Rapid access to (cycloalkyl)tellurophene oligomer mixtures and the first poly(3-aryltellurophene)

Article abstract of DOI:10.1039/c9cc07512g

Polytellurophenes represent an emerging class of π-conjugated polymers that possess important characteristics for flexible electronics, such as narrow HOMO-LUMO gaps (E_g) and charge mobilities that can be orders of magnitude larger than their ubiquitous polythiophene counterparts. Herein we use the reactivity of pinacolboronate (BPin) groups attached to tellurophene scaffolds to directly prepare new 2,5-diiodinated monomers, which are then polymerized to afford new low E_g oligomers and polymers. Specifically, mixtures of (cycloalkyl)tellurophene oligomers (of 5 to 12 repeat units) with ring-fused 5- and 6-membered cyclic side chains were prepared via Yamamoto coupling, with a dramatic narrowing of band gap noted when less bulky 5-membered cycloalkyl side groups were present, due to enhanced tellurophene ring coplanarity. Grignard metathesis (GRIM) was also used to gain access to the first poly(3-aryltellurophene) with 94% regioregularity, along with a low E_g of 1.3 eV and an onset of light absorption at ca. 1000 nm.

Full text of DOI:10.1039/c9cc07512g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Rapid access to (cycloalkyl)tellurophene oligomer
mixtures and the first poly(3-aryltellurophene)†
Cite this: Chem. Commun., 2019,
5, 14218
5
Bruno T. Luppi,
Eric Rivard
Robert McDonald,
Michael J. Ferguson,
Lingzi Sang* and
Received 24th September 2019,
Accepted 1st November 2019
*
DOI: 10.1039/c9cc07512g
rsc.li/chemcomm
Polytellurophenes represent an emerging class of p-conjugated
polymers that possess important characteristics for flexible phenes is efficient access to monomers. For example, Seferos
electronics, such as narrow HOMO–LUMO gaps (E ) and charge and coworkers have used Na Te-mediated ring formation to
A key challenge in the development of new polytelluro-
g
2
mobilities that can be orders of magnitude larger than their ubiquitous selectively
polythiophene counterparts. Herein we use the reactivity of pinacol- tellurophenes,
place
alkyl-substituents
while our group used zirconium-instigated
onto
monomeric
3
a,10
boronate (BPin) groups attached to tellurophene scaffolds to directly alkyne coupling, followed by Zr/Te metallacycle transfer, to
prepare new 2,5-diiodinated monomers, which are then polymerized yield tellurophenes with reactive pinacolboronate (BPin) groups
3
b,5,6,11,12
to afford new low E
g
oligomers and polymers. Specifically, mixtures at the 2- and 5-positions.
The latter procedure has
of (cycloalkyl)tellurophene oligomers (of 5 to 12 repeat units) with been used to synthesize (cycloalkyl)tellurophenes such as B-Te-
ring-fused 5- and 6-membered cyclic side chains were prepared 6-B (Scheme 1) that undergo Suzuki–Miyaura coupling to yield
6
b,11
3b
via Yamamoto coupling, with a dramatic narrowing of band gap phosphorescent molecules
or copolymers. Thus far, all
noted when less bulky 5-membered cycloalkyl side groups were copolymers synthesized from B-Te-6-B have wide optical band
3
b
present, due to enhanced tellurophene ring coplanarity. Grignard gaps (42.3 eV), suggesting substantial twisting of the tell-
metathesis (GRIM) was also used to gain access to the first urophene units away from coplanarity. In addition, low poly-
3
b
poly(3-aryltellurophene) with 94% regioregularity, along with a low mer molecular weights (3–7 kDa) are often observed due to
E
g
of 1.3 eV and an onset of light absorption at ca. 1000 nm.
competitive protodeboronation and accompanying polymer
termination.
Polytellurophenes, the heavier element (Te) containing analogues
In this Communication the reactivity of ring-bound BPin
of polythiophenes, have been gaining momentum as promising groups is used to advance tellurophene chemistry in two
1,2
optoelectronic materials. Replacement of S by Te within a distinct ways (Scheme 1): first, the efficient syntheses of the
heterocyclic building block can lead to increased quinoidal char- 2,5-iodonated tellurophene building blocks I-Te-5-I and I-Te-6-I
acter of the resulting polyheteroles (and a reduced HOMO–LUMO (Scheme 2) now enables their direct homopolymerization by
g
gap, E ), along with an increase in hole mobility. Accordingly
polytellurophenes are being explored for next generation organic
3,4
solar cells and field effect transistors. The presence of the heavy
element tellurium can also increase spin–orbit coupling, enabling
efficient phosphorescence to transpire, both in the solid state and
5,6
in the presence of the known quencher O
2
.
