Transparent polymer solar cells employing
a layered light-trapping architecture
Rafael Betancur
1
†
, Pablo Romero-Gomez
1
†
, Alberto Martinez-Otero
1
, Xavier Elias
1
,MarcMaymo
´
1
and Jordi Martorell
1,2
*
Organic solar cells have unique properties that make them very attractive as a renewable energy source. Of particular
interest are semi-transparent cells, which have the potential to be integrated into building fac¸ades yet not completely
block light. However, making organic cells transparent limits the metal electrode thickness to a few nanometres, drastically
reducing its reflectivity and the device photon-harvesting capacity. Here, we propose and implement an ad hoc path for
light-harvesting recovery to bring the photon-to-charge conversion up to almost 80% that of its opaque counterpart. We
report semi-transparent PTB7:PC
71
BM cells that exhibit 30% visible light transmission and 5.6% power conversion
efficiency. Non-periodic photonic crystals are used to trap near-infrared and near-ultraviolet photons. By modifying the
layer structure it is possible to tune the device colour without significantly altering cell performance.
T
he unique properties of the active material used to capture
solar photons in organic photovoltaics (OPVs)—light
weight, flexibility, semi-transparency, sensitivity to low light
levels or non-direct sunlight, and solution-processing—make
OPVs one of the most attractive photovoltaic technologies for the
development of electricity production units to be integrated into
everyday life. In recent years, it has been recognized by many
authors that the intrinsic semi-transparency of the active material
found in many OPV devices
1–27
opens a route to their integration
into transparent elements such as windows in buildings and auto-
mobiles or screens in electronic equipment. In most polymer cells,
the low charge carrier mobility in donor or acceptor organic
materials prevents the use of the thick active layers that would be
needed for very efficient photon harvesting. However, such low
charge mobility and the resulting associated semi-transparency of
the active layer may turn out to be the strongest asset for OPVs to
compete in the photovoltaic production of electrical energy. In a
conventional OPV device it is not possible to make use of the
active-layer semi-transparency, because the device is capped with
a non-transparent metal layer that serves a dual purpose, as one
of the electrodes of the cell and as a mirror to reflect non-absorbed
photons in the first pass back into the device. It has not been until
recently that semi-transparent top electrodes have been fabricated
with electrical properties similar to those where electrodes are
deposited directly on the substrate
28–36
. Reharvesting infrared or
ultraviolet photons that are lost in semi-transparent devices requires
either additional changes in the cell architecture or the use of
materials with an enhanced absorption in the near-ultraviolet and
especially in the near-infrared. A Bragg reflector has been used to
increase near-infrared photon harvesting, demonstrating that the
power conversion efficiency (PCE) of OPV cells could be increased
up to 1.7% for small-molecule cells
37
and up to 2.5% for polymer
cells
38
, and could improve transparency in thin-film a-Si:H cells
39
.
An alternative approach is to reharvest red light using cholesteric
liquid crystals as wavelength-dependent reflectors
40
. Enhancing
near-infrared photon harvesting is also achievable by considering
donor polymers with an absorption band infrared-shifted close to
800 nm. With such a strategy, a rather significant breakthrough in
semi-transparent OPVs was achieved recently, with PCE . 4%
reported
41,42
. The use of high-performance polymers or tandem
cells has also been considered in semi-transparent cells
43–45
.In
this case, either efficiencies above 5% can be achieved at the
expense of a rather limited luminosity, or high transmission in
the visible is obtained at the expense of reduced near-infrared
photon harvesting.
