Side chain modification on PDI-spirobifluorene-based molecular acceptors and its impact on organic solar cell performances

Article abstract of DOI:10.1039/c8nj04810j

N-Substitution in perylene diimide (PDI) n-type semiconductors is critical for their performance in organic bulk heterojunction solar cells. In this work we synthesized and compared three perylene diimide-spirobifluorene derivatives, N-substituted with different alkyl side groups. These molecular systems were obtained by cost-effective methods, using straightforward synthetic procedures and easy purification. These PDI derivatives were used as electron acceptor materials, blended with a benchmark quinoxaline-benzodithiophene based donor polymer, in bulk heterojunction solar cells. All devices were processed by blade coating in air, and also using non-chlorinated solvents. The three molecules presented identical HOMO/LUMO energy levels and very similar optical properties but different photovoltaic responses. These different performances were discussed through optical, electrical and morphological characterizations. The highest power conversion efficiency was achieved for the active layer based on the derivative with branched and longer alkyl groups in N-positions, which favored a morphology with reduced donor:acceptor phase segregation.

Full text of DOI:10.1039/c8nj04810j

NJC
View Article Online
View Journal | View Issue
PAPER
Side chain modification on PDI-spirobifluorene-
based molecular acceptors and its impact on
organic solar cell performances†
Cite this: New J. Chem., 2018,
42, 18633
a
d
Margherita Bolognesi,
Marinella Catellani,
Desta Gedefaw,bc Marco Cavazzini,
e
c
a
Mats R. Andersson,
Michele Muccini,
Erika Kozma *e and Mirko Seri
*
f
N-Substitution in perylene diimide (PDI) n-type semiconductors is critical for their performance in
organic bulk heterojunction solar cells. In this work we synthesized and compared three perylene
diimide-spirobifluorene derivatives, N-substituted with different alkyl side groups. These molecular
systems were obtained by cost-effective methods, using straightforward synthetic procedures and easy
purification. These PDI derivatives were used as electron acceptor materials, blended with a benchmark
quinoxaline–benzodithiophene based donor polymer, in bulk heterojunction solar cells. All devices were
processed by blade coating in air, and also using non-chlorinated solvents. The three molecules
presented identical HOMO/LUMO energy levels and very similar optical properties but different
photovoltaic responses. These different performances were discussed through optical, electrical and
morphological characterizations. The highest power conversion efficiency was achieved for the active
layer based on the derivative with branched and longer alkyl groups in N-positions, which favored a
morphology with reduced donor:acceptor phase segregation.
Received 21st September 2018,
Accepted 7th October 2018
DOI: 10.1039/c8nj04810j
rsc.li/njc
of common chlorinated solvents with alternative ones, having
reduced costs, toxicity and environmental impact;3 and
Introduction
Solution-processed organic solar cells (OSCs) are a possible (iii) competitive efficiency. Within solution-processed OSCs,
alternative technology for the generation of sustainable energy. lab-scale single layer bulk heterojunction (BHJ) devices have
OSCs offer benefits in terms of processing costs, large area reached efficiencies 410% due to the synergic development of
production, compatibility with flexible and light-weight supports, novel active materials,4 in depth understanding of polymeric
abundant and non-toxic elements and structural tunability at film microstructures5 and meticulous device optimization,6
different levels.1 Some strategic aspects of the OSC technology indicating their potential for future commercial applications.7
are: (i) the use of deposition methods, at laboratory-scale, compa-
The properties of active materials still play a crucial role for
tible with industrial production systems;2 (ii) the replacement the achievement of efficient OSCs.8 Fullerene derivatives have
been the most used acceptors thanks to their high and isotropic
electron mobility, good solubility, proper nanoscale domain
a Istituto per lo Studio dei Materiali Nanostrutturati (ISMN), Consiglio Nazionale
formation, etc.9 However, fullerene derivatives also have some
delle Ricerche (CNR), Via P. Gobetti 101, 40129 Bologna, Italy
b School of Biological and Chemical Sciences, The University of South Pacific,
drawbacks, such as low absorption coefficient and low spectral
coverage, hard purification and high cost, limited energy level
tunability, morphological instability, and inferior mechanical
flexibility.10 For these reasons it is not clear whether fullerene
derivatives are ultimately the best candidate materials, and
therefore many efforts have been focused on the development
of alternative non-fullerene acceptors. Various small molecules/
polymers having acceptor properties have been recently designed
and found to generate performances greater than those achieved
using fullerenes, with greater than 13% efficiency.11
Laucala Campus, Suva, Fiji
c Flinders Institute for NanoScale Science and Technology, Flinders University,
Sturt Road, Bedford Park, SA 5042, Australia
d Istituto di Scienze e Tecnologie Molecolari (ISTM), Consiglio Nazionale delle
Ricerche (CNR), Dipartimento di Chimica Organica e Industriale,
`
Universita degli Studi di Milano, 20133 Milano, Italy
e Istituto per lo Studio delle Macromolecole (ISMAC),
Consiglio Nazionale delle Ricerche (CNR), Via A. Corti 12, 20133 Milano, Italy.
