Trifluoromethylation of allylsilanes under photoredox catalysis

Article abstract of DOI:10.1021/ol400184t

A new catalytic method to access allylic secondary CF₃ products is described. These reactions use the visible light excited Ru(bpy) ₃Cl₂·6H₂O catalyst and the Togni or Umemoto reagent as the CF₃ source. The photoredox catalytic manifold delivers enantioenriched allylic trifluoromethylated products not accessible under Cu(I) catalysis.

Full text of DOI:10.1021/ol400184t

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Trifluoromethylation of Allylsilanes under
Photoredox Catalysis
†
Satoshi Mizuta, Keary M. Engle, Stefan Verhoog, Oscar Galicia-Lopez,
†
†
†
ꢀ
Miriam O’Duill, Maurice Medebielle, Katherine Wheelhouse, Gerasimos Rassias,
†
‡
§
§
ꢀ
Amber L. Thompson, and Veronique Gouverneur*
†
,†
ꢀ
Chemistry Research Laboratory, University of Oxford, Mansfield Road, Oxford OX1
ꢀ
3TA, U.K., Universite Claude Bernard Lyon I, 43 bd du 11 Novembre, 1918 Villeurbanne
69622, France, and GlaxoSmithKline Medicines Research Centre, Gunnels Wood Road,
Stevenage, Herts SG1 2NY, U.K.
veronique.gouverneur@chem.ox.ac.uk
Received January 21, 2013
ABSTRACT
A new catalytic method to access allylic secondary CF3 products is described. These reactions use the visible light excited Ru(bpy)3Cl2 6H2O
3
catalyst and the Togni or Umemoto reagent as the CF3 source. The photoredox catalytic manifold delivers enantioenriched allylic trifluoro-
methylated products not accessible under Cu(I) catalysis.
Modern drug discovery campaigns have created an in-
creasing demand for new methods of trifluoromethylation.1
To date, catalytic reactions to access CF3 products with
by Sodeoka and co-workers,5 established that alkenes
activated with a trimethylsilyl group on the allylic position
undergo trifluoromethylation with 1-(trifluoromethyl)-1,2-
benziodoxol-3(1H)-one (Togni reagent I) under Cu(I) catal-
ysis. This method relies on the regiodirecting trimethylsilyl
3
Csp ꢀCF3 stereogenicity from substrates not activated by
a carbonyl functionality are extremely limited.2 This syn-
thetic gap prompted us to address this challenge with a
research program focused on allylic trifluoromethylation.
The groups of Buchwald, Liu, Wang, and Qing reported
thatCu-catalyzedtrifluoromethylation of terminal alkenes
leads to linear allylic CF3 products, exclusively.3 Our
own contribution,4 along with a similar study reported
3
group to program access to allyl products with Csp ꢀCF3
stereogenicity. However, the substrate scope is restricted to
the formation of gem-disubstituted allyl CF3 products, a
limitation that hampers its overall utility. Herein, we
report that allylsilanes are amenable to trifluoromethyla-
tion through photoredox catalysis.6 This high-performing
catalytic system for generating CF3• species allows access
to enantioenriched allylic CF3 products inaccessible under
Cu(I) catalysis.
† University of Oxford.
‡
ꢀ
Universite Claude Bernard Lyon I.
§ GlaxoSmithKline Medicines Research Centre.
(1) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem. Soc.
Initial investigations focused on the trifluoromethyla-
tion of ethyl 3-(trimethylsilyl)hex-4-enoate (E)-1a to form
ethyl 6,6,6-trifluoro-5-methylhex-3-enoate 2a (Table 1).
The model allylsilane 1a was selected on the basis that it
offersa structural platform for studying the stereochemical
Rev. 2008, 37, 320–330.
(2) Yasu, Y.; Koike, T.; Akita, M. Angew. Chem., Int. Ed. 2012, 51,
9567–9571.
(3) (a) Parsons, A. T.; Buchwald, S. L. Angew. Chem., Int. Ed. 2011,
50, 9120–9123. (b) Xu, J.; Fu, Y.; Luo, D.-F.; Jiang, Y.-Y.; Xiao, B.; Liu,
Z.-J.; Gong, T.-J.; Liu, L. J. Am. Chem. Soc. 2011, 133, 15300–15303. (c)
Wang, X.; Ye, Y.; Zhang, S.; Feng, J.; Xu, Y.; Zhang, Y.; Wang, J.
J. Am. Chem. Soc. 2011, 133, 16410–16413. (d) Chu, L.; Qing, F.-L. Org.
Lett. 2012, 14, 2106–2109.
(6) For reviews on photocatalysis, see: (a) Narayanam, J. M. R.;
ꢀ
ꢀ
(4) Mizuta, S.; Galicia-Lopez,O.;Engle,K.M.;Verhoog,S.;Wheelhouse,
Stephenson, C. R. J. Chem. Soc. Rev. 2011, 40, 102–113. (b) Teply, F.
K.; Rassias, G.; Gouverneur, V. Chem.;Eur. J 2012, 18, 8583–8587.
(5) Shimizu, R.; Egami, H.; Hamashima, Y.; Sodeoka, M. Angew.
Chem., Int. Ed. 2012, 51, 4577–4580.
Collect. Czech. Chem. Commun. 2011, 76, 859–917. (c) Tucker, J. W.;
Stephenson, C. R. J. J. Org. Chem. 2012, 77, 1617–1622. (d) Xuan, J.;
Xiao, W.-J. Angew. Chem., Int. Ed. 2012, 51, 6828–6838.
r
10.1021/ol400184t
XXXX American Chemical Society

aspects of the reaction. Copper catalysis led to disappoint-
ing results; under our previously reported conditions
using Togni reagent I,4 the starting material was largely
consumed (>90%), but less than 5% of 2a was formed
to the identification of the best reaction conditions: 1 equiv
of allylsilane 1a, 1.8 equiv of reagent I (or III), 5 mol % of
Ru(bpy)3Cl2 6H2O, MeOH (EtOH for III) at rt with
exposure of the reaction vessel to one household 14 W
light bulb over 48 h (entry 6). Notably, the stereoselectivity
of the trifluoromethylation was found to be dependent on
the CF3 source with Umemoto reagent III affording 2a
with the most favorable E/Z ratio.
The next experiments were designed to probe the reac-
tion’s tolerance to various degrees of substitution and to
study its stereochemical course. Conditions B (Togni I)
and/or C (Umemoto III) were considered for comparative
purposes (Table 2).
3
(entry 1). Gratifyingly, the visible lightꢀexcited Ru(bpy)3Cl2
3
6H2O catalyst provided 2a under various conditions.
Control experiments established the authenticity of the
photoredox concept (entries 2 and 3) and the importance
of the silyl group.7 Togni reagent I and Umemoto reagent
III outperformed Togni reagent II, CF3I, and CF3SO2Cl
in terms of conversion and/or selectivity. For III, we
noted little influence of the counteranion and observed
transesterification of both 1a and 2a when the reaction
solvent was methanol (entries 13ꢀ17). CF3I led to tri-
fluoromethylation (E/Z ratio ∼1) but only in the pres-
ence of i-Pr2NEt (entries 18ꢀ19).8 This observation con-
trasts with the detrimental effect of i-Pr2NEt when using
Togni reagent I (entry 5). The optimum solvent for this
reaction is MeOH (EtOH for III). These initial studies led
The desired allyl CF3 product was predominant with
traces of unidentified but separable byproducts. Typically,
higher yields were observed with reagent I, but III was
superior in terms of E-selectivity. Careful analysis of
the crude reaction mixture for the trifluoromethylation
of anti R-substituted β-silyl-(E)-crotylsilanes 1eꢀh7 revealed
that, out of the four possible isomers that could be obtained,
the syn-(E) branched CF3 allylic products were predomi-
nantly formed. In this series, the E/Z ratio is typically high,
but significant quantities of the anti-isomers were formed.
The stereochemical outcome of the reaction was found to be
sensitive to the substitution pattern of the substrates and
the CF3 source. For the assignment of the relative stereo-
chemistry, 2f (major isomer) was subjected to sequential
alkene then ester reduction, followed by esterification.
The resulting saturated ester, 2-benzyl-6,6,6-trifluoro-5-
methylhexyl 3,5-dinitrobenzoate 3, was characterized
by single-crystal X-ray diffraction.7 The assignment of
the relative stereochemistry of syn-2f indicates that the
silyl group is regio- and stereodirecting, with the sense of
stereocontrol being consistent with the well-established
anti-SE20 mode of addition observed for electrophiles other
than the putative CF3• species formed under the reaction
conditions.9 The trifluoromethylation of 1h gave the CF3-
substituted amino ester 2h following selective Boc depro-
tection (entry 7). The syn β-silyl-(E)-crotylsilane 1i also
responded totrifluoromethylation and led preferentially to
anti-(E)-2i in >70% yield (entry 8). The trifluoromethyla-
tion was applied to allylsilanes that are structurally differ-
ent from 1aꢀi. Theterminal linearallylsilane (E)-1j reacted
under the reaction conditions B togivethe desired terminal
branched allylic CF3 product 2j in 44% yield along with
an additional separable silylated product identified as 4
(33%) (entry 9). This side reaction indicates that the
addition of the CF3 group is not regioselective for this
substrate; the allylsilanes 1aꢀi therefore benefit from the
steric constraint imposed by the proximal stereogenic sily-
lated carbon for optimum regiocontrol during CꢀCF3 bond
formation. For 1k, both the trimethylsilyl and the phenyl
Table 1. Trifluoromethylation of Allylsilane 1a
entry CF3
condb
CuCle
convc,d (%) yieldc,d (%) E/Z ratiod
a
1
I
91
<5
<5
45
30
90
18
42
89
<5
<5
<5
82i
74
75i
>99
76
<5
48
48
66
<5
<5
1.4
f
2
I
no Ru
f
3
I
no light
4
I
27
f
2.5
5
I
iPr2NEtg
48 h
6
Ih
I
55
17
22
1.7
2.2
1.7
1.8
7
MeCN
CH2Cl2
DMF
8
I
9
I
37
f
10
11
12
13
14
15
16
17
18
19
20
21
22
II
II
II
III
III
IIIk
III
III
IV
IV
V
f
f
MeCN
DMF
30j
38j
34j
7
5.3
3.4
4
EtOH, 48 h
MeCN
DMF
8
33
5.4
iPr2NEtg
44
17
10
1
4.2
3
V
DMF
V
MeCN
a 1.2 equiv for IꢀIII and V; 10 equiv for CF3I (IV). b Variation from
(7) For details (e.g., the synthesis of the starting materials), see the
Supporting Information.
(8) The tertiary amine could serve as a sacrificial reductant and/or
quench deleterious HI byproduct.
(9) (a) Masse, C. E.; Panek, J. S. Chem. Rev. 1995, 95, 1293–1316. (b)
Fleming, I.; Barbero, A.; Walter, D. Chem. Rev. 1997, 97, 2063–2192. (c)
Chabaud, L.; James, P.; Landais, Y. Eur. J. Org. Chem. 2004, 3173–
3199.
the standard conditions detailed in the scheme. c Conv refers to con-
sumption of 1a and yield refers to the formation of 2a. d Analysis by 19
F
NMR with C6H5F as internal standard. e Conditions for the reaction:
20 mol % of CuCl, MeOH, 70 °C, 2 h. f No product. g 2.0 equiv of
iPr2NEt. h 1.8 equiv of I. i Transesterification of 1a and 2a. j Yield of
isolated 2a. k BF4ꢀ as counteranion.
B
Org. Lett., Vol. XX, No. XX, XXXX

These results encouraged us to study the efficiency of
chirality transfer upon trifluoromethylation of enantioen-
riched allylsilanes (R,E)-1nꢀq (er >97:3) (Scheme 1).10
Applying the standard reaction conditions B (Togni re-
agent I), the major products (S,E)-2nꢀq were isolated
with some erosion of the er (>85:15). The minor isomers
(S,Z)-2o and (S,Z)-2q were formed in higher enantiomeric
ratio (98:2). The installation of a CF2CF3 group at the
allylic position was also successful with reagent VI11 giving
5 and 6 with good er (9:1). The absolute configuration of
the E isomers is assigned assuming an anti-SE0 mode of
addition with respect to the silyl group in analogy with the
trifluoromethylation of 1f.
Table 2. Trifluoromethylation of Allylsilanes 1bꢀm
Scheme 1. Trifluoromethylation of Enantioenriched 1nꢀq
To study the mechanism of this reaction, we focused
on the reaction conditions using the Togni reagent I
(Scheme 2). Cyclic voltammetry measurements are con-
sistent with a catalytic oxidative quenching cycle. The
reduction potential of Togni reagent I (ꢀ0.68 V vs SCE
in CH3CN)12 can be compatible with the reduction step
using excited state Ru(bpy)32þ*; this implies that single
electron transfer (SET) reduction of I would be concurrent
with the oxidation of Ru(bpy)32þ* to Ru(bpy)33þ (ꢀ0.81 V
vs SCE in CH3CN).6,13 The ensuing Togni I radical anion
a B: 5 mol % Ru(bpy)3Cl2 6H2O, 14 W light bulb, 1.8 equiv of
•
3
could collapse to generate the electrophilic CF3 , which is
reagent I, MeOH, rt, 24 h. C: 5 mol % of Ru(bpy)3Cl2 6H2O, 14 W light
3
bulb, 1.8 equiv of reagent III, EtOH, rt, 24 h. b E/Z ratio determined by
19F NMR. c Syn/anti ratio determined by 19F NMR. d Values in pa-
rentheses refer to the measured ratio following purification. e Yield of
isolated product. f For I, 2 ꢁ 1.5 equiv was used. g N-Boc deprotection
with HCl/EtOH. h Ratio refers to E/Z or anti/syn; relative stereochem-
istry assigned assuming anti mode of addition with respect to the silyl
group by analogy with 2f.
well suited to add regio- and stereoselectively to allylsilane
1a. The resultant radical species 714 could then undergo a
second SET with the strong oxidant Ru(bpy)33þ (þ1.29 V
vs SCE in CH3CN),6,13 an event regenerating the ground
state photocatalyst Ru(bpy)32þ and forming the stabilized
β-silyl cation 8.15 Desilylation of 8 with methanol provides
2a. A possible mechanistic scenario for the alternative
reductive quenching cycle could feature the allylsilane 1a
substituents contribute to direct the regiochemistry of the
trifluoromethylation (entry 10). The terminal branched
allylsilane 1l gave the linear allyl CF3 product 2l but only
under conditions C (Umemoto reagent III). No reaction
was observed with Togni reagent I (entry 11). This observa-
tion highlights the impact of the CF3 source on reactivity.
Finally, the formation of 2m from 1m demonstrates that the
silyl group is captured with methanol (entry 12).
(11) Li, Y.; Studer, A. Angew. Chem., Int. Ed. 2012, 51, 8221–8224.
(12) Reduction potential refers to peak potential (irreversible
reduction). See the Supporting Information.
(13) Haga, M.-A.; Dodsworth, E. S.; Eryavec, G.; Seymour, P.;
Lever, A. B. P. Inorg. Chem. 1985, 24, 1901–1906.
(14) Fragmentation of the β-silyl radical is documented and cannot
ꢀ
be ruled out; see: Mereau, R.; d’Antuono, P.; Castet, F.; Rouquet, G.;
Robert, F.; Landais, Y. Organometallics 2010, 29, 2406–2412.
(15) Silicon has the ability to stabilize β-carbocations (29ꢀ30
kcal/molꢀ1) through σSi‑Cfp hyperconjugation and to a lesser degree
β- carboradicals (2.6ꢀ4.5 kcal/molꢀ1): Lambert, J. B.; Zhao, Y.; Emblidge,
R. W.; Salvador, L. A.; Liu, X.; So, J.-H.; Chelius, E. C. Acc. Chem. Res.
1999, 32, 183–190.
