Side chain engineering of quinoxaline-based small molecular nonfullerene acceptors for high-performance poly(3-hexylthiophene)-based organic solar cells

Article abstract of DOI:10.1007/s11426-019-9618-7

Poly(3-hexylthiophene) (P3HT) is one of the most used semiconducting polymers for organic photovoltaics because it has potential for commercialization due to its easy synthesis and stability. Although the rapid development of the small molecular non-fulle

Full text of DOI:10.1007/s11426-019-9618-7

SCIENCE CHINA
Chemistry
•ARTICLES•
https://doi.org/10.1007/s11426-019-9618-7
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Side chain engineering of quinoxaline-based small molecular
nonfullerene acceptors for high-performance poly(3-
hexylthiophene)-based organic solar cells
Bo Xiao1,2†, Qianqian Zhang1,3†, Gongqiang Li3*, Mengzhen Du1,3, Yanfang Geng1*,
Xiangnan Sun1,2, Ailing Tang1, Yingliang Liu4, Qiang Guo4 & Erjun Zhou1,2,4*
1CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, CAS Center for Excellence in Nanoscience, National Center for
Nanoscience and Technology, Beijing 100190, China;
2Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, China;
3Key Laboratory of Flexible Electronic (KLOFE) & Institute of Advanced Materials (IAM), Jiangsu National Synergistic Innovation Center for
Advanced Materials (SICAM), Nanjing Tech University (Nanjing Tech), Nanjing 211816, China;
4Henan Institutes of Advanced Technology, Zhengzhou University, Zhengzhou 450003, China
Received August 30, 2019; accepted September 19, 2019; published online December 11, 2019
Poly(3-hexylthiophene) (P3HT) is one of the most used semiconducting polymers for organic photovoltaics because it has
potential for commercialization due to its easy synthesis and stability. Although the rapid development of the small molecular
non-fullerene acceptors (NFAs) have largely improved the power conversion efficiency (PCE) of organic solar cells (OSCs)
based on other complicated p-type polymers, the PCE of P3HT-based OSCs is still low. In addition, the design principle and
structure-properties correlation for the NFAs matching well with P3HTare still unclear and need to be investigated in depth. Here
we designed a series of NFAs comprised of acceptor (A) and donor (D) units with an A2-A1-D-A1-A2 configuration. These NFAs
are abbreviated as Qx3, Qx3b and Qx3c, where indaceno[1,2-b:5,6-b′]dithiophene (IDT), quinoxaline (Qx) and 2-(1,1-dicya-
nomethylene)rhodanine serve as the middle D, bridged A1 and the end group A2, respectively. By subtracting the phenyl side
groups appended on both IDT and Qx skeletons, the absorption spectra, energy levels and crystallinity could be regularly
modulated. When paired with P3HT, three NFAs show totally different photovoltaic performance with PCEs of 3.37% (Qx3),
6.37% (Qx3b) and 0.03% (Qx3c), respectively. From Qx3 to Qx3b, the removing of phenyl side chain in the middle IDT unit
results in the increase of crystallinity and electron mobility. However, after subtracting all the grafted phenyl side groups on both
IDT and Qx units, the final molecule Qx3c exhibits the lowest PCE of only 0.03%, which is mainly attributed to the serious
phase-separation of the blend film. These results demonstrate that optimizing the substituted position of phenyl side groups for
A2-A1-D-A1-A2 type NFAs is vital to regulate the optoelectronic property of molecule and morphological property of active layer
for high performance P3HT-based OSCs.
P3HT, non-fullerene acceptor, phenyl, quinoxaline, organic solar cells
Citation:
Xiao B, Zhang Q, Li G, Du M, Geng Y, Sun X, Tang A, Liu Y, Guo Q, Zhou E. Side chain engineering of quinoxaline-based small molecular
nonfullerene acceptors for high-performance poly(3-hexylthiophene)-based organic solar cells. Sci China Chem, 2019, 62, https://doi.org/10.1007/
s11426-019-9618-7
†These authors contributed equally to this work.
*Corresponding authors (email: iamgqli@njtech.edu.cn; gengyf@nanoctr.cn; zhouej@nanoctr.cn)
© Science China Press and Springer-Verlag GmbH Germany, part of Springer Nature 2019
chem.scichina.com link.springer.com

2
Xiao et al. Sci China Chem
1 Introduction
=7.37 mA/cm2, PCE=5.12% [29]), BTA2 (VOC=1.22 V, JSC
=6.15 mA/cm2, PCE=4.50% [30]) and BTA3 (VOC=0.90 V,
JSC=9.64 mA/cm2, PCE=5.64% [34]). Therefore, A2-A1-D-
A1-A2 type NFAs can match well with P3HT and it is also
vital to develop new electron-withdrawing A1 unit besides
BT and BTA.
Organic solar cells (OSCs) possess several advantages such
as low-cost, light weight and flexibility, and thus attract
immense attention in recent decades [1–5]. Thanks to the
emergence of diverse promising p-type and n-type organic
semiconductors, the power conversion efficiency (PCE) of
OSCs has been rapidly enhanced from 4%–5% in 2005 [6] to
over 16% in 2019 [7]. However, with the purpose of pursuing
good light-harvesting and suitable electron-donating cap-
ability, the chemical structures of p-type polymers become
more and more complicated [8], which would lead to in-
evitable problems such as high cost and bad reproducibility.