Access to photo-
excited triplet states can also yield longer exciton (electron–hole
pair) diffusion lengths which can increase solar cell efficiency and
7–9
simplify working device architectures.
Department of Chemistry, University of Alberta, 11227 Saskatchewan Dr,
Edmonton, Alberta, T6G 2G2, Canada. E-mail: lsang@ualberta.ca,
erivard@ualberta.ca
Scheme 1 Center: BPin-functionalized (cycloalkyl)tellurophenes B-Te-5-
B and B-Te-6-B, and the monoborylated tellurophene (BTe). Left: A
copolymer derived from B-Te-6-B (top) and an example of a protodebor-
onation side-product derived from B-Te-6-B (bottom). Right: New
tellurophene oligomers and polymers synthesized in this study.
†
Electronic supplementary information (ESI) available: Full experimental,
crystallographic and computational details. CCDC 1918983–1918985. For ESI
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c9cc07512g
1
4218 | Chem. Commun., 2019, 55, 14218--14221
This journal is ©The Royal Society of Chemistry 2019

View Article Online
Communication
ChemComm
Scheme 2 Synthesis and oligomerization of I-Te-5-I and I-Te-6-I.
Yamamoto coupling to form mixture of oligomers that show
different optical band gaps (E ) depending on the size of the cyclic
g
13
side group. Second, we report the first poly(3-aryltellurophene)
bearing cumenyl side groups with a low E of ca. 1.3 eV.
Cyclo)tellurophene oligomers (Oligo-Te5 and Oligo-Te6) and
g
(
influence of the cyclic side groups on optoelectronic properties. Our
initial attempts to form cycloalkyl-functionalized tellurophene
homopolymers involved subjecting the known 2,5-dibrominated
3
Fig. 1 UV-Vis spectra and images of Oligo-Te5 and Oligo-Te6 in CHCl .
1
1a
tellurophene Br-Te-6-Br
to widely used Grignard metathesis units in I-[Te6]
7
-I yields nearly perpendicular tellurophene
14
15
(
GRIM) and Yamamoto polymerization protocols. Unfortunately, rings (891 twist angle), while the less hindered Te5 units in
both routes only resulted in low isolated yields (ca. 5–10%) of the I-[Te5] -I enable a preferential coplanar arrangement of the
7
oligomeric mixture Oligo-Te6 (Scheme 2). In order to improve the tellurophene subunits (11 twist angle). As expected, the com-
1
4b
efficiency of polymerization,
I-Te-5-I and I-Te-6-I were prepared by reacting N-iodosuccinimide than in I-[Te6]
NIS) with B-Te-5-B and B-Te-6-B in DMF at 40 1C (25 and 31% yields, penalty of 90 kJ mol
respectively; Scheme 2; see Fig. S4 and S5, ESI,† for the X-ray within a tetrameric model for Oligo-Te6 (I-[Te6]
structures). To our knowledge, these are the first examples of direct arrangement (Fig. S2, ESI†). Closer inspection of the optimized
the 2,5-diiodinated tellurophenes puted E
g
value for planar I-[Te5]
7
-I (2.17 eV) is much smaller
1
9
7
-I (4.18 eV); we also computed an energetic
ꢀ
1
(
for twisting the tellurophene rings
-I) into a planar
4
–
BR to –I conversion on a tellurophene, however related chemistry structures indicates that unfavorable Te–H C (side group)
2
2
1
6,17
involving boronic acid-functionalized thiophenes is known.
interactions in this model for Oligo-Te6 are responsible
I-Te-5-I and I-Te-6-I were stored as solids at ꢀ30 1C in the dark for ring twisting. Thus, we are now investigating copolymers
1
8
to prevent their self-oligomerization.
