Robust Cobalt Catalyst for Nitrile/Alkyne [2+2+2] Cycloaddition: Synthesis of Polyarylpyridines and Their Mechanochemical Cyclodehydrogenation to Nitrogen-Containing Polyaromatics**

Article abstract of DOI:10.1002/anie.202017220

The transition-metal-catalyzed [2+2+2] cycloaddition of nitriles and alkynes is an established synthetic approach to pyridines; however, these cycloadditions often rely on the use of tethered diynes or cyanoalkynes as one of the reactants. Thus, examples of efficient, fully intermolecular catalytic [2+2+2] pyridine synthesis, especially those employing unactivated nitriles and internal alkynes leading to pentasubstituted pyridines, remain scarce. Herein, we report on simple and inexpensive catalytic systems based on cobalt(II) iodide, 1,3-bis(diphenylphosphino)propane, and Zn that promote [2+2+2] cycloaddition of various nitriles and diarylacetylenes for the synthesis of a broad range of polyarylated pyridines. DFT studies support a reaction pathway involving oxidative coupling of two alkynes, insertion of the nitrile into a cobaltacyclopentadiene, and C-N reductive elimination. The resulting tetra- and pentaarylpyridines serve as precursors to hitherto unprecedented nitrogen-containing polycyclic aromatic hydrocarbons via mechanochemically assisted multifold reductive cyclodehydrogenation.

Full text of DOI:10.1002/anie.202017220

Angewandte
Research Articles
Chemie
How to cite: Angew. Chem. Int. Ed. 2021, 60, 9627–9634
International Edition: doi.org/10.1002/anie.202017220
German Edition: doi.org/10.1002/ange.202017220
Cycloadditions
Robust Cobalt Catalyst for Nitrile/Alkyne [2+2+2] Cycloaddition:
Synthesis of Polyarylpyridines and Their Mechanochemical
Cyclodehydrogenation to Nitrogen-Containing Polyaromatics**
Chang-Sheng Wang+, Qiao Sun+, Felipe García, Chen Wang,* and Naohiko Yoshikai*
Dedicated to Professor E. Peter Kündig on the occasion of his 75th birthday
Abstract: The transition-metal-catalyzed [2+2+2] cycloaddi-
tion of nitriles and alkynes is an established synthetic approach
to pyridines; however, these cycloadditions often rely on the
use of tethered diynes or cyanoalkynes as one of the reactants.
Thus, examples of efficient, fully intermolecular catalytic
[2+2+2] pyridine synthesis, especially those employing unac-
tivated nitriles and internal alkynes leading to pentasubstituted
pyridines, remain scarce. Herein, we report on simple and
inexpensive catalytic systems based on cobalt(II) iodide, 1,3-
bis(diphenylphosphino)propane, and Zn that promote
[2+2+2] cycloaddition of various nitriles and diarylacetylenes
for the synthesis of a broad range of polyarylated pyridines.
DFT studies support a reaction pathway involving oxidative
coupling of two alkynes, insertion of the nitrile into a cobalta-
cyclopentadiene, and C-N reductive elimination. The resulting
tetra- and pentaarylpyridines serve as precursors to hitherto
unprecedented nitrogen-containing polycyclic aromatic hydro-
carbons via mechanochemically assisted multifold reductive
cyclodehydrogenation.
sively explored with the aid of various transition metal
catalysts including cobalt,[6,7] nickel,[8] iron,[9] ruthenium,[10]
rhodium,[11] and iridium,[12] in addition to stoichiometric
variants using zirconium[13] and titanium[14] reagents. Despite
the tremendous developments in the last decades, the
majority of efficient catalytic [2+2+2] pyridine formations
take advantage of using tethered diyne or cyanoalkyne as one
of the starting materials to reduce the entropic penalty as well
as to address the issues of chemoselectivity (e.g., competition
with alkyne [2+2+2] cyclotrimerization) and regioselectivity
(Scheme 1a). Thus, examples of fully intermolecular [2+2+2]
pyridine synthesis from unactivated nitriles and unactivated
internal alkynes are surprisingly scarce. This problem, espe-
cially with respect to the cycloaddition involving diarylacety-
lenes, warrants attention in light of the significant interest in
polyarylated pyridines for their optical/electronic properties
and relevant applications.[15] In fact, literature reports on the
synthesis of pentarylpyridines through [2+2+2] cycloaddition
have been limited to those promoted by Cp*Co(h3-allyl)[16] or
palladium catalysts,[17] with only one or two examples for each
catalyst (Scheme 1b). During the preparation of this manu-
script, Wang and co-workers reported a catalytic system
comprised of iron and cobalt complexes for such cyclo-
addition, albeit using only diphenylacetylene and a limited
scope of aryl nitriles.[18]
As a part of our pursuit of low-valent cobalt-diphosphine
catalyzed synthetic transformations,[19,20] we report herein
that cobalt(II) iodide, 1,3-bis(diphenylphosphino)propane
(dppp), and Zn in polar solvents give rise to simple,
inexpensive, and highly robust catalytic systems for the fully
intermolecular [2+2+2] cycloaddition of various aryl and
alkyl nitriles with diarylacetylenes (Scheme 1c). The present
systems allow for the synthesis of a diverse range of
polyarylated pyridines without using large excess of the
nitrile with respect to the alkyne (nitrile:alkyne = 1:1 to
1.5:1). Computational studies have suggested that the present
[2+2+2] cycloaddition is initiated by oxidative coupling of
two alkynes on cobalt, followed by nitrile insertion into
cobaltacyclopentadiene and C-N reductive elimination.
We have also demonstrated multifold cyclodehydrogenation
of tetra- and pentaarylpyridines enabled under mechano-
chemical conditions to form hitherto unprecedented types
of nitrogen-containing polycyclic aromatic hydrocarbons
(PAHs).
Introduction
Transition-metal-catalyzed [2+2+2] cycloaddition be-
tween a nitrile and two alkynes represents a straightforward
and atom-economical approach to the synthesis of multi-
substituted pyridines,[1] which has found a number of appli-
cations in the synthesis of pyridine-containing complex
natural products,[2] bioactive molecules,[3] and extended p-
conjugated materials,[4] to name a few. Since the pioneering
work of Wakatsuki and Yamazaki on CpCo complex-medi-
ated/catalyzed reactions,[5] this cycloaddition has been exten-
[*] Dr. C.-S. Wang,[+] Dr. Q. Sun,[+] Prof. F. García, Prof. N. Yoshikai
Division of Chemistry and Biological Chemistry, School of Physical
and Mathematical Sciences, Nanyang Technological University
Singapore 637371 (Singapore)
E-mail: nyoshikai@ntu.edu.sg
Prof. C. Wang
Zhejiang Key Laboratory of Alternative Technologies for Fine
Chemical Process, Shaoxing University
Shaoxing 312000 (P. R. China)
E-mail: wangchen@usx.edu.cn
[+] These authors contributed equally to this work.
[**] A previous version of this manuscript has been deposited on
a preprint server (https://doi.org/10.26434/chemrxiv.13465697.v1).
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202017220.
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634
ꢀ 2021 Wiley-VCH GmbH
9627

Angewandte
Research Articles
Chemie
Table 1: [2+2+2] Cycloaddition of benzonitrile (1a) and diphenylacety-
lene (2a).[a]
Entry
CoXn
Ligand
Solvent
T [8C]
Yield [%][b]
1
2
3
4
5
6
7
CoI2
CoI2
CoI2
CoI2
CoCl2
CoBr2
CoCl(PPh3)3
CoCl(PPh3)3
CoI2
CoI2
CoI2
CoI2
CoI2
CoI2
CoI2
dppp
dppp
dppp
dppp
dppp
dppp
dppp
dppp
dppm
dppe
dppb
dppbz
PPh3
NMP
DMF
toluene
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
NMP
120
120
120
80
80
80
80
80
80
80
80
80
80
80
80
53
47[c]
0
82 (80)
25
54
45
0
4
0
0
8[d]
9
10
11
12
13[e]
14[f]
15[g]
2
0
dppp
dppp
81 (79)
52
[a] Unless otherwise noted, the reaction was performed using 1a
(0.5 mmol), 2a (0.2 mmol), CoXn (10 mol%), ligand (10 mol%), and Zn
(50 mol%) in 0.3 mL of solvent for 24 h. dppp=1,3-bis(diphenylphos-
phino)propane; dppm=bis(diphenylphosphino)methane; dppe=1,2-
bis(diphenylphosphino)ethane; dppb=1,4-bis(diphenylphosphino)bu-
tane; dppbz=1,2-bis(diphenylphosphino)benzene. [b] Determined by
GC using n-tridecane as an internal standard; isolated yield is shown in
the parentheses. [c] The hydrocarbamoylation product (addition of DMF
to 2a) was obtained in 13% yield. [d] The reaction was performed
without Zn. [e] 20 mol% of PPh3 was used. [f] 0.2 mmol of 1a and 0.2 mL
of NMP were used. [g] 0.1 mmol of 1a and 0.2 mL of NMP were used.
isolated yield; entry 4). The use of CoCl2 or CoBr2 in place of
CoI2 led to a substantial drop in the yield of 3aa (entries 5 and
6). A well-defined CoI complex, CoCl(PPh3)3, also served as
a viable precatalyst under otherwise identical conditions
(entry 7).[21] However, this complex failed to promote the
reaction when Zn was omitted (entry 8), implying that
reduction of CoI to Co0 was necessary. Screening of various
ligands revealed that the present catalytic activity is unique to
dppp. Thus, common diphosphines such as dppm, dppe, dppb,
and dppbz as well as monophosphines such as PPh3 barely
promoted the formation of either 3aa or hexaphenylbenzene
(entries 9–13; see Table S1 for additional results). It was
particularly surprising to us that the CoI2/dppe/Zn system was
completely ineffective (entry 10), as it is rather similar to the
systems developed by Cheng and Okamoto for the pyridine
formation using tethered substrates.[7b–d] The ratio of 1a and
2a could be reduced to 1:1 without an apparent decrease in
the yield of 3aa (entry 14), while further reduction to 1:2
(which corresponds to the theoretical stoichiometry) led to
a drop to 52% yield (entry 15).
Scheme 1. Transition-metal-catalyzed [2+2+2] cycloaddition of nitrile
and alkynes.
Results and Discussion
The present study commenced with screening of reaction
conditions for the cycloaddition of benzonitrile (1a,
0.5 mmol) with diphenylacetylene (2a, 0.2 mmol). In the
initial attempts,
a catalytic system comprised of CoI2
(10 mol%), dppp (10 mol%), and Zn dust (50 mol%) was
found to promote the desired cycloaddition in NMP at 1208C
to afford, after 24 h, pentaphenylpyridine (3aa) in 53% yield
(Table 1, entry 1), along with a small amount (4%) of
competitive cyclotrimerization product of 2a (i.e. hexaphe-
nylbenzene). The reaction also took place in DMF to give 3aa
in a comparable yield, which was accompanied by a small
amount of byproduct arising from the addition of DMF to 2a
With the optimized catalytic system (Table 1, entry 14) in
hand, we set out to explore the scope of the [2+2+2]
cycloaddition. First, a variety of aryl nitriles were subjected to
the reaction with 2a (Table 2). The reaction of 1a could be
performed on a 3 mmol scale to afford 3aa in 83% yield
[20f]
À
via carbamoyl C H activation (entry 2).
No pyridine
formation was observed in toluene (entry 3). Lowering the
temperature to 808C improved the yield of 3aa to 82% (80%
9628 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634

Angewandte
Research Articles
Chemie
Table 2: [2+2+2] Cycloaddition of various aryl nitriles with 2a.[a]
Table 3: [2+2+2] Cycloaddition of 1a with various alkynes.[a]
Yield [%][b]
Entry
Ar
Product
Yield [%][b]
Entry
R
Product
1
2
3
4
5
6
7
8
Ph (1a)
3aa
3ba
3ca
3da
3ea
3 fa
3ga
3ha
3ia
3ja
3ka
3la
3ma
3na
83[c]
72
61
62
76
61[d]
64
78
43
52
68
54
65
76
1
4-MeC6H4 (2b)
4-tBuC6H4 (2c)
4-MeOC6H4 (2d)
4-FC6H4 (2e)
4-EtO2CC6H4 (2 f)
4-Me(O)CC6H4 (2g)
3-MeC6H4 (2h)
3-MeOC6H4 (2i)
3-Me2NC6H4 (2j)
3,5-(MeO)2C6H3 (2k)
Pr (2l)
3ab
3ac
3ad
3ae
3af
3ag
3ah
3ai
65
67
57
71
63
41
51
84
66
86
63
2[c]
3[c]
4
4-MeC6H4 (1b)
4-MeOC6H4 (1c)
4-F3CC6H4 (1d)
4-Me(O)CC6H4 (1e)
4-NCC6H4 (1 f)
3-MeC6H4 (1g)
3-MeOC6H4 (1h)
2-MeC6H4 (1i)
2-FC6H4 (1j)
1-naphthyl (1k)
2-thienyl (1l)
2-pyridyl (1m)
3-pyridyl (1n)
5
6
7
8
9
9
3aj
3ak
3al
10
11
12
13
14
10
11[d]
[a] Unless otherwise noted, the reaction was performed using 1a
(0.4 mmol), alkyne 2 (0.4 mmol), CoI2 (10 mol%), dppp (10 mol%), and
Zn (50 mol%) in 0.4 mL of NMP at 808C for 24 h. [b] Isolated yield.
[c] The reaction was performed using 0.6 mmol of 1a for 72 h. [d] The
reaction was performed using bis(dicyclohexylphosphino)methane
(dcypm) and DME in place of dppp and NMP, respectively.
[a] Unless otherwise noted, the reaction was performed using nitrile
1 (0.4 mmol), 2a (0.4 mmol), CoI2 (10 mol%), dppp (10 mol%), and Zn
(50 mol%) in 0.4 mL of NMP at 808C for 24 h. [b] Isolated yield.
[c] Isolated yield for a 3 mmol scale reaction. [d] 0.2 mmol of tereph-
thalonitrile (1 f) and 0.6 mmol of 2a were used. Twofold cycloadduct was
obtained in 7% yield.
of ethyl 4-((4-methoxyphenyl)ethynyl)benzoate afforded an
inseparable mixture of multiple pyridine regioisomers, de-
spite the electronic difference between the two aryl groups
(see Figure S2). The CoI2/dppp system failed to engage 4-
octyne (2l) in the cycloaddition with 1a. However, a modified
catalytic system employing bis(dicyclohexylphosphino)me-
thane (dcypm) and 1,2-dimethoxyethane (DME) promoted
the reaction to afford the desired pyridine 3al in 63% yield
(entry 11; see Table S2 for the optimization).
The present catalytic system also proved effective for the
cycloaddition of di(heteroaryl)acetylenes, as demonstrated by
the synthesis of structurally attractive, symmetric penta(he-
teroaryl)pyridines (Scheme 2). Thus, the cycloaddition of 2-
cyanothiophene (1l) and di(2-thienyl)ethyne (2m) afforded
(entry 1). Benzonitriles bearing various electron-donating or
-withdrawing substituents (1b–1 f) participated in the reac-
tion to afford the desired products 3ba–3 fa in 61–76% yields
(entries 2–6). Note that terephthalonitrile (1 f) was used as
the limiting reagent with respect to 2a in a ratio of 1:3,
affording the cycloadduct 3 fa in 61% yield along with a small
amount (7%) of twofold cycloaddition product (entry 6). The
reaction also tolerated meta- (1g and 1h) and ortho-substi-
tuted (1i and 1j) benzonitriles (entries 7–10), while the latter
reacted somewhat sluggishly probably due to the steric
hindrance. 1-Naphthonitrile (1k) smoothly underwent the
cycloaddition to give the product 3ka in 68% yield (entry 11).
Heteroaryl nitriles (1l–1n) were also well tolerated, affording
thienyl- or pyridyl-substituted pentaarylpyridines 3la–3na
(entries 12–14).
We next explored the reaction of 1a with various alkynes
(Table 3). Symmetrical diarylacetylenes bearing electron-
donating (alkyl, OMe) or -withdrawing (F, CO2Et, COMe)
substituents at the para-positions took part in the cyclo-
addition to afford the corresponding products 3ab–3ag in
moderate to good yields (entries 1–6), where only a small
amount of the alkyne cyclotrimerization product was indi-
cated by TLC analysis of the reaction mixture. Among these
substituents, t-butyl and methoxy groups made the reaction
somewhat sluggish, necessitating an increase in the nitrile/
alkyne ratio to 1.5:1 and a prolonged reaction time of 72 h
(entries 2 and 3). The reaction also tolerated diarylacetylenes
bearing meta-substituents (2h–2k), affording pentaarylpyr-
idines 3ah–3ak in moderate to high yields (entries 7–10),
while those bearing ortho-substituents such as 1,2-di-o-
tolylethyne failed to undergo the cycloaddition. The reaction
Scheme 2. Synthesis of penta(heteroaryl)pyridines.
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org 9629

Angewandte
Research Articles
Chemie
penta(2-thienyl)pyridine (3lm) in 80% yield, which was
previously synthesized via sequential Stille couplings[15a] or
fivefold Suzuki–Miyaura coupling of pentachloropyridine.[15c]
Penta(3-pyridyl)pyridine (3nn) could also be prepared from
3-cyanopyridine (1n) and di(3-pyridyl)ethyne (2n) albeit in
a modest yield.
The scope of the [2+2+2] cycloaddition was further
extended to alkyl nitriles (Table 4). We noted that the above
catalytic system was only moderately effective for the
reaction of MeCN (1o) and 2a. With further optimization
(see Table S3), a modified catalytic system comprised of CoI2
(5 mol%), dppp (5 mol%), and Zn (25 mol%) in DMA was
found to efficiently promote the reaction, at the expense of
large excess MeCN (1o:2a = 5:1), affording the desired
product 3oa in 90% yield. MeCN could also be coupled with
para-substituted diphenylacetylenes, affording 3ob and 3oc
in good yields. The reaction using CD3CN afforded 3oa-d3 in
92% yield. The reaction of propionitrile (1p) and 2a was
subjected to additional optimization (see Table S4) and was
achieved in high yield (89%) at a lower nitrile/alkyne ratio of
1.5:1. This ratio was effective for a variety of other alkyl
nitriles. Sterically unhindered alkyl and benzyl nitriles,
including those bearing an olefin moiety, participated in the
reaction with 2a to afford the desired pyridine derivatives
3qa–3ua in high yields. The reaction also tolerated sterically
hindered secondary alkyl nitriles (see 3va and 3wa) as well as
nitriles containing dialkylamino, ester, and pyridyl groups
(see 3xa–3za). Note that, among these 2-alkyl-3,4,5,6-tetraar-
ylpyridines, only 3oa, 3pa, and 3qa are known in the
literature, reflecting the scarcity of relevant [2+2+2] cyclo-
additions.[16,18] A cyanamide derivative such as morpholine-4-
carbonitrile (1aa) also reacted with 2a to afford the 2-
aminopyridine 3aaa in an excellent yield. The reaction of 1p
and 2l was achieved using the CoI2/dcypm system to give the
pentaalkylpyridine 3pl in a moderate yield.
Table 4: [2+2+2] Cycloaddition of alkyl nitriles.[a]
The applicability of the present catalytic system to
tethered substrates was briefly examined (Scheme 3). A
diyne capped with phenyl groups (4) underwent cycloaddition
with 1a to afford the pentasubstituted pyridine 5 in 89%
yield. Note that, despite the extensive previous studies,
bicyclic pyridines bearing 2,3,6-triaryl substituents such as 5
have not been synthesized by bimolecular [2+2+2] cyclo-
addition. The reaction between a tethered cyanoalkyne 6 and
2a also proved feasible, affording the 2,3,4-triarylpyridine
derivative 7 albeit in a modest yield. This type of bimolecular
cycloaddition is also rare.[8e]
Scheme 3. [2+2+2] Cycloaddition using tethered substrates.
To gain insight into the reaction pathway of the present
[2+2+2] cycloaddition, we turned to density functional theory
(DFT) calculations. The [2+2+2] pyridine formation pro-
moted by CpCoI catalysts has been studied computationally,
and reaction pathways initiated by oxidative coupling of two
alkyne molecules on CoI have been identified.[7f, 22] On the
other hand, the working mode of cobalt-diphosphine catalysts
remains unexplored, while oxidative coupling of alkynes on
CoI has been proposed.[7b–d] Here, we explored the cyclo-
addition of 1a and 2a on two model low-valent cobalt
[a] Unless otherwise noted, the reaction was performed using nitrile
1 (0.6 mmol), alkyne 2 (0.4 mmol), CoI2 (5 mol%), dppp (5 mol%), and
Zn (25 mol%) in 0.4 mL of DMA at 808C for 24 h. The yields refer to
isolated yields. [b] 2 mmol of nitrile was used. [c] The reaction was
performed for 72 h. [d] CoI2 (10 mol%), dppp (10 mol%), and Zn
(50 mol%) were used. [e] The reaction was performed using propionitrile
(2 mmol), 4-octyne (0.4 mmol), CoI2 (10 mol%), dcypm (10 mol%), and
Zn (50 mol%) in DME at 1008C.
9630 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634

Angewandte
Research Articles
Chemie
Scheme 4. Summary of the computationally revealed reaction pathway.
complexes, namely, [(dppp)Co0] and [(dppp)CoI]+ (see the
Supporting Information for detail).[23,24] In short, these studies
have shown the following three major points (Scheme 4).
Firstly, both the Co0 and CoI complexes can promote the
reaction via oxidative coupling of two molecules of 2a as the
initial step, followed by insertion of 1a into the cobaltacyclo-
pentadiene and C-N reductive elimination. The potentially
competing oxidative coupling of 1a and 2a requires a much
higher activation barrier. Secondly, the activation barrier for
the alkyne/alkyne oxidative coupling is distinctly lower for the
Co0 complex than for the CoI complex, suggesting that a Co0
species is more likely to serve as the active species. Thirdly,
the alkyne oxidative coupling and the nitrile insertion
proceed in a stepwise manner involving metastable metalla-
cycle intermediates. Given that nitrile/alkyne oxidative cou-
pling has been proposed to operate in some of the previously
reported nickel- and iron-catalyzed [2+2+2] pyridine synthe-
sis,[8a,9a] the difference among the neighboring first-row
transition metal triad (i.e., Fe/Co/Ni) would deserve further
theoretical investigation.
In light of Scholl-type cyclodehydrogenation of polyar-
ylbenzenes as a well-established strategy to construct poly-
cyclic aromatic hydrocarbons (PAHs),[25] it appears a natural
idea to use tetra- and pentaarylpyridines as precursors to aza-
PAHs[26] via analogous cyclodehydrogenation. However, to
our best knowledge, no such transformation has been
reported. The lack of precedence can be attributed to the
electron-deficient nature and the basicity of a pyridine ring
that would prevent the typical Scholl reaction employing
Lewis or Brønsted acid. In this context, the reductive
cyclodehydrogenation using potassium metal,[27] which has
recently been utilized by Ito, Itami and co-workers for the
conversion of azine-containing polyaromatic systems to aza-
PAHs,[28] appeared attractive.
Scheme 5. Mechanochemically assisted cyclodehydrogenation of tetra-
and pentaarylpyridines to nitrogen-containing PAHs.
yields, accompanied by tarry insoluble products with full
conversion of the starting materials (Scheme 5a). Important-
ly, the reaction of 3oa in toluene afforded 8a in only 14%
yield (NMR), which could not be isolated due to byproducts
arising from incomplete (single and twofold) cyclodehydro-
genations. Curiously, the reaction of analogous tetraarylpyr-
idine 3oc afforded cyclodehydrogenation/demethylation
product 9 as the only isolable compound (Scheme 5b).
Finally, the fourfold cyclodehydrogenation of pentaarylpyr-
idine 3ac afforded the aza-PAH 10 in 27% yield (Scheme 5c),
whereas no detectable amount of 10 could be obtained by the
reaction in toluene. The failure of the solution-phase syn-
theses of the aza-PAHs might be ascribed to the inhomoge-
neity of the reaction mixture and the decreased solubility of
partially cyclodehydrogenated intermediates.
The planarized structure of the novel aza-PAH 10 was
unambiguously confirmed by X-ray crystallographic analysis
(Figure 1a).[33] The C C bonds connecting the pyridine ring
À
and the peripheral benzene rings (bonds a–e) have average
length of 1.444 Š, which is shorter than that of the corre-
sponding bonds in penta(p-tolyl)pyridine (1.494 Š),[15c] re-
flecting the extended p-conjugation. Meanwhile, as expected
from the Clar structure[34] of 10, these and the newly formed
Adopting the reductive cyclodehydrogenation strategy
powered by mechanochemistry,[29–32] we have successfully
converted some of the tetra- and pentaarylpyridines de-
scribed above into the corresponding aza-PAHs (Scheme 5).
The threefold cyclodehydrogenation of 2,3,4,5-tetraaryl-6-
alkylpyridines 3oa, 3pa, and 3ob in the presence of potassium
(20 equiv) under ball milling (30 Hz, 2 h) afforded the aza-
PAHs 8a–8c as poorly soluble light-yellow solids in 19–37%
À
C C bonds (bonds f–i; average length 1.469 Š) are distinctly
À
longer than that of the C C bonds in the rings A–E (average
1.400 Š). The molecular packing of 10 displayed a face-to-
face antiparallel p-p stacked pair with a 3.31 Š-distance
between the C4 positions of the pyridine rings, which is
À
sandwiched by two other molecules of 10 through C H/p
interactions (Figure 1b and Figure S4).
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9631

Angewandte
Research Articles
Chemie
Conclusion
In summary, we have developed highly robust cobalt-
based catalytic systems for the fully intermolecular [2+2+2]
cycloaddition of nitrile and internal alkynes. Thus, simple and
inexpensive catalysts generated from CoI2, dppp, and Zn have
enabled the preparation of monocyclic tetra- and pentaryl-
pyridines with an unprecedented level of structural diversity.
DFT calculations supported a reaction pathway involving
oxidative coupling of two alkyne molecules, insertion of
nitrile into cobaltacyclopentadiene, and C-N reductive elim-
ination, while demonstrating stepwise nature of the oxidative
coupling and nitrile insertion steps. Furthermore, we have
also demonstrated the power of mechanochemistry in forging
hitherto unknown types of aza-PAHs through multifold
reductive cyclodehydrogenation of polyarylated pyridines.
Taken together, the present studies would stimulate further
development and mechanistic study of intermolecular
[2+2+2] pyridine formation as well as the use of polyarylated
pyridines as precursors to nitrogen-doped nanographenes.[24]
Acknowledgements
Figure 1. a) ORTEP representation of 10 (thermal ellipsoids at the
50% probability). Solvent molecule (THF) is omitted for clarity.
b) Molecular packing in the single-crystal X-ray structure of 10.
Distances are in Š.
This work was supported by the Singapore Ministry of
Education Academic Research Fund Tier 2 (MOE2016-T2-
2-043 to N.Y.) and Tier 1 (RG114/18 to N.Y.), Zhejiang
Provincial Natural Science Foundation of China (Grant
LY18B020007 to C.W.), and A*STAR (AME IRG
A1783c0003 and A2083c0050 to F.G.). We thank Dr. Yongxin
Li for his assistance with X-ray crystallographic analysis.
Figure 2 shows the UV-vis absorption and emission
spectra of the aza-PAHs 8a and 10. Similar to their hydro-
carbon analogues,[35] 8a and 10 displayed the absorption
spectra featuring high degree of fine structure in the ranges of
290–402 nm and 304–430 nm, respectively, where the absorp-
tion peaks may be designated as a, b, or para-bands. With TD-
DFT calculations of the optimized structures, the lowest
transition energies of 8a and 10 were calculated to be 379 nm
and 405 nm, respectively (see Tables S7 and S8). The emission
spectra of 8a and 10 were characterized by multiple peaks
(407, 432, and 457 nm for 8a; 435, 462, and 494 nm for 10)
with rather small Stokes shift (< 10 nm), reflecting their rigid
molecular structures. The emission quantum yields were
determined to be 0.27 and 0.47 for 8 and 10, respectively.
Conflict of interest
The authors declare no conflict of interest.
Keywords: cycloaddition ·homogeneous catalysis ·
mechanochemistry ·polycyclic aromatic compounds ·pyridines
[1] a) J. A. Varela, C. Saa, Chem. Rev. 2003, 103, 3787 –3801; b) P. R.
Chopade, J. Louie, Adv. Synth. Catal. 2006, 348, 2307 –2327; c) B.
Heller, M. Hapke, Chem. Soc. Rev. 2007, 36, 1085 –1094; d) G.
Domínguez, J. PØrez-Castells, Chem. Soc. Rev. 2011, 40, 3430 –
3444.
[2] a) N. Nicolaus, H.-G. Schmalz, Synlett 2010, 2071 –2074; b) C.
Yuan, C. T. Chang, A. Axelrod, D. Siegel, J. Am. Chem. Soc.
2010, 132, 5924 –5925.
[3] a) B. L. Gray, X. Wang, W. C. Brown, L. Kuai, S. L. Schreiber,
´
Org. Lett. 2008, 10, 2621 –2624; b) G. Schäfer, J. Milic, A.
Eldahshan, F. Gçtz, K. Zühlke, C. Schillinger, A. Kreuchwig,
J. M. Elkins, K. R. Abdul Azeez, A. Oder, M. C. Moutty, N.
Masada, M. Beerbaum, B. Schlegel, S. Niquet, P. Schmieder, G.
Krause, J. P. von Kries, D. M. Cooper, S. Knapp, J. Rademann,
W. Rosenthal, E. Klussmann, Angew. Chem. Int. Ed. 2013, 52,
12187 –12191; Angew. Chem. 2013, 125, 12409 –12413.
[4] J. Klívar, A. Jancarik, D. Saman, R. Pohl, P. Fiedler, L.
Bednarova, I. Stary, I. G. Stara, Chem. Eur. J. 2016, 22, 14401 –
14405.
Figure 2. UV-vis absorption (solid lines) and emission (dashed lines)
spectra of 8a and 10 in CH2Cl2 (excitation wavelengths: 301 nm for 8a
and 331 nm for 10).
9632 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634

Angewandte
Research Articles
Chemie
[5] a) Y. Wakatsuki, H. Yamazaki, J. Chem. Soc. Chem. Commun.
[13] a) T. Takahashi, F.-Y. Tsai, M. Kotora, J. Am. Chem. Soc. 2000,
1973, 280; b) Y. Wakatsuki, H. Yamazaki, Tetrahedron Lett.
1973, 14, 3383 –3384.
122, 4994 –4995; b) T. Takahashi, F. Y. Tsai, Y. Li, H. Wang, Y.
Kondo, M. Yamanaka, K. Nakajima, M. Kotora, J. Am. Chem.
Soc. 2002, 124, 5059 –5067.
[6] For early and recent reviews, see: a) H. Bçnnemann, Angew.
Chem. Int. Ed. Engl. 1978, 17, 505 –515; Angew. Chem. 1978, 90,
517 –526; b) K. P. C. Vollhardt, Angew. Chem. Int. Ed. Engl.
1984, 23, 539 –556; Angew. Chem. 1984, 96, 525 –541; c) H.
Bçnnemann, Angew. Chem. Int. Ed. Engl. 1985, 24, 248 –262;
Angew. Chem. 1985, 97, 264 –279; d) V. Gandon, Transition-
Metal-Mediated Aromatic Ring Construction (Ed.: K. Tanaka),
Wiley, Hoboken, 2013, chap. 1; e) T. Gläsel, M. Hapke, Cobalt
Catalysis in Organic Synthesis (Eds.: M. Hapke, G. Hilt), Wiley-
VCH, Weinheim, 2020, chap. 9.
[7] a) A. Gutnov, B. Heller, C. Fischer, H. J. Drexler, A. Spannen-
berg, B. Sundermann, C. Sundermann, Angew. Chem. Int. Ed.
2004, 43, 3795 –3797; Angew. Chem. 2004, 116, 3883 –3886;
b) H.-T. Chang, M. Jeganmohan, C.-H. Cheng, Org. Lett. 2007, 9,
505 –508; c) K. Kase, A. Goswami, K. Ohtaki, E. Tanabe, N.
Saino, S. Okamoto, Org. Lett. 2007, 9, 931 –934; d) A. Goswami,
K. Ohtaki, K. Kase, T. Ito, S. Okamoto, Adv. Synth. Catal. 2008,
350, 143 –152; e) A. Geny, N. Agenet, L. Iannazzo, M. Malacria,
C. Aubert, V. Gandon, Angew. Chem. Int. Ed. 2009, 48, 1810 –
1813; Angew. Chem. 2009, 121, 1842 –1845; f) M. Hapke, N.
Weding, A. Spannenberg, Organometallics 2010, 29, 4298 –4304;
g) P. Garcia, Y. Evanno, P. George, M. Sevrin, G. Ricci, M.
Malacria, C. Aubert, V. Gandon, Chem. Eur. J. 2012, 18, 4337 –
4344; h) I. Thiel, H. Jiao, A. Spannenberg, M. Hapke, Chem.
Eur. J. 2013, 19, 2548 –2554.
[8] a) M. M. McCormick, H. A. Duong, G. Zuo, J. Louie, J. Am.
Chem. Soc. 2005, 127, 5030 –5031; b) P. Kumar, S. Prescher, J.
Louie, Angew. Chem. Int. Ed. 2011, 50, 10694 –10698; Angew.
Chem. 2011, 123, 10882 –10886; c) R. M. Stolley, M. T. Maczka,
J. Louie, Eur. J. Org. Chem. 2011, 3815 –3824; d) R. M. Stolley,
H. A. Duong, D. R. Thomas, J. Louie, J. Am. Chem. Soc. 2012,
134, 15154 –15162; e) X. You, X. Xie, G. Wang, M. Xiong, R.
Sun, H. Chen, Y. Liu, Chem. Eur. J. 2016, 22, 16765 –16769.
[9] a) C. Wang, X. Li, F. Wu, B. Wan, Angew. Chem. Int. Ed. 2011,
50, 7162 –7166; Angew. Chem. 2011, 123, 7300 –7304; b) B. R.
DꢀSouza, T. K. Lane, J. Louie, Org. Lett. 2011, 13, 2936 –2939;
c) T. K. Lane, B. R. DꢀSouza, J. Louie, J. Org. Chem. 2012, 77,
7555 –7563; d) C. Wang, D. Wang, F. Xu, B. Pan, B. Wan, J. Org.
Chem. 2013, 78, 3065 –3072; e) N. A. Spahn, M. H. Nguyen, J.
Renner, T. K. Lane, J. Louie, J. Org. Chem. 2017, 82, 234 –242.
[10] a) Y. Yamamoto, R. Ogawa, K. Itoh, J. Am. Chem. Soc. 2001,
123, 6189 –6190; b) Y. Yamamoto, S. Okuda, K. Itoh, Chem.
Commun. 2001, 1102 –1103; c) J. A. Varela, L. Castedo, C. Saa,
J. Org. Chem. 2003, 68, 8595 –8598; d) Y. Yamamoto, K.
Kinpara, T. Saigoku, H. Takagishi, S. Okuda, H. Nishiyama, K.
Itoh, J. Am. Chem. Soc. 2005, 127, 605 –613; e) Y. Yamamoto, K.
Kinpara, H. Nishiyama, K. Itoh, Adv. Synth. Catal. 2005, 347,
1913 –1916; f) Y. Yamamoto, K. Kinpara, R. Ogawa, H.
Nishiyama, K. Itoh, Chem. Eur. J. 2006, 12, 5618 –5631; g) S.
Medina, G. Dominguez, J. Perez-Castells, Org. Lett. 2012, 14,
4982 –4985; h) J.-F. Tan, C. T. Bormann, F. G. Perrin, F. M.
Chadwick, K. Severin, N. Cramer, J. Am. Chem. Soc. 2019, 141,
10372 –10383.
[14] a) D. Suzuki, R. Tanaka, H. Urabe, F. Sato, J. Am. Chem. Soc.
2002, 124, 3518 –3519; b) D. Suzuki, Y. Nobe, Y. Watai, R.
Tanaka, Y. Takayama, F. Sato, H. Urabe, J. Am. Chem. Soc. 2005,
127, 7474 –7479; c) R. Tanaka, A. Yuza, Y. Watai, D. Suzuki, Y.
Takayama, F. Sato, H. Urabe, J. Am. Chem. Soc. 2005, 127, 7774 –
7780.
[15] a) S. L. Gholap, P. Hommes, K. Neuthe, H. U. Reissig, Org. Lett.
2013, 15, 318 –321; b) C. Doebelin, P. Wagner, F. Bihel, N.
Humbert, C. A. Kenfack, Y. Mely, J. J. Bourguignon, M. Schmitt,
J. Org. Chem. 2014, 79, 908 –918; c) S. Reimann, P. Ehlers, A.
Petrosyan, S. Kohse, A. Spannenberg, A. E. Surkus, T. V.
Ghochikyan, A. S. Saghyan, S. Lochbrunner, O. Kühn, R.
Ludwig, P. Langer, Adv. Synth. Catal. 2014, 356, 1987 –2008;
d) S. Suzuki, Y. Segawa, K. Itami, J. Yamaguchi, Nat. Chem.
2015, 7, 227 –233; e) T. Asako, W. Hayashi, K. Amaike, S.
Suzuki, K. Itami, J. Yamaguchi, Tetrahedron 2017, 73, 3669 –
3676.
[16] H. Nehl, Chem. Ber. 1994, 127, 2535 –2537.
[17] J. Uhm, H. W. An, J. Korean Chem. Soc. 2001, 45, 268 –273.
[18] Y. Xie, C. Wu, C. Jia, C.-H. Tung, W. Wang, Org. Chem. Front.
2020, 7, 2196 –2201.
[19] a) W. Hess, J. Treutwein, G. Hilt, Synthesis 2008, 3537 –3562;
b) K. Gao, N. Yoshikai, Acc. Chem. Res. 2014, 47, 1208 –1219;
c) P. Gandeepan, C. H. Cheng, Acc. Chem. Res. 2015, 48, 1194 –
1206; d) M. Moselage, J. Li, L. Ackermann, ACS Catal. 2016, 6,
498 –525; e) P. Rçse, G. Hilt, Synthesis 2016, 48, 463 –492; f) N.
Yoshikai, Synthesis 2019, 51, 135 –141.
[20] a) J. Yang, N. Yoshikai, J. Am. Chem. Soc. 2014, 136, 16748 –
16751; b) J. Yang, A. Rerat, Y. J. Lim, C. Gosmini, N. Yoshikai,
Angew. Chem. Int. Ed. 2017, 56, 2449 –2453; Angew. Chem.
2017, 129, 2489 –2493; c) C. Wu, N. Yoshikai, Angew. Chem. Int.
Ed. 2018, 57, 6558 –6562; Angew. Chem. 2018, 130, 6668 –6672;
d) J. Yang, Y. Shen, Y. J. Lim, N. Yoshikai, Chem. Sci. 2018, 9,
6928 –6934; e) W. Ding, N. Yoshikai, Angew. Chem. Int. Ed.
2019, 58, 2500 –2504; Angew. Chem. 2019, 131, 2522 –2526;
f) C. S. Wang, S. Di Monaco, A. N. Thai, M. S. Rahman, B. P.
Pang, C. Wang, N. Yoshikai, J. Am. Chem. Soc. 2020, 142, 12878 –
12889.
[21] For the use of CoCl(PPh3)3 as a catalyst for alkyne cyclo-
trimerization, see: a) P. Jungk, F. Fischer, I. Thiel, M. Hapke, J.
Org. Chem. 2015, 80, 9781 –9793; b) L. D. Field, A. J. Ward, P.
Turner, Aust. J. Chem. 1999, 52, 1085 –1092; c) K. P. Anger-
mund, P. Betz, H. Butenschçn, Chem. Ber. 1993, 126, 713 –724.
[22] a) G. Dazinger, M. Torres-Rodrigues, K. Kirchner, M. J. Cal-
horda, P. J. Costa, J. Organomet. Chem. 2006, 691, 4434 –4445;
b) A. A. Dahy, K. Yamada, N. Koga, Organometallics 2009, 28,
3636 –3649; c) A. A. Dahy, N. Koga, J. Organomet. Chem. 2010,
695, 2240 –2250; d) A. M. Rodriguez, C. Cebrian, P. Prieto, J. I.
Garcia, A. de la Hoz, A. Diaz-Ortiz, Chem. Eur. J. 2012, 18,
6217 –6224.
[23] All the DFT calculations were performed using Gaussian 09
program package. See the Supporting Information for the details
of the computational methods and results.
[11] a) Y. Komine, A. Kamisawa, K. Tanaka, Org. Lett. 2009, 11,
2361 –2364; b) T. Shibata, T. Uchiyama, K. Endo, Org. Lett.
2009, 11, 3906 –3908; c) Y. Komine, K. Tanaka, Org. Lett. 2010,
12, 1312 –1315; d) K. Murayama, Y. Sawada, K. Noguchi, K.
Tanaka, J. Org. Chem. 2013, 78, 6202 –6210; e) K. Kashima, M.
Ishii, K. Tanaka, Eur. J. Org. Chem. 2015, 1092 –1099.
[12] a) G. Onodera, Y. Shimizu, J. N. Kimura, J. Kobayashi, Y.
Ebihara, K. Kondo, K. Sakata, R. Takeuchi, J. Am. Chem. Soc.
2012, 134, 10515 –10531; b) R. Takeuchi, S. Fujisawa, Y. Yoshida,
J. Sagano, T. Hashimoto, A. Matsunami, J. Org. Chem. 2018, 83,
1852 –1860.
[24] For the generation of diphosphine-ligated CoI and Co0 species
through reduction by Zn, see: M. R. Friedfeld, H. Zhong, R. T.
Ruck, M. Shevlin, P. J. Chirik, Science 2018, 360, 888 –893.
[25] a) M. Grzybowski, K. Skonieczny, H. Butenschon, D. T. Gryko,
Angew. Chem. Int. Ed. 2013, 52, 9900 –9930; Angew. Chem.
2013, 125, 10084 –10115; b) M. Grzybowski, B. Sadowski, H.
Butenschon, D. T. Gryko, Angew. Chem. Int. Ed. 2020, 59, 2998 –
3027; Angew. Chem. 2020, 132, 3020 –3050.
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9633

Angewandte
Research Articles
Chemie
[26] a) X.-Y. Wang, X. Yao, A. Narita, K. Müllen, Acc. Chem. Res.
27891; c) Y. Zhao, S. V. Rocha, T. M. Swager, J. Am. Chem. Soc.
2016, 138, 13834 –13837.
[31] We also made attempts to perform the present [2+2+2] cyclo-
addition under mechanochemical conditions, which however did
not meet with success.
˙
´
´
2019, 52, 2491 –2505; b) M. StØpien, E. Gonka, M. Zyla, N.
Sprutta, Chem. Rev. 2017, 117, 3479 –3716.
[27] a) P. Schlichting, U. Rohr, K. Müllen, J. Mater. Chem. 1998, 8,
2651 –2655; b) D. T. Gryko, J. Piechowska, M. Galezowski, J.
Org. Chem. 2010, 75, 1297 –1300; c) M. Rickhaus, A. P. Belang-
er, H. A. Wegner, L. T. Scott, J. Org. Chem. 2010, 75, 7358 –7364.
[28] K. P. Kawahara, W. Matsuoka, H. Ito, K. Itami, Angew. Chem.
Int. Ed. 2020, 59, 6383 –6388; Angew. Chem. 2020, 132, 6445 –
6450.
[32] For other failed attempts on cyclodehydrogenation of 3oa, see
the Supporting Information.
[33] Deposition Number 2050264 (10) contains the supplementary
crystallographic data for this paper. These data are provided free
of charge by the joint Cambridge Crystallographic Data Centre
and Fachinformationszentrum Karlsruhe Access Structures
service www.ccdc.cam.ac.uk/structures.
´
[29] For selected reviews, see: a) T. Friscic, C. Mottillo, H. M. Titi,
Angew. Chem. Int. Ed. 2020, 59, 1018 –1029; Angew. Chem.
2020, 132, 1030 –1041; b) D. Tan, F. García, Chem. Soc. Rev.
2019, 48, 2274 –2292; c) J. L. Howard, Q. Cao, D. L. Browne,
Chem. Sci. 2018, 9, 3080 –3094; d) J. Andersen, J. Mack, Green
[34] M. Solà, Front. Chem. 2013, 1, 22.
[35] a) R. Rieger, K. Müllen, J. Phys. Org. Chem. 2010, 23, 315 –325;
b) A.-K. Steiner, K. Y. Amsharov, Angew. Chem. Int. Ed. 2017,
56, 14732 –14736; Angew. Chem. 2017, 129, 14926 –14931.
´
Chem. 2018, 20, 1435 –1443; e) J.-L. Do, T. Friscic, ACS Cent.
Sci. 2017, 3, 13 –19.
[30] For examples of mechanochemical synthesis of PAHs, see: a) G.
Bµti, D. Csókµs, T. Yong, S. M. Tam, R. R. S. Shi, R. D. Webster,
I. Pµpai, F. García, M C. Stuparu, Angew. Chem. Int. Ed. 2020,
59, 21620 –21626; Angew. Chem. 2020, 132, 21804 –21810; b) C.
Manuscript received: January 6, 2021
Accepted manuscript online: February 8, 2021
Wang, M. Hill, B. Theard, J. Mack, RSC Adv. 2019, 9, 27888 – Version of record online: March 12, 2021
9634 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9627 –9634