E-mail: erika.kozma@ismac.cnr.it
f
`
Istituto per la Sintesi Organica e la Fotoreattivita (ISOF),
Consiglio Nazionale delle Ricerche (CNR), Via P. Gobetti 101, 40129 Bologna,
Italy. E-mail: mirko.seri@isof.cnr.it
Within non-fullerene materials, perylene diimide derivatives
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8nj04810j (PDIs) are suitable candidates as acceptors in OSCs, because of
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 18633--18640 | 18633

View Article Online
Paper
NJC
their low cost, high absorption coefficient, good photochemical N-substitution with differently shaped alkyl groups, on this
stability and chemical versatility.12 Power conversion efficien- promising molecular structure.
cies (PCEs) beyond 8% were reported using helical PDI trimers
Considering the importance of understanding the structure–
blended with PTB7 in single-junction BHJ solar cells,13 and property relationship and the effect of structural modification
PCEs 410% were obtained with PDI derivatives in tandem on OSC performance, we report here the comparison of three
devices.14 Although the performances are rather high, it is molecular PDI derivatives (Fig. 1) having a sterically hindered
difficult to imagine that complicated synthetic procedures used spirobifluorene (SF) unit in the bay position and different alkyl
for the preparation and purification of helical PDIs and/or side chains at the imide position: 10-nonadecyl (PDI-SF-nd),20
complex solar cell architectures may be adapted eventually 2-ethylhexyl (PDI-SF-eh) and n-dodecyl (PDI-SF-d). The corres-
for scale-up. Therefore, it is mandatory to design low cost ponding optical, electrochemical, morphological and photovoltaic
materials, by employing straightforward synthetic and purifica- properties of all three PDIs were systematically investigated. Using
tion procedures for their use in cost-effective BHJ solar cells a benzodithiophene–quinoxaline based polymer (PFQBDT-TR1)21
with simple architectures.
as a p-type material, the resulting BHJ solar cells gave PCEs
Perylene diimides have two chemically active centers: imide ranging from 0.5% to 1.5% as a function of different designs of
positions and bay positions.15 The imide positions are usually the acceptor molecules.
substituted with linear or branched alkyl chains to increase
solubility and modulate the self-organization in the solid
state,16 while the bay substitution is used both to tune the
HOMO–LUMO electronic levels and, depending on the substi-
tuent, to twist the perylene core.17 Indeed, the presence of
sterically hindered moieties in the bay positions generates
out-of-plane PDI structures, hindering intra- and inter-molecular
interactions in the solid state.18 For photovoltaic applications, the
presence of strong intermolecular interactions could determine
the formation of (too) large crystalline domains which, through
exciton self-trapping, could limit charge generation.19
Experimental section
Materials
All starting materials were purchased from Aldrich and used as
received without further purification. Solvents and reagents
were dried and/or distilled by the usual methods and typically
used under an inert gas atmosphere. The PDI derivatives
were prepared by a Suzuki cross-coupling reaction and a
detailed procedure for their synthesis is reported in the ESI†
(Schemes S1 and S2).
In our previous studies several PDI derivatives based
on different building blocks were designed, studied and
characterized.20 These acceptor molecules were designed with
the aim to reduce PDI’s tendency to aggregate and crystallize in
the solid state, by the introduction of sterically demanding
moieties on the perylene core. Among them, PDI-SF-nd (Fig. 1)
containing a spirobifluorene moiety in the bay position
resulted in the best performing acceptor material in combi-
nation with P3HT, yielding a PCE of B1.5%. The presence of
the spirobifluorene unit reduces the aggregation in the solid
state due to its unique cross-shape rigid structure and strong
steric hindrance, while its electron donor ability contributes to
the extension of the absorption. This prompted us to further
investigate the impact of fine structural tuning, such as the PDI
Cyclic voltammetry
Cyclic voltammetry measurements were performed in dichloro-
methane solution containing tetrabutyl ammonium tetrafluoro-
borate (0.1 M) as the electrolyte, under a nitrogen atmosphere
with a computer controlled Amel 2053 (with Amel 7800
interface) electrochemical workstation, in a three electrode
single-compartment cell using a platinum plate as the working
electrode, a platinum wire as the counter electrode and an SCE
standard reference electrode. An Fc/Fc+ redox couple was used
as the internal standard. A scan rate of 50 mV sÀ1 was used for
all measurements.
Device fabrication and characterization
As already described in ref. 21 we report the main details
regarding the fabrication and characterization of all devices.
The materials used, PEDOT:PSS (poly(3,4-ethylenedioxythiophene):
poly(4-styrenesulfonate), Clevios P VP A1 4083, H. C. Starck),
PC61BM ([6,6]-phenyl-C61-butyric acid methyl ester, Solenne BV),
1,8-diiodooctane (DIO) and various anhydrous solvents (Sigma-
Aldrich) were purchased from commercial sources and used
without further purification. Patterned ITO-coated glasses
(Rs B 10 O sqÀ1) were cleaned in deionized water, acetone
and isopropanol baths under ultrasonication for 15 minutes.
After the final sonication step, substrates were dried and placed
in an oxygen plasma chamber for 5 min. A layer (B40 nm;
blade speed 2 mm sÀ1 at 30 1C) of PEDOT:PSS (or ZnO, provided
by Genes’Ink, for inverted structures, Table S1, ESI†) was blade-
coated in air on the top of the ITO and then annealed at 150 1C
Fig. 1 Chemical structure of PDI-SF-nd, PDI-SF-eh, and PDI-SF-d acceptors
and the PFQBDT-TR1 donor (R1 = n-octyl).
18634 | New J. Chem., 2018, 42, 18633--18640 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
for 15 min (or 80 1C for 5 min for ZnO). The active solutions have already shown that the PDI and the spirobifluorene
were prepared inside the glove-box, dissolved in ODCB : DIO moieties adopt a highly twisted conformation, with a dihedral
(97 : 3; % v/v) or o-xylene : DIO (97 : 3; % v/v) at a total concen- angle between them of approximately 501.23 This is expected
tration of 30 mg mLÀ1, and kept under stirring overnight at to provide the three acceptor molecules good solubility in
80 1C. Lower concentrations (25 or 20 mg mLÀ1) were also used common organic solvents (425 mg mLÀ1) and a low inter-
for 1 : 3 PFQBDT-TR1 : PDI-SF-nd based solutions. The resulting molecular aggregation tendency in the solid state. However, the
blends were deposited in air by blade-coating (blade speed of different alkyl chains at the imide positions can further tune
7.5 mm sÀ1 at 80 1C or 5 mm sÀ1 at 60 1C for ODCB and o-xylene the solubility and miscibility of the molecules with a polymeric
based solutions, respectively) on the top of the ITO/PEDOT:PSS donor, as discussed below.
(or ZnO) surface. Before cathode (or anode for inverted struc-
tures) deposition, the substrates were annealed or left as-cast.
Optical and electrochemical properties
Next, LiF (0.6 nm) and Al (100 nm) (or MoO3 (10 nm)/Ag
(100 nm) for inverted devices) were deposited in a vacuum
chamber. The current–voltage (J–V) characteristics of all devices
were recorded using a Keithley 236 source-measure unit under
AM1.5G simulated light, 100 mW cmÀ2 (Abet Technologies Sun
2000 Solar Simulator). The active area of the solar cell was
exactly 6 mm2. During testing, each cell was carefully masked
using a calibrated mask. All devices were tested, without
encapsulation, inside the glove box in an oxygen and moisture
free environment.
Electron mobility was calculated from the dark J–V curves of
electron-only devices (structure: glass/ITO/ZnO (40 nm)/active
layer/LiF (0.6 nm)/Al (100 nm)) based on the best performing
films, using the space-charge limited current (SCLC) method
through the Mott–Gurney equation:22 J = (9/8)mere0V2/d3, where
J is the current density, m is the mobility, er is the relative
permittivity of the blend, e0 is the permittivity of vacuum, d is
the thickness of the active layer, and V is the voltage. The
resulting curves are plotted as J1/2 vs. V (Fig. S2, ESI†).
As shown in Fig. 2a and listed in Table 1, the absorption spectra
of PDI-SF-nd, PDI-SF-eh, and PDI-SF-d acceptors in diluted
solutions are almost perfectly overlapped. The spectra have a
typical pattern of bay substituted PDIs and extend from 300 to
650 nm, with a weaker absorption band at 433–435 nm, a more
intense bandwidth maximum at 576–580 nm and one vibra-
tional shoulder. Passing from diluted solutions to thin films,
the spectra of the three PDI-based acceptors remain almost
overlapped (Fig. 2b).
In detail, for the three molecules the higher energy band still
lies at 432–438 nm, while the lower energy and more intense
band is blue-shifted by 15 nm to 562–565 nm, likely due to the
switch between the relative intensities of the two vibrational
features of the band. This shift corresponds also to a red-shift
of the absorption onset wavelength (lonset) of the molecules in
the solid state, which increases by around 20 nm passing from
PDI-SF-nd to PDI-SF-d. This is traduced into a reduction of the
optical band gaps following the trend PDI-SF-nd 4 PDI-SF-eh
4 PDI-SF-d. Most probably, the longer and branched alkyl side
chains of PDI-SF-nd hinder ordering, while a slightly higher
aggregation tendency is favored by the short branched alkyl
chains of PDI-SF-eh and even more by the long linear alkyl
chains of PDI-SF-d. However, the three structures present good
solubility in common organic solvents with a generally low
aggregation/ordering tendency in the solid state. In Fig. 2b the
normalized absorption spectrum of the polymer PFQBDT-TR1
is also reported. The blend of the polymer with any of the three
acceptors is expected to give a good coverage of the sun’s
spectrum, between 300 and 750 nm. Indeed, the absorption
spectra of the active layers of the optimized devices based on
PFQBDT-TR1 and the three acceptors well cover all wavelengths
up to 750 nm (see Fig. 4e).
Thin film characterization
All thin-film characterizations were performed in air. The
solution and film optical absorption spectra were recorded on
a JASCO V-550 spectrophotometer. The thickness of the various
films was measured using a profilometer (KLA Tencor, P-6).
Atomic force microscopy (AFM) images were recorded using a
Solver Pro (NT-934 MDT) scanning probe microscope in
tapping mode. All active films were characterized after deposi-
tion on glass/ITO/PEDOT:PSS sub-structures (and then corrected
for the background signal) in order to mimic the real working
conditions of the BHJ blends. The emission spectra were
measured using a PerkinElmer LS-50 spectrofluorometer in
reflection mode.
Results and discussion
Chemical structure
PFQBDT-TR1 and PDI-SF-nd were synthesized by Stille or
Suzuki coupling reactions as already reported in the literature,
while PDI-SF-eh and PDI-SF-d were synthesized using a similar
procedure reported in the ESI.†
The three acceptor molecules present the same PDI core
bay-substituted with a spirobifluorene unit, and different alkyl
chains at the imide positions. Quantum chemical calculations
Fig. 2 (a) Normalized absorption spectra of PDI-SF-nd, PDI-SF-eh, and
PDI-SF-d acceptors in diluted (E10À6 M) ortho-dichlorobenzene solution;
(b) normalized absorption spectra of PDI-SF-nd, PDI-SF-eh, PDI-SF-d
acceptors and the PFQBDT-TR1 donor in thin films.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 18633--18640 | 18635

View Article Online
Paper
NJC
Table 1 Absorption properties in solution and solid state, and the electrochemical characteristics, experimental HOMO and LUMO energy values (EHOMO
opt
gap
and ELUMO) and optical band gaps (E ) for PDI-SF-nd, PDI-SF-eh, PDI-SF-d and PFQBDT-TR1. Half-way reduction potentials (Ered1/2) are obtained from
the first reduction peaks of cyclic voltammetry, while the oxidation potentials were not observable. Eogappt was calculated from the onset of the absorption
in the solid state. ELUMO = Àe(Ered + 4.4) eV and EHOMO = Eogpt + ELUMO
1/2
Solution
Thin film
opt
gap
Material
lmax (nm)
lonset (nm)
lmax (nm)
lonset (nm)
E
(eV)
E0red (V)
EHOMO (eV)
ELUMO (eV)
PDI-SF-nd
PDI-SF-eh
PDI-SF-d
576
577
580
614
631
637
642
707
562
564
565
610
639
650
659
729
1.94
1.91
1.88
1.70
À0.68
À0.71
À0.69
À0.73
À5.66
À5.60
À5.59
À5.37
À3.72
À3.69
À3.71
À3.67
PFQBDT-TR1
The EHOMO and ELUMO of the acceptors, PDI-SF-nd, PDI-SF-eh, (PEDOT:PSS and the BHJ active layer) were deposited in air by
and PDI-SF-d, and of the donor, PFQBDT-TR1, were determined by blade-coating, which is an easy, reproducible and scalable
cyclic voltammetry (Fig. S1 ESI†) and calculated from their optical deposition method.26 The photovoltaic responses of PFQBDT-TR1:
band gaps (Table 1). Since the bay substitution moiety is the same PDI-SF-nd/PDI-SF-eh/PDI-SF-d based solar cells depend on the
for all three PDIs, the EHOMO and ELUMO are very similar for the active layer processing in terms of BHJ blend composition and
three acceptors. As known from the literature, indeed, PDIs’ solvent, deposition parameters and thickness, and thermal
frontier molecular orbitals have nodes at the nitrogen atom, and annealing. These variables are indeed critical to simultaneously
this reduces the coupling between the perylene core and the imide optimize the optical, electrical and morphological properties
substituents.24 The EHOMO and ELUMO positions indicate that of the BHJ blends. The processing conditions and the photo-
PDI-SF-nd, PDI-SF-eh, and PDI-SF-d have suitable energy levels to voltaic responses of the PFQBDT-TR1:PDI-SF-nd/PDI-SF-eh/
be used as acceptors in BHJ solar cells with PFQBDT-TR1 as the PDI-SF-d based solar cells prepared during optimization are
donor. Notably, the small offset between the ELUMO of the BHJ reported in Table S1 (ESI†). Additional details regarding the
components is not an issue for non-fullerene-based devices, since fabrication and characterization of the OSCs are reported in the
it is known that photogenerated excitons can separate very effi- experimental section. The optimized processing conditions and
ciently upon small and/or negligible driving energies, contributing the relative photovoltaic response curves and data, including
to high photovoltages and overall PCEs.11a,25
VOC, JSC, FF and PCE, are summarized in Table 2 and Fig. 3.
The photoluminescence (PL) spectra of the pristine polymer
The best D : A ratio was 1 : 2 (w/w) for all acceptors, different
and optimized blended films were also recorded (Fig. S3, ESI†). from the reference devices based on P3HT:PDI-SF-nd20 and
The PL spectra were collected by exciting the films both at PFQBDT-TR1 : PCBM21 for which the best ratio was 1 : 1 (w/w).
650 nm and 565 nm. At 650 nm the absorption of the polymer is Note that, by further increasing the D : A ratio of the PDI-based
predominant compared to the PDI-acceptor contributions (Fig. 2b). blends to 1 : 3 w/w a detrimental impact in terms of morpho-
The PL of the resulting BHJ films can be attributed to the polymer logical and photovoltaic properties was observed (Table S1, ESI†).
quenched emission. Energy transfer processes can be reasonably
Overall, the FF of the devices, even after thermal annealing,
excluded due to the unfavorable spectral overlap between the is quite low (o33%, Table 2) highlighting the presence of
polymer emission and the acceptors’ absorption and due to the significant internal resistances. This is most likely due to a
absence of acceptors’ emission features in the PL spectra of suboptimal charge transport/collection causing recombination
the blends (Fig. S3B, ESI†). Therefore, the charge transfer between processes to occur in the active layer. Within the three acceptors,
the excited polymer and the acceptor molecules is the main process the best PCE was obtained for PDI-SF-nd, with a value of 1.5%,
responsible for the PL quenching in the blends. In addition, by while a PCE of 0.5% and of 0.9% was obtained for PDI-SF-eh
exciting the blends at 565 nm (Fig. S3B, ESI†) we investigated the and PDI-SF-d, respectively. The observed PCE trend is mainly
emission signal arising from the acceptor molecules (E650 nm, as governed by the JSC of the solar cells, which follows the same trend
reported in ref. 20b). The resulting quenched emissions suggested (PDI-SF-nd 4 PDI-SF-d 4 PDI-SF-eh). Since the thickness and the
even in this case an efficient charge transfer (holes) from the optical properties of all BHJ devices are almost identical, the trend
excited acceptor(s) to the donor polymer.
of the resulting JSC could be related to the different nanoscale
Finally, while a similarly efficient photoinduced charge transfer morphologies and charge transport properties of the active
can be reasonably supposed for all the three optimized blends based blends, as discussed in the next section. Notably, by substituting
on PFQBDT-TR1 and the PDI-acceptor(s), their charge dynamics ODCB with the non-chlorinated and more environmentally
(mobility, recombinations, etc.) could be different, affecting the friendly o-xylene (o-xy), the photovoltaic parameters and the
resulting photovoltaic properties (see the following sections).
overall PCE of the best performing devices remain almost
unvaried, confirming the good solubility and processability of
this class of non-fullerene acceptors.
Photovoltaic properties
The photovoltaic properties of the three acceptors were tested
Interestingly, despite the significant recombination loss
in devices with standard configuration: glass/ITO/PEDOT:PSS/ processes and internal resistances within the different active
BHJ active layer/LiF/Al. The organic layers of the solar cells layers, as suggested by the shape of the corresponding J–V
18636 | New J. Chem., 2018, 42, 18633--18640 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
Table 2 Experimental conditions for the preparation of optimized devices based on BHJ blends of PFQBDT-TR1 with PDI-SF-nd, PDI-SF-eh or
PDI-SF-d. Photovoltaic parameters of the corresponding devices measured under 1 sun illumination: short circuit photocurrent (JSC), open circuit
voltage (VOC), fill factor (FF) and power conversion efficiency (PCE). The reported data are averaged over four different solar cells
Active layer
D : A ratio Solventa Annealing T (1C) Annealing time (min) Thickness (nm) JSC (mA cmÀ2) VOC (V) FF (%) PCE (%)
PFQBDT-TR1:PDI-SF-nd (1 : 2)
ODCB
ODCB
o-xy
—
140
140
—
20
20
120
120
115
2.6
6.1
5.6
0.79
0.76
0.76
29
32
33
0.6
1.5
1.4
(1 : 2)
(1 : 2)
PFQBDT-TR1:PDI-SF-eh (1 : 2)
ODCB
ODCB
o-xy
—
140
140
—
20
20
110
110
110
1.3
1.9
1.7
0.81
0.83
0.83
28
32
28
0.3
0.5
0.4
(1 : 2)
(1 : 2)
PFQBDT-TR1:PDI-SF-d (1 : 2)
ODCB
ODCB
o-xy
—
140
140
—
20
20
110
110
110
1.8
3.3
2.9
0.85
0.83
0.83
33
33
33
0.5
0.9
0.8
(1 : 2)
(1 : 2)
a
All active solutions contain 3% (v/v) of 1,8-diiodooctane (DIO) as a solvent additive.
correlate well with the photovoltaic responses previously discussed.
In the case of the reference PFQBDT-TR1:PCBM blend, a well
defined nanostructuring of the surface into small domains is
observed (Fig. 4d). This indicates an optimal donor:acceptor phase
segregation into nanocrystalline aggregates, which favors the
generation and transport of free charges in both the polymeric
donor and fullerene acceptor domains. This is traduced into a
highly performing solar cell.21 By contrast, the topography of the
best performing blend, within the series PFQBDT-TR1:PDI-SF-nd/
PDI-SF-eh/PDI-SF-d, shows nanostructures with undefined D:A
Fig. 3 J–V curves of the best devices based on PFQBDT-TR1 with
PDI-SF-nd (black), PDI-SF-eh (blue) or PDI-SF-d (green).
curves (Fig. 3), the VOC of all devices is still remarkable. This
indicates that despite significant voltage losses due to internal
inefficient mechanisms, the almost aligned LUMO levels of the
BHJ components guarantee satisfying VOC values, comparable
to the reference system based on PCBM (VOC: 0.83 V, Table S1,
ESI†).
Note that, the different VOC observed for PFQBDT-TR1:PDI-
SF-nd (E0.76 V) and PFQBDT-TR1:PDI-SF-eh/PDI-SF-d (E0.83 V)
based BHJ solar cells can be ascribed to different BHJ nano-
morphologies, as discussed below.27
Morphological properties
The nanomorphology of the active layer plays an important role
in determining the photovoltaic performance of a BHJ solar
cell. To better understand how the alkyl chains of the molecular
acceptors affect the solar cell characteristics, we investigated
by tapping-mode AFM the morphology of the PFQBDT-TR1:
PDI-SF-nd/PDI-SF-eh/PDI-SF-d blends processed under optimal
conditions (Fig. 4).
We compared these nanomorphologies with that of the refer-
ence PFQBDT-TR1:PCBM blend. Although the surface roughness is
almost the same for the four blends (root mean square between 0.7
and 1 nm), their morphological features are slightly different and
Fig. 4 AFM height images (10 mm  10 mm) of optimized solar cells with
BHJ blends processed from ODCB solutions: (a) PFQBDT-TR1:PDI-SF-nd
(1 : 2), (b) PFQBDT-TR1:PDI-SF-eh (1 : 2), (c) PFQBDT-TR1:PDI-SF-d (1 : 2),
(d) PFQBDT-TR1:PCBM (1 : 1). (e) Normalized absorption spectra of the
blended films.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 18633--18640 | 18637

View Article Online
Paper
NJC
domains (PFQBDT-TR1:PDI-SF-nd in Fig. 4a), whilst the lowest favor donor:acceptor miscibility, correlates well with the enhanced
performing blend presents more defined and finely nanostruc- PCE of the corresponding devices. On the other hand, the lower
tured domains (PFQBDT-TR1:PDI-SF-eh in Fig. 4b) and the electron mobility for PFQBDT-TR1:PDI-SF-d and PFQBDT-TR1:
intermediate performing an intermediate topography (PFQBDT-TR1: PDI-SF-eh films, combined with an enhanced donor:acceptor
PDI-SF-d in Fig. 4c). This different segregation tendency of the phase segregation, could be responsible for geminate recombi-
BHJ blends is also confirmed by the corresponding UV-Vis nation losses which limit the performances of the corresponding
spectra (Fig. 4e). In particular, the relative intensity of the devices.
shoulder in the longer-wavelength side of the spectra (ca.
