Benzopyrazinoisoindigo or Its Reduced Form? Synthesis, Clarification, and Application in Field-Effect Transistors

Article abstract of DOI:10.1002/ejoc.201600363

Benzopyrazinoisoindigo, a pigment reported 40 years ago, should be a good candidate for n-type semiconductors if the reported structure is correct. Reinvestigation of this molecule revealed that it is actually (4H,4′H)-benzopyrazinoisoindigo, which could be considered as the reduced form of benzopyrazinoisoindigo, and hence, it is a good candidate for p-type semiconductors. The route toward the synthesis of this molecule was optimized, and a mechanism was accordingly proposed. A field-effect transistor based on this material showed a hole mobility up to 2.5 × 10^-2 cm² V^-1 s^-1. An expedient route for the synthesis of unexpectedly alkylated (4H,4′H)-benzopyrazinoisoindigo is developed, which also clarifies the previously reported structure. A mechanism is accordingly proposed. A field-effect transistor based on this material shows a hole mobility up to 2.5 × 10^-2 cm² V^-1 s^-1.

Full text of DOI:10.1002/ejoc.201600363

DOI: 10.1002/ejoc.201600363
Communication
Organic Semiconductors
Benzopyrazinoisoindigo or Its Reduced Form? Synthesis,
Clarification, and Application in Field-Effect Transistors
[
a][‡]
[b,c][‡]
[d]
[d]
[c]
[a]
Xiao Wang,
Zhiyuan Zhao,
Na Ai, Jian Pei, Yunqi Liu, and Xiaobo Wan*
Abstract: Benzopyrazinoisoindigo,
a
pigment reported of benzopyrazinoisoindigo, and hence, it is a good candidate
40 years ago, should be a good candidate for n-type semicon- for p-type semiconductors. The route toward the synthesis of
ductors if the reported structure is correct. Reinvestigation of this molecule was optimized, and a mechanism was accordingly
this molecule revealed that it is actually (4H,4′H)-benzopyr- proposed. A field-effect transistor based on this material
azinoisoindigo, which could be considered as the reduced form showed a hole mobility up to 2.5 × 10– cm V s .
2
2
–1 –1
Introduction
1970s stating the successful synthesis of benzopyrazinoiso-
[
9]
indigo,
namely, [3,3′-bipyrrolo[2,3-b]quinoxalinylidene]-2,2′
Small molecular organic field-effect transistors (SM-OFETs) have
attracted much attention in the past decade owing to the exis-
tence of a large pool of tailorable conjugated molecules, the
versatile methods towards their synthesis, and their uniform
(1H,1′H)-dione (Figure 1, compound 1), and related compounds.
We realized that the combination of the electron-deficient char-
acter of both the pyrazine and isoindigo cores would make it an
even better candidate for electron transport. However, owing
to the poor solubility of those pigments, they were not well
characterized (only mass spectroscopic data and elemental
analysis data of N atom were provided), and the actual structure
remains in question. We envisioned that the synthesis of the
alkylated derivatives would cast light on its real structure, and
also, its OFET performance could be tested. Herein, we clarify
the structure of compound 1, which is actually its reduced form
[
1]
structure and performance. Although great progress has been
[
2]
achieved for SM-OFETs, there is still a need to discover novel
organic semiconductors that can be synthesized from commer-
cially affordable starting materials in an easy and scalable way
[
3]
[4]
to lower the costs of SM-OFETs. Recently, isoindigo and its
[
5]
[6]
derivatives were used to construct polymeric OFETs. It is
believed that strong dipolar interactions between the adjacent
isoindigo units favor close packing of the polymer chains, which
improves the performance. Theoretically, such a strong interac-
tion would also favor charge-carrier transport in small mol-
ecules based on the isoindigo core, provided that the length of
conjugation in such small molecules was long enough. How-
ever, SM-OFETs based on the isoindigo core structure are less
studied. Recently, it was reported that a cyanated isoindigo
SM-OFET exhibited ambipolar charge-transfer behavior with
(compound 2, Figure 1), namely, alkylated (4H,4′H)-benzopyr-
azinoisoindigo. Starting from abundant maleic anhydride, a fast
and efficient synthesis route toward this conjugated molecule
is developed and the mechanism is discussed. Contrary to isoin-
digo, this small molecule exhibits p-type semiconductivity, and
SM-OFETs based on this material show a hole mobility up to
2
2
–1 –1
2
.5 × 10– cm V s .
2
–1 –1
electron and hole mobilities of 0.11 and 0.045 cm V s , re-
[
7]
spectively, which implies potential room for further improve-
ment.
