Systematic investigation of isoindigo-based polymeric field-effect transistors: Design strategy and impact of polymer symmetry and backbone curvature

Article abstract of DOI:10.1021/cm300117x

Ten isoindigo-based polymers were synthesized, and their photophysical and electrochemical properties and device performances were systematically investigated. The HOMO levels of the polymers were tuned by introducing different donor units, yet all polymers exhibited p-type semiconducting properties. The hole mobilities of these polymers with centrosymmetric donor units exceeded 0.3 cm² V^-1 s^-1, and the maximum reached 1.06 cm² V^-1 s^-1. Because of their low-lying HOMO levels, these copolymers also showed good stability upon moisture. AFM and GIXD analyses revealed that polymers with different symmetry and backbone curvature were distinct in lamellar packing and crystallinity. DFT calculations were employed to help us propose the possible packing model. Based on these results, we propose a design strategy, called molecular docking , to understand the interpolymer π-π stacking. We also found that polymer symmetry and backbone curvature affect interchain molecular docking of isoindigo-based polymers in film, ultimately leading to different device performance. Finally, our design strategy maybe applicable to other reported systems, thus representing a new concept to design conjugated polymers for field-effect transistors.

Full text of DOI:10.1021/cm300117x

Article
pubs.acs.org/cm
Systematic Investigation of Isoindigo-Based Polymeric Field-Effect
Transistors: Design Strategy and Impact of Polymer Symmetry and
Backbone Curvature
Ting Lei, Yue Cao, Xu Zhou, Yang Peng, Jiang Bian, and Jian Pei*
Beijing National Laboratory for Molecular Sciences, the Key Laboratory of Bioorganic Chemistry and Molecular Engineering of
Ministry of Education, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
*
S Supporting Information
ABSTRACT: Ten isoindigo-based polymers were synthesized, and their photophysical and
electrochemical properties and device performances were systematically investigated. The
HOMO levels of the polymers were tuned by introducing different donor units, yet all
polymers exhibited p-type semiconducting properties. The hole mobilities of these polymers
2
−1 −1
with centrosymmetric donor units exceeded 0.3 cm V s , and the maximum reached 1.06
2
−1 −1
cm V s . Because of their low-lying HOMO levels, these copolymers also showed good
stability upon moisture. AFM and GIXD analyses revealed that polymers with different
symmetry and backbone curvature were distinct in lamellar packing and crystallinity. DFT
calculations were employed to help us propose the possible packing model. Based on these
results, we propose a design strategy, called “molecular docking”, to understand the
interpolymer π−π stacking. We also found that polymer symmetry and backbone curvature
affect interchain “molecular docking” of isoindigo-based polymers in film, ultimately leading to different device performance.
Finally, our design strategy maybe applicable to other reported systems, thus representing a new concept to design conjugated
polymers for field-effect transistors.
KEYWORDS: polymeric field-effect transistors, structure−property relationship, conjugated polymers, organic electronics
1
0
INTRODUCTION
more constituent molecules. Molecular docking has been
broadly used to understand the biological processes determined
by noncovalent interactions, especially for rational design of
■
Organic field-effect transistors (OFETs) are advancing rapidly
in terms of their applications in low-cost, large-area thin film
11
1
drugs. Electron-deficient unit, isoindigo, with branched alkyl
chains as acceptor, and ten different electron-rich units as
donor are combined to prepare ten isoindigo-based copoly-
mers. Because of the donor−acceptor interaction and the
spatial steric hindrance caused by the branched alkyl chains, we
envision that the small units dock into the cavity formed by the
isoindigo cores and branched alkyl chains (Figure 1c). We
systematically compare FET device performances, polymer
packings and film morphologies of the polymers, and find that
this strategy is efficient to obtain high-performance FETs.
Polymers with centrosymmetric donors exhibit systematically
higher field-effect mobility than those with axisymmetric ones.
By means of atomic force microscopy (AFM) and grazing-
incident X-ray diffraction (GIXD), we attribute the great
difference in device performance to the different symmetry and
backbone curvature of the polymers, which greatly affect the
interpolymer docking and lamellar packing. Density functional
theory (DFT) calculations support our proposed packing
model. Finally, we compare the device performances of
diketopyrrolopyrole (DPP) and naphthalenedicarboximide
transistors. Compared to small molecules, polymers offer great
advantages, such as solution processability, good mechanical
property, and thermal stability. Benefiting from new molecular
design and device fabrication improvement, polymeric field-
effect transistors (PFETs) have made significant progress in
past decade and some of them exhibit carrier mobility
2
2
−1 −1
3
approaching or surpassing 1 cm V s . However, upon
prolonged exposure to air or moisture, device performance
degradation is also a critical problem for practical application of
4
3a,c,5c,d
PFETs. With notably rare exceptions,
PFETs with high
carrier mobility and long-time stability are usually achieved in
low-humidity and inert atmosphere. Therefore, to design new
4
polymers for high-mobility PFETs with ambient stability is still
5
of great challenge. On the other hand, the structure−property
relationship is important for designing new organic materials
for optoelectronics. Although several design strategies and even
computational methods have been proposed for small
6
−9
molecular OFETs,
rational design strategy for PFETs are
seldom reported.