Poly(3-hexylthiophene) (P3HT), by contrast, is a good
choice to promote the commercialization of OSCs technol-
ogy as a simple photovoltaic polymer [9]. However, because
of the high-lying HOMO (highest occupied molecular orbi-
tal) energy level (~−5.0 eV) [10] and band gap of 1.9 eV,
P3HT could only realize a low open-circuit voltage (VOC) of
~0.6 V, short-circuit current (JSC) of ~10 mA/cm2 and PCE of
~5% when blending with PC61BM. The highest PCE of
P3HT-based solar cells increased to 7.40% by using a kind of
complicated indene bisadduct of fullerene derivative (IC70
BA) as early as 2012 [11], but it is at a standstill in recent
years due to the difficulty to fine-tune the properties of
fullerene.
As a generation of promising electron-accepting materi-
als, the small molecular non-fullerene acceptors (NFAs)
came into being as research hotspots due to their advantages
of well-defined molecular structure, easy modification,
facile synthesis and purification [12−14]. In combination
with many excellent polymer donors, a single-junction OSC
displayed an outstanding PCE over 16% [7], benefiting
from the rapid development of NFAs, such as Y6
[15−17]. Compared with the achievements of other polymer
donors, there is still a huge potential to develop high-per-
formance P3HT-based OSCs. The reported NFAs combin-
ing with P3HT have mainly used fluorine [18,19], carbazole
[20], dibenzosilole [21], or bifluorenylidene [22], spirobi-
fluorene [23−25], indeno[2,1-a]fluorine [26], and indaceno
[1,2-b:5,6-b′]dithiophene (IDT) [27−32] as the central core.
Among them, IDT-based A2-A1-D-A1-A2 type NFAs de-
monstrate the best photovoltaic performance, where the
bridged A1 groups are mostly confined to benzothiadiazole
(BT) and benzo[d][1,2,3]triazole (BTA) segments (as
shown in Scheme 1). The A2-BT-IDT-BT-A2 type small
molecules with different end-capped groups could achieve
high JSC and PCE, such as IDT-2BR (VOC=0.84 V; JSC
=8.91 mA/cm2; PCE=5.12% [27]), BT2b (VOC=0.92 V; JSC
=10.02 mA/cm2; PCE=6.08% [33]) and O-IDTBR
(VOC=0.72 V; JSC=13.9 mA/cm2; PCE=6.4% [28]). On the
other hand, A2-BTA-IDT-BTA-A2 based NFAs could realize
high VOC of PCE, such as BTA1 (VOC=1.02 V, JSC
From the viewpoint of chemical structure, Qx unit is also a
kind of benzene-based electron-deficient two-membered
aromatic ring [35–37], similar with BTand BTA. Qx exhibits
medium electron-withdrawing ability between BT and BTA
(BT>Qx>BTA) [28,29,38] and also provides the possibility
of introducing different substituents on the 2- and 3-posi-
tions. Thus, Qx has been widely utilized to construct p-type
photovoltaic polymers with the highest PCE of 9.2% [39]
and 11.7% [40] when blending with fullerene acceptor or
NFA, repectively. However, the reports of Qx-based n-type
photovoltaic materials are very limited. In 2014, Zhang et al.
[41] synthesized a Qx modified C60 derivative of TQMA and
achieved a PCE of 2.8% when blending with P3HT. In 2018,
we synthesized two Qx-based NFAs and named as Qx1 and
Qx1b, where IDT and rhodanine were adopted as the central
and terminal segment. OSCs based on P3HT:Qx1 and P3HT:
Qx1b showed PCEs of 4.03% and 4.81%, respectively [39].
The priliminary results indicate that A2-Qx-IDT-Qx-A2
based materials are effective to match with P3HT, which also
privides a chance to investigate the effect of side chains on
both IDT and Qx units. Therefore, here we synthesized three
NFAs and named as Qx3, Qx3b and Qx3c, respectively,
where 2-(1,1-dicyanomethylene)rhodanine (RCN) was used
as the end group A2 and the molecular backbone was fixed to
RCN-Qx-IDT-Qx-RCN (as shown in Scheme 1(b)). By ad-
justing the numbers of the aromatic side groups in both IDT
an Qx, the optoelectronic properties of three NFAs could be
fine-tuned.
The systematical studies of UV-vis absorption, photo-
luminescence (PL), cyclic voltammetry (CV), space charge
limit current (SCLC) mobility, atomic force microscopy
(AFM) and grazing incident wide-angle X-ray scattering
(GIWAXS) revealed the strong relationship between aro-
matic side groups and molecular conformation, optoelec-
tronic, morphological and photovoltaic properties. Firstly,
the presence of eight phenyl groups attached to both IDT and
Qx units decrease the crystallinity and electron mobility of
Qx3, thus resulting in a low JSC and PCE. Secondly, the
removal of all the eight phenyl groups induces totally planar
molecular conformation of Qx3c, which instead tends to
form worse phase-separation, leading to much lower JSC and
fill factor (FF). The highest PCE of 6.37% was realized for
Qx3b upon only taking away the phenyl side groups attached
to the IDT units, which is one of the highest values for P3HT-
based solar cells with non-fullerene acceptors. Our results
demonstrate that side chain engineering play a vital role in
A2-A1-D-A1-A2 type NFAs and indicate that Qx unit is an

3
Xiao et al. Sci China Chem
Scheme 1 (a) The A2-A1-D-A1-A2 type NFAs with BT, BTA and Qx as the bridged A1 unit; (b) the general molecular backbone of RCN-Qx-IDT-Qx-RCN;
(c) three specific chemical structures of Qx-based NFAs Qx3, Qx3b and Qx3c; (d) the structures of small molecule IDT-2BR, BT2b, O-IDTBR, BTA1, BTA2
and BTA3 (color online).
effective building block to construct NFAs to combine with
P3HT.