In our continuous efforts to design novel derivatives based
[
8]
on the isoindigo core, we noticed a report published in the
[
a] CAS Key Laboratory of Bio-based Materials, Qingdao Institute of
Bioenergy & Bioprocess Technology, Chinese Academy of Sciences,
Qingdao 266101, P. R. China
E-mail: wanxb@qibebt.ac.cn
http://english.qibebt.cas.cn/rh/rs/bic/bbpm/
b] Northeast Normal University,
Changchun 130024 P. R. China
c] Institute of Chemistry, Chinese Academy of Sciences,
Beijing 100080 P. R. China
Figure 1. The structure of the reported benzopyrazinoisoindigo derivative
compound 1) and the actual structure (compound 2).
[
[
[
[
(
d] Peking University,
Beijing 100049, P. R. China
Results and Discussion
In the reported procedure,[9] compound 1a was synthesized in
‡] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejoc.201600363.
three steps (Scheme 1, route A): the reaction of dichloromaleic
Eur. J. Org. Chem. 2016, 2603–2607
2603
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 1. Synthetic route toward the target molecule; NBS = N-bromosuccinimide, DEAD = diethyl azodicarboxylate.
anhydride (3) with ammonia afforded 1H-pyrrole-2,5-dione 4, benzene also worked (Table 1, entry 7). It is clear that amines
which upon treatment with 1,2-diaminobenzene gave an in- promote this reaction, but their role is unclear. However, we can
completely characterized compound, 3-chloro-1H-pyrrolo- confirm that they do not act as a base, as there is no proton to
[
2,3-b]quinoxalin-2(4H)-one (5). Compound 5 was then heated abstract in compound 6. Upon heating the starting material
at refluxed in DMF to effect dimerization to give compound 1a under acidic conditions (Table 1, entry 8), the same product
in 21 % yield. However, the last step is questionable: at first was obtained in 51 % yield, which further testified to the fact
glance, the oxidation of the enamine-like moiety followed by that a base was not necessary for this reaction. Given that
homocoupling accounts for the formation of 1a, but it is un- amines are prone to oxidation, it is possible for them to act as
clear how this process could be realized only by heating 5 in an electron donor under heating conditions. Thus, other com-
DMF. This procedure was repeated in our laboratory, and the pounds that could act as an electron donor were tested
structure of compound 5 was confirmed. Unfortunately, heating (Table 1, entries 9–11). Upon treating the substrate with PPh3
compound 5 in DMF gave a purplish-red powder, which was at room temperature, the desired product was obtained in 10 %
impure and difficult to characterize even after sublimation.
To solve this problem, alkylation of compound 5 was at- The use of P(NEt )
2 3
yield. The yield was improved to 40 % under reflux conditions.
, a more powerful one-electron donor, gave
tempted. Unfortunately, alkylation with an alkyl bromide gave a satisfactory yield of the product at room temperature. Reac-
a mixture, as alkylation of the undesired N atom could not be
avoided. Alkylation under Mitsunobu conditions also failed as
a result of the poor solubility of compound 5. Alternatively,
Table 1. Optimization of the synthesis of 2a.
bromination of compound 5 afforded the geminal halogenated
compound, which was soluble under Mitsunobu conditions and
was successfully alkylated to give compound 6 in good yield
(
Scheme 1, route B). However, subjecting compound 6 to reflux
conditions in DMF resulted in a complex mixture; one of the
products was identified as 2a, but it was afforded in only low
1
yield (ca. 5 %). The H NMR spectrum of 2a clearly shows a
signal at δ = 12.98 ppm, which is assigned to the N–H proton
of the enamine, and high-resolution mass spectrometry explic-
itly gave the mass of the protonated molecule (m/z = 929.7337,
see the Supporting Information). Other side products failed to
be purified by flash chromatography and could not be identi-
fied.
