Key words
arenes - condensation - electron acceptors - non-fullerene acceptors - organic electronics
- organic solar cells
Introduction
Tailor-made N-heteroarenes[1] are electron-deficient semiconductors with applications in organic field-effect
transistors,[2] organic light-emitting diodes[3] and bulk hetero- junction (BHJ) solar cells.[4] Negative charges are stabilized in their radical anions;[5] these are the charge-transporting species in thin-film transistors.[6] N-Heteroarenes are attractive electron acceptors for organic photovoltaics (OPVs).
The hydrocarbon analogues,[7] electron donors, cannot be used with fullerene acceptors due to Diels–Alder adduct
formation;[8] azaacenes and azaarenes, however, can be combined with donors such as PTB7,[4]
P3HT and PCPDTBT
.
[4a] Triisopropylsilyl (TIPS)-ethynylated (hetero)acenes and (hetero)arenes display a
strong tendency to crystallize,[4]
[9] and are therefore problematic in BHJ cells as phase segregation instead of BHJ formation
occurs in thin–film blends.[4c]
To control aggregation and morphology in bulk or mixtures, azaarenes are equipped
with morphology-dominating substituents such as iptycenes[4a] (e.g. A, [Figure 1]) or they are oligomerized;[4b]
[4d]
[10] this improves photovoltaic performance. The resulting amorphous species balance
acceptor properties, charge transport and percolation pathways in thin films. They
are also quite soluble and processable.
Figure 1 Selected acenothiadiazoles used as acceptor materials for BHJ solar cells.[4a]
[4b]
Others and we investigated spirobridged (hetero)-acenes and acenothiadiazoles for
intramolecular singlet fission[11] and photovoltaics (e.g. B, [Figure 1]).[4b] This 1,1′-spiro linkage was popularized by McKeown[12a]
[12b] and others[12c]
[12d] in polymers of intrinsic microporosity.[12e]
Thiadiazoles performed better than tetra- and diazapentacenes in solar cells, but
the effect of the linker was not clear. Herein, we study phenazinothiadiazoles, connected
via two different linkers.[12]
[13] We compare their optoelectronic properties and OPV performance to those of thiadiazole
6 and B to study the role of the linker. In heteroarene dimers, the linker impacts bulk crystallinity.
Results and Discussion
2,5-Hexanedione reacts with catechol, and subsequent NaIO4 oxidation gives ethanoanthracenetetraone 1. Spirobisindenetetraone 2 is obtained after oxidation of its tetraol precursor,[13] but decomposes within 1 d. Its surprising reactivity might be attributed to better
steric accessibility of 2 when compared to 1 and the precursor quinone for B – both stable under ambient conditions. Hence, 2 was used without purification. Condensation[14] of quinones 1 and 2 with diamine 5 furnishes the target dimers 3 (77% yield)[15] and 4 (22% yield)[16] as greenish-blue amorphous solids ([Scheme 1]). We ascribe the lower yield of 4 to the poorer solubility and purity as well as higher reactivity of quinone 2 compared to 1.
Scheme 1 Synthesis of dimers 3 and 4. a
2 was generated in situ from the tetraol and immediately used (see [SI] for details). Conditions: CHCl3, HOAc, 50 °C, 15 h.
Compared to the monomer 6 (λ
max = 642 nm), the absorption maxima of 3 (λ
max = 640 nm) and 4 (λ
max = 636 nm) are slightly blue-shifted by 2/0.01 and 6/0.02 nm/eV [see [Figure 2] for photographs and [Figure 3] (top) and [Table 1] for spectra and spectral data]. The extinction coefficient of the chromophores 3 and 4 are twice as high as that of 6. In thin films, spun-cast from chloroform, the absorption features are narrowed compared
to that of monomer 6 and resemble the spectra in solution, suggesting less intermolecular interactions
and π-stacking, or reduced domain boundary scattering (amorphous film) as evidenced
by the polarization micrographs ([Figure 4], bottom). In contrast, thin films of the monomer 6 are crystalline.