bearing the less hindered Te5 subunit as only cyclopentane
Yamamoto polymerization of I-Te-5-I and I-Te-6-I in the substitution creates a planar structure for oligo- and polychalco-
0
presence of stoichiometric Ni(COD)
2
/bipy (bipy = 2,2 -bipyridine; genophenes. The computed UV-Vis spectra (by TD-DFT) for
COD = 1,5-cyclooctadiene) (Scheme 2) afforded the corresponding trimeric models of Oligo-Te5 and Oligo-Te6 reproduced the
oligomer mixtures Oligo-Te5 and Oligo-Te6 as red and orange overall spectral features found by experiment (Fig. S3, ESI†). In
solids in 49 and 52% purified yields, respectively. These products each case, the HOMO–LUMO transition has C–C p character
were characterized by NMR (Fig. S11 and S12, ESI†), MALDI-MS in the HOMO (with a small iodine(lp) admixture) and quinoidal
and TGA (Fig. S32, S33 and S42, ESI†); high molecular weight C–C p and Te(lp) contributions to the LUMO (Fig. 2).
polymers were not formed due to a lack of solubility, however
A poly(3-aryl)tellurophene and its small band gap. Soluble
these oligomers provide insight into the influence of the cycloalkyl polytellurophenes with long alkyl side chains have been
3
a,10
side groups on optoelectronic properties. The formation of longer developed by Seferos and coworkers.
The corresponding
(
more soluble) chains of Oligo-Te6 (n r 12 in Scheme 2) versus
Oligo-Te5 (n r 5) is likely due to enhanced backbone ring twisting
in Oligo-Te6, leading to reduced interchain polymer interactions.
Oligo-Te6 yields a lmax at 295 nm in CHCl , while a sub-
3
stantially red-shifted lmax at 471 nm is seen for Oligo-Te5
(Fig. 1). Thus removing one –CH
2
– group from the cyclic side
chain leads to an E decrease to a value of 1.97 eV for Oligo-Te5
g
in solution (estimated from the onset of absorption) that is
significantly red-shifted compared to Oligo-Te6 (2.72 eV) and
other known tellurophene copolymers with Te6 subunits (e.g.,
3
b
Scheme 1). As expected, the E is further reduced to 2.22 and
g
1.82 eV for films of Oligo-Te6 and Oligo-Te5 (Fig. S28, ESI†).
To understand this significant difference in absorption,
density functional theory (DFT) computations [B3LYP/6-31G(d,p)
with LANL2DZ for Te and I)] were performed on the oligomeric
models I-[Te5] -I and I-[Te6]
-I (Fig. S1, ESI†). The presence of Te6 Oligo-Te5.
(
Fig. 2 DFT calculated orbitals for trimeric models of Oligo-Te6 and
7
7
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun., 2019, 55, 14218--14221 | 14219

View Article Online
ChemComm
Communication
Scheme 3 Synthesis of the poly(3-aryltellurophene) PolyTe-cumenyl.
poly(3-aryltellurophenes) are unknown and could yield enhanced
crystallinity/charge mobility in the solid state due to added p–p
stacking interactions. Towards this goal, the known 3-borylated Fig. 3 (a) MALDI-MS and (b) cyclic voltammogram (annealed film) of
1
1a
PolyTe-cumenyl.
tellurophene BTe (Scheme 3) was coupled with 4-iodocumene
to yield the new 3-arylated tellurophene Te-cumenyl (see Fig. S6,
ESI,† for its X-ray structure); in principle, this procedure could
be used to graft a wide range of different aryl groups onto a
tellurophene. Iodination of Te-cumenyl with NIS in DMF at 70 1C
affords the requisite polymer precursor I-Te-cumenyl-I as an
air-stable orange oil (57%) after purification (Scheme 3).
We then polymerized I-Te-cumenyl-I via the GRIM method
i
using stoichiometric PrMgClꢁLiCl for activation (ꢀ78 1C to
2
room temperature), followed by addition of Ni(dppp)Cl as a
pre-catalyst (dppp = Ph
resulting polymer PolyTe-cumenyl (M
2 2 3 2
P(CH ) PPh ) and heating to 80 1C. The
2
0a
w
= 8 kDa, PDI = 1.1)
was obtained an air- and moisture-stable deep blue-purple
solid, while the same route afforded the thiophene analogue
and PolyS-cumenyl (Mw = 12 kDa, PDI = 1.1) as a dark purple
solid. Integration of the major and minor isopropyl signals in
1
the H NMR spectrum of PolyTe-cumenyl indicates 94% head–
tail regioregularity, which matches values found within known
3
a,10
poly(3-alkyltellurophenes).
End group analysis by MALDI-MS shows two major sets of
peaks for PolyTe-cumenyl, assigned to chains with either H/I or
H/H end groups (Fig. 3a). For controlled GRIM (also known as
1
4,21,22
catalyst transfer polymerization)
of a mono-activated
monomer (e.g., RMg-Te-cumeny-I) one would expect only H/I
end groups after work-up. Thus the presence of H/H suggests
that some double activation/metallation of the tellurophene
monomer is occurring, which has been confirmed in our model/
stoichiometric activation studies on I-Te-cumenyl-I (Fig. S25, ESI†).
Similar H/H and H/I end groups are seen in the MALDI-MS
of PolyS-cumenyl (Fig. S36, ESI†). Polymerization trials with
Fig. 4 UV-Vis spectra and photos of PolyS-cumenyl (top) and
PolyTe-cumenyl (bottom) normalized at the most red-shifted absorption
maxima. Films were drop-cast from CHCl
with CHCl (at 35 1C).
3
and solvent vapor annealed
3
Ni(dppe)Cl2 (dppe = Ph P(CH ) PPh ) did not improve end compared to 1.72 eV in solution. Overall, both polymers
2
2 2
2
group control and led to a significant reduction in regio- have film E values that are lower than their respective poly(3-
g
regularity (Fig. S26 and S38, ESI†).
hexylthiophene) (1.9 eV) and poly(3-hexyltellurophene) (1.4 eV)
3
a
The UV-Vis spectra for PolyS-cumenyl and PolyTe-cumenyl counterparts, showing that the presence of an aryl side group
reveal important differences. For PolyS-cumenyl (Fig. 4, top), can help induce planarity in the solid state, presumably via
there is an expected shift in absorption maximum from 455 nm enhanced p–p stacking. Attempts to verify crystallinity by powder
in THF to 467 nm as a film and finally to 542 nm (E
g
= 1.77 eV) XRD were hindered by low signal intensity (Fig. S48, ESI†).
A cyclic voltammogram on a film of PolyTe-cumenyl is
after vapor annealing with CHCl at 35 1C. This observation
3
points to improved polymer backbone ring coplanarity in the shown in Fig. 3b (inset), which contains two redox pairs at
+
film. For PolyTe-cumenyl (Fig. 4, bottom) the lmax shifts from 0.3 and ꢀ2.0 V (vs. Fc/Fc ). From their onset of oxidation and
2
3
6
06 nm in THF to 631 nm as a film, but annealing does not reduction, respectively, we extract an E
g
value of 1.55 V; this
for PolyS-cumenyl
tailing of the absorption profile of PolyTe-cumenyl in the solid (1.94 eV, Fig. S46, ESI†). Additionally, the redox onsets of
have a pronounced effect. However, one sees a substantial value is lower than the corresponding E
g
2
0b
state with an estimated E
g
of 1.30 eV (Fig. S31, ESI†),
PolyTe-cumenyl indicate a HOMO that is 0.20 eV higher and a
1
4220 | Chem. Commun., 2019, 55, 14218--14221
This journal is ©The Royal Society of Chemistry 2019

View Article Online
Communication
ChemComm
3
a,23a
LUMO that is 0.19 eV lower in comparison to PolyS-cumenyl.
The larger value of E obtained by CV compared to UV-Vis is a
6 (a) G. He, B. D. Wiltshire, P. Choi, A. Savin, S. Sun, A. Mohammadpour,
M. J. Ferguson, R. McDonald, S. Farsinezhad, A. Brown, K. Shankar and
E. Rivard, Chem. Commun., 2015, 51, 5444–5447; (b) W. Torres Delgado,
C. A. Braun, M. P. Boone, O. Shynkaruk, Y. Qi, R. McDonald,
M. J. Ferguson, P. Data, S. K. C. Almeida, I. de Aguiar, G. L. C. de Souza,
A. Brown, G. He and E. Rivard, ACS Appl. Mater. Interfaces, 2018, 10,
g
23b
common observation for polychalcogenophenes. As a final point,
polymer degradation was noted upon repeated cycling (Fig. S47,
3
a
ESI†), likely due to the formation of reactive radicals.