The formation of 2a was surprising, as it seemed that the
corresponding product was formed by the reduction of com-
pound 6 followed by homocoupling, without any reducing rea-
gents. To improve the yield and also to cast light on the mecha-
nism, various conditions were tested, as listed in Table 1. Simply
switching to other solvents did not seem to lead to any im-
provements (Table 1, entries 1 to 4). Inspired by a patent,[10]
Entry
Conditions
Yield [%]
1
2
3
4
5
6
7
8
DMF, 140 °C, 5 h
<5[a]
DMSO, 140 °C, 24 h
xylene, Ar, 110 °C, 12 h
xylene, 110 °C, 24 h
decomposition
no reaction
no target molecule
48[
[a]
a]
xylene, NBu
3
, 110 °C, 12 h
NMP, NBu , 130 °C, 12 h
3
65
xylene, 1,2-diaminobenzene, Ar, 110 °C, 3 h 69
AcOH/THF/H O, 125 °C; then, evaporate to 51[
a]
2
dryness
9
PPh
PPh
3
, THF, r.t., 48 h
, THF, reflux
10
40
58
10
37
71
1
1
0
1
3
2 3 2 2
P(NEt ) , CH Cl , –78 °C to r.t., 48 h
Cu powder, DMSO, r.t., 24 h
CuI, THF/DMF, ascorbic acid, r.t., 24 h
NBu was then added, and the yield increased to 48 % (Table 1,
3
12
13
entry 5). NBu was also used in more polar solvents such as N-
3
1
4
3
nBu SnH/toluene, 60 °C, 24 h
methylpyrrolindone (NMP), and the yield was further improved
to 65 % (Table 1, entry 6). Other amines such as 1,2-diamino-
[a] Reaction was performed in an open system.
Eur. J. Org. Chem. 2016, 2603–2607
www.eurjoc.org
2604
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
0
tions involving Cu were also tested (Table 1, entry 12 and 13), tures and, thus, goes through a similar process to give the de-
and the product was obtained in moderate yield in each case. sired compound, as described in Scheme 2. It is also under-
All these results imply that a one-electron reduction process or standable that switching to more reductive 1,2-diamino-
radical species is involved in this reaction. With this assumption benzene improves the yield, and a more powerful reducing rea-
in mind, a radical initiator, tributyltin hydride (nBu SnH), was gent such as P(NEt ) even allows this reaction to be performed
3
2 3
introduced into the reaction. To our delight, the reaction oc- as room temperature. However, at this stage we are not sure
curred at room temperature and the rate was increased at an why this process can still occur upon heating in acidic solutions.
elevated temperature (60 °C) to deliver the product in a satisfac-
tory yield (71 %; Table 1, entry 14).
With the clarified mechanism in mind, we further optimized
the reaction conditions. To circumvent the problematic alkyl-
The mechanism became clear at that point, as shown in ation of compound 5, dichloromaleic anhydride (3) was con-
Scheme 2, by using nBu SnH as an example. The organotin radi- verted into N-alkylated pyrrole-2,5-dione 4b. Upon treatment
3
cal abstracts the bromine atom from compound 6 to afford with 1,2-diaminobenzene (2 equiv., added portionwise) in xyl-
radical intermediate I, which grabs a hydrogen atom from ene at 110 °C, compound 4b was converted into final product
nBu SnH to give compound 7 with regeneration of the organo- 2b in 51 % yield, as shown in Scheme 3 (a). Furthermore, given
3
tin radical. The chlorine atom of compound 7 is then abstracted that this reaction is not sensitive to solvents, the product was
by the tin radical to form radical intermediate II. Dimerization even obtained in a “quasi-one-pot” manner from one of the
of II affords the final product. During this process, at least most abundant chemicals, maleic anhydride (8), as shown in
2
equivalents of nBu SnH is needed. In reality, an excess Scheme 3 (b). Treatment of compound 8 with an alkylamine
3
amount of nBu SnH was needed, which is in accordance with gave the corresponding N-alkylated pyrrole-2,5-dione, which
3
the proposed mechanism. Although compound 7 could not be was then treated with Br2 to afford 3,4-dibromo-1-octyl-1H-
directly prepared from compound 5, we managed to synthesize pyrrole-2,5-dione. The excess amounts of Br and HBr generated
2
octylated derivative 7b in an alternative way (see the Support- in situ were then evaporated, and without isolation, 1,2-di-
ing Information, p. S6), which was converted into dimerized aminobenzene (2 equiv.) was added portionwise to the mixture,
product 2b upon treatment with nBu SnH. This confirmed that and the mixture was heated to reflux. Finally, an adequate
3
compound 7 was the intermediate involved in the reaction.
amount of water was added to the mixture, and the solvent was
slowly evaporated. The residue was then crystalized in MeOH
to give 2b in an overall 10 % yield. No isolation or chromatogra-
phy was needed before the final crystallization step, which
makes this route suitable for scale up.
Scheme 2. Proposed mechanism of the reaction.
Scheme 3. Optimized synthetic conditions.
0
Other conditions (e.g., amines, phosphines, Cu ) also pro-
mote this radical process. At first glance, the role of the trialk-
Compounds 2a and 2b are deep-red crystalline powders,
ylamine in this reaction is not clear, as the redox ability of tri- which show a maximum UV/Vis absorption band at around
alkylamines is quite weak. However, it is known that trialkyl- 506 nm, as shown in Figure 2 (a) and Table 2. Relative to that
amines can abstract the iodine atom from alkyl iodides to gen- observed for isoindigo, which shows λmax = 496 nm, the inten-
erate an alkyl radical and initiate a living radical polymeriza- sity of the band is largely increased, which reflects the effect of
[
11]
tion.