zole by converting methyl group to aldehyde group (the
route 3) [42]. Three small molecular acceptors, Qx3, Qx3b
and Qx3c, were obtained through condensation reaction
between di-aldehyde compounds and RCN, respectively, as
shown in Scheme 1. The detailed synthesis procedure was
shown in Supporting Information online. The structures of
three NFAs were confirmed through nuclear magnetic re-
sonance (NMR) and matrix-assisted laser desorption/ioni-
zation time of flight (MALDI-TOF) as shown in Figure S1
(Supporting Information online). The thermal gravimetric
(TGA) analysis shown in Figure S2(a) reveals that Qx3,
Qx3b and Qx3c possess good thermal stability with the
deposition temperature up to 390, 385 and 365 °C, re-
spectively, corresponding to 5% weight loss. In differential
scanning calorimetry (DSC) spectra (Figure S2(b)), Qx3
did not show any peaks at the first cooling and second
rising. However, Qx3b and Qx3c exhibited the alkyl chains
melting and backbone melting peaks at 282/311 °C and
208/299 °C, respectively. In addition, the exothermic peak
2 Results and discussion
2.1 Synthesis of materials
To synthesize three RCN-Qx-IDT-Qx-RCN type NFAs, the
most important intermidiates are 8-bromo-2,3-diphe-
nylquinoxaline-5-carbaldehyde (intermediate 3) and 8-
bromoquinoxaline-5-carbaldehyde (intermediate 5). Inter-
mediate 3 could be easily synthesized by converting one
bromine in 5,8-bromo-2,3-diphenylquinoxaline (inter-
mediate 2) to aldehyde group, as shown in Scheme 2(a) (the
route 1). However, the same method failed to synthesize
intermediate 5, probably due to the high reactivity of H
atoms in 2,3-position of quinoxaline with n-butyllithium
(the route 2). Alternatively, the intermediate 5 could be
obtained from 4-bromo-7-methylbenzo[c][1,2,5]thiadia-

4
Xiao et al. Sci China Chem
Scheme 2 (a) Synthetic routes for the intermidiates; (b) synthetic route of Qx-based small molecules Qx3, Qx3b and Qx3c (color online).
corresponding to recrystallization was observed in Qx3b
(257 °C) and Qx3c (240 °C). These results indicate that
Qx3b and Qx3c may adopt a better interchain packing than
Qx3.
Qx3b and Qx3c demonstrate red-shifted and fine-structured
absorption band in the range from 450 to 800 nm, with two
peaks at 706 and 686 nm, respectively, suggesting that the
removal of phenyl side groups induced more planar con-
jugated backbone. Qx3b and Qx3c exhibit obvious should-
ers near 650 nm, indicating that Qx3b and Qx3c possess a
highly ordered molecular packing due to the intense inter-
molecular interactions compared with Qx3. In addition,
Qx3b and Qx3c show more complementary absorption with
donor P3HT than Qx3. The optical gaps of Qx3, Qx3b and
Qx3c are estimated to be 1.64, 1.59 and 1.60 eV, respec-
tively, from the absorption edges of their low-energy ab-
sorption bands. These results indicate that the adjustment of
phenyl side groups on the IDT units has distinct effect on the
absorption both in solution and thin film.
2.2 Characterization
The UV-vis absorption spectra of Qx3, Qx3b and Qx3c in
chloroform solutions and in films are shown in Figure 1. In
solution, they exhibit two similar absorption bands with the
peak at ~400 and ~650 nm, respectively, which are assigned
to the π-π* and intramolecular charge transfer transition. Qx3
shows an absorption maximum at 613 nm with an absorption
coefficient of 1.03×105 L mol−1 cm−1, while Qx3b and Qx3c
show red-shifted maximum peaks at 642 and 636 nm as well
as slightly higher absorption coefficient of 1.04×105 and
1.10×105 L mol−1 cm−1, respectively. The absorption onsets
of Qx3c and Qx3b in solution sequentially red-shift com-
pared with Qx3. In the thin films, Qx3 exhibits an absorption
band from 500 to 750 nm with maximum at 654 nm, while
The electrochemical properties of Qx3, Qx3b and Qx3c
were evaluated by cyclic voltammetry (CV) measurements
as depicted in Figure 1(c), with thin films on a platinum plate
in acetonitrile solution of 0.1 mol L−1 Bu4NPF6 (Bu=butyl) at
scan rate of 50 mV s−1. The lowest unoccupied molecular

5
Xiao et al. Sci China Chem
Figure 1 UV-vis absorption spectra of Qx3, Qx3b and Qx3c in chloroform solution (a) and thin film together with P3HT absorption in film (b). (c) CV
curves of Qx3, Qx3b and Qx3c. (d) Molecular energy level diagrams of donor and acceptor obtained from CV measurements (color online).
orbitals (LUMO) and HOMO energy levels ELUMO and
EHOMO are estimated according to the equation EHOMO/LUMO
and IDT as well as the Qx and the RCN in Qx3b is esti-
mated to be 11° and 1.1°, respectively, which are slightly
larger than those in Qx3. The dihedral angles between the
phenyl unit attached to the Qx and Qx backbone remain
45°, implying that the phenyl groups attached to IDT shows
negligible influence on the backbone structure. Qx3c pos-
sesses a more planar conjugated backbone with smaller
dihedral angles of 0.52° and 0.23° along the molecular
backbone. Therefore, the phenyl groups attached to the IDT
units prevent the intermolecular π-stacking and play a
major role in impacting the molecular planarity. Figure S3
illustrates the electron distributions of their frontier mole-
cular orbitals (LUMO and HOMO). Their HOMO levels are
located on the electron-rich central part, while LUMO are
concentrated onto the electron-deficient peripheral part,
which will benefit charge transfer. Although the attached
phenyl side groups to both IDT and Qx units cannot con-
tribute to the electron distributions of LUMO and HOMO,
the tailoring phenyl side groups slightly affect the energy
levels of LUMO and HOMO due to their influence on the
molecular conformations. Finally, the LUMO/HOMO en-
ergy levels of Qx3, Qx3b and Qx3c are calculated as −3.17
eV/−5.16 eV, −3.20 eV/−5.19 eV, −3.32 eV/−5.30 eV, re-
spectively.
