Vincent 2018
Vincent 2018
Vincent 2018
Premkumar Vincent, Sang-Chul Shin, Ji Soo Goo, Young-Jun You, Boram Cho,
Suwoong Lee, Dong-Wook Lee, Se Ra Kwon, Kwun-Bum Chung, Jae-Joon Lee, Jin-
Hyuk Bae, Jae Won Shim, Hyeok Kim
PII: S0143-7208(18)30672-7
DOI: 10.1016/j.dyepig.2018.06.025
Reference: DYPI 6830
Please cite this article as: Vincent P, Shin S-C, Goo JS, You Y-J, Cho B, Lee S, Lee D-W, Kwon SR,
Chung K-B, Lee J-J, Bae J-H, Shim JW, Kim H, Indoor-type photovoltaics with organic solar cells
through optimal design, Dyes and Pigments (2018), doi: 10.1016/j.dyepig.2018.06.025.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Indoor-type Photovoltaics with Organic Solar Cells through Optimal
Design
Premkumar Vincent1,+, Sang-Chul Shin2,+, Ji Soo Goo2, Young-Jun You2, Boram Cho3,
Suwoong Lee3, Dong-Wook Lee3, Se Ra Kwon4, Kwun-Bum Chung4, Jae-Joon Lee5, Jin-
PT
Hyuk Bae1*, Jae Won Shim2*, and Hyeok Kim6*
1
School of Electronics Engineering, Kyungpook National University, 80 Daehakro, Bukgu,
RI
Daegu, 702-701, Republic of Korea
SC
2
Research Center for Photoenergy Harvesting & Conversion Technology (PHCT), Division of
Electronics and Electrical Engineering, Dongguk University, Seoul 04620, Republic of Korea
U
3
Korea Institute of Industrial Technology (KITECH), 320 Techno-sunhwanro, Yuga-myeon,
AN
Dalseong-gun, Daegu 42990, Republic of Korea
4
Division of Physics and Semiconductor Science, Dongguk University, Seoul 100715,
M
Republic of Korea
D
5
Research Center for Photoenergy Harvesting & Conversion Technology (PHCT),
TE
Republic of Korea
EP
6
Department of Electrical Engineering, ERI, Gyeongsang National University, 501
jwshim@dongguk.edu, hyeok.kim@gnu.ac.kr
AC
ABSTRACT:
Indoor solar cells are an attractive option to install in buildings to harvest the energy emitted
1
ACCEPTED MANUSCRIPT
energy harvesting unit. Previous research has utilized amorphous silicon and dye-sensitized
solar cells for power generation from a fluorescent light source. In our study, we evaluated
light spectrum from an indoor light-emitting diode (LED). Because the absorption peak of
poly(3-hexylthiophene) overlaps the LED spectrum peak, it serves as a good candidate for
PT
indoor light harvesting under an LED source. We have extracted the power conversion
RI
efficiency of the solar cell under different LED illuminance values, such as 200, 800, 1000,
and 2000 lx. With the help of finite-difference time domain simulations, we optimized the
SC
solar device structure for each illuminance. The results detailed in this article were intended
to assess the applicability of organic solar cells as indoor photovoltaics, and to provide a
U
methodology to optimize the photovoltaic structure to maximize its efficiency.
AN
Keywords: indoor photovoltaic, low-light harvesting, FDTD, P3HT:ICBA, optical
M
modelling.
D
TE
C EP
AC
2
ACCEPTED MANUSCRIPT
1. INTRODUCTION
While it has been continuously researched to maximize the power harvested from the sun
PT
[1–8], there is limited investigation into harnessing the microwatt power output from indoor
lighting [9–13]. Microelectronic devices have a power demand in the microwatt range, and thus,
RI
indoor photovoltaics (IPVs) will be suitable for use in such micro-energy harvesting applications
SC
[10,14]. Hybrid energy harvesting methods may utilize IPV devices to power applications such
U
micro-embedded systems require energy harvested from multiple environmental sources. IPVs
AN
have been used in such multi-harvesting systems to achieve, ideally, an unlimited energy supply
to the device.
M
To the best of our knowledge, there are no standardized indoor lighting conditions similar
D
to the AM1.5G spectrum (1000 W/m2). Indoor light can be comprised of only artificial lighting,
TE
or a mixture of natural and artificial lighting. The irradiance of the artificial indoor illumination
spectrum falls in the range of 0–1000 mW/m2 for the light emitting diode (LED) source used in
EP
this study. Some studies had used ray-tracing programs such as RADIANCE and DAYSIM, along
with computer-aided design (CAD) models, to simulate the irradiance due to natural and
C
artificial illumination sources inside a building [20–24]. 3DS MAX DESIGN had also shown to
AC
have measured results similar to the more widely used DAYSIM software [25]. Different indoor
illumination conditions which include parameters such as time of the day, seasonal changes,
shadowing of light due to blinds, etc., had been researched in past studies [26–30]. These studies
intent to realize an efficient lighting system for their modelled environment. Another study had
3
ACCEPTED MANUSCRIPT
conditions [31]. Dye-sensitized solar cells (DSSC) had already been commercialized (GCell
indoor solar cells) for their high power conversion efficiency (PCE) under indoor illumination
[32–35]. Studies that used thin film silicon modules and DSSC modules have shown that by
optimizing the structure for specific indoor light spectra, the efficiency can be maximized
PT
[36,37]. Silicon based IPVs showed the best efficiency under naturally lit environment, whereas
RI
DSSCs showed higher efficiencies under compact fluorescent lamp and LED illumination. A
study comparing the efficiency of inorganic and organic indoor solar cells under fluorescent and
SC
incandescent lighting using a maximum power point (MPP) tracking algorithm concluded that
DSSCs exhibited the best power density under fluorescent lighting, and poly-Si showed superior
U
performance under incandescent lighting [38]. This was due to DSSCs having better spectral
AN
matching with the illumination source’s spectra. Since indoor artificial light sources emit light
only in the visual wavelengths, photovoltaic devices that absorb maximum photons in those
M
wavelengths will provide the best efficiency. Similar MPP tracking algorithm studies have been
D
extensively investigated for IPV applications [39–42]. Organic PVs, however, show a steady
TE
trend of improving the efficiency and lifetime of the device, which makes them promising
The Shockley-Queisser limit on solar cell efficiency had been deduced as 58.4% for an
C
RGB white LED source for an optimal bandgap of ~1.9 eV by a previous study [10]. The study
AC
derived conclusion that absorber materials with bandgaps of 1.9 – 2.0 eV are the most suitable
for RGB white LED energy harvesting. Thus, wide-bandgap light-absorbing materials would
exhibit higher performance under LED illumination. In our study, we optimized a wide-bandgap
organic photovoltaic (OPV) to maximize the absorption of the LED spectrum. P3HT absorbs
4
ACCEPTED MANUSCRIPT
wavelengths of light from 400 nm to 650 nm effectively [23]. This makes it a suitable candidate
for indoor solar cells under LED illumination as the LED white light spectrum peaks at ~455 nm
and spans from ~400 nm to ~650 nm. This coincides with the absorption coefficient of P3HT,
thus making P3HT as an ideal absorber for the white LED light spectrum. The matching of
PT
absorption window and the illumination source spectrum had been found to be more beneficial
RI
through another study [43].
PCE of the P3HT:ICBA IPVs were derived under different light intensities, such as 200,
SC
800, 1000, and 2000 lx. The solar cell device structure was optimized to obtain the maximum
U
short-circuit current density (Jsc,max) through finite-difference time-domain (FDTD) simulations.
AN
We utilized Lumerical, an FDTD solution software, which has a built-in CAD module for
drawing the solar cell structure, to solve for the distribution of the electric field due to light
M
inside the IPV. The originality of our study is that we demonstrate that a common organic
absorber (P3HT) based solar cell, which when structurally optimized using optical simulation,
D
can exhibit efficiencies up to ~8% under low-power white LED irradiance. The active layer
TE
thickness at which the maximum photocurrent was extracted from the solar cell was also
obtained by comparing the simulation results with the semi-empirically fitted, experimentally
EP
P3HT (4002E, Rieke Metals, Lincoln, NE, U.S.A.) and ICBA (Luminescence
Technology Corp., New Taipei City, Taiwan) were mixed in 1,2-dicholorobenzene (DCB)
5
ACCEPTED MANUSCRIPT
(Aldrich, St. Louis, MO, U.S.A.) at 1:1 weight ratio. Six different total concentrations (20, 30, 40,
50, 60, and 80 mg/ml) of P3HT:ICBA were prepared. The solution was stirred overnight, in a
nitrogen-filled glove box, at 70 °C, and 0.4 wt.% of ethoxylated polyethylenimine (PEIE) [80%
ethoxylated, Mw ~70,000 g/mol, 35–40 wt.% in water purchased from Aldrich, St. Louis, MO,
PT
U.S.A.] was diluted in 2-Methoxyethanol (Aldrich, St. Louis, MO) and stirred overnight in
RI
ambient air.
SC
Indium tin oxide (ITO)-coated glass with a sheet resistivity of 10 Ω/sq (OMNI.0782, All
U
for LAB, Republic of Korea) was used as the substrate. Substrates were ultrasonicated in a
AN
detergent bath (Liqui-Nox® Phosphate-Free Liquid Detergent, Alconox, Inc., White Plains, NY,
U.S.A.), rinsed with deionized (DI) water, and followed by sequential cleaning in ultrasonic
M
baths of DI water, acetone, and 2-propanol. The cleaned substrates were blow-dried using a
D
nitrogen gun. A layer of 350-nm-thick SiOx (Silicon Monoxide, Kurt J. Lesker, Clairton, PA,
U.S.A.) was deposited on ITO through the e-beam deposition process (KVE-E2000 series, Korea
TE
Vacuum Tech. Ltd., Republic of Korea). It was deposited through a shadow mask (to electrically
EP
isolate almost half of the glass/ITO area), at a rate of 0.5–0.8 nm/s, while maintaining a base
pressure of ~1 × 10-6 Torr. The substrates were ultrasonicated in 2-propanol for 10 minutes and
C
dried. PEIE solution was filtered through a 0.2-µm-pore PTFE filter, and a 10-nm layer was
AC
spin-coated onto the substrates, in ambient atmosphere, at a speed of 5000 rpm for 1 min. The
PEIE-cast substrates were dried at 110 ºC for 10 minutes in air, on a hotplate. The substrates
were then transferred into the glove box for further coating. The P3HT:ICBA films were spin
cast at 700 rpm for 30 s, after filtering through a 0.2-µm-pore PTFE filter. 20-, 30-, 40-, 50-, 60-,
and 80-mg/ml concentrations of P3HT:ICBA produced 80-, 130-, 200-, 230-, 280-, and 400-nm-
6
ACCEPTED MANUSCRIPT
thick layers. The coated substrates were then annealed at 150 ºC for 10 minutes under a nitrogen
atmosphere. 10 nm of MoOx was deposited onto the active layers using a vacuum thermal
glove box. It was deposited through a shadow mask at a rate of 0.1–0.15 nm/s, while maintaining
PT
a base pressure of ~6 µPa. A 150-nm Ag layer was vacuum deposited at a rate of 0.1–0.2 nm/s
RI
while maintaining the same pressure in the chamber. The active area of the PV device was 0.1
cm2.
SC
2.1.3 Characterization:
U
The active layer thicknesses were measured using a profilometer (a-step2000, Tencor
AN
Instruments, Milpitas, CA, U.S.A). Dark and illuminated current density-voltage (J-V)
characteristics were measured using 2401, Keithley Instruments, Cleveland, OH, source meter,
M
operated by the K730 program by McScience, Republic of Korea. The AM1.5 G solar simulator
D
used was Xe55 which was purchased from McScience, Republic of Korea. The AM1.5G had a
total irradiance of 100 mW/cm2. The indoor LED light simulator was also purchased from
TE
McScience, Republic of Korea. The LED lamp used was white linear COB LED from
EP
McScience. The irradiance of the LED lamps for different illuminance values is provided in
supplementary figure 1(a). The LED power controller used was MLP-1230 from M-VISION,
C
Republic of Korea. The LED lamp output illuminances of 200, 800, 1000, and 2000 lx were used
AC
2D simulations were executed on the solar cell structure designed using Lumerical, an FDTD
solution software. The device was constructed with complex refractive index values
7
ACCEPTED MANUSCRIPT
implemented on each layer. The complex refractive index values of each layer is provided in
supplementary figure 1(b). The structure was meshed to such an extent that further increases in
mesh density did not affect the results of the simulation. FDTD simulations computed the
distribution of the normalized electric field due to light inside the solar cell structure. Using this,
PT
the ideal photo-generated short-circuit current density (Jsc,ideal) from the active layer of the solar
RI
cell was derived. Jsc,ideal is simulated in FDTD with the assumption that the IQE is 100 % similar
to other cited references [44,45]. The simulation setup had two perfectly matched layers (PML):
SC
one below the bottom electrode, and one above the top electrode. Periodic boundaries were
provided along the x-axis as the light source was input at 90° to the ITO electrode. A continuous-
U
wave (CW) normalized source was used to obtain the impulse response of the system. To
AN
simulate the Jsc,ideal due to a user-defined power spectrum (AM1.5G spectrum or indoor LED
light spectrum), the impulse response was multiplied with the respective source spectrum.
M
the supplementary under the “CW-Normalization in FDTD” section. The thickness of the active
TE
layer was varied from 50 nm to 400 nm with an incrementing step size of 10 nm. In order to take
into account the losses due to recombination and resistance, we used the method used in
EP
reference [45] and reference [46] to take into account the recombination and parasitic resistance
effects on the PCE. By fitting the data to a trend line, we found a good enough approximation for
C
the optimized active layer thickness. This minimized the extensive number of experiments
AC
required to find the optimized active layer thickness, but at the same time, provided a satisfying
approximate optimized thickness result. Empirical fitting of the internal quantum efficiency
(IQE), fill factor (FF), and open-circuit voltage (Voc) was used to predict the respective values
8
ACCEPTED MANUSCRIPT
between the active layer thickness ranges of 130–400 nm. Finally, the PCE was calculated
theoretically from the simulated values, and the most optimized structure was devised.
3. RESULTS:
PT
The structure in Figure 1(a) was designed in Lumerical, and FDTD simulations were run
to calculate the Jsc,ideal that could be obtained from the IPV. The IQE of the IPV was assumed to
RI
be 100% for the FDTD simulation in order to find the ideal short-circuit current density [44,47].
SC
The active layer of the solar cell was varied in order to maximize the electric field due to light
concentrated inside it. This allowed more photons to be absorbed in the active layer, yielding a
U
higher Jsc,ideal. The indoor LED (1000 lx) light spectrum, and the absorption of 180-nm-thick
AN
P3HT:ICBA active layer, is shown in Figure 1(b). Since P3HT has high absorption in the
wavelength range of ~400–650 nm, it shows good absorption of the indoor LED spectral
M
irradiance. Jsc,ideal vs the active layer thickness is plotted in Figure 2. It follows a similar
D
oscillating trend to that explained in [44] due to the interference of light reflected off the top
electrode in the active layer. The simulated Jsc,ideal values for different illuminations have been
TE
plotted in Supplementary Figure 2(a). The experimental results of the fabricated devices,
EP
illuminated under AM1.5 G and 1000-lx LED, are provided in Table 1 (a) and 1 (b), respectively.
Experimentally extracted J–V values (Supplementary Figure 2(b)) were used to calculate the real
C
To obtain the most optimized structure providing the highest PCE, we used 2nd order polynomial
fitting of the IQE scatter plot, as shown in Figure 3(a). We opted for 2nd order polynomial fit for
IQE in order to utilize a fitting trend similar to the one used in reference [45], and to use the
9
ACCEPTED MANUSCRIPT
same fitting scheme for all the different IQE plots under different illumination sources. Although
the IQE points under LED illumination could be curve fit using linear fitting, using a 2nd order
polynomial fit does not skew the data as it will provide a fit close to linear fitting, by reducing
the coefficient of the 2nd order, in the case that the data points show a linear trend. We utilized
PT
error bars in order to provide weights to data points during fitting. Instrumental weighing
RI
function was used to calculate the weight of each data point used in the curve fitting. We also
used 2nd order polynomial fitting for Voc as it showed good fit with the data points. FF does not
SC
have seem to share a common trend between these studies. [46,48–53]. Thus, we utilizing the
least polynomial order function to fit the curve. The fill factor was fit using 3rd order polynomial
U
fitting as 2nd order polynomial fitting lead to under-fitting of the data points in few plots. FF and
AN
Voc plots are shown in shown in Figure 3(b). These fittings hold true only within the active layer
thicknesses of 130–400 nm, and are not necessarily accurate outside this range. However, our
M
study focuses on optimizing the active layer thickness within the fitted range. We calculated the
D
practical short-circuit current density (Jsc,cal) that could be obtained from the solar cell structure
TE
by utilizing the fitted IQE values and the simulated Jsc,ideal values in Equation 1. The reason as to
why Jsc,cal is calculated is provided in the supplementary. This allowed us to calculate the short-
EP
circuit current density, which was closer to that which would have been obtained through
experiments. Figure 4(a) shows a comparison between the calculated short-circuit current density
C
(Jsc,cal) and the experimentally obtained Jsc,exp under AM1.5 G and 1000 lx LED illumination
AC
where Voc,fit is the open-circuit voltage obtained after curve fitting, FFfit is the fill factor
of the solar cell obtained after curve fitting, and Pin is the input power of the illumination source,
10
ACCEPTED MANUSCRIPT
which in this case, was a 1000-lx indoor LED source. Jsc and PCE optimization for indoor LED
source with an illuminance of 200, 800, and 2000 lx, are provided in the supplementary section.
The PCE obtained theoretically (PCEcal) was compared to the experimentally extracted PCE
(PCEexp) in Figure 4(b). The best PCE was obtained theoretically as 5.38% at an active layer
PT
thickness of 210 nm. Although the Jsc,cal was maximum at an active layer thickness of 320 nm for
RI
an illumination source of 1000 lx LED, due to the decrease of FFfit after ~230 nm of thickness,
the PCE did not improve. It should also be noted that the PCE for devices illuminated under a
SC
2000-lx indoor LED spectrum reached ~8.8% (Supplementary Table 1(c)). This was due to the
absorption peak of P3HT (~500 nm) nearly aligning with the spectral irradiance maximum at
~455 nm.
U
AN
The electronic behavior of the solar cell was studied using the equivalent circuit of the
M
solar cell (Figure 5(a)) [54–56]. The photo-generated current produced by the solar cell can be
modeled by a current source, Jph, which will depend on the parasitic resistances. These parasitic
D
resistances, series resistance (Rs) and shunt resistance (Rsh), can be derived from the dark current
TE
vs. voltage graph of the solar cell, which is shown in figure 5(b). The Rs and Rsh values were
calculated from the J-V curves under different independent illuminations and are provided in
EP
Tables 2 (a), 2 (b), and Supplementary Tables 2 (a)-(c). High shunt resistance and low series
C
resistance are indications of a good solar cell device. Lower series resistance indicates that the
AC
contacts between the active layer and the electrodes are good with few to no defects. High shunt
resistance can determine that there is a low leakage of current from the device. The parasitic
resistances values provided in Table 2 (a), (b) indicate that the IPV devices fabricated had good
11
ACCEPTED MANUSCRIPT
Active layer thickness optimization was required respective to each illuminance intensity of
the light source. Thus, our study provided a methodology to optimize the IPV structure using
FDTD simulation for different illumination scenarios, which would hopefully aid in fabricating
PT
4. DISCUSSION
RI
The discussion section will cover the relevance of this work, limitations, and possible
SC
applications (future works).
4.1 RELEVANCE:
U
The relevance of our work is to emphasize the usage of a common organic semiconductor solar
AN
cell design in indoor light harvesting application. The indoor LED spectra has its maximum
M
irradiance peaks at ~455 nm and outputs considerable irradiance from ~500 nm to ~650 nm
wavelengths. The absorption range of P3HT correlates well with the LED’s spectrum. When the
D
solar cell’s spectral sensitivity matches the radiant spectra of the light source, maximum power
TE
could be harvested. Different solar cell technologies show differences in their performance when
illuminated under different light sources. Y. Li et al. had researched the performances of three
EP
different solar cell modules made from amorphous silicon, polycrystalline silicon, and DSC [57].
The study found that most power was harvested under incandescent light source, followed by
C
compact fluorescent light (CFL), and then by LED. As LEDs are considered as future energy
AC
saving solutions for lighting, it compels the need for photovoltaic devices which can efficiently
harvest its radiant spectra. C.L. Cutting et al. had observed that organic photovoltaics that
perform significantly worse than silicon solar cells, can outperform the later under indoor LED
light conditions [58]. Another work studied the efficiency of P3HT:PC60BM based solar cell
12
ACCEPTED MANUSCRIPT
under low light conditions and achieved ~7% under 1000 lx fluorescent lamp [59].
b0]dithiophene-2,6-diyl-alt-(4-(2-ethylhexyl)-3-fluorothieno[3,4-b]thiophene)-2-carboxylate-2-
PT
under an inorganic LED and a fluorescent tube with the aim of achieving high efficiency
RI
polymer photovoltaic devices for low-power indoor illumination [60]. It showed that
P3HT:ICBA can reach high efficiencies of ~13% when illuminated under 500 lx LED light with
SC
an intensity of 0.5 – 1 mW cm-2. Since the white LED in our study outputs a lower intensity, we
realize ~8% PCE for 2000 lx LED which has an intensity of about 712 µW cm-2.The study
U
conducted by B. Minnaert et al. concluded that the efficiency of the solar cells depends on the
AN
illumination source [43]. They conducted simulations to validate the usage of chalcogenide
photovoltaic cells under LED, incandescent lamp, and a “warm” and a “cool” fluorescent tube.
M
Their conclusion was that the broadening of the absorption window of the solar cell was not
D
beneficial in all cases. Their results back up our choice of active material for the indoor
TE
photovoltaic as it fits the LED’s spectra. The article also discusses the lowering of FF and Voc
with decreasing light intensities. We also examined a drop in the Voc and the FF values under
EP
lower light intensities. This could be due to the dependence of parasitic resistances of the cells on
the irradiance of the illumination source as explained in reference [61]. As explained in the study
C
by A. Virtuani et al., the effect of parasitic current is significant under low irradiances and thus,
AC
cannot be neglected. This effect was observed in our devices which showed changes in FF and
Voc with difference irradiance (Table 1 (a), (b) and Supplementary Tables 1 (a)-(c)). Roland
Steim et al. also conducted experiments to illustrate the impact of series and parallel resistances
in the photovoltaic device for low light conditions [59]. Their solar cell was based on P3HT and
13
ACCEPTED MANUSCRIPT
PC60BM which exhibited decrease in FF and Voc with lowering irradiance of the light source as
expected. Their best PCE was 7% under fluorescent lamp emitting an illuminance of 1000 lx. A.
Sacco et al. researched the performance of different photovoltaic modules under incandescent
lamp, halogen lamp, and fluorescent lamp [62]. Their studies showed that P3HT based organic
PT
photovoltaic module produced low PCE under these light sources (with maximum efficiency of
RI
2.06 % under 5000 lux fluorescent tube lamp). Our solar cell device achieved 5.4 % PCE under
1000 lx white LED light source. It reached a PCE of 8.8 % under 2000 lx. This result is not far-
SC
fetched as Roland Steim et al. and S.-S. Yang had concluded that OPVs exhibited high
performance when illuminated under low light [59,60]. This could be explained by matched
U
spectral absorption in combination with lower parasitic resistances at higher illumination
AN
irradiance. The originality of our work is that we use FDTD optical simulations to optimize the
solar cell layer thickness in order to absorb the most amount of photons from the low-power
M
4.2 LIMITATIONS:
TE
There are few technical limitations that are not part of the scope of this article. The active area of
EP
the solar cell is about 0.1 cm2 which will not reproduce the same result as a manufactured
module due to the variation of PCE with device area [63,64]. Although the semi-empirical fitting
C
follows a reported method [45], it does not legitimate the curve fitting using theoretical
AC
procedure. It is based of experimental data, and thus, the fitted curve does not hold true outside
its boundaries.
14
ACCEPTED MANUSCRIPT
Unlike for outdoor conditions where there is a standard illumination condition, indoor lighting
does not have a standardized spectrum. This is partly because indoor lighting is a mixture of the
amount of sunlight that illuminated the room and the artificial lighting in the room. As this is
unique to a building structure, predicting the illuminance inside the architecture is required.
PT
Indoor illumination using software like DAYSIM, which uses RADIANCE, had been used to
RI
study and control the amount of light in a modelled environment under different scenarios
[30,65]. 3DS MAX DESIGN also uses a similar ray tracing technique as RADIANCE to measure
SC
the incoming light flux in the modelled surrounding [22,65]. By using such simulation software,
we intent to find the extent of power saving that could be achieved with the use of our optimized
The application of solar cells to harvest indoor lighting is a promising field for integrated self-
D
powering circuits. Our study presented the utilization of one of the most common organic photo-
absorber, P3HT, in IPV devices. We studied the IPV’s characteristics under different indoor
TE
LED illuminance and compared it to the simulation results. From theoretical calculations, it was
EP
determined that the P3HT:ICBA-based solar cell could reach a maximum PCE of ~5.38% under
1000-lx illuminance in the indoor LED spectrum used in the study. As P3HT’s absorption peak
C
absorber for the LED spectrum. Thus, when the illuminance was increased to 2000 lx, the PCE
reached ~8.74%, which is close to the commercialization efficiency of 10%. This study had
Similar methodology could be expanded to find optimum structures under different LED
15
ACCEPTED MANUSCRIPT
bandgap could be fine-tuned to absorb the whole spectrum of artificial light efficiently. Since
most IPV materials are semi-transparent, they could also serve as energy-harvesting tinted
windows. Such possibilities of IPV attracts the improvisation of its efficiency and lifetime
PT
through research.
RI
5. Acknowledgements
H.K. would like to acknowledge the Ministry of Trade, Industry & Energy (MOTIE,
SC
Korea) under the Industrial Technology Innovation Program (No. 10063473), and this research
U
(2017-0878) has been conducted with the support of the fund for new professor research
AN
foundation program, Gyeongsang National University, 2017. This research was also supported
by the Technology Development Program to Solve Climate Changes of the NRF funded by the
M
Ministry of Science, ICT & Future Planning (NRF-2016M1A2A2940912), and Basic Science
Research Program through the National Research Foundation of Korea (NRF), funded by the
D
REFERENCES:
C
[1] Brabec CJ, Sariciftci NS, Hummelen JC. Plastic solar cells. Adv Funtional Mater
AC
2001;11:15–26. doi:10.1002/1616-3028(200102)11:1<15::AID-ADFM15>3.0.CO;2-A.
[2] Günes S, Neugebauer H, Sariciftci NS. Conjugated polymer-based organic solar cells.
16
ACCEPTED MANUSCRIPT
crystalline silicon for solar cell applications. Part 1: Native substrates. Sol Energy Mater
[4] Catchpole KR, Mccann MJ, Weber KJ, Blakers AW. A review of thin- " lm crystalline
PT
silicon for solar cell applications . Part 2 : Foreign substrates. Sol Energy Mater
2001;68:173–215. doi:http://dx.doi.org/10.1016/S0927-0248(00)00246-4.
RI
[5] Liang Y, Wu Y, Feng D, Tsai S-T, Son H-J, Li G, et al. Development of New
SC
Semiconducting Polymers for High Performance Solar Cells. J Am Chem Soc
2009;131:56–7. doi:10.1021/ja808373p.
[6]
U
Green MA, Emery K, Hishikawa Y, Warta W, Dunlop ED. Solar cell efficiency tables
AN
(Version 45). Prog Photovoltaics Res Appl 2015;23:1–9. doi:10.1002/pip.2573.
M
[7] Peet J, Kim JY, Coates NE, Ma WL, Moses D, Heeger AJ, et al. Efficiency enhancement
in low-bandgap polymer solar cells by processing with alkane dithiols. Nat Mater
D
2007;6:497–500. doi:10.1038/nmat1928.
TE
2005;4:864–8. doi:10.1038/nmat1500.
C
[9] Steim R, Ameri T, Schilinsky P, Waldauf C, Dennler G, Scharber M, et al. Solar Energy
AC
Materials & Solar Cells Organic photovoltaics for low light applications. Sol Energy
[10] Müller MF, Freunek M, Reindl LM. Maximum efficiencies of indoor photovoltaic devices.
17
ACCEPTED MANUSCRIPT
[11] Lee HKH, Li Z, Durrant JR, Tsoi WC. Is organic photovoltaics promising for indoor
[12] Müller MF. Indoor Photovoltaics: Efficiencies, Measurements and Design. Sol. Cell
PT
Nanotechnol., John Wiley & Sons, Inc.; 2013, p. 203–22.
doi:10.1002/9781118845721.ch8.
RI
[13] Corazza M, Krebs FC, Gevorgyan SA. Lifetime of organic photovoltaics: Linking outdoor
SC
and indoor tests. Sol Energy Mater Sol Cells 2015;143:467–72.
doi:10.1016/j.solmat.2015.07.037.
[14]
U
Vullers RJM, Schaijk R Van, Doms I, Hoof C Van, Mertens R. Solid-State Electronics
AN
Micropower energy harvesting. Solid State Electron 2009;53:684–93.
doi:10.1016/j.sse.2008.12.011.
M
[15] Tan YK, Panda SK. Energy harvesting from hybrid indoor ambient light and thermal
D
energy sources for enhanced performance of wireless sensor nodes. IEEE Trans Ind
TE
[16] Teran AS, Wong J, Lim W, Kim G, Lee Y, Blaauw D, et al. AlGaAs Photovoltaics for
EP
Indoor Energy Harvesting in mm-Scale Wireless Sensor Nodes. IEEE Trans Electron
C
[17] Warneke BA, Pister KSJ. MEMS for distributed wireless sensor networks. 9th Int Conf
[18] Hande A, Polk T, Walker W, Bhatia D. Indoor solar energy harvesting for sensor network
18
ACCEPTED MANUSCRIPT
[19] Yu H, Yue Q. Indoor Light Energy Harvesting System for Energy-aware Wireless Sensor
[20] Leite Didoné E, Wagner A. Semi-transparent PV windows: A study for office buildings in
PT
Brazil. Energy Build 2013;67:136–42. doi:10.1016/j.enbuild.2013.08.002.
RI
photovoltaics in commercial buildings. Sol Energy 2015;115:120–32.
SC
doi:10.1016/j.solener.2015.02.016.
[22] Reich NH, van Sark WGJHM, Turkenburg WC, Sinke WC. Using CAD software to
U
simulate PV energy yield - The case of product integrated photovoltaic operated under
AN
indoor solar irradiation. Sol Energy 2010;84:1526–37. doi:10.1016/j.solener.2010.05.015.
M
[23] Müller M, Wienold J, Walker WD, Reindl LM. Characterization of indoor photovoltaic
devices and light. Conf Rec IEEE Photovolt Spec Conf 2009:000738–43.
D
doi:10.1109/PVSC.2009.5411178.
TE
[24] Müller M, Hildebrandt H, Walker WD, Reindl LM. Simulations and measurements for
indoor photovoltaic devices. Proc. 24th Eur. Photovolt. Sol. Energy Conf., Hamburg: 2009,
EP
p. 4363–7. doi:10.4229/24thEUPVSEC2009-5BV.3.45.
C
[25] Reinhart C, Breton P-F. Experimental Validation of Autodesk® 3ds Max® Design 2009
AC
[26] Newsham GR. Manual Control of Window Blinds and Electric Lighting: Implications for
doi:10.1177/1420326X9400300307.
19
ACCEPTED MANUSCRIPT
[27] Lee ES, Selkowitz SE. The Design and Evaluation of Integrated Envelope and Lighting
[28] Bodart M, De Herde A. Global energy savings in offices buildings by the use of
PT
daylighting. Energy Build 2002;34:421–9. doi:https://doi.org/10.1016/S0378-
7788(01)00117-7.
RI
[29] Mardaljevic J. Validation of a lighting simulation program under real sky conditions. Int J
SC
Light Res Technol 1995;27:181–8. doi:10.1177/14771535950270040701.
[30] Reinhart CF. Lightswitch-2002: a model for manual and automated control of electric
U
lighting and blinds. Sol Energy 2004;77:15–28.
AN
doi:https://doi.org/10.1016/j.solener.2004.04.003.
M
doi:10.1002/ese3.110.
TE
solar cells for efficient power generation under ambient lighting. Nat Photonics
EP
2017;11:372–8. doi:10.1038/nphoton.2017.60.
C
[33] Brown TM, De Rossi F, Di Giacomo F, Mincuzzi G, Zardetto V, Reale A, et al. Progress
AC
in flexible dye solar cell materials, processes and devices. J Mater Chem A 2014;2:10788–
817. doi:10.1039/C4TA00902A.
[34] De Rossi F, Pontecorvo T, Brown TM. Photovoltaics for indoor light harvesting:
20
ACCEPTED MANUSCRIPT
flexible dye solar cells to deliver superior efficiency. Appl Energy 2015;156:413–22.
doi:http://dx.doi.org/10.1016/j.apenergy.2015.07.031.
[35] Yoon S, Tak S, Kim J, Jun Y, Kang K, Park J. Application of transparent dye-sensitized
PT
solar cells to building integrated photovoltaic systems. Build Environ 2011;46:1899–904.
doi:10.1016/j.buildenv.2011.03.010.
RI
[36] Foti M, Tringali C, Battaglia A, Sparta N, Lombardo S, Gerardi C. Efficient flexible thin
SC
film silicon module on plastics for indoor energy harvesting. Sol Energy Mater Sol Cells
2014;130:490–4. doi:https://doi.org/10.1016/j.solmat.2014.07.048.
[37]
U
De Rossi F, Pontecorvo T, Brown TM. Characterization of photovoltaic devices for indoor
AN
light harvesting and customization of flexible dye solar cells to deliver superior efficiency
doi:https://doi.org/10.1016/j.apenergy.2015.07.031.
D
[38] Sridhar N, Freeman D. A study of Dye Sensitized Solar Cells under Indoor and Low Level
TE
Outdoor Lighting: Comparison to Organic and Inorganic Thin Film Solar Cells and
Methods to Address Maximum Power. Conf. Proc. 26th EU-PVSEC, Hamburg: 2011, p.
EP
232–6. doi:10.4229/26thEUPVSEC2011-1CO.10.5.
C
[39] Xiao W, Dunford WG. A modified adaptive hill climbing MPPT method for photovoltaic
AC
power systems. PESC Rec - IEEE Annu Power Electron Spec Conf 2004;3:1957–63.
doi:10.1109/PESC.2004.1355417.
[40] Kim H, Kim S, Kwon CK, Min YJ, Kim C, Kim SW. An energy-efficient fast maximum
power point tracking circuit in an 800-µ W photovoltaic energy harvester. IEEE Trans
21
ACCEPTED MANUSCRIPT
[41] Cabal C, Alonso C, Cid-Pastor A, Estibals B, Seguier L, Leyva R, et al. Adaptive digital
MPPT control for photovoltaic applications. IEEE Int Symp Ind Electron 2007:2414–9.
doi:10.1109/ISIE.2007.4374985.
PT
[42] Daraban S., Petreus D., Morel C. A novel MPPT (maximum power point tracking)
RI
maximum power point in photovoltaic systems affected by partial shading. Energy
2014;74:374–88. doi:10.1016/j.energy.2014.07.001.
SC
[43] Minnaert B, Veelaert P. Efficiency simulations of thin film chalcogenide photovoltaic
U
cells for different indoor lighting conditions. Thin Solid Films 2011;519:7537–40.
AN
doi:https://doi.org/10.1016/j.tsf.2011.01.362.
[44] Gilot J, Barbu I, Wienk MM, Janssen RAJ. The use of ZnO as optical spacer in polymer
M
solar cells: Theoretical and experimental study. Appl Phys Lett 2007;91:113520.
D
doi:10.1063/1.2784961.
TE
[45] Jan G, M. WM, J. JRA. Optimizing Polymer Tandem Solar Cells. Adv Mater
2009;22:E67–71. doi:10.1002/adma.200902398.
EP
[46] Kim T, Kim H, Park J, Kim H, Yoon Y, Kim SM, et al. Triple-junction hybrid tandem
C
solar cells with amorphous silicon and polymer-fullerene blends. Sci Rep 2014;4:7154.
AC
doi:10.1038/srep07154.
[47] Beiley ZM, McGehee MD. Modeling low cost hybrid tandem photovoltaics with the
doi:10.1039/c2ee23073a.
22
ACCEPTED MANUSCRIPT
[48] Sievers DW, Shrotriya V, Yang Y. Modeling optical effects and thickness dependent
doi:10.1063/1.2388854.
PT
[49] Shariff NSM, Saad PSM, Affendi IHH, Mahmood MR, Shariffudin SS. Thickness and
RI
heterojunction organic solar cells. AIP Conf Proc 2016;1733:20052.
doi:10.1063/1.4948870.
SC
[50] Moulé AJ, Bonekamp JB, Meerholz K. The effect of active layer thickness and
U
composition on the performance of bulk-heterojunction solar cells. J Appl Phys
AN
2006;100:94503. doi:10.1063/1.2360780.
[51] Min Nam Y, Huh J, Ho Jo W. Optimization of thickness and morphology of active layer
M
for high performance of bulk-heterojunction organic solar cells. Sol Energy Mater Sol
D
Influence of Mobility, Lifetime, and Space Charge. J Phys Chem Lett 2012;3:3470–5.
doi:10.1021/jz301639y.
C
AC
[53] Xue J, Uchida S, Rand BP, Forrest SR. 4.2% Efficient Organic Photovoltaic Cells With
[54] Cheknane A, Hilal HS, Djeffak F, Benyoucef B, Charles JP. An equivalent circuit
doi:http://dx.doi.org/10.1016/j.mejo.2008.01.053.
23
ACCEPTED MANUSCRIPT
[55] Qi B, Wang J. Fill factor in organic solar cells. Phys Chem Chem Phys 2013;15:8972.
doi:10.1039/c3cp51383a.
[56] Han L, Koide N, Chiba Y, Mitate T. Modeling of an equivalent circuit for dye-sensitized
PT
solar cells. Appl Phys Lett 2004;84:2433–5. doi:10.1063/1.1690495.
[57] Li Y, Grabham NJ, Beeby SP, Tudor MJ. The effect of the type of illumination on the
RI
energy harvesting performance of solar cells. Sol Energy 2015;111:21–9.
SC
doi:https://doi.org/10.1016/j.solener.2014.10.024.
[58] Cutting CL, Bag M, Venkataraman D. Indoor light recycling: a new home for organic
U
photovoltaics. J Mater Chem C 2016;4:10367–70. doi:10.1039/C6TC03344J.
AN
[59] Steim R, Ameri T, Schilinsky P, Waldauf C, Dennler G, Scharber M, et al. Organic
M
photovoltaics for low light applications. Sol Energy Mater Sol Cells 2011;95:3256–61.
doi:https://doi.org/10.1016/j.solmat.2011.07.011.
D
[60] Yang S-S, Hsieh Z-C, Keshtov ML, Sharma GD, Chen F-C. Toward High-Performance
TE
2017;1:1700174–n/a. doi:10.1002/solr.201700174.
EP
[61] Virtuani A, Lotter E, Powalla M. Performance of Cu(In,Ga)Se2 solar cells under low
C
6090(03)00184-6.
doi:https://doi.org/10.1016/j.apenergy.2012.07.001.
24
ACCEPTED MANUSCRIPT
[63] Carlé JE, Helgesen M, Hagemann O, Hösel M, Heckler IM, Bundgaard E, et al.
Overcoming the Scaling Lag for Polymer Solar Cells. Joule 2017;1:274–89.
doi:https://doi.org/10.1016/j.joule.2017.08.002.
PT
[64] Xiao X, Lee K, Forrest SR. Scalability of multi-junction organic solar cells for large area
RI
[65] Horvat M, Dubois M-C. Tools and Methods for Solar Design–An Overview of IEA SHC
SC
Task 41, Subtask B. Energy Procedia 2012;30:1120–30.
doi:https://doi.org/10.1016/j.egypro.2012.11.125.
FIGURE CAPTIONS:
U
AN
Figure 1(a). Indoor photovoltaic device structure with P3HT:ICBA as its active layer. P3HT
M
absorbs the LED light and transfers electrons to the ICBA acceptor. These electrons (and holes)
are collected at the respective electrodes through the buffer layers. In the FDTD simulation, the
D
thickness of the active layer was varied to find the optimum thickness.
TE
Figure 1(b). 1000-lx LED light spectrum is plotted in black. The absorption of a 180-nm-thick
P3HT:ICBA active layer is plotted in red. As shown, P3HT’s absorption band satisfactorily
EP
overlapped the spectral irradiance maxima of the LED light spectrum. Thus, it makes it a viable
C
Figure 2. FDTD simulations provided the highest possible short-circuit current density that
could be obtained assuming the solar cell exhibited 100% IQE. Jsc,ideal obtained under 1000-lx
illumination is plotted against various active layer thicknesses ranging within 50–400 nm.
25
ACCEPTED MANUSCRIPT
Figure 3(a). IQE % was fit for active layer thicknesses in the range of 130–400 nm using
polynomial fitting. Polynomial fit was used, because it provided the best fit for the distribution of
PT
Figure 3(b). Fill factor was fit for active layer thicknesses from 130 nm to 400 nm using
polynomial fitting. [Insert]: Voc was fit using linear fitting for the same active layer thickness
RI
range.
SC
Figure 4(a). Experimentally obtained short-circuit current density (Jsc,exp) under 1000-lx and
AM1.5 G illuminations is plotted in dashed blue lines. The short-circuit current density
U
calculated (Jsc,cal) from dividing fitted IQE by Jsc,ideal is plotted in solid blue line. The Jsc,cal was
AN
highest at an active layer thickness of 320 nm for 1000 lx illumination.
Figure 4(b). Experimentally obtained power conversion efficiency (PCE (exp)) under 1000-lx
M
and AM1.5 G illumination is plotted in dashed red lines. The PCE (cal) calculated from equation
D
2 using polynomial fitted FF, linear fitted Voc, and Jsc,cal is plotted as a solid red line. The
TE
maximum PCE (cal) was obtained at an active layer thickness of 210 nm for 1000 lx illumination.
Figure 5(a). Equivalent circuit of the solar cell is provided. Jph is the photo-generated current.
EP
DIPV is the indoor photovoltaic diode, with current JD flowing through it. Parasitic resistances are
Figure 5(b). Dark J–V characteristic is provided for different active layer thicknesses.
TABLE CAPTIONS:
Table 1(a): Experimentally extracted data (averaged over 5 devices) for different active layer
26
ACCEPTED MANUSCRIPT
Table 1 (b): Experimentally extracted data (averaged over 5 devices) for different active layer
Table 2 (a): Shunt and series resistance values (averaged over 5 devices) for different active
PT
layer thicknesses under AM1.5 G illumination.
Table 2 (b): Shunt and series resistance values (averaged over 5 devices) for different active
RI
layer thicknesses under 1000 lx LED illumination.
SC
Table 3: The optimized active layer thicknesses and the respective PCEcal for different LED
illuminance values.
FIGURES:
U
AN
Figure 1(a):
M
D
TE
C EP
AC
27
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 1(b):
AN
M
D
TE
C EP
AC
28
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 2:
AN
M
130
D
120
Jsc,ideal (µA/cm )
TE
2
110
EP
90
AC
80
70
100 200 300 400
Active layer thickness (nm)
29
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 3(a):
AN
M
D
TE
C EP
AC
30
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 3(b):
AN
M
D
TE
C EP
AC
31
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 4(a):
AN
M
D
TE
C EP
AC
32
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 4(b):
AN
M
D
TE
C EP
AC
33
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 5(a):
AN
M
D
TE
C EP
AC
34
ACCEPTED MANUSCRIPT
PT
RI
U SC
Figure 5(b):
AN
M
100µ
10µ
D
1µ
TE
Log 10 |J| (J: A/cm2)
100n
EP
10n
1n
C
80 nm
100p 130 nm
AC
200 nm
10p 230 nm
280 nm
1p 400 nm
100f
-1.0 -0.5 0.0 0.5 1.0
Voltage (V)
35
ACCEPTED MANUSCRIPT
PT
RI
U SC
TABLES:
AN
Table 1 (a):
M
Thickness & Concentration Voc (mV) Jsc (mA/cm2) FF (%) PCE (%)
D
Table 1 (b):
Jsc,exp
Thickness & Concentration Voc (mV) FF (%) PCE (%)
(µA/cm2)
36
ACCEPTED MANUSCRIPT
PT
Table 2 (a):
RI
Thickness & Concentration Rsh (Ω·cm2) Rs (Ω·cm2)
SC
3
80 ± 5 (nm), (20 mg/ml) 1.3 × 10 4±1
4
130 ± 10 (nm), (30 mg/ml) 6.9 × 10 4±1
5
280 ± 10 (nm), (60 mg/ml) 7.1 × 10 6±1
6
400 ± 15 (nm), (80 mg/ml) 1.6 × 10 10±1
D
TE
Table 2 (b):
EP
4
200 ± 10 (nm), (40 mg/ml) 6.0 × 10 12±1
AC
4
230 ± 10 (nm), (50 mg/ml) 5.5 × 10 13±1
4
280 ± 10 (nm), (60 mg/ml) 4.8 × 10 17±1
4
400 ± 15 (nm), (80 mg/ml) 4.5 × 10 26±1
Table 3:
37
ACCEPTED MANUSCRIPT
PT
1000 210 5.35
2000 210 8.92
RI
U SC
AN
M
D
TE
C EP
AC
38