Side chain effect on poly(beznodithiophene-co-dithienobenzoquinoxaline) and their applications for polymer solar cells

Article abstract of DOI:10.1016/j.polymer.2015.11.052

Two conjugated polymers (PBDT86-TQ and PBDT88-TQ) with different side chains, but with the same polymer backbone, have been prepared for polymer solar cells. The side chains in the two polymers induced different steric hindrance. PBDT86-TQ showed the 0-0 absorption peak at ～740 nm, and it was blue-shifted to ～660 nm for PBDT88-TQ as a result of strong steric hindrance from two dioctyl chains. Further, the HOMO energy levels of PBDT86-TQ (-5.10 eV) and PBDT88-TQ (-5.54 eV) also showed a large difference. The hole mobility of PBDT88-TQ is ～70 times higher than that of PBDT86-TQ. The photovoltaic cells based on PBDT86-TQ gave a very low PCE of 0.69-0.80%, and for PBDT88-TQ-based cells, the PCEs were improved to 3.70-4.50% with a V_oc of up to 0.98 V and improved J_sc and FF. These results show the importance of side chains on the design of conjugated polymers for high performance polymer solar cells.

Full text of DOI:10.1016/j.polymer.2015.11.052

Polymer 82 (2016) 228e237
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Side chain effect on poly(beznodithiophene-co-
dithienobenzoquinoxaline) and their applications for polymer solar
cells
, d,
Jingzhe Fan a, Yong Zhang a *, Caili Lang a, Meng Qiu b, Jinsheng Song c, Renqiang Yang b,
Fengyun Guo a, Qingjiang Yu a, Jinzhong Wang a, Liancheng Zhao a
a School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China
b CAS Key Laboratory of Bio-based Materials, Qingdao Institute of Bioenergy and Bioprocess Technology, Chinese Academy of Sciences, Qingdao 266101,
China
c Key Laboratory for Special Functional Materials of Ministry of Education, Henan University, Kaifeng 475004, China
d State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou 510640, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Two conjugated polymers (PBDT86-TQ and PBDT88-TQ) with different side chains, but with the same
polymer backbone, have been prepared for polymer solar cells. The side chains in the two polymers
induced different steric hindrance. PBDT86-TQ showed the 0-0 absorption peak at ~740 nm, and it was
blue-shifted to ~660 nm for PBDT88-TQ as a result of strong steric hindrance from two dioctyl chains.
Further, the HOMO energy levels of PBDT86-TQ (ꢀ5.10 eV) and PBDT88-TQ (ꢀ5.54 eV) also showed a
large difference. The hole mobility of PBDT88-TQ is ~70 times higher than that of PBDT86-TQ. The
photovoltaic cells based on PBDT86-TQ gave a very low PCE of 0.69e0.80%, and for PBDT88-TQ-based
cells, the PCEs were improved to 3.70e4.50% with a Voc of up to 0.98 V and improved Jsc and FF. These
results show the importance of side chains on the design of conjugated polymers for high performance
polymer solar cells.
Received 12 August 2015
Received in revised form
22 October 2015
Accepted 25 November 2015
Available online 28 November 2015
Keywords:
Conjugated polymer
Polymer solar cell
Side chain effect
© 2015 Elsevier Ltd. All rights reserved.
1. Introduction
improved performance. A feasible strategy is to design the electron-
rich donor and electron-withdrawing acceptor to synthesize donor-
Over the past decade, conjugated polymer solar cells (PSCs) have
been widely investigated as one of the potential renewable energy
sources with the advantages of such as low-cost, light-weight,
flexible and large area through ink-jet printing technology, in
comparison with the traditional silicon-based solar cells [1e5]. The
most successful polymer solar cell is based on the solution pro-
cessable bulk heterojunction (BHJ) photovoltaic cell, which has a
blend film of conjugated polymer and a fullerene derivative as the
active layer. In the early years, the most efficient BHJ cells were
based on regioregular poly(3-hexylthiophene) (P3HT) and [6,6]-
phenyl-C61-butyric acid methyl ester (PC61BM) blend [6]. However,
due to the limit of P3HT on light harvesting, the researchers have
developed a lot of other low band-gap conjugated polymers with
acceptor (DeA) alternating conjugated polymer. The power con-
version efficiencies (PCEs) based on these conjugated polymers
have gradually increased to ~10% after the complicated optimiza-
tions on materials and device engineering optimizations [7e13].
Benzodithiophene (BDT) is one of the promising donor units in
D-A type conjugated polymers with good performances. The opti-
cal, electrochemical and photovoltaic properties of BDT-based
conjugated polymers can be finely tuned by copolymerizing BDT
with different acceptors, such as benzothiadiazole, naphtha[1,2-
c:5,6-c]bis [1,2,5]thiadiazole, thieno[3,4-b]dithiophene, diketo-
pyrrolopyrrole and N-alkylthieno[3,4-c]pyrrole-4,6-dione, etc.
[14e18] The intramolecular charge transfer from BDT to these
electron acceptors can effectively reduce the band-gap of conju-
gated polymers, resulting an enhanced light harvesting. For
example, Yu and coauthors reported a series of D-A conjugated
polymer based on BDT and thieno[3,4-b]thiophene, which exhibi-
ted lower band-gap of 1.60e1.80 eV, broader absorption of up to
~800 nm and better photovoltaic properties with PCEs of 6%e7%
* Corresponding author. School of Materials Science and Engineering, Harbin
Institute of Technology, Harbin 150001, China.
E-mail address: yongzhang@hit.edu.cn (Y. Zhang).
http://dx.doi.org/10.1016/j.polymer.2015.11.052
0032-3861/© 2015 Elsevier Ltd. All rights reserved.

J. Fan et al. / Polymer 82 (2016) 228e237
229
[16]. Hou and You et al. also synthesized a lot of BDT-based poly-
mers by either modifying BDT or changing acceptors, combining
with the device engineering, the PCEs of these polymers-based
PSCs could be up to 9% with open circuit voltage (Voc) of
0.6e0.8 V and high current density [11,12,19]. In addition, because
gilding the devices, the oxide layer was passivated with a thin
divinyltetramethyldisiloxane-bis(benzocyclobutene) (BCB) buffer
layer. BCB precursor solution in toluene was spun onto the silicon
oxide and subsequently annealed at 250 ꢁC overnight. The total
capacitance density measured from parallel-plate capacitors was
10.6 nF/cm2. Polymer solution in CB was first filtered through
of the better planarity of BDT unit, the pep stacking among poly-
mer chains would be enhanced through selecting the appropriate
acceptor units.