Figure 2 Photographs of monomer 6 (left), dimers 3 (middle) and 4 (right) under daylight (top) and UV light with excitation at 365 nm (bottom) in n-hexane.
Figure 3 Top: Normalized absorption spectra of 3, 4 and 6 in n-hexane (black, dilute solution) and spun-cast thin-films on glass (red, from chloroform,
c = 10 mg mL−1). Bottom: Microscopic images of spin-coated thin films of 3 (middle), 4 (bottom) and 6 (top) on glass (10 mg mL−1, chloroform) under crossed polarizers.
Table 1
Optical, electrochemical, and calculated (gas-phase) properties of phenazinothiadiazole
dimers 3, 4, and their consanguine monomeric counterpart 6 in solution (n-hexane).
|
Cmpd
|
λ
max, abs [nm]
|
λ
max, em [nm]
|
ε
[a] [x 104 M−1cm−1]
|
Φ
|
E
(0/−) [V][c]
|
IP/HOMO [eV][d] meas/calcd[e]
|
EA/LUMO [eV][f] meas/calcd[e]
|
Gap [eV][g] meas/calcd[e]
|
|
6
|
642
|
651
|
5.87
|
0.46
|
−0.85
|
−5.83/− 5.76
|
−3.97/− 3.77
|
1.86/1.99
|
|
3
|
640
|
647
|
13.6
|
–[b]
|
−0.90
|
−6.06/− 5.72
|
−4.20/− 3.79
|
1.86/1.93
|
|
4
|
636
|
646
|
13.5
|
0.02
|
−0.92
|
−6.06/− 5.73
|
−4.18/− 3.72
|
1.88/2.00
|
a At λ
max, abs.
b Non-emissive.
c First reduction potentials from cyclic voltammetry (CV) in DCM at room temperature
with Bu4NPF6 as the electrolyte against Fc/Fc+ as an internal standard (−5.10 eV) at 0.2 Vs−1.
d IPmeas = EA
meas − gap
meas.
e Obtained from DFT calculations (Gaussian 16: geometry optimization: B3LYP/def2-SVP//
B3LYP/def2-TZVP; FMO calculations: B3LYP/def2-TZVP; TMS groups were used instead of
TIPS).[20]
f EAmeas = − e(5.1 V + E
(0/−)).[21]
g Gapmeas calculated from λ
onset in n-hexane.
Figure 4 DFT-optimized geometries of B (a), 3 (b) and 4 (c) (B3LYP/def2-SVP // B3LYP/def2-TZVP) and calculated distance between center points
of closest aromatic rings.[19] Solid-state structure of 3 (d) and its crystal packing (e). The angle between the azaarene subunits is approximately
121° (angle estimated from best-fit planes of closest aromatic rings; ellipsoids set
to 50% probability level).
The normalized absorption spectra of the dimers 3 and 4 are indistinguishable, but only 4 is faintly fluorescent (quantum yield = 2%). In 3 the chromophores are closer together than in spiro linked dimer 4; also the geometry and rigidity differ ([Figure 4], top). Density functional theory geometry optimizations of 3 and 4 furnish reduced (3: 4.53 Å) and enlarged (4: 6.59 Å) closest distances of the acenothiadiazoles (center points of closest aromatic
rings) compared to B (5.08 Å). The homodimer 3 is C
2v-symmetric, 4 (and B) are of C
2 symmetry, due to the cyclopentylene moieties of the dimer 4 preferably adopting an envelope conformation.
Figure 5 (a) Schematic of the OPV device architecture, (b) chemical structures of the materials
used as the BHJ active layer, (c) external quantum efficiency and (d) current density–voltage
characteristics of the best OPVs.
Recently, Mastalerz et al. disclosed homoaromaticity of triptycene-based chromophores,[17] expressed as individually observed reduction waves of each arm under cyclic voltammetry
conditions – not observed for the dimers 3 or 4. The first reduction potentials are 6 (−0.85 V) > 3 (−0.90 V) > 4 (−0.92 V, all vs. Fc/Fc+), due to the electron-releasing effect of the linkers, comparable to those of two
alkyl groups.
Dimer 4 did not crystallize, while 3 gave specimen suitable for X-ray crystal structure analysis – note that various attempts
and variation of crystallization conditions were necessary to obtain viable crystals.
The often-observed S–N interactions are absent in this structure.[18] Compound 3 packs in one-dimensional stacks. Molecules of each stack face the same direction.
Adjacent stacks are oriented in opposite directions. The TIPS-ethynyl substituents
protrude into the neighboring stacks and prevent intermolecular π–π interactions.
One of 3's phenazinothiadiazole arms is slightly bent due to packing effects.
To investigate the feasibility of the phenazinothiadiazole dimers 3, 4 and B as electron acceptors in OPV devices, they were mixed with the donor polymer poly({4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl}{3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl}) (PTB7) to form BHJ solar cells.[20] A schematic of the solar-cell architecture and the chemical structures of the donor
and acceptors are shown in [Figure 5a] and [b]. [Figure 5c] shows the external quantum efficiency (EQE) of the resultant OPV devices. Since
PTB7 only weakly absorbs below 500 nm, the EQE features in this spectral range can be
attributed to the acceptors, in agreement with their absorption ([Figure 3], top), demonstrating their contribution to the generation of photocurrent. [Figure 5d] displays the current density–voltage characteristics of the best devices and [Table 2] summarizes the average and highest achieved OPV performances. Overall, the three
types of solar cells show similar performance. The choice of linker of the dimers
has no significant influence on the OPV performance. As expected from the EQE spectra,
OPVs with compound 4 exhibit the highest short-circuit current on average. The open-circuit voltage consistently
reaches values of about 0.7 V for all OPVs which is related to the similarity in the
acceptor LUMO positions (see [Table 1] and Reference[4b]). In all three cases, the fill factors are low and might be the result of unfavorable
active layer morphologies. Lastly, it should be noted that the solar cell fabrication
parameters have not been extensively optimized. However, the presented results are
consistent with previous results for solar cells with acceptor B fabricated with slightly different parameters.[4b]
Table 2
Photovoltaic parameters of the OPVs utilizing compounds 3, 4, and B as acceptors and PTB7 as donor. The highest achieved values are given in brackets representing the J–V curves shown in [Figure 5d].
|
Cmpd
|
V
OC [V]
|
J
SC [mA cm−2]
|
FF [%]
|
PCE [%]
|
|
3
|
0.73 ± 0.01 (0.73)
|
−4.62 ± 0.30 (−5.10)
|
38.44 ± 2.18 (40.56)
|
1.29 ± 0.15 (1.52)
|
|
4
|
0.70 ± 0.02 (0.74)
|
−5.10 ± 0.07 (−5.15)
|
38.56 ± 1.89 (40.94)
|
1.39 ± 0.10 (1.55)
|
|
B
|
0.70 ± 0.01 (0.71)
|
−4.85 ± 0.14 (−5.05)
|
37.82 ± 1.73 (38.50)
|
1.28 ± 0.10 (1.38)
|
Note: V
OC = open-circuit voltage, J
SC = short-circuit current, FF = fill factor, PCE = power conversion efficiency.
Conclusions
In conclusion, we prepared two TIPS-ethynylated phenazinothiadiazole dimers 3 and 4, covalently linked via non-conjugated alkylene bridges and compared them to related
dimer B and their consanguine monomeric counterpart 6. Both 3 and 4 form amorphous thin films which, when employed as acceptors in BHJ solar cells, improve
device performance compared to that of the crystalline monomer 6, as was the case for B.
[4b] Despite different intramolecular geometries (distance and relative orientation of
chromophores), the dimers share similar optical and electrochemical properties –homoconjugation
is vanishingly small in the phenazinothiadiazole dimers. This assumption is strengthened
by comparing 3, 4, and B in BHJ photovoltaic devices – we find that choice of linker has only minor influence
with close to equal device performances. Despite the only moderate power conversion
efficiencies (up to 1.6%), our results provide valuable insight into the role of linker
in homodimers.
