Iron-catalysed regioselective thienyl C–H/C–H coupling

Article abstract of DOI:10.1038/s41929-021-00653-7

Regioselective thienyl–thienyl coupling is arguably one of the most important transformations for organic electronic materials. A prototype of ideal organic synthesis to couple two thienyl groups by cutting two C–H bonds requires formal removal of two hydrogen atoms with an oxidant, which often limits the synthetic efficiency and versatility for oxidation-sensitive substrates (for example, donor and hole-transporting materials). Here, we found that diethyl oxalate, used together with AlMe₃, acts as a two-electron acceptor in an iron-catalysed C–H activation. We describe the regioselective thienyl C–H/C–H coupling with an iron(III) catalyst, a trisphosphine ligand, AlMe₃ and diethyl oxalate under mild conditions. The efficient catalytic system accelerated by ligand optimization polymerizes thiophene-containing monomers into homo- and copolymers bearing a variety of electron-donative π motifs. The findings suggest the versatility of iron catalysis for the synthesis of functional polymers, for which the potential of this ubiquitous metal has so far not been fully appreciated. [Figure not available: see fulltext.].

Full text of DOI:10.1038/s41929-021-00653-7

Articles
https://doi.org/10.1038/s41929-021-00653-7
Iron-catalysed regioselective thienyl C–H/C–H
coupling
ꢀ✉
ꢀ✉
Takahiro Dobaꢀ ꢀ, Laurean Iliesꢀ ꢀ, Wataru Sato, Rui Shangꢀ ꢀ and Eiichi Nakamuraꢀ ꢀ
Regioselective thienyl–thienyl coupling is arguably one of the most important transformations for organic electronic materials.
A prototype of ideal organic synthesis to couple two thienyl groups by cutting two C–H bonds requires formal removal of two
hydrogen atoms with an oxidant, which often limits the synthetic efficiency and versatility for oxidation-sensitive substrates
(for example, donor and hole-transporting materials). Here, we found that diethyl oxalate, used together with AlMe3, acts as
a two-electron acceptor in an iron-catalysed C–H activation. We describe the regioselective thienyl C–H/C–H coupling with
an iron(III) catalyst, a trisphosphine ligand, AlMe3 and diethyl oxalate under mild conditions. The efficient catalytic system
accelerated by ligand optimization polymerizes thiophene-containing monomers into homo- and copolymers bearing a variety
of electron-donative π motifs. The findings suggest the versatility of iron catalysis for the synthesis of functional polymers, for
which the potential of this ubiquitous metal has so far not been fully appreciated.
irect C–H/C–H coupling to connect two organic mol- of the polymers are truncated with a hydrogen atom, hence obviat-
ecules is a straightforward C–C bond-forming strategy ing the end-capping procedure that is typically required to remove
D
that has long been known for biaryl synthesis and is still the potentially detrimental halide or organometallic termini in
utilized for the synthesis of condensed aromatics and thiophene the conventional optoelectronic polymer synthesis (for example,
polymers for use in materials applications1–6. However, strongly polymerization by C–M/C–X coupling or C–H/C–X coupling)21–27
.
oxidizing conditions are often required for the formal removal of We anticipate that the present four-component iron catalytic system
two hydrogen atoms, limiting substrate versatility and reaction developed for C–H/C–H coupling sets a cornerstone for the future
selectivity. For example, the classical reactions through cationic or development of efficient iron catalysis and polymerization.
radical cationic mechanisms (such as the Scholl reaction and oxi-
dative aromatic coupling) are limited in scope, producing isomers Results
and undesired side products (Fig. 1a)7. A modern variation using Development of thienyl C–H/C–H homocoupling. Figure 2 sum-
palladium-catalysed C–H activation has broadened the scope, cre- marizes key reaction parameters and our investigation to identify
ating a simple route to thiophene polymers, but the requirement of readily available and inexpensive diethyl oxalate as the most effec-
an oxidant (for example, Cu(II) or Ag(I)) to turn over the Pd(II)/ tive two-electron acceptor in the presence of AlMe3. A representa-
Pd(0) cycle with a large redox potential (E0 (PdII/Pd0)=+0.92V ver- tive example for iron(III)-catalysed regioselective thienyl C–H/C–H
sus the normal hydrogen electrode) still limits its synthetic versatil- coupling is shown in Fig. 2a using benzo[b]thiophene (1) as the
ity, especially in the synthesis of electron-rich and highly conjugated substrate. The C2 dimerization product (2) was obtained quantita-
molecules (Fig. 1a)8,9. Towards better synthetic design, we chose to tively using Fe(acac)3 (10mol%), a conjugated tridentate phosphine
avoid the strongly oxidative condition in its entirety and to achieve ligand (TP2; 11mol%) previously used in an iron-catalysed C–H
high selectivity using an iron-catalysed C–H activation protocol methylation28, AlMe3 (100mol%) and DEO (50mol%) in a tetrahy-
(Fig. 1b, I and II)10–15. We recognized that diethyl oxalate (DEO) drofuran and toluene mixed solvent at 70°C. The use of 50mol%
functions as a latent oxidant in combination with Al(III) species for AlMe3 resulted in incomplete conversion of 1. Neither the product
the regeneration of an iron(III) catalyst by accepting two electrons, of C3 activation nor the one due to C–H methylation formed. The
probably via inner-sphere electron transfer (Fig. 1b, III), to form exclusive C2 activation agrees with a deprotonative C–H activation
an aluminium enediolate (Fig. 1b, IV). The strong oxophilicity of mechanism that prefers the more acidic C2 proton than the C3 pro-
Al(III) to form the thermodynamically stable five-membered ring ton12, and probably reflects the steric effects of the tridentate phos-
aluminium enediolate (Fig. 1b, IV) contributes as the driving force phine ligand (see below). Using C2-deuterated 1, we ascertained
for Fe(III) regeneration, and oxophilic Al(III) sequesters reduced that the methyl group of AlMe3 cleaves the C2–H bond (CH3D was
oxalate from Fe(III) to ensure effective catalyst turnover. A readily detected using 1H NMR; see Supplementary Fig. 12 for details).
available tridentate phosphine ligand, tridentate phosphine 1 (TP1)
(Fig. 1c), controls the coordination environment of the iron centre C2–H cleavage because AlMe3 alone did not generate C2 metalation
to sense the steric nature of the incoming substrates. product in the absence of an iron catalyst (see Supplementary Fig. 7
Control experiments showed that an iron catalyst is essential for
Here,wedocumentacatalyticC–HactivationforC2-regioselective for details). Kinetic isotope experiments of parallel reactions using
thienyl C–H/C–H coupling using an iron(III) catalyst, TP, DEO 1 and 1-2-d (benzo[b]thiophene-2-d) showed no primary kinetic
and AlMe3 under mild conditions (Fig. 1b). The reaction acceler- isotope effect (kH/kD =1), indicating that C–H cleavage is not the
ated by the newly designed ligand has achieved a straightforward turnover-limiting step in the catalytic cycle (see Supplementary Fig.
C–H/C–H coupling approach to polymeric semiconductive materi- 11 for details). The reaction using benzil in place of DEO produced
als useful for electronic device applications (Fig. 1d)16–20. The termini 2 and benzoin (Fig. 2b and entry 7 in Fig. 2c), providing evidence
✉
Present address: Department of Chemistry, The University of Tokyo, Tokyo, Japan. e-mail: rui@chem.s.u-tokyo.ac.jp; nakamura@chem.s.u-tokyo.ac.jp
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
631

NATuRe CATAlysIs
Articles
a
entry 2)29. Fe(acac)3 is essential (entry 3). Fe(acac)2 as a precata-
lyst resulted in reduced reactivity (entry 4), as was reported pre-
S
S
S
S
H
S
viously in some iron-catalysed C–H activation reactions28,29
.
[O]
Without DEO (entry 5), the reaction produced 2 in 7% yield,
indicating less than one turnover of the catalyst. Other latent
oxidants—N,N,N′,N′-tetramethyloxalamide, benzil (Fig. 2b) and
1,4-benzoquinone30—were much less effective (entries 6–8). DEO
was found to be the highest yielding among oxalate esters, as
dimethyl, diisopropyl and di-tert-butyl oxalates gave 2 in 45–86%
yields (see Supplementary Fig. 1 for details). We partly attributed
the poor performance of 1,2-dichloropropane (entry 9) to the
inhibitory effects of the propene side product coordinating to the
iron/TP catalyst. In addition, we attributed the poor conversion of
simple thiophene and some mono-substituted thiophenes (such as
3-methoxythiophene and 3-hexylthiophene) to a similar process.
1-Octadecene heavily suppressed the reaction (entry 10), while
trans-stilbene did not (entry 11). This contrast suggests that the
iron/TP coordination site is sensitive to steric effects of the incom-
ing π substrates, which agrees with internal double bonds and π
units in conjugated polymers remaining intact (Fig. 3–5). AlMe3
was essential and could not be replaced by either AlEt3 (entries
12 and 13), MeLi, MeMgBr or MeZnBr (entry 14). AlEt3 failed as
a base, probably because of the formation of inactive iron hydride
species after β-hydride elimination of the ethyl group28. MeLi,
MeMgBr and MeZnBr failed because of their strong nucleophilic-
ity, which is incompatible with oxalate, and high reducing ability
that reduces catalytically active Fe(III) species29. The tridentate
phosphine TP2 was essential (entry 15). PPh3 (33mol%), dppbz
and TetraP in place of TP2 were entirely ineffective, giving back the
starting material (entries 16–18)31. All other ligands investigated,
such as bipyridines, phenanthrolines and N-heterocyclic carbene
ligands, were ineffective (see Supplementary Table 1 for details).
TP1 and TP2 were equally effective for the dimerization (entry 19),
while TP1 was essential in polycondensation (Fig. 3). As found in
the reported crystal structures of iron/TP2 complexes32, TP blocks
one apical and two equatorial sites, leaving three sites available for
catalysis. Steric effects of TP’s bulk may be large in light of the inhi-
bition by 1-octadecene (Fig. 2c, entry 10) but not by trans-stilbene.
For all entries in Fig. 2c where low yields are shown, 1 was largely
recovered. We note that the reported palladium-catalysed dimer-
ization of 1 to 2 was rather moderate yielding: 41% yield under
the original Mori conditions using PdCl2(PhCN)2 (3mol%) and
AgF (2.0 equiv.) in dimethyl sulfoxide (DMSO) at 60°C; and 50%
yield under the recent Stahl conditions using Pd(OAc)2 (5mol%),
or Pd(II)/[O]
S
H
S
S
and so on
b
cat. Fe(III)/TP
AlMe3 (1.0 equiv.)
DEO (0.50 equiv.)
S
S
+
H
H
THF/PhMe
70°C, 15 h
S
S
(ArH)
[Al]III
P
[Fe]I
P
[Fe]III
O
O
P
P
P
P
[Al]III
EtO
OEt
O
O
III
O
O
+ AlMe3
EtO
OEt
+
IV
EtO
OEt
DEO
H
Me
P
S
+ ArH, AlMe3
– 2Me–H
[Fe]III
P
P
[Fe]III
Ar
P
P
P
Ar
II
I
c
PPh2
PPh2
2
Ph2P
P
Li
O
P(OPh)3
Li
TP1
O
(1.80g, 90% over two steps)
d
cat. Fe(III)/TP1
S
S
H
H
π
π
H
H
AlMe3, DEO
70°C
S
S
n
DP = up to 46
PDI = 1.8–2.9
Fig. 1 | Iron-catalysed regioselective thienyl C–H/C–H coupling enabled
by the TP/alMe3/DEO system. a, Conventional thienyl C–H/C–H coupling
under oxidizing conditions. [O], oxidant. b, Iron-catalysed regioselective
thienyl C–H/C–H coupling via deprotonative C–H activation with AlMe3
and regeneration of the Fe(III) catalyst with DEO/Al(III). Metal atoms
in brackets indicate the omission of coordination details for clarity. c,
Structure of the most effective TP and its facile synthesis. d, Regioselective
C–H/C–H polycondensation of bisthienyl monomers enabled by a
four-component iron catalytic system. DP, degree of polarization.
1,10-phenanthroline-5,6-dione
(5mol%),
Mn(OAc)2·4H2O
(5mol%) and 1,4-benzoquinone (5mol%) under 1.1atm O2 in
DMSO at 120°C33,34. Note also that an attempted dimerization of a
benzo[b]thiophene derivative with 2,3-dichloro-5,6-dicyano-1,4-b
enzoquinone was reported to produce a cyclotrimerization product
instead of 2 (ref. 35).
Development of thienyl C–H/C–H polycondensation.
Polycondensation through metal-catalysed C–H bond activation
for the formation of the enediolate IV, and hence supporting the imposes much higher hurdles than small-molecule synthesis as to
proposed role of DEO as a two-electron acceptor. Attempts to detect the catalyst efficiency and selectivity. Only a handful of examples
IV in situ or ethyl glyoxalate after hydrolysis have failed so far, prob- of electron-deficient polymers have been reported using transition
ably because of the aggregation and high oxophilicity of Al(III) spe- metal-catalysed C–H/C–H polycondensation, and none have been
cies. This aggregation behaviour facilitates product isolation—an reported for electron donor-type semiconducting polymers4. As
unexpected bonus of the use of DEO as a mild and inexpensive a benchmark (Fig. 3a), we chose the conversion of a 2,7-dithienyl
two-electron acceptor that is compatible with the redox-sensitive carbazole (DTC) (3) to a thiophene–carbazole polymer (4) used
functionalities in the product.
for organic thin-film transistors36. The conventional synthesis of
The data in Fig. 2 illustrate some intriguing characteristics of 4 is an archetype of the synthesis of functional polymers, relying
the present C–H/C–H coupling reaction. 9,10-Dihydroanthracene on the Migita–Kosugi–Stille coupling of 5,5′-bis(trimethylstannyl)-
(1 equiv.)—a typical hydrogen donor—was fully recovered with- 2,2′-bithiophene with 2,7-dibromocarbazole monomers (number
out affecting the reaction at all. Thus, the reaction involves nei- average molecular weight (Mn)=10–17kDa; polydispersity index
ther oxidative nor free radical or radical cation species (Fig. 2c, (PDI)=2.6–3.5; degree of polymerization(DP) =16–30). No C–H/
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
632

NATuRe CATAlysIs
Articles
sulfur atom. The electronic properties of the TP ligand affect the
reaction efficiency. Installation of an electron-rich aryl group on
the central phosphine atom decreased the degree of polymeriza-
tion (entries 1–5; Fig. 3b). We also found, by mass spectral analysis,
mono- and dimethylated products in the recovered TP2 ligand, and
suspected catalyst deactivation by ortho-methylation of phenyl on
the central phosphine atom of the ligand37. We replaced the cen-
tral phenyl group with a five-member heteroaryl group to suppress
the ligand degradation (TP7, TP1 and TP8). We ascribe the poor
performance of TP7 to the electron-rich indole group and steric
hindrance of the N-methyl substituent. TP1 and TP8 produced
polymer 4 with degrees of polymerization of 23 and 21 and small
PDIs, which were isolated in 86 and 88% isolated yield, respectively.
Notably, branched polymers were not generated, as judged by analy-
Fe(acac)3 (10 mol%)
TP2 (11 mol%)
AlMe3 (0.20 mmol)
a
S
S
H
DEO (0.10 mmol)
THF/PhMe
(vol/vol = 2/1, 0.30 ml)
70°C, 15 h
S
1
2
0.20 mmol
b
O
O
O
OH
Ph
As above
+
+
1
2
Except no DEO
Ph
Ph
Ph
0.20 mmol
0.10 mmol
0.046 mmol
0.016 mmol
c
Entry
Variation from standard condition
Yield (%)
1
sis of H NMR spectra, because a conspicuous singlet peak would
1
2
None
100
100
0
appear in the aromatic region in the case of formation of a branched
product (see Supplementary Fig. 4 for details).
With 9,10-DHA added (1.0 equiv.)
Without Fe(acac)3
Kinetic experiments indicated a step-growth polycondensation
mechanism38,39. Figure 3c shows the correlation between the Mn and
PDI values as a function of monomer conversion in the synthesis of
polymer 5 (F8T2 polymer; Fig. 5). Here, we note that the Mn value
increases sharply at a high conversion of monomer, while the PDI
value slowly approaches 2—a typical value for a step-growth poly-
condensation reaction.
3
Fe(acac)2 instead of Fe(acac)3
4
47
7
5
Without DEO
Tetramethyloxalamide instead of DEO
Benzil instead of DEO
26
46
16
61
9
6
7
8
1,4-Benzoquinone instead of DEO
1,2-Dichloropropane instead of DEO
Substrate scope. Figure 4 illustrates the scope of thienyl C–H/
C–H coupling. The reaction showed exclusive C2 (or C5) selec-
tivity as to the thiophene ring, and dimers of benzo[b]thiophene
(2), thiophenes (6 and 7) and benzo[1,2-b:4,5-b′]dithiophene (11)
were obtained in excellent yield. C2 methylation instead of dimer-
ization barely took place (detected only for 18–20 in <1% yield).
The reaction cleanly produced the desired product, which was
easily isolated and purified either by silica gel chromatography or
filtration and reprecipitation from the reaction mixture. A hexyl
substituent on the C3 position inhibited C–H coupling on the adja-
cent C2 position, and the coupling produced dialkylated tetrathio-
phene (9) as an exclusive product in excellent yield. Electron-rich
3,4-ethylenedioxythiophene (18) reacted in moderate yield due to
steric hindrance by the 3,4-substituents. π motifs commonly found
in optoelectronic materials, such as benzo[b]furan (10), vinylene
(12), fluorene (16), carbazoles (19 and 20) and electron-rich tri-
arylamines (21 and 22), survived the reaction conditions well. An
oligo-thienylenevinylene product, which was susceptible to elec-
tron transfer side reactions, was obtained in high yield, and stereo-
isomerization of the E-vinylene structure was not observed (12).
The reaction tolerated triisopropylsilyl (11–13), tributyltin (14) and
pinacoboronate (15) on arenes, which serve as useful handles for
synthetic elaborations for materials synthesis. Oligothiophenes with
silicon and tin protection at the terminal position were obtained
in high yields (13 and 14). The nonoxidative condition kept intact
the multiaromatic array in 17, which was potentially susceptible
to the conventional oxidative aromatic coupling conditions40. The
hole-transporting dithiophene 22 has been used for perovskite solar
cells41. The reaction was easily scalable to a gram quantity (19).
Figure 5 illustrates examples of straightforward conversion of
monomers to the corresponding polymers that are inaccessible
either by the conventional oxidative aromatic C–H/C–H coupling
9
With 1-octadecene added (1.0 equiv.)
With trans-stilbene added (1.0 equiv.)
Without AlMe3
10
11
12
13
14
15
16
17
18
19
94
0
0
AlEt3 instead of AlMe3
MeLi, MeMgBr or MeZnBr instead of AlMe3
Without TP2
0
0
0
PPh3 (33 mol%) instead of TP2
Dppbz instead of TP2
0
0
TetraP instead of TP2
100
TP1 instead of TP2
PPh2
PPh2
H2
C
PPh2
PPh2
Ph2P
P
Ph2P
P
C
H2
PPh2
Dppbz
9,10-DHA
TP2
TetraP
Fig. 2 | Investigation of reaction parameters. a, Iron-catalysed
regioselective thienyl C–H/C–H homocoupling of benzo[b]thiophene
under standard reaction conditions. b, Detection of benzoin, generated by
two-electron reduction of benzil. c, Table of the key reaction parameters.
9,10-DHA, 9,10-dihydroanthracene. yields were determined with 1H NMR
using 1,3,5-trimethoxybenzene as an internal standard. yields of 0% were
determined using gas chromatography.
C–H coupling route to 4 has been reported. We noticed first that or by the palladium-catalysed C–H/C–H coupling. Homo- and
the acetylacetonate ligand reduces the catalyst efficiency, and found copolymers were obtained with a degree of polymerization of up
air-stable iron(III) chloride hexahydrate instead to be an iron source to 46 with a unimodal distribution and a PDI of 1.75–2.87 (2.29
of choice (see Supplementary Table 2 for details).
averaged over 11 polymers). The excellent C2-regioselective thie-
Figure 3 summarizes the study on DTC polymer synthesis, nyl C–H activation permitted the production of linear polymers
which culminated in the design of TP1 as a key enabler of the pres- uncontaminated by polymers branched at the C3 position. Poly(
ent study (see Fig. 1c) after structural and electronic tuning (Fig. 9,9-dioctylfluorene-alt-bithiophene) (5; abbreviated as F8T2)—a
3b). Analysis of H and 13C NMR spectra showed that the carbon commercially available semiconductive polymer (Mn = 16 kDa;
1
chain extension took place exclusively at the carbon next to the PDI = 2.76) that may contain 200 ppm Pd—is widely used for
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
633

NATuRe CATAlysIs
Articles
FeCl3•6H2O (5.0 mol%)
a
b
C8H17
C8H17
C8H17
C8H17
TP (10 mol%)
AlMe3 (3.0 equiv.)
N
N
DEO (2.0 equiv.)
THF/PhMe
(vol/vol = 1/1, 0.60 ml)
70°C, 24 h
S
S
S
S
n
3
4
0.20 mmol
DTC polymer
Mn
(kg mol−1
Mw
(kg mol−1
)
a
DPb
Entry
TP
Mw/Mn
Yield (%)c
a
a
)
–
–
–
–
–
–
1
2
3
4
5
6
7
8
TP2
TP3
TP4
TP5
TP6
TP7
TP1
TP8
5.1
1.6
10.8
2.0
2.12
1.23
1.99
2.09
2.07
1.30
1.70
1.76
9
3
3.0
5.9
5
5.0
10.4
13.2
2.2
9
6.4
11
3
1.7
13.3
11.9
22.6
21.0
23
21
86
88
–R
PPh2
PPh2
X
X
NMe
O
–NMe2
–OMe
–F
TP3
TP7
Ph2P
P
R
Ph2P
P
TP4
TP5
TP6
TP1
TP8
S
–CF3
c
14
12
10
8
C8H17
C8H17
S
S
n
5
6
4
2
2
1
0
0
0
40
50
60
70
80
90
100
Conversion (%)
Fig. 3 | Investigation of the ligand effect on polycondensation. a, Iron-catalysed regioselective thienyl C–H/C–H polycondensation to produce the DTC
polymer. b, Ligand effect on polycondensation. aMn and the weight average molecular weight (Mw) were determined with GPC using polystyrene standards
for the crude polymers. bThe degree of polymerization was calculated using Mn. cIsolated yields after precipitation. yields were not determined for
short oligomers. c, Mn and Mw/Mn values of 5 as a function of monomer conversion. Reaction conditions: monomer (1.0 equiv.; 0.20ꢀmmol), FeCl3·6H2O
(5.0ꢀmol%), TP1 (10ꢀmol%), AlMe3 (3.0 equiv.), DEO (2.0 equiv.), THF (0.30ꢀml) and PhMe (0.30ꢀml) at 70ꢀ°C.
organic light-emitting diodes, photovoltaics and field-effect and the degree of polymerization to 42–46 (23 and 24). A new
transistors42. The iron catalysis afforded 5 in 91% isolated yield polymer containing 3-hexylthiophene and triarylamine motifs
with a large molecular weight (Mn = 20.3 kDa; degree of polym- (25) was made in 93% yield with a degree of polymerization of 26
erization = 37) and a PDI of 2.61 (a value indicating narrower and a PDI of 2.01. The ionization potential of 25 (–5.4 eV; film)
molecular weight distribution compared with commercially was close to that of methylammonium lead iodide perovskite,
available F8T2). The iron content in F8T2 could be reduced to suggesting its application in solar cells43. The excellent C2 regi-
43 ppm by treatment with silica metal scavenger (Fuji Silysia SH). oselectivity of C–H/C–H coupling allowed us to synthesize a
Increasing the solubility by installation of alkyl substituents on polythiophene (26) from a 3,3′′-dihexyl-2,2′:5′,2′′-terthiophene
monomers for F8T2 synthesis increased the value of Mn to 33 kDa monomer. This polymer (26) contains an average of 69 thiophene
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
634

NATuRe CATAlysIs
Articles
Fe(acac)3 (10 mol%)
TP1 or TP2 (11 mol%)
AlMe3 (1.0 equiv.)
S
S
H
DEO (0.50–0.55 equiv.)
THF/PhMe
(vol/vol = 2/1, 0.67 M)
70°C, 15 h
S
0.20–0.50 mmol
C6H13
Ph
Ph
S
Ph
Ph
Ph
C6H13
Ph
S
S
S
S
S
S
S
S
S
S
S
S
2
99%
Ph
Ph
C6H13
6
87%
7
83%
8
9
C6H13
94% (99%)a,b
88%a
S
TIPS
S
TIPS
O
S
S
S
O
S
S
TIPS
S
10
76%
11
95%
12
86%
TIPS
S
O
B
O
S
S
S
S
S
TIPS
Bu3Sn
TIPS
SnBu3
S
S
S
S
S
O
B
13
14
15
92%
79%
82%
O
O
O
OMe
S
S
S
S
S
S
18
33%
MeO
16
91%
17
77%
O
O
Ph
N
N
S
S
S
S
N
N
20
81%
19
83% (40%)b,c
90% (1.16 g, 4.0-mmol scale)d
Ph
OMe
OMe
MeO
N
MeO
N
S
S
S
S
N
N
OMe
OMe
21
22
96%
94%
MeO
OMe
Fig. 4 | Scope of iron-catalysed regioselective thienyl C–H/C–H homocoupling. All yields are of isolated products unless otherwise noted. See the
Supplementary Methods and Supplementary Discussion for experimental details. aWith FeCl3·6H2O (5.0ꢀmol%), TP1 (10ꢀmol%), AlMe3 (1.5 equiv.) and
DEO (1.0 equiv.). byields were determined with 1H NMR using 1,1,2,2-tetrachloroethane as an internal standard. cUsing PdCl2(PhCN)2 (3.0ꢀmol%), AgF (2.0
equiv.) and DMSO at 60ꢀ°C for 5ꢀh (Mori’s conditions). dWith Fe(acac)3 (5.0ꢀmol%) and TP8 (5.5ꢀmol%). TIPS, triisopropylsilyl.
units bearing hexyl side chains at predetermined positions44. Figs. 5 and 6 for details). Copolycondensation of two and three
Polymers containing ester and amide groups as electron-acceptor different monomers produced polymers with a ratio of monomers
π motifs formed efficiently as well (27 and 28). The reaction of in the polymers identical to the initial reactant ratio (30 and 31).
4,8-dialkoxybenzo[1,2-b:4,5-b′]dithiophene afforded a polymer The current limitation is that polymerization methodology based
with a degree of polymerization of 22 and a PDI of 2.16 (29). on C–H/C–H homocoupling is not applicable to alternating copo-
The degrees of polymerization of 26 and 29, as determined from lymerization. Alternating C–H/C–H copolymerization is expect-
1
end-group signals in H NMR spectra, were 21 and 28, respec- edly achievable by further development of homocoupling-free
tively, which also corroborate the generation of long polymers twofold C–H activation/cross-couplings between two different
according to degree-of-polymerization values measured by gel C–H substrates using our reported transient connection strategy
permeation chromatography (GPC) analysis (see Supplementary in the presence of linchpin groups13.
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
635

NATuRe CATAlysIs
Articles
FeCl3•6H2O (10 mol%)
TP1 (15 mol%)
AlMe3 (3.0 equiv.)
S
S
H
π
π
H
DEO (2.0 equiv.)
THF/PhMe
S
S
n
(vol/vol = 1/1, 0.60 ml)
70°C, 24 h
0.20 mmol
C8H17
C8H17
C8H17 C8H17
C8H17 C8H17
N
S
S
S
S
S
S
n
n
n
C6H13
C6H13
23
4
5 (F8T2 for organic electronics)
Mn = 20.3 kg mol−1
Mw = 52.9 kg mol−1
Mw/Mn = 2.61
M
n = 32.9 kg mol−1
Mn = 13.8 kg mol−1
Mw = 24.2 kg mol−1
Mw/Mn = 1.75
DP = 24
Mw = 89.4 kg mol−1
Mw/Mn = 2.72
DP = 46
DP = 37
91% yieldb,c
68% yield
86% yielda
C8H17 C8H17
S
S
S
S
N
S
n
n
S
S
C6H13
C6H13
C2H5
C2H5
n
C4H9
C4H9
C6H13
C6H13
26
24
25
Mn = 9.6 kg mol−1
M
n = 32.6 kg mol−1
Mn = 16.0 kg mol−1
Mw = 32.3 kg mol−1
Mw/Mn = 2.01
DP = 26
Mw = 19.5 kg mol−1
Mw/Mn = 2.02
DP = 23
Mw = 69.9 kg mol−1
Mw/Mn = 2.15
DP = 42
60% yield
86% yieldd
93% yield
C10H21
C8H17
C10H21
O
O
O
N(C8H17)
2
C12H25
O
S
S
S
S
S
S
n
O
n
n
C6H13
C6H13
C6H13
C6H13
C8H17
C10H21
27
29
28
Mn = 14.9 kg mol−1
Mw = 42.9 kg mol−1
Mn = 17.2 kg mol−1
Mw = 37.0 kg mol−1
Mw/Mn = 2.16
DP = 22
M
n = 12.2 kg mol−1
Mw = 30.2 kg mol−1
Mw/Mn = 2.49
DP = 18
Mw/Mn = 2.87
DP = 19
85% yield
88% yield
91% yieldc,e
C8H17
C8H17
C8H17
C8H17
N
Mn = 19.7 kg mol−1, Mw = 38.0 kg mol−1
Mw/Mn = 1.93, DP = 35, 98% yield
S
S
S
S
0.5
0.5
C8H17
C8H17
30
C8H17 C8H17
C8H17 C8H17
N
S
S
S
S
S
S
0.33
0.33
0.33
C6H13
C6H13
31
Mn = 23.1 kg mol−1, Mw = 56.3 kg mol−1, Mw/Mn = 2.44, DP = 38, 56% yield
Fig. 5 | Scope of iron-catalysed regioselective thienyl C–H/C–H polycondensation. All yields are of isolated products unless otherwise noted. See the
Supplementary Methods and Supplementary Discussion for experimental details. Mn and Mw were determined with GPC using polystyrene standards
for the polymers isolated by either precipitation or GPC. The degree of polymerization was calculated using Mn. aWith FeCl3·6H2O (5.0ꢀmol%) and TP1
(10ꢀmol%). bWith THF (0.15ꢀml). cFor 48ꢀh. dAt 60ꢀ°C. eWith TP1 (20ꢀmol%).
Conclusions
synthesis of π-conjugated small molecules and macromolecules of
The work presented above highlights the benefits of cooperative use importance in organic electronic device applications. We consider
of two Earth-abundant metals (that is, iron and aluminium) for the that the structural and electronic elements built in the TP ligand
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
636

NATuRe CATAlysIs
Articles
homogeneous reaction mixture was diluted with THF (2ml), quenched by the
addition of a dioxane solution of HCl (4moll−1; 0.50ml) and stirred for 1h. The
crude product was purified by GPC or precipitation.
and DEO/Al(III) combination have struck a good balance for con-
trol of the valence and spin state of organoiron intermediates. The
C–H/C–H polycondensation represents an addition to a still limited
repertoire of organoiron catalysis in polymer synthesis45–48, and will
serve as a model for further development of practically useful iron
catalysis.
Data availability
All of the data supporting the findings of this study, including experimental
procedures and compound characterization, are available within the paper and its
Supplementary Information, or from the authors upon reasonable request.
Methods
General. All air- or moisture-sensitive reactions were performed in a dry reaction
vessel under an argon atmosphere. Air- or moisture-sensitive liquids and solutions
were transferred with a syringe or Tefon cannula. Te water content of solvents
was confrmed to be <30ppm by Karl Fischer titration performed with MKC-
210 (Kyoto Electronics Manufacturing). Analytical thin-layer chromatography
was performed with a glass plate coated with 0.25-mm, 230- to 400-mesh silica
gel containing a fuorescence indicator. Organic solutions were evacuated with a
diaphragm pump through a rotary evaporator. Flash column chromatography was
performed as described by Still et al.49. Preparative-recycling GPC was performed
with an LC-92XX II NEXT instrument (Japan Analytical Industry) equipped with
JAIGEL-2HR polystyrene columns using chloroform as an eluent at a fow rate
Received: 25 January 2021; Accepted: 15 June 2021;
Published online: 19 July 2021
References
1. Handbook of Tiophene-Based Materials: Applications in Organic Electronics
and Photonics Vols 1 and 2 (eds Perepichka, I. F. & Perepichka, D. F.)
(Wiley, 2009).
2. Stuart, D. R. & Fagnou, K. Te catalytic cross-coupling of unactivated arenes.
Science 316, 1172–1175 (2007).
3. Lyons, T. W. & Sanford, M. S. Palladium-catalyzed ligand-directed C–H
functionalization reactions. Chem. Rev. 110, 1147–1169 (2010).
4. Yang, Y., Lan, J. & You, J. Oxidative C–H/C–H coupling reactions between
two (hetero)arenes. Chem. Rev. 117, 8787–8863 (2017).
of 7.5mlmin−1
.
Gas chromatography was performed with a GC-2014 instrument (Shimadzu)
equipped with an ULBON HR-1 capillary column (0.25mm I.D. ×25ml;
0.25μm; Shinwa Chemical Industries). Mass spectra (by gas chromatography–
mass spectrometry) were taken with a Parvum 2 instrument (Shimadzu).
High-resolution mass spectra were taken with an LCMS-IT-TOF mass
5. Yang, Y., Nishiura, M., Wang, H. & Hou, Z. Metal-catalyzed C–H activation
for polymer synthesis and functionalization. Coord. Chem. Rev. 376,
506–532 (2018).
6. Gao, G.-L., Xia, W., Jain, P. & Yu, J.-Q. Pd(II)-catalyzed C3-selective arylation
of pyridine with (hetero)arenes. Org. Lett. 18, 744–747 (2016).
7. Grzybowski, M., Skonieczny, K., Butenschön, H. & Gryko, D. T. Comparison
of oxidative aromatic coupling and the Scholl reaction. Angew. Chem. Int. Ed.
52, 9900–9930 (2013).
8. Takahashi, M. et al. Palladium-catalyzed C–H homocoupling of
bromothiophene derivatives and synthetic application to well-defned
oligothiophenes. J. Am. Chem. Soc. 128, 10930–10933 (2006).
9. Wang, L. & Carrow, B. P. Oligothiophene synthesis by a general C–H
activation mechanism: electrophilic concerted metalation–deprotonation
(eCMD). ACS Catal. 9, 6821–6836 (2019).
10. Arndtsen, B. A., Bergman, R. G., Mobley, T. A. & Peterson, T. H. Selective
intermolecular carbon–hydrogen bond activation by synthetic metal
complexes in homogeneous solution. Acc. Chem. Res. 28, 154–162 (1995).
11. Bauer, I. & Knölker, H.-J. Iron catalysis in organic synthesis. Chem. Rev. 115,
3170–3387 (2015).
12. Shang, R., Ilies, L. & Nakamura, E. Iron-catalyzed C–H bond activation.
Chem. Rev. 117, 9086–9139 (2017).
13. Doba, T., Matsubara, T., Ilies, L., Shang, R. & Nakamura, E.
Homocoupling-free iron-catalysed twofold C–H activation/cross-
couplings of aromatics via transient connection of reactants. Nat. Catal. 2,
400–406 (2019).
14. Yu, R. P., Hesk, D., Rivera, N., Pelczer, I. & Chirik, P. J. Iron-catalysed
tritiation of pharmaceuticals. Nature 529, 195–199 (2016).
15. Gandeepan, P. et al. 3d Transition metals for C–H activation. Chem. Rev. 119,
2192–2452 (2019).
16. Guo, X., Baumgarten, M. & Müllen, K. Designing π-conjugated polymers for
organic electronics. Prog. Polym. Sci. 38, 1832–1908 (2013).
17. Bujak, P. et al. Polymers for electronics and spintronics. Chem. Soc. Rev. 42,
8895–8999 (2013).
spectrometer (Shimadzu) using reserpine (MW 608.2734) as an internal standard.
Analytical GPC was performed with a Prominence instrument (Shimadzu)
equipped with a KF-805L (Shodex) column at 40°C, using chloroform as an eluent
at a flow rate of 1.0mlmin−1. Calibration curves were obtained with ReadyCal-Kit
(Polymer Standards Service) standard polystyrenes. The melting points of solid
compounds were measured on a Mel-Temp capillary melting point apparatus and
were uncorrected. NMR spectra were taken with an ECZ-500 (JEOL) at room
temperature (unless otherwise noted) and reported in ppm. 1H NMR spectra were
internally referenced to tetramethylsilane (0ppm), CHCl3 (7.26ppm), CHDCl2
(5.32ppm), C2HDCl4 (5.97ppm) or (CHD2)(CD3)SO (2.50ppm). 13C NMR spectra
were internally referenced to tetramethylsilane (0ppm), CDCl3 (77.0ppm), CD2Cl2
(53.8ppm), C2D2Cl4 (73.8ppm) or (CD3)2SO (39.5ppm). 19F NMR spectra were
internally referenced to C6F6 (−164.9ppm). 31P NMR spectra were internally
referenced to (CH3O)3PO (2.1ppm). Inductively coupled plasma (ICP) analysis was
performed on Shimadzu ICPS-7510 equipment.
Materials. Unless otherwise noted, reagents were purchased from Tokyo Chemical
Industry, Sigma–Aldrich, FUJIFILM Wako Pure Chemical and other commercial
suppliers, and were used as received. Anhydrous tetrahydrofuran and diethyl ether
were purchased from Kanto Chemical and purified before use using a solvent
purification system (GlassContour) equipped with columns of activated alumina
and supported copper catalyst50. Fe(acac)3 (99.9% trace metal basis) was purchased
from Sigma–Aldrich and used as received. FeCl3·6H2O (over 99.0%) was purchased
from FUJIFILM Wako Pure Chemical and used as received. Diethyl oxalate
was purchased from Tokyo Chemical Industry, degassed by freeze–pump–thaw
cycling three times, dried with 4-Å molecular sieves and kept in a storage flask.
Concentrated sulfuric acid (for trace analysis) and nitric acid (density=1.42gml−1
for trace analysis), as well as chloroform and methanol (both Primepure; for ICP
analysis), were purchased from Kanto Chemical and used as received.
;
18. Shirota, Y. & Kageyama, H. Charge carrier transporting molecular materials
and their applications in devices. Chem. Rev. 107, 953–1010 (2007).
19. Cheng, Y.-J., Yang, S.-H. & Hsu, C.-S. Synthesis of conjugated polymers for
organic solar cell applications. Chem. Rev. 109, 5868–5923 (2009).
20. Wang, C., Dong, H., Hu, W., Liu, Y. & Zhu, D. Semiconducting π-conjugated
systems in feld-efect transistors: a material odyssey of organic electronics.
Chem. Rev. 112, 2208–2267 (2012).
A general procedure for thienyl C–H/C–H homocoupling. A thiophene
(0.40mmol), TP2 (28mg; 0.044mmol) and a THF (tetrahydrofuran) solution of
Fe(acac)3 (0.10moll−1; 0.40ml; 0.040mmol) were added to an oven-dried Schlenk
tube. Then, a toluene solution of AlMe3 (2.0moll−1; 0.20ml; 0.40mmol) was added
dropwise at room temperature and the reaction mixture was stirred for 5min
to give a clear, dark reddish-brown solution. Diethyl oxalate (27μl; 0.20mmol)
was added and the reaction mixture was stirred at 70°C for 15h. The reaction
mixture was cooled to room temperature, diluted with ethyl acetate (4ml) and
quenched carefully with methanol (0.2ml). A saturated aqueous solution of
potassium sodium tartrate (2ml) was added and the mixture was stirred vigorously
until clear phase separation was observed. The aqueous layer was extracted with
dichloromethane and the combined organic layers were washed with brine and
dried over Na2SO4. The solvent was removed under reduced pressure and the crude
product was purified by silica gel column chromatography or filtration.
21. Pouliot, J.-R., Grenier, F., Blaskovits, J. T., Beaupré, S. & Leclerc, M. Direct
(hetero)arylation polymerization: simplicity for conjugated polymer synthesis.
Chem. Rev. 116, 14225–14274 (2016).
22. Mercier, L. G. & Leclerc, M. Direct (hetero)arylation: a new tool for polymer
chemists. Acc. Chem. Res. 46, 1597–1605 (2013).
23. Rudenko, A. E. & Tompson, B. C. Optimization of direct arylation
polymerization (DArP) through the identifcation and control of defects in
polymer structure. J. Polym. Sci. A Polym. Chem. 53, 135–147 (2015).
24. Bura, T., Blaskovits, J. T. & Leclerc, M. Direct (hetero)arylation
polymerization: trends and perspectives. J. Am. Chem. Soc. 138,
10056–10071 (2016).
A general procedure for thienyl C–H/C–H polycondensation. A monomer
(0.20mmol) and TP1 (20mg; 0.030mmol) were added in an oven-dried Schlenk
tube, then the tube was evacuated and refilled with argon three times. Next, a
THF solution of FeCl3·6H2O (0.067moll−1; 0.30ml; 0.020mmol) and a toluene
solution of AlMe3 (2.0moll−1; 0.30ml; 0.60mmol) were added by rinsing the wall
of the Schlenk tube at room temperature. Finally, diethyl oxalate (54μl; 0.40mmol)
was added and the reaction mixture was degassed by freeze–pump–thaw cycling
three times. After stirring with an agitation rate of 500r.p.m. at 70°C for 24h, the
25. Blaskovits, J. T. & Leclerc, M. C–H activation as a shortcut to conjugated
polymer synthesis. Macromol. Rapid Commun. 40, 1800512 (2019).
26. Dudnik, A. S. et al. Tin-free direct C–H arylation polymerization for high
photovoltaic efciency conjugated copolymers. J. Am. Chem. Soc. 138,
15699–15709 (2016).
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
637

NATuRe CATAlysIs
Articles
27. Yoon, K.-Y. & Dong, G. Modular in situ functionalization strategy:
multicomponent polymerization by palladium/norbornene cooperative
catalysis. Angew. Chem. Int. Ed. 57, 8592–8596 (2018).
28. Shang, R., Ilies, L. & Nakamura, E. Iron-catalyzed ortho C–H methylation of
aromatics bearing a simple carbonyl group with methylaluminum and
tridentate phosphine ligand. J. Am. Chem. Soc. 138, 10132–10135 (2016).
29. Shang, R., Ilies, L. & Nakamura, E. Iron-catalyzed directed C(sp2)–H and
C(sp3)–H functionalization with trimethylaluminum. J. Am. Chem. Soc. 137,
7660–7663 (2015).
30. Salazar, C. A. et al. Tailored quinones support high-turnover Pd catalysts for
oxidative C–H arylation with O2. Science 370, 1454–1460 (2020).
31. Boddien, A. et al. Efcient dehydrogenation of formic acid using an iron
catalyst. Science 333, 1733–1736 (2011).
32. Li, P., de Bruin, B., Reek, J. N. H. & Dzik, W. I. Photo- and thermal
isomerization of (TP)Fe(CO)Cl2 [TP=bis(2-diphenylphosphinophenyl)
phenylphosphine]. Organometallics 34, 5009–5014 (2015).
33. Masui, K., Ikegami, H. & Mori, A. Palladium-catalyzed C–H homocoupling
of thiophenes: facile construction of bithiophene structure. J. Am. Chem. Soc.
126, 5074–5075 (2004).
34. Tereniak, S. J., Bruns, D. L. & Stahl, S. S. Pd-catalyzed aerobic oxidative
coupling of thiophenes: synergistic benefts of phenanthroline dione and a Cu
cocatalyst. J. Am. Chem. Soc. 142, 20318–20323 (2020).
35. Cho, H. Y. & Scott, L. T. Oxidative cyclotrimerization of unsaturated
compounds with DDQ and trific acid: an efcient synthetic route to
triply-fused benzene rings. Tetrahedron Lett. 56, 3458–3462 (2015).
36. Wakim, S., Blouin, N., Gingras, E., Tao, Y. & Leclerc, M. Poly(2,7-carbazole)
derivatives as semiconductors for organic thin-flm transistors. Macromol.
Rapid Commun. 28, 1798–1803 (2007).
37. Ikariya, T. & Yamamoto, A. Preparation and properties of methyliron
complexes with tertiary phosphine ligands and their decomposition pathways
through the formation of carbenoid intermediates. J. Organomet. Chem. 118,
65–77 (1976).
38. Carothers, W. H. Polymerization. Chem. Rev. 8, 353–426 (1931).
39. Flory, P. J. Fundamental principles of condensation polymerization. Chem.
Rev. 39, 137–197 (1946).
40. Watson, M. D., Fechtenkötter, A. & Müllen, K. Big is beautiful—“aromaticity”
revisited from the viewpoint of macromolecular and supramolecular benzene
chemistry. Chem. Rev. 101, 1267–1300 (2001).
41. Wu, Y. et al. Infuence of nonfused cores on the photovoltaic
performance of linear triphenylamine-based hole-transporting
materials for perovskite solar cells. ACS Appl. Mater. Interfaces 10,
17883–17895 (2018).
42. Donat-Bouillud, A. et al. Light-emitting diodes from fuorene-based
π-conjugated polymers. Chem. Mater. 12, 1931–1936 (2000).
43. Calió, L., Kazim, S., Grätzel, M. & Ahmad, S. Hole-transport
materials for perovskite solar cells. Angew. Chem. Int. Ed. 55,
14522–14545 (2016).
44. Roncali, J. Conjugated poly(thiophenes): synthesis, functionalization, and
applications. Chem. Rev. 92, 711–738 (1992).
45. Ando, T., Kamigaito, M. & Sawamoto, M. Iron(II) chloride complex for living
radical polymerization of methyl methacrylate. Macromolecules 30,
4507–4510 (1997).
46. Matyjaszewski, K., Wei, M., Xia, J. & McDermott, N. E. Controlled/“living”
radical polymerization of styrene and methyl methacrylate catalyzed by iron
complexes. Macromolecules 30, 8161–8164 (1997).
47. Bullock, R. M. et al. Using nature’s blueprint to expand catalysis with
Earth-abundant metals. Science 369, eabc3183 (2020).
48. Nakamura, E. & Sato, K. Managing the scarcity of chemical elements. Nat.
Mater. 10, 158–161 (2011).
49. Still, W. C., Kahn, M. & Mitra, A. Rapid chromatographic technique for
preparative separations with moderate resolution. J. Org. Chem. 43,
2923–2925 (1978).
50. Pangborn, A. B., Giardello, M. A., Grubbs, R. H., Rosen, R. K. & Timmers, F.
J. Safe and convenient procedure for solvent purifcation. Organometallics 15,
1518–1520 (1996).
acknowledgements
We thank Mitsubishi Chemical Corporation for partial financial support. We thank S.
Kobayashi and T. Yasukawa for generous help with the ICP analysis of the contents of
metal residues. This research is supported by MEXT KAKENHI grant number 19H05459
(to E.N.) and JSPS KAKENHI grant number 19K15555 (to R.S.). T.D. thanks JSPS for the
predoctoral fellowship.
author contributions
E.N. and R.S. guided the research and wrote the manuscript. T.D. performed the
experiments to develop the reaction and study the scope, application and mechanism.
L.I. contributed to the optimization of dimerization. W.S. helped with polymer design,
purification and characterization of properties. All authors contributed to designing the
experiments, analysing the data and editing the manuscript.
Competing interests
E.N., R.S., T.D. and W.S. are inventors on Japanese patent application number 2020-
130678, submitted by The University of Tokyo, which covers synthetic methods
described in this manuscript. L.I. declares no competing interests.
additional information
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s41929-021-00653-7.
Correspondence and requests for materials should be addressed to R.S. or E.N.
Peer review information Nature Catalysis thanks Filipe Vilela and the other, anonymous,
reviewer(s) for their contribution to the peer review of this work.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2021
NaTuRE CaTaLySIS | VOL 4 | JULy 2021 | 631–638 | www.nature.com/natcatal
638