0.2 mm syringe filters, spin-coated onto the BCB-modified substrate,
and then annealed at 110 ꢁC for 10 min under nitrogen. Output and
transfer characteristics of the transistor devices were performed in
The long or branched alkyl side chains are commonly intro-
duced onto polymer backbone to improve the solubilities in aiming
for the solution porcessable PSCs. Beside this, the side chains also
have significant impact on the optical, electrical properties and the
morphology of the films of conjugated polymers, which will result
in the various performances of PSCs [20e30]. Therefore, a fine se-
lection and tune for the side chains are crucial to improve the
overall performance of PSCs. Recenlty, more and more studies are
focusing on the exploration of side chains, which have been
extended from the pure alkyl chains to the alkyl-substituted aro-
matic side chains with the improved photovoltaic performances.
Although the side chain engineering is becoming one of the key
a
N2-filled glovebox using an Agilent 4155B semiconductor
parameter S6 analyzer. The field-effect mobility was calculated in
the saturation regime from the linear fit of (Ids)1/2 vs VGS. The
threshold voltage (VTH) was estimated as the x intercept of the
linear section of the plot of (Ids)1/2 vs VGS
The PSCs were fabrication as following ITO-coated glass sub-
strates (15 /sq.) were cleaned with detergent, de-ionized water,
acetone, and isopropyl alcohol. A thin layer (ca. 40 nm) of
PEDOT:PSS (Baytron® P VP AI 4083, filtered at 0.45
m) was first
.
U
m
spin-coated on the pre-cleaned ITO-coated glass substrates at
5000 rpm and baked at 140 ꢁC for 10 min under ambient condi-
tions. The substrates were then transferred into an argon-filled
glove-box. Subsequently, the polymer:PC71BM active layer (ca.
90 nm) was spin-coated on the PEDOT:PSS layer from a homoge-
neously blended solution. The solution was prepared by dissolving
the polymer at a blend weight ratio of 1:3 in o-dichlorobenzene and
issues in the development of new conjugated polymers,
a
comprehensive understanding of the relationship between the side
chains of conjugated polymers and their photovoltaic perfor-
mances is still needed. In this work, we report two copolymers of
BDT and dithienobenzoquinoxaline (TQ) units, PBDT86-TQ and
PBDT88-TQ, which have different side chains and substitute posi-
tions, to investigate the side chain effect on their optical and
photovoltaic properties. We found that the two polymers showed a
significant difference on the optical, electrochemical, field-effect
transistor (FET) and photovoltaic properties even that they have
the same polymer backbone. The absorption maxima at low energy
band for PBDT86-TQ polymer is at ~740 nm, and but the peak for
PBDT88-TQ is blue-shifted to ~660 nm. The HOMO energy level of
PBDT86-TQ is measured to be only ꢀ5.10 eV, and it is downshifted
to ꢀ5.54 eV for PBDT88-TQ. The photovoltaic devices based on
PBDT86-TQ showed very low PCEs of 0.69%e0.80%, but PCEs for
PBDT88-TQ-based devices increased by 5 times to 3.70%e4.54%
with higher Voc of up to 0.98 V and improved current density.
filtered with a 0.2 mm PTFE filter. At the final stage, the substrates
were pumped down to high vacuum (<6 ꢂ 10ꢀ7 Torr), and calcium
(30 nm) topped with aluminum (100 nm) was thermally evapo-
rated onto the active layer through shadow masks to define the
active area of the devices. The un-encapsulated solar cells were
tested under ambient conditions using a Keithley 2400 SMU and an
Oriel Xenon lamp (450 W) with an AM1.5 filter. A mask was used to
define the device illumination area of 10.08 mm2 to minimize
photocurrent generation from the edge of the electrodes. The light
intensity was calibrated to 100 mW/cm2 using a calibrated silicon
solar cell with a KG5 filter, which has been previously standardized
as the National Renewable Energy Laboratory.
2.3. Synthesis
2. Experimental
2.3.1. 2-(20-Hexyldecyl)thiophene (1)
2.1. Materials and general characterization methods
To a solution of thiophene (5 g, 60 mmol) in THF (50 mL) was
added n-BuLi (25 mL, 2.5 M in hexane) at ꢀ78 ꢁC. The solution was
stirred for 30 min at ꢀ78 ꢁC and 1 h at 0 ꢁC. The 2-hexyldecyl
bromide (20 g, 66 mmol) was then added at 0 ꢁC and the mixture
was stirring for overnight before pouring into water and extracting
with hexane. The hexane phase was dried over Na2SO4. After
removing the solvent, the crude product was purified by distillation
All reagents were purchased from commercial sources without
further purification. Benzo[1,2-b:4,5-b0]dithiophene-4,8-dione (2),
2,3-dioctylthiophene (5) and compound 8 were prepared by
following literature procedures. UVevis spectra were obtained
using a PerkineElmer Lambda-9 spectrophotometer. The 1H and
13C NMR spectra were collected on a Bruker AV 300 and AV 500
spectrameter operating at 300 MHz and 125 MHz in deuterated
chloroform (or chloroform/TCB) solution with TMS as reference,
respectively. Cyclic voltammetries of polymer films were conducted
to give a colorless liquid (13.7 g, 75%). 1H NMR (300 MHz, CDCl3,
d):
7.15 (dd, 1H), 6.96 (t, 1H), 6.77 (m, 1H), 2.80 (d, J ¼ 6.66 Hz, 2H), 1.65
(br, 1H), 1.29 (br, 24H), 0.93e0.89 (m, 6H). 13C NMR (125 MHz,
CDCl3,
30.20, 29.85, 29.56, 26.79, 22.91, 14.35. GCeMS (m/z): calcd for
20H36S, 308.56; found, 308.
d): 144.57, 126.75, 125.18, 123.12, 40.25, 34.45, 33.39, 32.12,
in acetonitrile with 0.1
M of tetrabutylammonium hexa-
fluorophosphate using a scan rate of 100 mV sꢀ1. ITO, Ag/AgCl and
Pt mesh were used as working electrode, reference electrode and
counter electrode, respectively.
C
2.3.2. Compound 3
To a solution of 2-(20-hexyldecyl)thiophene (2.7 g, 8.8 mmol) in
THF (15 mL) was added n-BuLi (4.4 mL, 2.5 M in hexane) at 0 ꢁC. The
mixture was stirred for 2 h at 0 ꢁC, and 1 h at 30 ꢁC. Then, benzo[1,2-
b:4,5-b0]dithiophene-4,8-dione (2) (0.7 g, 3.2 mmol) was added in
one portion at 30 ꢁC and the resulted mixture was stirred at 30 ꢁC
for 1.5 h. The mixture was further stirred for 1 h after adding
aqueous SnCl2 2H2O (6 g in 2 mL conc. HCl and 8 mL H2O). The
mixture was then poured into water and extracted with hexane.