=
−e(Eonset,ox/red+4.80 eV) from the onset oxidation potential
Eonset,ox and reduction potential Eonset,red and collected in Table
1. The EHOMO/LUMO of Qx3, Qx3b and Qx3c is −5.38 eV/
−3.56 eV, −5.27 eV/−3.63 eV and −5.30 eV/−3.59 eV, re-
spectively. Thus, in comparison with Qx3, the removing
phenyl groups both on IDT and Qx units cause LUMO en-
ergy level down-shift and HOMO energy level up-shift.
Therefore, the electrochemical band gaps decrease, which is
in accordance with the optical results. The results indicate
that the phenyl side groups linking to both D and A1 units
have an influence on the molecular energy level of A2-A1-D-
A1-A2 type NFAs, which may originate from the extent of
molecular planarity. In addition, there is enough driving
force for exciton dissociation between P3HT (−3.10 eV/
−5.00 eV) and these NFAs.
We also performed density functional theory (DFT) cal-
culations through Gaussian 09 at B3LYP/6-31G(d) level to
visualize the chemical geometries and electronic properties
of the prepared acceptors. Figure 2 shows the top view and
side view of the minimum energy conformations of Qx3,
Qx3b and Qx3c. After subtracting the phenyl groups at-
tached to the IDT unit, the dihedral angle between the Qx

6
Xiao et al. Sci China Chem
Table 1 The optoelectronic properties of Qx3, Qx3b and Qx3c
λmax (nm)
Compound
Td (°C)
ε (M−1cm−1)a)
Eg (eV)b)
HOMO (eV)c)
LUMO (eV)c)
Solution
613
Film
654
706
686
Qx3
Qx3b
Qx3c
390
385
365
1.03×105
1.04×105
1.10×105
1.64
1.59
1.60
−5.38
−5.27
−5.30
−3.56
−3.63
−3.59
642
636
a) Molar absorptivity at λmax in solution. b) Estimated from the absorption onset in thin films. c) HOMO and LUMO energy levels were estimated from the
onset oxidation and reduction potentials in the CV measurements carried out on the as-cast thin film with 0.1 mol L−1 Bu4NPF6 electrolyte acetonitrile.
Figure 2 Side view and top view of Qx3, Qx3b and Qx3c optimized geometries (color online).
2.3 Photovoltaic properties
but CN showed positive effect on P3HT:Qx3b device. The
optimal D/A weight ratios were found to be 2:1 and 5:4 for
P3HT:Qx3 and P3HT:Qx3b devices, respectively. Upon pre-
thermal annealing at 120 °C, the highest PCE was achieved
for P3HT:Qx3 devices, while the P3HT:Qx3b device
showed the highest performance after increasing the an-
nealing temperature to 160 °C. The average thickness of
active layer of P3HT:Qx3, P3HT:Qx3b and P3HT:Qx3c
device is 93, 114 and 101 nm, respectively. Detailed per-
formance can be found in Figure S4.
The typical current density-voltage (J-V) curves are de-
picted in Figure 3, and the photovoltaic parameters are
compiled in Table 2. After optimization, P3HT:Qx3 device
afforded an optimal PCE of 3.37% with a VOC of 0.89 V, JSC
of 5.57 mA cm−2, and fill factor (FF) of 0.68. When the
phenyl side groups on the IDT unit were subtracted, the PCE
of devices based on P3HT:Qx3b dramatically increased to
6.37%, with largely improved JSC to 12.87 mA cm−2 and
slightly decreased VOC to 0.75 V and FF to 0.66. A con-
tinuous removal of phenyl side groups on the Qx units re-
sulted in the lowest PCE of 0.03% for P3HT:Qx3c devices
with a largely decreased JSC to 0.14 mA cm−2 and a quite low
The photovoltaic performances of these new acceptors were
evaluated by adopting a conventional architecture of indium
tin oxide (ITO)/PEDOT:PSS/active layer/Ca/Al. The number
average molecular weight (Mn) value of P3HT is 25.2 k with
polydisperisity index of 2.21. In order to exploit the poten-
tials of these new acceptors for the high performance P3HT-
based OSCs, a detailed optimization was carried out through
varying the weight ratio, additive and annealing temperature.
These newly synthesized NFAs have good solubility in
common solvents such as o-dichlorobenzene (ODCB),
chlorobenzene (CB) and chloroform (CF), and thus the
processing solvents were carefully detected. The optimal
solvent was found to be ODCB and CF for P3HT:Qx3 and
P3HT:Qx3b devices, respectively. It is worth noting that
P3HT:Qx3c devices exhibit quite low PCE whatever sol-
vents are used. Taking the common additives used in P3HT-
based devices into consideration, 1,8-diiodooctane (DIO),
acetonitrile (CN) and diphenyl ether (DPE) were selected as
the processing additives. It was found that these additives
had little effect on the performance of P3HT:Qx3 devices,

7
Xiao et al. Sci China Chem
Figure 3 J-V characteristics (a) and EQE curves (b) of P3HT:Qx3 (1:0.5, w/w), P3HT:Qx3b (1:0.8, w/w) and P3HT:Qx3c (1:1, w/w) devices (color online).
