Zinc catalysed electrophilic C-H borylation of heteroarenes

Article abstract of DOI:10.1039/d1sc01883c

Cationic zinc Lewis acids catalyse the C-H borylation of heteroarenes using pinacol borane (HBPin) or catechol borane (HBCat). An electrophile derived from [IDippZnEt][B(C₆F₅)₄] (IDipp = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene) combined withN,N-dimethyl-p-toluidine (DMT) proved the most active in terms of C-H borylation scope and yield. Using this combination weakly activated heteroarenes, such as thiophene, were amenable to catalytic C-H borylation using HBCat. Competition reactions show these IDipp-zinc cations are highly oxophilic but less hydridophilic (relative to B(C₆F₅)₃), and that borylation proceedsviaactivation of the hydroborane (and not the heteroarene) by a zinc electrophile. Based on DFT calculations this activation is proposed to proceed by coordination of a hydroborane oxygen to the zinc centre to generate a boron electrophile that effects C-H borylation. Thus, Lewis acid binding to oxygen sites of hydroboranes represents an under-developed route to access reactive borenium-type electrophiles for C-H borylation.

Full text of DOI:10.1039/d1sc01883c

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Zinc catalysed electrophilic C–H borylation of
heteroarenes†
Cite this: Chem. Sci., 2021, 12, 8190
*
All publication charges for this article
have been paid for by the Royal Society
of Chemistry
Matthew E. Grundy, Kang Yuan, Gary S. Nichol
and Michael J. Ingleson
Cationic zinc Lewis acids catalyse the C–H borylation of heteroarenes using pinacol borane (HBPin) or
catechol borane (HBCat). An electrophile derived from [IDippZnEt][B(C6F5)4] (IDipp ¼ 1,3-bis(2,6-
diisopropylphenyl)imidazol-2-ylidene) combined with N,N-dimethyl-p-toluidine (DMT) proved the most
active in terms of C–H borylation scope and yield. Using this combination weakly activated
heteroarenes, such as thiophene, were amenable to catalytic C–H borylation using HBCat. Competition
reactions show these IDipp–zinc cations are highly oxophilic but less hydridophilic (relative to B(C6F5)3),
and that borylation proceeds via activation of the hydroborane (and not the heteroarene) by a zinc
electrophile. Based on DFT calculations this activation is proposed to proceed by coordination of
a hydroborane oxygen to the zinc centre to generate a boron electrophile that effects C–H borylation.
Thus, Lewis acid binding to oxygen sites of hydroboranes represents an under-developed route to
access reactive borenium-type electrophiles for C–H borylation.
Received 5th April 2021
Accepted 12th May 2021
DOI: 10.1039/d1sc01883c
rsc.li/chemical-science
the solution to which would be facilitated by identifying novel
mechanisms to catalyse (hetero)arene C–H borylation.
Introduction
One earth abundant high PDE element that has had limited
use in catalysing arene borylation is zinc.27–30 While zinc cata-
lysed C–H borylation of alkynes has been reported,30,31 no zinc
complex that catalyses the C–H borylation of (hetero)arenes
using hydroboranes has been reported to our knowledge. As
zinc complexes are distinct to Fe/Co (zinc is redox neutral
during catalysis) and to boron Lewis acids, different borylation
mechanisms maybe accessible. Comparing zinc and boron
Lewis acids, zinc Lewis acids are “harder”, i.e. more oxophilic
and less hydridophilic, and contain more polarised d+Zn–Hdꢀ
and d+Zn–Cdꢀ bonds (than B–H/B–C). The latter facilitated
NHC–zinc catalysed alkyne C–H borylation with pinacolborane
(HBPin), with B–C bond formation proposed to occur via alkyne
deprotonation by a zinc hydride and then s-bond metathesis of
the Zn–C species with HBPin (Fig. 1A).31 However, another C–H
As it enables rapid construction of complex molecules, C–H
functionalisation has become increasingly important.1,2 Among
the most useful moieties made via C–H functionalisation are
C–B containing units, owing in part to the power of the Suzuki–
Miyaura reaction.3,4 This has made organoboranes ubiquitous
nucleophiles and provides continued incentive to develop new
routes to organoboranes,5 particularly via C–H borylation.6
Presently, the most common transformations of this type are
based on noble metal catalysts, most oen iridium.7–11 Base
metal catalytic alternatives are preferable,12,13 with arguably the
most powerful processes developed to date using cobalt cata-
lysts.14,15 However, both cobalt and iridium have extremely low
permitted daily exposure (PDE) values of 50 mg and 100 mg (oral
intake).16 Alternative C–H borylation catalysts based on earth
abundant elements that have higher PDE values are desirable.
