Photochromic organic solar cells based on diarylethenes

Article abstract of DOI:10.1039/d0ra04508j

Photovoltaic devices that switch color depending on illumination conditions may find application in future smart window applications. Here a photochromic diarylethene molecule is used as sensitizer in a ternary bulk heterojunction blend, employing poly(4-butylphenyldiphenylamine) (poly-TPD) and [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) for the transport of holes and electrons, respectively. Sandwiched between two electrodes, the blend creates a photochromic photovoltaic device that changes color, light absorption, and photon-to-electron conversion efficiency in the visible spectral range after having been illuminated with UV light. This journal is

Full text of DOI:10.1039/d0ra04508j

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Photochromic organic solar cells based on
diarylethenes†
Cite this: RSC Adv., 2020, 10, 30176
ab
Bart W. H. Saes,a Martijn M. Wienka and Rene A. J. Janssen
*
´
Photovoltaic devices that switch color depending on illumination conditions may find application in future
smart window applications. Here a photochromic diarylethene molecule is used as sensitizer in a ternary
bulk heterojunction blend, employing poly(4-butylphenyldiphenylamine) (poly-TPD) and [6,6]-phenyl-
C61-butyric acid methyl ester (PC61BM) for the transport of holes and electrons, respectively. Sandwiched
between two electrodes, the blend creates a photochromic photovoltaic device that changes color, light
absorption, and photon-to-electron conversion efficiency in the visible spectral range after having been
illuminated with UV light.
Received 20th May 2020
Accepted 11th August 2020
DOI: 10.1039/d0ra04508j
rsc.li/rsc-advances
which at least one is transparent to light. At the interface
between the donor and acceptor, photoinduced electron trans-
1. Introduction
fer generates holes and electrons that can be collected at the
electrodes aer being transported via the donor and acceptor
materials. As most existing photochromic dyes are not capable
of efficient charge transport, it is more appropriate to incor-
porate photochromic molecules as sensitizers in a ternary blend
in which the donor and acceptor materials ensure adequate
charge transport, but are weakly absorbing in the visible range
of the spectrum (Fig. 1). Aer absorbing UV light, the colorless
photochromic dye rst converts to the colored isomer. Subse-
quent absorption of visible light by the colored isomer of the
photochromic dye generates an excited state which dissociates
into a hole and electron on the donor and acceptor materials
(Fig. 1a).
Organic bulk heterojunction (BHJ) solar cells have received
considerable attention in recent years as they promise to
combine high performance and device exibility with low
cost.1–3 However, in terms of large scale photovoltaic energy
production, alternative technologies are better suited at
present.4,5 It is therefore also of interest to explore alternative
application areas for organic solar cells to exploit the unique
properties of organic molecules. Here we propose a proof of
concept for a photochromic BHJ organic solar cell (POSC). A
POSC should change its absorption spectrum under the inu-
ence of selective illumination. Even more ideal, a future prac-
tically useful POSC would be able to provide full transparency
during low-light conditions and transform into
a light-
For efficient charge generation, the energy levels of the
highest occupied molecular orbital (HOMO) of the colored
isomer of the photochromic dye should be positioned below the
HOMO level of the donor, while the lowest unoccupied molec-
ular orbital (LUMO) should be higher than that of the acceptor.
In addition, the photochromic dye has to be spatially arranged
to be simultaneously in electronic contact with the donor and
acceptor materials. In this way, holes and electrons generated
on the dye can be transported to the electrodes via the HOMO of
the donor and LUMO of the acceptor. In this concept, the donor
and acceptor govern the effective transport gap and thereby the
open-circuit voltage. The short-circuit current is determined by
the fraction of absorbed photons by the dye and the quantum
efficiency by which excitons on the dye can generate charges
that can be collected.
absorbing state and solar cell in daylight by using a reversible
photochromic system matched to the day/night cycle of
sunlight. Such device does not exist. Some previous work on
photochromic solar cells has been reported, yet to our knowl-
edge none utilizing the BHJ concept,6–9 so we took it as our aim
to explore the possibilities for a POSC and develop a rst
example of a working device that changes color under illumi-
nation, provides a photovoltaic effect, and can be switched
between two states.
In a conventional BHJ solar cell, the photoactive layer
consists of a binary blend of organic semiconductors with
electron donating and accepting properties, respectively. The
photoactive layer is sandwiched between two electrodes of
aMolecular Materials and Nanosystems, Institute for Complex Molecular Systems,
Eindhoven University of Technology, 5600 MB Eindhoven, The Netherlands. E-mail:
r.a.j.janssen@tue.nl
Irie's diarylethene (DAE) dyes are a versatile class of photo-
chromic molecules.10 DAEs possess two thermally stable
isomers (Fig. 1b) that can be interconverted selectively by illu-
mination with UV or visible light. The colorless open-ring
isomer and colored closed-ring isomer can retain switchability
bDutch Institute for Fundamental Energy Research, De Zaale 20, 5612 AJ Eindhoven,
The Netherlands
† Electronic supplementary information (ESI) available. See DOI:
10.1039/d0ra04508j
30176 | RSC Adv., 2020, 10, 30176–30185
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
Fig. 1 (a) Energy diagram of a photochromic organic solar cell. The photoactive layer is a ternary blend containing a photo-switchable dye (DAE)
and organic donor and acceptor semiconductors for hole (poly-TPD) and electron transport (PC61BM). After switching the DAE dye with UV light
from the colorless open-ring to the colored closed-ring isomer, absorption of visible light generates holes on the donor and electrons on the
acceptor. (b) Structures of a DAE dye, poly-TPD, and PC61BM.
and fatigue resistance of the interconversion in the solid state enables analyzing the properties of the uncolored and various
and when dispersed in a polymer matrix.11–21 Hence, DAEs photostationary states.
promise to be suitable candidates to construct photochromic
The colorability of DAEs is less in thin lm than in solu-
devices. So far, DAE dyes have been used in functional elec- tion.23,41,42 This has been explained by considering that the
tronic devices such as organic optical memories,11,21–27 organic open-ring isomer of DAE photochromes exists as a near 1 : 1
electronic switches,12,26,28–34 transistors,12,35 and organic light- mixture of two conformational isomers. One in which the aryl
emitting diodes.36,37 In a bilayer structure with N,N0-di(1-naph- rings are oriented parallel to each other and one with these
thyl)-N,N0-diphenylbenzidine
a photocurrent.38
(NPB),
DAEs
generate rings in an anti-parallel orientation (Fig. 3). Only the anti-
parallel conformation can convert to the closed-ring isomer
In this study three DAEs 1–3 were selected (Fig. 2). Their ring- upon UV illumination.41,43–51 in accordance with a conrotatory
open isomers (1a–3a) absorb light in the UV range at wave- mechanism for the photoinduced ring-closing.52–54 In solution,
lengths l < 350 nm, while the corresponding ring-closed thermal interconversion between both the two conformers of
isomers (1b–3b) absorb light across the entire visible part of the open-ring isomer proceeds rapidly at room temperature, but
the spectrum, i.e. at 450 nm < l < 650 nm. For 1–3, the quantum decreased orientational freedom in the solid state can restrict
yields both for the cyclization (4O–C) and cycloreversion (4C–O
)
this interconversion.
reactions in various solvents and in solid matrices have been
The aim of this work is to provide a proof-of-principle for
reported.11,39,40 DAEs 1a–3a exhibit high cyclization quantum a photochromic organic solar cell based on a ternary blend bulk
yields (4O–C) of 0.40–0.65 in solution, but the cycloreversion heterojunction using DAE dyes. We will show that one of the
quantum yields (4C–O) of the corresponding 1b–3b isomers are three dyes selected (1), exhibits the desired functionality, giving
considerably lower: 0.01–0.02. This is a useful asset because a clear difference in photoresponse in the visible range aer UV
a high 4O–C/4C–O ratio is required to reach a signicantly illumination converted the ring-open (1a) to the ring-closed (1b)
colored lm in ambient sunlight which has relatively little isomer. The other two dyes do not exhibit this behavior, which
intensity in the UV range compared to the visible range. The is explained by either a poor photoconversion (2), or inappro-
open-ring and closed-ring isomers of 1–3 have been reported to priate positioning of the energy levels to act as a sensitizer (3).
be thermally stable up to temperatures of 100 ꢀC.11,39,40 This
Fig. 2 Structures of the open-ring (1a–3a) and closed-ring (1b–3b) DAE isomers. Selective illumination with UV or visible light results in their
interconversion.
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 30176–30185 | 30177

View Article Online
RSC Advances
Paper
Fig. 3 The parallel and antiparallel conformations of 1a that are interswitchable in solution but are less likely to interconvert in the solid phase.
Correspondingly in the solid state the photochromic ring closure to 1b will occur from the antiparallel conformation of 1a.
potential. Interestingly, aer oxidation of 1a and 3a, the signals
in the subsequent reduction waves occur at potentials also
found in the cyclic voltammograms of the ring-closed isomers
2. Results and discussion
2.1 Optical and electrochemical properties
1b and 3b (Fig. 4a and c). This indicates that 1a and 3a undergo
ring closure upon oxidation.59,60 For 2a this occurs to minor
extent only. The reduction of the open-ring isomers 1a–3a
occurs at potentials well below ꢁ2 V vs. Fc/Fc+. Only for 2a
a reduction wave was observed (Fig. 4b), which is irreversible.
Of the closed-ring isomers, 1b and 3b show reversible
oxidation waves. For 1b the oxidation involves two consecutive
one-electron oxidations (Fig. 4a), while for 3b three closely
spaced one-electron oxidation waves can be seen (Fig. 4c). The
much lower oxidation potential of 3b compared to 1b is related
to the electron-withdrawing uorines on the cyclopentene ring
in 3b. This also causes the occurrence of the third oxidation in
3b to the triple cation at relatively low potentials. The oxidation
of 2b occurs at a higher potential than for 1b, because of steric
hindrance by the methyl groups on the 3-position of the thienyl
rings. In contrast to 1b and 3b, oxidation of 2b is irreversible,
likely as a consequence of the absence of methyl groups on the
para positions of the two phenyl rings. These methyl groups
effectively hinder dimerization via carbon–carbon bond
formation at the para positions aer oxidation in 1b and 3b.
Aer oxidation of 2b, the subsequent reduction wave shows
signals at much lower potentials that probably result from the
reduction of oxidized dimers of 2b that were coupled at the para
positions. At negative potentials, 1b and 2b show an irreversible
reduction wave. For 3b no reduction wave is observed and the
reduction potential is probably low and outside the potential
window of acetonitrile electrolyte used, as consequence of the
lack of uorine atoms. The onsets of the oxidation and reduc-
tions waves were used to estimate the energy levels of the
HOMO and LUMO for the ring-open and ring-closed isomers of
1–3 and are collected in Table 2.
Table 1 lists the wavelength of maximum absorption (lmax), the
onset of the optical absorption (lg), the optical band gap (Eg),
the molar absorption coefficients (3) and the quantum yields
(4O–C and 4C–O) inferred from measurements on solutions of
the ring-open and ring-closed isomers of 1–3 before and aer
UV illumination in either hexane (1, 2) or acetonitrile (3).17,40,55
In all cases, the ring-open isomer is colorless and only absorbs
light with wavelength l < 400 nm.
For a working photochromic solar cell, the dyes must have
appropriate energy levels for the HOMO LUMO compared to
donor and acceptor materials used for charge transport to be
able to generate free charges (Fig. 1a) upon illumination. In this
study poly(4-butylphenyldiphenylamine) (poly-TPD) and [6,6]-
phenyl C61 butyric acid methyl ester (PC61BM) are used as
donor as acceptor, respectively. Poly-TPD has an optical band
gap of about 2.9 eV and, hence, does not absorb in the visible
spectral range. PC61BM has a much lower optical band gap of
1.75 eV, but the absorption coefficient is low in the visible
range. Hence poly-TPD and PC61BM are suitable materials for
a proof-of-concept. Cyclic voltammetry was used to determine
the energy levels for 1–3,55–58 and to compare them to those of
poly-TPD and PC61BM.
The open-ring isomers 1a–3a, all show irreversible oxidation
waves (Fig. 4). For 3a, the onset of the oxidation is at lower
potential than for 1a and 2a as result of the absence of the six
electron withdrawing uorine substituents on the cyclopentene
ring. Compared to 1a, the onset of oxidation of 2a occurs at
a slightly higher potential. A likely explanation is that the
phenyl and thienyl rings cannot be coplanar in 2a because of
steric hindrance by the methyl groups on the 3-position of the
thienyl rings. The loss of coplanarity increases the oxidation
Table 1 Optical properties and quantum yields for interconversiona
Open-ring isomer
Closed-ring isomer
lmax
lmax
(nm)
lg/Egb (nm eVꢁ1
)
3 (Mꢁ1 cmꢁ1
)
4O–C (at lmax
)
(nm)
lg/Egb (nm eVꢁ1
)
3 (Mꢁ1 cmꢁ1
)
4C–O (at 492 nm)
1
2
3
285
262
279
330/3.76
310/4.00
340/3.65
3.99 ꢂ 104
2.80 ꢂ 104
3.45 ꢂ 104
0.59
0.46
0.43
580
562
522
680/1.82
655/1.89
600/2.07
1.81 ꢂ 104
1.10 ꢂ 104
1.97 ꢂ 104
0.013
0.015
0.01
a
From ref. 17, 40 and 55. b Optical bandgaps were determined from the onset of the absorption. For 1 and 2 solutions in hexane;17,40 for 3 solution in
acetonitrile.55
30178 | RSC Adv., 2020, 10, 30176–30185
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
Fig. 4 (a–c) Cyclic voltammograms of open-ring (black solid lines) and closed-ring isomers (colored solid lines) of 1–3 in acetonitrile. Potentials
are versus Fc/Fc+. (d) Comparison of energy levels of 1–3 with those of poly-TPD and PC61BM.
Based on the measured energy levels, some predictions for holes on poly-TPD aer photoexcitation. For 1b and 2b the
the performance in the solar cells can be made. Compounds 1b offsets (DEHOMO) of the HOMO energies with poly-TPD are 0.30
and 2b have HOMO and LUMO energies that lie between those and 0.40 eV respectively, which is likely sufficient for charge
of poly-TPD and PC61BM, but for 3b the HOMO energy is higher transfer. The offsets (DELUMO) of the LUMO energies with
than that of poly-TPD. As consequence 3b will unlikely generate PC61BM are 0.42 and 0.23 eV for 1b and 2b, respectively. A
DELUMO of 0.23 eV for 2b is less than ꢃ0.3 eV that is oen used
as phenomenological threshold for charge transfer to occur.61
For several covalently linked DAE-C60 adducts it has been
Table 2 HOMO and LUMO levels derived from the onsets of the redox
wavesa
shown that the photochromism is quenched severely when
compared to the non-covalently linked DAE equivalents in
isolation.62 It was concluded that, due their close proximity,
electron transfer from the DAE moiety to the fullerene cage is
the cause of the decreased photochromism. To determine the
photoswitchability of DAE dyes 1–3 in blends with poly-TPD and
PC61BM, UV-vis absorption measurements were performed
consecutively on DAE/poly-TPD/PC61BM blends in pristine
lms, aer UV illumination, and aer illumination with visible
light (Fig. 5). In these experiments, the weight ratio between
poly-TPD and PC61BM and was kept constant at 1 : 4. Fig. 5
shows the UV-vis spectra and photographs for lms of the three
dyes for a DAE to poly-TPD/PC61BM blend weight ratio of the of
EHOMO (eV)
ELUMO (eV)
ECgV (eV)
Poly-TPD
1a
1b
2a
2b
3a
3b
PC61BM
ꢁ4.66
ꢁ5.57
ꢁ4.96
ꢁ5.69
ꢁ5.06
ꢁ5.23
ꢁ4.41
ꢁ5.60
>ꢁ2.3b
>ꢁ2.3b
ꢁ3.09
ꢁ2.39
ꢁ3.28
>ꢁ2.3b
>ꢁ2.3b
ꢁ3.51
—
—
1.87
3.29
1.78
—
—
2.09
a
b
Conversion using E(Fc/Fc+) ¼ ꢁ4.59 eV against vacuum. No signal
observed within the measured range. The LUMO energy must be
>ꢁ2.3 eV.55–58
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 30176–30185 | 30179

View Article Online
RSC Advances
Paper
Fig. 5 UV-vis absorption spectra and photographs of ternary blends of 1–3, PC61BM, and poly-TPD. The ratio DAE/poly-TPD/PC61BM weight
ratio was 10 : 1 : 4. The samples were consecutively measured as pristine samples, after UV illumination, and after illumination with visible light as
detailed in the Experimental section. The UV-vis spectra are offset vertically for clarity.
2 : 1. The corresponding UV-vis spectra and photographs for discoloration. Apparently, under these conditions, the UV light
DAE to poly-TPD/PC61BM blend weight ratios of 1 : 1 and 1 : 2 present in the white light has a higher efficiency for the cycli-
are shown in Fig. S1 and S2 ESI†.
zation of 3a that the visible light has for the cycloreversion of 3b.