The combined hexane layer was dried over Na2SO4, and after
2.2. Device fabrication and characterization
Polymer FETs were fabricated through the top-contact and
bottom-gate geometry. A thermally grown 300 nm thickness of
SiO2 was purchased from Montco Silicon Technologies INC and used
as the gate dielectric. The source/drain regions were defined by a
50 nm thickness of Au through a shadow mask, and the channel
length (L) and width (W) was 30 and 1000 mm, respectively. Before

230
J. Fan et al. / Polymer 82 (2016) 228e237
removing solvent, the crude product was purified by silica column
C65H90S4Sn2, 1130.3969; found, 1130.3948.
using hexane as eluent to give a yellow liquid (1.9 g, 76%). 1H NMR
(300 MHz, CDCl3,
d
): 7.66 (d, J ¼ 5.70 Hz, 2H), 7.48 (d, J ¼ 5.70 Hz,
2.3.7. Synthesis of PBDT86-TQ
2H), 7.33 (d, J ¼ 3.48 Hz, 2H), 6.91 (d, J ¼ 3.48 Hz, 2H), 2.89 (d,
Monomer 4 (310 mg, 0.27 mmol), 8 (112 mg, 0.25 mmol),
Pd2(dba)3 (9 mg) and P(o-tol)3 (23 mg) were charged into a 25-mL
flask fitted with a condenser under N2 protection. Dry toluene
(6 mL) was added into the flask. The resulting mixture was then
degassed twice and heated up to 120e130 ꢁC for 48 h. After cooling
to room temperature, the mixture was poured into acetone. The
precipitate was then dissolved into chloroform and passing from a
short column. Then, the concentrated chloroform solution was
poured into acetone and was further purified by Soxhlet extraction
with acetone and hexane for 12 h, respectively. The black solid was
collected and dried overnight under vacuum (250 mg, 91%). 1H
J ¼ 6.66 Hz, 4H), 1.75 (br, 2H), 1.38e1.32 (br, 48H), 0.92e0.90 (br,
12H). 13C NMR (125 MHz, CDCl3,
d): 145.95, 139.24, 137.44, 136.74,
127.91, 127.64, 125.60, 124.32, 123.64, 40.27, 34.91, 33.61, 32.16,
32.13, 31.82, 30.23, 29.87, 29.60, 26.88, 22.93, 14.36. HRMS (ESI):
calcd for C50H74S4, 802.4673; found, 802.4677.
2.3.3. Compound 4
To a solution of compound 3 (0.53 g, 0.66 mmol) in THF (10 mL)
was added n-BuLi (0.8 mL, 2.5 M in hexane) at ꢀ78 ꢁC, the mixture
was stirred for 2 h at ꢀ78 ꢁC. Then, the trimethyltin chloride
(2.4 mL, 1 M in hexane) was added in one portion. The cooler was
removed and the mixture was stirred at room temperature for
overnight. The resulted mixture was poured into water and
extracted with hexane. The combined hexane layer was dried over
Na2SO4. After removing solvent, the title compound was achieved
as a yellow liquid and used directly without further purification. 1H
NMR (300 MHz, CDCl3,
d) 8.80e6.80 (br, 12H), 3.20 (br, 4H),
1.80e0.65 (br, 62H). Mn: 17.3k, Mw: 26.2k, PDI ¼ 1.5.
2.3.8. Synthesis of PBDT88-TQ
PBDT88-TQ was synthesized with monomer
7 (250 mg,
0.22 mmol) and 8 (96 mg, 0.21 mmol) by following a similar pro-
NMR (300 MHz, CDCl3,
d) 7.70 (t, 2H), 7.34 (d, J ¼ 3.45 Hz, 2H), 6.92
cedure to that of PBDT86-TQ (217 mg, 93%). 1H NMR (500 MHz,
(d, J ¼ 3.48 Hz, 2H), 2.90 (d, J ¼ 6.63 Hz, 4H), 1.76 (br, 2H), 1.38e1.31
CDCl3, d) 8.80e7.00 (br, 10H), 2.90e2.50 (br, 8H), 1.90e0.60 (br,
(br, 48H), 0.89e0.91 (br, 12H), 0.41 (t, 18H). 13C NMR (125 MHz,
60H). Mn: 25.0k, Mw: 46.4k, PDI ¼ 1.9.
3. Results and discussion
3.1. Synthesis
CDCl3,
d): 145.57, 143.50, 142.43, 138.21, 137.52, 131.39, 127.75,
125.51, 122.64, 40.27, 34.93, 33.66, 32.15, 30.28, 29.93, 29.59, 26.95,
22.93, 14.37, ꢀ8.13. HRMS (ESI): calcd for C65H90S4Sn2, 1130.3969;
found, 1130.3926.
2.3.4. Compound 2,3-dioctylthiophene (5)
The synthetic routes of monomers are shown in Scheme 1.
Compounds 1 and 2 were prepared as previous reported [31,32].
Compound 3 was synthesized by the double nucleophilic addition
of 2-(20-hexyl)decylthiophene (1) into benzo[1,2-b:4,5-b0]dithio-
phene-4,8-dione (2) in the yield of 76%. Then, compound 3 was
lithiated by n-butyllithium followed by quenching with trime-
thyltin chloride to afford the distannyl compound 4 with 85% yield.
The distannyl compound 7 was synthesized starting from 2,3-
dioctylthiophene (5) in a similar manner. The main difference be-
tween 4 and 7 is the alkyl chain substituted position. Compound 4
has two branched hexyldecyl chains, and compound 7 has the four
straight octyl chains. Both long alkyl chains ensure good solubility
in the resulting polymers. The dithienobenzoquinoxaline (TQ)
acceptor 8 was prepared by following our previous report [33].
Scheme 2 shows the synthetic route of the polymers, PBDT86-
TQ and PBDT88-TQ. The polymers were synthesized via Stille
polycondensation from the distannyl compounds (4 or 7) and 8
using toluene as the solvent and tris(dibenzylideneacetone)dipal-
ladium (Pd2(dba)3) and tri(o-tolyl)phosphine (P(o-tol)3) as cata-
lysts. The polymers were purified by Soxhlet extraction with
acetone and hexane to remove small molecular and residual cata-
lyst. Both polymers could be dissolved in chlorobenzene (CB) and o-
dichlorobenzene (DCB), etc, but they have moderate to limited
solubility in common solvents such as chloroform and THF. It is
noted that PBDT86-TQ has better solubility in chlorinated solvents
than PBDT88-TQ. The gel permeation chromatography (GPC) was
used to measure the molecular weight using DCB as eluent solvent.