The formation of an iodine radical–trialkylamine com- extending the chromophore. Contrary to isoindigo, which is an
plex and/or a more complicated complex is the driving force. electron acceptor, 2a and 2b are good electron donors, as
Although abstraction of a bromine atom by NBu is more diffi- shown by the cyclic voltammetry diagrams (Figure 2, b) Two
3
cult, we postulate that it could be initiated at elevated tempera- reversible oxidation peaks located at 0.54 and 0.95 eV are ob-
Eur. J. Org. Chem. 2016, 2603–2607
www.eurjoc.org
2605
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
served. No reductive peak is observed upon scanning towards
negative potentials. Thus, 2a and 2b can be considered the
reduced forms of benzopyrazinoisoindigo (1). Notably, 2a is
very stable under electrochemical conditions. Even after
500 scans, little decay of its electrochemical activity is detected,
as shown in Figure 2 (b). However, all attempts to convert com-
pounds 2a/2b into 1a/1b by using various oxidants failed [or-
ganic oxidants such as PhI(OAc) , 7,7,8,8-tetracyanoquinodi-
2
methane, 2,3-dichloro-5,6-dicyanobenzo-1,4-quinone; inorganic
oxidants such as PbO , I , MnO , AgNO ], and we then came to
2
2
2
3
the conclusion that compound 1 reported in the literature is
actually reduced form 2.
Figure 3. (a) Output characteristics and (b) transfer characteristics of 2a OFET
devices; VDS = drain-to-source voltage; VGS = gate-to-source voltage; device
dimensions: ratio of channel width to channel length: 175; device was meas-
ured under ambient conditions.
Conclusions
In conclusion, aiming at alkylated benzopyrazinoisoindigo,
(4H,4′H)-benzopyrazinoisoindigo, was unexpectedly obtained.
The reaction conditions were optimized, and C -(4H,4′H)-benzo-
8
pyrazinoisoindigo could even be obtained in a “quasi-one-pot”
manner from maleic anhydride, one of the most abundant
chemicals, which may lower the cost of producing such conju-
gated molecules for OFETs. A mechanism involving a one-elec-
tron reduction/radical process was accordingly proposed. This
work also clarified the structure of an long-known pigment re-
ported 40 years ago. Contrary to isoindigo, (4H,4′H)-benzo-
pyrazinoisoindigo is a p-type semiconductor with excellent
electrochemical stability, and its HOMO energy level matches
well with the work function of Au and is suitable to fabricate
p-type OFETs. A field-effect transistor based on this material
Figure 2. (a) UV/Vis absorption spectra of isoindigo (black line) and 2a (red
–
5
line) in CH
nBu NPF /CH
black line) and 500th cycle (red line) are shown.
2
Cl
2
solution (1 × 10
M); (b) cyclic voltammetry diagrams of 2a in
solution (scan rate: 0.1 V s–1, vs. Ag/AgCl), the 1st cycle
4
6
2
Cl
2
(
Table 2. Optical (UV/Vis) properties and electrochemical data.[a]
ox
onset
opt
λ
max [nm]
E
E
g
LUMO
[eV]
HOMO
[eV]
in solution
[eV]
[eV]
2a
506
0.54
2.05
–2.89
–4.94
showed a hole mobility up to 2.5 × 10– cm V s .
2
2
–1 –1
The HOMO energy level of compound 2 was calculated by
ox
using the equation E
= –(Eonset – ferrocene) – 4.8 eV, in
HOMO
ox
which Eonset is the onset of the oxidation potential. The LUMO
energy level was deduced from the equation E = EHOMO
Acknowledgments
This work was supported by the “100 Talents Program” from
the Chinese Academy of Sciences and the National Science
Foundation of China (NSFC) (grant numbers 21402220,
+
LUMO
opt
Eg , in which Eg
opt
is the optical band gap. The HOMO/LUMO
levels of 2a are –4.94 and –2.89 eV, respectively. The suitable
HOMO level and its excellent electrochemical stability indicate
that 2a is an idea material to fabricate p-type OFETs.
5
1573204).
To evaluate the OFET performance, (4H,4′H)-benzopyrazino-
isoindigo with three different alkyl chains (branched C , n-C ,
2
0
8
Keywords: Conducting materials · Dyes/Pigments · Organic
field-effect transistors · Semiconductors · Synthetic methods
and branched C ; compounds 2a–c) were synthesized, and top-
8
gate/bottom-contact OFET devices were fabricated (Table 3).