Table 2 Photovoltaic performance of P3HT-based OSCs under the illumination of AM 1.5 G, 100 mW cm−2
PCE
Acceptor
VOC (V)
JSC (mA cm−2)
FF
µh (cm2 V−1 s−1)
µe (cm2 V−1 s−1)
Best
Ave.a)
3.22%
6.26%
0.03%
Qx3
Qx3b
Qx3c
0.89
0.75
0.75
5.57
12.87
0.14
0.68
0.66
0.30
3.37%
6.37%
0.03%
1.1×10−5
1.9×10−4
–
4.2×10−6
2.0×10−5
1.3×10−4
a) The average PCE values are calculated from eight devices.
FF to 0.30. The relatively high VOC of P3HT:Qx3 device is
ascribed to the largest LUMO-HOMO offset between Qx3
and P3HT. The removal of phenyl groups decreases VOC
from 0.95 to 0.75 V, which is consistent with the lower-lying
LUMO energy levels of Qx3b and Qx3c. The influence of
phenyl side groups on JSC and FF will be discussed in the
following part. Clearly, the removal of phenyl groups at-
tached to the central IDT unit significantly enhances the PCE
to 6.37%, which is one of the highest values for P3HT-based
non-fullerene OSCs. The photon energy loss (Eloss) values
devices are 34%, 57% and 1.2%, respectively. The JSC values
of devices calculated from integration of the EQE spectra are
in good agreement with the measured JSC values from J-V.
Our results demonstrate that side chain engineering of A2-
A1-D-A1-A2 type NFAs plays a vital role to affect the pho-
tocurrent response.
The exciton dissociation ability in the blend films is also
probed by employing photoluminescence (PL) spectroscopy.
As shown in Figure S6, when excited at 520 nm, efficient PL
quenching occurred for three devices, which indicates the
electron transfer from P3HT to NFA is effective. However,
P3HT:Qx3c blend film reveals relatively incomplete PL
quenching under excitation at 520 nm compared with P3HT:
Qx3 and P3HT:Qx3b devices. It has been accepted that the
significant PL quenching of the NFA emission indicates ef-
ficient hole transfer from acceptor to P3HT through the light
absorption of acceptor, which also can contribute to the
current generation. Therefore, the incomplete PL quenching
for P3HT:Qx3c might be related, at least in part, with the low
JSC.
are calculated by using the common formula Eloss=Eg−eVOC
.
The Eloss values are 0.75, 0.84 and 0.85, respectively, which
suggest there are still large room to improve the photovoltaic
performance via material design.
The external quantum efficiency (EQE) curves as shown in
Figure 3(b) demonstrate that P3HT:Qx3b devices exhibit
obviously broader photocurrent response in the range of
300–800 nm than that of P3HT:Qx3 device from 300 to
750 nm, which is consistent with their UV-vis absorption
spectra of blend films as shown in Figure S5. In addition, the
EQE value of P3HT:Qx3b is much larger than that of P3HT:
Qx3 and P3HT:Qx3c devices, indicating that P3HT:Qx3b
exhibits more efficient photoelectron conversion after sub-
tracting the phenyl side groups on IDT unit and remaining
the phenyl side groups on Qx unit. The quite weak photo-
response of P3HT:Qx3c may be ascribed to the low hole
mobility and the morphology, which will be discussed later.
The maximum EQE values of Qx3, Qx3b and Qx3c-based
In order to further understand the impact of phenyl side
groups on the photovoltaic parameters JSC and FF, the charge
carrier (hole and electron) mobility of these systems were
obtained by using the SCLC method as shown in Figure S7.
The electron mobility of P3HT:Qx3, P3HT:Qx3b and P3HT:
Qx3c systems was determined to be 4.2×10−6, 2.0×10−5 and
1.3×10−4 cm2 V−1 s−1, respectively. Thus, it can be seen that
the removing all the attached phenyl side groups increased

8
Xiao et al. Sci China Chem
the electron mobility, which may be ascribed to the strong
intermolecular π-stacking. The hole mobility of P3HT:Qx3
and P3HT:Qx3b blend was estimated to be 1.1×10−5 and
1.9×10−4 cm2 V−1 s−1, respectively; however the hole mobi-
lity of P3HT:Qx3c is so low that no SCLC region can be
observed in the current density-voltage curve. Thus, P3HT:
Qx3b exhibits relatively balanced electron and hole mobility,
which is attributed to the high JSC value. The relatively small
JSC of P3HT:Qx3 may be ascribed to low charge carrier
mobility and large overlapping light absorption with P3HT
discussed above. The quite low JSC and FF of P3HT:Qx3c
device are mainly due to the unbalanced electron and hole
mobility as well as the incomplete PL quenching.
film demonstrates strong out-of-plane lamellar peak at
0.42 Å−1 (d=1.50 nm) and π-stacking peak at 1.81 Å−1 (d=
0.347 nm), therefore the crystalline Qx3c adopts both edge-
on and face-on orientation, suggesting the removal of phenyl
side groups attached on the Qx unit largely affects the mo-
lecular orientation in thin film. Compared with the GIWAXS
patterns of Qx3b and Qx3c, the weaker scattering features of
Qx3 neat film suggest the lower crystallinity of Qx3, in-
dicating that the substituted phenyl side groups on IDT and
Qx units decreases the crystallinity. From the full width at
half maximum of the (100) peak in the qxy direction, their
crystal coherence length values are estimated according to
the Scherrer’s equation [43,44], to be 10.4 nm for Qx3, 33.9
nm for Qx3b and 39.8 nm for Qx3c , confirming the highly
crystalline Qx3b and Qx3c.