Herein, employing the concept of molecular docking, we
design and synthesize ten copolymers for PFETs to investigate
their structure−property relationship (Figure 1). In the
molecular modeling field, docking is a method to predict the
structure of intermolecular complex formed between two or
Received: January 11, 2012
Revised: April 26, 2012
Published: May 2, 2012
©
2012 American Chemical Society
1762
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
Figure 1. (a) Structures of isoindigo-based copolymers; (b) centrosymmetric and axisymmetric donor of the copolymers; (c) proposed interpolymer
docking model.
a
Scheme 1. Synthesis of Monomers and Isoindigo-Based Copolymers.
a
Reagent and conditions: (a) AcOH/HCl, reflux, 24 h, 86%; (b) 1-iodo-2-octyldodecane, K CO , DMF, 74%; (c) Pd (dba) , P(o-tol) , toluene, 110
2
3
2
3
3
°C, 48 h.
(
NDI) based polymers reported by other groups, and we
in order to obtain p-type isoindigo-based polymers, electron-
rich groups should be introduced to increase the HOMO level
of the desired polymers. The donor−acceptor interactions of
electron-deficient isoindigo skeleton and electron-rich units
conclude that the molecular docking strategy is also applicable
to those systems, and thus representing a new concept to
design conjugated polymers for PFETs.
1
4,15
enhance the interchain π−π stacking.
Selenophene- (4) and
RESULTS AND DISCUSSION
thiophene-containing (5−11) units were employed to con-
struct these copolymers (Scheme 1). Units 4−7 and 10 are
centrosymmetric, and units 8, 9, and 11 are axisymmetric. To
prove our molecular docking strategy, we also synthesized 3,3′-
■
Design and Synthesis. Electron-deficient isoindigo unit is
12
a new building block for low bandgap polymeric solar cells.1
3
As isoindigo homopolymer is a typical n-type semiconductor,
1
763
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
dimethyl-2,2′-bithiophene (10) to compare with 2,2′-bithio-
phene. We envision that the methyl groups may hinder π−π
b′]dithiophene (NDT), IID-BDT and IID-NDT showed
obvious blue-shift of the absorption spectra compared with
other polymers. Computational results show that the HOMOs
are well delocalized along the polymer chain, but the LUMOs
are mostly localized on the isoindigo core (Figure 3a).
3b
16 17 18
stacking and interchain docking. Compounds 4, 5, 6, 7,
1
9
20
8
,
and 9 were synthesized according to literatures.
Monomer IID were obtained from commercially available 6-
bromoisatin (1) and 6-bromooxindole (2) only two steps with
high yield. All monomers were prepared in 10-g scale. The
copolymers were obtained through Stille-coupling polymer-
ization using Pd (dba) and P(o-tol) as catalysts. After the
2
3
3
polymerization, strongly complexing ligand N,N-diethylpheny-
lazothioformamide was added to remove any residual catalyst.
All polymers were purified by Soxhlet extraction (sequentially
by methanol, hexane, and CHCl ), and afforded dark or dark
3
blue solids. After introducing the branched alkyl chains, all
polymers (except IID-T3) showed good solubility in common
solvents, such as CHCl , toluene, and THF. IID-T3 dissolved
3
in chloro-containing solvents such as CHCl , trichloroethylene
3
(
TCE), and dichlorobenzene (DCB).
Molecular weights of all polymers were evaluated by gel
permeation chromatography (GPC) using 1,2,4-tricholoroben-
zene (TCB) as eluent at 150 °C. Large PDI remained at 150
°
C, presumably because of strong aggregation of the polymers
by strong interchain interactions. All polymers showed good
thermal stability with decomposition temperatures over 350 °C.
No phase transition was observed by differential scanning
calorimetry (DSC) before decomposition.
Figure 3. (a) Calculated molecular orbitals of the IID-BDT trimer;
(b) HOMO and LUMO energy levels are calculated for oligomers (n
=
1, 2, and 3) of IID-BDT, IID-NDT, IID-TT, and IID-T3. The
energy levels of polymers can be extrapolated from the values. Alkyl
chains are replaced with methyl groups for computational simplicity,
and all the calculations are performed at B3LYP/6-31G (d) levels.
Photophysical and Electrochemical Properties. The
absorption spectra of all polymers in dilute solution, in thin film
and in annealed film are shown in Figure 2 and Figure S1 in the
The absorption peaks of all polymers (except for polymer
IIDDT-Me) in film showed no obvious red-shift in comparison
with those in solution. This suggests that the polymers adopted
similar geometry both in solution and in film. Scrutiny of
spectra reveals that the 0−0 vibrational transition increased,
whereas 0−1 decreased in film, suggesting that polymers
become more planar or form some J aggregates in film. After
annealing the film at 150 °C for 30 min, a further increase of
the 0−0 transition was observed, indicating the packing and
planarity of polymers were improved. Two methyl groups of
IIDDT-Me resulted in a great torsion angle, which broke the
conjugation, thus leading to an obvious blue-shift in solution.
This torsion angle was suppressed in solid state because of
interchain π−π stacking, hence red-shifting the absorption
maximum significantly (over 100 nm) to a similar region as
other polymers (Table 1).
