Content uploaded by Marinela Miclau
Author content
All content in this area was uploaded by Marinela Miclau on May 12, 2023
Content may be subject to copyright.
RESEARCH ARTICLE
Highly efficient dye-sensitized solar cells for wavelength-
selective greenhouse: A promising agrivoltaic system
Daniel Ursu
1
| Melinda Vajda
1,2
| Marinela Miclau
1
1
Condensed Matter Department, National
Institute for Research and Development
in Electrochemistry and Condensed
Matter, Timis¸oara, Romania
2
Applied Chemistry and Engineering of
Inorganic Compounds and Environment
Department, Politehnica University
Timis¸oara, Timis¸oara, Romania
Correspondence
Marinela Miclau, National Institute for
Research and Development in
Electrochemistry and Condensed Matter,
Dr. A. P
aunescu-Podeanu Street
144, Timis¸oara 300569, Romania.
Email: marinela.miclau@gmail.com
Funding information
Romanian National Authority for
Scientific Research and Innovation,
Grant/Award Numbers: PN-III-
P2-2.1-PED-2019-2091, PNCDI III
Summary
In trying to solve simultaneously the energy and food crisis, the concept pro-
posed by Agriculture 4.0, namely a combination between the agriculture and
the solar energy could provide a possible solution. The wavelength-selective
greenhouse could be a promising agrivoltaic system if the trade-off between
photovoltaic roofs and plants will be achieved. Using less studied solar cells as
an electricity source for an autonomous greenhouse, this study has demon-
strated experimentally that the requirements imposed by a greenhouse can be
provided by a DSSC using an affordable commercial yellow dye. The successful
implementation of DSSC in autonomous greenhouses is conditioned by three
main requirements, and that are the transparency of the entire Photosynthetic
Active Radiation (PAR) domain along with high UV absorption, high effi-
ciency of the solar cell, and sustainability during the whole year. The best
DSSC has proved a high absorption of UV radiation, closely by 90%, a transpar-
ency of the DSSCs preserved on the whole PAR domain and achieved a photo-
voltaic efficiency two times higher than the best efficiency reported for this
dye so far. Furthermore, by preserving the maximum efficiency of almost 5%
under the light intensity in the range 50 to 100 mW/cm
2
, the sustainability of
our DSSC over the whole year has been demonstrated.
KEYWORDS
agrivoltaic system, commercial yellow dye, dye-sensitized solar cell, photoanode
architecture, wavelength-selective greenhouse
1|INTRODUCTION
Diversification in the use of solar cells is of increasing
interest and not only from the perspective of the energy
crisis. Regardless of the kind of solar cells, their design
and efficiency have been made to increase the photovol-
taic efficiency, considering the large spaces where it can
be positioned. The use of existing spaces for electricity
production was made to the detriment of the agricultural
crops. However, resolving the energy crisis must consider
and be contingent on solving the food crisis. Such possi-
ble options could be vertical arrangement of the solar
cells or the intelligent use of the same space for different
needs and functions. The second category includes agri-
culture 4.0 which proposes the combination of the agri-
culture with the generation of photovoltaic energy and
the development of “agrivoltaics”or “solar farming”.An
agrivoltaic system can mitigate the competition for land
and spatial constraints that condition the development of
the solar energy, providing an intelligent solution for the
durable development of the future energy. The forecast is
that worldwide demand would be compensated if even
less than 1% of cultivated land will be devoted to the
implementation of the agrovoltaic systems.
1-3
Received: 22 April 2022 Revised: 22 June 2022 Accepted: 20 July 2022
DOI: 10.1002/er.8469
18550 © 2022 John Wiley & Sons Ltd. Int J Energy Res. 2022;46:18550–18561.wileyonlinelibrary.com/journal/er
The advantages of the integration of the solar cells in
the greenhouse have been highlighted in recent research,
such as (i) a “zero-impact greenhouse”by diminishing
the cost of production through the elimination of any
energy provided by the conventional sources
4
; (ii) low
water usage in irrigation by the diminution of the evapo-
transpiration process
5
; (iii) decreasing the competition
between the land for food and energy production
6
;
(iv) regulate the microclimate, photovoltaic system being
also a passive cooling system
7
; (v) local electricity genera-
tion especially in rural and decentralized areas where
access to energy is lacking; (vi) high absorption of solar
UV radiation leads to a beneficial impact on plant growth
and the diminution of the plant pathogenicity.
8
Light is one of the crucial factors for the plant growth.
From the perspective of the plants grown in a greenhouse,
the solar radiation consists of Ultraviolet (UV), Photosyn-
thetic Active Radiation (PAR), and Near-Infrared (NIR).
PAR, defined as light with a 430 to 700 nm wavelength
range, is essentially required by plant growth and photo-
synthesis of the plants. Chlorophylls are the primary pig-
ments used in photosynthesis, characterized by four
important absorption peaks of chlorophyll aand
bpositioned in the red and blue regions, corresponding to
625 to 675 nm and 430 to 475 nm respectively.
9
A photovoltaic (PV) greenhouse have to ensure simul-
taneously two opposite demands, namely maximizing the
accessibility of PAR radiation into the greenhouse by
diminishing the shading effect caused by PV panels
10
and
enhancing the generation of the energy accompanied by
increasing the opaque surfaces of the panels. The biggest
challenge of a PV greenhouse is to find a compromise
between PV roofs and plants. First and second genera-
tions of solar cells are based on crystalline silicon and
III-V compound semiconductors. Although first tested,
the main limitations of the integration in greenhouses
concern the fact that these PV cells do not transmit sun-
light resulting in a permanent shadow region with the
detrimental effects on production, the declining crop
growth or the amount of biomass.
10-13
Usually, the roof of the greenhouse built with
wavelength-selective photovoltaic (WSPV) panels consists
of WS luminescent absorbers positioned between Si-PV
panels with the coverage rate (ratio of the horizontal sur-
face of the greenhouse covered by solar panels situated
on the roof) between 10% and 60% depending on the
requested light crops.
14
The power efficiency of a WSPV
module is conditioned by the power efficiency of the PV
cells and luminescent material and is limited by low
absorption of the solar radiation and the self-absorption
of the emitted light from the luminescent absorber. Opti-
mizing the geometry and cell type of WSPV, the best con-
version efficiency of 5% has been reported.
15
Even though less efficient in the production of elec-
tricity, because of the construction and the principle of
operation, dye-sensitized solar cells (DSSCs) allow an
easy adaptation to the conditions imposed by the optimal
operation of greenhouses. Simplicity, low cost, and
modeling light absorption through dye selectivity are the
attractive advantages of DSSCs.
16-20
Using less studied solar cells as an electricity source
for an autonomous greenhouse, and, generally, from the
perspective of theoretical demonstration of the concept,
our paper aims to demonstrate experimentally that the
requirements imposed by a greenhouse can be provided
by a DSSC.
21-23
To achieve this, we started from the fol-
lowing premises: (i) the choice of an affordable commer-
cial dye that absorbs UV and is transparent to PAR,
namely DN-F01 dye (Dyenamo Yellow) with the best
energy conversion efficiency of 2.39%
24
; (ii) optimization
of an inexpensive common electrolyte; (iii) designing and
optimization of the photoanode based on TiO
2
poly-
morphs for this dye and electrolyte; (iv) enhancing the
energy performance of the DSSC based on UV dye for
both full sun and shading conditions, without
impacting PAR.
The optimization of the photoanode based on TiO
2
using a complex architecture has successfully allowed
(i) achievement of the photovoltaic efficiency two times
higher than the best efficiency reported for the same dye
so far; (ii) transparency on the entire PAR domain;
(iii) the capability of DSSCs to achieve the coverage rate
of 100%, and (iv) the same maximum efficiency for 1-sun
illumination (summer) and shading, particularly for
60 mW/cm
2
corresponding to the maximum light inten-
sity in the three seasons, namely, spring, autumn, and
winter (as small variations due to the positioning on the
planet).
2|EXPERIMENTAL SECTION
2.1 |Preparation of TiO
2
nanoparticles
(TiO
2
_NP)
1 mL of acetic acid (Sigma-Aldrich) and 2.4 mL of tita-
nium (IV) isopropoxide (Sigma-Aldrich) were added
under stirring for 20 min at room temperature. The
obtained solution was mixed with 14 mL H
2
O under fast
stirring (800 rot/min), to obtain instantaneously a white
precipitate. Stirring is continued for 1.5 h achieving a
complete hydrolysis reaction. After that, 5 mL of H
2
O
and 200 μL of nitric acid (Sigma-Aldrich) are added and
stirred for 90 min at 70 C. The resulting mixture was
kept in a 60 mL autoclave and heated at 100C for 12 h.
The white sediment of TiO
2
_NP was obtained by
URSU ET AL.18551
centrifugation and was washed with DI water and etha-
nol for five times and dried at 40C for 12 h.
2.2 |Preparation of TiO
2
particles for
light scattering (TiO
2
_LS)
2 mL of titanium (IV) isopropoxide (Sigma-Aldrich), 3.6 g
of NaOH (Sigma-Aldrich), and 30 mL H
2
O were stirred
for 60 min at room temperature. The mixture obtained
has been kept in a 60 mL autoclave and heated at 220C
for 24 h. The white sediment was washed with DI water
and ethanol and dried at 60C for 12 h. Finally, TiO
2
_LS
obtained was calcined at 500C at a speed of 2C/min
and kept for 1 h in air.
The powder X-ray diffraction (XRD) PW 3040/60
X'Pert PRO using Cu-Kαradiation (λ=1.5418 Å) and the
range 2θ=10to 80allowed to determine the structure
of the products at room temperature. Using a Scanning
Electron Microscope Inspect S (SEM), the morphology
was highlighted. UV-visible analysis of the samples was
performed using a UV/Vis/NIR spectrophotometer. The
specific surface and average pore size of anatase TiO
2
_NP
nanoparticles were measured using a Quantachrome
Nova 1200e Pore Size Analyzer. Ahead of the BET mea-
surement, the sample has been degassed at 200C for 4 h
to remove the already adsorbed moisture.
2.3 |Fabrication of DSSCs
For the preparation of the substrate, the FTO glass (13 Ω/
sq, Sigma Aldrich) was cleaned with ethanol, acetone,
and water for 30 min, followed by 25 min of UV-Ozone
treatment (Ossila). The compact TiO
2
blocking layer
(TiO
2
_BL) was prepared on the FTO glass using titanium
diisopropoxide bis(acetylacetonate) (75 wt. % in isopropa-
nol) solution and tert-butanol (99.8%, Sigma-Aldrich)
with 1:5 concentration by spin-coating method at
5000 rpm for 30 s followed by calcination at 450C for
60 min with 1C/min. The preparation of pastes based on
the TiO
2
_NP and TiO
2
_LS materials has been described
in detail in our previous work.
18
The TiO
2
_NP and
TiO
2
_LS layers were printed on FTO glass using the doc-
tor blade method followed by calcination at 500C for
60 min with 1C/min for each layer. Finally, the photo-
electrode was dipped in a 40 mM TiCl
4
solution at 70C
for 30 min, followed by calcinations at 450C for 60 min
with 1C/min heating rate.
Based on the above films, four photoanodes were
designed and made with distinct functionalized layers.
Photoanode 1 (PhA 1) consists of a single layer of
TiO
2
_NP, photoanode 2 (PhA 2) consists of TiO
2
_BL and
TiO
2
_NP layers, photoanode 3 (PhA 3) consists of
TiO
2
_BL, TiO
2
_NP and TiO
2
_LS layers, and photoanode
4 (PhA 4) consists of TiO
2
_BL layer, TiO
2
_NP layer,
TiO
2
_LS layer and TiCl
4
treatment.
All photoanodes obtained were introduced into
0.3 mM of DN-F01 dye (Dyenamo Yellow) solution in
absolute ethyl alcohol for 5 h. The encapsulation together
of the photoanode and platinized counter electrode
(CE) (H
2
PtCl
6
at 400C for 30 min on FTO substrate) has
been achieved using a polymer seal (Meltonix 1170-60).
18
Two different electrolytes were used to fill the space
between the electrodes which consist of a solution of
0.6 M 1-butyl-3-methyl-immidazolium iodide, 0.03 M I
2
,
0.10 M guanidiniumthiocyanate, and 0.5 M 4 tertbutylpyr-
idine in acetonitrile/valeronitrile (85/15) for electrolyte
1 (E1) and 2 M 1-butyl-3-methyl-immidazolium iodide,
0.05 M I
2
, 0.10 M guanidiniumthiocyanate, and 0.5 M
4 tertbutylpyridine in acetonitrile for electrolyte 2 (E2).
The DSSCs built with the four photoanodes defined
above will henceforth be named as DSSC 1, DSSC 2, DSSC
3, and DSSC 4 respectively.
Current-Voltage curves have been carried out using a
Keithley 2450 SourceMeter under AM 1.5G simulated
sunlight (100 mW/cm
2
). The electrochemical investiga-
tions of the DSSCs were performed using a Voltalab
potentiostat model PGZ 402, with VoltaMaster 4 (version
7.09) software, in the potential range: 10.2 V, with a
0.05 V potential step, a fixed frequency of 1 kHz and
0.02 V amplitude of AC potential. All potential values are
related to the I
/I
3
redox potential, the measurements
being performed in E1 and E2 electrolytes. Electrochemi-
cal impedance spectroscopy (EIS) analysis was also stud-
ied under illumination conditions, the EIS was
performed using a bias potential of 0.62, 0.68, 0.84, and
0.81 V for DSSC 1, DSSC 2, DSSC 3, and DSSC 4 respec-
tively. The frequency range is from 0.001 to 10 kHz, using
0.01 V of magnitude for the modulation signal. Using
adsorption-desorption technique, the dye loading capac-
ity of each photoanode was highlighted according to the
literature.
25
The calibration curve for DN-F01 dye was
determined using five different dye concentrations in
10 mM NaOH, recording UV-Vis absorption spectra. The
photoanodes were immersed in 10 mM NaOH solution
and kept for 2 h to desorb the dye molecules from the
substrate surface which were measured by UV-Vis
absorption spectroscopy using 3 mL of solution for each
sample.
3|RESULTS AND DISCUSSIONS
XRD pattern of the TiO
2
nanoparticles is presented in
Figure 1A. All diffraction peaks corresponded to anatase
18552 URSU ET AL.
TiO
2
with a tetragonal structure (space group: I41/amd).
The crystalline size of TiO
2
_NP was calculated using the
Scherrer formula:
Dhkl ¼Kλ
Bcosθð1Þ
where k-constant (1), B-full width at half maximum
(FWHM), λ- wavelength of X-ray, and θ- diffraction
angle. The crystalline size of TiO
2
_NP is 18 nm. The
proposed brookite polymorph of TiO
2
microparticles for
the light scattering layer is highlighted by the characteris-
tic XRD pattern of orthorhombic symmetry with Pbca
space group (Figure 1B).
Figure 2reveals the SEM image of the obtained
brookite TiO
2
with quasi-microcube-like morphology
having relative uniform particle sizes in the range of
330 to 950 nm with a mean particle size of 659 nm.
The optical direct band gap of the TiO
2
photoanode
material is calculated from the diffuse reflectance spectra
(Figure 3). Intercepting the extrapolated linear fit of the
plotted experimental data of [αhν]
2
vs incident photon
energy (hν) near the absorption edge, values of 3.13 eV
for anatase TiO
2
_NP and 3.32 eV for brookite TiO
2
_LS,
respectively, were obtained, which are comparable to the
reported values.
26
In agreement with the 18 nm of the crystalline size
estimated by the Scherrer method, anatase TiO
2
_NPs are
characterized by a high Brunauer-Emmett-Teller (BET)
specific surface area with a value of 286 m
2
/g,
27
much
larger than the screen-printable transparent TiO
2
paste
from Dyenamo (approx. 85 m
2
/g). From the adsorption-
desorption spectra (Figure 4), a type IV(a) isotherm and
H1-type hysteresis loop are confirmed for anatase
TiO
2
_NPs suggesting a characteristic material with meso-
porosity and high molecular adsorption energy
28
and
shows a narrow pore size distribution centered at 3.5 nm
(inset of Figure 4), having a homogeneous pore size in
the sample.
Both analyses showed that the hydrothermal synthe-
sis proposed at very low temperature provided the ana-
tase nanoparticles with a higher surface area for dye
loading.
In accordance with the preparation of the photoa-
nodes presented in the experimental section, the crystal-
line structures, morphology, and uniformity of the films
deposed on FTO substrate are preserved and shown in
Figures 5and 6. Furthermore, PhA3 consists of TiO
2
_NP
film of 0.80 μm thickness and TiO
2
_LS film of 2.20 μm
thickness. The additional treatment with TiCl
4
resulted
in an increase of 0.90 μm in the thickness of PhA4
(Figure 6).
The successful implementation of DSSC in autono-
mous greenhouses is conditioned by three main require-
ments, and that are the transparency on full PAR domain
along with high UV absorption, high photovoltaic effi-
ciency, and sustainability over the whole year. Based on
FIGURE 1 Room-temperature X-ray diffraction pattern of
(A) TiO
2
anatase nanoparticles and (B) TiO
2
brookite
microparticles
FIGURE 2 (A) SEM image
of brookite TiO
2
_LS particles;
(B) Particle size distribution of
brookite TiO
2
_LS particles
URSU ET AL.18553
the intrinsic mechanism of a DSSC, four different archi-
tectures of TiO
2
photoactive electrodes using DN-F01 dye
and E1 electrolyte were designed and tested to determine
the optimal configuration that ensures their simulta-
neous achievement.
Figure 7shows the transmittance spectra of DSSCs
based on four different photoanodes. The transmit-
tance spectra were similar for all DSSCs, highlighted a
high absorption of UV radiation, closely by 98% and a
similar transparency of all DSSCs on the full PAR
domain with those reported in the literature.
29
As
expected, the deposition of the LS layer diminished the
transparency of the DSSC 3 and DSSC 4 but remains
over 25%.
The J-V characteristics of the DSSCs using E1 electro-
lyte are presented in Figure 8and the detailed photovol-
taic parameters (J
SC
,V
OC
, FF, and η) are presented in
Table 1. A significant enhancement of the efficiency is
due to the addition of TiO
2
_LS layer in DSSC 3 (92%
more) and DSSC 4 (242% more) relative to DSSC 1.
Moreover, for a deeper understanding of the charge
transfer mechanisms responsible by the photovoltaic per-
formance of our DSSCs and the effect of TiO
2
_LS, electro-
chemical impedance spectroscopy (EIS) analysis was
performed at V
OC
and under 1 sun illumination. Based
on these results, an equivalent circuit has been proposed
in the inset of Figure 9being composed of R1, the ohmic
serial resistance and representing the intrinsic resistance
FIGURE 3 Band gap of (A) anatase TiO
2
_NP and (B) brookite TiO
2
_LS calculated from the diffuse reflectance spectra
FIGURE 4 N
2
adsorption-desorption spectra and insert shows
the pore size distribution of anatase TiO
2
_NPs FIGURE 5 Room-temperature X-ray diffraction pattern of the
TiO
2
photoanodes on FTO substrate
18554 URSU ET AL.
of the assembled cells, R2, the charge transfer resistance
at the CE/electrolyte interface, and R3, the charge trans-
fer resistance at the TiO
2
photoelectrode/dye/electrolyte
interface. The obtained values of the resistance represent-
ing each interfacial process of the DSSC are summarized
in Table 1. In addition, the desorption study highlighted
the amount of adsorbed dye and the correlation between
the photoanode architectures and the dye loading capac-
ity being also presented in Table 1. In the light of the
above results, the evolution of the photovoltaic parame-
ters for each architecture of the photoanodes can be
explained. In case of PhA3 architecture, the light scatter-
ing layer has a dual effect on the photovoltaic parameters
reflecting in the increase of J
SC
(36% more than without
LS) and V
OC
(21% respectively). An expected diminution
of the dye loading capacity is caused by decreasing the
surface area of the microcrystalline brookite TiO
2
particles for dye anchoring. Therefore, the beneficial evo-
lution of J
SC
is directly determined by the brookite
TiO
2
_LS microcubes characterized by the mean particle
size of 659 nm and high refractive index (n
brookite
=2.64)
reflects the incident sunlight on the dye. Facilitating light
capture in the DSSC by elongating the optical path of the
light inside PhA3 has led to an expansion of the DSSC's
absorption range from 435 to 470 nm covering nearly
entirely the possible interaction between the incident
photons and dye molecules (Figure 7). The architecture
proposed for PhA4 is based on a synergistic effect of both
high dye loading capacity promoted by TiCl
4
treatment
and high light scattering effect enabled by brookite
TiO2_LS microcubes. Indeed, the increase of J
SC
(2.76
times) is directly determined by both microcrystalline
brookite TiO
2
particles increasing the light-absorbing
capability of the dye. and TiCl
4
treatment, which
FIGURE 6 SEM images of
the TiO
2
photoanodes on FTO
substrate
FIGURE 7 The transmittance spectra of all DSSCs using E1
electrolyte and the absorption of DN-F01 dye is in the inset FIGURE 8 J-V measurements of DSSCs using E1 electrolyte at
100 mW/cm
2
irradiation
URSU ET AL.18555
improves the dye loading capacity of the DSSC (from
1.98 10
7
to 3.24 10
7
mol cm
2
). Moreover, DSSC
3 displayed a slightly higher R3 than DSSC 2, the scatter-
ing layer affecting the electrolyte diffusion and therefore
the larger recombination and slower diffusion have stabi-
lized. The same beneficial synergy between the scattering
layer and TiCl
4
treatment was highlighted in the charge
transfer processes of DSSC 4. R3 decreased dramatically
at 5.42 Ω, suggesting easier electron transfer due to the
high crystallization of TiO
2
_LS particles and a facile elec-
trolyte diffusion fostered by the strong adhesion and
homogeneity of the TiO
2
nanoparticle layer resulting in
decrease of the charge recombination process. R2 has
almost the same value, indicating that the charge recom-
bination between the electrolyte and CE is similar for all
DSSCs.
In addition, the improvement of V
OC
(more than
0.120 V) is determined by the high bandgap energy (Eg)
TABLE 1 Photovoltaic performances and EIS parameters obtained from the fitted Nyquist plots of DSSCs
a
DSSC type
TiO
2
thickness (μm) J
SC
(mA) V
OC
(V) FF (%) η(%)
R1
(Ω)
R2
(Ω)
R3
(Ω)
Dye loading
(mol cm
2)
DSSC 1 0.73 3.36 ± 0.10 0.660 ± 0.00 61.1 ± 0.09 1.35 ± 0.04 7.49 2.78 19.56 2.16 10
7
DSSC 2 0.97 3.50 ± 0.12 0.698 ± 0.01 70.7 ± 0.10 1.72 ± 0.08 7.49 2.13 10.71 2.22 10
7
DSSC 3 3 4.76 ± 0.11 0.844 ± 0.04 64.7 ± 0.14 2.60 ± 0.19 7.22 2.63 12.77 1.98 10
7
DSSC 4 3.9 9.69 ± 0.10 0.827 ± 0.03 57.8 ± 0.09 4.62 ± 0.22 7.49 4.22 5.42 3.24 10
7
Best
record
22
- 5.23 0.710 64 2.39 - - - -
a
Device average values and SD are based on five devices.
FIGURE 9 Impedance spectra of DSSCs using E1 electrolyte
under 1 sun illumination with equivalent circuit in the inset
FIGURE 10 The transmittance spectra of DSSC 4_E1 and
DSSC 4_E2
FIGURE 11 J-V measurements of the best DSSC 4_E1 and
DSSC 4_E2 under 1 sun (100 mW/cm
2
) illumination
18556 URSU ET AL.
of the TiO
2
brookite polymorph compared to anatase
phase (Figure 3) which causes only the most negative
conduction band (CB) level of brookite TiO
2
_LS because
due to their identical elementary composition, the energy
level of the valence band (VB) remains similar for both
polymorphs.
The complex architecture of PhA4 is validated by a
photovoltaic efficiency two times higher than the best
efficiency reported for this dye so far, and the future
optimization of the electrolyte to improve the fill factor
(FF) is further proposed. E2 electrolyte consists of high
I
3
species, I
3
/I
ratio having the value of 40/1 com-
paredto20/1incaseofE1electrolyte.Thetransparency
ontheentirePARdomainalongwithUVabsorption
and photovoltaic efficiency of DSSC 4 using both elec-
trolytes were analyzed. As seen in Figures 10 and 11,the
transparencyontheentirePARdomainissimilarfor
both DSSCs, a small decrease in UV absorption and V
OC
isobservedinthecaseofDSSC4_E2.BasedontheMott-
Schottky analysis, the TiO
2
photoanode is characterized
byapositiveslope,accordingtotheexpectedn-type
semiconductor characteristics. In accordance with Mott-
Schottky equation,
30
the flat band potential of the semi-
conductor (E
FB
)isobtainedbyextrapolating1/C
2
to
0, having a value of 0.66 V for DSSC 4_E1 and
0.58 V for DSSC 4_E2 respectively (Figure 12). As it is
known, the V
OC
is closely correlated to the difference
between the redox potential of the electrolyte and the
Fermi level of a semiconductor, a positive shift in the
flat band potential has led to decrease in the V
OC
value
of DSSC 4_E2.
Nonetheless, the expected improvement of FF and η
(Figure 11) is achieved by increasing I
3
species that
favored the charge transfer at the CE/electrolyte inter-
face, R2 decreased more than three times in DSSC 4_E2
and therefore the diminution of the cell serial resistance
(Figure 15 and Table 3).FIGURE 12 Mott-Schottky plot of DSSC 4_E1 and DSSC 4_E2
TABLE 2 The photovoltaic performances of the best DSSC 4_E1 and DSSC 4_E2 under different irradiance
Cell P
in
(mW/cm
2
) Electrolyte Dye J
SC
(mA) V
OC
(V) P (μW) FF (%) η(%)
DSSC 4 20 E1 DN-F01 1.12 0.765 555 64.8 2.77
30 1.20 0.773 629 67.8 2.09
40 1.87 0.793 966.40 65.2 2.41
50 4.34 0.828 2158.74 60.1 4.31
60 5.30 0.834 2644.40 59.8 4.40
70 5.55 0.829 2431 52.8 3.47
80 6.75 0.837 3308.20 58.5 4.13
90 7.23 0.838 3496.32 57.7 3.88
100 9.69 0.827 4619.64 57.6 4.62
DSSC 4 20 E2 DN-F01 1.15 0.738 599.04 70.6 2.99
30 1.30 0.741 685.91 71.2 2.28
40 2.01 0.761 1075.12 70.3 2.68
50 4.86 0.797 2653.1 68.5 5.30
60 5.92 0.802 3200 67.4 5.33
70 6.14 0.804 3312.3 67.1 4.73
80 7.19 0.806 3797.3 65.5 4.74
90 8.02 0.804 3973.25 61.6 4.41
100 9.61 0.805 4912.6 63.5 4.91
URSU ET AL.18557
As important as the transparency on the whole PAR
domain along with UV absorption and high photovoltaic
efficiency, the sustainability of our DSSCs throughout the
year is a defining precondition for the successful imple-
mentation in an agrivoltaic system. In this context, the
correlation between the different irradiance (ranging
from 20 to 100 mW/cm
2
) and the photovoltaic perfor-
mance of DSSC4_E1 and DSSC 4_E2 was investigated
(Table 2). As shown in Figures 13 and 14, for both
DSSCs, a similar evolution of J
SC
is highlighted under dif-
ferent light intensities, from approximately 1.1 to
9.6 mA/cm
2
with a steep rise to 50 mW/cm
2
. EIS analysis
have revealed that the behavior of J
SC
is directly corre-
lated with R3 (Table 3), for example, in the case of
50 mW/cm
2
and 100 mW/cm
2
, the doubling of J
SC
is
caused by decreasing twice the charge transfer resistance
at the TiO
2
photoelectrode/dye/electrolyte interface.
More linear, open circuit potential increases slightly as
the illumination increases and becomes near constant
after 50 mW/cm
2
. The evolution of the output power
(P =IV) as a function of voltage under different light
intensities is shown in Figure 13. The power conversion
of both DSSCs increases with the increase in the light
intensity until 4912.6 μW for DSSC 4_E2 and 4619.64 μW
for DSSC 4_E1, respectively, at 100 mW/cm
2
. Even
though FF decreases slightly as the light intensity
increases, DSSC 4_E2 demonstrates nearly 5% light-to-
electricity conversion efficiency for illumination intensity
between 50 and 100 mW/cm
2
with a maximum at
60 mW/cm
2
. Our DSSCs have highlighted excellent per-
formance even under conditions of not optimum illumi-
nation, a real operation mode of the outdoor
applications, such as a greenhouse throughout the day,
month, or season.
FIGURE 13 The photovoltaic parameters of the best DSSC4_E1 under different irradiance: (A) J-V curves; (B) The calculated power P
as function of voltage; (C) Light intensity dependence of V
OC
and J
SC
and (D) Evolution of the efficiency and FF as function of the light
intensity
18558 URSU ET AL.
4|CONCLUSION
In a trade-off between transparency and photovoltaic effi-
ciency of DSSC, a complex architecture of the photoa-
node is designed, built, and optimized for loading with
an affordable commercial UV dye, DN-F01 dye. The
architecture consists of TiO
2
_BL layer, TiO
2
_NP layer,
TiO
2
_LS layer, and TiCl
4
treatment. The future
improvement of FF and ηis achieved by increasing I
3
species that favored the charge transfer at the
CE/electrolyte interface. The best DSSC has achieved a
photovoltaic efficiency of nearly 5% at an irradiance
between 50 and 100 mW/cm
2
.
In conclusion, the preservation of the maximum effi-
ciency throughout all four seasons, as well as the selec-
tive absorption of UV, a harmful radiation to plant DNA,
while maintaining the transparent PAR imposed by the
growing of plants in the greenhouse, have demonstrated
experimentally that the implementation of DSSC in
autonomous greenhouses can be taken into consideration
as a viable and inexpensive concept for future series pro-
duction of greenhouses. In future work, we will study the
stability of these DSSCs in “outdoor”conditions and the
evolution of plants in the conditions imposed by the
greenhouse roof with DSSCs based on UV selective
absorption, together with the optimization of the
production cost.
TABLE 3 Impedance parameters obtained from the fitted
Nyquist plots of DSSC 4_E1 and DSSC 4_E2 under 50 and
100 mW/cm
2
of the light intensity
DSSC R1 [Ω]R2[Ω]R3[Ω]
DSSC 4_E1 under 50 mW/cm
2
7.49 5.34 10.29
DSSC 4_E1 under 100 mW/cm
2
7.49 4.2 5.42
DSSC 4_E2 under 50 mW/cm
2
7.6 1.6 9.7
DSSC 4_E2 under 100 mW/cm
2
7.6 1.3 5.15
FIGURE 14 The photovoltaic parameters of the best DSSC4_E2 under different irradiance: (A) J-V characteristics; (B) The calculated
power P as function of voltage; (C) Light intensity dependence of V
OC
and J
SC
and (D) Evolution of the efficiency and FF as function of the
light intensity
URSU ET AL.18559
ACKNOWLEDGEMENT
This work was supported by a grant of the Romanian
Autoritatea National
a pentru Cercetare Stiintific
a and
Innovation, UEFISCDI Project No. PN-III-P2-2.1-PED-
2019-2091, PNCDI III.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are avail-
able from the corresponding author upon reasonable
request.
ORCID
Marinela Miclau https://orcid.org/0000-0003-4749-
9518
REFERENCES
1. Sekiyama T, Nagashima A. Solar sharing for both food and
clean energy production: performance of Agrivoltaic Systems
for Corn, a typical shade-intolerant crop. Environments. 2019;
6:65.
2. Weselek A, Ehmann A, Zikeli S, Lewandowski I, Schindele S,
Högy P. Agrophotovoltaic systems: applications, challenges,
and opportunities. A review. Agron Sustain Dev. 2019;39:39-58.
3. Adeh EH, Good SP, Calaf M, Higgins CW. Solar PV power
potential is greatest over croplands. Sci Rep. 2019;9:11442.
4. Marucci A, Cappuccini A. Dynamic photovoltaic greenhouse:
energy efficiency in clear sky conditions. Appl Energy. 2016;
170:362-376.
5. Ahemd HA, Al-Faraj AA, Abdel-Ghany AM. Shading green-
houses to improve the microclimate, energy and water saving
in hot regions: a review. Sci Hortic. 2016;201:36-45.
6. Urena-Sanchez R, Callejon-Ferre AJ, Perez-Alonso J, Carreno-
Ortega A. Greenhouse tomato production with electricity gen-
eration by roof-mounted flexible solar panels. Sci Agric. 2012;
69:233-239.
7. Dominguez A, Kleissl J, Luvall JC. Effects of solar photovoltaic
panels on roof heat transfer. Sol Energy. 2011;85:2244-2255.
8. Raviv M, Antignus Y. UV radiation effects on pathogens and
insect pests of greenhouse-grown crops. Photochem Photobiol.
2007;79:219-226.
9. Gross J. Pigment in Vegetables: Chlorophylls and Carotenoids.
New York: Springer Science & Business Media; 1991.
10. Moretti S, Marucci A. A photovoltaic greenhouse with variable
shading for the optimization of agricultural and energy produc-
tion. Energies. 2019;12:2589.
11. Hollingsworth JA, Ravishankar E, O'Connor B, Johnson JX,
DeCarolis JF. Environmental and economic impacts of
solar-powered integrated greenhouses. JIndEcol. 2019;24:
234-247.
12. Chen J, Xu F, Ding B, Wu N, Shen Z, Zhang L. Performance
analysis of radiation and electricity yield in a photovoltaic
panel integrated greenhouse using the radiation and thermal
models. Comput Electron Agric. 2019;164:104904.
13. Wang T, Wu G, Chen J, et al. Integration of solar technology to
modern greenhouse in China: current status, challenges and
prospect. Renew Sust Energ Rev. 2017;70:1178-1188.
14. Aira J-R, Gallardo-Saavedra S, Eugenio-Gozalbo M, Alonso-
G
omez V, Muñoz-García M- ´
A, Hern
andez-Callejo L. Analysis
of the viability of a photovoltaic greenhouse with semi-
transparent amorphous silicon (a-Si) glass. Agronomy. 2021;11:
1097.
15. Corrado C, Leow SW, Osborn M, Chan E, Balaban B,
Carter SA. Optimization of gain and energy conversion effi-
ciency using front-facing photovoltaic cell luminescent solar
concentratordesign. Sol Energy Mater Sol Cells. 2013;111:74-81.
16. Ursu D, Miclau M, Banica R, Vaszilcsin N. Impact of Fe doping
on performances of CuGaO
2
p-type dye-sensitized solar cells.
Mater Lett. 2015;143:91-93.
FIGURE 15 Impedance spectra of (A) the best DSSC 4_E1 and (B) DSSC 4_E2 under 50 and 100 mW/cm
2
of irradiance. The equivalent
circuit diagram used to fit the experimental impedance spectra is presented in the inset
18560 URSU ET AL.
17. Ursu D, Vaszilcsin N, B
anica R, Miclau M. Effect of Al doping
on performance of CuGaO
2
p-type dye-sensitized solar cells.
J Mater Eng Perform. 2015;25:59-63.
18. Ursu D, Vajda M, Miclau M. Investigation of the p-type dye-
sensitized solar cell based on full Cu
2
O electrodes. J Alloys
Compd. 2019;802:86-92.
19. Vajda M, Ursu D, Duteanu N, Miclau M. Low lying valence
band edge materials based on copper oxide for tandem dye-
sensitized solar cells. Mater Lett. 2020;275:128151.
20. Vajda M, Ursu D, Mosoarca C, Duteanu N, Miclau M. Experi-
mental investigation of hydrogen insertion in copper oxide on
photovoltaic performance of p-type dye-sensitized solar cell. Int
J Energy Res. 2020;45:5309-5317.
21. Lu L, Ya'acob ME, Anuar MS, et al. Thermal analysis of a por-
table DSSC mini greenhouse for botanical drugs cultivation.
Energy Rep. 2020;6:238-253.
22. Roslan N, Ya'acob ME, Radzi MAM, Hashimoto Y,
Jamaludin D, Chen G. Dye sensitized solar cell (DSSC) green-
house shading: new insights for solar radiation manipulation.
Renew Sust Energ Rev. 2018;92:171-186.
23. Ntinas GK, Kadoglidou K, Tsivelika N, et al. Performance and
hydroponic tomato crop quality characteristics in a novel
greenhouse using dye-sensitized solar cell Technology for Cov-
ering Material. Horticulturae. 2019;5:42.
24. Marinado T, Nonomura K, Nissfolk J, et al. How the nature of
Triphenylamine-polyene dyes in dye-sensitized solar cells
affects the open-circuit voltage and electron lifetimes. Lang-
muir. 2010;26:2592-2598.
25. Chang S-M, Lin C-L, Chen Y-J, Wang H-C, Chang W-C, Lin L-
Y. Improved photovoltaic performances of dye-sensitized solar
cells with ZnO films co-sensitized by metal-free organic sensi-
tizer and N719 dye. Org Electron. 2015;25:254-260.
26. Ardizzone S, Bianchi CL, Cappelletti G, Gialanella S,
Pirola C, Ragaini V. Tailored Anatase/Brookite Nanocrystal-
line TiO
2
. The optimal particle features for liquid and gas-
phase photocatalytic reactions. JPhysChemC. 2007;111:
13222-13231.
27. Archana J, Harish S, Sabarinathan M, et al. Highly efficient
dye-sensitized solar cell performance from template derived
high surface area mesoporous TiO
2
nanospheres. RSC Adv.
2016;6:68092-68099.
28. Thommes M, Kaneko K, Neimark AV, et al. Physisorption of
gases, with special reference to the evaluation of surface area
and pore size distribution (IUPAC technical report). Pure Appl
Chem. 2015;87:1051-1069.
29. Nogueira AF. Shades of transparency. Nature.Energy. 2020;5:
428-429.
30. Gelderman K, Lee L, Donne SW. Flat-band potential of a semi-
conductor: using the Mott–Schottky equation. J Chem Educ.
2007;84:685-688.
How to cite this article: Ursu D, Vajda M,
Miclau M. Highly efficient dye-sensitized solar cells
for wavelength-selective greenhouse: A promising
agrivoltaic system. Int J Energy Res. 2022;46(13):
18550‐18561. doi:10.1002/er.8469
URSU ET AL.18561
