and Secondary Alcohols 4a/
b with TBSCl (1a)
3
4
conversion. Addition of 1.2 equiv of Et N 3b to the reaction
3
mixture at that point neutralizes the HCl byproduct and allows
the reaction to go to completion. The reaction rates show little
variation with the particular type of auxiliary base, as is
demonstrated by reaction half-lives of 7.4 ± 0.6 min for Et N
3
3b and 7.5 ± 0.2 min for imidazole 3a. Even increasing the
concentration of imidazole 3a to 1.8 equiv does not influence the
reaction rate (t1/2 = 7.4 ± 0.4 min) (Figure 1). These results are,
together with those for pyridines 6 and 7, most easily rationalized
by assuming that DMF-d is the only catalytically active Lewis
7
Received: May 26, 2015
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b01536
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
ence observed for reagents 1a and 1c (Figure 2). This implies
that the strongly activated reagent 1b undergoes a direct (that is,
uncatalyzed) reaction with substrate alcohol 4b.
Figure 1. Comparison of auxiliary bases in DMF-d for the silylation of
7
4
b with 1a.
base present under these conditions, while all other bases (3a, 3b,
, 7) merely act as auxiliary bases. This is distinctly different from
the alternative mechanism involving imidazole as the catalytically
Figure 2. Influence of catalyst concentration for various silylation
reagents in CDCl3.
6
4
,9
active Lewis base.
An analogous set of measurements has been performed in
The influence of the leaving group on the reaction rate was first
explored for reagents containing the TBS protecting group such
as tert-butyldimethylsilyl chloride (TBSCl, 1a), tert-butyldime-
DMF-d as the solvent and Et N 3b as the auxiliary base. This
7
3
leads to practically the same absolute rates as compared to CDCl3
for silyl triflate 1b that hardly depend on DMAP concentration.
This is also true for silyl chloride 1a, silyl cyanide 1c, and silyl
amide 1e whose respective reaction rates remain practically
unchanged after addition of 30 mol % DMAP (Figure 3). It
10
thylsilyl triflate (TBSOTf, 1b), tert-butydimethylsilyl cyanide
1
1
(
1
(
TBSCN, 1c), tert-butyldimethylsilyl imidazole (TBSImi,
12
d), and tert-butyldimethylsilyl-N-methyltrifluoroacetamide
13
MTBSTFA, 1e). These measurements were performed
using the previously developed procedure for secondary alcohol
b using 30 mol % of DMAP and Et N (1.2 equiv) in CDCl
4
(
3
3
7
Table 1).
Table 1. Reaction Half-Lives for the Silylation of 4b with 30
mol % of DMAP Catalyst for Various Leaving Groups in
CDCl and DMF-d
3
7
Figure 3. Influence of catalyst concentration for various silylation
reagents in DMF-d7.
a
−1 −1
b
c
keff in l·mol ·s . Half-life in min. Based on 4.4% conversion after
should be added here that turnover curves for 1c deviate from the
second-order behavior observed for all other reagents such as to
indicate (partial) autocatalysis through product cyanide (see the
Supporting Information). Even under these conditions, very little
turnover can be detected for silyl imidazole 1d. Because of the
higher reactivity of primary alcohol 4a, absolute reaction rates
could only be determined for the reagents 1a and 1c in CDCl3.
When a catalyst loading of 4 mol % of DMAP is used, silyl
chloride 1a is approximately 1 order of magnitude faster than silyl
cyanide 1c. For both reagents, the rate of reaction depends
linearly on the catalyst concentration (see the Supporting
Information for details).
d
1
0 d in CDCl . Based on 6.4% conversion after 120 min in DMF-d .
3
7
Very little conversion can be observed under these basic
conditions for silyl imidazole 1d, which is estimated to react 1
order of magnitude slower than 1e. For the more reactive
reagents 1a, 1c, and 1e, the rate in DMF-d is increased by 2
7
orders of magnitude compared to that in CDCl , while for triflate
3
1
b reaction rates are quite comparable in both solvents.