The cyclic voltammetries (CV) of all polymers in thin films
were measured to evaluate their electronic energy levels (Figure
4
and Figure S2 in the Supporting Information). Most
Figure 2. Normalized UV−vis absorption spectra of (a) IID-Se, (b)
polymers showed much stronger oxidative peaks than reductive
ones, almost 1 order of magnitude higher, which indicated that
these polymers were more easily oxidized. All polymers showed
similar reductive potentials with LUMO levels around −3.70
eV. HOMO levels of polymers vary with electron-donating
properties of donor units. For example, the HOMO/LUMO
level of IID-Se is −5.65/−3.79 eV, and that of IID-TT is
−5.70/−3.73 eV. Polymers containing less weak donors had
lowered HOMO levels (−5.84 eV for IID-BDT and −5.90 eV
for IID-NDT). These results were consistent with the
calculation that LUMO levels of all polymers were mostly
localized on the isoindigo core and HOMO levels were
distributed along the polymer chains (Figure 3a). Thus HOMO
levels of all polymers were more easily affected by the electronic
structure of the donor unit, whereas the LUMO levels remained
−5
IID-NDT, (c) IIDDT-Me, and (d) IID-T3 in CHCl (1 × 10 M),
3
in thin film (spin-cast from TCE solution, 1 mg/mL), and in annealed
film (150 °C for 30 min).
Supporting Information. All polymers show typically dual band
absorption both in solution and in film: band I (500−800 nm)
and band II (300−500 nm). Varying with different polymers,
band II was attributed to the absorption of the donor part. In
contrast, these polymers had similar features in band I region
with three obvious vibrational peaks 0−0, 0−1, and 0−2, thus
associating the shape-persistent isoindigo core. In addition,
band I is a typical charge-transfer absorption from the donor
part to the isoindigo core. After introducing relatively weak
donors, benzodithiophene (BDT) and naphtho[1,2-b:5,6-
1
764
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
Table 1. Optical and Electrochemical Properties of Polymers
M (kDa)
λmax0 sol
−0
λmax0−1 sol
λmax0−0 film
λmax0−1 film
Egopt
E
E
LUMO
e
Egcv
(eV)
n
HOMO
(eV)
a
b
b
c
c
d
e
f
polymers
/PDI
(nm)
(nm)
(nm)
(nm)
(eV)
(eV)
IIDDT
33.7/5.4
26.7/3.5
27.3/4.7
24.8/3.5
25.5/3.5
20.1/3.9
19.6/3.1
38.8/3.4
26.6/3.7
11.2/2.5
706
716
674
666
723
604
691
N.A.
719
658
647
655
646
616
666
N.A.
644
628
662
613
701
715
676
662
720
725
697
682
720
668
637
649
620
610
656
670
645
628
656
615
1.59
1.56
1.66
1.69
1.55
1.54
1.58
1.58
1.55
−5.65
−5.56
−5.84
−5.90
−5.70
−5.54
−5.80
−5.48
−5.44
−5.55
−3.78
−3.79
−3.77
−3.75
−3.73
−3.78
−3.81
−3.70
−3.72
−3.70
1.87
1.77
2.07
2.15
1.97
1.76
1.99
1.78
1.73
1.85
IID-Se
IID-BDT
IID-NDT
IID-TT
IIDDT-Me
IIDT
g
g
IID-T3
IID-TTT
IID-TBT
1.65
−5
a
b
c
GPC versus polystyrene standards, TCB as eluent, at 150 °C. Solution absorption spectra (1 × 10 M in chloroform). Thin film absorption
d
e
spectra from pristine film spin-cast from 5 mg/mL TCE solution. Optical energy gap estimated from the onset of the film absorption. cyclic
+
f
CV
g
voltammetries were determined with Fc/Fc (E
weak.
= −4.80 eV) as external reference. E = ELUMO − EHOMO. The first reduction peak was
HOMO
g
TCB a maximum mobility up to 1.06 cm V s−1 and average
2
−1
2
−1 −1
mobility of 0.66 cm V
s
were obtained (Figure 5a). This
result is among the highest values reported to date. Use of TCB
as solvent to improve PFETs device performance was also
reported in the fabrication of P3HT based device, because TCB
with high boiling point provided P3HT with better interchain
2
1,22
packing and more condensed uniform film.
Because the
OTS treated SiO surface is very hydrophobic and TCB has a
2
relatively higher contact angle than TCE, spin-coating on the
surface was only achieved for polymers IIDDT, IID-Se, or IID-
NDT in TCB, presumably because of their higher molecular
weight and viscosity. For other polymers, spin-coating from
their TCB solutions did not provide continuous films. We also
found that solvent-annealing in TCE atmosphere in Petri dish
effectively improved the device performance after spin-coating
from TCE solutions. Thus solvent-annealing with TCE in Petri
dish was also performed for polymers after spin-coating from
TCE. To our delight, all polymers with centrosymmetric
donors showed excellent hole-transporting properties with
Figure 4. Cyclic voltammogram of (a) IID-Se; (b) IID-NDT; (c)
IIDDT-Me; and (d) IID-T3 in thin film drop-casting on a glassy
carbon electrode and tested in n-Bu NPF /CH CN solution (scan
4
6
3
−1
2
−1 −1
rate: 50 mV s ).
mobilities over 0.1 cm V s . (Figure 5, Table 2). Such
systematically high mobilities rarely appeared in literature.
Although selenophene-containing polymers were reported to
show a better carrier-transporting performance in OFETs than
unchanged. Interestingly, IID-TTT showed two oxidative peaks
and a HOMO level of −5.44 eV, indicating the strong electron-
donating property of dithieno[3,2-b:2′,3′-d]thiophene (see the
Supporting Information). Using DFT calculation, we obtained
theoretical HOMO/LUMO levels of polymers by extrapolating
the calculated energy levels of oligomers (n = 1−3) (Figure
3b
that of the thiophene-containing polymers, our selenophene-
containing polymer IID-Se only showed a comparable result
with that of the thiophene-containing polymer IIDDT (with
2
−1 −1
the maximum hole mobility up to 0.66 cm V
s
and average
2
−1 −1
mobility of 0.46 cm V s ). IIDDT-Me showed the worst
3b). Although DFT calculations usually provide higher values
2
−1
performance (with the maximum hole mobility of 0.11 cm V
than electrochemical results, both methods gave similar trends
of energy-level variations. As shown in Figure 3b, the HOMO
level of IID-T3 was much higher than those of other polymers
and that of IID-TT was relatively higher than those of IID-
BDT and IID-NDT. However, their LUMO levels showed no
obvious change. All photophysical and electrochemical data are
summarized in Table 1.
Field-Effect Transistor Fabrication and Character-
ization. Bottom-gate/top-contact (BG/TC) devices were
fabricated by spin-coating the polymer solutions (4 mg/mL)
−1
2
−1 −1
s
and average mobility of 0.091 cm V s ), about 1 order of
magnitude lower than that of IIDDT. Methyl groups hinder the
interchain π−π stacking and thereby reduce the mobility. In
sharp contrast, all polymers with axisymmetric donors showed
much lower hole mobilities (see Figure S3 in the Supporting
2
Information). IIDT had a maximum hole-mobility of 0.019 cm
−1
−1
V
s , and IID-T3 had a maximum hole-mobility of 0.061
cm V s . IID-TTT and IID-TBT showed poor hole-
2
−1 −1
−3
−4
2
−1 −1
mobility in the range of 1 × 10 to 1 × 10 cm V s .
Oxidative potentials of polymers correlate the threshold
voltages of their devices. IID-Se and IID-TT have higher
HOMO levels than IIDDT, and their threshold voltage were
more positive (−10 V for IID-Se and −6 V for IID-TT).
Similarly, IID-BDT and IID-NDT have lower HOMO levels
and their threshold voltages were more negative (−28 V for
onto octadecyltrimethoxysilane (OTS) treated SiO (300 nm)
2
on heavily doped Si substrate. Previously, we used TCE as
solvent to fabricate thin film transistors. After optimizing
conditions, we found that using TCB provided better device
performance. For example, with TCE as solvent, IIDDT only
2
−1 −1
gave an average hole-mobility of 0.42 cm V s , whereas with
1
765
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
Figure 5. Transfer and output characteristics of (a, b) IIDDT, (c, d) IID-Se, (e, f) IID-BDT, (g, h) IID-NDT, (i, j) IID-TT, and (k, l) IIDDT-Me
devices (spin-casted from TCB or TCE solutions, 4 mg/mL) at VSD = −50 V (L = 60 μm, W = 3.0 mm) after thermal annealing.
V− s−1 for IID-BDT and 0.32 cm V s−1 for IID-NDT).
Interestingly these two polymers show very small hysteresis in
transfer characteristics, indicating less trapping in polymer
1
2
−1
Table 2. Bottom-Gate/Top-Contact (BG/TC) OFET Device
Performances for Isoindigo-Based Polymers Measured
under Ambient Condition (R = 50−60%)
H
5
films.
2
−1
a
polymers
Tannealing (°C) μ (cm V s−1
)
Vt (V)
I /I
on off
Note that IIDDT exhibit good stability for six months under
ambient condition (see Figure S4 in the Supporting
Information). After the storage, only a slight increase of the
off-current was observed. Other centrosymmetric polymers
b
6
7
6
7
7
6
6
IIDDT
150
160
150
200
150
150
150
150
200
150
1.06 (0.66)
0.66 (0.46)
0.48 (0.37)
0.32 (0.25)
0.34 (0.31)
0.11 (0.091)
−18
−10
−28
−30
−6
−5
−20
−4
10 −10
b
5
IID-Se
10 −10
c
6
IID-BDT
10 −10
b
6
IID-NDT
10 −10
(
IID-Se, IID-TT, and IIDDT-Me) were stable under this
c
5
IID-TT
10 −10
condition for 2 months (see Figure S4 in the Supporting
Information). All isoindigo polymers show good ambient
stability, which may be attributed to their low-lying HOMO
c
5
IIDDT-Me
10 −10
c
5
IIDT
0.019 (0.015)
0.061 (0.048)
10
c
5
IID-T3
10
5
levels.
c
−3
4
IID-TTT
1.03 × 10
−4
−16
10
Film Morphology and Molecular Packing. Tapping-
mode atomic force microscopy (TM-AFM) was employed to
investigate morphologies of polymer films. Figure 6 and Figure
S5 in the Supporting Information show height images of
polymer films prepared under the same condition as their
device fabrications. Root-mean-square (rms) analyses of height
images are also employed to reflect the roughness of films.
Thermal annealing at 150 °C for 30 min resulted in negligible
change in film morphologies. All polymers with centrosym-
metric units, even IIDDT-Me, showed obviously crystallized
zones or fibrillar intercalating networks in films. These
networks were likely due to good intermolecular packing as
c
−4
3
IID-TBT
1.35 × 10
10
a
Maximum values of hole mobility and average mobilities are shown in
parentheses (more than 10 devices were tested for each polymer).
b
Devices were fabricated using TCB as solvent without solvent
c
annealing. Devices were fabricated using TCE as solvent, and
annealed in TCE solvent atmosphere for 2 h.
IID-BDT and −30 V for IID-NDT). This correlation is related
to the hole-injection barrier from the gold electrode to the
organic semiconducting layer. Gold electrode has a work
function of 5.1 eV, and lowering the HOMO levels of polymers
increase the barrier. Hence threshold voltages became more
negative. IID-BDT and IID-NDT showed threshold voltages2
up to −30 V, nonetheless, their mobilities were high (0.48 cm
2
d,3
observed in other high performance PFET materials.
contrast, polymers with axisymmetric units showed almost
smooth and amorphous films. Polymers with centrosymmetric
In
1
766
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
IIDDT and IIDT-Se films showed four out-of-plane diffraction
peaks as spin-casted from their TCE solutions, which were
attributed to 100, 200, 300, and 400 diffractions. After the films
were thermally annealed at 150 °C for 30 min, the intensity of
diffraction peaks increased, and diffraction points became more
centered. IID-BDT and IID-NDT showed only three
diffraction peaks before annealing, but after being annealed
the fourth-order diffraction peaks appeared. Although these
polymer films are relatively more amorphous than P3HT films,
they also had ordered packing that was perpendicular to the
film plane. According to the out-of-plane GIXD, first diffraction
peaks of IIDDT, IIDDT-Se, IID-BDT, and IID-NDT were
very strong at 2θ of 3.58, 3.60, 3.59, and 3.65°, corresponding
to a d-spacing of 19.85, 19.74, 19.79, and 19.46 Å, respectively.
These results are consistent with molecular models that
polymer films had an edge-on lamellar packing in film. Only
three diffraction peaks were observed for IIDDT-Me after
thermal annealing and the 100 diffraction became much
broader compared to that of IIDDT, indicating a relatively
disordered lamellar packing. Presumably the methyl groups in
IIDDT-Me hindered the interpolymer π−π stacking, which
may deteriorate out-of-plane lamellar packing. In sharp
contrast, all polymers containing axisymmetric donor showed
poor lamellar packing. IID-T3 only displayed one diffraction
peak, and the peak of IIDT was weak (Figure 7k, l, and Figure
S7 in the Supporting Information). Polymers IID-TTT and
IID-TBT showed no obvious diffraction peak. Interestingly, the
first diffraction peak of IIDT was at 2θ of 3.25°, corresponding
to a d-spacing of 21.86 Å, significantly larger than those of the
centrosymmetric polymers.
Figure 6. Tapping-mode AFM height images of (a) IIDDT, (b) IID-
BDT, (c) IIDT, and (d) IID-TTT films spin-casted from TCE
solution (4 mg/mL) on OTS-treated SiO /Si substrate and annealed
2
at 150 °C for 30 min.
donors had higher surface roughness (3.46 nm for IIDDT and
2
.83 nm for IID-TT) than those with axisymmetric donors
(
1.07 nm for IIDT and 0.54 nm for IID-TTT).
To understand the device performances from a molecular
level, we employed GIXD to investigate polymer packing in
thin film. Figure 7 shows the two-dimensional GIXD (2D-
GIXD) patterns of polymer films prepared by spin-casting the
TCE solution of polymers onto OTS-treated SiO /Si substrate.
2
Figure 7. 2D-GIXD patterns of (a, e) IIDDT, (b, f) IID-Se, (c, g) IID-BDT, (d, h) IID-NDT, (i) IIDDT-Me, (k) IIDT, and (l) 6 in. thin films on
OTS-treated SiO /Si substrate (all polymers were spin-coated from 8 mg/mL TCE solution followed by solvent annealing). (a−d) spin-cast film
2
without thermal annealing and (e−i, k, l) annealed at 150 °C for 30 min at nitrogen atmosphere. (j) Schematic representation of the steric hindrance
of IIDDT-Me for molecular docking.
1
767
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
To further demonstrate our proposed packing model of
polymers, we performed ab initio DFT calculations. Specifically,
ωB97X-D functional, which includes empirical dispersion, was
used in order to correctly describe the weak intermolecular
interaction. Long alkyl chains were replaced with isobutyl
groups. We used oligomers (n = 2) of IIDDT to explore the
π−π stacking of two polymer chains and several local minimum
structures were found from different starting structures. Direct
overlap of isoindigo cores exhibited much higher energy
because the strong repulsion caused by alkyl chains. The most
stable intermolecular stacking is shown in Figure 8a and the
potential of the dimer. The HOMO of dimers was delocalized
on both oligomers which indicated the strong interaction of
their frontier orbitals (see Figure S8 in the Supporting
Information). However, after incorporating methyl groups,
oligomers of IIDDT-Me showed larger π−π stacking distance,
and the association energy was 49.9 kcal/mol, smaller than that
of IIDDT. Interestingly, the π−π stacking suppressed the
torsion angles of IIDDT-Me from 29° to 7.8°, which is also
consistent with the absorption spectra (see Figure S9 in the
Supporting Information). Polymers with axisymmetric donor,
such as IIDT, cannot form good π−π stacking and the
association energy of IIDT dimer was only 43.5 kcal/mol. The
association energy among these three polymers will be much
larger. Therefore, these computational results prove our
molecular packing models and support the “molecular docking”
strategy that small donor units can dock into large isoindigo
cores to enhance interchain π−π stacking.
2
3
Molecular Docking Strategy. Figure 9a shows a cartoon
27
representation of conjugated polymer films, which is very
similar to the film morphology of IIDDT (Figure 6a). In a
typical polymer film, three possible carrier transporting
pathways are marked with arrows: (1) intrachain carrier
transport, which is very fast and efficient; (2) interchain carrier
transport at well-ordered sites, which may adopt a hopping
mechanism like ordered small molecules; (3) interchain carrier
transport at loosely contacted sites, which is very slow.
Therefore, the carrier transport is obviously limited by loosely
contact sites, and improve the carrier mobility at pathway (3) is
crucial for high-performance PFETs. One effective way to
reduce zones of pathway (3) is to increase interpolymer chain
interactions and thus increasing the polymer packing orders.
Traditional polymers designed for PFETs, such as P3HT,
always contain one or two alkyl chains on each unit (Figure
9b). Typical alkanes (C H ) (n = 6−9) have inter alkyl chain
Figure 8. (a) Calculated dimer structures of the IIDDT oligomers (n
=
2). Structural optimization was performed at ωB97X-D/6-31G(d)
level. (b) Side view of the electrostatic potential of the dimer.
association energy of the dimer is 55.1 kcal/mol. Bithiophene
units partly overlapped with the isoindigo cores with π−π
stacking distances in the range of 3.4−3.5 Å. These results
agreed with the absorption spectra that the polymer formed
some J-aggregates in film. Figure 8b shows the electrostatic
n
2n+2
24
distances of 3.6−3.8 Å in single crystal, larger than a typical
9,25
π−π distance of 3.4 Å.
Hence, steric hindrance may exist
Figure 9. (a) Cartoon representation of conjugated polymer film, and its three possible charge transport pathways. (b) Traditional polymer contains
one or two alkyl chains on each unit. Steric hindrance may exist when polymers are stacked with each other (black circle). (c) Molecular docking
strategy to avoid steric hindrance and improve the interpolymer π−π interaction. The cartoon representation of copolymers with different
symmetries and their film packings; (d) polymers with centrosymmetric donors and (e) polymers with axisymmetric donors.
1
768
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
Scheme 2. Diketopyrrolopyrole and Naphthalenedicarboximide-Based Large Aromatic Core Polymers and Their FET
Performances
−1
when polymers are stacked with each other. Removal of alkyl
chains from small units will avoid this steric hindrance (Figure
s . Polymers with centrosymmetric units, such as PDBP-co-
TT, PDQT, P(DPP-alt-DTBSe) and n-type P(NDI2OD-T2),
showed excellent mobilities. In contrast, those with axisym-
9
c). Starting from this rough analysis, we propose the molecular
29
docking design strategy: (1) reducing the steric hindrance of
alkyl chains and making the small unit “dock into” a large
aromatic core, similar to the “brick layer” packing of TIPS-
pentacene; (2) large aromatic core can also reduce reorganiza-
tion energy of polymers, which has been widely used in the
metric units and zigzag backbones (PDP3T and PDPP-
30
TBT ) had relatively low mobilities. Similar to IID-T3,
although P1 contains axisymmetric thiophenes, it has an almost
linear backbone, thus showing good mobility. With an identical
backbone to PDQT yet two additional dodecyl chains,
6,7
28
2
−1
design of high-performance small molecule OFETs.
BBTDPP1 showed hole mobility as high as 0.1 cm V
−1
As described above, device performances of the polymers
apparently correlated with their film morphologies and polymer
packings. Polymers with centrosymmetric units exhibited better
crystallinity and lamellar packing in thin films, because of good
interchain π−π stacking. As shown in Figure 9d,e, for those
polymers with centrosymmetric units, the “small unit” can
readily dock into the cavity formed by the “large core” and
branched alkyl chains. The donor−acceptor interaction
between the isoindigo core and the donor unit further
enhances the docking. In addition, the polymers backbones
with centrosymmetric donors are almost linear and parallel to
the substrate. The decent π−π stacking and the linear backbone
of polymers led to good lamellar phase in polymer films, as
proved by GIXD experiment (Figure 9d). Even for methyl-
substituted IIDDT-Me, the backbone remains linear. IIDDT-
Me shows good lamellar packing yet poor maximum mobility
s , almost 1 order of magnitude lower than PDQT. This
example is comparable to IIDDT-Me.
CONCLUSION
■
In summary, we introduce the “molecular docking” strategy to
design new polymers for high-performance PFETs. The key
concept of molecular docking is that the small conjugated units
can dock into the cavity formed by large aromatic cores and
alkyl chains. This docking leads to increased interchain π−π
stacking of the polymer. According to the strategy, a series of
isoindigo-based polymers have been designed and synthesized
for high-performance PFETs. Excitingly, the molecular docking
design provides us systematically high device performances. Six
polymers with centrosymmetric units show hole-mobilities over
2
−1 −1
0
.1 cm V s , in which five polymers show hole-mobilities
2
−1 −1
over 0.3 cm V s . IIDDT gives a maximum hole-mobility
2
−1 −1
2
−1 −1
(
0.11 cm V s ), presumably because of poor π−π stacking
over 1 cm V s . In addition, these polymers show good
stability upon moisture due to low HOMO levels. By using
AFM and GIXD analyses, we have attributed the obvious
difference in device performance to the different symmetry and
backbone curvature of polymers, which greatly affect the
interchain π−π stacking, lamellar packing and crystallinity.
Furthermore, DFT calculations support our proposed packing
model and the molecular docking strategy. Finally, we compare
the device performances of several DPP and NDI based
polymers reported by other groups, and find that the molecular
docking strategy is also applicable in their systems. The
molecular docking strategy is effective and general in polymer
design for PFETs. We are currently extending this strategy to
other systems with large aromatic cores.
caused by methyl groups.
Polymer backbones containing axisymmetric donors, none-
theless, become zigzag, inhibiting the docking and lamellar
packing in polymer film, as shown in Figure 9e. With the most
linear backbone, IID-T3 shows the best hole-mobility up to
2
−1 −1
0.061 cm V s . This result indicates the importance of
26
backbone curvature. The important effect of symmetry and
backbone curvature
groups.
1
6,27
on PFETs are also reported by other
The “molecular docking” strategy is applicable to other
3b,d,e,g,28−30
systems, such as diketopyrrolopyrole (DPP)
naphthalenedicarboximide (NDI). Scheme 2 presents several
recently developed polymers with mobilities around 1 cm V
and
2d
2
−1
1
769
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770

Chemistry of Materials
Article
(
10) Ho
̈
ltje, H.-D.; Folkers, G. Molecular Modeling: Basic Principles
ASSOCIATED CONTENT
■
and Applications; Wiley-VCH: Weinheim, Germany, 1996.
11) Kitchen, D. B.; Decornez, H.; Furr, J. R.; Bajorath, J. Nat. Rev.
*
S
Supporting Information
(
Monomers/polymers synthesis, characterization, and device
fabrication details. This material is available free of charge via
the Internet at http://pubs.acs.org.
Drug Discovery 2004, 3, 935.
(12) (a) Wang, E.; Ma, Z.; Zhang, Z.; Vandewal, K.; Henriksson, P.;
Inganas
̈
, O.; Zhang, F.; Andersson, M. R. J. Am. Chem. Soc. 2011, 133,
4244. (b) Zhang, G.; Fu, Y.; Xie, Z.; Zhang, Q. Macromolecules 2011,
4, 1414. (c) Mei, J.; Graham, K. R.; Stalder, R.; Reynolds, J. R. Org.
1
4
AUTHOR INFORMATION
■
*
Lett. 2010, 12, 660. (d) Stalder, R.; Mei, J.; Reynolds, J. R.
Macromolecules 2010, 43, 8348.
(13) Stalder, R; Mei, J. G.; Subbiah, J.; Grand, C.; Estrada, L. A.; So,
Corresponding Author
E-mail: jianpei@pku.edu.cn (J. Pei).
F.; Reynolds, J. R. Macromolecules 2011, 44, 6303.
Notes
(
(
14) Guo, X. G.; Watson, M. D. Org. Lett. 2008, 10, 5333.
15) Chen, Z.; Zheng, Y.; Yan, H.; Facchetti, A. J. Am. Chem. Soc.
The authors declare no competing financial interest.
2
009, 131, 8.
ACKNOWLEDGMENTS
(16) Rieger, R.; Beckmann, D.; Mavrinskiy, A.; Kastler, M.; Mullen,
K. Chem. Mater. 2010, 22, 5314.
̈
■
This work was supported by the Major State Basic Research
Development Program (No 2009CB623601) from the Ministry
of Science and Technology, and National Natural Science
Foundation of China. We thank Prof. Xiao-Yu Cao (Xiamen
University) for his suggestion and insightful discussion. The
authors thank beamline BL14B1 (Shanghai Synchrotron
Radiation Facility) for providing the beam time.
(
17) Osaka, I.; Abe, T; Shinamura, S.; Miyazaki, E.; Takimiya, K. J.
Am. Chem. Soc. 2010, 132, 5000.
18) Sato, M.; Asami, A.; Maruyama, G.; Kosuge, M.; Nakayama, J.;
Kumakura, S.; Fujihara, T.; Unoura, K. J. Organomet. Chem. 2002, 654,
6.
19) Li, J.; Qin, F.; Li, C. M.; Bao, Q.; Chan-Park, M. B.; Zhang, W.;
Qin, J.; Ong, B. S. Chem. Mater. 2010, 22, 5314.
20) Rieger, R.; Beckmann, D.; Pisula, W.; Steffen, W.; Kastler, M.;
Mullen, K. Adv. Mater. 2010, 22, 83.
21) Majewski, L. A.; Kingsley, J. W.; Balocco, C.; Songa, A. M. App.
Phys. Lett. 2006, 88, 222108.
22) Chang, J.-F.; Clark, J.; Zhao, N.; Sirringhaus, H.; Breiby, D. W.;
(
5
(
(
̈
REFERENCES
■
(
(
1) (a) Katz, H. E. Chem. Mater. 2004, 16, 4748. (b) Zaumseil, J.;
Sirringhaus, H. Chem. Rev. 2007, 107, 1296. (c) Tang, Q.; Jiang, L.;
Tong, Y.; Li, H.; Liu, Y.; Wang, Z.; Hu, W.; Liu, Y.; Zhu, D. Adv.
Mater. 2008, 20, 2947. (d) Wen, Y.; Liu, Y. Adv. Mater. 2010, 22, 1331.
e) Guo, Y.; Yu, G.; Liu, Y. Adv. Mater. 2010, 22, 4427.
2) (a) Garnier, F.; Hajlaoui, R.; Yassar, A.; Srivastava, P. Science
994, 265, 1864. (b) Sirringhaus, H.; Tessler, N.; Friend, R. H. Science
998, 208, 1741. (c) Sirringhaus, H.; Kawase, T.; Friend, R. H.;
Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E. P. Science 2000, 290,
123. (d) Yan, H.; Chen, Z.; Zheng, Y.; Newman, C.; Quinn, J. R.;
Doltz, F.; Kestler, M.; Facchetti, A. Nature 2009, 457, 679.
3) (a) Guo, X.; Ortiz, R. P.; Zheng, Y.; Kim, M.-G.; Zhang, S.; Hu,
Y.; Lu, G.; Facchetti, A.; Marks, T. J. J. Am. Chem. Soc. 2011, 133,
(
Andreasen, J. W.; Nielsen, M. M.; Giles, M.; Heeney, M.; McCulloch,
I. Phys. Rev. B 2006, 74, 115318.
(
(
23) Cha, D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10,
615.
24) Boese, R.; Weiss, H. C.; Blas
8, 988.
25) (a) Liang, Z.; Tang, Q.; Xu, J.; Miao, Q. Adv. Mater. 2011, 23,
535. (b) Lei, T.; Zhou, Y.; Cheng, C.-Y.; Cao, Y.; Peng, Y.; Bian, J.;
(
1
1
6
(
̈
er, D. Angew. Chem., Int. Ed. 1999,
3
(
2
1
Pei, J. Org. Lett. 2011, 13, 2642. (c) Liu, Y.-Y.; Song, C.-L.; Zeng, W.-
J.; Zhou, K.-G.; Shi, Z.-F.; Ma, C.-B.; Yang, F.; Zhang, H.-L.; Gong, X.
J. Am. Chem. Soc. 2010, 132, 16349.
(
1
1
3685. (b) Ha, J. S.; Kim, K. H.; Choi, D. H. J. Am. Chem. Soc. 2011,
33, 10364. (c) Lei, T.; Cao, Y.; Fan, Y.; Liu, C.-J.; Yuan, S.-C.; Pei, J.
(
26) He, M.; Li, J.; Tandia, A.; Sorensen, M.; Zhang, F.; Fong, H. H.;
Pozdin, V. A.; Smilgies, D.-M.; Malliaras, G. G. Chem. Mater. 2010, 22,
770.
27) Osaka, I.; Abe, T.; Shinamura, S.; Takimiya, K. J. Am. Chem. Soc.
011, 133, 6852.
28) Burgi, L.; Turbiez, M.; Pfeiffer, R.; Bienewald, F.; Kirner, H.-J.;
Winnewisser, C. Adv. Mater. 2008, 20, 2217.
29) Bijleveld, J. C.; Zoombelt, A. P.; Mathijssen, S. G. J.; Wienk, M.
M.; Turbiez, M.; Leeuw, D. M.; Janssen, R. A. J. J. Am. Chem. Soc.
009, 131, 16616.
30) Sonar, P.; Singh, S. P.; Li, Y.; Soh, M. S.; Dodabalapur, A. Adv.
Mater. 2010, 22, 5409.
J. Am. Chem. Soc. 2011, 133, 6099. (d) Bronstein, H.; Chen, Z.; Ashraf,
R. S.; Zhang, W.; Du, J.; Durrant, J. R.; Tuladhar, P. S.; Song, K.;
Watkins, S. E.; Geerts, Y.; Wienk, M. M.; Janssen, R. A. J.;
Anthopoulos, T.; Sirringhaus, H.; Heeney, M.; McCulloch, I. J. Am.
Chem. Soc. 2011, 133, 3272. (e) Li, Y.; Sonar, P.; Singh, S. P.; Soh, M.
S.; van Meurs, M.; Tan, J. J. Am. Chem. Soc. 2011, 133, 2198. (f) Tsao,
H. N.; Cho, D. M.; Park, I.; Hansen, M. R.; Mavrinskiy, A.; Yoon, D.
2
(
2
(
(
Y.; Graf, R.; Pisula, W.; Spiess, H. W.; Mu
̈
llen, K. J. Am. Chem. Soc.
011, 133, 2605. (g) Li, Y.; Singh, S. P.; Sonar, P. Adv. Mater. 2010,
2, 4862.
2
(
2
2
(
4) Sirringhaus, H. Adv. Mater. 2009, 21, 3859.
(
5) (a) Osaka, I.; Sauve, G.; Zhang, R.; Kowalewski, T.; McCullough,
R. D. Adv. Mater. 2007, 19, 4160. (b) Osaka, I.; Zhang, R.; Sauve, G.;
Smilgies, D.-M.; Kowalewski, T.; McCullough, R. D. J. Am. Chem. Soc.
2
009, 131, 2521. (c) Osaka, I.; Zhang, R.; Liu, J.; Smilgies, D.-M.;
Kowalewski, T.; McCullough, R. D. Chem. Mater. 2010, 22, 4191.
d) Osaka, I.; Takimiya, K.; McCullough, R. D. Adv. Mater. 2010, 22,
993.
6) Coropceanu, V.; Jer
Silbey, R.; Bredas, J.-L. Chem. Rev. 2007, 107, 926.
7) Takimiya, K.; Shinamura, S.; Osaka, I; Miyazaki, E. Adv. Mater.
011, 23, 4347.
8) (a) Anthony, J. E. Angew. Chem., Int. Ed. 2008, 47, 452.
b) Anthony, J. E. Chem. Rev. 2006, 106, 5028.
9) Sokolov, A. N.; Atahan-Evrenk, S.; Mondal, R.; Akkerman, H. B.;
Sanchez-Carrera, R. S.; Granados-Focil, S.; Schrier, J.; Mannsfeld, S. C.
B.; Zoombelt, A. P.; Bao, Z.; Aspuru-Guzik, A. Nature Commun. 2011,
, 437.
(
4
(
́ ̂
ome, C.; da Silva Filho, D. A.; Olivier, Y.;
́
(
2
(
(
(
́
2
1
770
dx.doi.org/10.1021/cm300117x | Chem. Mater. 2012, 24, 1762−1770