RESEARCH ARTICLE
www.solar-rrl.com
Increased Efficiency of Organic Solar Cells by Seeded
Control of the Molecular Morphology in the Active Layer
Md Habibur Rahaman, Jason Holland, Md. Anower Hossain, Leiping Duan,
Bram Hoex, Pablo Mota-Santiago, Valerie D. Mitchell, Ashraf Uddin,*
and John A. Stride*
acceptor materials, including weak absorp-
The performance of non-fullerene, polymer bulk heterojunction (BHJ) organic tion in the visible range, costly synthesis,
photovoltaic devices has a significant correlation with the molecular morphology of limited tunability of energy levels, and
[3]
the donor and acceptor. The authors show that small organic molecules coordi- spherical shape. In contrast, NFAs have
nated to a metal oxide, an electron transport seed layer (ETSL), can profoundly easily designed molecular skeletons and
flexibility with respect to side-chain
modify the donor:acceptor molecular morphology of inverted organic photovoltaic functionalization, allowing for tuning of
(OPV) devices. Using grazing incidence wide angle X-ray scattering (GIWAXS), the the electronic properties, prerequisites for
authors show that a PTB7-Th:IEICO-4F BHJ active layer has a higher degree of face- enhancing organic photovoltaic (OPV)
on molecular alignment on ETSL-1 (biphenyl-4,4 0-dicarboxylic acid, coordinated to device efficiency.[3] Recent studies have
ZnO), whilst for naphthalene-2,6-dicarboxylic acid coordinated to ZnO (ETSL-2), it shown that NFA–polymer bulk heterojunc-
is reduced. Devices of PTB7-Th:IEICO-4F BHJ prepared on ETSL-1 had a 19.91% tions (BHJs) can result in improved solar
cell efficiency when using strategies such
increase in the average power conversion efficiency (PCE), a 1.56% increase in the as the novel blending of small molecule
fill factor (FF), and a 16.66 0.2% enhancement in the short circuit current density. acceptor–polymer donors, charge transport
The observed improvements are believed to be due to significant modifications to layer passivation, and solvent-induced BHJ
the oxide-BHJ interfacial region of ETSL-1, namely the elimination of nano-ridges molecular morphological modification.[4–7]
and defect centers, along with an enhanced wettability. These factors can be Among all these strategies, very few
correlated with the enhanced device performances, leading to the conclusion that studies have focused on correlating the
heterointerface of the electron charge
the modulation of the molecular morphology of donor:acceptor blends by ETSL-1 transport layer and its impact on the
has a broad impact on improving OPV cell efficiencies. molecular morphology of the BHJ to the
OPV device performance.[5,8,9] For
example, NFA–polymer combinations,
such as PTB7-Th:ITIC, PTB7-Th:ITIC-4F,
1. Introduction and PTB7-Th:IEICO-4F, have regularly been employed in
inverted OPV configurations with electron extraction layers
Nonfullerene small molecule acceptors blended with conjugated (ETLs) such as ZnO and TiO2, but lack a detailed
polymer semiconductor donors are being extensively correlation between the ETL and BHJ heterointerface, its effect
studied, largely aimed at improving the efficiency of organic solar on the donor:acceptor molecular morphology, and the overall
[1,2] OPV device performance.[5,8,9]cells compatible with wet-chemistry solution processing. To In addition, controlled reduction
date, nonfullerene acceptor (NFA) materials have received more of the chemical and physical defects, metal ion vacancy, and sur-
focus, largely due to significant disadvantages of fullerene face roughness and uniformity of these aforementioned ETL
M. H. Rahaman, J. Holland, J. A. Stride M. H. Rahaman, M. A. Hossain, B. Hoex, A. Uddin
School of Chemistry School of Photovoltaic and Renewable Energy Engineering
The University of New South Wales The University of New South Wales
Sydney, NSW 2052, Australia Sydney NSW 2052, Australia
E-mail: J.Stride@unsw.edu.au E-mail: a.uddin@unsw.edu.au
The ORCID identification number(s) for the author(s) of this article L. Duan
can be found under https://doi.org/10.1002/solr.202200184. School of Engineering
Australian National University
© 2022 The Authors. Solar RRL published by Wiley-VCH GmbH. This is an Canberra ACT 0200, Australia
open access article under the terms of the Creative Commons Attribution
License, which permits use, distribution and reproduction in any medium, P. Mota-Santiago, V. D. Mitchell
provided the original work is properly cited. Australian Synchrotron
ANSTO
DOI: 10.1002/solr.202200184 800 Blackburn Road, Clayton, VIC 3168, Australia
Sol. RRL 2022, 2200184 2200184 (1 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
layers may impact the molecular morphology of the donor:accep- morphology of the PTB7-Th:IEICO-4F BHJ layer. We also
tor blends and yet any correlation to the OPV device efficiency is observed a high energy offset between the LUMO level of
rare to find. Self-assembled monolayers (SAMs) of organic mol- ETSL-2 and the acceptor of the BHJ. The enhanced face-on molec-
ecules have been found to passivate metal oxide surfaces by ular morphology was highly correlated to the device performances
greatly reducing the defect centers, but may result in an increase with improved performance in device D1 over D0 and D2. The
in the interfacial energy barrier, which could reduce the OPV champion device (glass/ITO/ETSL-1/PTB7-Th:IEICO-4F/MoO3/
performance.[10–12] Numerous studies have used this approach, Ag (D1)) showed a power conversion efficiency (PCE) of
including use of the chelating agent ethylenediaminetetraacetic 12.53%, with a fill factor (FF) of 65.11%, a short-circuit current
acid (EDTA) on ZnO, fullerene (C60) molecules as a passivation density (J ) of 26.6mA cm2sc , and open-circuit voltage (Voc) of
layer on ZnO, while benzoic acid derivatives, tert-butyl amines, 0.71 V, compared to 10.74% PCE, 63% FF, 24.35mA cm2 Jsc,
and halogenated benzoic acids have also been used as organic and 0.69 Voc for the reference device D0 (glass/ITO/ZnO/
linkers on the ETL, all demonstrating some enhancement of PTB7-Th:IEICO-4F/MoO3/Ag). An average increase of 19.91%
the OPV device performance, but lacking in any correlation to in the PCE of D1 (11.20% up from 9.34%) was recorded compared
the impact on the molecular morphology within the BHJs.[9,13] to D0. Whereas device D2 (glass/ITO/ETSL-2/PTB7-Th:IEICO-
One of the most commonly used linkers in coordination chem- 4F/MoO3/Ag) showed a reduction of 23.12% in the average
istry is the deprotonated carboxylic acid group (COOH), as PCE (down to 7.18% from 9.34%). Using the same BHJ combi-
found in biphenyl-4,4 0-dicarboxylic acid (BPDC) and nation, Xin Song et al. recently reported a higher face-on molecu-
naphthalene-2,6-dicarboxylic acid (NDC). In both of these small lar orientation, improving the average PCE from 9.41% to 12.81%
organic molecules, the carboxylate groups are oriented at opposite by optimizing the additive concentration.[5] Luye Cao et al.
