A. Aguiar, et al.
Dyes and Pigments 172 (2020) 107842
respect to PC60BM in BHJ photovoltaic cells because their HOMO and
LUMO lie above the corresponding levels of PC60BM i.e. −6.1 eV and
BDP6-based devices. Devices based on BDP1 and BDP5 showed a
slightly lower efficiencies. BDP3 and BDP4 led to OPV with much worst
performance. The current density versus voltage, in the dark and under
illumination, for the best BODIPYs:PC60BM cells are shown in the
supplementary information (Fig. S34).
−
3.7 eV, respectively, thereby forming a type II heterojunction.
3.6. Theoretical calculations
As mentioned above, the optimization process showed that the best
thickness of the active layer should be around 70 nm. However, under
the optimal conditions, BDP3 has 110 nm of active layer thickness.
When the active layer thickness was reduced to 77 nm (using 35 mg/ml
solution concentration) the PCE remained at 0.01%. The thinner layer
caused a small increase in short circuit current (JSC) but also a decrease
of open circuit voltage (VOC).
To better understand the change of photophysical and electro-
chemical properties among the various BODIPYs, the structures and
energies of the frontier orbitals were calculated by Density Functional
Theory (DFT) in vacuum. The optimized ground-state geometries and
electronic distribution in HOMO-LUMO levels are presented in Table 6.
All calculations were carried out using the B3LYP method and a
6
–31 G* basis set.
The presented BODIPY-based OPVs show high VOC values, some of
which are similar to those of the best BODIPY-based solar cells, but the
main limitation factors in these systems are JSC and fill factor (FF).
These parameters are related to light harvesting ability, recombination,
charge transport and collection. Since the proposed BODIPY dyes show
high absorption coefficients and absorbance in a favourable range of
the spectrum, we believe that the main restriction of this system is re-
lated to a non-optimal active layer morphology or to a poor charge
transport.
The six-member ring of BODIPY core is almost planar. Around the
boron atom, in a tetrahedral coordination, the two fluorine atoms are
perpendicular to the BODIPY core. The meso-substituted groups are
nearly perpendicular to the plane of the BODIPY core.
The π-electrons in the HOMOs are delocalized over the entire
BODIPY backbone π-systems. Similarly, the LUMO orbitals in all
BODIPYs, with the exception of BDP4, are delocalized also within the
BODIPY backbone but with higher intensity at the meso carbon. Due to
the less perpendicular symmetry of meso group on BDP3 there is a very
small delocalization between the backbone and the meso-substituting
group. In the case of BDP4, the LUMO is entirely localised in the ni-
trophenyl substituent. The absence of spatial overlap between HOMO
and LUMO of BDP4, indicates that the HOMO – LUMO transition has a
charge transfer character, which is in line with the fluorescence results
discussed above.
Among the five tested BODIPYs, the largest difference between the
corresponding devices is in the current. The best BODIPY is the one
with the free-meso-position and the worst (BDP3 and BDP4) are the
BODIPYs with electron-withdrawing groups at the meso-position. These
results show that the nature of the group at the meso position sig-
nificantly affects the photovoltaic performance. Since the meso group is
perpendicular to the BODIPY core and it does not have a considerable
influence on the optical-electronics properties, the differences in device
performance are likely related to the intermolecular interactions and
molecular packing.
The HOMO energies obtained from cyclic voltammetry are very si-
milar to the theoretical ones. However, a somewhat higher deviation is
found for the LUMO energy. The dissimilarity between the experimental
LUMO energy and the calculated can be justified by the lack of solvent
stabilisation effects. Nevertheless, the experimental and calculated data
follows the same trend, in what concerns HOMO, LUMO and energy-gap,
and confirm that all the six BODIPYs are suitable candidates to be used as
electron donors in OPVs when blended with PC60BM.
The external quantum efficiencies (EQE) were also measured for
BDP1:PC60BM, BDP5:PC60BM, BDP6:PC60BM (Fig. 5). The EQE profiles
are similar to the corresponding absorption spectra of the films, with a
broad response in the 375–600 nm range. This response is mainly due to
the BODIPY absorption profile but some contribution of PC60BM at
3
00–400 nm is also detected. The stronger response of the BDP6-based
3.7. Photovoltaic performance evaluation and surface morphology
OPV is consistent with a higher short-circuit current.
The morphology of the OPVs’ active layers can provide valuable
information about the photoelectronic response, being related with
exciton dissociation efficiency and, consequently, with the obtained
photocurrent. For that reason, the film's surface of BDP1:PC60BM,
BDP3:PC60BM, BDP4:PC60BM, BDP5: PC60BM and BDP6:PC60BM were
characterized by AFM (Fig. 6 and Fig. S35). All films exhibit very si-
milar surface topography without noticeable phase domains or ag-
gregates. They are very flat (root mean squared (RMS) inferior to
0.40 nm). Usually low roughness and small phase domains are in-
dicators of good miscibility between the donor and the acceptor and are
associated with good photovoltaic efficiencies. These results demon-
strate that there is a good miscibility between the BODIPY dyes and the
PC60BM, with a good donor:acceptor interpenetrating network within
the blend films.
All BODIPYs exhibit the fundamental requirements to be used as
electron-donor materials in PC60BM-based OPVs. Fluorescence studies
of the film prepared with the BDP6:PC60BM blend (Fig. 3) reveal a
complete quenching of the BODIPY emission, which is consistent with
an excited state electron transfer from BODIPY to fullerene, thereby
providing good perspectives for application in OPVs.
Solution-processed bulk heterojunction OPVs were manufactured
with a typical multilayer structure: ITO/PEDOT:PSS/active layer/LiF/Al.
The PEDOT:PSS layer was spin-coated at 1800 rpm and dried for
1
0 min at 125 °C to obtain a film thickness of about 40 nm. Several
conditions of the active layer were tested, aiming to achieve the best
power conversion efficiency (PCE). The optimization procedures, Table
S2 (Supplementary Information), were carried out with BDP6-based
devices and then replicated in the OPVs based on the remaining BOD-
IPYs. The adjusted parameters were: donor/acceptor weight ratio, active
layer solution concentration and back (top) electrode (Ca/Al or LiF/Al).
The optimized BDP6-based OPV was prepared from a BDP6:PC60BM
blend with a weight ratio of 1:3, dissolved in dichlorobenzene at 40 mg/
ml, and deposited by spin-coating at 1200–1300 rpm with LiF/Al top
electrode.
The AFM results show a good intermixing of the two active layer
materials (BODIPY and PCBM). This morphology facilitates the exciton
dissociation (by virtue of a large donor/acceptor interface) but tends to
limit the charge transport, as it does not allow the formation of per-
colation paths for the generated charges to reach the electrodes.
The hole mobility in films of neat BODIPYs was calculated from the
current-voltage characteristics of hole-only devices, with the structure
Devices based on BDP2 were poorly characterized due to the BDP2
low solubility, which led to films with poor quality and irregular cur-
rent-voltage (J–V) curves. For this reason, BDP2 was not included in
the photovoltaic studies.
ITO/PEDOT:PSS/BODIPY/MoO /Al in the space charge limited current
3
−7 2
(SCLC) regime [39]. The calculated values were BDP1: 3 × 10 cm /
−6
2
−7
2
V.s; BDP3: 6 × 10 cm /V.s; BDP4: 6 × 10 cm /V.s; BDP6:
−6
2
1 × 10 cm /V.s. In case of BDP5, we could not observe such SCLC
regime, and the mobility could not be calculated using this method.
Despite the similarity of the calculated mobilities to some reported
BODIPY-based OPVs [40,41], these values are much lower than both
The parameters characterizing the photovoltaic cells performance
are given in Table 7 and the J–V curves are shown in Fig. 4. For the
optimized conditions, the best PCE values were obtained with the
9