2.4 Molecular orientation
With regard to the blend films, P3HT:Qx3 exhibits an
obvious in-plane π-stacking peak at 1.66 Å−1 (d=0.378 nm)
and an enhanced out-of-plane lamellar peak at 0.36 Å−1 (d=
1.75 nm) due to the overlap of the corresponding peaks of
P3HT and Qx3, indicating that P3HT prefers an increased
crystallinity and an edge-on orientation. The broad lamellar
peak at 0.35 Å−1 and π-stacking peak at 1.25 Å−1 in the qxy
direction similar with that of Qx3 neat film reveal that Qx3
remains its crystalline structure after blending with P3HT,
which may induce relatively large phase-separation. The
clear π-stacking peak of P3HT after blending with Qx3 also
indicates a better crystalline P3HT due to the weak inter-
molecular interaction between P3HT and Qx3. For the
P3HT:Qx3b blend film, the intensities of the in-plane la-
mellar peak at 0.36 Å−1 and the out-of-plane π-stacking peak
at 1.67 Å−1 are enhanced, indicating P3HT and Qx3b adopts
face-on orientation, leading to the stronger hole and electron
mobility in P3HT:Qx3b than that of P3HT:Qx3. The in-
creased intensities of the out-of-plane lamellar peak at 0.36
Å−1 and the in-plane lamellar peak at 0.48 Å−1 indicate the
The influence of phenyl side groups on the molecular
packing in pristine and blend films was evaluated by two-
dimensional grazing incidence wide-angle X-ray scattering
(2D GIWAXS) as shown in Figure 4 and Figure S8. The
scattering profiles of in-plane and out-of-plane obtained
from GIWAXS patterns are shown in Figure 5. From Figure
S8, the P3HT pristine thin film exhibits broad in-plane la-
mellar peak at 0.33 Å−1 (d=1.90 nm) and out-of-plane π-
stacking peak at 1.75 Å−1 (d=0.359 nm), indicating P3HT
adopts a dominant face-on orientation in thin film. For the
Qx3, two broad scattering peaks at 0.33 and 1.25 Å−1 (d=
0.503 nm) in the qxy direction are observed in the neat film,
indicating an amorphous molecular packing. The pristine
film of Qx3b shows a lamellar peak at 0.38 Å−1 (d=1.65 nm)
and broad scattering peak 1.69 Å−1 (d=0.372 nm) in the qz
direction, suggesting Qx3b adopts both edge-on and face-on
orientation, in which the face-on packing relative to the
substrate is dominant because a strong in-plane lamellar peak
at 0.36 Å−1 (d=1.75 nm) can be also observed. Qx3c neat
Figure 4 2D GIWAXS patterns of the pristine films of Qx3 (a), Qx3b (b) and Qx3c (c); as well as the blend films of P3HT:Qx3 (d), P3HT:Qx3b (e), and
P3HT:Qx3c (f) (color online).

9
Xiao et al. Sci China Chem
Figure 5 In-plane (a–c) and out-of-plane (d–f) GIWAXS profiles of the pristine films of Qx3, Qx3b and Qx3c as well as the blend films of P3HT:Qx3,
P3HT:Qx3b and P3HT:Qx3c (color online).
stronger edge-on orientation of Qx3b after blending P3HT.
Consequently, the largely enhanced scattering peaks of
P3HT:Qx3b blend film compared with Qx3b neat film in-
dicate the better cocrystalization of P3HT and Qx3b due to
the strong intermolecular interaction between P3HT and
Qx3b, which is consistent with the nanoscale phase-se-
paration of P3HT:Qx3b blend film discussed later. By con-
trast, the P3HT:Qx3c blend film exhibits dramatically
different scattering pattern from that of Qx3c neat film. The
broad lamellar peak ranging from 0.275 to 0.375 Å−1 in the
qxy direction becomes weakened and narrowed to a range of
0.275–0.32 Å−1, suggesting that P3HT adopts the amorphous
structure due to the absence of lamellar peak at 0.33 Å−1. In
the qz direction, the lamellar peak at 0.42 Å−1 and π-stacking
peak at 1.80 Å−1 down-shift to 0.36 and 1.66 Å−1, respec-
tively, indicating that the introduction of P3HT not only
decreases the crystallinity of Qx3c but also increases the
lamellar and π-stacking distance. The highly crystalline
Qx3c domain in the P3HT:Qx3c blend film leads to high
electron mobility. However the serious imbalance hole/
electron mobility results in quite low JSC and FF.
In short, a large fraction of face-on orientation in P3HT:
Qx3b blend film is benefit for the vertical charge transport,
while the edge-on orientation in P3HT:Qx3 and P3HT:Qx3c
blend film is detrimental to the charge transport. Therefore,
the removal of phenyl side groups attached to the Qx unit
properly regulates the morphological properties of active
layers, while the absence of phenyl side groups possibly
induces excessive aggregation due to the weak steric hin-
drance effect.
2.5 Film morphology
Atomic force microscopy (AFM) was employed to under-
stand the impact of phenyl side groups on the morphology of
the blend films, which not only revealed the molecular
crystallinity but also explained the charge transport property
through the phase-separation size. The height and phase
images of blend films were obtained from AFM character-
izations as shown in Figure 6. Compared with P3HT:Qx3
blend film exhibiting large phase separation with a root-
mean-square roughness (RMS) value of 5.93 nm, P3HT:
Qx3b shows a nanoscale homogeneous surface with a RMS
of 1.84 nm, indicating that subtraction of phenyl groups from
the IDT core unit dramatically increases the miscibility be-
tween P3HT and acceptor, which may also induce the en-
hancement of light harvesting of blend film [45]. The
relatively larger RMS of P3HT:Qx3 blend film may be as-
cribed to the steric molecular geometry of Qx3 due to the
presence of eight phenyl side groups, which prevents the
formation of bicontinuous interpenetrating structure. By
contrast, P3HT:Qx3c blend film exhibits obviously large
aggregations unevenly distributed on the surface because of
the planar molecular conformation. The optical microscopy
images as shown in Figure 7 also reveal much smoother
surface of P3HT:Qx3 and P3HT:Qx3b blend film as well as
an excessive aggregation in P3HT:Qx3c blend film. Notice
that the poor morphology of P3HT:Qx3c cannot be further
improved by the solvent additive and annealing temperature.