12124–12134; (c) E. Hupf, Y. Tsuchiya, W. Moffat, L. Xu, M. Hirai, Y. Zhou,
In conclusion, readily available pinacolboronate (BPin)-
substituted tellurophenes were used for the synthesis of new
tellurophene oligomers and polymers. The effect of the cyclo-
alkyl substituents in Oligo-Te5 and Oligo-Te6 was examined
by theory and experiment, revealing that the less hindered
M. J. Ferguson, R. McDonald, T. Murai, G. He and E. Rivard, Inorg. Chem.,
2019, 58, 13323–13336.
7
8
9
R. D. Pensack, Y. Song, T. M. McCormick, A. A. Jahnke, J. Hollinger,
D. S. Seferos and G. D. Scholes, J. Phys. Chem. B, 2014, 118, 2589–2597.
L. Yang, W. Gu, L. Lv, Y. Chen, Y. Yang, P. Ye, J. Wu, L. Hong,
A. Peng and H. Huang, Angew. Chem., Int. Ed., 2018, 57, 1096–1102.
B. T. Luppi, D. Majak, M. Gupta, E. Rivard and K. Shankar, J. Mater.
Chem. A, 2019, 7, 2445–2463.
5-membered side groups in Oligo-Te5 yields a substantially
more planar backbone and a large reduction in E . The first
g
10 S. Ye, M. Steube, E. I. Carrera and D. S. Seferos, Macromolecules,
2016, 49, 1704–1711.
1 (a) W. Torres Delgado, F. Shahin, M. J. Ferguson, R. McDonald,
G. He and E. Rivard, Organometallics, 2016, 35, 2140–2148; (b) C. A.
Braun, N. Martinek, Y. Zhou, M. J. Ferguson and E. Rivard, Dalton
Trans., 2019, 48, 10210–10219.
2 For related studies, see: (a) B. Jiang and T. D. Tilley, J. Am.
Chem. Soc., 1999, 121, 9744–9745; (b) S. M. Parke, E. Hupf, G. K.
Matharu, I. de Aguiar, L. Xu, H. Yu, M. P. Boone, G. L. C. de Souza,
R. McDonald, M. J. Ferguson, G. He, A. Brown and E. Rivard, Angew.
Chem., Int. Ed., 2018, 57, 14841–14846; (c) S. Das and S. S. Zade,
Chem. Commun., 2010, 46, 1168–1170.
3 (a) K. Ohshimizu, A. Takahashi, Y. Rho, T. Higashihara, M. Ree and
M. Ueda, Macromolecules, 2011, 44, 719–727; (b) For examples of
monomeric (unpolymerized) 3-aryltellurophenes, see: V. Krishna
Karapala, H.-P. Shih and C.-C. Han, Org. Lett., 2018, 20, 1550–1554.
4 (a) I. Osaka and R. D. McCullough, Acc. Chem. Res., 2008, 41,
1202–1214; (b) M. A. Baker, C.-H. Tsai and K. J. T. Noonan,
Chem. – Eur. J., 2018, 50, 13078–13088.
poly(3-aryltellurophene) PolyTe-cumenyl was also synthesized,
1
showing high degree of regioregularity (94%) and small E
g
(1.30 eV, onset of absorption at ca. 1000 nm). The corresponding
3-arylated polythiophene PolyS-cumenyl also yields a smaller E
g
1
when compared to the ubiquitous poly(3-alkylthiophene), thus
chalcogenophene-bound aryl groups can yield advantageous
optoelectronic properties in the solid state. The polymer design
concepts introduced herein should be of great value to those
seeking new optoelectronic materials with thermal stability and
1
2
4
narrow optical band gaps.
This work was supported by NSERC (Discovery Grants to
E. R. and L. S., CREATE (ATUMS) to B. L. and E. R.) as well as
the University of Alberta Future Energy Systems (FES), Compute
Canada and WestGrid. Dr William Torres Delgado, Dr Emanuel
Hupf, Wayne Moffat, Jennifer Jones and Jing Zheng are thanked
for their valuable assistance.
1
1
5 T. Yamamoto, A. Morita, Y. Miyazaki, T. Maruyama, H. Wakayama,
Z.-h. Zhou, Y. Nakamura, T. Kanbara, S. Sasaki and K. Kubota,
Macromolecules, 1992, 25, 1214–1223.
1
1
1
6 C. Thiebes, G. K. Surya Prakash, N. A. Petasis and G. A. Olah, Synlett,
1
998, 141–142.