Additional experiments in CDCl demonstrate that the rate of
3
reaction of silyl triflate 1b is independent of catalyst
concentration, in significant contrast to the first-order depend-
B
DOI: 10.1021/acs.orglett.5b01536
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Reactions in DMF-d were found to be too fast for accurate
7
direct rate measurements for all reagents 1a−e, and the reactivity
of primary alcohol 4a was therefore quantified through
competition experiments with secondary alcohol 4b. These
experiments employ equimolar mixtures of alcohols 4a and 4b,
and the underlying reaction kinetics are thus directly comparable
14
to those of kinetic resolution experiments. Turnover curves in
these experiments measure the chemoselectivity (expressed as C
=
([5a] − [5b])/([5a] + [5b]) as a function of turnover (of both
substrate alcohols 4a and 4b). For the highly selective silylation
in CDCl using silyl chloride 1a with 4 mol % of DMAP (6) as
3
catalyst and Et N (3b, 1.2 equiv) as the auxiliary base, we find
3
that primary alcohol 4a turns over almost completely before that
of secondary alcohol 4b commences at conversions >50%
(
Figure 4). The corresponding turnover curve is characterized by
Figure 5. Temperature-dependent competition experiments with silyl
triflate 1b in CD Cl .
2
2
transformations can be achieved by highly reactive reagents even
at low temperature.
In order to rationalize the influence of the leaving groups on
relative reaction rates, reaction enthalpies (ΔHRxn) for the
silylation of secondary alcohol 4b have been calculated at the
MP2/G3MP2large//MPW1K/6-31+G(d) level of theory in
combination with the SMD continuum solvation model in
chloroform. For the sake of brevity, trimethylamine was used as
auxiliary base in the calculation. For all reagents 1a−e a
satisfactory correlation can be found between reaction rates in
CDCl (k ) against the reaction enthalpy (ΔHRxn). This
3
eff
correlation can be used to predict reaction rates for other
reagents such as TBS perchlorate and azide (Figure 6).
Figure 4. Competition experiments performed for 1a, 1b, and 1c in
CDCl with 4 mol % of DMAP 6.
3
chemoselectivities C just below 1.0 for the first 50% turnover and
a subsequent systematic decline to C = 0.0 afterward. The data
points located in the critical region between 30 and 70% turnover
can nicely be fitted with a selectivity value S = 120 obtained from7
previous direct kinetic measurements for alcohols 4a and 4b,
thus confirming the validity of the relative rate measurements
obtained here. The same high selectivity S was measured under
these conditions for silyl cyanide 1c, while that for silyl triflate 1b
is much lower at S = 4. Changing to the Lewis basic solvent DMF-
d , the reaction of silyl chloride 1a becomes significantly less
7
7
selective with S = 20, again in line with previous observations. In
conclusion, these results show that the most reactive reagent
(
triflate 1b) is the least selective in differentiating between
primary and secondary alcohols 4a and 4b. Comparatively low
selectivities are also found when the (catalytically active) Lewis
base solvent DMF-d is employed.
7
Figure 6. Correlation of reaction enthalpy ΔH vs log(k ) with 4b
Rxn
eff
Whether the selectivity of the highly reactive silyl triflate 1b
can be increased through moving to lower reaction temperatures
was finally addressed in competition experiments using 1:1
mixtures of alcohols 4a and 4b in CD Cl at +20, 0, and −78 °C
and DMAP (30 mol %) in CDCl .
3
Three different mechanistic scenarios for the silylation of
alcohols emerge from the current results as a function of leaving
groups, solvents, and Lewis bases (Figure 7). The fastest and
least selective reactions are observed for TBS triflate 1b.
Reactions show only small solvent effects in this case and hardly
respond to Lewis base catalysis. This can best be rationalized
through direct (that is uncatalyzed) reaction of alcohols with 1b,
whose properties may approximately be understood as those of a
2
2
(
Figure 5). This change in solvent away from CDCl is expected
3
7
to have only a minor influence on reaction rates but allows
reliable selectivity measurements at much lower temperatures. A
small increase in selectivity was observed when lowering the
reaction temperature from 20 °C (S = 4) to 0 °C (S = 6).
Lowering the reaction temperature further to −78 °C leads to S =
1
5. It can thus be concluded that only moderatly selective
C
DOI: 10.1021/acs.orglett.5b01536
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(9) (a) Clayden, J.; Greeves, N.; Warren, S.; Wothers, P. Organic
Chemistry, 1st ed.; Oxford University Press: Oxford, 2001. (b) Kocienski,
P. J. Protecting Groups, 3rd ed.; Thieme: Stuttgart, 2005.
(
1
(
2
2
(
10) Corey, E. J.; Cho, H.; Ru
981, 22, 3455−3458.
11) (a) Treichel, P. M.; Shaw, D. B. J. Organomet. Chem. 1977, 139,
1−30. (b) Renzetti, A.; Koga, N.; Nakazawa, H. Bull. Chem. Soc. Jpn.
014, 87, 59−68.
12) Tanabe, Y.; Murakami, M.; Kitaichi, K.; Yoshida, Y. Tetrahedron
Lett. 1994, 35, 8409−8412.
13) Mawhinney, T. P.; Madson, M. A. J. Org. Chem. 1982, 47, 3336−
339.
̈
cker, C.; Hua, D. H. Tetrahedron Lett.
(
3
(
14) Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1988, 18, 249−330.
(15) Akhani, R. K.; Moore, M. I.; Pribyl, J. G.; Wiskur, S. L. J. Org.
Chem. 2014, 79, 2384−2396.
Figure 7. Overview of various pathways for the silylation reaction
depending on the choice of solvent and leaving group.
contact ion pair. Better selectivities at somewhat slower rates are
obtained in DMF as a Lewis basic solvent for the less reactive
reagents 1a, 1c, and 1e. These reactions are likely to involve
silylated DMF as transient intermediates of the catalytic cycle.
Best selectivities and slowest rates are obtained in apolar organic
solvents such as CDCl and CD Cl for the Lewis base catalyzed
3
2
2
reaction of reagents 1a, 1c, and 1e. That reaction rates correlate
so systematically with selectivities is likely due to the steric
demands of the respective transition states: while the uncatalyzed
reaction of alcohols with triflate 1b proceeds through transition
states composed only of these two reactants, the Lewis base
catalyzed pathways have to accommodate the presence of either a
15
small (such as DMF) or a larger (e.g., DMAP) Lewis base. This
qualitative rationale also implies that the development of
sterically more encumbered Lewis bases may lead to still larger
selectivities for the Lewis base catalyzed processes.
ASSOCIATED CONTENT
Supporting Information
■
*
S
Experimental details, explanation of methods, theoretical
calculations, and time conversion plots. The Supporting
Information is available free of charge on the ACS Publications
website at DOI: 10.1021/acs.orglett.5b01536.
AUTHOR INFORMATION
Corresponding Authors
■
*
patch@cup.lmu.de.
zipse@cup.lmu.de.
*
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, 2006. (b) Kocienski, P. J.
Protecting Groups, 3rd ed.; Thieme: Stuttgart, 2005.
(
2) White, J. D.; Carter, R. G. Silyl Ethers in Science of Synthesis, Fleming,
I., Ed.; Thieme: Stuttgart, 2002.
(
(
3) Crouch, D. Synth. Commun. 2013, 43, 2265−2279.
4) Corey, E. J.; Venkateswarlu, A. J. Am. Chem. Soc. 1972, 94, 6190−
6191.
(
5) Chaudhary, E. J.; Hernandez, O. Tetrahedron Lett. 1979, 20, 99−
102.
(
6) Yoshida, K.; Takao, K.-i. Tetrahedron Lett. 2014, 55, 6861−6863.
(
7) Patschinski, P.; Zhang, C.; Zipse, H. J. Org. Chem. 2014, 79, 8348−
8
357.
(
8) Mayr, H.; Ofial, A. Angew. Chem., Int. Ed. 2006, 45, 1844−1854.
D
DOI: 10.1021/acs.orglett.5b01536
Org. Lett. XXXX, XXX, XXX−XXX