The number-average molecular weights (Mn) of PBDT86-TQ and
PBDT88-TQ were found to be 17.3 and 25.0 kDa, with a poly-
dispersity index (PDI) of 1.5 and 1.9, respectively.
To a solution of 2-bromo-3-octylthiophene (11 g, 40 mmol) and
Ni(ddpf)2Cl2 (150 mg) in THF (40 mL) at 0 ꢁC was added n-
C8H17MgBr (60 mmol in 30 mL THF, prepared from 1-bromooctane
and Mg). The resulting mixture was then heated to reflux for
overnight. After cooling to room temperature, the mixture was
poured into an aqueous HCl and extracted with hexane. The hexane
layer was dried over Na2SO4. After removing solvent, the crude
product was purification by distillation to give a colorless liquid.
(8.0 g, 65%) 1H NMR (300 MHz, CDCl3,
d
): 7.05 (d, J ¼ 5.13 Hz, 1H),
6.84 (d, J ¼ 5.13 Hz,1H), 2.74 (t, J1 ¼7.56 Hz, J2 ¼ 7.86 Hz, 2H), 2.53 (t,
J1 ¼ 7.53 Hz, J2 ¼ 7.86 Hz, 2H), 1.68e1.55 (m, 4H), 1.31 (br, 20H),
0.93e0.89 (m, 6H). 13C NMR (125 MHz, CDCl3,
d): 139.02, 137.91,
128.90, 121.09, 32.19, 32.12, 32.11, 31.10, 29.76, 29.72, 29.62, 29.60,
29.51, 29.47, 28.44, 27.99. GCeMS (m/z): calcd. for C20H36S, 308.56;
found, 308.
2.3.5. Compound 6
Compound 6 was synthesized from compound 2 and 5 by
following similar procedure of compound 3. 1H NMR (300 MHz,
CDCl3,
d
): 7.72 (d, J ¼ 5.70 Hz, 2H), 7.47 (d, J ¼ 5. 70 Hz, 2H), 7.24 (s,
2H), 2.85 (t, J1 ¼ 7.53 Hz, J2 ¼ 7.92 Hz, 4H), 2.64 (t, J1 ¼ 7.41 Hz,
J2 ¼ 7.77 Hz, 4H), 1.79e1.66 (m, 8H), 1.33 (br, 40H), 0.94e0.91 (m,
12H). 13C NMR (125 MHz, CDCl3,
d): 140.38, 139.04, 138.31, 136.55,
135.28, 130.06, 127.47, 124.33, 123.80, 32.14, 32.11, 32.07, 31.05,
29.75, 29.74, 29.69, 29.63, 29.56, 29.50, 28.53, 28.22, 22.92, 14.35.
HR-MS (ESI): calcd for C50H74S4, 802.4763; found, 802.4662.
2.3.6. Compound 7
Compound 7 was prepared as a yellow solid from compound 6
using similar procedure of compound 4. 1H NMR (300 MHz, CDCl3,
d
) 7.74 (t, 2H), 7.25 (s, 2H), 2.85 (t, J1 ¼7.56 Hz, J2 ¼ 7.89 Hz, 4H), 2.64
3.2. Optical properties
(t, J1 ¼ 7.35 Hz, J2 ¼ 7.74 Hz, 4H), 1.79e1.69 (m, 8H), 1.44e1.28 (br,
40H), 0.92e0.89 (m, 12H), 0.42 (t, 18H). 13C NMR (125 MHz, CDCl3,
The UVevis absorption spectra of PBDT86-TQ and PBDT88-TQ in
solution and film states were shown in Fig. 1. In chlorobezene so-
lution, PBDT86-TQ shows a broad absorption from 300 nm to
900 nm with two peaks at low energy bands and a few stronger
d): 143.27, 142.12, 140.01, 138.20, 137.36, 135.94, 131.62, 130.02,
122.68, 32.19, 32.13, 32.05, 31.06, 29.92, 29.83, 29.74, 29.67, 29.56,
29.53, 28.54, 28.25, 22.92, 14.37, ꢀ8.13. HRMS (ESI): calcd for

J. Fan et al. / Polymer 82 (2016) 228e237
231
Scheme 1. Synthetic routes of monomers.
Scheme 2. Synthesis of polymers PBDT86-TQ and PBDT88-TQ.

232
J. Fan et al. / Polymer 82 (2016) 228e237
behaviors between PBDT86-TQ and PBDT88-TQ, UVevis spectra of
both polymers in cholorbenzene solution at different temperature
were measured. Fig. 2 shows UVevis spectra of PBDT86-TQ and
PBDT88-TQ in chlorobenzene solution at room temperature (25 ꢁC),
60 ꢁC, 80 ꢁC and 90 ꢁC. As seen from Fig. 2a, the absorption shoulder
at ~790 nm for PBDT86-TQ decreased with the temperature
increased from 25 ꢁC to 90 ꢁC. This decrease is a typical behavior of
polymer packing peak due to the breakup of polymer pep stacking
(aggregation) upon heating. The 0-0 absorption peak for PBDT86-
TQ shows a very slight blue-shift with increasing temperature,
and there is almost no change on the short wavelength absorption
part. These results indicate that the intramolecular charge transfer
interactions between donor and acceptor for PBDT86-TQ are rela-
tively stronger and are little affected by external environment. It is
also possible that the polymer backbone of PBDT86-TQ is encap-
sulated by the long branched hexlydecyl chains, which will prevent
the free move of polymer chains during the heating, and therefore
there is a small effect on the intrinsic pep stacking of PBDT86-TQ
backbone at different temperature. However, the absorption peak
for PBDT88-TQ shows a large blue-shift with increasing tempera-
ture as shown in Fig. 2b. Moreover, the short wavelength absorp-
tion also increased while increasing temperature. This means that
PBDT88-TQ polymer chains at hot solution will be fully dissolved
and the chains can be freely rotated compared to that at cold so-
lution (i. e. at room temperature), and therefore leading to a weak
charge transfer state and short effective conjugation length [35].
It can see that PBDT86-TQ and PBDT88-TQ only show the dif-
ference on the side chains. PBDT86-TQ has two branched hex-
yldecyl side chain, and PBDT88-TQ has four octyl side chains.