Representative transfer and output curves are given in Figure 3.
Films of 2a gave the best performance, with the highest hole
[
[
1] a) J. Mei, Y. Diao, A. L. Appleton, L. Fang, Z. Bao, J. Am. Chem. Soc. 2013,
135, 6724–6746; b) H. L. Dong, X. L. Fu, J. Liu, Z. R. Wang, W. P. Hu, Adv.
Mater. 2013, 25, 6158–6183.
2] H. Minemawari, T. Yamada, H. Matsui, J. Tsutsumi, S. Haas, R. Chiba, R.
Kumai, T. Hasegawa, Nature 2011, 475, 364–367.
–2
2
–1 –1
mobility (μ ) of 2.5 × 10 cm V s , the lowest threshold volt-
h
age (V ) value of 3 V, and the highest on/off current ratio (I /
T
on
6
Ioff) of 10 , which indicate that side-chain engineering also has
an important influence on device performance.
[3] T. V. Pho, J. D. Yuen, J. A. Kurzman, B. G. Smith, M. Miao, W. T. Walker, R.
Seshadri, F. Wudl, J. Am. Chem. Soc. 2012, 134, 18185–18188.
[
4] a) E. G. Wang, W. Mammo, M. R. Andersson, Adv. Mater. 2014, 26, 1801–
826; b) R. Stalder, J. G. Mei, K. R. Graham, L. A. Estrada, J. R. Reynolds,
1
Table 3. OFET performance of 2a–c thin films under ambient conditions.
Chem. Mater. 2014, 26, 664–678; c) T. Lei, J.-Y. Wang, J. Pei, Acc. Chem.
Res. 2014, 47, 1117–1126.
μ
h
[cm V s–1
2
–1
]
V
T
[V]
Ion/Ioff
[
5] a) R. S. Ashraf, A. J. Kronemeijer, D. I. James, H. Sirringhaus, I. McCulloch,
Chem. Commun. 2012, 48, 3939–3941; b) I. Meager, M. Nikolka, B. C.
Schroeder, C. B. Nielsen, M. Planells, H. Bronstein, J. W. Rumer, D. I. James,
R. S. Ashraf, A. Sadhanala, P. Hayoz, J.-C. Flores, H. Sirringhaus, I. McCu-
6
2
2
2
a
b
c
0.025
0.0057
0.004
3
–40
–23
10
10
10
6
5
Eur. J. Org. Chem. 2016, 2603–2607
www.eurjoc.org
2606
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
lloch, Adv. Funct. Mater. 2014, 24, 7109–7115; c) T. Hasegawa, M. Ashi-
zawa, H. Matsumoto, RSC Adv. 2015, 5, 61035–61043; d) W. Yue, R. S.
Ashraf, C. B. Nielsen, E. Collado-Fregoso, M. R. Niazi, S. A. Yousaf, M.
Kirkus, H. Y. Chen, A. Amassian, J. R. Durrant, I. McCulloch, Adv. Mater.
[8] a) N. Zhao, L. Qiu, X. Wang, Z. J. An, X. Wan, Tetrahedron Lett. 2014, 55,
1040–1044; b) S. Xu, N. Ai, J. Zheng, N. Zhao, Z. Lan, L. Wen, X. Wang, J.
Pei, X. Wan, RSC Adv. 2015, 5, 8340–8344.
[9] H. Nishi, K. Kitahara, S. Tokita, Nippon Kagaku Kaishi 1977, 146–149.
[10] D. Peter, D. Kar, BASF AG, FR 1410733 (A), 1965.
2015, 27, 4702–4707.
[
[
6] G. Kim, S.-J. Kang, G. K. Dutta, Y.-K. Han, T. J. Shin, Y.-Y. Noh, C. Yang, J.
Am. Chem. Soc. 2014, 136, 9477–9483.
7] W. Yue, T. He, M. Stolte, M. Gsanger, F. Wurthner, Chem. Commun. 2014,
[11] a) Goto, T. Suzuki, H. Ohfuji, M. Tanishima, T. Fukuda, Y. Tsujii, H. Kaji,
Macromolecules 2011, 44, 8709–8715; b) A. Goto, A. Ohtsuki, H. Ohfuji,
M. Tanishima, H. Kaji, J. Am. Chem. Soc. 2013, 135, 11131–11139.
5
0, 545–547.
Received: March 23, 2016
Published Online: May 3, 2016
Eur. J. Org. Chem. 2016, 2603–2607
www.eurjoc.org
2607
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim