Nickel-catalyzed Mizoroki-Heck reaction of aryl sulfonates and chlorides with electronically unbiased terminal olefins: High selectivity for branched products

Article abstract of DOI:10.1002/anie.201308391

Achieving high selectivity in the Heck reaction of electronically unbiased alkenes has been a longstanding challenge. Using a nickel-catalyzed cationic Heck reaction, we were able to achieve excellent selectivity for branched products (≥19:1 in all cases) over a wide range of aryl electrophiles and aliphatic olefins. A bidentate ligand with a suitable bite angle and steric profile was key to obtaining high branched/linear selectivity, whereas the appropriate base suppressed alkene isomerization of the product. Although aryl triflates are traditionally used to access the cationic Heck pathway, we have shown that, by using triethylsilyl trifluoromethanesulfonate, we can effect a counterion exchange of the catalytic nickel complex, such that cheaper and more stable aryl chlorides, mesylates, tosylates, and sulfamates can be used to yield the same branched products with high selectivity. Branching out: A Ni-catalyzed Heck reaction for the preparation of 1,1-disubstituted alkenes is presented. High selectivity for the branched products is achieved with electronically unbiased aliphatic terminal olefins. Regioselectivities remain consistently high (≥19:1) throughout. TESOTf=triethylsilyl trifluoromethanesulfonate. Copyright

Full text of DOI:10.1002/anie.201308391

.
Angewandte
Communications
DOI: 10.1002/anie.201308391
Mizoroki–Heck Chemistry
Nickel-Catalyzed Mizoroki–Heck Reaction of Aryl Sulfonates and
Chlorides with Electronically Unbiased Terminal Olefins: High
Selectivity for Branched Products**
Sarah Z. Tasker, Alicia C. Gutierrez, and Timothy F. Jamison*
Abstract: Achieving high selectivity in the Heck reaction of
electronically unbiased alkenes has been a longstanding chal-
lenge. Using a nickel-catalyzed cationic Heck reaction, we were
able to achieve excellent selectivity for branched products (ꢀ
19:1 in all cases) over a wide range of aryl electrophiles and
aliphatic olefins. A bidentate ligand with a suitable bite angle
and steric profile was key to obtaining high branched/linear
selectivity, whereas the appropriate base suppressed alkene
isomerization of the product. Although aryl triflates are
traditionally used to access the cationic Heck pathway, we
have shown that, by using triethylsilyl trifluoromethanesulfo-
nate, we can effect a counterion exchange of the catalytic nickel
complex, such that cheaper and more stable aryl chlorides,
mesylates, tosylates, and sulfamates can be used to yield the
same branched products with high selectivity.
S
ince the 1970s, the Mizoroki–Heck reaction[1] has afforded
synthetic chemists a powerful way to synthesize more
substituted olefins from aryl or benzyl electrophiles and
simpler alkenes.[2] Although much less well studied than its
Pd-catalyzed counterpart, the Ni-catalyzed Heck reaction[3]
can offer several distinct advantages in addition to the low
cost of nickel, including: faster oxidative addition (allowing
for the use of a wide range of electrophile classes), more facile
olefin insertion, and a more controlled steric environment
owing to shorter Ni-ligand bond lengths.[4] These advantages
seem to be underutilized in the Heck reaction, compared to
the more prevalent use of Ni in other cross-coupling
reactions.[5] Recently, our group demonstrated some of these
features by showing that benzyl chlorides could react with
ethylene and terminal olefins in a highly selective manner (ꢀ
19:1 in most cases) to afford branched products using
[Ni(cod)2] and PCy2Ph (cod = 1,5-cyclooctadiene, Cy = cyclo-
hexyl).[6] This report represented the first example of
a catalyst-controlled, highly branched-selective Heck reaction
of electronically unbiased alkenes. Herein, we report the
Scheme 1. Branched-selective Heck reactions.
expansion of this highly regioselective reaction to more
widely used aryl electrophiles.
For terminal alkenes, there are two possible regiochemical
outcomes of the Heck reaction (Scheme 1). Classically,
electron-poor olefins have been utilized for Heck couplings,
yielding almost exclusively linear products. However, by the
mid-1990s, a series of developments by Cabri and Candiani,[7]
as well as others,[8] allowed access to the cationic Heck
pathway to provide high selectivity for branched products
with electron-rich olefins. By manipulation of the reaction
conditions, counterion dissociation from the metal center is
favored, which can cause a reversal in overall branched/linear
(br/ln) selectivity.[9]
Unfortunately, electronically unbiased olefins have given
more modest br/ln ratios (e.g. 5.2:1), even under cationic
conditions.[10] In the past year, this field has seen renewed
interest. Zhou and co-workers reported the Pd-catalyzed
Heck reaction of terminal olefins[11] and vinylarenes[12] with
[*] S. Z. Tasker, Dr. A. C. Gutierrez, Prof. T. F. Jamison
Department of Chemistry, Massachusetts Institute of Technology
Cambridge, MA 02139 (USA)
E-mail: tfj@mit.edu
Homepage: http://web.mit.edu/chemistry/jamison
[**] Financial support was provided by the NIGMS (GM62755), NSF
(Graduate Research Fellowship to S.Z.T.) and NIH (Postdoctoral
Fellowship to A.C.G.). We are grateful to Eric Standley and Dr. Kim
Lebek Jensen for helpful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201308391.
1858
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1858 –1861

Angewandte
Chemie
of monodentate ligands of varying electronic and steric
properties provided little improvement in yield or br/ln
selectivities; thus, the search was expanded to bidentate
ligands. Dcypp (6) demonstrated excellent regioselectivity,
but only moderate reactivity (entry 2). Elevated temperatures
led to formation of visible nickel(0) particles and deterio-
ration of yields. We hypothesized that, while oxidative
addition proceeded smoothly, insertion of the phenyl group
to the alkene remained slow. In order to promote this step, we
investigated ligands of a wider bite angle, as larger bite angles
have been shown, both theoretically[16] and experimentally,[17]
to increase the rate of migratory insertion.
Figure 1. Key intermediate structures for benzyl (left)[6a] and aryl (right)
electrophiles in the cationic Heck mechanism.
aryl triflates in good yields. Stahl and co-workers also
reported the oxidative Heck reaction of vinyl boronic acids
and terminal olefins using Pd and a phenanthroline-type
ligand.[13] In these reactions, high levels of regioselectivity for
a wide substrate scope are of the utmost importance, as in
nearly all cases the resulting alkene regioisomers are insep-
arable by column chromatography.[14] However, the work
from the Zhou and Stahl groups displays a wide range of
regioselectivities, with the former method typically needing
ortho-subsituted aryl triflates to achieve excellent levels of
regioselectivity (defined here as ꢀ 19:1, i.e. ꢀ 95:5).
Indeed, bidentate ligands with larger bite angles[18] such as
dippf (7) increased conversion and were more stable at higher
temperatures (entry 3). A four-carbon bridge seemed most
promising overall. Dcypb (10) showed low conversion at
room temperature, but was stable at 608C, providing excellent
br/ln product ratios, good yields, and no formation of side
products (entry 6). Use of cyclopentyl rather than cyclohexyl
groups in the ligand (1) provided the best results with high
conversion, excellent br/ln selectivity, and good overall r.r.
Building on our understanding of the Ni-catalyzed
cationic Heck reaction,[6a,15] we set out to directly access this
pathway with phenyl triflate (2a) and 1-octene. The ligand
previously used for benzyl chlorides, PCy2Ph (5), produced
only a moderate yield and regioisomeric ratio (r.r.) of the
desired product, even after optimization of reaction condi-
tions (Table 1, entry 1). Comparing the proposed intermedi-
ates in Figure 1, we see that a benzyl electrophile can
transition from h1 to h3.[6a] However, the ligand sphere for
aryl electrophiles is quite different, presumably containing
two ligated phosphines, and providing a fresh challenge for
inducing high levels of regiocontrol. Extensive investigation
À
(entry 7). Finally, if a persistent Ni H species is responsible
for erosion of the r.r. of the product after b-H elimination,
changing the base might decrease isomerization. Gratifyingly,
1,4-diazabicyclo[2.2.2]octane (DABCO) provided superior
yields and regioselectivities, and additionally allowed reduc-
tion of the catalyst loading to 10 mol% (entry 8). Further-
more, the amount of 1-octene necessary in the reaction could
be reduced to 1.5 equiv or even to 1.1 equiv (entry 9), making
this transformation attractive for reactions in which both
reaction components are valuable.
Even more significantly, the
combination of DABCO and
ligand 1 allowed the use of aryl
Table 1: Ligand optimization: tether length and steric demand.
chlorides for the first time (Table 2).
Although our group had previously
demonstrated that TESOTf could
be used to perform a counterion
À
À
exchange of Ni Cl to Ni OTf for
benzyl chlorides in order to enter
the cationic Ni-Heck pathway, pre-
vious attempts to use aryl chlorides
had failed. But, with some modifi-
cation of reaction conditions, good
yields of product 3a could be
obtained, not only for aryl chlorides
(entry 1), but also aryl mesylates
(entry 2), tosylates (entry 3), and
sulfamates (entry 4).[19] Although
bromides and iodides underwent
oxidative addition, counterion
exchange did not occur. This use
of electrophiles that are tradition-
ally viewed as unreactive with Pd
demonstrates the power of Ni-cata-
lyzed reactions to access products
made from less expensive, more
stable, and more readily available
chlorides and phenol derivatives.[5]
Entry Ligand
Natural
Base (equiv) T [8C] Conversion [%][b] Yield [%][c]
br/ln r.r.[d]
bite angle [8][a]
1[e]
2
PCy2Ph (5)
dcypp (6) 104
dippf (7)
diop (8)
dppb (9)
–
Et3N (3)
Et3N (3)
Et3N (3)
Et3N (3)
Et3N (5)
Et3N (5)
Et3N (5)
RT
RT
60
RT
RT
60
60
87
20
50
86
65
47
22
41
70
48
67
86
99[g]
92
11.4:1 2.7
>100:1 9.9
>100:1 9.8
5.2:1 2.8
3[e]
4
120
105
102
5
7.2:1 2.8
6
dcypb (10) 111
68
89
>100:1 9.2
>100:1 7.5
87:1 32.9
7
ligand 1
ligand 1
ligand 1
107
107
107
8[f]
9[h]
DABCO (3) 60
DABCO (3) 60
>95
94
>100:1 36.5
À
[a] Calculated according to Ref. [18] using molecular mechanics MMF force field and a Ni P bond length
of 2.28 ꢀ. [b] GC conversion with respect to PhOTf. [c] Yield of branched product 3a only, as determined
by GC analysis. [d] Ratio of br/all isomers. [e] Ligand (30 mol%). [f] [Ni(cod)2] (10 mol%), 1 (12 mol%),
1-octene (1.5 equiv), THF (1m). [g] Yield of isolated product. [h] 1-octene (1.1 equiv). cod=1,5-
cyclooctadiene, DABCO=1,4-diazabicyclo[2.2.2]octane.
Angew. Chem. Int. Ed. 2014, 53, 1858 –1861
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1859

.
Angewandte
Communications
Table 2: Reaction of aryl halides and sulfonates with TESOTf.[a]
Entry
X
Yield [%]
br/ln
r.r.[b]
1
2
3
4
5
6
Cl
OMs
OTs
OSO2NMe2
Br
I
81
91
72
61
4[c]
2[c]
62:1
60:1
37:1
59:1
–
35.7
30.4
19.6
19.6
–
–
–
[a] All yields shown are of isolated products, unless otherwise noted.
[b] Ratio of br/all isomers. [c] Yield determined by GC analysis.
Ms=methanesulfonyl, Ts=para-toluenesulfonyl.
With these optimized conditions in hand, we sought to
explore the scope of this transformation. A range of
substituted aryl electrophiles were subjected to the reaction
conditions (Scheme 2). A variety of substituents were toler-
ated, from electron-rich (3b) to electron-poor (3c, 3d), with
electron-poor substrates providing slightly slower reaction
rates, but excellent regioselectivities. Very electron-rich
products, however, were prone to isomerization upon purifi-
cation. Therefore, regioselectivities are reported before and
after purification for 3b and 3t.[20] Gratifyingly, reactions
involving counterion exchange with TESOTf to access the
cationic intermediate worked only slightly less efficiently than
simply beginning with the aryl triflate. Substitution at the
para- and meta-positions was, for the most part, also well
tolerated. Ortho-substitution resulted in lower yield and
a slightly reduced r.r. (3g).
Scheme 2. Scope of aryl electrophile with 1-octene. All yields shown
are of isolated products. Reaction conditions: for X=OTf: [Ni(cod)2]
(10 mol%), 1 (12 mol%), DABCO (3 equiv), THF (1m), 608C, 24 h.
For X =Cl, OMs, OTs: [Ni(cod)2] (10 mol%), 1 (12 mol%), DABCO
(5 equiv), TESOTf (2 equiv), PhMe (0.5m), 608C, 24 h. [a] r.r. before
purification. [b] 48 h. [c] [Ni(cod)2] (15 mol%), 1 (18 mol%), 1-octene
(3 equiv), 48 h. [d] TIPSOTf (2 equiv), 48 h. br/ln=ratio of 3 to linear
product 4 (GC), r.r. =ratio of 3 to all other isomers, mainly olefin
isomerization (GC). TIPS=triisopropylsilyl.
Although the electrophile scope was broad, we found that
substrates with para-alkyl groups (2e) suffered from reduced
yields and required longer reaction times. This puzzling
observation did not seem to stem from the presence of
À
benzylic C H bonds, as para-tBuPhOTf (2k) resulted in
almost no conversion, suggesting a steric phenomenon.
Nakamura and co-workers have proposed an explanation
for just such an effect: the rate-limiting precoordination of the
least hindered portion of the arene, which forms a p-complex
prior to oxidative addition, for Ni-catalyzed cross-coupling
reactions.[4,21] We sought to explore this mechanistic feature
further by preparing a variety of para-substituted aryl triflates
and subjecting them to the reaction conditions (Scheme 3).
Overall, the results are consistent with steric crowding of the
ligand cyclopentyl groups and the group in the para-position.
Substrates with groups in the meta-position (2 f, 2h) can
coordinate on the less sterically hindered side of the arene,
and react more quickly than those with para-substituents (as
also observed by Nakamura). However, when two meta-
substituents (2p) are introduced, the substrate can no longer
coordinate well, and the reaction does not proceed.
Scheme 3. Investigation of steric substitution effects. Results shown
are the conversion of triflate under standard conditions (24 h), as
determined by GC analysis.
and amine functional groups were compatible (3t, 3u),
although the presence of acidic protons (such as free alcohols
or ketones with enolizable protons) in the alkene led to
complete inhibition of the reaction. The transformation was
selective for terminal olefins in the presence of more
substituted alkenes (3s).
A range of aliphatic alkenes also successfully underwent
the desired transformation (Scheme 4), again using triflates or
chlorides/sulfonates with TESOTf. The presence of increased
steric bulk at the allylic position was well tolerated, although
extended reaction times were needed (3r). Protected alcohol
In summary, we have successfully developed a Ni-cata-
lyzed Mizoroki–Heck coupling of aryl triflates, chlorides, and
1860 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1858 –1861

Angewandte
Chemie
[2] a) The Mizoroki – Heck Reaction (Ed.: M. Oestreich), Wiley,
Chichester, UK, 2009; b) I. P. Beletskaya, A. V. Cheprakov,
Chem. Rev. 2000, 100, 3009 – 3066.
[3] a) S. Iyer, C. Ramesh, A. Ramani, Tetrahedron Lett. 1997, 38,
8533 – 8536; b) G. P. Boldrini, D. Savoia, E. Tagliavini, C.
Trombini, A. Umani Ronchi, J. Organomet. Chem. 1986, 301,
C62 – C64; c) A. R. Ehle, Q. Zhou, M. P. Watson, Org. Lett. 2012,
14, 1202 – 1205.
[4] B.-L. Lin, L. Liu, Y. Fu, S.-W. Luo, Q. Chen, Q.-X. Guo,
Organometallics 2004, 23, 2114 – 2123.
[5] B. M. Rosen, K. W. Quasdorf, D. A. Wilson, N. Zhang, A.-M.
Resmerita, N. K. Garg, V. Percec, Chem. Rev. 2011, 111, 1346 –
1416.
[6] a) R. Matsubara, A. C. Gutierrez, T. F. Jamison, J. Am. Chem.
Soc. 2011, 133, 19020 – 19023; b) E. A. Standley, T. F. Jamison, J.
Am. Chem. Soc. 2013, 135, 1585 – 1592.
[7] W. Cabri, I. Candiani, Acc. Chem. Res. 1995, 28, 2 – 7.
[8] a) F. Ozawa, A. Kubo, T. Hayashi, J. Am. Chem. Soc. 1991, 113,
1417 – 1419; b) J. Ruan, J. Xiao, Acc. Chem. Res. 2011, 44, 614 –
626; c) G. D. Daves, Jr., A. Hallberg, Chem. Rev. 1989, 89, 1433 –
1445.
[9] For example: 2:3 br/ln for enamides under traditional Heck
conditions, to > 100:1 under cationic Heck conditions. See
Ref. [7].
[10] K. Olofsson, M. Larhed, A. Hallberg, J. Org. Chem. 1998, 63,
5076 – 5079.
[11] L. Qin, X. Ren, Y. Lu, Y. Li, J. Zhou, Angew. Chem. 2012, 124,
6017 – 6021; Angew. Chem. Int. Ed. 2012, 51, 5915 – 5919.
[12] Y. Zou, L. Qin, X. Ren, Y. Lu, Y. Li, J. Zhou, Chem. Eur. J. 2013,
19, 3504 – 3511.
Scheme 4. Scope of the alkene coupling partner. All yields shown are
of isolated products. Reaction conditions: for X=OTf: [Ni(cod)2]
(10 mol%), 1 (12 mol%), DABCO (3 equiv), THF (1m), 608C, 24 h.
For X =Cl, OMs, OTs: [Ni(cod)2] (10 mol%), 1 (12 mol%), DABCO
(5 equiv), TESOTf (2 equiv), PhMe (0.5m), 608C, 24 h. [b] r.r. before
purification. [c] 48 h. br/ln=ratio of 3 to linear product 4 (GC),
r.r. =ratio of 3 to all other isomers, mainly olefin isomerization (GC).
[13] C. Zheng, D. Wang, S. S. Stahl, J. Am. Chem. Soc. 2012, 134,
16496 – 16499.
other sulfonates with electronically unbiased alkenes in good
yields. This reaction displays excellent branched/linear selec-
tivity for the coupled product, with overall regioselectivities
(of the desired product to all other isomers) that are ꢀ 19:1 in
all cases. This universally highly branched-selective Heck
reaction also leverages the intrinsic properties of Ni to allow
for the use of inexpensive, stable, and synthetically practical
chlorides and sulfonates as coupling partners. Though the cost
of [Ni(cod)2] is not insignificant, we hope to continue to
develop alternative catalysts or precatalysts from inexpensive
Ni sources.[6b,22] These developments continue to show the
promise of the Ni-catalyzed Heck reaction as a viable, highly
selective alternative to its Pd-catalyzed counterpart.
[14] In reporting regioselectivities of alkene addition, two methods
are common. In this account, we will use both the regioisomeric
ratio (r.r.), which refers to the ratio of the desired (branched)
product to all other isomers, and the branched/linear (br/ln)
ratio, which refers to the ratio of desired 3 to linear 4. Ratios
were determined by GC analysis, and the peak corresponding to
the linear product was confirmed by running the reaction in the
absence of phosphine ligand, (br/ln ratio of ca. 1:1).
[15] a) R. Matsubara, T. F. Jamison, J. Am. Chem. Soc. 2010, 132,
6880 – 6881; b) T. M. Gøgsig, J. Kleimark, S. O. Nilsson Lill, S.
Korsager, A. T. Lindhardt, P.-O. Norrby, T. Skrydstrup, J. Am.
Chem. Soc. 2012, 134, 443 – 452.
[16] D. L. Thorn, R. Hoffmann, J. Am. Chem. Soc. 1978, 100, 2079 –
2090.
[17] J. Ledford, C. S. Shultz, D. P. Gates, P. S. White, J. M. DeSimone,
M. Brookhart, Organometallics 2001, 20, 5266 – 5276.
[18] C. P. Casey, G. T. Whiteker, Isr. J. Chem. 1990, 30, 299 – 304.
[19] T. K. Macklin, V. Snieckus, Org. Lett. 2005, 7, 2519 – 2522.
[20] See the Supporting Information for details.
Received: September 25, 2013
Revised: November 19, 2013
Published online: January 8, 2014
[21] For theoretical and KIE studies, see: N. Yoshikai, H. Matsuda, E.
Nakamura, J. Am. Chem. Soc. 2008, 130, 15258 – 15259.
[22] A cod-free precatalyst of 1 (see Ref. [6b]) was synthesized, but
purification proved difficult. Impure samples of the precatalyst
were tested, but yields were significantly lower than with
[Ni(cod)2]/1.
Keywords: alkenes · Heck reaction · nickel · regioselectivity
.
[1] a) T. Mizoroki, K. Mori, A. Ozaki, Bull. Chem. Soc. Jpn. 1971,
44, 581; b) R. F. Heck, J. P. Nolley, Jr., J. Org. Chem. 1972, 37,
2320 – 2322.
Angew. Chem. Int. Ed. 2014, 53, 1858 –1861
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1861