Therefore, it is reasonable to believe that the side chain plays a
critical role on the polymer's molecular geometry besides
improving the solubility. The molecular geometry of the polymer
will be able to affect the electronic communications of donor and
acceptor, thus exhibiting a different charge transfer behaviors. In
literature, there are a few reports that focus on the side chain effect
for polymer solar cells, but most of them only show a slight dif-
ference on the optical properties of the resulting polymer [20e30].
To our best of knowledge, this is the first report to have such sig-
nificant effect of side chain on the resulting polymers' optical
properties. Considering the side chains length and substituted
position in PBDT86-TQ and PBDT88-TQ, it is reasonable to speculate
that the side chains will be able to induce a certain degree of steric
hindrance between BDT donor and TQ acceptor, which will affect
the polymer backbone stacking and the charge transfer. To prove
this speculation, molecular simulation was carried out for both
PBDT86-TQ and PBDT88-TQ using density functional theory (DFT)
at the B3LYP/6-31G (d) level [36]. As discussed above, the side
Fig. 1. The UVevis spectra of PBDT86-TQ and PBDT88-TQ in solution and solid states.
peaks at high energy bands. The high energy peaks between
300 nm and 500 nm come from the localized
low energy peaks are attributed to the intramolecular charge
transfer (ICT) or the stacking (aggregation) bands of polymer.
pep* transition. The
pep
The absorption of PBDT86-TQ in solution exhibited a 0-0 absorption
peak at ~740 nm and a well-resolved shoulder peak as 0e1 at
~660 nm [34]. In addition, there is shoulder at ~790 nm, which is
believed to be from the
pep stacking (aggregation) band of
PBDT86-TQ backbone. However, the absorption spectrum of
PBDT88-TQ in solution only shows a low energy absorption band
with a maximum peak at ~660 nm, which is attributed to the 0-
0 absorption band. It is amazing that the absorption of PBDT88-TQ
shows a large blue-shift compared to that of PBDT86-TQ. Consid-
ering the same polymer backbone in both polymers, the blue-shift
should come from the side chain effect. The UVevis spectra of
PBDT86-TQ and PBDT88-TQ in film states have similar features with
that in solution. The low energy peaks for PBDT86-TQ film have the
maximum at ~660 and ~740 nm, which are same with that in so-
lution. The film of PBDT88-TQ shows a maximum of ~660 nm at low
energy band, and there is no change compared to the solution state.
The onset of absorption in film states for PBDT86-TQ and PBDT88-
TQ are ~890 nm and ~740 nm, corresponding to an optical band-gap
of 1.39 eV and 1.62 eV, respectively.
To deep understand the reason for the different absorption
Fig. 2. The UVevis spectra changes of PBDT86-TQ and PBDT88-TQ in CB solution at different temperature.

J. Fan et al. / Polymer 82 (2016) 228e237
233
chains play an important role on the optical properties; here a full
length side chain at one side of each repeat unit for both polymers
was kept for more accurate calculation. Fig. 3 shows the optimized
geometries of PBDT86-TQ and PBDT88-TQ. The dihedral angles are
susceptible to the substitutes. As shown in Fig. 3, the dihedral angle
between BDT and TQ units is 27.7ꢁ for PBDT86-TQ, and this angle
increases to 31.2ꢁ in PBDT88-TQ. This means the backbone of
PBDT88-TQ are more twisted than PBDT86-TQ, in other words,
PBDT86-TQ backbone is more planar. As seen from the optimized
geometry (Fig. 3), this large dihedral angle in PBDT88-TQ is mainly
caused by the interactions between octyl side chains on BDT unit
and the quinoxaline part on TQ unit, the steric hindrance of octyl
side chain prohibits TQ unit to form a more planar backbone. In
addition, the dihedral angles bewteen BDT and the side thienyl unit
for both polymers also show a slight difference. The dihedral angles
are 51.5ꢁ and 49.4ꢁ for PBDT86-TQ and PBDT88-TQ, respectively
(Fig. 3). This indicates that the side chains significantly affect the
planarity of the polymer backbone. Compared PBDT86-TQ and
PBDT88-TQ, the smaller dihedral angle between BDT and TQ units
in PBDT86-TQ results a more planarity of polymer backbone,
therefore leading to a longer effective conjugation length. In
contrast, PBDT88-TQ backbone becomes more twisted and difficult
to stack because of the large dihedral angle and large steric hin-
drance volume. This can explain the observed large difference on
the optical properties of both polymers. This also provides a new
insight to understand the importance of side chain in the devel-
opment of conjugated polymers.
levels of PBDT86-TQ and PBDT88-TQ were calculated to be ꢀ5.10 eV
and ꢀ5.54 eV, respectively. It is unexpected that PBDT86-TQ and
PBDT88-TQ show a large difference on HOMO energy level. It is
well-known that the HOMO energy level is mainly determined by
the donor part, therefore the large difference is believed to be
affected by the alkyl chains of polymers that affect the effective
conjugation length and the packing behaviors. The HOMO energy
level of PBDT88-TQ is ~0.1e0.2 eV lower compared with other
PBDT-based polymers, which will benefit to have the higher open
circuit voltage in the photovoltaic device. The LUMO energy levels
of PBDT86-TQ and PBDT88-TQ were ꢀ3.51 eV and ꢀ3.67 eV,
respectively, which shows a similar trend with HOMO energy level.
These results demonstrate the significant impact and importance of
the alkly chain substitute position to tune energy levels of the
PBDT-based polymers. The band-gaps measured from CV for
PBDT86-TQ and PBDT88-TQ are 1.59 eV and 1.87 eV, respectively.
Fig. 5 shows the DFT calculation for the distributions of the
frontier molecular orbitals of PBDT86-TQ and PBDT88-TQ polymers.
For simplification, two repeat units (dimer) with the methyl side
chains were used to perform the DFT calculation. It is noted that the
calculated HOMO and LUMO energy levels of ꢀ4.90 eV
and ꢀ3.25 eV, respectively, are same for both polymers, which
values are largely different with that from CV measurements since
only the dimer and the short side chain were used in the DFT cal-
culations. As shown in Fig. 5, both polymers showed similar dis-
tribution of frontier molecular orbitals and the electronic wave-
function of the HOMO was distributed almost entirely over the
conjugated main chains. Whereas after light irradiation, the elec-
trons efficiently flowed to the acceptor and the electron density of
LUMO was mainly localized on the TQ segment in both polymers.
This kind of delocalization is similar with the PBDT polymers re-
ported previously. The similar quantum calculated results further
prove that the significant differences on optical and electro-
chemical properties between PBDT86-TQ and PBDT88-TQ polymers
are mainly attributed to the effect of side chains.
3.3. Electrochemical properties
The electrochemical properties of the polymers in the film states
were investigated by cyclic voltammetry (CV) with a Pt counter
electrode and a Ag/Agþ reference electrode in 0.1 M tetrabuty-
lammonium hexfluorophosphate in acetonitrile at a scan rate of
100 mV sꢀ1 [37]. Fig. 4 shows the cyclic voltammetric curves and
the energy levels of the polymers. The ferrocene was used as the
reference for the saturated calomel electrode (SCE). The highest
occupied molecular orbital (HOMO) and the lowest unoccupied
molecular orbital (LUMO) energy levels were calculated from the
onset oxidation potentials and the onset reduction potentials,
respectively, using equations HOMO ¼ ꢀ[Eox ꢀ Eferrocene þ 4.80] eV
and LUMO ¼ ꢀ[Ered ꢀ Eferrocene þ 4.80] eV [38]. The HOMO energy
3.4. Field-effect transistor properties
The charge carrier mobility of polymer is one of the important
parameters for polymer solar cells [39]. The higher mobility will be
beneficial to increase the current density and decrease the unfa-
vorable excition recombination. Normally a charge carrier mobility
Fig. 3. The optimized geometries of PBDT86-TQ and PBDT88-TQ monomer.

234
J. Fan et al. / Polymer 82 (2016) 228e237
Fig. 4. (a) The CV curves of PBDT86-TQ and PBDT88-TQ. (b) The schematic energy levels of the polymers, PCBM and cathodes (Ca, Al).
Fig. 5. The HOMO and LUMO wave functions of polymers PBDT86-TQ and PBDT88-TQ.
of higher than 10ꢀ4 cm2 Vꢀ1 Sꢀ1 is reasonable for high performance
polymer solar cells. The polymer field-effect transistors (FETs) were
fabricated with a top-contact and bottom-gate geometry. The gate
dielectric was SiO2 layer and the buffer layer was
divinyltetramethyldisiloxane-bis(benzocyclobutene) (BCB) thin
slight higher molecular weight of PBDT88-TQ may also contribute
to the better hole mobility compared to that of PBDT86-TQ. In
addition, both polymers also exhibited ambipolar property, the
measured electron mobility reached to 2 ꢂ 10ꢀ5 cm2 Vꢀ1 Sꢀ1 and
3.5 ꢂ 10ꢀ4 cm2 Vꢀ1 Sꢀ1 for PBDT86-TQ and PBDT88-TQ, respec-
tively. The higher mobility and ambipolar behavior from PBDT88-
TQ will be beneficial to an improved photovoltaic performance [40].
layer. The charge carrier mobility (
lated from the following equation:
m) of FET devices can be calcu-
3.5. Photovoltaic properties
1
2
W
L
IDS
¼
ꢂ
mCiðVGS ꢀ VTHÞ2
The photovoltaic properties of the polymers were evaluated
with the device configuration of ITO/PEDOT:PSS/polymer:PC71BM/
Ca/Al. Detailed fabrication and characterization are shown in
Experimental Section. It is known that the thermal and solvent
annealing treatments can improve the device performance effi-
ciently. Herein, we fabricated devices with the thermal annealing
and vapor annealing, respectively. The JeV curves of PBDT86-TQ/
PC71BM and PBDT88-TQ/PC71BM devices with thermal and vapor
annealing were shown in Fig. 7a. PBDT86-TQ/PC71BM device with
thermal annealing showed a Voc of 0.64 V, a short-circuit current
density (Jsc) of 2.85 mA cmꢀ2, and a fill factor (FF) of 0.44, resulting
an unexpected low PCE of 0.80%. Vapor annealing for PBDT86-TQ
W and L are the channel width and channel length, respectively.
Ci is the gate dielectric capacitance per unit. VGS is the gate voltage
and VTH is the threshold voltage.
Fig. 6 shows the output and transfer characteristics of FET de-
vices with PBDT86-TQ and PBDT88-TQ as the channel materials.
The hole mobilities of PBDT86-TQ and PBDT88-TQ extracted from
the saturated region are 1.5
ꢂ
10ꢀ5 cm2 Vꢀ1 Sꢀ1 and
1.0 ꢂ 10ꢀ3 cm2 Vꢀ1
S
ꢀ1, with an on/off ratio of 104 (Table 1). The
hole mobility of PBDT88-TQ shows a ~70 times higher than that of
PBDT86-TQ. The higher mobility may be from the more desirable
backbone stacking of PBDT88-TQ compared with PBDT86-TQ. The

J. Fan et al. / Polymer 82 (2016) 228e237
235
Fig. 6. Output at different gate voltage (Vgs) (a, PBDT86-TQ; c, PBDT88-TQ) and transfer characteristics (b, PBDT86-TQ; d, PBDT88-TQ) in the saturation regime under constant source
drain voltage (Vds ¼ ꢀ100 V) for FET devices.
Table 1
Table 2
The hole and electron mobility of PBDT86-TQ and PBDT88-TQ.
The device performances of PBDT86-TQ/PC71BM and PBDT88-TQ/PC71BM under
thermal and vapor annealing.
polymer
mh (cm2 Vꢀ1 sꢀ1
)
me (cm2 Vꢀ1 sꢀ1
)
On/Off
Vt (V)
Polymer/PC71BM
PBDT86-TQ
Annealing
Voc (V)
Jsc (mA cmꢀ2
)
FF
PCE (%)
PBDT86-TQ
PBDT88-TQ
1.5 ꢂ 10ꢀ5
~2 ꢂ 10ꢀ5
1.0 ꢂ 104
ꢀ5
1.0 ꢂ 10ꢀ3
~3.5 ꢂ 10ꢀ4
3.6 ꢂ 104
17
Thermal
Vapor
0.64
0.67
0.98
0.94
2.85
2.51
8.53
9.33
0.44
0.41
0.44
0.52
0.80
0.69
3.70
4.54
PBDT88-TQ
Thermal
Vapor
device gave an lower PCE of 0.60% with a Voc of 0.67 V, a Jsc of
2.51 mA cmꢀ2 and a FF of 0.41 (Table 2). However, the PCE of
PBDT88-TQ/PC71BM device with thermal annealing reaches up to
3.70% with a Voc of 0.98 V, a Jsc of 8.53 mA cmꢀ2 and a FF of 0.44
(Table 2). The Voc is one of the highest values for the single-junction
polymer solar cells. Further improved device performance via vapor
annealing for PBDT88-TQ/PC71BM device was achieved; the device
showed a Voc of 0.94 V, a Jsc of 9.33 mA cmꢀ2 and a FF of 0.52, giving
the PCE of 4.54% (Table 2), benefiting from the improved Jsc and FF.
Fig. 7b shows the external quantum efficiencies (EQEs) of these
devices. The PBDT86-TQ/PC71BM devices for both thermal and va-
por annealing show very low photoresponse from 350 nm to
800 nm, the EQE values are less than 25% in most wavelength
ranges, and the values decrease to less 10% between 650 nm and
800 nm. The main contribution to EQE is from PC71BM and the
Fig. 7. The JeV and EQE curves.

236
J. Fan et al. / Polymer 82 (2016) 228e237
Fig. 8. AFM figs. of PBDT86-TQ/PC71BM (a) and PBDT88-TQ/PC71BM (b).
short wavelength absorption of PBDT86-TQ. This result is consis-
tent with the measured low Jsc in PBDT86-TQ/PC71BM devices. The
EQE values of PBDT88-TQ/PC71BM devices are much higher than
that of PBDT86-TQ/PC71BM devices in the range of
350 nme750 nm. The EQE values for PBDT88-TQ/PC71BM devices
are more than 30% between 350 nm and 650 nm with highest
values of ~52% at ~400 nm. The EQE results are in line with the
higher Jsc for PBDT88-TQ/PC71BM devices and the much lower Jsc for
the PBDT86-TQ/PC71BM devices, which also indicate the accuracy of
the measurements.
major role in inducing the stronger bimolecular recombination in
devices. Combined with the low charge carrier mobility of PBDT86-
TQ, a lower Jsc in the photovoltaic device is expected. For PBDT88-
TQ/PC71BM film, a favorable morphology could be observed, and
therefore resulted in a higher Jsc. The large phase separation in
PBDT86-TQ blend film may be due to the worse miscibility of
polymer and PC71BM, suggesting that the proper introduction of
the side chain is very important to achieve the better miscibility of
the polymer and fullerene acceptor.
The Voc of the PSC is mostly dependent on the difference be-
tween HOMO energy level of the donor polymer and the LUMO
energy level of PCBM. Therefore, the measured high Voc of
0.94e0.98 V for PBDT88-TQ devices are attributed to the lower
lying HOMO energy level of PBDT88-TQ (ꢀ5.54 eV), and the lower
Voc for the PBDT86-TQ devices is due to the high lying HOMO en-
ergy level (ꢀ5.10 eV). The theoretical Voc can be roughly calculated
4. Conclusion
In conclusion, we have reported the copolymers of BDT and TQ,
PBDT86-TQ and PBDT88-TQ, with different alkyl side chains and
substituted positions. The results showed that the side chains of
both polymers have significant effect on the optical, electro-
chemical, FET and photovoltaic properties. The absorption peak of
PBDT86-TQ is on ~740 nm and it was blue-shifted to ~660 nm
because of the strong steric hindrances from the dioctyl chains
substitutes. The band-gaps for PBDT86-TQ and PBDT88-TQ are
1.39 eV and 1.62 eV, respectively. The hole mobility of PBDT88-TQ is
~70 times higher than that of PBDT86-TQ, which may come from
the more desirable polymer chain orientations and/or the inter-/
intra-molecular charge transport compared with the polymer
PBDT86-TQ. The photovoltaic devices of PBDT86-TQ/PC71BM gave a
very low PCE of 0.69e0.80%, and that was increased to 3.70e4.50%
for PBDT88-TQ/PC71BM-based photovoltaic devices with higher Voc
of up to 0.98 V and higher current density as a result of lower
HOMO energy level, high charge carrier mobility and better
miscibility with PCBM. These results indicate that selecting the
proper side chains on the design of conjugated polymers is highly
of importance to achieve high performance PSCs. This work also
provides a new insight to understand the relationship between side
chains of conjugated polymer and the performance.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
using the equation of Voc ¼ 1=eð Edonor ꢀ EPCBM Þ ꢀ 0:3V regard-
ꢀ
ꢀ
ꢀ
ꢀ
HOMO
LUMO
less of other reasons that affect the Voc [41]. The theoretical Voc for
PBDT86-TQ- and PBDT88-TQ-based devices are 0.94 V and 0.5 V,
which are similar with measured values, since other factors such as
the cathode, interface resistance and excition nonradiative
recombination, etc also play a role in determining the Voc of the
resulting device.
The recombination of photogenerated charge carriers in the
active layer of PSCs is the main loss of the short circuit current. The
recombination rate is relevant to the charge carrier mobility and
the morphology of the BHJ blend. The higher charge carrier mo-
bilities of PBDT88-TQ should have a large contribution to the higher
Jsc for PBDT88-TQ/PC71BM devices. Similar, one of the reasons for
the low Jsc in PBDT86-TQ/PC71BM devices is due to the low charge
carrier mobility as shown and discussed above. In addition, the
proper and careful control of the nanoscale morphology of the
polymer/PCBM blend is also a major way to improve the Jsc. It is not
favorable for the efficient charge separation if the domains are too
large or too small, therefore will lead to the severe charge recom-
bination. Fig. 8 shows the AFM images of the thin film blends of
PBDT86-TQ and PBDT88-TQ with PC71BM. As shown in Fig. 8, the
PBDT86/PC71BM blend film showed a rougher surface and a large
mean square surface roughness (Rq) compared to that of PBDT88/
PC71BM blend film. There are a lot of large sized pin-holes existed in
the PBDT86-TQ/PC71BM films, and it was mainly caused by
immiscibility of polymer and PCBM. These pin-holes may play a
Acknowledgments
The authors thank the support of the National Natural Science
Foundation of China (51403044, 51173199, U1204212 and
21404031), the Fundamental Research Funds for the Central Uni-
versities (Harbin Institute of Technology), and the Open Fund of
State Key Laboratory of Luminescent Materials and Devices (South
China University of Technology, 2015-skllmd-12). The authors

J. Fan et al. / Polymer 82 (2016) 228e237
(2014) 18988e18997.
237
thank Dr. Jingyu Zou, Dr. Chu-Chen Cheuh and Prof. Alex Jen from
University of Washington for the help on the device
characterizations.
[21] P.P. Khlyabich, B. Burkhart, A.E. Rudenko, B.C. Thompson, Polymer 54 (2013)
5267e5298.
[22] J. Lee, M. Kim, B. Kang, S.B. Jo, H.G. Kim, J. Shin, et al., Adv. Energy Mater. 4
(2014) 1400087.
[23] S. Liu, X.C. Bao, W. Li, K.L. Wu, G.H. Xie, R.Q. Yang, et al., Macromolecules 48
(2015) 2948e2957.
References
[24] L.J. Huo, L. Ye, Y. Wu, Z.J. Li, X. Guo, M.J. Zhang, et al., Macromolecules 45
(2012) 6923e6929.
[25] S.Q. Zhang, L. Ye, W.C. Zhao, D.L. Liu, H.F. Yao, J.H. Hou, Macromolecules 47
(2014) 4653e4659.
[1] G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270 (1995)
1789e1791.
[2] L. Ye, S.Q. Zhang, L.J. Huo, M.J. Zhang, J.H. Hou, Acc. Chem. Res. 47 (2014)
1595e1603.
[3] G. Li, R. Zhu, Y. Yang, Nat. Photonics 6 (2012) 153e161.
[4] P.M. Beaujuge, J.M.J. Frechet, J. Am. Chem. Soc. 133 (2011) 20009e20029.
[5] Y.J. Cheng, S.H. Yang, C.S. Hsu, Chem. Rev. 109 (2009) 5868e5923.
[6] G. Li, V. Shrotriya, J.S. Huang, Y. Yao, T. Moriarty, K. Emery, et al., Nat. Mater. 4
(2005) 864e868.
[7] Z.C. He, B. Xiao, F. Liu, H.B. Wu, Y.L. Yang, S. Xiao, et al., Nat. Photonics 9 (2015)
174e179.
[8] Z.C. He, C.M. Zhong, S.J. Su, M. Xu, H.B. Wu, Y. Cao, Nat. Photonics 6 (2012)
591e595.
[9] L.J. Huo, T. Liu, X.B. Sun, Y.H. Cai, A.J. Heeger, Y.M. Sun, Adv. Mater. 27 (2015)
2938e2944.
[10] S.J. Liu, K. Zhang, J.M. Lu, J. Zhang, H.L. Yip, F. Huang, et al., J. Am. Chem. Soc.
135 (2013) 15326e15329.
[11] S. Zheng, S.Q. Zhang, M.J. Zhang, K. Zhao, L. Ye, Y. Chen, et al., Adv. Mater. 27
(2015) 1189e1194.
[12] X. Guo, M.J. Zhang, W. Ma, L. Ye, S.Q. Zhang, S.J. Liu, et al., Adv. Mater. 26
(2014) 4043e4049.
[13] M.J. Zhang, X. Guo, W. Ma, H. Ade, J.H. Hou, Adv. Mater. 27 (2015) 4655e4660.
[14] L.J. Huo, S.Q. Zhang, H.Y. Chen, Y. Yang, Angew. Chem. Int. Ed. 49 (2010)
1500e1503.
[15] M. Wang, X.W. Hu, P. Liu, W. Li, X. Gong, F. Huang, et al., J. Am. Chem. Soc. 133
(2011) 9638e9641.
[16] Y.Y. Liang, D.Q. Feng, Y. Wu, S.T. Tsai, G. Li, C. Ray, et al., J. Am. Chem. Soc. 131
(2009) 7792e7799.
[17] L.T. Dou, J. Gao, E. Richard, J. You, C.C. Chen, K.C. Cha, et al., J. Am. Chem. Soc.
134 (2012) 10071e10079.
[18] Y. Zhang, S.K. Hau, H.L. Yip, Y. Sun, O. Acton, A.K.Y. Jen, Chem. Mater. 22 (2010)
2696e2698.
[26] Y. Dong, X.W. Hu, C.H. Duan, P. Liu, S.J. Liu, L.Y. Lan, D.C. Chen, L. Ying, S.J. Su,
X. Gong, F. Huang, Y. Cao, Adv. Mater. 25 (2013) 3683e3688.
[27] J.H. Dou, Y.Q. Zheng, T. Lei, S.D. Zhang, Z. Wang, W.B. Zhang, J.Y. Wang, J. Pei,
Adv. Funct. Mater. 24 (2014) 6270e6278.
[28] T. Lei, J.H. Dou, J. Pei, Adv. Mater. 24 (2012) 6457e6461.
[29] T. Lei, J.Y. Wang, J. Pei, Chem. Mater. 26 (2014) 594e603.
[30] H.F. Yao, H. Zhang, L. Ye, W.C. Zhao, S.Q. Zhang, J.H. Hou, ACS Appl. Mater.
Interfaces (2015), http://dx.doi.org/10.1021/acsami.5b07311.
[31] Z.Q. Wu, C.G. Qi, N. Liu, Y. Wang, J. Yin, Y.Y. Zhu, L.Z. Qiu, H.B. Lu, J. Polym. Sci.
Poly. Chem. 51 (2013) 2939e2947.
[32] J.H. Hou, M.H. Park, S.Q. Zhang, Y. Yao, L.M. Chen, J.H. Li, et al., Macromolecules
41 (2008) 6012e6018.
[33] Y. Zhang, J.Y. Zou, H.L. Yip, K.S. Chen, J.A. Davies, Y. Sun, et al., Macromolecules
44 (2011) 4752e4758.
[34] Z.H. Chen, P. Cai, J.W. Chen, X.C. Liu, L.J. Zhang, L.F. Lan, et al., Adv. Mater. 26
(2014) 2586e2591.
[35] Y. Zhang, J.Y. Zou, C.C. Cheuh, H.L. Yip, A.K.Y. Jen, Macromolecules 45 (2012)
5427e5435.
[36] T.M. McCormick, C.R. Bridges, E.I. Carrera, P.M. DiCarmine, G.L. Gibson,
J. Hollinger, et al., Macromolecules 46 (2013) 3879e3886.
[37] Y. Zhang, J.Y. Zou, H.L. Yip, K.S. Chen, D.F. Zeigler, Y. Sun, et al., Chem. Mater.
23 (2011) 2289e2291.
[38] J. Pommerehne, H. Vestweber, W. Guss, R.F. Mahrt, H. Bassler, M. Porsch, et al.,
Adv. Mater. 7 (1995) 551e554.
[39] R.P. Qin, W.W. Li, C.H. Li, C. Du, C. Veit, H.F. Schleiermacher, et al., J. Am. Chem.
Soc. 131 (2009) 14612e14613.
[40] Y. Zhang, J.Y. Zou, C.C. Cheuh, H.L. Yip, A.K.Y. Jen, Macromolecules 45 (2012)
5427e5435.
[41] M.C. Scharber, D. Wuhlbacher, M. Koppe, P. Denk, C. Waldauf, A.J. Heeger, et
al., Adv. Mater. 18 (2006) 789e794.
[19] S.C. Price, A.C. Stuart, L.Q. Yang, H.X. Zhou, W. You, J. Am. Chem. Soc. 133
(2011) 4625e4631.
€
[20] X.F. Xu, Z.J. Li, O. Backe, K. Bini, D.I. James, E. Olsson, et al., J. Mater. Chem. A 2