ends of the molecule and as such may act as anchoring sites to reported OPV devices having PTB7-Th:IEICO-4F as the BHJ
bind with the metal oxide surface through surface metal ions showing an average PCE of 11.39%, or 12.72% when using a third
(Zn2þ) or oxygen defect centers, and may eventually form a flat component in a ternary system.[19] Jianqiu Wang et al. reported a
or tilted thin layers across the metal oxide surface.[14,15] Studies PCE around 11.13% by optimizing the donor acceptor ratio of
show that BPDC binds to the surface with one anchoring carbox- PTB7-Th:IEICO-4F.[20] Approaches such as solvent additive con-
ylate group in an upright or tilted standing orientation, while NDC centrations, donor–acceptor ratio optimizations, and third compo-
coordinates to the ZnO surface at both carboxylate groups, result- nent adding in BHJ could improve the overall device performance
ing in a planar arrangement.[15] Small organic molecules such as for PTB7-Th:IEICO-4F, whereas this ETSL approach could be
benzoic acid, biphenyl-4-carboxylic acid, or terphenyl-4-carboxylic applied for different nonfullerene polymer donor–acceptor BHJ
acid on metal oxides may not form highly ordered self-assembled molecular conformation improvement for enhancing the organic
monolayers due to the absence of a second carboxylic acid group solar cell efficiency.
which would benefit the formation of intermolecular hydrogen
bonds and leading to long-range crystallinity.[16] In the case of
BPDC, an upright standing orientation on different metals or 2. Results and Discussion
metal oxide films is feasible, in which a free carboxylic acid
group is retained, facilitating H-bonding and increased First-principles density functional theory (DFT) calculations were
crystallinity.[15–17] In addition, enhanced H-bond formation may carried out using the projector augmented wave (PAW)[21]
be feasible between the free carboxylic acid to the fluorine atoms pseudopotentials approach as implemented in the Vienna Ab
present in either the donors or acceptors of the BHJ.[18] However, initio Simulation Package (VASP).[22] The interaction between
in NDC and other planar conjugated phenyl ring systems may dis- the valence electrons was first described by the generalized
play a tendency to form a flat orientation onmetal oxides, resulting gradient approximation (GGA) as formulated in the
in a dense long-range self-assembled layer.[15–17] Therefore, BPDC Perdew–Burke–Ernzerhof (PBE) density functional followed by
and NDC potentially provide ideal contrasts that allow for a the HeydScuseriaErnzerhof (HSE06) hybrid functional.[23]
detailed study of the nature of the molecular coordination to A kinetic energy cutoff of 420 eV was used throughout. This
the ETL and the impact of the heterointerface on the molecular yielded a total energy convergence within 1meV. Integration
morphology in the BHJ layer PTB7-Th:IEICO-4F, allowing us over the Brillouin zone for the 76 atoms ZnO supercell with
to correlate this with the OPV device performance. Here, we have BPDC and NDC molecules on the slabs of ZnO with a vacuum
focused primarily on the coordination of the BPDC and on a sol– of 20 Å was performed using a 3 3 1 k-point mesh generated
gel-prepared ZnO film to produce electron transport seed layer-1 within the Monkhorst–Pack scheme, with a Gaussian smearing
(ETSL-1) and compared this to an equivalent ZnO film, electron of 0.01 eV. The electronic density of states (DOS) was computed
transport seed layer-2 (ETSL-2), and the capacity of both ETSLs to using the tetrahedron scheme. It is well known that surface mod-
influence the molecular morphology of a PTB7-Th:IEICO-4F BHJ ification of a metal oxide with interfacial layers as well as the
active layer. We found that ETSL-1 led to a higher face-on molecu- incorporation of dopants is beneficial in improving the charge
lar order in the PTB7-Th:IEICO-4F layer with a smoother and carrier collection in OPVs.[11,12] Both bulk and surface dopants
more uniform surface morphology in addition to a potential tem- or the use of organic molecules on the metal oxide surface play
plating of the BHJ. ETSL-1 was also found to reduce the interfacial a crucial role in enhancing the charge carrier collection efficiency
energy barrier between the ETSL lowest unoccupied molecular by suppressing carrier recombination; thus, the defect engineer-
orbital (LUMO) and the BHJ acceptor state. In contrast, bulk pas- ing significantly affects the electronic band structure
sivation of ETSL-2 was observed due to the agglomerated planar properties.[10–12] Therefore, DFT calculations were used to study
molecular stacking of NDC, reducing the face-on molecular the surface chemistry of ZnO, ETSLs. The optimized geometry
Sol. RRL 2022, 2200184 2200184 (2 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
of BPDC and NDC on the ZnO surface (Figure S1, Supporting Bader charge analysis was conducted to get more insight into
Information) in which the –COOH groups bind more strongly the charge transfer and distribution around the surface of ZnO
with the ZnO surface atoms. The interaction at ZnO/NDC upon adsorption of BPDC and NDC molecules. It shows higher
and ZnO/BPDC interfaces, indicating charge transfer of NDC numerical values of charge distribution for BPDC and NDC in
and BPDC with ZnO. As shown in Figure S1, Supporting vertical and horizontal orientations on the ZnO surface, respec-
Information, the plane-averaged charge density difference (Δρ) tively (Table S1, Supporting Information). Thus, BPDC favors a
for BPDC/NDC was calculated using Equation (1) vertical geometric attachment upon adsorption on a ZnO surface,
whereas NDC attaches horizontally.
Δρ ¼ ρðZnO=BPDCorNDCÞ  ρðZnOÞ  ρðBPDCorNDCÞ An inverted BHJ device configuration was chosen to facilitate
(1) the study of the ETSL/BHJ interface in addition for a superior
device stability. A sol–gel-prepared ZnO electron extraction layer
(30 nm) was treated with BPDC and NDC to generate ETSL-1 and
where ρ(ZnO/BPDC or NDC), ρ(ZnO), and ρ(BPDC or NDC) are ETSL-2, respectively. The PTB7-Th:IEICO-4F (100 nm) layer was
the charge densities of ZnO/BPDC, NDC heterostructure, and spin-coated on the BPDC/NDC-modified ZnO layer. A MoO3
its constituent reference ZnO and BPDC, NDC in the same het- hole transport layer (HTL, 10 nm) was deposited on the active
erostructure, respectively. Figure S1, Supporting Information, layer using a shadow mask in a high vacuum thermal evaporator
shows that charge distribution at the ZnO/NDC and ZnO/ followed by Ag metal layer deposition. The cell area was
BPDC interfaces, indicating a strong interaction of the BPDC 0.045 cm2. The schematic of a device is given in Figure 2.
and NDC with ZnO. However, the BPDC in horizontal and The detailed experimental procedure is summarized in
NDC in vertical configurations show electron depletion; thus, Section S1, Supporting Information.
repulsive charge interaction which was found to be pronounced To understand and evaluate the detailed molecular morphol-
for the NDC than BPDC (Table S1, Supporting Information). ogy of the PTB7-Th:IEICO-4F layer grown on ZnO, ETSL-1, and
However, as shown in Figure 1, the DOS for both BPDC and ETSL-2, synchrotron radiation-based grazing-incidence wide-
NDC on the surface of ZnO results in many defects states within angle X-ray scattering (GIWAXS) characterization analysis was
the bandgap region compared to pristine ZnO. The hybrid DFT- performed (Figure 3). The 2D GIWAXS patterns are shown in
calculated bandgap of pristine ZnO was found to be 3.4 eV, close Figure 3a–c, with 1D in-plane (IP) and out-of-plane (out of plane)
to the experimental value of 4 eV, as shown in Figure S3, profiles in Figure 3d–e. In the PTB7-Th:IEICO-4F film grown on
Supporting Information. While adsorption of both the BPDC ZnO (ZnO/PTB7-Th:IEICO-4F), we found a broad out-of-plane
and NDC on ZnO resulted in more electronic states compared peak at q ¼ 1.77 Å1z corresponding to the (010) planes (π–π
to pristine ZnO, it was found that BPDC significantly lowers the stacking) of IEICO-4F, a small (100) peak at 0.3 Å1 (out of plane)
density of defect states compared to NDC. The combination of from the PTB7-Th, and a (100) peak (lamellar stacking) at 0.3 Å1
altered electronic states and the tendency to template the BHJ (in-plane) due to PTB7-Th, indicating a mixed face-on/
molecular morphology may provide a significant improvement edge-on molecular morphology of the film (Figure 3).[24,25]
in the carrier separation in OPV devices and, thus, greater Some higher-order scattering peaks at 0.6 and 0.9 Å1 corre-
efficiency. sponding to the (200) and (300) planes are also present in the
1D in-plane line profile (Figure 3d), suggesting lamellar stacking
of the PTB7-Th; this may enhance the edge-on stacking in the
ZnO/PTB7-Th:IEICO-4F film.[24] For the ETSL-1/PTB7-Th:
IEICO-4F film, the (010) π–π stacking peak of IEICO-4F was
shifted higher at q ¼ 1.8 Å1z with a complete absence of the
(200) and (300) higher orders of PTB7-Th lamellar stacking,
along with the presence of the IP (100) peak at 0.3 Å1
(Figure 3d). A more pronounced (100) lamellar peak at
0.3 Å1 in the out-of-plane direction is clearly observable due
to a highly mixed face-on/edge-on molecular morphology.[24–26]
A key finding is the absence of the (200, 300) higher order lamel-
lar peaks, which indicates less lamellar stacking, corresponding
to lower edge-on and an increased face-on molecular morphol-
ogy. In contrast, the reduced intensity of the (100) peak at
0.3 Å1 in the out-of-plane direction observed for ETSL-2/
PTB7-Th:IEICO-4F films indicates a decreased face-on stacking.
We calculated the full width at half maximum (FWHM) and
crystal coherence length (CCL) of the significant peaks in both
the in-plane and out-of-plane directions from their GIWAXS
analysis. All the FWHM values, CCL, and peak positions for
the (010) and (100) peaks are summarized in Table 1. The
Figure 1. The total DOS of BPDC in vertical and NDC in horizontal π–π CCL was calculated using the Scherrer equation
covered ZnO surface compared to a pure ZnO surface calculated using
hybrid DFT. CCL ¼ 2πK=β (2)
Sol. RRL 2022, 2200184 2200184 (3 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
Figure 2. a) Device schematic of glass/ITO (cathode)/ETSL (electron transport seed layer)/PTB7-Th:IEICO-4F (BHJ)/MoO3 (hole transport layer)/Ag
(anode) inverted OPV device, b) nonfullerene acceptor (IEICO-4F), c) chemical structure of the donor (PTB7-Th), d) biphenyl dicarboxylic acid (BPDC),
and e) naphthalene dicarboxylic acid (NDC).
Figure 3. 2D GIWAXS patterns of a) ZnO/PTB7-Th:IEICO-4F (CN additive), b) ETSL-1/PTB7-Th:IEICO-4F (CN additive), and c) ETSL-2/PTB7-Th:IEICO-
4F (CN additive). 1D line profiles for d) in-plane figures for ZnO/PTB7-Th:IEICO-4F, ETSL-1/PTB7-Th:IEICO-4F, and ETSL-2/PTB7-Th:IEICO-4F (CN
additive). e) Out-of-plane figures for ZnO/PTB7-Th:IEICO-4F, ETSL-1/PTB7-Th:IEICO-4F, and ETSL-2/PTB7-Th:IEICO-4F (CN additive). f ) Face-on/
edge-on ratio for ZnO/PTB7-Th:IEICO-4F, ETSL-1/PTB7-Th:IEICO-4F, and ETSL-2/PTB7-Th:IEICO-4F (CN additive).
Sol. RRL 2022, 2200184 2200184 (4 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
Table 1. Crystal coherence length analysis from GIWAXS experiment optimize the selectivity of the additives, we performed
for ZnO/PTB7-Th:IEICO-4F (CN additive), ETSL-1/PTB7-Th:IEICO-4F GIWAXS to evaluate the effects of two of the mostly used dop-
(CN additive), and ETSL-2/PTB7-Th:IEICO-4F (CN additive). ants, chloroform (CF), and CN in the molecular ordering of the
PTB7-Th:IEICO-4F BHJ layer. We found a more negative HOF
Film Peak position FWHM Coherence value of 0.06 0.01 for the ETSL-1/PTB7-Th:IEICO-4F (CF
[In-plane (100)/ length [Å] additive) (Figure S2, Supporting Information). Thus, CN was
out-of-plane (010)]
found to be a more suitable additive to the PTB7-Th:IEICO-4F
ZnO/PTB7-Th:IEICO-4F 0.30 (in-plane) 0.03 186.40 BHJ for high face-onmolecular order and used for all subsequent
ETSL-1/PTB7-Th:IEICO-4F 0.30 (in-plane) 0.05 111.84 samples. The GIWAXS analysis shows that the molecular mor-
ETSL-2/PTB7-Th:IEICO-4F 0.31 (in-plane) 0.04 139.80 phology of the PTB7-Th:IEICO-4F is highly correlated with the
ZnO/PTB7-Th:IEICO-4F 1.77 (out-of-plane) 0.41 13.63 ETSL-1 heterointerface. Thus, the interaction of the BPDC/NDC
ETSL-1/PTB7-Th:IEICO-4F 1.81 (out-of-plane) 0.48 11.65 with ZnO is crucial for controlling the PTB7-Th:IEICO-4F molec-
ular morphology.
ETSL-2/PTB7-Th:IEICO-4F 1.82 (out-of-plane) 1.32 4.32 To understand the surface chemistry of the small organic mol-
ecules with the ZnO surface, attenuated total reflectance-Fourier
transform infrared (ATR-FTIR) spectroscopy was employed to
where K is the dimensionless shape factor (0.9) and β is the evaluate the interfacial composition of the ETSL. The out-of-
FWHM (Å1).[24,26] It was found that the CCL (111.84 Å) for plane deformation mode of the phenyl rings, γoop(CH) at
the (100) lamellar peak in the in-plane mode of ETSL-1/PTB7- 764 cm1 (ETSL-2), is polarized perpendicular to the aromatic
Th:IEICO-4F was lower than ZnO/PTB7-Th:IEICO-4F ring and originates from the planar agglomeration of NDC on
(186.40 Å). A close CCL values of the (010) π–π stacking peak the ZnO surface (Figure 4).[14,15,17] Meanwhile, the lower inten-
(11.65 and 13.63 Å) in the out-of-plane were observed for both sity of γoop(CH) in ETSL-1 indicates a tilted arrangement of the
samples. The CCL of the (100) lamellar peak reached to BPDC to the ZnO surface.[14,15,17] The band at 1400 cm1, which
139.80 Å, with a big reduction to just 4.32 Å for the (010) π–π contains contributions from both the OH deformation mode
stacking peak in the ETSL-2/PTB7-Th:IEICO-4F. A lower CCL δ(OH) of the free acid and the symmetric OCO stretching mode
value of the out-of-plane (010) π–π stacking peak and a higher νs(OCO) of the carboxylate, shows a high intensity in ETSL-2 and
value for the (100) in-plane lamellar peak in ETSL-2/PTB7-Th: a lower intensity in ETSL-1, which may indicate a lower amount
IEICO-4F are dominant with increased degree of edge-on orien- of free acid and carboxylate. The high intensity of the ν(C═O,
tation, whereas the smallest CCL of the in-plane (100) lamellar stretching) band at 1663 cm1 is consistent with H-bonded
peak of ETSL-1/PTB7-Th:IEICO-4F is consistent with less edge- chains of head-to-head coupled NDCs lying flat on the ZnO in
on orientation, critical in the device efficiency. We attempted to ETSL-2. This band has a lower intensity for ETSL-1 due to
quantify the face-on/edge-on ratio in the BHJ film using the reduced H-bonding between the molecules.
Hermans’ orientation function (HOF),[24] defined as To further evaluate the BPDC/NDC coordination on the ZnO
surface, we employed X-ray photoemission spectroscopy (XPS) to
HF 2hkl ¼ 3ðcos φhklÞ–1=2 (3)Z Z evaluate the ZnO, ETSL-1, and ETSL-2 spectra and determine theπ π corresponding binding energies, allowing us to identify the pres-
ð 2 2cos2φ 2hklÞ ¼ IðjÞsinj cos jdj= IðjÞsinj dj (4) ence of oxygen vacancies and to correlate it to the coordination
0 0 properties observed in ATR-FTIR and DFT studies.
where φ is the azimuthal angle relative to the incident X-ray Figure 5 demonstrates the XPS analysis of ZnO, ETSL-1, and
beam and IðjÞ is the intensity of the (hkl) reflection along ETSL-2. In Figure 5a, two binding energies are observed in the
φ.[24] HOF determines the orientation of a crystallite having a
specific Miller index reflection by fitting Equations (3) and (4)
to the intensity distribution along φ relative to the thin film sub-
strate, where φ¼ 0 is normal to the substrate.[24] Values of HOF
range from 1 to 0.5. If all the signal intensity for a given Miller
index reflection is observed exclusively out of plane to the sub-
strate (φ¼ 0), then the HOF will yield a value of 1.[24] If all of the
signal intensity is IP to the substrate (φ¼ 90), then the HOF will
be equal to 0.5.[24] HOF values obtained from the films were
around 0.157 0.01 for the ZnO/PTB7-Th:IEICO-4F film
and 0.016 0.01 and 0.143 0.01 for the modified ETSL-
1/PTB7-Th:IEICO-4F and ETSL-2/PTB7-Th:IEICO-4F films
(Figure 3f ).
The HOF value for ETSL-1/PTB7-Th:IEICO-4F is closest to 1,
which relates to the previous observation of the high face-on
orientation within the mixed face-on/edge-on molecular mor-
phology. It should be noted that chloronapthalene (CN) was used
as additives inside the active layer for all the samples. To Figure 4. ATR-FTIR spectrum of the ETSL-1 and ETSL-2.
Sol. RRL 2022, 2200184 2200184 (5 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
Figure 5. XPS analysis of a,d) O 1s, Zn 2p3 core label spectra of ZnO, b,e) O 1s, Zn 2p3 core label spectra of ETSL-1, and c,f ) O 1s, Zn 2p3 core label
spectra of ETSL-2.
deconvolution of the O 1s region (O 1s B to higher energy and O  0.02 eV (Figure S4, Supporting Information). To evaluate the
1s A 1.5 eV lower in energy). The O 1s A peak (530.13 eV, conduction band (CB) and valence band (VB) positions, we
Figure 5a) is associated with oxygen atoms bonded to the nearest obtained the ultraviolet photoelectron spectra (UPS) of ZnO,
Zn atoms in the ZnO lattice, which shifts to 530.32 eV in ETSL-1 ETSL-1, and ETSL-2 films. The cutoff binding energy (Ecutoff )
(Figure. 5b) due to the interaction between the carboxylate group measured (Figure S4a, Supporting Information) for ZnO,
in BPDC.[27] A much higher binding energy, 530.92 eV, for O 1s ETSL-1, and ETSL-2 films was 17.85, 18.05, and 18.80 eV, respec-
A was observed for ETSL-2, indicating an enhanced coordination tively, while the Fermi levels (EF) of ZnO, ETSL-1, and ETSL-2
in NDC/ZnO (Figure 5c), which reflects the extent of bonding thin films were calculated to be 3.37, 3.17, and 2.42 eV,
between the carboxylate group of the NDC to Zn2þ. The O 1s respectively, using E ¼ (21.22 eV E ).[30]F — cutoff The Fermi edge
B peak (531.63 eV) relates to the oxygen vacancies or oxygen- (EF,edge) was estimated to be around 5.31, 5.31, and 5.28 eV
deficient regions in ZnO, which shifts to 531.84 eV in ETSL-1 for ZnO, ETSL-1, and ETSL-2 (Figure S4b, Supporting
and 532.18 eV in ETSL-2, indicating a stronger interaction with Information), from which the VB was calculated by using the for-
the carboxylate group in ETSL-2 (Figure 5a–c).[28,29] Overall, the mula VB¼ (E [30]F—EF,edge). Thus, 8.68, 8.48, and 7.7 eV
O 1s peaks increase in BE by 0.2 eV between ZnO and ETSL-1 were the VB values and the CB energies were calculated to be
and 0.6 eV between ETSL-1 and ETSL-2. Also, the ratio of 4.68, 4.48, and 3.8 eV in ZnO, ETSL-1, and ETSL-2, respec-
the relative intensities of O 1s A and O 1s B is 1.5 in ZnO, tively, according to the formula of CB¼ (VBþ Eg).[30] The analy-
1.9 in ETSL-1, and 2.1 in ETSL-2, indicating that oxygen vacan- sis of the energy levels for all the materials (Figure S4e,
cies are more pronounced in ETSL-2 than ETSL-1. The peak at Supporting Information) used in this study has shown that
1021.41 eV (Figure 5d) for the sol–gel-deposited ZnO may origi- the CB energy around 4.48 eV (ETSL-1) is much nearer to
nate from the Zn2þ state, which shifts to slightly higher energies the LUMO of the IEICO-4F acceptor (4.19 eV), which is
in ETSL-1, 1021.71 eV (Figure 5e) and ETSL-2, 1022.17 eV expected to be significant in reducing interface barrier energy
(Figure 5f ).[27] Our DFT analysis also shows a similar observation and exciton recombination. Whereas, in case of ETSL-2, a high
for the BPDC and NDC coordination modes to ZnO. The interfacial barrier energy to PTB7-Th:IEICO-4F BHJ can be seen
bandgap energies (Eg) for ZnO, ZnO/BPDC, and ZnO/NDC in Figure S4e, Supporting Information. Thus, an enhanced
films were calculated using a Tauc plot (Figure S3, energy barrier in ETSL-2 does not provide an efficient electron
Supporting Information) obtained from their UV–vis absorption extraction and may reduce device efficiency. Also, strongly
spectra (Figure S4g, Supporting Information); a similar bandgap agglomerated NDC molecules could abruptly change the wetta-
was observed for ZnO and ETSL-1 but a slightly smaller bandgap bility with PTB7-Th:IEICO-4F, whereas an optimized smooth
was observed for ETSL-2. The bandgaps of the ZnO, ETSL-1, and heterointerface may occur in the case of ETSL-1. We employed
ETSL-2 were 4, 4, and 3.9 eV, respectively, each with an error of atomic force microscopy (AFM) and contact angle (CA) analysis
Sol. RRL 2022, 2200184 2200184 (6 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
(Figure 6c) displayed an abrupt change in the surface morphol-
ogy, with an even lower RMS roughness of 5.2 1.5 nm. From
Figure 7a, it can be observed that a highly rough ZnO surface
with numerous nanoridges may not form an ideal seed layer
for the PTB7-Th:IEICO-4F π–π molecular morphology, but
ETSL-1 without nanoridges and having a smoother surface
(Figure 7b) is an ideal seed layer for a π–π stacked molecular mor-
phology in the BHJ. Thus, the lower surface roughness of ETSL-2
led to an overall decrease in the crystallinity in the BHJ. From the
GIWAXS analysis (Figure 3), it was observed that crystallinities
of the BHJ are strongly correlated with surface morphologies of
the ZnO/ETSL-1/ETSL-2 layers. In case of ETSL-2, a planar and
flat orientation of the NDC molecules on ZnO significantly
reduces the surface roughness. The flat NDC molecules may
have higher interaction with the BHJ, which could negatively
impact the crystallinity.[33,34] Also, this enhanced interaction
between the BHJ and the flat-capped NDC molecules of ETSL-
2 may reduce donor–acceptor phase separation, along with
reduced crystallinities of the BHJ.[10,35] However, in the case
of ETSl-1, a tilted conformation of the BPDC molecules with
COOH functional groups could initiate H bonding with the
BHJ which would improve its crystallinity.[36] However, a greater
face-on morphology is significant for the OPV device perfor-
mance and was observed in the PTB7-Th:IEICO-4F layer. In
Figure 6. AFM (phase, 3D view) images of a,d) ZnO, b,e) ETSL-1, and
c,f ) ETSL-2. addition, a similar surface energy of ETSL-1 with the PTB7-
Th:IEICO-4F layer may result in an enhanced heterointerface
in comparison to both ZnO and ETSL-2.[10]
The surface energies of the ZnO, ETSL-1, ETSL-2, and PTB7-
to measure the surface roughness and the surface energy of the Th:IEICO-4F layers weremeasured by taking the CA (Figure 8a–h).
thin films of ZnO, ETSL-1, ETSL-2, and PTB7-Th:IEICO-4F, It can be observed that the values for ZnO, ETSL-1, and ETSL-2
respectively. Figure 6a shows the surface morphology of the were measured to be 94, 92.5, and 84.5 using water droplets.
ZnO ETL to be nonuniform and consisting of multiple As the surface roughness is proportional to the CA, this agrees
nanoridges. It should be noted that this nonuniform surface with our AFM analysis. ETSL-1 does not drastically change the
containing nanoridges[31] may not be an optimal interface with surface roughness of the ZnO thin film, but a complete passiv-
an upper BHJ active layer; this may reduce the charge extraction ation by NDC greatly reduces the surface roughness and CA
and increase the defect-related recombination properties.[32] Our value (Figure 8c). Most importantly, the CA for ZnO/PTB7-
XPS data also show that the ZnO surface also contains multiple Th:IEICO-4F, ETSL-1/PTB7-Th:IEICO-4F, and ETSL-2/PTB7-
Zn2þ metal ions and oxygen defect zones. The root mean square Th:IEICO-4F (Figure S5a–c) follows a similar trend with ZnO,
(RMS) roughness of the ZnO surface was measured to be ETSL-1, ETSL-2 (Figure 8a–c). Thus, the growth of the PTB7-
23.43 1.5 nm. In contrast, ETSL-1 showed a significant reduc- Th:IEICO-4F strongly depends on the ZnO, ETSL-1, or ETSL-
tion in the nanoridges with a more uniform surface and an RMS 2 surface morphology, which was also found in the GIWAXS
roughness of 15.90 1.5 nm (Figure 6b). ETSL-2 by contrast analysis. An optimal heterointerface between the BHJ and charge
Figure 7. Schematic representation of the morphology of BHJ layer on a) ZnO (rough surface) and b) ETSL-1, and c) ETSL-2.
Sol. RRL 2022, 2200184 2200184 (7 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
Figure 8. CA analysis of a) ZnO, b) ETSL-1, c) ETSL-2, and d) PTB7-Th:IEICO-4F using deionized water droplets, and e) ZnO, f ) ETSL-1, g) ETSL-2, and
h) PTB7-Th:IEICO-4F using diiodomethane droplets.
collecting layer is obtained with a minimal surface energy differ- Table 2. CA analysis of ZnO, ETSL-1, ETSL-2, and PTB7-Th:IEICO-4F by
ence in the corresponding films (ETSL-1, PTB7-Th:IEICO-4F).[10] using DI water and diiodomethane droplets.
All of the surface energies were calculated based on Fowkes
theory.[37] Sample Contact Contact angle Surface Surface energy Surface energy
    angle [diiodomethane] energyγS (polar)γpS (dispersive)γd
D D 1
S
γ γ 2 þ γPγP 1S S 2l l ¼ γlðcos θ þ 1Þ=2 (5) [water] [mNm1] [mNm1] [mNm1]
ZnO 94 38.5 36.70 0.170 36.53
Initially, we used diiodomethane liquid which has a relatively
1 ZnO/ 92.5 44.9 37.32 0.268 37.06high overall surface tension of 50.8mNm and no polar com- ETSL-1
ponent to its overall surface tension. Thus, γDS could be measured ZnO/ 84.5 42.3 39.81 1.38 38.43
based on the CA and the dispersive component determined. ETSL-2
γD ¼ ðγ =4Þðcos θ þ 1Þ2 (6) PTB7-Th: 101 43.1 38.37 0.36 38.01S l
IEICO-4F
In the second step, a nondispersive (polar) component such as
water which has γp ¼ 46.4mNm1l , and γDl ¼ 26.4mNm1,
along with the CA on the solid surface could be used to find
out γPS . Thus, the total surface energy could be measured by using
Equation (5). shown in Figure 9a. D1 showed a high efficiency, with an average
of 11.20% PCE (PCE max 12.53%) and an average Voc of 0.70 V
γ D Ps ¼ γS þ γS (7) and Jsc of 25.2 mA cm2 and FF of 62.40%, whereas a PCE of
9.34%, 0.66 Voc, 21.80mA cm
2 Jsc, 61.44% FF were
where γs, γ Dl, γS , γPS , γD
p
l , γl represent the surface energies of solid, measured for D0. A dramatic lowering of the PCE to around
liquid, dispersive, and polar components of the solid and liquid, 7.18% PCE, 22.85mA cm2 Jsc, and 46.11% FF was
respectively. All of the CA images for diiodomethane solvent measured for D2. Consequently, it can be concluded that the
have been provided in Figure 8e–h for ZnO, ETSL-1, ETSL-2, ETSL-1/PTB7-Th:IEICO-4F configuration enhanced the device
and PTB7-Th:IEICO-4F thin films. The surface energy, polar, performance. The averaged data for all the cells (eight cells) is
and dispersive values have been provided in Table 2. We found summarized in Table 3. We found that series resistance (RSh)
that the surface energy for ETSL-1 is much closer to the surface of D2 was much higher than D1 and D0, which suggests
energy of PTB7-Th:IEICO-4F. Thus, a similar surface energy less n-type conductivity of the electron extraction layer and a less
between ETSL-1 and PTB7-Th:IEICO-4F thin film relates to an efficient charge extraction at the interface of the BHJ and the
enhanced heterointerface between them and is more favorable electron extraction layer, consistent with the lower average FF
for a higher face-on arrangement of the PTB7-Th:IEICO-4F thin value (Table 3).[27]
film.[10] D1 demonstrated around 1.08% and 54.7% lower average
series resistance (RS) compared to D0 and D2, respectively
2.1. Device Performance (D0: 4.60Ω cm2, D1: 4.55Ω cm2, D2: 10.06Ω cm2) and
nearly 6.4% and 162.06% higher average shunt resistance
We fabricated three different OPV devices D0: (glass/ITO/ (RSh) compared to D0 and D2 (D0: 500Ω cm2, D1:—
ZnO/PTB7-Th:IEICO-4F/MoO3/Ag), D1: (glass/ITO/ETSL-1/ 532Ω cm
2, D2: 203Ω cm2). Generally, the FF of a solar cell
PTB7-Th:IEICO-4F/MoO3/Ag), and D2: (glass/ITO/ETSL- can be mathematically expressed by the following empirical
2/PTB-Th:IEICO-4F/MoO3/Ag), with their (J–V ) characteristics formula.
[27,38]
Sol. RRL 2022, 2200184 2200184 (8 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
Figure 9. a) J–V analysis, b) EQE analysis, c) PL spectrum analysis of ZnO/PTB7-Th:IEICO-4F (CN additive), ETSL-1/PTB7-Th:IEICO-4F (CN additive),
and ETSL-2/PTB7-Th:IEICO-4F (CN additive).
Table 3. Photovoltaic performances of D0, D1, D2 devices (average of eight cells) under AM1.5 G 100mA cm2 Device.
Device Composition Voc [V] Jsc [mA cm
2] FF [%] PCE [%] Rs [Ω cm2] R 2shunt [Ω cm ]
D0: aver.best Glass/ITO/ZnO/PTB7-Th: 0.66 (0.05)0.69 21.60 (0.20)24.35 61.44 (0.15)63 9.34 (0.10) 10.74 4.60 (0.05)5.22 500 (5)637
IEICO-4F/MoO3/Ag
D1: aver.best Glass/ITO/ETSL-1/PTB7-Th: 0.70 (0.05)0.71 25.2 (0.2)26.6 62.40 (0.15)65.11 11.20 (0.10)12.53 4.55 (0.05)4.63 532 (5)661
IEICO-4F/MoO3/Ag
D2: aver.best Glass/ITO/ETSL-2/PTB7-Th: 0.71 (0.05)0.72 23 (0.2)22.85 48.10 (0.15)46.11 7.18 (0.10)7.81 10.06 (0.05)9.86 203 (5)233
IEICO-4F/MoO3/Ag
  
¼  1.1RS þ R
2
s where RS, RCH, Voc, and RSh denote series resistance, the char-FF FF0 1
 R 5.4R2  	 acteristic resistance (RCH¼ VOC/JSC), normalized open-circuitCH CHþ (8)V 0.7 FF R 1.1R R2 voltage (Voc¼ VOC/nVT, VT, and n refer to thermal voltage 1 OC  0 CH 1 s þ s2 and ideality factor of the cell), and shunt resistance of the solarVOC RSh RCH 5.4R CH cell, respectively. A reduced RS and enhanced RSh value in D1
devices led to a larger FF than D0 and D2 devices
(Table 3).[27] We found that a reduction of series resistance,
V
FF ¼ OC  lnðVOC þ 0.7Þ (9) Rs, in the D1 device was related to the enhanced face-on molecu-0 VOC þ 1 lar morphology of PTB7-Th:IEICO-4F film atop the ETSL-1, as
Sol. RRL 2022, 2200184 2200184 (9 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
observed from GIWAXS analysis. Also, the high values of RSh, in D1. By contrast, D2 (glass/ITO/ETSL-2/PTB7-Th:IEICO-4F/
JSC, and FF in D1 compared to D0 and D2 indicate less charge MoO3/Ag) showed a reduction of 23.12% in the average PCE
carrier recombination in the active layer from our J–V analy- (down to 7.18% from the 9.34% reference). The reduction in per-
sis.[39] In addition, an external quantum efficiency (EQE) analy- formance of D2 was correlated to the decreased face-on molecu-
sis, Figure 9b, was employed to see the spectral response of the lar order in the BHJ, as identified in the GIWAXS analysis.
films. A strong spectral response from all the films was seen, Devices prepared using another nonfullerene acceptor donor
whereas an increased intensity was observed for device D1 (PTB7-Th:ITIC-4F) BHJ layer also showed improved
between 600 and 900 nm wavelength. This observation matches performance when using ETSL-1 compared to bare ZnO.
our J–V analysis where we can see an increasing trend of the Consequently, this work clearly shows that ETSLs can signifi-
current density ( Jsc). Steady-state photoluminescence (PL) analy- cantly improve the performance of OPV solar cells by serving
sis (Figure 9c) was carried out to analyze the quenching behavior as a BHJ molecular ordering modulator.
of the as-prepared devices. A stronger quenching at 885 nm for
ITO/ETSL-1/PTB7-Th/IEICO-4F (D1) could be observed due to a
high dissociation of the excitons. This strong quenching origi- 4. Experimental Section
nates from the enhanced heterointerface, which increases the
charge transfer between the subsequent electrodes and reduces Materials: The prepatterned ITO glass with an area of 12 12mm was
[40] purchased from Lumtec. PTB7-Th, IEICO-4F, and ITIC-4F were purchasednonradiative recombination processes. D1 showed a 42% from 1-Materials. Zinc acetate dihydrate (Zn(CH3COO)2·2H2O, Sigma-
decrease of the PCE after 18 days inside the glove box, but a high Aldrich, >99.0%), 2-methoxyethanol (CH3OCH2CH2OH, Sigma-Aldrich,
reduction in PCE (46%) for D0 was measured under the same 99.8%, anhydrous), ethanolamine (NH2CH2CH2OH, Sigma-Aldrich,
conditions (Figure S6, Supporting Information). We performed >99.5%), 1-chloronaphthalene (CN), biphenyl-4,4 0-dicarboxylic acid
a similar device fabrication approach with another BHJ (PTB7- (C14H10O4, 99%), naphthalene-2,6-dicarboxylic acid (C12H8O4, 99 %),
Th:ITIC-4F) system, where two kinds of devices were prepared, CF, and MoO3 were purchased from Sigma-Aldrich.
Device Fabrication: An inverted device structure of ITO glass/ZnO/
labeled as D3 (ITO/ZnO/PTB7-Th:ITIC-4F) and D4 (ITO/ETSL- PTB7-Th:IEICO-4F/MoO3/Ag (reference device) was used in the study,
1/PTB7-Th:ITIC-4F), respectively. We observed that D4 showed with slight modification based on our previous studies.[41] The ITO glass
enhanced device performance (Figure S7, Supporting substrate was cleaned with soapy deionized (DI) water, pure DI water,
Information) over D3 (Table S2, Supporting Information), and acetone, and isopropanol after an ultrasonication for 10min in sequence.
their GIWAXS analysis (Figure S8, Supporting Information) A 0.4 M sol–gel ZnO solution was prepared by dissolving zinc
showed a higher face-on molecular morphology. Thus, ETSL-1 acetate dihydrate (Zn(CH3COO)2·2H2O, Sigma-Aldrich, >99.0%) in
2-methoxyethanol (CH3OCH2CH2OH, Sigma-Aldrich, 99.8%, anhydrous)potentially enhances the molecular morphologies of the BHJ with an additive ethanolamine (NH2CH2CH2OH, Sigma-Aldrich,>99.5%)
layer, along with passivation of the defect centers of the metal and kept at 250 rpm stirring inside the glove box for 12 h. PTB7-Th (10mg)
oxide, significant aspects in improving the OPV device efficiency. and IEICO-4F (15mg) in a 1:1.5 wt ratio in a 25mgmL1 chlorobenzene
solution with 4 vol% CN were prepared and kept at 200 rpm stirring under
a dark condition inside the glove box for 12 h. After stirring, ZnO solution
3. Conclusions was spin-coated on ITO substrate at 4000 rpm for 60 s. A hot plate air
annealing at 170 C for 40min was applied to the ZnO thin films. Two
In conclusion, ETSL-1was used as a modified electron extraction 18.5mM solution of DMF, biphenyl-4,4
0-dicarboxylic acid (compound
layer to control the molecular orientation and crystallization of A) and DMF, naphthalene-2,6-dicarboxylic acid (compound B) were pre-
pared after an ultrasonication of 1 h, 30min heat treatment at 120 C
the PTB7-Th:IEICO-4F BHJ layer. As a result of the controlled inside the glove box. Compound A was spin-coated on the ZnO thin film
molecular order of the BHJ layer and a good interface with by spin coating at 3000 rpm for 40 s and kept at room temperature con-
ETSL-1, the photon absorption and charge dissociation were ditions for 10min inside the glove box for ETSL-1 layer preparation.
improved, which are beneficial for overall device performance. Compound B was also spin-coated on the ZnO thin film by following
The enhanced device performances were also highly correlated the exact conditions for ETSL-2 preparation. The stirred active layer solu-
to the increased degree of face-on molecular ordering in the tion inside the N2-filled glove box was spin-coated at a rate of 2800 rpm for
1min onto three kinds of samples (ZnO, ZnO:/ETSL-1, and ZnO/ETSL-2)
BHJ layer, shown by GIWAXS analysis. In addition, the elimina- on ITO glass substrate to reach 100 nm thickness. Right after that, the
tion of multiple surface nanoridges in ETSL-1 played a great role coated samples were put into a vacuum chamber at a pressure of
in improving the wettability with the BHJ layer, while the more 105 Pa. The 10 nm-thick film of MoO3 and 100 nm-thick film of silver were
similar surface energies of ETSL-1 and PTB7-Th:IEICO-4F BHJ deposited on the sample surface through a shadow mask by thermal evap-
made for a more effective interface, lowering the carrier recom- oration. All the fabrication steps excluding the BPDC and NDC spin coat-
bination (high shunt resistance) and increasing the PL quench- ing here in this report were followed from our previous publication with
little modification.[41] The recipe for sol–gel-prepared ZnO usually provides
ing, giving rise to an increased current density and PCE for an average 30–40 nm-thick layer, whereas the PTB7-Th: IEICIO-4F BHJ
device D1 (glass/ITO/ETSL-1/PTB7-Th:IEICO-4F/MoO3/Ag) with 1:1.6 in CB gives around 100 nm. We focused for optimizing the spin
in comparison to D0 (glass/ITO/ZnO/PTB7-Th:IEICO-4F/ coater RPM speed of BPDC and NDC on ZnO which was found to be more
MoO3/Ag) and D2 (glass/ITO/ETSL-2/PTB-Th:IEICO-4F/ straightforward way to control the thickness. The device area fabricated
2
MoO3/Ag). An enhanced average PCE of around 11.20% was 0.045 cm . Devices for PTB7-Th:ITIC-4F were also prepared in the
(PCE ¼ 12.53%), 0.7 V V , and a J of 25.2mA cm2, with same manner with the 1:1.5 D/A ratio with 0.4 vol% CN additive andmax oc sc
an FF of 62.40% were recorded for device D1, whereas D0 coated on ZnO and ZnO/BPDC with MoO3 HTL (10 nm) and Ag(100 nm) layer formation.
showed an average PCE of 9.34%, 0.66 V , 21.80mA cm2oc Device Characterization: The current density–voltage ( J–V ) measure-
Jsc, and 61.44% FF. In total, a 19.91% increase in the average ments were conducted by using a solar cell I–V testing system from PV
PCE, from a reference value of 9.34% to 11.20%, was achieved Measurements, Inc. (using a Keithley 2400 source meter) under
Sol. RRL 2022, 2200184 2200184 (10 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
illumination power of 100mW cm2 by an AM1.5 G solar simulator. The Open access publishing facilitated by University of New South Wales,
device temperature was measured by a GM1350 50:1 LCD infrared ther- as part of the Wiley - University of New South Wales agreement via the
mometer digital gun and maintained at around 25C. GIWAXS analysis Council of Australian University Librarians.
was done at the SAXS/WAXS beamline of the Australian Synchrotron, a
part of ANSTO. All the films were deposited on ITO substrates. A
Pilatus detector was placed at628.85mm from the sample position with Conflict of Interest
13.003 keV X-rays used to give a q-range out to 2 Å1. The sample to detec-
tor distance was fixed at 628.859mm which has been kept same for all the The authors declare no conflict of interest.
samples for a q range 0.2–2 Å1. We used (q¼ 4π sin(2θ/2)/lambda),
where 2θ is the scattering angle, and lambda denotes the X-ray wavelength
(0.9537 Å) for calculating the length of the scattering vector q range which Data Availability Statement
depends on the sample to detector distance. The detector pixel size was
0.172mm. The sample-to-detector distance and X-ray wavelength were cal- The data that support the findings of this study are available from the
ibrated by silver behenate standard. Several angular steps of 0.01–0.05 corresponding author upon reasonable request.
were taken near the critical angle of 0.15 to have a higher scattering inten-
sity for bulk of the active layer film for each sample, which was determined
as the angle of maximum scattered intensity. We used Igor Pro software,
Nika macros, SAS 2D for the data reduction. Several steps such as mask- Keywords
ing, beam center refinement, and sector analysis were performed sequen-
tially applied to generate and correct the 2D pattern and 1D line profiles. bulk heterojunctions, electron transport layers, grazing-incidence wide-
GIWAXS sample alignment was done based on the detector to sample angle X-ray scattering (GIWAXS), molecular order, organic photovoltaics
position for enabling the sufficient recording of X-ray scattering at wide
angles, at (λ¼ 0.9537 Å) a particular wavelength for several incident angle Received: February 27, 2022
of the light beam. A vertical and horizontal movement of the detector Published online:
required for the sample alignment and the selective positioning of the scat-
tering pattern on the detector. An attenuated Fourier transform infrared
spectroscopy (Perkin Elmer) was used to study thin films vibrational prop- [1] G. J. Zhao, Y. J. He, Y. Li, Adv. Mater. 2010, 22, 4355.
erties. The X-ray photoelectron spectroscopy (XPS) and UPS were analyzed [2] H. Aqoma, S. Park, H. Y. Park, W. T. Hadmojo, S. H. Oh, S. Nho,
by ESCALAB250Xi instrument (Thermo Scientific, UK), with an D. H. Kim, J. Seo, S. Park, D. Y. Ryu, S. Cho, S. Y. Jang, Adv. Sci.
X-ray source of monochromated Al Kα (energy 1486.68 eV) at 160W
 2018, 5, 201700858.(14.5 11 mA). A Bruker Dimension ICON SPMwith a scan size of “5mm
 [3] S. Xiong, L. Hu, L. Hu, L. Sun, F. Qin, X. Liu, M. Fahlman, Y. Zhou,5 mm” at a scan rate of 0.512Hz was used as the AFM to characterize Adv. Mater. 2019, 31, 1.
the film surface morphology. The CA was measured by a CA goniometer
using an optical subsystem to capture the profile of DI water (polar) and [4] Y. Chang, X. Zhang, Y. Tang, M. Gupta, D. Su, J. Liang, D. Yan, K. Li,
diiodomethane (nonpolar) (5 μL) droplet on thin films for calculating the X. Guo, W. Ma, H. Yan, C. Zhan, Nano Energy 2019, 64, 103934.
surface energies. All the CAs were measured at room temperature within [5] X. Song, N. Gasparini, L. Ye, H. Yao, J. Hou, H. Ade, D. Baran, ACS
the first 30 s of depositing the droplets (water, diiodomethane) on the thin Energy Lett. 2018, 32018, 669.
film surface. A QEX10 spectral response system from PV measurements [6] H. C. Liao, C. C. Ho, C. Y. Chang, M. H. Jao, S. B. Darling, W. F. Su,
was used as a characterization machine for highly sensitive EQE measure- Today 2013, 16, 326.
ments. Optical characterization was conducted using a UV–vis–NIR [7] H. Yao, Y. Cui, D. Qian, C. S. Ponseca, A. Honarfar, Y. Xu, J. Xin,
spectrometer (Perkin Elmer-Lambda 950). The PL characterization was Z. Chen, L. Hong, B. Gao, R. Yu, Y. Zu, W. Ma, P. Chabera,
conducted using a pulsed OPO laser at 328 nm. For steady-state spectral T. Pullerits, A. Yartsev, F. Gao, J. Hou, J. Am. Chem. Soc. 2019,
PL, the signal was detected by a Glacier X TE Cooled CCD spectrometer 141, 7743.
with a detection range of 200–1050 nm. [8] X. Li, X. Liu, W. Zhang, H. Q. Wang, J. Fang, Chem. Mater. 2017, 29,
4176.
[9] W. Li, M. Chen, J. Cai, E. L. K. Spooner, H. Zhang, R. S. Gurney, D. Liu,
Supporting Information Z. Xiao, D. G. Lidzey, L. Ding, T. Wang, Joule 2019, 3, 819.
Supporting Information is available from the Wiley Online Library or from [10] H. Liu, Z. X. Liu, S. Wang, J. Huang, H. Ju, Q. Chen, J. Yu, H. Chen,
the author. C. Z. Li, Adv. Energy Mater. 2019, 9, 201900887.
[11] A. Samavati, Z. Samavati, A. F. Ismail, M. H. D. Othman,
M. A. Rahman, I. S. Amiri, RSC Adv. 2018, 8, 1418.
Acknowledgements [12] C. Tozlu, A. Mutlu, M. Can, A. K. Havare, S. Demic, S. Icli, Appl. Surf.
Sci. 2017, 422, 1129.
The authors would like to highly express their sincere gratitude to [13] T. Stubhan, M. Salinas, A. Ebel, F. C. Krebs, A. Hirsch, M. Halik,
Australian Nuclear Science and Technology Organization (ANSTO) and C. J. Brabec, Adv. Energy Mater. 2012, 2, 532.
Dr. Valerie Mitchell for helping with the GIWAXS experiment. This [14] P. M. Jones, J. A. May, E. I. Solomon, Inorg. Chim. Acta 1998,
research was undertaken on the SAXS/WAXS beamline at the 275–276, 327.
Australian Synchrotron, part of ANSTO. The authors would also like to [15] S. Mohr, T. Schmitt, T. Döpper, F. Xiang, M. Schwarz, A. Görling,
gratefully acknowledge the support from the staff of the School of M. A. Schneider, J. Libuda, Langmuir 2017, 33, 4178.
Photovoltaic and Renewable Energy Engineering (SPREE), Dr. Yin Yao
from the Electron Microscope Unit (EMU), and Dr. Anne Rich and Dr. [16] H. Aitchison, H. Lu, S. W. L. Hogan, H. Früchtl, I. Cebula,
Bill Bin Gong from surface analysis laboratory, Solid State and M. Zharnikov, M. Buck, Langmuir 2016, 32, 9397.
Elemental Analysis Unit (SSEAU) under Mark Wainwright Analytical [17] M. A. S. T. Schmitt, L. Hammer, J. Phys. Chem. C 2016, 120, 1043.
Centre, University of New South Wales (UNSW). This work was also [18] H. Zhang, X. Du, Y. Tang, X. Lu, L. Zhou, C. Zheng, H. Lin, Front.
undertaken with the assistance of resources and services from the Chem. 2020, 8, 1.
National Computational Infrastructure (NCI), which is supported by the [19] L. Cao, X. Du, H. Lin, C. Zheng, Z. Chen, S. Tao, J. Mater. Chem. C.
Australian Government, and through the UNSW-NCI partner trial scheme. 2020, 8, 17429.
Sol. RRL 2022, 2200184 2200184 (11 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH
www.advancedsciencenews.com www.solar-rrl.com
[20] J. Wang, S. Xie, D. Zhang, R. Wang, Z. Zheng, H. Zhou, Y. Zhang, [31] Z. Ma, Z. Tang, E. Wang, M. R. Andersson, O. Inganäs, F. Zhang,
J. Mater, Chem. A. 2018, 6, 19934. J. Phys. Chem. C. 2012, 116, 24462.
[21] P. E. Blöchl, Phys. Rev. B. 1994, 50, 17953. [32] J. Huang, Z. Yin, Q. Zheng, Energy Environ. Sci. 2011, 4, 3861.
[22] G. Kresse, J. Hafner, Phys. Rev. B. 1993, 47, 558. [33] W. Huang, B. Zhu, S. Y. Chang, S. Zhu, P. Cheng, Y. T. Hsieh,
[23] B. Hammer, L. B. Hansen, J. K. Nørskov, Phys. Rev. B Condens. Matter L. Meng, R. Wang, C. Wang, C. Zhu, C. McNeill, M. Wang,
Mater. Phys. 1999, 59, 7413. Y. Yang, Nano Lett. 2018, 18, 5805.
[24] M. Chen, D. Liu, W. Li, R. S. Gurney, D. Li, J. Cai, E. L. K. Spooner, [34] H. W. Cheng, P. Raghunath, K. L. Wang, P. Cheng, T. Haung, Q. Wu,
R. C. Kilbride, J. D. McGettrick, T. M. Watson, Z. Li, R. A. L. Jones, J. Yuan, Y. C. Lin, H. C. Wang, Y. Zou, Z. K. Wang, M. C. Lin,
D. G. Lidzey, T. Wang, ACS Appl. Mater. Interfaces 2019, 11, 26194. K. H. Wei, Y. Yang, Nano Lett. 2020, 20, 715.
[25] Y. Zhu, A. Gadisa, Z. Peng, L. Ye, Z. Xu, S. Zhao, H. Ade, Adv. Energy [35] G. Shi, J. Yang, Y. Ding, H. Fang, ChemPhysChem. 2014, 15, 2588.
Mater. 2019, 10, 1900376. [36] Q. Wu, M. Bhattacharya, S. E. Morgan, ACS Appl. Mater. Interfaces
[26] E. F. Manley, J. Strzalka, T. J. Fauvell, N. E. Jackson, M. J. Leonardi, 2013, 5, 6136.
N. D. Eastham, T. J. Marks, L. X. Chen, Adv. Mater. 2017, 29, 1. [37] F. M. Fowkes, Ind. Eng. Chem. 1964, 56, 40.
[27] A. Mahmud, N. K. Elumalai, D. Wang, A. Mahboubi, M. Wright, C. Xu, [38] M. A. Green, Solar cells: operating principles, technology, and system
F. Haque, A. Uddin, ACS Appl. Mater. Interfaces, 2017, 10, 7b09153. applications, Prentice-Hall, Englewood Cliffs, NJ 1982, p. 274.
[28] T. C. L. Zhang, D. Wett, R. Szargan, Surf. Interface Anal. 2004, 36, [39] L. Duan, B. Sang, M. He, Y. Zhang, M. A. Hossain, M. H. Rahaman,
1479. Q. Wei, Y. Zou, A. Uddin, B. Hoex, Sol. RRL 2020, 4, 202000497.
[29] D. E. Pugel, R. D. Vispute, S. S. Hullavarad, Appl. Surf. Sci. 2008, 254, [40] X. Ma, Y. Mi, F. Zhang, Q. An, M. Zhang, Z. Hu, X. Liu, J. Zhang,
2220. W. Tang, Adv. Energy Mater. 2018, 8, 201702854.
[30] W. Zhu, Z. Zhang, W. Chai, Q. Zhang, D. Chen, Z. Lin, J. Chang, [41] L. Duan, M. Guli, Y. Zhang, H. Yi, F. Haque, A. Uddin, Energy Technol.
J. Zhang, C. Zhang, Y. Hao, ChemSusChem. 2019, 12, 2318. 2020, 1901401, 1.
Sol. RRL 2022, 2200184 2200184 (12 of 12) © 2022 The Authors. Solar RRL published by Wiley-VCH GmbH