(10) For the synthesis of enantioenriched allylsilanes, see: Li, D.;
Tanaka, T.; Ohmiya, H.; Sawamura, M. Org. Lett. 2010, 12, 3344–3347.
Org. Lett., Vol. XX, No. XX, XXXX
C

in MeOH with the reductive quencher i-Pr2NEt (Table 1,
entry 5). Moreover, the reaction was successfully per-
formed in CH3CN, DMF and CH2Cl2 in the absence of
MeOH (entries 7ꢀ9, Table 1). None of these solvents can
act as a reductive quencher to Ru(bpy)32þ*. The mecha-
nism depicted in Scheme 2 does not account for the
different E/Z and syn/anti ratios observed for the products
depending on the CF3 source. The mechanistic details
remain unclear, but we verified that 1a and 2a do not
undergo E/Z isomerization in the presence of 5 mol % of
Scheme 2. Mechanistic Considerations
Ru(bpy)3Cl2 6H2O under visible light activation. Monitor-
3
ing of product distribution over time also confirmed that
all isomers are stable under the reaction conditions.
Control “light/dark” experiments verified the necessity of
light and are not supportive of a radical chain propagation
mechanism.18 In addition, a silylated product resulting
from addition of Togni reagent I (or II) across the alkene
functionality could never be detected.
This work discloses a photoredox-based catalytic meth-
od for the synthesis of enantioenriched branched allylic
CF3 products. The silyl group in the starting material is an
important entity to control the regioselectivity of the
trifluoromethylation. It also programs the stereochemical
outcome of the reaction in synergy with the CF3 reagent.
On a more fundamental level, this work highlights the
influence of the CF3 source on photoredox catalyzed
processes in terms of both reactivity and stereoselectivity.
We are continuing work to gain further insight into the
reaction mechanism as this information should enable
future applications and facilitate the design of new CF3
reagents.
Acknowledgment. We thank the EU (Grant No. PIIF-
GA-2010-274903toS.M.), the Skaggs-OxfordScholarship
Program and NSF GRFP (predoctoral fellowships to
K.M.E.), the EPSRC and GSK (studentship to S.V.),
the CONACyT Mexico Program (Grant No. 205456 to
O.G.L.), the BBSRC (studentship to M.O’D.) for generous
funding, Dr. K. Vincent and I. McPherson (Oxford
University) for helping with cyclic voltammetry measure-
ments, and the Diamond Light Source for an award of
instrument time on I19(MT7768) and the instrument
scientists for support.
as a reductive quencher to Ru(bpy)32þ* leading to a radical
cation species 9.16 This reactive intermediate could com-
bine with CF3• (radicalꢀradical coupling, path A) emerg-
ing from Togni reagent I via SET, an event concurrent
with the oxidation of Ru(bpy)3þ, which is a strong redu-
cing agent.6,14 This pathway was dismissed based on the
high oxidation potential of allylsilane 1a (>þ1.8 V vs SCE
in CH3CN)17 making the formation of the radical cation
thermodynamically quite challenging. If an alternative
reductive quencher to Ru(bpy)32þ* is present, CF3• would
add to 1a to give the radical species 7. Oxidation with the
mild oxidant Ru(bpy)32þ* (þ0.77 V vs SCE in CH3CN)6,14
and desilylation with MeOH would afford 2a (Path B). No
product 2a was obtained when the reaction was conducted
Supporting Information Available. Experimental pro-
cedures and spectral data. This material is available free
of charge via the Internet at http://pubs.acs.org.
(18) This process could have potentially accounted for the differences
in E/Z selectivity, with oxidation of 7 into 8 triggered by the trifluor-
(16) Desilylation of 9 with MeOH may lead to an allyl radical, which
can combine with CF3 . In a control experiment, no TEMPOꢀallyl
3þ
•
omethylation reagent instead of Ru(bpy)3
.
adduct was observed.
(17) Oxidation potential refers to peak potential (irreversible oxidation).
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX