Synthesis, characterization and photovoltaic properties of azadipyrromethene-based acceptors: Effect of pyrrolic substituents

Article abstract of DOI:10.1039/c4ta05765a

Azadipyrromethene derivatives are conjugated molecules with high absorptivity in the visible to near-IR region and high electron affinity that have great potential as electron acceptors for solar harvesting applications. To fully take advantage of these m

Full text of DOI:10.1039/c4ta05765a

Journal of
Materials Chemistry A
View Article Online
View Journal
PAPER
Synthesis, characterization and photovoltaic
properties of azadipyrromethene-based acceptors:
Cite this: DOI: 10.1039/c4ta05765a
effect of pyrrolic substituents†
Wasana Senevirathna, Jia-yu Liao, Zhenghao Mao, Jun Gu, Matthew Porter,
*
Chunlai Wang, Roshan Fernando and Genevieve Sauve
`
´
Azadipyrromethene derivatives are conjugated molecules with high absorptivity in the visible to near-IR
region and high electron affinity that have great potential as electron acceptors for solar harvesting
applications. To fully take advantage of these molecules, it is necessary to understand their structure–
property relationships. We recently showed that phenylethynyl pyrrolic substituents red-shifted the
visible absorption band, increased the electron affinity, and significantly improved the device
performance in organic photovoltaic devices. Here, we synthesized and characterized a series of
azadipyrromethene derivatives and their chelates, where the pyrrolic substituents were chosen to
examine the effect of ethynyl group, aryl group and alkyl solubilizing group. Using thienylethynyl or
alkoxyphenylethynyl substituents caused a small red-shift of the absorption compared to phenylethynyl
substituents. The zinc(II) complexes were blended with regioregular poly(3-hexylthiophene) as the
electron donor and tested in organic solar cells. All fullerene-free solar cells studied showed good
photovoltaic properties, with power conversion efficiencies ranging from 2% to 4%. The blend films all
had similar AFM images with no evidence of large-scale segregation. When comparing solubilizing
groups on the pyrrolic substituents, we find that the performance varied as follows: H > t-butyl > 2-
ethylhexyl. These results pave the way for developing higher performance non-fullerene acceptors for
organic photovoltaic devices.
Received 27th October 2014
Accepted 23rd December 2014
DOI: 10.1039/c4ta05765a
www.rsc.org/MaterialsA
limit achievable efficiencies for OPVs.7 Recent efforts
demonstrated that non-fullerene acceptors can be used to
overcome some of these disadvantages.11–23 Most successful
non-fullerene organic materials include sub-phthalocya-
nines, rylene diimide-based molecules and polymers, as well
as molecules and conjugated polymers incorporating elec-
tron-withdrawing groups such as imides, cyano, uorine and
benzothiadiazoles.21,24–29
Introduction
Organic photovoltaic devices (OPVs) hold great promise for
providing sustainable, low cost electrical energy.1,2 In
particular, solution-processed bulk heterojunction (BHJ)
OPVs can be produced on a large scale using current printing
technology.3,4 Power conversion efficiencies have climbed
rapidly over the past 15 years, reaching ꢀ10% for single layer
devices.5,6 The active layer of BHJ OPVs typically consists of a
conjugated polymer (or molecule) electron donor blended
with a fullerene derivative as the electron acceptor. Fullerene
derivatives have been the electron acceptor of choice because
they work well: fullerenes can accept several electrons
reversibly, have good electron transport properties,7 have a
large spherical shape,8,9 and can form favorable nano-scale
phase separation with typical electron donor materials.10
However, fullerenes do not signicantly contribute to light
harvesting at wavelengths greater than 600 nm and their
energy levels cannot be easily tuned, factors that seriously
One overlooked candidate for designing non-fullerene
acceptors is azadipyrromethene (Fig. 1). Azadipyrrome-
thenes are bidentate ligands with strong visible absorption,
tunable properties through substitution and chelation, and
+
high electron affinity.30–32 Their BF2 chelates have been
investigated for photodynamic therapy applications,33 as
well as donors for organic solar cells and acceptors in
+
molecular triads.34 The BF2 chelate of related dipyrrome-
thene dyes has also been demonstrated as an acceptor in
OPVs.35 We recently demonstrated that the homoleptic zin-
c(II) complex of di(phenylethynyl) tetraphenylazadipyrro-
methene is an effective electron acceptor for organic solar
cells, with a power conversion efficiency of as high as 4.1%
when blended with the common donor poly(3-hexylth-
iophene) (P3HT).36,37 Interestingly, the power conversion
efficiency of this acceptor was three times higher than when
Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106,
USA. E-mail: genevieve.sauve@case.edu
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c4ta05765a
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Experimental
Materials and instrumentation
Solvents and reagents were purchased from commercial
suppliers, Fischer or Aldrich, and used without further puri-
cation unless indicated. Azadipyrromethene (H(ADP)) and
iodinated H(ADP) (ADPI2) were synthesized according to pub-
1
lished methods. H NMR, and 19F NMR spectra were recorded
on a Varian AS-400 spectrometer. Chemical shis (1H) are
reported in parts-per-million relative to Si(CH3)4. Elemental
analysis (C, H, and N) was performed using the optimum
combustion conditions by Robertson Microlit Laboratories. UV-
Vis spectra were collected on a Cary 500 spectrophotometer in
HPLC grade chloroform or on a 1 cm wide glass plate. MALDI-
TOF MS spectra were acquired in reective negative mode on a
Bruker autoex III smartbeam MALDI-TOF TOF spectrometer.
Thermal gravimetric analysis (TGA) was performed on a TA
instrument Q500 thermogravimetric analyzer.
Fig. 1 Series of ligands with various pyrrolic substituents explored in
this report.
we used a Zn(II) complex of unmodied tetraphenylazadi-
pyrromethene as the acceptor. The better acceptor had
additional phenylethynyl ‘conjugated arms’ at the pyrrolic
positions (Fig. 1), which not only extended the conjugation
system in different directions, increased electron affinity
and lowered the optical energy gap, but also prevented
Synthesis
Tributyl(4-tert-butylphenylethynyl)tin (A1). 4-tert-Butylphenylacety-
lene (3.5 g, 22.1 mmol) was added into a 250 mL Schlenk ask.
The ask was evacuated and relled (ꢁ3) with N2. Then 20 mL of
dry diethyl ether was added and the solution was degassed for 5
min. The ask was cooled using a CHCl3/dry ice bath. n-BuLi
(8.85 mL, 22.1 mmol) was added dropwise via a syringe. Aer
ꢀ10–15 min, the cold bath was removed and the reaction mixture
was allowed to reach room temperature. The solution became
pale brown. Aer 30 min, the reaction mixture was cooled using
an acetonitrile/dry ice bath. Bu3SnCl (6.0 mL, 22.1 mmol) was
added dropwise via a syringe. Aer ꢀ10 min, the cold bath was
removed and the reaction mixture was allowed to reach room
temperature. Upon warming, a white suspension was formed.
Aer stirring at room temperature for ꢀ1 h, the mixture was dried
by rotary evaporation. The oily residue was dissolved in hexanes
and vacuum ltered. The ltrate was taken and the solvent was
removed by vacuum evaporation. A yellow oil was obtained (A1).
Yield: 9.6 g, 98%. 1H NMR (400 MHz, CDCl3, d): 7.36–7.38 (m, 2H),
7.27–7.29 (m, 2H), 1.55–1.65 (m, 6H), 1.28 (s, 9H), 1.32–1.40 (m,
6H), 1.01–1.05 (t, J ¼ 8 Hz, 6H), 0.86–0.92 (t, J ¼ 4 Hz, 9H).
(5-(2-ethylhexyl-2-thiophene)trimethyl)tin (A2). This was
synthesized according to a literature procedure.40 A pale yellow
crystallization and promoted
a
favorable nanoscale
morphology in blends with P3HT. The idea of breaking
planarity to decrease aggregation and obtain better nano-
scale morphology in blends has also been reported for per-
ylene diimides-based acceptors and other systems.20,23,38,39
With our metal complexes, we aim to not only design non-
planar structures without impacting conjugation, but also
obtain isotropic charge transport by having conjugated arms
pointing in several directions.
Here, we performed a structure–property study of azadi-
pyrromethene ligands and their complexes, focusing on the
impact of the pyrrolic substituents on optical, electro-
chemical and photovoltaic properties. We synthesized a
series of ligands (WS1–5, Fig. 1) designed to examine the
effect of ethynyl, aryl and alkyl groups in the pyrrolic
substituents. We introduced thiophene in the pyrrolic
substituents either directly on the ADP core (WS2) or via an
ethynyl spacer (WS5) to understand the effect of ethynyl
spacer on the properties. Also, thiophene groups are electron
rich and should contribute to tuning the energy levels and
push the absorption to the near infrared region. We also
compared phenylethynyl pyrrolic substituents with various
alkyl substitutions at the para-position of the phenyl rings:
no alkyl group (WS3), t-butyl solubilizing group (WS1) and
electron rich solubilizing group 2-ethylhexyloxy (WS4). The
1
oil was obtained (A2). Yield: 5.04 g. 71%. H NMR (400 MHz,
CDCl3, d): 6.82–6.84 (d, J ¼ 8 Hz, 1H), 6.49–6.50 (d, J ¼ 4 Hz, 1H),
2.63–2.67 (t, J ¼ 8 Hz, 2H), 1.50 (m, 1H), 1.21–1.32 (m, 8H), 0.80–
0.89 (m, 9H).
Tributyl(phenylethynyl)tin (A3). The same procedure was used
as that for A1, using ethynylbenzene. A yellow oil was obtained
(A3). Yield: 18.5 g, 97%. 1H NMR (400 MHz, CDCl3, d): 7.43–7.46
(m, 2H), 7.25–7.28 (m, 3H), 1.55–1.64 (m, 6H), 1.33–1.42 (m,
6H), 1.04–1.08 (t, J ¼ 8 Hz, 6H), 0.91–0.94 (t, J ¼ 4 Hz, 9H).
Tributyl(4-(2-ethylhexyloxy)phenylethynyl)tin (A4). The same
procedure was used as that for A1, using 1-((2-ethylhexyl)oxy)-4-
ethynylbenzene (iv) (synthesis of iv can be found in ESI†).
+
BF2 and Zn(II) chelates for all ligands were synthesized and
characterized. We found that the Zn(II) chelates all performed
well as acceptors in organic photovoltaic devices, with power
conversion efficiencies ranging from 2 to 4%. The best
performance was obtained with the phenylethynyl pyrrolic
substituents that have no solubilizing groups.
1
A yellow oil was obtained (A4). Yield: 10.0 g, ꢀ99%. H NMR
(400 MHz, CDCl3, d): 7.35–7.37 (m, 2H), 6.77–6.79 (m, 2H), 3.80–
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Journal of Materials Chemistry A
+
3.81 (m, 2H), 1.68–1.74 (m, 1H), 1.32–1.44 (m, 10H), 1.26–1.31
(m, 4H), 1.56–1.64 (m, 6H), 1.01–1.05 (m, 6H), 0.86–0.96 (m,
15H).
Synthesis of BF2 chelates, BF2(WS1–5). BF2(WS1) was
synthesized using the literature procedure.37
BF2(WS2). The same procedure was used as for BF2(WS1),
using H(WS2). Yield: 88 mg, 83%. 1H NMR (400 MHz, CDCl3, d)
7.55–7.59 (d, J ¼ 8 Hz, 4H), 7.50–7.52 (d, J ¼ 8 Hz, 4H), 7.25–7.34
(m, 12H), 6.48–6.49 (d, J ¼ 4 Hz, 2H), 6.40–6.41 (d, J ¼ 4 Hz, 2H),
2.59–2.61 (d, J ¼ 8 Hz, 4H), 1.38–1.44 (m, 2H), 1.17–1.25 (m,
16H), 0.77–0.86 (m, 12H). 19F NMR (400 MHz, CDCl3, d) 132.5–
132.8 (q). MALDI-TOF MS: m/z calcd for C56H58BF2N3S2 886.02,
found 884.46. Elem. anal. calcd: C, 75.91; H, 6.60; N, 4.74;
found: C, 76.11; H, 6.89; N, 4.53.
Tributyl((5-(2-ethylhexyl)thiophen-2-yl)ethynyl)tin (A5). The
same procedure was used as that for A1, using 2-(2-ethylhexyl)-5-
ethynylthiophene (vii) (synthesis of vii can be found in ESI†). A
1
yellow oil was obtained (A5). Yield: 8.4 g, 97%. H NMR (400
MHz, CDCl3, d): 6.98–6.99 (d, J ¼ 4 Hz, 1H), 6.55–6.56 (d, J ¼ 4
Hz, 1H), 2.67–2.68 (d, J ¼ 4 Hz, 2H), 1.54–1.57 (m, 1H), 1.30–1.37
(m, 26H), 0.85–0.92 (m, 15H).
Synthesis of free ligands, H(WS1–5)
H(WS1). Compound A1 (0.581 g, 1.311 mmol) and ADPI2
(0.307 g, 0.437 mmol) were charged into a dry 50 mL ask and
put under vacuum for ꢀ10 min. The ask was evacuated and
relled (ꢁ3) with N2. Dry m-xylene (50 mL) was added via a
syringe and stirred for ꢀ5 min under inert conditions. The
Pd(PPh3)4 (50 mg, 10% mmol) catalyst was added inside a glove
box. The reaction mixture was heated to 125 ꢂC under N2 for 36–
48 h. Upon heating, the solution color changed from purple to
deep blue. The solvent was removed by rotary evaporation and
the product was triturated with cold methanol (10 mL) and
washed with cold ether (10 mL ꢁ 3), until a pale blue coloration
appeared. The crude product was puried by column chroma-
tography with silica gel and dichloromethane (DCM). Yield: 238
mg, 71%. 1H NMR (400 MHz, CDCl3, d): 8.31–8.33 (d, J ¼ 8.0 Hz,
4H), 8.16–8.19 (d, J ¼ 8.0 Hz, 4H), 7.43–7.52 (m, 6H), 7.35–7.42
(m, 14H), 1.31 (s, 18H). MALDI-TOF MS: m/z calcd for C56H47N3
761.38, found 760.38.
BF2(WS3). The same procedure was used as for BF2(WS1),
using H(WS3). Yield: 63 mg, 49%. 1H NMR (400 MHz, CDCl3,
d): 8.24–8.27 (d, J ¼ 8 Hz, 4H), 8.02–8.04 (d, J ¼ 8 Hz, 4H),
7.49–7.53 (m, 12H), 7.29–7.30 (m, 10H). MALDI-TOF MS: m/z
calcd for C48H30BF2N3 697.25, found 696.23. Elem. anal.
calcd: C, 82.6 4; H, 4.33; N, 6.02, found: C, 82.38; H, 4.61;
N, 5.92.
BF2(WS4). The same procedure was used as for BF2(WS1),
using H(WS4). Yield: 86 mg, 68%. 1H NMR (400 MHz, CDCl3, d):
8.22–8.24 (d, J ¼ 8 Hz, 4H), 7.98–8.00 (d, J ¼ 8 Hz, 4H), 7.42–7.47
(m, 12H), 7.18–7.20 (d, J ¼ 8 Hz, 4H), 6.76–6.78 (d, J ¼ 8 Hz, 4H),
3.77–3.79 (d, J ¼ 8 Hz, 4H), 1.67 (m, 2H), 1.35–1.40 (m, 8H),
1.25–1.27 (m, 8H), 0.83–0.89 (m, 12H). 19F NMR (400 MHz,
CDCl3, d): 131.6–131.8 (q). MALDI-TOF MS: m/z calcd for
C
64H62BF2N3O2 953.49, found 952.22. Elem. anal. calcd: C,
80.57; H, 6.55; N, 4.40, found: C, 80.34; H, 6.55; N, 4.33.
BF2(WS5). The same procedure was used as for BF2(WS1),
using H(WS5). Yield: 51 mg 77%. 1H NMR (400 MHz, CDCl3, d):
8.20–8.22 (d, J ¼ 8 Hz, 4H), 7.98–8.00 (d, J ¼ 8 Hz, 4H), 7.42–7.49
(m, 12H), 6.92–6.93 (d, J ¼ 4 Hz, 4H), 6.59–6.60 (d, J ¼ 4 Hz, 4H),
2.68–2.69 (d, J ¼ 4 Hz 4H), 1.53 (m, 2H), 1.25–1.31 (m, 16H),
0.83–0.87 (m, 12H). 19F NMR (400 MHz, CDCl3, d): 131.4–131.7
(q). MALDI-TOF MS: m/z calcd for C60H58BF2N3S2 933.41, found
932.22. Elem. anal. calcd: C, 77.15; H, 6.26; N, 4.50, found: C,
77.29; H, 6.37; N, 4.47.
H(WS2). The same procedure was used as that for H(WS1),
1
using A2 and ADPI2. Yield: 262 mg, 88%. H NMR (400 MHz,
CDCl3, d): 7.53–7.58 (m, 8H), 7.32–7.37 (m, 6H), 7.20–7.21 (m,
6H), 6.64 (s, 4H), 2.69–2.71 (d, J ¼ 8.0 Hz, 4H), 1.50 (m, 2H),
1.25–1.32 (m, 16H), 0.84–0.87 (m, 12H). MALDI-TOF MS: m/z
calcd for C56H59N3S2 837.42, found 836.39. Elem. anal. calcd: C,
80.24; H, 7.09; N, 5.01; S, 7.65, found: C, 80.03; H, 7.18; N, 4.88;
S, 7.44.
H(WS3). The same procedure was used as that for H(WS1),
using A3 and ADPI2. Yield: 135 mg, 28%. 1H NMR was not
available due to poor solubility. MALDI-TOF MS: m/z calcd for
Synthesis of Zn(II) complexes, Zn(WS1–5)2. Zn(WS1)2 was
synthesized using the literature procedure.37
Zn(WS2)2. The same procedure was used as for Zn(WS1)2,
using H(WS2). Synthesis involved the reaction between H(WS2)
(0.404 g, 0.482 mmol) and zinc acetate [Zn(OAc)2] (0.115 g, 0.628
mmol) reux in 50 mL of 1-butanol at 125 ꢂC for 2 h. The solvent
was evaporated and the crude was puried by column chro-
matography on silica gel with 1 : 1 DCM/hexanes. The solvent
was evaporated and the solid residue was dissolved in a
minimum amount of DCM. The dark blue solid was isolated by
slow evaporation of DCM. Yield: 83 mg, 92%. 1H NMR (400
MHz, CDCl3, d): 7.31–7.33 (m, 8H), 7.18–7.20 (m, 12H), 7.00–
7.13 (m, 20H), 6.51 (d, J ¼ 4 Hz, 4H), 6.38 (d, J ¼ 4 Hz, 4H), 2.62–
2.64 (m, 8H), 1.45–1.47 (m, 4H), 1.22–1.30 (m, 32H), 0.81–0.85
(m, 24H). MALDI-TOF MS: m/z calcd for C112H116N6S4Zn
1737.75, found 1738.63. Elem. anal. calcd: C, 77.32; H, 6.72; N,
4.83; found: C, 77.34; H, 6.82; N, 4.75.
C
48H31N3 649.25, found 648.18. Elem. anal. calcd: C, 88.72; H,
4.81; N, 6.47, found: C, 87.80; H, 5.19; N, 6.23.
H(WS4). The same procedure was used as that for H(WS1),
1
using A4 and ADPI2. Yield: 340 mg, 66%. H NMR (400 MHz,
CDCl3, d): 8.30–8.32 (d, J ¼ 8 Hz, 4H), 8.18–8.20 (d, J ¼ 8 Hz, 4H),
7.50–7.56 (m, 6H), 7.37–7.40 (m, 12H), 6.86–6.88 (d, J ¼ 8.0 Hz,
4H), 3.83–3.85 (dd, J ¼ 8.0 Hz 4H), 1.72 (m, 2H), 1.43–1.50 (m,
8H), 1.30–1.32 (m, 8H), 0.88–0.92 (m, 9H). MALDI-TOF MS: m/z
calcd for C64H63N3O2 905.49, found 904.32. Elem. anal. calcd: C,
84.82; H, 7.01; N, 4.64, found: C, 84.78; H, 6.86; N, 4.61.
H(WS5). The same procedure was used as that for H(WS1), using
A5 and ADPI2. Yield: 173 mg, 65%.1H NMR (400 MHz, CDCl3, d):
8.25–8.29 (d, J ¼ 8 Hz, 4H), 8.15–8.17 (d, J ¼ 8 and 4 Hz, 4H), 7.49–
7.58 (m, 6H), 7.35–7.47 (m, 6H), 7.06–7.07 (d, J ¼ 4.0 Hz, 2H), 6.67–
6.68 (d, J ¼ 4.0 Hz, 2H), 2.69–2.71 (d, J ¼ 8.0 Hz, 4H), 1.47–1.65 (m,
2H), 1.18–1.42 (m, 16H), 0.74–0.99 (m, 12H). MALDI-TOF MS: m/z
calcd for C60H59N3S2. 885.42, found 884.30. Elem. anal. calcd: C,
81.31; H, 6.71; N, 4.74, found: C, 81.11; H, 7.00; N, 4.71.
Zn(WS3)2. The same procedure was used as for Zn(WS2)2,
1
using H(WS3). Yield: 176 mg, 54%. H NMR (400 MHz, CDCl3,
d): 7.97–7.99 (m, 8H), 7.78–7.76 (m, 8H), 7.45–7.46 (m, 12H),
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
7.35–7.37 (m, 8H), 7.27–7.29 (m, 12H), 7.21–7.23 (m, 12H). 20 mg mLꢃ1) in o-dichlorobenzene was prepared. The ZnO layer
MALDI-TOF MS: m/z calcd for C96H60N6Zn 1361.42, found was spin-coated on a pre-cleaned ITO substrate at 4000 rpm
1360.77. Elem. anal. calcd: C, 84.60; H, 4.44; N, 6.17, found: C, from a ZnO precursor solution of zinc acetate dihydrate (0.25 M)
84.42; H, 4.36; N, 6.08.
and monoethanolamine (0.25 M) in 2-methoxyethanol (solvent),
Zn(WS4)2. The same procedure was used as for Zn(WS2)2, and the substrate was subsequently annealed at 150 ꢂC for
1
using H(WS4). Yield: 123 mg, 96%. H NMR (400 MHz, CDCl3, 5 min. The photoactive layer of P3HT : Zn(WS1–5)2 (ꢀ90 nm)
d): 7.98–8.00 (m, 8H), 7.75–7.77 (m, 8H), 7.43–7.45 (m, 14H), was prepared by spin-coating the blend solution at 1000 rpm for
7.26–7.28 (m, 8H), 7.19–7.21 (m, 10H), 6.80–6.82 (m, 8H), 3.81– 40 s and then 2000 rpm for 2 s, followed by annealing at 120 ꢂC
3.82 (d, J ¼ 4 Hz, 8H), 1.68–1.73 (m, 4H), 1.36–1.48 (m, 16H), for 30 min. Finally, the process was completed by depositing
1.28–1.30 (t, J ¼ 4 Hz, 16H), 0.88–0.92 (t, J ¼ 8 Hz 24H). MALDI- MoO3 (10 nm) and Ag (80 nm) in a sequence under a vacuum
TOF MS: C128H124N6O4Zn m/z calcd 1873.90, found 1872.60. pressure of 2 ꢁ 10ꢃ6 Torr using an Angstrom Engineering Evo-
Elem. anal. calcd: C, 81.96; H, 6.66; N, 4.48, found C, 82.26; H, vac Deposition System. The measurements of solar cells were
6.89; N, 4.43.
carried out in the glove box (Pure LabHE) under 1.5 AM solar
Zn(WS5)2. The same procedure was used as for Zn(WS2)2, illumination at an intensity of 100 mW cmꢃ2 using an Oriel
1
using H(WS5). Yield: 180 mg, 80%. H NMR (400 MHz, CDCl3, Sol2A solar simulator and a Keithley 2400 IV station. The total
d): 7.92–7.94 (m, 8H), 7.69–7.71 (m, 8H), 7.41–7.43 (m, 12H), effective area of the lm was 0.17 cm2. The IPCE was measured
7.19–7.20 (m, 12H), 6.92–6.93 (d, J ¼ 4 Hz, 4H), 6.59–6.60 (d, J ¼ with a QEX10 Quantum Efficiency Measurement System.
4 Hz, 4H), 2.67–2.69) (d, J ¼ 8 Hz, 8H), 1.51 (m, 4H), 1.26–1.31
(m, 16H), 0.84–0.88 (m, 24H). MALDI-TOF MS: C120H116N6S4Zn
m/z calcd 1833.75, found 1833.72. Elem. anal. calcd: C, 78.51; H,
6.37; N, 4.58, found C, 78.74; H, 6.49; N, 4.43.
Results and discussion
Synthesis
Azadipyrromethene dyes with various pyrrolic substituents
Electrochemistry
(H(WS1–5)) were synthesized in moderate yield by Stille
coupling of iodinated azadipyrromethene (ADPI2) with the
corresponding tributyltin derivative (Scheme 1). In our previous
publication,37 we had synthesized ligand H(WS1) using Sono-
gashira coupling, but purication and isolation of H(WS1) were
difficult due to the presence of signicant amounts of mono-
substituted byproducts. Here, MALDI-TOF MS of the crude
H(WS1) made by Stille coupling showed the desired product as
the dominant peak, with much smaller amounts of mono-
substituted byproducts than with the previous method. The
higher conversion to the di-substituted product could result
from a higher boiling point solvent used (m-xylene,ꢂ125 ꢂC) than
in the previous method (xylene/triethylamine, 70 C). Further-
more, we found that washing with cold methanol and diethyl
ether removed most impurities, including the mono-
substituted byproduct, allowing the use of smaller columns for
the nal purication. These changes combined to increase the
overall yield for H(WS1) from 51% using our previous method to
71% using this method. All pure ligands (H(WS1–5)) were iso-
Cyclic voltammetry experiments were conducted using an Auto-
Lab-PGSTAT 302N, Eco Chemie potentiostat at room tempera-
ture. All glassware was thoroughly cleaned and oven-dried
overnight before use. Dichloromethane (DCM) was distilled
over calcium hydride and stored in a nitrogen glovebox. Ferro-
cene was puried by sublimation. A typical three-electrode
conguration was used, with a glassy carbon electrode as the
working electrode and two platinum wires as counter and
pseudo-reference electrodes. Ferrocene/ferrocenium was used
as an internal standard. The measurements were performed
with a scan rate of 0.1 V sꢃ1 in degassed DCM with tetra-n-
butylammoniumhexauorophosphate (0.1 M TBAPF6) as the
supporting electrolyte. The electrolyte was purged with nitrogen
for ꢀ10 min prior to each measurement.
Film characterization
Samples for the lm absorption were prepared by spin coating
(400 rpm for 1 min), a ltered (0.2 mm PTFE lter) solution of
P3HT (10 mg mLꢃ1) for the neat P3HT lm and solutions of
ADP-based materials (10 mg mLꢃ1) for neat lms, or solutions
of 10 mg mLꢃ1 of P3HT and 10 mg mLꢃ1 acceptor for the 1 : 1 w/
w blend lms. All solutions were prepared using HPLC grade
CHCl3. UV-Vis spectra were collected on a Cary 500 spectro-
photometer. Fluorescence emission spectra were recorded
using a Cary Eclipse uorescence spectrometer. Atomic force
microscopy (AFM) measurements were performed on an Agilent
Technologies 5500 instrument in tapping mode.
1
lated as air stable blue powders, and were characterized by H
NMR spectroscopy, MALDI-TOF MS and elemental analysis.
Ligand H(WS3) could not be puried as well as the other ligands
due to its limited solubility in organic solvents.
+
The BF2 chelates were synthesized by reacting the free
ligands H(WS1–5) with triuoroboron etherate and diisopro-
pylethylamine (Scheme 1). The chelates were puried by liquid–
liquid extraction, followed by column chromatography with
1 : 1 DCM/hexanes to afford air-stable purple-blue powders in
good yield. 19F spectroscopy showed a quartet around ꢀ131–132
ppm, conrming the tetracoordination state of the boron atom.
Homoleptic Zn(II) complexes were synthesized by reacting
Solar cell fabrication
The photovoltaic properties of Zn(WS1–5)2 series were studied H(WS1–5) with zinc acetate in 1-butanol (Scheme 1). All
in a device structure of ITO/ZnO/P3HT : Zn(WS1–5)2/MoO3/Ag complexes were puried by chromatography and characterized
1
(ITO ¼ indium tin oxide, R ¼ 15 U sqꢃ1). A blend solution by MALDI-TOF MS, H NMR, and elemental analysis. MALDI-
P3HT : Zn(WS1–5)2 (w/w with
a
total concentration of TOF MS conrmed the 1 : 2 stoichiometry ratio of metal to
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Journal of Materials Chemistry A
Scheme 1 Synthesis of azadipyrromethene dyes with various pyrrolic substituents.
ligand for all complexes, Zn(WS1–5)2. The thermal properties of complexes, ꢀ0.14 V to 0.26 V, indicating that the second
ADP-based molecules were examined by thermal gravimetric reduction occurs almost as easily as the rst. The ability of the
analysis (Fig. S13 and Table S1 in the ESI section†). All Zn(II) complexes to easily accept two electrons makes them
compounds were thermally stable, with 5% weight loss onsets promising electron acceptors.
occurring in the range of 280–517 ꢂC.
The effect of the pyrrolic substituents on the electrochemical
properties of ADP was as expected. Phenylethynyl substituent
(3) anodically (positively) shied both the rst oxidation and the
rst reduction potentials of the BF2+ chelate by 0.18 V and 0.20
Electrochemistry
The electrochemical properties of the ligands and complexes V, respectively. Similarly, the rst oxidation and reduction
were investigated by cyclic voltammetry in dichloromethane, potentials of Zn(WS3)2 were anodically shied by 0.10 V and
using ferrocene/ferrocinium (Fc/Fc+) as the internal reference. 0.17 V, respectively, relative to Zn(ADP)2. The resulting energy
The electrochemical properties of all compounds are summa- gap is reduced due to the extended conjugation through the
rized in Table 1, and the cyclic voltammograms of the WS4 phenylethynyl groups. Adding electron-donating groups at the
series are shown in Fig. 2. The cyclic voltammograms of the para-position of the phenylethynyl substituents (1, 4) further
other compounds can be found in the ESI section.† All free ne-tuned the electrochemical potentials by slightly shiing the
+
ligands and BF2 chelates displayed two reversible reductions potentials cathodically.
and one or two irreversible oxidations, while the Zn(II)
The thienyl pyrrolic substituent (2) had little effect on the
complexes exhibited two closely spaced reversible reductions rst oxidation potential of the free ligand and Zn(II) complex,
and two reversible oxidations. The potential difference between but cathodically shied that of the BF2+ complex by 0.15 V, due
the 1st E1/2(reduction) and the 2nd E1/2(reduction) for the free to the electron-rich character of thiophene. DFT calculations
ligands (H(WS1–5)) varied from 0.55 V to 0.65. This potential indicate that conjugation is not signicantly extended into the
+
difference for BF2 chelates was ꢀ0.80 V, implying that the thiophene groups because they are out-of-plane relative to the
second reduction is much harder than the rst reduction. ADP conjugated plane (dihedral angle between 56ꢂ and 60ꢂ).41
Interestingly, this potential difference is very small for Zn(II) Adding an ethynyl spacer between the ADP core and thiophene
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
rst oxidation potentials of the free ligands H(WS2) and H(WS5)
were similar, the additional ethynyl group anodically shied the
+
rst reduction potential of H(WS2) by 0.13 V. In the case of BF2
complexes, the ethynyl group anodically shied the oxidation
and reduction potentials of BF2(WS2) by 0.09 V and 0.15 V,
respectively. A similar trend was observed for the Ep,a and Ep,c of
the Zn(II) complexes.
Fig. 3 summarizes the HOMO and LUMO energy levels esti-
mated using cyclic voltammetry. The HOMO for Zn(WS3)2 was
ꢃ5.6 eV and the LUMO was ꢃ3.85 eV. By comparison, the LUMO
estimated for PCBM in our laboratory using the same method
was ꢃ3.99 eV. The HOMO and LUMO of Zn(WS3)2 are lower
than those of the Zn(ADP)2 analogue by 0.10 eV and 0.17 eV,
respectively, suggesting that the pyrrolic phenylethynyl moieties
of WS3 stabilize both frontier molecular orbitals. Zn(WS2)2 had
higher LUMO energy levels than Zn(ADP)2 due to the electron
donating strength of thiophene and the fact that the thiophene
is out-of-plane with ADP. Fig. 3 also includes the estimated
energy levels for the common conjugated polymer donor,
regioregular poly(3-hexylthiophene) (P3HT). Although the esti-
mated energy levels are not directly comparable (lm versus
solution measurements), the lower HOMO and LUMO energy
levels of the Zn(II) complexes compared to P3HT should allow
electron transfer from P3HT to the complex and hole transfer
from the complex to P3HT. However, more accurate estimates
of HOMO and LUMO levels are required to determine the exact
energy gaps for these charge transfer processes.
2
Cyclic voltammograms representing H+, BF2 and Zn+2
+
Fig.
chelates of WS4 in 0.1 M TBAPF dichloromethane solution and Fc/Fc+
as the internal standard.
Optical properties
The solution absorption spectra of the azadipyrromethene-
based ligands and complexes are shown in Fig. 4 and the optical
properties are summarized in Table 2. Fig. 4a compares the
absorption spectra of the free ligands. They all showed a strong
absorption peak in the visible range (594–633 nm) due to a
group (5) reduces the steric hindrance between thiophene
hydrogens and ADP phenyl rings, thereby allowing the conju-
gated pyrrolic substituents to lay in the same plane as the ADP
core and to extend the size of the conjugation system. While the
Table 1 Electrochemical properties in dichloromethane
E1/2 oxidation (V vs. Fc/Fc+)
E(p,a) (V vs. Fc/Fc+)
E1/2 reduction (V vs. Fc/Fc+)
E(p,c) (V vs. Fc/Fc+)
H(ADP)
—
—
0.56, 0.96
0.90
0.46, 0.76
0.67, 0.91
0.90
0.57, 0.84
0.63, 0.95
0.75
0.45, 0.69
—
ꢃ1.20, ꢃ1.85
ꢃ0.81, ꢃ1.61
ꢃ1.42, ꢃ1.56
ꢃ1.05, ꢃ1.65
ꢃ0.71, ꢃ1.51
ꢃ1.27, ꢃ1.48
ꢃ1.23, ꢃ1.78
ꢃ0.90, ꢃ1.74
ꢃ1.51, ꢃ1.77
—
ꢃ1.25, ꢃ1.90
ꢃ0.84, ꢃ1.64
ꢃ1.48, ꢃ1.62
ꢃ1.18, ꢃ1.77
ꢃ0.78, ꢃ1.57
ꢃ1.33, ꢃ1.54
ꢃ1.29, ꢃ1.86
ꢃ0.96, ꢃ1.80
ꢃ1.58, ꢃ1.86
—
BF2(ADP)
Zn(ADP)2
H(WS1)
BF2(WS1)
Zn(WS1)2
H(WS2)
BF2(WS2)
Zn(WS2)2
H(WS3)
0.40, 0.69
—
—
0.47, 0.74
—
—
0.39, 0.63
a
BF2(WS3)
Zn(WS3)2
H(WS4)
BF2(WS4)
Zn(WS4)2
H(WS5)
—
1.08
ꢃ0.79, ꢃ1.59
ꢃ1.25, ꢃ1.47
ꢃ1.12, ꢃ1.70
ꢃ0.85, ꢃ1.62
ꢃ1.30, ꢃ1.51
ꢃ1.10, ꢃ1.67
ꢃ0.75, ꢃ1.53
—
ꢃ0.95, ꢃ1.75
ꢃ1.33, ꢃ1.55
ꢃ1.23, ꢃ1.79
ꢃ0.95, ꢃ1.58
ꢃ1.37, ꢃ1.75
ꢃ1.18, ꢃ1.75
ꢃ0.81, ꢃ1.60
ꢃ1.27, ꢃ1.09
0.50, 0.77
—
—
0.58, 0.86
0.58, 0.85
0.87
0.49, 0.74
0.64
0.44, 0.68
—
—
—
BF2(WS5)
Zn(WS5)2
0.84
0.52, 0.76
a
Could not obtain a good cyclic voltammogram of H(WS3) due to limited solubility and purity.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 3 Estimated HOMO and LUMO energy levels obtained by cyclic
voltammetry from the E1/2 values in dichloromethane solution, using
the value of ꢃ5.1 eV for Fc/Fc+.42 For Zn(WS5)2, Ep,a and Ep,c were used
as an approximation to E1/2. The HOMO and LUMO levels of P3HT films
were estimated in our laboratory from the oxidation onset and the
optical gap, and are included for comparison.
pyrrolic ring p–p* transition, and a strong absorption peak near
300 nm. The molar absorptivities for the visible absorption
band ranged from 32 ꢁ 103 Mꢃ1 cmꢃ1 to 47 ꢁ 103 Mꢃ1 cmꢃ1
,
where H(WS2) had the lowest and H(WS1, 4) had the highest
molar absorptivity. The presence of phenylethynyl or thienyle-
thynyl groups at pyrrolic positions lead to a bathochromic shi
of ꢀ37 nm compared to the lmax of H(ADP). H(WS2) has a
similar absorption maximum (lmax ¼ 594 nm) than H(ADP)
(lmax ¼ 596 nm), suggesting that the thiophene groups are not
conjugated with the ADP core. On the other hand, when an
ethynyl spacer was incorporated between the ADP core and the
thiophene group, the maximum absorption red shied by 39
nm to 633 nm for H(WS5), due to the reduced steric hindrance
caused by the spacer, resulting in extended conjugation. Larger
red-shis were observed when comparing the lonset. Relative to
H(ADP), the lonset for H(WS2) red-shied by 14 nm due to the
electron donating nature of thiophene, H(WS3) red-shied by
31 nm due to extended conjugation, and H(WS1, 4, 5) red
shied by 70–90 nm, due to a combination of extended conju-
gation and the presence of electron donating groups in the
pyrrolic substituents.
Fig. 4b compares the absorption spectra of the BF2+ chelates
in chloroform. The absorption spectra showed a broad intense
absorption peak in the near infrared region. Using thienyl
pyrrolic substituents (2) resulted in the smallest lmax of 671 nm,
which is 21 nm red-shied compared to BF2(ADP) (650 nm).
This is consistent with the thiophenes acting as electron rich
substituents that are not conjugated with the ADP core. Using
phenylethynyl (3) or p-alkylphenylethynyl (1) substituents
increased lmax to 732 nm and 736 nm, respectively, consistent
with conjugation between the phenylethynyl groups and the
ADP core. Adding electron-donating groups in the pyrrolic
substituents further increased lmax to 757 nm and 758 nm for p-
alkoxyphenylethynyl (4) and thienylethynyl (5) substituents,
respectively. These lmax values are signicantly larger than
those of the respective free ligands, with a change of: 111 nm for
+
Fig. 4 Absorption spectra of (a) H(ADP) and H(WS1–5) (b) BF2
chelates, and (c) Zn(II) complexes in chloroform solutions.
BF2(WS1), 77 nm for BF2(WS2), 122 nm for BF2(WS3), 128 nm
for BF2(WS4), and 135 nm for BF2(WS5). These red-shis are all
larger than that observed for BF2(ADP) (50 nm). These signi-
cant bathochromic shis are attributed to the enforced
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Table 2 Summary of optical properties of ADP-based molecules in
solution and film
To evaluate how the absorption of our acceptors comple-
ment that of P3HT, we obtained the thin lm absorption spectra
of 1 : 1 w/w mixtures of ADP derivatives and P3HT. Fig. 5a shows
UV-Vis absorption spectra of free ligands mixed with P3HT. The
absorption of the mixture was similar to that of neat P3HT,
partly because the absorption of the free ligands overlaps with
that of P3HT. Nevertheless, there were a few differences: the
lmax for P3HT/H(WS2) was blue shied compared to neat P3HT,
while others have red-shied absorption by ꢀ10–30 nm. A
signicant absorption near ꢀ300 nm, which is not present in
neat P3HT, is attributed to the absorption from the ADP
ligands. Fig. 5b depicts the absorption spectra of P3HT mixed
Solution
lmax
Film
visible
max
,
lonset
,
l
,
lonset,
nm
nm (3, ꢁ103 Mꢃ1 cmꢃ1
)
nm
nm
H(ADP)
302 (38), 596 (42)
310 (20), 649 (50)
302 (57), 591 (78), 645 (51)
302 (46), 625 (47)
307 (32), 741 (48)
300 (82), 650 (91), 683 (94)
305 (45), 594 (32)
275 (23), 671 (36)
278 (51), 620 (84), 676 (66)
309 (39), 610 (42)
306 (25), 732 (49)
309 (39), 674 (105), 705 (93)
309 (51), 629 (47)
645
688
713
715
798
764
659
800
737
676
782
757
725
830
780
736
850
780
640
711
650
612
763
707
528
654
622
585
763
696
597
780
709
645
750
722
740
810
735
765
835
780
675
795
745
775
845
785
820
880
785
870
890
815
BF2(ADP)
Zn(ADP)2
H(WS1)
BF2(WS1)
Zn(WS1)2
H(WS2)
BF2(WS2)
Zn(WS2)2
H(WS3)
BF2(WS3)
Zn(WS3)2
H(WS4)
BF2(WS4)
Zn(WS4)2
H(WS5)
BF2(WS5)
Zn(WS5)2
+
with BF2 chelates. Compared to neat P3HT, the blends had
extra absorption bands near 300 nm and in the near-IR range
+
ꢀ650–900 nm, due to the absorption of the BF2 chelates that
nicely complement that of P3HT. Fig. 5c illustrates the lm
absorption spectra of the P3HT blend with Zn(II) complexes. The
absorption spectra of the Zn(II) complexes also complement
those of P3HT, extending the absorption of P3HT near 300 nm
and in the ꢀ650–800 nm for a better overlap with the solar
spectrum.
310 (44), 757 (55)
310 (81), 655 (88), 693 (90)
315 (38), 633 (37)
320 (34), 768(41)
318 (84), 665(98), 702 (97)
One pre-requisite for good organic solar cells is efficient
exciton dissociation, which can be probed using uorescence
quenching experiments. Fig. S17† compares the uorescence of
planarity and extended conjugation through aryl groups. The the neat P3HT lm with that of blend lms (spin-coated from
BF2+ chelates in chloroform weakly emitted in the near-infrared 1 : 1 w/w mixtures). The data for ADP and WS1 series were
(lmax 794–853 nm, see ESI section†).
previously published in a separate study and are included in
Fig. 4c shows the absorption spectra of the Zn(II) Fig. S17† for comparison.37 The lms were excited at 360 nm.
complexes. Generally, the Zn(II) complexes have a broad Although azadipyrromethene dyes can also be excited at 360
+
absorption in the visible to near-infrared region, with very nm, free ligands and Zn(II) complexes do not uoresce and BF2
high molar absorptivity ranging from 88 ꢁ 103 Mꢃ1 cmꢃ1 to chelates have very weak emission at longer wavelengths than
98 ꢁ 103 Mꢃ1 cmꢃ1. The Zn(WS1, 3–5)2 complexes exhibited P3HT, and thus do not interfere with the uorescence of P3HT.
two overlapping absorptions, resulting from a combination All free ligands and chelates quenched the uorescence of
of ligand planarization upon Zn(II) chelation and exciton P3HT, consistent with exciton dissociation. Quenching was not
coupling of the two planar ligands.43 All Zn(II) complexes were complete for the free ligands H(ADP), H(WS2), H(WS3) and
non-uorescent.
H(WS4). Incomplete quenching could be due to either lower
To evaluate the optical properties in the lm, we fabricated quenching ability or to a larger phase separation between P3HT
lms on glass slides by spin-coating from chloroform solutions. and acceptor. Our previous work reported that H(ADP) and
The Zn(II) complex lms were more uniform than those of the H(WS1) have similar Stern–Volmer quenching constants in
free ligands or BF2+ chelates. The lms were bright blue for free solution (and both are slightly higher than that of PCBM). We
ligands and BF2+ chelates, and blue-red for the Zn(II) complexes. therefore concluded that the lower quenching efficiency of
The optical properties of the lms are summarized in Table 2. H(ADP) than H(WS1) in blend lms was caused by a larger
The low energy absorption extended well into the near-infrared phase separation scale (also observed by AFM).37 Based on these
region. Depending on the pyrrolic substituents, onsets ranged results, it is reasonable to hypothesize that the incomplete
from 675 nm to 870 nm for free ligands, 810 nm to 890 nm for quenching of the free ligands H(WS2), H(WS3) and H(WS4) are
+
BF2 chelates and 735 nm to 815 nm for Zn2+ chelates. The also caused by larger phase separation between the ligand and
absorption spectra of solution and lm are compared in P3HT.
Fig. S15† and the shis upon lm formation are summarized in
Table S3.† In general, the absorption spectra of lms were
Photovoltaic properties
broader than in solution and the absorption onset shied to
longer wavelengths upon lm formation (with the exception of The photovoltaic properties of the azadipyrromethene-based
BF2(WS2), which slightly blue-shied upon lm formation). On dyes were studied in inverted organic solar cells using P3HT as
the other hand, the low energy absorption maximum either red- the electron donor. The device conguration was ITO/ZnO/
shied or blue-shied upon lm formation, depending on the P3HT : Zn(WS1–5)2 (90 nm)/MoO3 (10 nm)/Ag (80 nm). In our
+
pyrrolic substituent and chelation used. This indicates that the hands, free ligands and BF2 chelates did not give any signi-
pyrrolic substituents and chelation affect molecular packing in cant photovoltaic effects, most likely due to unfavorable blend
the solid state.
morphologies. On the other hand, the zinc(II) chelates all
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 5 Film absorption spectra of the P3HT blend with ADP-based materials.
showed good photovoltaic properties. We performed some suggesting that the t-butyl groups are slightly better than 2-
device processing optimization, which involved a screening of ethylhexyl. This is somewhat surprising, as t-butyl solubi-
total concentrations (lm thickness), annealing conditions and lizing groups are usually avoided in the design of conjugated
blend weight ratios using ortho-dichlorobenzene as the solvent. materials, as they are thought to be bulky. The best cell was
We found that the best results were obtained with a total obtained from Zn(WS3)2 with no alkyl groups, with a PCE of
concentration of 20 mg mLꢃ1 and annealing at 120 C for 30 4.1%.
ꢂ
min prior to top electrode deposition. The optimal blend ratio
(by weight) varied from 1 : 0.5 to 1 : 1 P3HT : acceptor.
Fig. 6 shows the current–voltage curves obtained for the
best P3HT : Zn(WS1–5)2 devices, and Table 3 summarizes the
device parameters, including the parameters for the best
devices and the parameter averages of 8 to 10 devices. All
current–voltage curves show a diode behavior. The PCE
obtained for the zinc acceptors varied between ꢀ2% and 4%.
Interestingly, all had a higher PCE than Zn(ADP) with no
pyrrolic substituents. This suggests that the pyrrolic
substituents are benecial for a higher PCE. Zn(WS2)2 with
thiophene pyrrolic substituents showed a PCE of 2.1%, a Jsc
of 4.8 mA cmꢃ2, a Voc of 0.9 V, and a FF of 49%. This acceptor
had the largest Voc due to its higher LUMO energy level.
Adding an ethynyl spacer between the ADP core and thio-
phene increased the PCE to 2.7% (Zn(WS5)2). Zn(WS4)2 had a
similar PCE of 2.6%, despite changing the aryl group to
Fig. 6 Current–voltage curves for the best P3HT : Zn(WS1–5)2 solar
cells.
phenyl. Zn(WS1)2 had a slightly higher PCE of 2.9%,
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Table 3 Photovoltaic performance of blends of P3HT with zinc complexes (20 mg mLꢃ1, annealed 120 ꢂC, 30 min)
Donor : acceptor
Weight ratio
Voc (V)
Jsc (mA cmꢃ2
)
FF (%)
PCE (%)
P3HT : ADP
1 : 0.7
1 : 0.5
1 : 1
1 : 0.7
1 : 0.5
1 : 0.7
0.82 (0.81 ꢄ 0.01)
0.83 (0.82 ꢄ 0.01)
0.91 (0.89 ꢄ 0.02)
0.77 (0.76 ꢄ 0.01)
0.82 (0.82 ꢄ 0.45)
0.84 (0.82 ꢄ 0.02)
2.8 (3.0 ꢄ 0.3)
6.7 (6.5 ꢄ 0.2)
5.0 (4.5 ꢄ 0.5)
9.1 (8.8 ꢄ 0.5)
6.2 (5.7 ꢄ 0.5)
5.9 (5.9 ꢄ 0.3)
60 (59 ꢄ 1)
52 (51 ꢄ 2)
49 (47 ꢄ 3)
59 (57 ꢄ 3)
51 (49 ꢄ 2)
55 (53 ꢄ 0.1)
1.4 (1.3 ꢄ 0.1)
2.9 (2.7 ꢄ 0.1)
2.2 (1.9 ꢄ 0.2)
4.1 (3.9 ꢄ 0.2)
2.6 (2.3 ꢄ 0.1)
2.7 (2.6 ꢄ 0.1)
P3HT : Zn(WS1)2
P3HT : Zn(WS2)2
P3HT : Zn(WS3)2
P3HT : Zn(WS4)2
P3HT : Zn(WS5)2
Zn(WS2)2 extended to shorter wavelength than the others, ꢀ750
nm, which partly explains its lower Jsc and PCE compared to
other complexes. However, the absorption spectra alone cannot
explain the trends in Jsc and PCE observed, because all lms had
a thickness of ꢀ90 nm. The maximum IPCEs at 510 nm were
26%, 30%, 31%, 34%, and 37% for Zn(WS2)2, Zn(WS5)2,
Zn(WS4)2, Zn(WS1)2, andZn(WS3)2, respectively. The ascending
order for maximum IPCE correlates with that of Jsc and PCE.
Because device performance strongly depends on the
morphology, we looked at the surface morphology of the active
layer by atomic force microscopy (AFM). Images were taken
directly from the best photovoltaic devices and are shown in
Fig. 8. All blends had a similar nanoscale surface morphology,
with no large-scale aggregates. This is in contrast to
P3HT : Zn(ADP)2, where large aggregates of Zn(ADP)2 could be
clearly seen.36 This suggests that all pyrrolic substituents
studied (including thienyl) prevent the formation of large
aggregates. The surface roughness (root-mean-square) was 10.0,
9.8, 13.0, 9.7 and 8.4 nm for blends with Zn(WS1)2, Zn(WS2)2,
Zn(WS3)2, Zn(WS4)2 and Zn(WS5)2, respectively. These values
are all similar and are comparable to the reported surface
roughness of efficient P3HT:PCBM active layers.44 Based on the
photovoltaic and AFM results, we hypothesize that all Zn(WS1–
5)2 acceptors phase separate on a similar scale with P3HT, and
that differences in performance arise from differences in local
donor–acceptor interfaces that would be affected by the solu-
bilizing groups. Additional studies are required to test this
hypothesis.
Fig. 7 External quantum yield spectra for the best P3HT : Zn(WS1–5)2
solar cells.
It appears that the differences in the PCE between the
acceptors is mainly governed by Jsc, which can be inuenced by
several factors including the absorption spectra, as well as
charge separation and charge extraction efficiency. Fig. 7
depicts the incident photon-to-current efficiency (IPCE) spectra
of the best solar cells. All P3HT : Zn(WS1–5)2 cells show a
photoresponse beyond the absorption onset of P3HT, up to 800
nm, indicating that excitons generated in the acceptors
contribute to photocurrent production. The spectral response of
Fig. 8 Tapping-mode AFM height images (a–e) and phase images (f–j) for P3HT : acceptor blends taken directly from the active layers of the
best cells reported in Table 3. The acceptor was: (a and f) Zn(WS1)2; (b and g) Zn(WS2)2; (c and h) Zn(WS3)2; (d and i) Zn(WS4)2; (e and j) Zn(WS5)2.
All images are 5 mm ꢁ 5 mm. The height scales were between 60 nm and 80 nm and the phase scale was between 30 and 40 degrees. Images with
the scales are available in the ESI section.†
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Journal of Materials Chemistry A
8 A. M. Nardes, A. J. Ferguson, P. Wolfer, K. Gui, P. L. Burn,
P. Meredith and N. Kopidakis, ChemPhysChem, 2014, 15,
1539–1549.
9 R. D. Pensack, C. Guo, K. Vakhshouri, E. D. Gomez and
J. B. Asbury, J. Phys. Chem. C, 2012, 116, 4824–4831.
10 J. Peet, M. L. Senatore, A. J. Heeger and G. C. Bazan, Adv.
Mater., 2009, 21, 1521–1527.
11 A. Facchetti, Mater. Today, 2013, 16, 123–132.
12 K. Cnops, B. P. Rand, D. Cheyns, B. Verreet, M. A. Empl and
P. Heremans, Nat. Commun., 2014, 5, 3406.
13 Y. Zang, C.-Z. Li, C.-C. Chueh, S. T. Williams, W. Jiang,
Z.-H. Wang, J.-S. Yu and A. K. Y. Jen, Adv. Mater., 2014, 26,
5708–5714.
14 Y. Zhou, T. Kurosawa, W. Ma, Y. Guo, L. Fang, K. Vandewal,
Y. Diao, C. Wang, Q. Yan, J. Reinspach, J. Mei, A. L. Appleton,
G. I. Koleilat, Y. Gao, S. C. B. Mannsfeld, A. Salleo, H. Ade,
D. Zhao and Z. Bao, Adv. Mater., 2014, 26, 3767–3772.
15 Z. Lu, B. Jiang, X. Zhang, A. Tang, L. Chen, C. Zhan and
J. Yao, Chem. Mater., 2014, 26, 2907–2914.
16 W. Jiang, L. Ye, X. Li, C. Xiao, F. Tan, W. Zhao, J. Hou and
Z. Wang, Chem. Commun., 2014, 50, 1024–1026.
17 T. Earmme, Y. J. Hwang, N. M. Murari, S. Subramaniyan and
S. A. Jenekhe, J. Am. Chem. Soc., 2013, 135, 14960–14963.
18 J. T. Bloking, T. Giovenzana, A. T. Higgs, A. J. Ponec,
E. T. Hoke, K. Vandewal, S. Ko, Z. Bao, A. Sellinger and
M. D. McGehee, Adv. Energy Mater., 2014, 4, 1301426.
19 B. Verreet, K. Cnops, D. Cheyns, P. Heremans, A. Stesmans,
G. Zango, C. G. Claessens, T. Torres and B. P. Rand, Adv.
Energy Mater., 2014, 4, 1301413.
Conclusions
In this study, we explored the effects of pyrrolic substituents on
the optical, electrochemical and photovoltaic properties of new
electron acceptors based on azadipyrromethene complexes. The
various modications studied ne-tuned the optical and elec-
trochemical properties of the ligands and complexes. In solar
cells, the modications of the zinc complexes did not dramat-
ically affect the device performance and blend morphology.
This shows the robustness of these acceptors when blended
with P3HT as the donor. We made progress in identifying some
of the design principles for electron accepting materials based
on azadipyrromethene dyes. For example, comparing the
performance of Zn(WS3)2, Zn(WS1)2 and Zn(WS4)4 suggests that
no solubilizing groups is better than t-butyl solubilizing groups,
which is slightly better than 2-ethylhexyl solubilizing groups.
Other less bulky solubilizing groups should be studied in the
future to optimize solubility and performance. In addition,
comparing the performance of Zn(WS5)2 and Zn(WS2)2 points
to the importance of an ethynyl group between the azadipyrro-
methene core and the aryl group. We are currently exploring the
effect of using a vinyl instead of an ethynyl group. Preliminary
results with other high efficiency polymer donors (optimized for
PCBM and PC70BM acceptors) have yielded poor results. The
polymer donors tested had deeper HOMO and LUMO energy
levels than P3HT, resulting in unfavorable energy matches with
our acceptors. We are currently tuning the energy levels of
Zn(WS3)2 to test this hypothesis, continuing to learn how to
design better non-fullerene acceptors for the next generation of
organic photovoltaic devices.
20 X. Zhang, Z. Lu, L. Ye, C. Zhan, J. Hou, S. Zhang, B. Jiang,
Y. Zhao, J. Huang, S. Zhang, Y. Liu, Q. Shi, Y. Liu and
J. Yao, Adv. Mater., 2013, 25, 5791–5797.
Acknowledgements
21 P. Cheng, L. Ye, X. Zhao, J. Hou, Y. Li and X. Zhan, Energy
Environ. Sci., 2014, 7, 1351–1356.
22 Y. Zhou, Y.-Z. Dai, Y.-Q. Zheng, X.-Y. Wang, J.-Y. Wang and
J. Pei, Chem. Commun., 2013, 49, 5802–5804.
23 Y. Lin, Y. Wang, J. Wang, J. Hou, Y. Li, D. Zhu and X. Zhan,
Adv. Mater., 2014, 26, 5137–5142.
24 A. a. F. Eaiha, J.-P. Sun, I. G. Hill and G. C. Welch, J. Mater.
Chem. A, 2014, 2, 1201–1213.
We are grateful to the National Science Foundation (CHEM
1148652) for funding this project. This material is based upon
work supported by the National Science Foundation under
Grant MRI-28.5 0821515 (for the purchase of the MALDI-TOF/
TOF).
25 Y. Kim and E. Lim, Polymers, 2014, 6, 382–407.
26 J. E. Anthony, A. Facchetti, M. Heeney, S. R. Marder and
X. Zhan, Adv. Mater., 2010, 22, 3876–3892.
References
1 C. Brabec, S. Gowrisanker, J. Halls, D. Laird, S. Jia and
S. Williams, Adv. Mater., 2010, 22, 3839–3856.
27 P. Hudhomme, EPJ Photovoltaics, 2013, 4, 40401–40411.
2 Organic Photovoltaics: Materials, Device Physics, and 28 J. E. Anthony, Chem. Mater., 2011, 23, 583–590.
Manufacturing Technologies, ed. C. Brabec, V. Dyakonov and 29 P. Sonar, J. P. Fong Lim and K. L. Chan, Energy Environ. Sci.,
U. Scherf, WILEY-VCH Verlag GmbH
Weinheim, 2008.
&
Co. KGaA,
2011, 4, 1558–1574.
30 A. Loudet and K. Burgess, Chem. Rev., 2007, 107, 4891–4932.
3 F. C. Krebs, S. A. Gevorgyan and J. Alstrup, J. Mater. Chem., 31 D. V. Partyka, N. Deligonul, M. P. Washington and T. G. Gray,
2009, 19, 5442–5451.
Organometallics, 2009, 28, 5837–5840.
4 N. Espinosa, M. Hosel, M. Jørgensen and F. C. Krebs, Energy 32 A. Palma, J. F. Gallagher, H. Mueller-Bunz, J. Wolowska,
¨
Environ. Sci., 2014, 7, 855–866.
5 J. Nelson, Mater. Today, 2011, 14, 462–470.
E. J. L. McInnes and D. F. O'Shea, Dalton Trans., 2009,
273–279.
6 L. Dou, J. You, Z. Hong, Z. Xu, G. Li, R. A. Street and Y. Yang, 33 A. Gorman, J. Killoran, C. O'Shea, T. Kenna, W. M. Gallagher
Adv. Mater., 2013, 25, 6642–6671. and D. F. O'Shea, J. Am. Chem. Soc., 2004, 126, 10619–10631.
7 Y. He and Y. Li, Phys. Chem. Chem. Phys., 2011, 13, 1970– 34 V. Bandi, K. Ohkubo, S. Fukuzumi and F. D'Souza, Chem.
1983.
Commun., 2013, 49, 2867–2869.
This journal is © The Royal Society of Chemistry 2015
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
35 A. M. Poe, A. M. Della Pelle, A. V. Subrahmanyam, W. White, 40 C. Xia, X. Fan, J. Locklin and R. C. Advincula, Org. Lett., 2002,
G. Wantz and S. Thayumanavan, Chem. Commun., 2014, 50,
2913–2915.
36 Z. Mao, W. Senevirathna, J. Y. Liao, J. Gu, S. V. Kesava,
4, 2067–2070.
41 R. Gresser, H. Hartmann, M. Wrackmeyer, K. Leo and
M. Riede, Tetrahedron, 2011, 67, 7148–7155.
C. Guo, E. D. Gomez and G. Sauve, Adv. Mater., 2014, 26, 42 C. M. Cardona, W. Li, A. E. Kaifer, D. Stockdale and
6290–6294. G. C. Bazan, Adv. Mater., 2011, 23, 2367–2371.
37 W. Senevirathna and G. Sauve, J. Mater. Chem. C, 2013, 1, 43 S. G. Telfer, T. M. McLean and M. R. Waterland, Dalton
6684–6694. Trans., 2011, 40, 3097–3108.
38 S. Rajaram, R. Shivanna, S. K. Kandappa and K. S. Narayan, J. 44 G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery
Phys. Chem. Lett., 2012, 3, 2405–2408.
39 Q. Yan, Y. Zhou, Y.-Q. Zheng, J. Pei and D. Zhao, Chem. Sci.,
2013, 4, 4389–4394.
and Y. Yang, Nat. Mater., 2005, 4, 864–868.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2015