The 1a/poly-TPD/PC61BM blends show an increased absor- The photographs of the lms corresponding conrm the color
bance at 600 nm upon UV illumination, corresponding to the change to blue for 1 and to red for 3, but no change for 2,
conversion of 1a to 1b. The conversion is reversible through consistent with the changes in the UV-vis spectra. All lms also
illumination with white light from a tungsten halogen lamp. show an additional brownish color, which is primarily due to
The cyclization and cycloreversion occur for each of the three the absorption of PC61BM. The surface topology of the blend
blend compositions tested and the absorption at 600 nm is lms was checked before and aer UV illumination using
largest for the 2 : 1 blend. In contrast, the 2a/poly-TPD/PC61BM atomic force microscopy (AFM) (Fig. S3 ESI†). The blend lms
blend shows almost no photochromism. Even for the 2 : 1 are very smooth with root-mean-square surface roughness (Rq)
weight ratio of 2a to poly-TPD/PC61BM, the change in absorp- of 0.26 nm # Rq # 0.70 nm before and 0.28 nm # Rq # 0.37 nm
tion upon UV illumination is marginal. At present, we have no aer UV illumination.
conclusive explanation for the different behavior of 1a and 2a. It
might be that the methyl groups on 3-position of the thienyl
rings in 2a cause too much steric hindrance in the solid state for
2.2 Photochromic photovoltaic devices
Solar cells were made by sandwiching ternary blend lms,
consisting of the ring-open isomers 1a–3a, poly-TPD, and
PC61BM between a transparent indium tin oxide (ITO) (180 nm)/
the aromatic rings to become coplanar with the bridging
cyclopentene ring. The 3a/poly-TPD/PC61BM blend showed
a response to UV light in between that of 1a and 2a, generating
a new band at ꢃ550 nm. Like for 1, the absorbance at 550 nm
increases with the concentration of the dye in the blend. In
contrast to 1, however, illumination with white light resulted in
further coloration of the layers instead of the expected
poly(3,4-ethylenedioxythiophene):polystyrene
sulfonate
(PEDOT:PSS) (40 nm) front contact and an opaque LiF (1 nm)/Al
(100 nm) back contact. The photovoltaic performance was rst
tested for pristine devices before any illumination induced
changes to the system, and then aer UV illumination (l z 365
30180 | RSC Adv., 2020, 10, 30176–30185
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
nm) for 30 min. The assumption was made that the simulated summarized in Table 3. The results obtained for a device
AM1.5 G light used to evaluate the devices does not signicantly comprising a binary blend of poly-TPD and PC61BM are
affect the performance during the measurement. Fig. 6 shows included in Table 1 as a reference and the performance is
the current density–voltage (J–V) characteristics and the corre- similar to previous reports.63
sponding external quantum efficiency (EQE) spectra for the
When compared to the poly-TPD/PC61BM binary reference,
three different blends. The photovoltaic parameters are the inclusion of either 1a, 2a, or 3a in the corresponding ternary
Fig. 6 (a, c and e) Current density–voltage (J–V) characteristics of ternary DAE/poly-TPD/PC61BM solar cells before and after UV illumination to
convert the ring-open isomer to the ring-closed isomer, measured under simulated AM1.5 G solar light. (b, d and f) Corresponding EQE spectra.
(a and b) For DAE 1, (c and d) for DAE 2, (e and f) for DAE 3.
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 30176–30185 | 30181

View Article Online
RSC Advances
Paper
Table 3 Photovoltaic parameters for DAE/poly-TPD/PC61BM ternary blends before and after illumination with UV lighta
DAE
1
Ratio
d (nm)
107 ꢄ 2
109 ꢄ 7
117 ꢄ 13
160
UV
JSC (mA cmꢁ2
)
VOC (V)
FF
PCE (%)
2 : 1 : 5
3 : 1 : 5
3 : 1 : 5
0 : 1 : 5
No
Yes
No
Yes
No
Yes
No
Yes
1.13 ꢄ 0.02
1.18 ꢄ 0.05
0.89 ꢄ 0.06
0.84 ꢄ 0.06
0.64 ꢄ 0.08
0.47 ꢄ 0.04
1.79
0.79 ꢄ 0.06
0.76 ꢄ 0.02
0.79 ꢄ 0.04
0.75 ꢄ 0.04
0.83 ꢄ 0.03
0.74 ꢄ 0.02
0.77
0.40 ꢄ 0.03
0.39 ꢄ 0.01
0.37 ꢄ 0.01
0.37 ꢄ 0.01
0.33 ꢄ 0.01
0.31 ꢄ 0.00
0.43
0.36 ꢄ 0.05
0.35 ꢄ 0.03
0.27 ꢄ 0.02
0.24 ꢄ 0.02
0.18 ꢄ 0.02
0.11 ꢄ 0.01
0.59
2
3
—
1.77
0.76
0.43
0.58
a
The average values and standard deviations were obtained over 13 devices. The poly-TPD/PC61BM reference cell data were measured on a single
device.
blend results in a decrease in short-circuit current (JSC), a small 3 in solution has been studied by van Esch et al.65 They nd that
increase in open-circuit voltage (VOC), and a decrease in ll the perhydrocyclopentene is very stable towards photochemical
factor (FF) to result in an overall lower power conversion effi- decomposition, but roughly 2–3 times less than the corre-
ciency (PCE) (Table 3). These results can be understood by sponding peruorocyclopentene. Because of the small differ-
considering that the open-ring isomers of the DAE dyes have an ence and bearing in mind that 1 survives several cycles of
overlapping absorption with both poly-TPD and PC61BM and 30 min. UV illumination, we consider it unlikely that 3 degrades
dilute the blend, which reduces charge generation as absorp- signicantly already in the rst cycle. Also the red color and the
tion by the open-ring isomers of the DAEs does not result in corresponding change in the UV-vis spectrum aer UV illumi-
charge generation. The 10% increase in VOC possibly relates to nation (Fig. 5) show that 3 is largely converted to the closed
the fact that 1a–3a have a signicantly lower HOMO than poly- form.
TPD in the binary blend system.64
Fig. 5 shows that prolonged irradiation with visible light
UV illumination converts 1a to 1b and 3a to 3b, but does not converts 1b back to 1a in blends with poly-TPD and PC61BM.
ring close 2a to 2b in blends with poly-TPD/PC61BM (Fig. 5). For This suggests that solar cells based on dye 1 can be switched
a solar cell based on a ternary blend of 1a with poly-TPD and between two states. Fig. 7 shows the effect of exposing 1/poly-
PC61BM, UV illumination results in small increase in JSC from TPD/PC61BM solar cells to alternating 30 min periods of UV and
1.13 ꢄ 0.02 to 1.18 ꢄ 0.05 mA cmꢁ2 that mainly originates from visible illumination and recording the J–V characteristics aer
photons absorbed around 600 nm as shown by the EQE spec- each illumination period. The results show that JSC increases
trum in Fig. 6b. This spectral range corresponds to the aer each UV illumination and then decreases aer subsequent
absorption prole of 1b (Table 1 and Fig. 5) that is formed from exposure to visible light. The switching of JSC occurs over at least
1a upon UV illumination. The small, but distinct contribution ve consecutive cycles (total 5 hour illumination). Some oscil-
of 1b to the photocurrent in the EQE spectrum demonstrates latory behavior can also be seen for the FF and VOC. Overall the
that for DAE dye 1 charge generation occurs as shown in Fig. 1a, parameters decrease with consecutive cycles. Fig. S4 (ESI†)
and that for excited 1b charge generation in the blend competes
with the cycloreversion reaction. Whereas JSC has increased
aer UV illumination, the VOC decreases, and the FF remains
virtually the same. The VOC (0.76 ꢄ 0.02 V) is now similar to the
VOC (0.76 V) of the cell with the binary poly-TPD/PC61BM blend.
As expected from the absence of a photochromic conversion
from 2a to 2b in a blend with poly-TPD and PC61BM as inferred
from Fig. 5, the J–V characteristics and EQE of a photovoltaic
device consisting of a 2a/poly-TPD/PC61BM ternary blend are
virtually identical before and aer UV illumination (Fig. 6c and
d). The only relevant change is a small decrease of VOC that is
seen for all three DAE/poly-TPD/PC61BM cells.
For 3a, an overall decrease in photovoltaic performance is
observed aer UV illumination (Fig. 6e and f). The EQE spec-
trum shows a decrease in photocurrent generation over the
entire spectrum. A plausible explanation is that the shallow
HOMO energy of 3b impedes charge transport via poly-TPD by
acting as a trap for the transport of holes. An alternative
explanation for the reduced performance could be UV degra-
Fig. 7 Evolution of the normalized photovoltaic parameters of solar
cells based on 1/poly-TPD/PC61BM blends in five cycles of 30 min of
dation of 3. The switching and photostability of analogs of 1 and
UV illumination followed by 30 min exposure to visible light.
30182 | RSC Adv., 2020, 10, 30176–30185
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
shows the J–V characteristics of the pristine cell, and of the cell
4. Experimental section
in the 1st and 5th cycle recorded in this experiment.
¨
DAE dyes 1 and 3 were provided by Michael Patzel and Stefan
¨
Hecht of the Humboldt-Universitat zu Berlin. DAE dye 2 was
obtained from TCI and used as received. Poly-TPD with
a molecular weight of 10–120 kg molꢁ1 was obtained from
American Dye Source, Inc. PC61BM was purchased from Solenne
BV. Processing solvents (SAF) were used as received.
UV-vis spectroscopy was conducted on a PerkinElmer
Lambda 1050 spectrophotometer equipped with a 3D WB PMT/
InGaAs/PbS detector module. UV illumination of samples was
performed with a benchtop UV light (TL 8W BLB 1FM/10X25CC,
l z 365 nm, 30 min at a distance of z10 cm). Illumination with
visible light was achieved with light from an unltered OSRAM
64610 HLX G6.35 50 W 12 V tungsten halogen lamp, for 30 min
at a distance of z10 cm. Cyclic voltammetry was carried out
using an Autolab PGSTAT 30 in inert atmosphere. The electro-
3. Conclusions
DAE dyes 1–3 were evaluated to investigate the possibility of
creating a photochromic organic solar cell. In combination with
poly-TPD and PC61BM for hole and electron transport, DAE dye
1 afforded a photochromic layer that can be reversibly inter-
converted between a colorless ring-open (1a) and a colored ring-
closed (1b) isomer. An organic solar cell based on a ternary
blend of 1a/poly-TPD/PC61BM exhibited a distinct induced
contribution at 600 nm in the EQE spectrum of the ring-closed
isomer 1b aer illuminating the cell with UV light. Hence, the
ternary 1a/poly-TPD/PC61BM blend provides
a proof-of-
principle for a bulk heterojunction photochromic organic
solar cell, exhibiting an increased EQE in the visible region of
the solar spectrum aer illumination with UV light to convert 1a
to 1b. Despite the increased EQE, there is no increase in power
conversion efficiency because of a concomitant reduction of
open-circuit voltage and ll factor. The other two DAE dyes (2
and 3) did not yield photochromic organic solar cells. Dye 2 has
a very limited photoswitchability from 2a to 2b in a blend with
poly-TPD and PC61BM, such that UV illumination does not
change the photovoltaic characteristics. For 3, the HOMO
energy of the colored closed-ring isomer 3b is higher than that
of poly-TPD which, such that it can act a trap for hole transport.
While the 1a/poly-TPD/PC61BM shows the desired effect, the
PCE of 0.35 ꢄ 0.03% is very modest. There are some obvious
reasons for the low overall PCE. First, the poly-TPD/PC61BM blend
is not optimal for charge transport as evidenced by the low FF of
the reference devices without DAE dyes. Second, the concentration
of the 1 in the optimized blend is only 25% (by weight) and with an
optimized layer thickness of 110 nm, not all light will be absorbed
even if the conversion of 1a to 1b would be, and remain, quanti-
tative. Third, for an optimal charge generation the DAE dyes
should be solely positioned at the interface between poly-TPD and
PC61BM. It is unlikely that this condition is met in the blends.
Future optimization of charge transport materials and
photochromic dyes may alleviate some of the present limita-
tions. In a later stage, aer it has become clear how to increase
the PCE of photochromic organic solar cells, the dyes can
possibly be tuned to become thermally unstable in the closed-
ring isomer. The future, perfect photochromic dye would be
one that switches irreversibly to a colored form by using the UV
part of sunlight and only relaxes via a thermal process to
a colorless isomer in the absence of sunlight. In combination
with transparent electrodes, this would enable creating a device
that absorbs light and produces electrical energy in sunlight but
returns to a colorless and transparent conguration in dim light
and in the dark. Methods to tune the thermal stability of DAE
dyes have been described in literature to such an extent that the
thermal stability of both the both ring-open en ring-closed
isomers can be selectively altered,23,66–70 even independently of
the electronic properties of the DAE.68,71,72
lyte consisted of 0.1
M
tetrabutylammonium hexa-
uorophosphate (TBAPF6) in acetonitrile. A silver rod was used
as counter electrode and a silver chloride coated silver rod (Ag/
AgCl) as quasi-reference electrode. The measurements were
performed at a scan speed of 0.1 V sꢁ1 and potentials are quoted
against the ferrocene/ferrocenium (Fc/Fc+) redox couple as
external standard. For conversion to energy levels versus
vacuum, E(Fc/Fc+) ¼ ꢁ4.59 eV was used, based on a recent
study.73 Values in the literature vary between ꢁ4.4 and
ꢁ5.4 eV.74 Cyclic voltammetry (CV) was conducted both in the
dark and under continuous illumination using a benchtop UV
light (TL 8W BLB 1FM/10X25CC, l z 365 nm) as the light
source. The assumption was made that the in situ generated
concentration of ring-closed product would be sufficient for
detection in CV.
Photovoltaic devices with an active area of 0.09 and 0.16 cm2
were
fabricated
by
spin
coating
poly(ethylenedioxy
thiophene):poly(styrenesulfonate) (PEDOT:PSS) (Clevios P, VP
Al4083) on pre-cleaned, patterned indium tin oxide (ITO) glass
substrates (Naranjo Substrates). The photoactive layer was spin
coated at 2000 rpm in air from a chlorobenzene solution at 100 ꢀC
always at a total concentration of 30 mg mLꢁ1. For optimized
devices, the weight ratios between DAE/poly-TPD/PC61BM were
2 : 1 : 5, 3 : 1 : 5 and 3 : 1 : 5, for 1, 2, and 3 respectively. For opti-
mized poly-TPD/PC61BM reference devices, a range of ratios from
1 : 2 to 1 : 5 poly-TPD : PC61BM was tested. Ratios of 1 : 4 and 1 : 5
gave similar performance. The back electrode was deposited using
high vacuum (z5 ꢂ 10ꢁ7 mbar) thermal evaporation and consisted
of a layer of LiF (1 nm) and Al (100 nm). Cells with active areas of
0.09 or 0.16 cm2 gave virtually identical results and showed no
consistent trends in size-dependent differences.
J–V characteristics were measured using a Keithley 2400
source meter under z100 mW cmꢁ2 white light illumination
from a tungsten–halogen lamp ltered by a Hoya LB120 daylight
lter. Short-circuit current densities and PCEs were calculated by
integrating the AM1.5 G solar spectrum and the spectral response
of the cells. EQEs were determined using modulated mono-
chromatic light from a 50 W tungsten–halogen lamp (Philips
Focusline) passing through a monochromator (Oriel Cornerstone
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 30176–30185 | 30183

View Article Online
RSC Advances
Paper
130) with the use of a mechanical beam chopper. The response 15 J. Frisch, M. Herder, P. Herrmann, G. Heimel, S. Hecht and
was recorded as a voltage over a 50 U resistor (Stanford Research N. Koch, Appl. Phys. A: Mater. Sci. Process., 2013, 113, 1–4.
Systems SR570) using a lock-in amplier (Stanford Research 16 Q. Wang, J. Frisch, M. Herder, S. Hecht and N. Koch,
Systems SR830). A calibrated silicon cell with a known spectral ChemPhysChem, 2017, 18, 722–727.
response was used as a reference. During the measurement, the 17 M. Irie, T. Lia, S. Kobatake and N. Kato, J. Am. Chem. Soc.,
device was kept in a nitrogen-lled box behind a quartz window. 2000, 122, 4871–4876.
UV illumination to switch the DAE dyes to the closed-ring 18 R. Schmidt, M. Pars, T. Weller, M. Thelakkat and J. Kohler,
¨
¨
isomers was afforded by a benchtop UV light (TL 8W BLB 1FM/
10X25CC, l z 365 nm, 30 min at a distance of z10 cm).
The thicknesses of the active layers on the devices were
Appl. Phys. Lett., 2014, 104, 013304.
¨
19 C. Schorner, D. Wolf, T. Schumacher, P. Bauer, M. Thelakkat
and M. Lippitz, J. Phys. Chem. C, 2017, 121, 16528–16532.
determined using a Veeco Dektak 150 prolometer, subtracting 20 T. Kato, K. Okano and T. Yamashita, J. Photopolym. Sci.
the thicknesses of any underlying layers.
Technol., 2010, 23, 241–244.
21 K. Shibata, K. Muto, S. Kobatake and M. Irie, J. Phys. Chem. A,
2002, 106, 209–214.
22 S. H. Kawai, S. L. Gilat, R. Ponsinet and J. M. Lehn, Chem.–
Eur. J., 1995, 1, 285–293.
23 S. L. Gilat, S. H. Kawai and J. M. Lehn, Chem.–Eur. J., 1995, 1,
275–284.
Conflicts of interest
There are no conicts to declare.
Acknowledgements
24 M. Irie, Chem. Rev., 2000, 100, 1685–1716.
25 T. Tsujioka and H. Kondo, Appl. Phys. Lett., 2003, 83, 937–
939.
26 M. Irie, T. Fukaminato, K. Matsuda and S. Kobatake, Chem.
Rev., 2014, 114, 12174–12277.
27 M. Irie, T. Lia, S. Kobatake and N. Kato, J. Am. Chem. Soc.,
2000, 122, 4871–4876.
28 S. H. Kawai, S. L. Gilat and J. M. Lehn, Eur. J. Org. Chem.,
1999, 2359–2366.
¨
We thank Martin Herder, Stefan Hecht, and Michael Patzel for
providing dyes 1 and 3 and for stimulating discussions and
Haijun Bin for recording the AFM data. We acknowledge
funding from the European Research Council (Grant Agreement
No. 339031) and the NWO Spinoza prize awarded to R. A. J.
Janssen by the Netherlands Organization for Scientic Research
(NWO), and the Ministry of Education, Culture and Science
(Gravity program 024.001.035).
29 J. Areephong, W. R. Browne, N. Katsonis and B. L. Feringa,
Chem. Commun., 2006, 3930–3932.
30 H. Logtenberg, J. H. M. Van der Velde, P. De Mendoza,
J. Areephong, J. Hjelm, B. L. Feringa and W. R. Browne, J.
Phys. Chem. C, 2012, 116, 24136–24142.
31 T. Tsujioka and M. Irie, J. Photochem. Photobiol., C, 2010, 11,
1–14.
References
1 C. Cui and Y. Li, Energy Environ. Sci., 2019, 12, 3225–3246.
¨
2 O. Inganas, Adv. Mater., 2018, 30, 1800388.
3 V. Vohra, Chem. Rec., 2019, 19, 1166–1178.
`
4 W. Hou, Y. Xiao, G. Han and J. Y. Lin, Polymers, 2019, 11, 143. 32 E. Orgiu and P. Samorı, Adv. Mater., 2014, 26, 1827–1845.
˜
´
´
5 J. A. Luceno-Sanchez, A. M. Dıez-Pascual and R. P. Capilla, 33 N. Zhang, W. Y. Lo, A. Jose, Z. Cai, L. Li and L. Yu, Adv.
Int. J. Mol. Sci., 2019, 20, 976.
Mater., 2017, 33, 1701248.
6 J. J. Wu, M. D. Hsieh, W. P. Liao, W. T. Wu and J. S. Chen, 34 Z. Liu, H. I. Wang, A. Narita, Q. Chen, Z. Mics,
¨
¨
ACS Nano, 2009, 3, 2297–2303.
D. Turchinovich, M. Klaui, M. Bonn and K. Mullen, J. Am.
7 W. Wu, J. Wang, Z. Zheng, Y. Hu, J. Jin, Q. Zhang and J. Hua,
Sci. Rep., 2015, 5, 8592.
Chem. Soc., 2017, 139, 9443–9446.
¨
35 M. El Gemayel, K. Borjesson, M. Herder, D. T. Duong,
´
8 S. Ma, H. Ting, Y. Ma, L. Zheng, M. Zhang, L. Xiao and
Z. Chen, AIP Adv., 2015, 5, 057154.
9 N. M. Johnson, Y. Y. Smolin, C. Shindler, D. Hagaman,
J. A. Hutchison, C. Ruzie, G. Schweicher, A. Salleo,
Y. Geerts, S. Hecht, E. Orgiu and P. Samori, Nat. Commun.,
2015, 6, 6330.
M. Soroush, K. K. S. Lau and H. F. Ji, AIMS Mater. Sci., 36 Z. Zhang, X. Liu, Z. Li, Z. Chen, F. Zhao, F. Zhang and
2015, 2, 503–509.
C. H. Tung, Adv. Funct. Mater., 2008, 18, 302–307.
¨
10 M. Irie and M. Mohri, J. Org. Chem., 1988, 53, 803–808.
11 M. Irie, T. Lia and K. Uchida, Mol. Cryst. Liq. Cryst., 1997,
297, 81–84.
37 P. Zacharias, M. C. Gather, A. Kohnen, N. Rehmann and
K. Meerholz, Angew. Chem., Int. Ed., 2009, 48, 4038–4041.
38 T. Tsujioka, M. Yamamoto, K. Shoji and K. Tani, Photochem.
Photobiol. Sci., 2010, 9, 157–161.
¨
12 E. Orgiu, N. Crivillers, M. Herder, L. Grubert, M. Patzel,
¨
J. Frisch, E. Pavlica, D. T. Duong, G. Bratina, A. Salleo, 39 S. Fredrich, R. Gostl, M. Herder, L. Grubert and S. Hecht,
N. Koch, S. Hecht and P. Samori, Nat. Chem., 2012, 4, 675–
679.
13 M. Hamano and M. Irie, Jpn. J. Appl. Phys., 1996, 35, 1764–
1767.
Angew. Chem., Int. Ed., 2016, 55, 1208–1212.
40 M. Irie, K. Sakemura, M. Okinaka and K. Uchida, J. Org.
Chem., 1995, 60, 8305–8309.
¨
41 T. Kawai, N. Fukuda, D. Groschl, S. Kobatake and M. Irie,
14 T. Leydecker, M. Herder, E. Pavlica, G. Bratina, S. Hecht,
Jpn. J. Appl. Phys., 1999, 38, L1194–L1196.
`
E. Orgiu and P. Samorı, Nat. Nanotechnol., 2016, 11, 769–775.
30184 | RSC Adv., 2020, 10, 30176–30185
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
42 M. Irie, O. Miyatake, K. Uchida and T. Eriguchi, J. Am. Chem. 58 M. Yoshida, K. Suemori, S. Uemura, S. Hoshino, N. Takada,
Soc., 1994, 116, 9894–9900.
43 S. Fukumoto, T. Nakashima and T. Kawai, Angew. Chem., Int.
Ed., 2011, 50, 1565–1568.
44 T. Yamaguchi and M. Irie, J. Photochem. Photobiol., A, 2006,
178, 162–169.
T. Kodzasa and T. Kamata, Jpn. J. Appl. Phys., 2010, 49,
04DK09.
59 W. R. Browne, J. J. D. De Jong, T. Kudernac, M. Walko,
L. N. Lucas, K. Uchida, J. H. Van Esch and B. L. Feringa,
Chem.–Eur. J., 2005, 11, 6414–6429.
45 A. Goldberg, A. Murakami, K. Kanda, T. Kobayashi, 60 W. R. Browne, J. J. D. De Jong, T. Kudernac, M. Walko,
S. Nakamura, K. Uchida, H. Sekiya, T. Fukaminato,
T. Kawai, S. Kobatake and M. Irie, J. Phys. Chem. A, 2003,
107, 4982–4988.
L. N. Lucas, K. Uchida, J. H. Van Esch and B. L. Feringa,
Chem.–Eur. J., 2005, 11, 6430–6441.
¨
61 M. C. Scharber, D. Muhlbacher, M. Koppe, P. Denk,
46 M. Boggio-pasqua, M. Ravaglia, M. J. Bearpark, M. Garavelli
and M. A. Robb, J. Phys. Chem. A, 2003, 107, 11139–11152.
C. Waldauf, A. J. Heeger and C. J. Brabec, Adv. Mater.,
2006, 18, 789–794.
47 S. Kobatake, K. Uchida, E. Tsuchida and M. Irie, Chem. 62 A. R. Tuktarov, A. A. Khuzin, A. R. Akhmetov,
Commun., 2002, 2804–2805.
V. A. Barachevsky, O. V. Venidiktova and U. M. Dzhemilev,
48 S. Nakamura, T. Kobayashi, A. Takata, K. Uchida, Y. Asano,
Tetrahedron Lett., 2015, 56, 7154–7157.
A. Murakami, A. Goldberg, D. Guillaumont, S. Yokojima, 63 L. Blankenburg, L. S. Sensfuss, H. Schache, J. Marten,
¨
S. Kobatake and M. Irie, J. Phys. Org. Chem., 2007, 20, 821–
829.
R. Milker and M. Schrodner, Synth. Met., 2015, 199, 93–104.
64 Q. An, F. Zhang, J. Zhang, W. Tang, Z. Deng and B. Hu,
49 D. Guillaumont, T. Kobayashi, K. Kanda, H. Miyasaka,
Energy Environ. Sci., 2016, 9, 281–322.
K. Uchida, S. Kobatake, K. Shibata, S. Nakamura and 65 J. J. D. de Jong, L. N. Lucas, R. Hania, A. Pugzlys,
M. Irie, J. Phys. Chem. A, 2002, 106, 7222–7227.
R. M. Kellogg, B. L. Feringa, K. Duppen and J. H. van Esch,
50 K. Uchida, D. Guillaumont, E. Tsuchida, G. Mochizuki,
Eur. J. Org. Chem., 2003, 1887–1893.
M. Irie, A. Murakami and S. Nakamura, J. Mol. Struct., 66 S. Nakamura and M. Irie, J. Org. Chem., 1988, 53, 6136–6138.
2002, 579, 115–120.
51 Y. Asano, A. Murakami, T. Kobayashi, A. Goldberg,
67 S. Takami, S. Kobatake, T. Kawai and M. Irie, Chem. Lett.,
2003, 32, 892–893.
D. Guillaumont, S. Yabushita, M. Irie and S. Nakamura, J. 68 S. Kawai, T. Nakashima, K. Atsumi, T. Sakai, M. Harigai,
Am. Chem. Soc., 2004, 126, 12112–12120.
52 J. Clayden, N. Greeves, S. Warren and P. Wothers, Organic
Chemistry, Oxford University Press, 2009.
Y. Imamoto, H. Kamikubo, M. Kataoka and T. Kawai,
Chem. Mater., 2007, 19, 3479–3483.
69 D. Chen, Z. Wang and H. Zhang, J. Mol. Struct.: THEOCHEM,
2008, 859, 11–17.
53 J. E. McMurry, Organic Chemistry, Thomson, 2008.
54 R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc., 1965, 70 S. Nakamura, S. Yokojima, K. Uchida, T. Tsujioka,
87, 395–397.
A. Goldberg, A. Murakami, K. Shinoda, M. Mikami,
T. Kobayashi, S. Kobatake, K. Matsuda and M. Irie, J.
Photochem. Photobiol., A, 2008, 200, 10–18.
¨
55 M. Herder, B. M. Schmidt, L. Grubert, M. Patzel, J. Schwarz
and S. Hecht, J. Am. Chem. Soc., 2015, 137, 2738–2747.
56 M. Herder, F. Eisenreich, A. Bonasera, A. Gra, L. Grubert, 71 S. Kobatake, K. Uchida, E. Tsuchida and M. Irie, Chem. Lett.,
¨
M. Patzel, J. Schwarz and S. Hecht, Chem.–Eur. J., 2017, 23,
2000, 29, 1340–1341.
3743–3754.
72 S. Kobatake and M. Irie, Chem. Lett., 2003, 32, 1078–1079.
57 M. Herder, Improving the Design of Diarylethene 73 R. E. M. Willems, C. H. L. Weijtens, X. de Vries, R. Coehoorn
Photoswitches and their Exploitation as Remote-Controlled and R. A. J. Janssen, Adv. Energy Mater., 2019, 9, 1803677.
Building Blocks, PhD Thesis, Humboldt-Universitat Berlin, 74 C. M. Cardona, W. Li, A. E. Kaifer, D. Stockdale and
¨
2015.
G. C. Bazan, Adv. Mater., 2011, 23, 2367.
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 30176–30185 | 30185