The last decade has seen some notable progress in this area,
including in Fe catalysed arene C–H borylation,17–21 and in
catalytic C–H borylation using boron based frustrated Lewis
pairs (FLPs) and/or borenium cations.22–26 To date, these cata-
lytic processes are either limited in scope or require the arene
substrate to be present in super stoichiometric amounts.
Addressing these limitations remains a signicant challenge,
EaStCHEM School of Chemistry, University of Edinburgh, Edinburgh, EH9 3FJ, UK.
E-mail: michael.ingleson@ed.ac.uk
† Electronic supplementary information (ESI) available: Full experimental details,
NMR spectra, crystallographic data and Cartesian coordinates for all calculations.
CCDC 2074217–2074219. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/d1sc01883c
Fig.
borylation.
1 Proposed mechanisms for zinc catalysed alkyne C–H
8190 | Chem. Sci., 2021, 12, 8190–8198
© 2021 The Author(s). Published by the Royal Society of Chemistry

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Edge Article
Chemical Science
borylation mechanism may be more relevant for substrates less cationic zinc complexes. Mechanistic studies are consistent
acidic than alkynes, with no (hetero)arene deprotonation by with C–H borylation proceeding through coordination of a zinc
a zinc-hydride reported to date to the best of our knowledge.32 Lewis acid to an oxo site of a hydroborane to generate a bore-
One alternative mechanism involves activation of a hydro- nium equivalent. When combined with the optimal base this
borane by a zinc Lewis acid. This would enhance electrophilicity leads to a catalytic C–H borylation methodology applicable to
at boron and form a functional equivalent of a borenium cation, less activated heteroarenes (e.g. thiophenes) than normally
species well documented to borylate p systems.33 Such a mech- observed using main group (redox inactive) catalysts.
anism was outlined in zinc catalysed alkyne borylation with
HBDan (HBDan ¼ 1,8-naphthalenediaminatoborane), with zinc
Results and discussion
Borylation with pinacolborane (HBPin)
proposed to interact with the B–H unit (Fig. 1B), with pyridine
then enabling C–H deprotonation.30 This can be viewed as
a frustrated Lewis pair (FLP) borylation mechanism.23 It should
be noted that related intermediates (e.g. [Zn]/H/SiR3) have
been proposed in zinc catalysed C–H silylation of (hetero)
arenes.34,35
Due to its efficacy in the C–H borylation of terminal alkynes,31
7DippZnH(NTf2) (1-H, Fig. 3) was trialled for catalysing heter-
oarene C–H borylation (7Dipp
¼
1,3-bis(2,6-diisopropyl-
N–Me–
phenyl)-4,5,6,7-tetrahydro-1,3-diazepin-2-ylidene).
Hydroboranes, such as HBPin, have multiple Lewis basic
sites, specically the B–H and 2x O, thus Lewis acids also can
coordinate at an oxygen of HBPin. Indeed, Lewis acid coordi-
nation to oxygen has been proposed in the hydroboration of
ketones catalysed by highly oxophilic Lewis acids (e.g. Ca and
Mg complexes, Fig. 2A).36 Regardless of binding site, to enable
Lewis acid coordination to weakly Lewis basic hydroboranes
requires exclusion of signicantly stronger Lewis bases.
However, bases oen play a crucial role in facilitating depro-
tonation steps during C–H borylation using main group elec-
trophiles.23,26 Therefore, identifying a Brønsted base that when
combined with a zinc Lewis acid functions as a FLP maybe
essential to facilitate C–H deprotonation and ultimately bor-
ylation (particularly given the lack of precedence for Zn–H units
acting as effective Brønsted bases towards heteroarenes).32
While some notable work on zinc based FLPs has been reported,
their use to date has been limited (Fig. 2B),37 and they have not
been applied in arene C–H functionalisation to our knowledge.
Herein we present the rst, to our knowledge, report of zinc
complexes that catalyse the C–H borylation of heteroarenes
using HBPin and HBCat. This process utilises NHC-supported
Indole (2a) was chosen as the initial substrate due to its high
nucleophilicity (N ¼ +5.75 on the Mayr nucleophilicity index)
and high reactivity in SEAr reactions.38 The stoichiometric
reaction of 1-H (generated in situ from 1-Ph and HBPin)31 with
2a and HBPin led to C–H borylation (by 1H and 11B NMR
spectroscopy), with effectively full conversion of 2a to the C3
borylated indole (3a) at 100 ꢁC (Table 1, entry 1). 1H NMR
spectroscopy conrmed the presence of 1-H in the reaction
mixture post full consumption of 2a indicating the feasibility of
turnover. At 10 mol% loading of 1-Ph 78% borylation of 2a in
chlorobenzene (PhCl) was observed over 36 h (entry 2).
As NTf2 is known to coordinate to NHC–zinc cations in hal-
oarene solvents we decided to probe anion effects.31 The meth-
odology of Dagorne and co-workers39 was utilised to access
[7DippZnEt][B(C6F5)4] (4) and [IDippZnEt][B(C6F5)4] (5) (Fig. 3) by
alkyl abstraction using [CPh3][B(C6F5)4] from the respective
NHCZnEt2 precursors. Compound 5 has been shown previously
by Dagorne et al. to be active in catalytic hydrosilylation with
reactivity presumably proceeding via [Zn]/H/SiR3 species that
react as silicon based electrophiles.39 Assessment of 4 and 5 as
(pre)catalysts showed that both effect catalytic C–H borylation
and are more active than the NTf2 analogue, with compound 5
providing the better outcome (entries 3 and 4). Further testing
showed that changing the solvent to C6D6 or toluene using 5 had
minimal effect (entries 5 and 6), whilst the loading of 5 could be
Fig. 2 (A) Hydroboration mediated by oxophilic Lewis acids. (B)
Previous work on zinc based FLPs. (C) This work, [Zn]/NR3 catalysed
C–H borylation of a range of heteroarenes.
Fig. 3 Complexes 1, 4–6.
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 8190–8198 | 8191

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Chemical Science
Edge Article
Table 1 Catalyst optimization
indolines, were also observed by NMR spectroscopy as minor
by-products in a number of cases alongside formation of 3x. The
borylation of other less nucleophilic substrates using 5/HBPin
was unsuccessful, with borylation not observed for anisole, 2-
methylthiophene and 2-methylfuran under these conditions.
Furthermore, N–Me–pyrrole, 3,4-ethylene-dioxythiophene and
6-uoro-N–Bn–indole gave only 5%, 14% and 4% borylation,
ꢁ
respectively, aer 18 h at 80 C.
Entry [Cat] Catalyst loading/mol% T/ꢁC Solv. Time/h Conv.a/%
Other main group catalysed arene C–H borylation reactions
are facilitated by addition of a Brønsted base that deprotonates
the arene in a concerted or stepwise SEAr mechanism.23,26
Indeed, as noted by Oestreich and co-workers, the in situ
formation of indolines during indole borylation may well be
vital in enabling catalytic C–H borylation of indoles.25b This is
potentially due to indolines being stronger Brønsted bases
(than indoles) and thus enable access to a FLP type borylation
mechanism.23 Therefore we investigated basic additives in this
C–H borylation reaction. The addition of a hindered base, 2,6-
di-tert-butyl-4-methyl-pyridine (DBP), to preclude any Zn ) Npy
bond formation, was found to reduce the conversion of 2a to 3a
(Table 1, entry 12). This is attributed to DBP sequestering
a proton from an SEAr reaction forming a weaker Brønsted acid
(relative to that otherwise present during catalytic electrophilic
borylation). This is hypothesised to slow a dehydrocoupling
step of [Base-H]+ with a Zn–H or a B–H species (forming H2) that
leads to regeneration of an on-cycle electrophile. In contrast, the
addition of the weaker base N,N-dimethyl-p-toluidine (DMT,
Scheme 1) at a 1 : 1 ratio with respect to 5 led to an acceleration
in the borylation of 2a (Table 1, entry 8 vs. 13). In addition, this
combination enabled the borylation of the less nucleophilic
substrate 2-methyl thiophene with formation of 31% of 7
(Scheme 2) aer 18 h. In contrast, no formation of 7 was
observed in the absence of DMT (by 1H NMR spectroscopy). This
conrms the importance of DMT, presumably functioning as
1
2
3
4
5
6
7
8
1-H
1-Ph
4
5
5
5
5
5
100
10
10
10
10
10
5
5
10
5
—
5
5
10
5
100 PhCl 18
100 PhCl 36
80 PhCl 36
80 PhCl 18
80 C6D6 18
80 C7H8 18
80 PhCl 18
80 PhCl 10
80 PhCl 18
80 PhCl 18
80 PhCl 18
80 PhCl 18
80 PhCl 10
80 PhCl 18
80 PhCl 18
80 PhCl 18
100
78b
47
97
80
91
94
82
0
9
6
10
11
12
13
14
15
16
ZnCFc
—
0
0
5
62d
92e
0
5
[8]+f
g
LBH3
IDipp 10
0
0
a
Conversions estimated in situ by 1H NMR spectroscopy of diagnostic
b
product resonances versus those of N–Me–indole. 2.3 equivalents of
c
d
HBPin used. ZnCF ¼ Zn(C6F5)2. With 1 equivalent of DBP added.
e
f
With 5 mol% of N,N-dimethyl-p-toluidine added. [8][B(C6F5)4] was
g
generated in situ before addition of HBpin/2a. L ¼ THF used as
a 1 M THF solution.
decreased to 5 mol% with high yielding borylation still observed
(entries 7 and 8). Use of the neutral complex IDippZnEt2 (6) as
a catalyst showed that the cationic nature of 5 is important for
catalytic borylation (entry 9). While use of a NHC free, neutral
zinc Lewis acid, Zn(C6F5)2, also led to no borylation of 2a (entry
10). In this case signicant Zn–C/H–B metathesis (to form C6F5–
BPin) was observed along with the precipitation of an insoluble
solid, presumably ZnH2. Finally, in the absence of a zinc catalyst
no borylation was observed (entry 11).
Next, the scope of borylation using 5 was surveyed, and
found to be limited to activated indoles (e.g. 3a–3d, Fig. 4), with
unactivated indoles (e.g. brominated indoles) not borylated. It
should be noted that products from the reduction of 2x, N–R–
Scheme 1 DMT dependent borylation of 2-methylthiophene with
HBPin.
Scheme 2 Left) No borylation observed using [IDippBPin]+. (Right) The
Fig. 4 C3–H borylation, yields by NMR spectroscopy versus an structure of the cationic portion of [8][NTf2], ellipsoids at 50% and H
internal standard.
atoms omitted.
8192 | Chem. Sci., 2021, 12, 8190–8198
© 2021 The Author(s). Published by the Royal Society of Chemistry

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Edge Article
Chemical Science
a Brønsted base during the borylation process. Furthermore, it
is a rare example of a main group electrophile catalysing the
C–H borylation of a less activated heteroarene, such as 2-methyl
thiophene (N ¼ +1.35),40 using HBPin directly.23 However, the
low yield of 7 aer 18 h with 5/DMT indicates a limit in the
heteroarene nucleophiles viable for borylation using this
system. Thus mechanistic insight was sought to enable expan-
sion of scope.
Scheme 3 Attempted metalation reactions between [Zn-R]+ and 2a.
Mechanistic studies
zinc centres, suggesting an oligomeric structure for putative
[IDippZnH]+.45
Zinc catalysed C–H borylation can proceed via a Zn electro-
phile interacting with 2a or HBPin. With the former, this would
form a species related to A (Scheme 3). In this the Brønsted
acidity of the indole–C3 proton will be enhanced.
The rst key question to answer was if the catalysis was Zn
mediated or simply Zn initiated (e.g. forming a boron electro-
phile which is an on-cycle species, e.g. a [PinB(base)]+/base
FLP).23 It should be noted that while [B(C6F5)4]ꢀ is used as the
counterion in 5, no decomposition of this anion to B(C6F5)3
(which can catalyse electrophilic C–H borylation and silylation)
11
was observed based on 19F and B NMR spectroscopy.41
Deprotonation (by Zn–H or another base, e.g. N–Me–indo-
line) would form a zinc indolyl complex (e.g. B or C). Subsequent
metathesis with HBPin then would generate the indole boronic
ester, 3a. With no Zn–H species derived from 5 isolable in our
hands 1-H was used in stoichiometric reactions. Combining 1-H
with 2a led to no C–H metalation, even under forcing condi-
tions. Furthermore, combining 5, 2a and DMT (1 : 1 : 1) also led
Furthermore, C–H borylation also proceeded with NTf2 as the
counterion (e.g. with 1-H), disfavouring a process mediated by
B(C6F5)3. At this point in the study a zinc Lewis acid acting as
a hydride acceptor enabling formation of a borenium cation
could not be precluded. Indeed, crystals of borenium ion
[IDippBPin][NTf2] ([8][NTf2]) were isolated from one catalytic
reaction mixture, albeit in a small amount. This indicates that
borenium electrophiles can be formed under catalytic condi-
tions (it should be noted [PinB(L)]+ boreniums are relatively
weak hydridophiles, thus hydride transfer from PinBH(L) to
a zinc cation is feasible – vide infra).42 [8]+ is presumably formed
by NHC dissociation from zinc, NHC coordination to HBPin,
followed by hydride abstraction from (NHC)HBPin – all re-
ported steps.31,43 The structure of [8]+ (Scheme 2, right) is
unremarkable compared to [(NHC)BCat]+,44 excluding shorter
B–O bonds in [8]+ due to the improved p donor ability of
pinacol relative to catechol. Importantly, a control reaction
using 10 mol% of [8]+, generated in situ from (IDipp)HBPin/
B(C6F5)3 (by 19F/11B NMR spectroscopy), as catalyst led to no
borylation of 2a with HBPin under identical catalytic conditions
(table 1 entry 14). Furthermore, 5 mol% BH3-THF (1 M in THF,
entry 15), and 10 mol% of IDipp (entry 16), as potential initia-
tors led to no borylation of 2a under the catalytic conditions.
This indicates a zinc complex is an on cycle species, with
a [IDippZnH]+ species proposed to be key using 5. This is due to
metathesis of the Zn–Et unit in 5 with HBPin being rapid rela-
tive to C–H borylation (formation of EtBPin occurs before
formation of signicant amounts of 3a). However, despite
numerous attempts, in our hands formation of an isolable
[NHCZnH][B(C6F5)4] species from combinations of 5 (or 4) with
HBPin (and via other routes) proved elusive. Formulation as
a zinc hydride was supported by studies combining 5 with
HBPin and DBPin; using HBPin new singlets grow in as
ꢁ
to no C–H zincation of 2a on heating to 80 C. These observa-
tions disfavour a borylation mechanism in which activation of
N–Me–indole occurs rst by interaction with a Lewis acidic zinc-
species.
It has been previously observed that 1-H reacts with HBPin
on heating to give unidentied zinc containing species,31b
indicating an interaction between a NHC–Zn Lewis acid and
HBPin can occur. To probe the interaction between 1-H and
HBPin further, 1-H was reacted with DBPin which led to H/D
scrambling at room temperature. The Zn–H resonance of 1-H
(3.66 ppm) decreases in intensity with the concurrent appear-
ance of a singlet resonance, assigned as Zn-D in 1-D, in the 2H
NMR spectrum (3.69 ppm). Additionally, the 11B NMR spectrum
now showed two species, one assigned as H-BPin (doublet) and
the other D-BPin (broad singlet). This conrms the borane and
complex 1-H interact at room temperature. Hydrogen scram-
bling between 1-H and HBPin can feasibly occur via H–B or O-
bound isomers (e.g. D or E, Scheme 4).
As C–H borylation is more effective in the presence of DMT
the reactivity of DMT towards 5 also was explored. The addition
of varying equivalents (from 0.5 to 2 equivalents) of DMT to 5
1
metathesis proceeds at 2.7 and 1.5 ppm in the H NMR spec-
trum, tentatively assigned as Zn–H. Notably, these resonances
1
are not observed in the H NMR spectrum when using DBPin,
although the by-product from metathesis, EtBPin, is present.
The putative hydride chemical shis are comparable to other
cationic NHC–ZnH species in which hydrides are bridging two
Scheme 4 H/D exchange at room temperature via possible inter-
mediates D or E.
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 8190–8198 | 8193