We conclude that the fine modifications of the molecular
650 nm), originating from intermolecular interactions within structure of the acceptors have a strong influence on their
polymeric chains, gradually increases passing from PDI-SF-nd, physical properties, such as solubility, super-saturation thermo-
to PDI-SF-d to PDI-SF-eh based blends, suggesting a gradual dynamics, precipitation kinetics and miscibility with the donor
increase of the corresponding D:A phase segregation. This is material, for example. These properties in turn strongly influence
in perfect agreement with the previous consideration: the the BHJ nanomorphology and consequently the photovoltaic
longer and branched alkyl side chains of PDI-SF-nd partially response of the solar cells.
hinder intermolecular ordering. From the morphological and
UV-Vis data on the blends, we can add that the side groups of
PDI-SF-nd favor its miscibility with the PFQBDT-TR1 polymeric
chains.
Conclusions
In contrast, the short branched alkyl chains of PDI-SF-eh are In this work we studied the structure–property relationship and
not sufficiently long to favor the same donor:acceptor intermixing. the photovoltaic characteristics of three perylene diimide-
Rather, they favor intermolecular ordering both within the spirobifluorene derivatives, N-substituted with different alkyl
acceptor domains and, as a consequence, also within the side groups: 10-nonadecyl (PDI-SF-nd), 2-ethylhexyl (PDI-SF-eh)
PFQBDT-TR1 domains. The (single) long and linear alkyl chain and n-dodecyl (PDI-SF-d). These molecules were synthesized by
of PDI-SF-d favors an intermediate interpenetration of the cost-effective methods, using a straightforward synthesis and
donor:acceptor domains. Of note, the supposed enhanced an easy purification. The optical and electrochemical properties
self-organization passing from PDI-SF-nd, to PDI-SF-d to of the three molecules were very similar. However, the photo-
PDI-SF-eh based blends correlates well with the increase of voltaic properties of the resulting blade-coated BHJ solar cells,
the corresponding VOC, which is known to strongly depend on based on their BHJ blends with a benchmark quinoxaline–
nanostructural factors of the active layer.28 The enhanced benzodithiophene donor polymer (PFQBDT-TR1), were different.
nanoscale organization of the active blends (PDI-SF-nd o In detail, the highest PCE (1.5%) was achieved for the PDI-SF-nd
PDI-SF-d o PDI-SF-eh) determines a reduction of the PCE of derivative having branched and longer alkyl groups in N-positions
the corresponding devices (PDI-SF-nd 4 PDI-SF-d 4 PDI-SF-eh). (10-nonadecyl), while the lowest PCE (0.5%) was achieved for
This could be explained by the different n and/or p charge the PDI-SF-eh derivative, bearing branched and short alkyl
mobilities of the PDI-SF-nd, PDI-SF-eh and PDI-SF-d based BHJ chains in N-positions (2-ethylhexyl). We demonstrated that
blends. In our previous studies on PDI-spirobifluorene based the different photovoltaic responses of the three molecules
molecules, we demonstrated that the different chemical struc- were due to their different miscibility/self-aggregation within
tures of the acceptors did not have any influence on their electron the BHJ blend, as a function of length and conformation of
mobility in pure films (10À5 cm2 VÀ1 sÀ1).20 Rather, we observed a the alkyl side chains. This was reflected in a different nano-
reduction of the electron mobility of the PDI-SF molecules when morphology of the active layers, with a donor:acceptor phase
blended with P3HT, in particular those blends showing a more segregation increasing when passing from long branched
segregated/coarser nanomorphology. Therefore, unbalanced elec- alkyl chains (PDI-SF-nd) to long linear ones (PDI-SF-d) to short
tron and hole mobility caused non-geminate recombinations that branched ones (PDI-SF-eh). The increasing phase segregation of
drastically limited the JSC, FF and the PCE of the corresponding the blends, accompanied by a reduction of the electron mobility,
devices. In this work, the phase segregation/nanostructuring was probably the origin of internal resistances and recombination
increases with a trend which reflects the JSC, the FF and the losses, which limited the photovoltaic response of the corres-
PCE in the corresponding solar cells. Similarly to our previous ponding solar cells.
work, we verified that also in this case the increase in the phase
In conclusion, this work provides a contribution to the
segregation is accompanied by a slight decrease of the electron development of design rules for the engineering of the side
mobility of the blend films. We measured the electron mobility chains in the N-positions of PDI-based molecular acceptors,
using electron-only devices based on PFQBDT-TR1:PDI-SF-nd, through easy synthetic methods, towards efficient and scalable
PFQBDT-TR1:PDI-SF-d and PFQBDT-TR1:PDI-SF-eh blends (Fig. S2, non-fullerene-based BHJ organic solar cells.
ESI†). The resulting mobilities were: B6 Â 10À5 cm2 VÀ1 sÀ1
,
3 Â 10À5 cm2 VÀ1 sÀ1 and 2 Â 10À5 cm2 VÀ1 sÀ1 respectively for
PFQBDT-TR1:PDI-SF-nd, PFQBDT-TR1:PDI-SF-d and PFQBDT-
TR1:PDI-SF-eh films. The higher electron mobility, favored by the
Conflicts of interest
longer and branched alkyl side chains of PDI-SF-nd, which likely There are no conflicts to declare.
18638 | New J. Chem., 2018, 42, 18633--18640 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
Energy Mater., 2017, 1700465; (b) A. Facchetti, Chem. Mater.,
2011, 23, 733.
Acknowledgements
The authors thank Flinders University and the Australian
Research Council (DP160102356) for financial support. The
authors also wish to thank Vincenzo Ragona (CNR-ISMN) and
Federico Prescimone (CNR-ISMN) for the precious technical
support.
9 (a) B. M. Savoie, A. Rao, A. A. Bakulin, S. Gelinas,
B. Movaghar, R. H. Friend, T. J. Marks and M. A. Ratner,
J. Am. Chem. Soc., 2014, 136, 2876; (b) R. D. Pensack, C. Guo,
K. Vakhshouri, E. D. Gomez and J. B. Asbury, J. Phys. Chem.
C, 2012, 116, 4824; (c) S. R. Cowan, N. Banerji, W. L. Leong
and A. J. Heeger, Adv. Funct. Mater., 2012, 22, 1116; (d) B. A.
Gregg, J. Phys. Chem. Lett., 2011, 2, 3013; (e) J. Roncali,
P. Leriche and A. Cravino, Adv. Mater., 2007, 19, 2045;
( f ) Y. Kanai and J. C. Grossman, Nano Lett., 2007, 7, 1967.
10 (a) M. Muccini, R. Danieli, R. Zamboni, C. Taliani, H. Mohn,
W. Mu¨ller and H.U. Ter Meer, Chem. Phys. Lett., 1995,
245, 107; (b) H. Schlaich, M. Muccini, J. Feldmann,
References
1 (a) G. Li, R. Zhu and Y. Yang, Nat. Photonics, 2012, 6, 153;
(b) F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394;
(c) R. R. Søndergaard, M. Hosel and F. C. Krebs, J. Polym.
Sci., Part B: Polym. Phys., 2013, 51, 16.
¨
¨
¨
H. Bassler, E. O. Gobel, R. Zamboni, C. Taliani, J. Erxmeyer
and A. Weidinger, Chem. Phys. Lett., 1995, 236, 135; (c) Y. Lin
and X. Zhan, Mater. Horiz., 2014, 1, 470; (d) H. Li,
T. Earmme, G. Ren, A. Saeki, S. Yoshikawa, N. M. Murari,
S. Subramaniyan, M. J. Crane, S. Seki and S. A. Jenekhe, J. Am.
Chem. Soc., 2014, 136, 14589; (e) D. Rodriquez, S. Savagatrup,
E. Valle, C. M. Proctor, C. McDowell, G. C. Bazan, T. Q. Nguyen
and D. J. Lipomi, ACS Appl. Mater. Interfaces, 2016, 8, 11649;
( f ) D. J. Lipomi, Adv. Mater., 2016, 28, 4180; (g) S. Savagatrup,
A. D. Printz, T. F. O’Connor, A. V. Zaretski, D. Rodriquez,
E. J. Sawyer, K. M. Rajan, R. I. Acosta, S. E. Root and
D. J. Lipomi, Energy Environ. Sci., 2015, 8, 55; (h) Y. He and
Y. Li, Phys. Chem. Chem. Phys., 2011, 13, 1970.
2 (a) R. Søndergaard, M. Hosel, D. Angmo, T. T. Larsen-Olsen
´
and F. C. Krebs, Mater. Today, 2012, 15, 36; (b) I. Burgues-
`
Ceballos, M. Stella, P. Lacharmoise and E. Martınez-Ferrero,
J. Mater. Chem. A, 2014, 2, 17711.
3 C.-C. Chueh, K. Yao, H.-L. Yip, C.-Y. Chang, Y.-X. Xu,
K.-S. Chen, C.-Z. Li, P. Liu, F. Huang, Y. Chen, W.-C. Chen
and A. K.-Y. Jen, Energy Environ. Sci., 2013, 6, 3241.
4 (a) L. Dou, J. You, Z. Hong, Z. Xu, G. Li, R. A. Street and
Y. Yang, Adv. Mater., 2013, 25, 6642; (b) H. Yip and A. K.-Y.
Yen, Energy Environ. Sci., 2012, 5, 5994; (c) M. Seri,
M. Bolognesi, Z. Chen, S. Lu, W. Koopman, A. Facchetti and
M. Muccini, Macromolecules, 2013, 46, 6419; (d) D. Gedefaw,
M. Tessarolo, M. Prosa, M. Bolognesi, P. Henriksson,
W. Zhuang, M. Seri, M. Muccini and M. R. Andersson,
Sol. Energy Mater. Sol. Cells, 2016, 144, 150.
5 (a) M. T. Dang, L. Hirsch, G. Wantz and J. D. Wuest, Chem.
Rev., 2013, 113, 3734; (b) F. Zhao, C. Wang and X. Zhan,
Adv. Energy Mater., 2018, 1703147; (c) M. Bolognesi,
M. Prosa, M. Tessarolo, G. Donati, S. Toffanin, M. Muccini
and M. Seri, Sol. Energy Mater. Sol. Cells, 2016, 155, 436.
6 (a) Y.-S. Cheng and S.-A. Chen, Adv. Mater., 2013, 25, 4766;
(b) N. Zhou, X. Guo, R. Ponce Ortiz, S. Li, S. Zhang,
R. P. H. Chang, A. Facchetti and T. J. Marks, Adv. Mater.,
2012, 24, 2242; (c) D. Gedefaw, M. Tessarolo, M. Bolognesi,
11 (a) J. Zhang, H. S. Tan, X. Guo, A. Facchetti and H. Yan, Nat.
Energy, 2018, 3, 720; (b) W. Zhao, S. Li, H. Yao, S. Zhang,
Y. Zhang, B. Yang and J. Hou, J. Am. Chem. Soc., 2017,
139, 7148.
12 C. Huang, S. Barlow and S. R. Marder, J. Org. Chem., 2011,
76, 2386.
13 Y. Zhong, M. T. Trinh, R. Chen, G. E. Purdum, P. P.
Khlyabich, M. Sezen, S. Oh, H. Zhu, B. Fowler, B. Zhang,
W. Wang, C.-Y. Nam, M. Y. Sfeir, C. T. Black, M. L.
Steigerwald, Y.-L. Loo, F. Ng, X.-Y. Zhu and C. Nuckolls,
Nat. Commun., 2015, 6, 8242.
M. Prosa, R. Kroon, W. Zhuang, P. Henriksson, K. Bini, 14 S. Chen, G. Zhang, J. Liu, H. Yao, J. Zhang, T. Ma, Z. Li and
E. Wang, M. Muccini, M. Seri and M. R. Andersson, Beilstein
H. Yan, Adv. Mater., 2017, 29, 1604231.
J. Org. Chem., 2016, 12, 1629; (d) M. Seri, D. Gedefaw, 15 J. L. Segura, H. Herrera and P. Bauerle, J. Mater. Chem., 2012,
M. Prosa, M. Tessarolo, M. Bolognesi, M. Muccini and
22, 8717.
M. R. Andersson, J. Polym. Sci., Part A: Polym. Chem., 2017, 16 Z. Chen, U. Baumeister, C. Tschierske and F. Wu
55(2), 234; (e) D. Gedefaw, M. Prosa, M. Bolognesi, M. Seri
Chem. – Eur. J., 2007, 13, 450.
and M. R. Andersson, Adv. Energy Mater., 2017, 7, 1700575; 17 (a) F. Wurthner, Chem. Commun., 2004, 1564; (b) N. V.
¨
¨rtner,
( f ) M. Prosa, A. Sagnella, T. Posati, M. Tessarolo,
M. Bolognesi, S. Cavallini, S. Toffanin, V. Benfenati,
Handa, K. D. Mendoza and L. D. Shirtcliff, Org. Lett.,
2011, 17, 4724.
M. Seri, G. Ruani, M. Muccini and R. Zamboni, RSC Adv., 18 F. Zhang, Y. Ma, Y. Chi, H. Yu, Y. Li, T. Jiang, X. Wei and
2014, 4, 44815.
J. Shi, Sci. Rep., 2018, 8, 8208.
7 (a) C. L. Cutting, M. Bag and D. Venkataraman, J. Mater. 19 T. Ye, R. Singh, H.-J. Butt, G. Floudas and P. E. Keivanidis,
Chem. C, 2016, 4, 10367; (b) F. C. Krebs, N. Espinosa, ACS Appl. Mater. Interfaces, 2013, 5, 11844.
M. Hosel, R. R. Søndergaard and M. Jørgensen, Adv. Mater., 20 (a) D. Kotowski, S. Luzzati, G. Scavia, M. Cavazzini, A. Bossi,
2014, 26, 29; (c) N. Espinosa, M. Hosel, M. Jørgensen and
F. C. Krebs, Energy Environ. Sci., 2014, 7, 855.
8 (a) J. Min, Y. N. Luponosov, C. Cui, B. Kan, H. Chen, X. Wan,
Y. Chen, S. A. Ponomarenko, Y. Li and C. J. Brabec, Adv.
M. Catellani and E. Kozma, Dyes Pigm., 2015, 120, 57;
(b) E. Kozma, D. Kotowski, M. Catellani, S. Luzzati,
M. Cavazzini, A. Bossi, S. Orlandi and F. Bertini, Mater.
Chem. Phys., 2015, 163, 152.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 New J. Chem., 2018, 42, 18633--18640 | 18639

View Article Online
Paper
NJC
21 M. Tessarolo, D. Gedefaw, M. Bolognesi, F. Liscio,
P. Henriksson, W. Zhuang, S. Milita, M. Muccini, E. Wang,
M. Seri and M. R. Andersson, J. Mater. Chem. A, 2014, 2, 11162.
G. Yang, J. Zhao, J. Bergqvist, F. Zhang, W. Ma, H. Ade,
¨
O. Inganas, K. Gundogdu, F. Gao and H. Yan, Nat. Energy,
2016, 1, 16089.
´
22 V. D. Mihailetchi, J. Wildeman and P. W. M. Blom, Phys. 26 I. Burgues-Ceballos, M. Stella, P. Lacharmoise and
´
Rev. Lett., 2005, 94, 126602. E. Martınez-Ferrero, J. Mater. Chem. A, 2014, 2, 17711.
23 C. B. Nielsen, S. Holliday, H.-Y. Chen, S. J. Cryer and 27 (a) N. K. Elumalai and A. Uddin, Energy Environ. Sci., 2016,
I. McCulloch, Acc. Chem. Res., 2015, 48, 2803.
24 E. Kozma and M. Catellani, Dyes Pigm., 2013, 98, 160.
9, 391–410; (b) D. M. Stevens, J. C. Speros, M. A. Hillmyer
and C. D. Frisbie, J. Phys. Chem. C, 2011, 115, 20806.
¨
25 (a) J. Hou, O. Inganas, R. H. Friend and F. Gao, Nat. Mater., 28 K. Vandewal, S. Himmelberger and A. Salleo, Macromolecules,
2017, 17, 119; (b) J. Liu, S. Chen, D. Qian, B. Gautam, 2013, 46, 6379.
18640 | New J. Chem., 2018, 42, 18633--18640 This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018