7 M. Kitano, T. Tanaka and A. Osuka, Organometallics, 2017, 36,
559–2564.
Conflicts of interest
2
8 Both I-Te-5-I and I-Te-6-I yield oligomers (n r 4) over time.†
The authors declare no conflict of interest.
This process takes months for I-Te-6-I but only days for I-Te-5-I.
2,5-Dibromoselenophenes are known to undergo the same process
References
in the solid state: A. Patra, Y. H. Wijsboom, S. S. Zade, M. Li,
Y. Sheynin, G. Leitus and M. Bendikov, J. Am. Chem. Soc., 2008, 130,
6734–6736.
1
2
E. I. Carrera and D. S. Seferos, Macromolecules, 2015, 48, 297–308.
(a) S. M. Parke, M. P. Boone and E. Rivard, Chem. Commun., 2016, 19 S. S. Zade and M. Bendikov, Chem. – Eur. J., 2007, 13, 3688–3700.
2, 9485–9505, and references therein; (b) G. C. Hoover and 20 (a) PolyTe-cumenyl absorbs strongly at the wavelength used for
5
D. S. Seferos, Chem. Sci., 2019, 10, 9182–9188.
(a) A. A. Jahnke, B. Djukic, T. M. McCormick, E. Buchaca Domingo,
C. Hellmann, Y. Lee and D. S. Seferos, J. Am. Chem. Soc., 2013, 135,
light scattering detection (670 nm), thus molecular weights were
determined with refractive index detectors; (b) The optical E of
PolyTe-cumenyl film was determined by the onset of absorption
(1.30 eV) and by Gaussian fit (1.50 eV)†.
3
g
9
51–954; (b) G. He, L. Kang, W. Torres Delgado, O. Shynkaruk,
M. J. Ferguson, R. McDonald and E. Rivard, J. Am. Chem. Soc., 2013, 21 The regioselectivity of the initial C–I metallation step was verified
i
1
35, 5360–5363; (c) L. Lv, X. Wang, X. Wang, L. Yang, T. Dong,
Z. Yang and H. Huang, ACS Appl. Mater. Interfaces, 2016, 8,
4620–34629; (d) M. Al-Hashimi, Y. Han, J. Smith, H. S. Bazzi,
for both I-S-cumenyl-I and I-Te-cumenyl-I by adding PrMgClꢁLiCl,
followed by quenching by HCl, and analysis of the product(s) by
1
3
H NMR spectroscopy.
S. Y. A. Alqaradawi, S. E. Watkins, T. D. Anthopoulos and M. Heeney, 22 For additional leading reviews on this topic, see: (a) K. Okamoto and
Chem. Sci., 2016, 7, 1093–1099.
(a) S. Ye, L. Janasz, W. Zajaczkowski, J. G. Manion, A. Mondal,
T. Marszalek, D. Andrienko, K. M u¨ lllen, W. Pisula and D. S. Seferos,
Macromol. Rapid Commun., 2019, 40, 1800596; (b) W. Xing, P. Ye, 23 (a) A. Patra, Y. H. Wijsboom, G. Leitus and M. Bendikov, Org. Lett.,
J. Lu, X. Wu, Y. Chen, T. Zhu, A. Peng and H. Huang, J. Power
Sources, 2018, 401, 13–19.
G. He, W. Torres Delgado, D. J. Schatz, C. Merten, A. Mohammadpour,
L. Mayr, M. J. Ferguson, R. McDonald, A. Brown, K. Shankar and
E. Rivard, Angew. Chem., Int. Ed., 2014, 53, 4587–4591.
C. K. Luscombe, Polym. Chem., 2011, 2, 2424–2434; (b) J. P. Lutz,
M. D. Hannigan and A. J. McNeil, Coord. Chem. Rev., 2018, 376,
225–247.
4
5
2009, 11, 1487–1490; (b) T. Johansson, W. Mammo, M. Svensson,
M. R. Andersson and O. Ingan ¨a s, J. Mater. Chem., 2003, 13, 1316–1323;
(c) Y. Li, Y. Cao, J. Gao, D. Wang, G. Yu and A. J. Heeger, Synth. Met.,
1999, 99, 243–248.
24 S. R u¨ hle, Sol. Energy, 2016, 130, 139–147.
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun., 2019, 55, 14218--14221 | 14221