Therefore, other donor materials with a strong intermolecular
force to Qx3c might be suitable for the absorption and en-

10
Xiao et al. Sci China Chem
Figure 6 AFM height (top) and phase (bottom) images (5 μm×5 μm) of P3HT:Qx3 (a, d), P3HT:Qx3b (b, e) and P3HT:Qx3c (c, f) (color online).
Figure 7 Optical microscopy images of P3HT:Qx3, P3HT:Qx3b and P3HT:Qx3c blend films (color online).
ergy level matching.
retical calculation further confirms that the effects of phenyl
groups attached to IDT and Qx on molecular conformations
and electronic properties are different. The attached phenyl
groups decrease the miscibility of Qx3 with P3HT, leading to
a large phase-separation in the P3HT:Qx3 blend film. Qx3b
has better cocrystallization with P3HT and shows enhanced
face-on molecular orientation, which thus facilitate the im-
provement of JSC. The absence of phenyl groups on both IDT
and Qx increases the molecular planarity but leads to ex-
cessive aggregation of Qx3c in the blend film. The P3HT:
Qx3b device shows the highest PCE of 6.37%, which is one
of the currently reported top PCE of P3HT-based devices
with non-fullerene acceptors. The results indicate that side
chain engineering plays a vital role to A2-A1-D-A1-A2 type
NFAs to balance the miscibility and crystallinity with P3HT
for achieving bicontinuous interpenetrating network in active
layers. In additions, our works also prove that Qx unit is also
an effective electron-deficient building block, like traditional
BT and BTA, to construct NFAs to combine with P3HT.
3 Conclusions
In summary, we designed three kinds of Qx-based non-
fullerene molecules, Qx3, Qx3b and Qx3c, in order to pur-
sue suitable acceptors matching with the traditional p-type
polymer P3HT. After sequential tailoring the phenyl side
groups attached to the molecular main chain, their optical,
electronic, crystallinity, charge-transport and photovoltaic
properties are systematically investigated. Compared with
parent Qx3, Qx3b shows a red-shifted absorption, a de-
creased LUMO level, an increased HOMO level and finally a
reduced bandgap after removal of four phenyl groups at-
tached to the IDT core. The continuous tailoring four phenyl
rings on the Qx resulted in Qx3c possessing a slightly blue-
shifted absorption, a slight increased LUMO level and de-
creased HOMO level in comparison with Qx3b. The theo-

11
Xiao et al. Sci China Chem
8238
Acknowledgements
This work was supported by the Key Research
21 Lin Y, Li Y, Zhan X. Adv Energy Mater, 2013, 3: 724–728
Program of Frontier Sciences, Chinese Academy of Sciences (QYZDB-
SSW-SLH033), the National Key Research and Development Program of
China (2017YFA0206600) and the National Natural Science Foundation of
China (51673048, 21875052, 51773046, 21602040, 51873044).
22 Gupta A, Rananaware A, Srinivasa Rao P, Duc La D, Bilic A, Xiang
W, Li J, Evans RA, Bhosale SV, Bhosale SV. Mater Chem Front,
2017, 1: 1600–1606
23 Yan Q, Zhou Y, Zheng YQ, Pei J, Zhao D. Chem Sci, 2013, 4: 4389–
4394
24 Liu F, Zhou Z, Zhang C, Vergote T, Fan H, Liu F, Zhu X. J Am Chem
Soc, 2016, 138: 15523–15526
25 Qiu N, Yang X, Zhang H, Wan X, Li C, Liu F, Zhang H, Russell TP,
Chen Y. Chem Mater, 2016, 28: 6770–6778
Supporting information The supporting information is available online at
http://chem.scichina.com and http://link.springer.com/journal/11426. The
supporting materials are published as submitted, without typesetting or
editing. The responsibility for scientific accuracy and content remains en-
tirely with the authors.
26 Baran D, Ashraf RS, Hanifi DA, Abdelsamie M, Gasparini N, Röhr
JA, Holliday S, Wadsworth A, Lockett S, Neophytou M, Emmott
CJM, Nelson J, Brabec CJ, Amassian A, Salleo A, Kirchartz T,
Durrant JR, McCulloch I. Nat Mater, 2017, 16: 363–369
27 Wu Y, Bai H, Wang Z, Cheng P, Zhu S, Wang Y, Ma W, Zhan X.
Energy Environ Sci, 2015, 8: 3215–3221
28 Holliday S, Ashraf RS, Wadsworth A, Baran D, Yousaf SA, Nielsen
CB, Tan CH, Dimitrov SD, Shang Z, Gasparini N, Alamoudi M,
Laquai F, Brabec CJ, Salleo A, Durrant JR, McCulloch I. Nat Com-
mun, 2016, 7: 11585
1
2
Yu G, Gao J, Hummelen JC, Wudl F, Heeger AJ. Science, 1995, 270:
1789–1791
Yan C, Barlow S, Wang Z, Yan H, Jen AKY, Marder SR, Zhan X. Nat
Rev Mater, 2018, 3: 18003
3
4
Tang A, Zhan C, Yao J, Zhou E. Adv Mater, 2017, 29: 1600013
Geng Y, Tang A, Tajima K, Zeng Q, Zhou E. J Mater Chem A, 2019,
7: 64–96
29 Xiao B, Tang A, Zhang J, Mahmood A, Wei Z, Zhou E. Adv Energy
Mater, 2017, 7: 1602269
30 Xiao B, Tang A, Yang J, Wei Z, Zhou E. ACS Macro Lett, 2017, 6:
410–414
31 Tang A, Xiao B, Wang Y, Gao F, Tajima K, Bin H, Zhang ZG, Li Y,
Wei Z, Zhou E. Adv Funct Mater, 2018, 28: 1704507
32 Tang A, Xiao B, Chen F, Zhang J, Wei Z, Zhou E. Adv Energy Mater,
2018, 8: 1801582
33 Xiao B, Tang A, Cheng L, Zhang J, Wei Z, Zeng Q, Zhou E. Sol RRL,
2017, 1: 1700166
5
6
7
8
9
Yang J, Xiao B, Tang A, Li J, Wang X, Zhou E. Adv Mater, 2018, 4:
1804699
Li G, Shrotriya V, Huang J, Yao Y, Moriarty T, Emery K, Yang Y. Nat
Mater, 2005, 4: 864–868
Fan B, Zhang D, Li M, Zhong W, Zeng Z, Ying L, Huang F, Cao Y.
Sci China Chem, 2019, 62: 746–752
Hildner R, Köhler A, Müller-Buschbaum P, Panzer F, Thelakkat M.
Adv Energy Mater, 2017, 7: 1700314
Po R, Bernardi A, Calabrese A, Carbonera C, Corso G, Pellegrino A.
Energy Environ Sci, 2014, 7: 925–943
34 Xiao B, Tang A, Zhang Q, Li G, Wang X, Zhou E. ACS Appl Mater
Interfaces, 2018, 10: 34427–34434
10 Liu Y, Summers MA, Edder C, Fréchet JMJ, McGehee MD. Adv
Mater, 2005, 17: 2960–2964
35 Yuan J, Ouyang J, Cimrová V, Leclerc M, Najari A, Zou Y. J Mater
Chem C, 2017, 5: 1858–1879
11 Guo X, Cui C, Zhang M, Huo L, Huang Y, Hou J, Li Y. Energy
Environ Sci, 2012, 5: 7943
36 Gedefaw D, Prosa M, Bolognesi M, Seri M, Andersson MR. Adv
Energy Mater, 2017, 7: 1700575
12 Zhai W, Tang A, Xiao B, Wang X, Chen F, Zhou E. Sci Bull, 2018, 63:
845–852
37 Liu M, Gao Y, Zhang Y, Liu Z, Zhao L. Polym Chem, 2017, 8: 4613–
4636
13 Mahmood A, Yang J, Hu J, Wang X, Tang A, Geng Y, Zeng Q, Zhou
E. J Phys Chem C, 2018, 122: 29122–29128
38 Xiao B, Tang A, Yang J, Mahmood A, Sun X, Zhou E. ACS Appl
Mater Interfaces, 2018, 10: 10254–10261
14 Yang L, Gu W, Lv L, Chen Y, Yang Y, Ye P, Wu J, Hong L, Peng A,
Huang H. Angew Chem Int Ed, 2018, 57: 1096–1102
15 Yuan J, Zhang Y, Zhou L, Zhang G, Yip HL, Lau TK, Lu X, Zhu C,
Peng H, Johnson PA, Leclerc M, Cao Y, Ulanski J, Li Y, Zou Y.
Joule, 2019, 3: 1140–1151
16 Yang J, Cong P, Chen L, Wang X, Li J, Tang A, Zhang B, Geng Y,
Zhou E. ACS Macro Lett, 2019, 8: 743–748
17 Chen Y, Geng Y, Tang A, Wang X, Sun Y, Zhou E. Chem Commun,
2019, 55: 6708–6710
39 Wang T, Lau TK, Lu X, Yuan J, Feng L, Jiang L, Deng W, Peng H, Li
Y, Zou Y. Macromolecules, 2018, 51: 2838–2846
40 Zheng Z, Awartani OM, Gautam B, Liu D, Qin Y, Li W, Bataller A,
Gundogdu K, Ade H, Hou J. Adv Mater, 2017, 29: 1604241
41 Chen W, Salim T, Fan H, James L, Lam YM, Zhang Q. RSC Adv,
2014, 4: 25291–25301
42 Qi X, Lo YC, Zhao Y, Xuan L, Ting HC, Wong KT, Rahaman M,
Chen Z, Xiao L, Qu B. Front Chem, 2018, 6: 260
43 Rivnay J, Noriega R, Kline RJ, Salleo A, Toney MF. Phys Rev B,
2011, 84: 045203
18 Winzenberg KN, Kemppinen P, Scholes FH, Collis GE, Shu Y, Bi-
rendra Singh T, Bilic A, Forsyth CM, Watkins SE. Chem Commun,
2013, 49: 6307–6309
44 Zhu C, Zhao Z, Chen H, Zheng L, Li X, Chen J, Sun Y, Liu F, Guo Y,
Liu Y. J Am Chem Soc, 2017, 139: 17735–17738
45 Bi YG, Feng J, Ji JH, Yi FS, Li YF, Liu YF, Zhang XL, Sun HB.
Nanophotonics, 2017, 7: 60
19 Holliday S, Ashraf RS, Nielsen CB, Kirkus M, Röhr JA, Tan CH,
Collado-Fregoso E, Knall AC, Durrant JR, Nelson J, McCulloch I. J
Am Chem Soc, 2015, 137: 898–904
20 Kim Y, Song CE, Moon SJ, Lim E. Chem Commun, 2014, 50: 8235–