An approach is needed that combines high-performance
polymer photovoltaic materials with a photonic configuration to
reharvest near-ultraviolet and near-infrared photons when the top
electrode is thinned down, to take the PCE of visible transparent
cells to the upper-limit efficiencies recently established based
on Schockley–Queisser theory
46
. In the scientific literature,
the single-junction opaque OPV cells that currently exhibit the
highest efficiency are cells made with bulk heterojunctions of
PTB7:PC
71
BM
47
. Although lowering the bandgap of such types of
material is a desirable objective to increase the efficiency for trans-
parent as well as non-transparent cells, attempts to do so have not
yet yielded efficiencies higher than those measured with
PTB7:PC
71
BM for direct
47,48
or inverted cells
47
. The extinction coef-
ficient of this blend (Fig. 1a) for 300–800 nm averages 0.2. With
such an extinction coefficient and a 100-nm-thick layer, approxi-
mately half the photons at any given wavelength are not absorbed
during the first pass. As indicated, when the mirror effect of the
back electrode is removed, photonic management becomes essen-
tial. The use of only a Bragg reflector to reflect the infrared and
transmit the visible would not produce an optimal result. A periodic
multilayer optimizes the interference to obtain maximum reflectiv-
ity at the wavelength satisfying the Bragg condition. In a photovol-
taic device, interference must be optimal at all wavelengths for the
part of the solar spectrum being absorbed by the active material.
In principle, broader light management could be achieved by imple-
menting a two- or three-dimensional nanostructuring of the refrac-
tive index. However, such a high level of nanostructuring may not be
desirable in solar cells, for obvious reasons.
An alternative route to the goal of broadband photonic control
using simple one-dimensional structures is to increase the degrees
of freedom relative to the periodic one-dimensional structure. In
1
ICFO-Institut de Ciencies Fotoniques, Mediterranean Technology Park, 08860 Castelldefels (Barcelona), Spain,
2
Departament de Fı
´
sica i Enginyeria Nuclear,
Universitat Polite
`
cnica de Catalunya, Terrassa, Spain;
†
These authors contributed equally to this work.
*
e-mail: jordi.martorell@icfo.es
ARTICLES
PUBLISHED ONLINE: 20 OCTOBER 2013 | DOI: 10.1038/NPHOTON.2013.276
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 1
© 2013 Macmillan Publishers Limited. All rights reserved.
the current work we report on the design and fabrication of a five-
layer photonic structure that provides optimal interference at each
wavelength to obtain the largest efficiency while at the same time
maintaining a very good transparency in most the visible wave-
lengths. Although the resulting multilayer structure lacks period-
icity, the optical control it provides is, in many ways, similar to
that provided by photonic crystals, and we will use this term to
describe it throughout. Such photonic crystals must be numerically
designed while noting that the photonic crystal and all other layers
in the device form an integral cell architecture.
We consider two types of reference device, which we term Tk-Ag
and Tn-Ag, and compare these with the device that includes the
photonic crystal. The architecture of the two reference cells is the
same except for the thickness of the top silver electrode and a pro-
tective 15-nm-thick LiF layer included in the Tn-Ag cell. The use of
such a protective layer is necessary to extend the lifetime for such
devices. Note that none of the devices used throughout the work
were encapsulated. For the Tk-Ag device the thickness of the evap-
orated top silver electrode is 100 nm, whereas for the Tn-Ag cell,
shown schematically in Fig. 1b, the thickness of the top electrode
is only 10 nm. The active material is in all cases a 90-nm-thick
bulk heterojunction of PTB7:PC
71
BM blend. As has been shown
recently, a 9.5% increase in PCE can be achieved if the calcium
layer, which is most commonly used as the hole-blocking layer for
such cells, is replaced by a highly transparent 3.5 nm layer of batho-
cuproine (BCP)
48
. As seen in Table 1 and Fig. 1c, the fabricated Tk-
Ag cell with 330 nm indium tin oxide (ITO) yielded a short-circuit
current of 14 mA cm
22
and a PCE of 7.3%, which are taken as the
reference cell parameters for comparison with the semi-transparent
cells. The experimentally measured external quantum efficiency
(EQE) for the Tk-Ag cell is shown in Fig. 2a. When the thickness
of the top electrode is reduced to 10 nm, the collection of photons
becomes less effective and the EQE is reduced at all wavelengths
(Fig. 2a). In Fig. 3a, the cell is shown to become semi-transparent
with a homogeneous transmission close to 30% in the near-
ultraviolet, visible and near-infrared ranges. Such a gain in trans-
parency is quite detrimental to cell performance, particularly in
the near-infrared, because this is precisely the spectral region
where the number of solar photons is largest. It is worth noting
that the wavelength dependence seen for the EQE of the Tn-Ag
cell does not exhibit any pronounced oscillation, as the reflectivity
reduction for the last electrode prevents the formation of standing
0.4
a
b
c
10
−10
−1,000 1,000−500 5000
0
0.3
0.2
0.1
0.0
400 600 800
Wavelength (nm)
Voltage (mV)
Extinction coecient
Current density (mA cm
−2
)
LiF
Ag
BCP
15 nm
10 nm
3.5 nm
90 nm
PTB7:PC
71
BM
ITO
120 nm
30 nm
PEDOT
LiF
Ag
BCP
15 nm
10 nm
3.5 nm
90 nm
PTB7:PC
71
BM
ITO
120 nm
30 nm
PEDOT
LiF
LiF
MoO
3
MoO
3
MoO
3
102 nm
102 nm
102 nm
102 nm
136 nm
Tk-Ag
Tn - A g
Tn - A g
PC-Tn-Ag
PC-Tn-Ag
Figure 1 | Polymer blend and device. a,PTB7:PC
71
BM extinction coefficient. b, Schematic view of the fabricated photovoltaic cells. Two stripes of IT O were
crossed by eight stripes of silver layers defining 9 mm
2
cells. The other layers between the electrodes wer e deposited to cover the entire substra te. A long
metal stripe at the back was deposited to contact the ITO. Using a mask, the photonic crystal was grown to cover, as in the schematic, half of the devices.
The layer thicknesses given for the photovoltaic part of the structure are the ones used for all semi-transpar ent devices in the current work. The thicknesses
for the photonic-crystal layers correspond to a device for which the curr ent density–v oltage curve would be similar to the one shown in c. Layer thicknesses
are not fully to scale. c, Current density–voltage curves for the Tk-Ag (black), Tn-Ag (red) and PC-Tn-Ag (green) cells fabricated with 330 nm ITO.
Table 1 | Photovoltaic parameters for the fabricated devices.
ITO J
sc
(mA cm
22
) Relative J
sc
(%) V
oc
(mV) FF (%) PCE (%) Luminosity (%)
Tk-Ag 14.0 100 724 71 7.3 0
Tn-Ag 330 nm 8.5 61 688 61 3.6 32
PC-Tn-Ag 10.7 77 716 68 5.2 29
Tk-Ag
120 nm
14.1 100 744 73 7.7 0
PC-Tn-Ag 10.9 77 733 70 5.6 28
ARTICLES
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.276
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics2
© 2013 Macmillan Publishers Limited. All rights reserved.
waves. Proper light management becomes an essential issue if it is
desirable to extract the largest current from such thin-film semi-
transparent devices.
To recover the efficiency in light harvesting but not lose transpar-
ency in the visible, ideally one would like to have a perfect mirror for
the near-infrared and near-ultraviolet, and zero reflectivity in the
visible. As indicated above, to achieve such a goal we consider
one-dimensional photonic crystals, incorporating an additional
degree of freedom by removing their inherent periodicity. The prac-
tical implementation of this typically requires the application of
numerical inverse problem-solving methods to find the optimal
design for the nanolayer structure, knowing the target solution.
For a transparent photovoltaic window there are essentially two par-
ameters that determine its performance: the efficiency of converting
light to electricity and the device visible transmission or luminosity.
The luminosity of a semi-transparent device corresponds to the
integral of the transmission weighted by the product of the
human eye photopic spectral response with illumination from the
white standard illuminant CIE-D65 (Supplementary Section S1),
normalized by the integral of the product of the photopic curve
and illuminant shown in Fig. 3a.
In our numerical calculation we sought to design a photonic
multilayer to maximize the contribution to J
sc
for wavelengths
below 400 nm and above 650 nm and where luminosity is at least
90% of that of the Tn-Ag cell. Figure 4a shows that J
sc
increases
rapidly when we increase the number of layers in the photonic
crystal, and saturates beyond five layers. Although J
sc
for the Tn-Ag
cell is 63.8% that of the opaque cell, the J
sc
calculated for the
five-layer photonic crystal (PC)-Tn-Ag cell is 76.3% that of the
opaque cell. The photonic-crystal structure we designed combines
layers of a low-refractive-index material (LiF) with layers of high-
refractive-index material (MoO
3
). The increase in J
sc
from a periodic
or Bragg structure was also calculated (Fig. 4a). For a Bragg reflector
of six layers formed combining three bilayers of LiF and MoO
3
, the
maximum J
sc
predicted corresponds to 72% that of the opaque cell.
This confirms the relevance of increasing the degrees of freedom to
reach a high level of tailoring for the EQE and, eventually, improve
the performance of the device. As shown in Fig. 2b, when
we compare the calculated EQE of the PC-Tn-Ag and Tn-Ag cells
we observe that contributions to the current from the near-infrared
as well as near-ultraviolet photons are clearly enhanced for the
PC-Tn-Ag device. For some near-infrared photons the EQE for
the PC-Tn-Ag is enhanced to such a level as to match the EQE of
the opaque cell. On the other hand, the contribution from visible
photons to the EQE remains similar to the contribution that we
obtain from the same type of photons for the Tn-Ag cell. This
results from the remarkable capacity of the photonic crystal to
trap near-infrared and near-ultraviolet light, as can be seen in the
field intensity contour plot in Fig. 4b. An enhanced field amplitude
or localization is clearly observed in the near-ultraviolet and
80
60
40
EQE (%)
20
0
80
60
40
EQE (%)
20
0
400 600
Wavelength (nm)
800
400 600
Wavelength (nm)
800
a
b
Tk-Ag
Tn - A g
PC-Tn-Ag
Figure 2 | External quantum efficiency. a, Experimentally measured EQEs
for the opaque cell (Tk-Ag, black circles), the semi-transpar ent cell where
the top electrode is a 10-nm-thick silver layer (Tn-Ag, red circles) and the
cell incorporating the photonic crystal (PC-Tn-Ag, green circles). The
trapping for the near-ultra violet is very effective and the J
sc
resulting from
photons at wa v elengths in the 300–400 nm range corresponds to 98% of
the opaque cells. For near-infrared photons in the 650–900 nm range, the
J
sc
corresponds to 91% that of opaque cells. b, Numerically calculated EQEs
for the opaque cell (black line), the semi-transparent cell where the top
electrode is a 10-nm-thick silver layer (red line) and the cell that
incorporates the photonic crystal (green line).
40
20
Transmission (%)
0
40
1.0
0.5
0.0
20
Transmission (%)
0
400 600 800
Wavelength (nm)
400 600 800
Wavelength (nm)
a
b
Tn - A g
PC-Tn-Ag
Figure 3 | Device transmission. a, Experimentally measured transmission for
the semi-transparent cell where the top electrode is a 10-nm-thick silver
layer (Tn-Ag, red circles) and for the cell that incorporates the photonic
crystal (PC-Tn-Ag, green circles). The product of the photopic curve and
CIE-65D illuminant is normalized to one (blue line). Note that this product is
relatively small in the 400–500 nm range. This is because the sensitivity of
the human eye to blue colours is rather low. However, to cause minimal
alteration to the colours of the image being observed through the PC-Tn-Ag
devices, it is important to keep the same level of transparency in that
wavelength range as in the 500–650 nm range. b, Numerically calculated
transmission for the semi-transparent cell where the top electrode is a
10-nm-thick silver layer (red line) and for the cell that incorporates the
photonic crystal (green line).
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.276
ARTICLES
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 3
© 2013 Macmillan Publishers Limited. All rights reserved.
near-infrared regions at several positions in the active layer, while in
the visible region the field intensity change is minimal at almost all
positions within this layer. The calculated transmissions (Fig. 3a) for
the Tn-Ag and PC-Tn-Ag cells are in the 30–45% range for most
visible wavelengths. However, for near-ultraviolet and near-infrared
wavelengths, the predicted transmission is below 10% for the
PC-Tn-Ag, whereas it stays above 30% for the Tn-Ag cell.
To fabricate the photonic crystal we followed a procedure (see
Methods) that imposed no degradation on the parameters of the
cell (such as fill factor (FF) and V
oc
, which are weakly dependent
on light-harvesting capacity). Indeed, as seen in Table 1, the V
oc
decreased from 724 mV to 716 mV, and the FF from 71% to 68%.
The slight decrease in both parameters may be attributed to the
silver electrode, which is only 10 nm thick. The experimentally
measured EQE for the PC-Tn-Ag cell is compared to the Tn-Ag
and Tk-Ag reference cells in Fig. 2a. In very good agreement with
the theoretical prediction, the near-ultraviolet and near-infrared
photon trapping is very effective. Indeed, the J
sc
resulting from
photons in the 300–400 nm wavelength range corresponds to 98%
of the original opaque cell, while for the near-infrared photons in
the 650–900 nm wavelength range the J
sc
corresponds to 91% that
of the original opaque cell. Experimentally, the overall increase in
J
sc
relative to the Tn-Ag cell is close to 26%, while transparency in
the visible is maintained close to 30%. The maximum efficiency
for a semi-transparent cell could be further increased using
120 nm ITO substrates as electrodes. For such a device, as summar-
ized in Table 1, we measured a PCE of 5.6% for a cell device with a
luminosity very close to 30%. The J
sc
of the transparent cell was 77%
that of the opaque device.
When looking through a thin layer of PTB7:PC
71
BM blend, one
does not perceive any significant alteration of the colour of any
image behind. Combining such fairly flat transmission (cf.
Fig. 1a) with the photonic crystal, we demonstrated that we can
broadly tune the colour of the device with a very limited PCE altera-
tion. As explained in detail in Supplementary Section S2, from the
predicted transmission curve of a given device one can predict its
colour. We fabricated four different devices: green, blue, red and col-
ourless (Fig. 5a). As reported in Supplementary Table S1, optimal
harvesting of the near-infrared and near-ultraviolet photons is
obtained with the blue and green devices (see EQE curves in
Fig. 5b). For the red device, transmission for long wavelengths
increases, resulting in less effective trapping of near-infrared
photons (cf. Fig. 5b). The EQE of the colourless device shows the
weakest wavelength dependence in the visible and near-
infrared intervals.
To summarize, in organic transparent cells, light harvesting
diminishes because the reflectivity of the top layer is reduced and
the device loses its capacity for photon trapping. We have opened
a route to effective light-harvesting recovery when thinning down
50
45
76
72
68
Relative J
sc
64
60
40
35
25
800
600
400
02040
Position in the active layer (nm)
60 80
60
−60
30
−30
0
01234567
30
Number of layers in the PC
Wavelength (nm)
Luminosity (%)
a
b
Tn - A g
PC-Tn-Ag
Figure 4 | Photonic crystal design. a, As a function of the number of layers,
numerically determined short-circuit current (black open circles) and
luminosity (red open circles) for devices incorporating the photonic crystal
and short-circuit current (black open squares) and luminosity (red open
squares) for devices incorporating the optimal periodic Bragg reflector. The
experimentally measured short-circuit current (black filled circles) and
luminosity (red filled circles) for the Tn-Ag and PC-Tn-Ag fabricated devices
are also shown. Note that the zero in the x-axis corresponds to the case
including a thin LiF protective layer. b, Contour plot of the percentage change
in the active layer field amplitude for the five-lay er PC-Tn-Ag relative to the
Tn-Ag devices. The horizontal axis indicates the position in the active layer
relative to the end of the PEDOT:PSS layer (0 nm) and the beginning of the
BCP layer (90 nm).
a
b
80
60
40
EQE (%)
20
0
400 600
Wavelength (nm)
800
Figure 5 | Device colour control. a, Four PC-Tn-Ag devices fabricated with
different relative layer thickness to tune the colour of the device. The ICFO
logo is located under the devices, and illumination is pro vided from behind
the logo. The size of each device in the image was 2.5 cm × 2.5 cm. The
purpose of this is to show the colour tuning effect pro vided by the photonic
crystal. Although the entire sequence of lay ers was deposited, no special
design was made for the extra ction of charges. The sequence and
thicknesses of all the layers from the ITO to the protective LiF layer is the
same in all four cases. All devices wer e cover ed with a five-layer photonic
crystal, where the only difference between one photonic crystal and another
is the rela tiv e thicknesses of the photonic-crystal layers. b, Experimentally
measured EQEs for four devices equivalent to those in a.
ARTICLES
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.276
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics4
© 2013 Macmillan Publishers Limited. All rights reserved.
the top electrode. For the types of device considered here, a semi-
transparent top electrode implies that light-harvesting capacity is
only 61% that of the corresponding opaque device. We demonstrate
that the photon management provided by a non-periodic one-
dimensional structure of five layers is sufficient to return the
light-harvesting capacity of the device to 77% that of the opaque
cell. When such photon trapping was applied to high-performance
bulk heterojunctions of OPV materials, we fabricated close to 30%
transparent cells with efficiencies above 5.6%. In a recent study of
the theoretical efficiency limits of transparent cells it was deter-
mined that cells with 30% luminosity have the potential to exhibit
a PCE close to 88% that of the Shockley–Queisser limit
46
.
Combining the photonic control proposed here with the use of red-
shifted absorption, low-bandgap polymers, we expect that the per-
formance of semi-transparent cells may come very close to such
theoretical predictions. Finally, the results reported, which are
essentially based on a photonic effect, have interest well beyond
the photonics community. They open the door to the design and
fabrication of new types of photovoltaic modules with great poten-
tial to be incorporated into buildings (Supplementary Section S3) as
windows, thereby leading to very good integration of electrical
power generation sources in highly populated urban areas.
Methods
Fabrication of photovoltaic cells. To fabricate the devices we used two different
kinds of ITO-patterned substrates, one with 120-nm-thick ITO and another with
330-nm-thick ITO. The thinner ITO provides a better performance in terms of
photovoltaic parameters, but with significantly larger inhomogeneity in the relative
performance of cells within the same substrate. Other details on the preparation of
the Tk-Ag device can be found elsewhere
48
.
In Supplementary Table S3, the specific architectures of the fabricated devices
are provided. Note that all devices use the same thickness for the PTB7:PC
71
BM
active layer. This indicates that the optical extinction coefficient (or absorption of the
active material) pl ays a secondary role in the final form of the EQEs (cf. Fig. 2),
which exhibit significantly different features that can be directly linked to the field
distribution imposed by the specific architecture in each case.
Design and fabrication of the photonic nanolayer structure. Given the simplicity
of the layer architecture of an OPV cell, the numerical inverse integration method we
chose here computed the outcome of all possible solutions for a given range of
parameters (which essentially include the number of layers and thickness for some
or all layers) and then numerically selected the solution that best matched the
objective or target solution (set in terms of one or several variables, including short-
circuit current, average transparency at a given wavelength range, colour or
luminosity of the devices). When the optical extinction coefficient and refractive
index for all the layers, including the substrate, are known, it is straightforward, using
a transfer matrix method
49
, to determine numerically the EQE (Fig. 2b) and
transmission (Fig. 3b). From this point, all other required variables, such as J
sc
and
luminosity, can be calculated.
The fabricated photonic-cr ystal structure comprises a non-periodic alternation
of low- and high-index-refraction layers with thicknesses of 100 nm, as shown
schematically in Fig. 1b and Supplementary Table S13. The fabrication of such a
photonic-crystal structure requires the use of methods and materials that do not
introduce damage or degradation to the electrical properties of the cell underneath.
Accordingly, the photonic multilayer structure was fabricated by thermal
evaporation of low-refractive-index (LiF) and high-refractive-index (MoO
3
)
materials. The photonic-cr ystal and semi-transparent back electrode of 10 nm of
silver were deposited using a thermal evaporation system (Mini SPECTRO, Kurt J.
Lesker Company). The deposition rate for silver was 5.5Å s
21
and was carried out in
a home-made cooled holder to decrease silver surface diffusion and therefore
prevent three-dimensional island formation by alteration of the standard nucle ation
process
33
. The silver electrode was deposited using masks made with laser beam
cutting technology, which yielded well-defined areas. The deposition rate for the
photonic-crystal materials was 1Å s
21
. Pellets/stones were used as material for
evaporation. The residual vacuum pressure was kept below 1 × 10
26
torr to prevent
contamination. The photonic crystal was deposited close to normal incidence to
prevent any optical problem related to a refractive index decrease due to an
increment in the porosity of the film. The thicknesses of all layers were monitored
using a crystal oscillator during deposition, and were later verified from transmission
curves adjusted using an electromagnetic field transmission model. To improve
electrical contact with the top thin silver layer, a further 100 nm layer of silver
was deposited on top of the part of the silver contact not covered by the
photonic crystal.
Note that the FF and V
oc
(Table 1) are significantly better for the PC-Tn-Ag cell
than for the Tn-Ag cell. This, however, should be attributed to the rapid degradation
of the Tn-Ag cell, which is less protected from the effects of oxygen or moisture
because it lacks a photonic crystal on top. The devices were not encapsulated and all
measurements were performed in air.
Measuring the performance of photovoltaic devices. The PCE of the fabricated
devices was determined from current density–voltage curve measurements obtained
under 1 sun, AM 1.5G spectrum illumination from a solar simulator (Abet
Technologies, model Sun 3000). The solar simulator illumination intensity was
monitored using a monocrystal silicon reference cell (Rera Systems) calibrated
against a National Renewable Energy Laboratory calibrated reference cell. In all
measurements of the semi-transparent cells, illumination was from the ITO side.
EQE values were measured using a QEX10 Quantum Efficiency Measurement
System (PV Measurements). For EQE measurements, the devices were illuminated
using monochromatic light from a xenon lamp passing through a monochromator.
A calibrated mono-silicon diode with known spectral response was used as a
reference. The light transmission spectra for the fabricated devices were recorded
using an ultraviolet–vis–near-infrared spectrophotometer (Lambda950, PerkinElmer).
Received 21 December 2012; accepted 10 September 2013;
published online 20 October 2013
References
1. Bailey-Salzman, R. F., Rand, B. P. & Forrest, S. R. Semitransparent organic
photovoltaic cells. Appl. Phys. Lett. 88, 233502 (2006).
2. Ng, G.-M. et al. Optical enhancement in semitransparent polymer photovoltaic
cells. Appl. Phys. Lett. 90, 103505 (2007).
3. Chen, F.-C., Wu, J.-L., Hsieh, K.-H., Chen, W.-C. & Lee, S.-W. Polymer
photovoltaic devices with highly transparent cathodes. Org. Electron. 9,
1132–1135 (2008).
4. Huang, J., Li, G. & Yang, Y. A semi-transparent plastic solar cell fabricated by a
lamination process. Adv. Mater. 20, 415–419 (2008).
5. Koeppe, R. et al. Organic solar cells with semitransparent metal back contacts for
power window applications. ChemSusChem 2, 309–313 (2009).
6. Meiss, J. et al. Ef ficient semitransparent small-molecule organic solar cells. Appl.
Phys. Lett. 95, 213306 (2009).
7. Schmidt, H. et al. Efficient semitransparent inverted organic solar cells with
indium tin oxide top electrode. Appl. Phys. Lett. 94, 243302 (2009).
8. Tanaka, S. et al. Semitransparent organic photovoltaic cell with carbon
nanotube–sheet anodes and Ga-doped ZnO cathodes. Synth. Metals 159,
2326–2328 (2009).
9. Bedeloglua, A., Demirb, A., Bozkurta, Y. & Sariciftci, N. S. Photovoltaic
properties of polymer based organic solar cells adapted for non-transparent
substrates. Renew. Energy 35, 2301–2306 (2010).
10. Ameri, T. et al. Fabrication, optical modeling, and color characterization of
semitransparent bulk-heterojunction organic solar cells in an inverted structure.
Adv. Funct. Mater. 20, 1592–1598 (2010).
11. Dong, Q. et al. All-spin-coating vacuum-free processed semi-transparent
inverted polymer solar cells with PEDOT:PSS anode and PAH-D interfacial
layer. Org. Electron. 11, 1327–1331 (2010).
12. Lee, J.-Y., Connor, S. T., Cui, Y. & Peumans, P. Semitransparent organic
photovoltaic cells with laminated top electrode. Nano Lett. 10, 1276–1279 (2010).
13. Nickel, F. et al. Cathodes comprising highly conductive
poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) for semi-transparent
polymer solar cells. Org. Electron. 11, 535–538 (2010).
14. Pandey, A. K. & Samuel, I. D. W. Photophysics of solution-processed transparent
solar cells under top and bottom illumination. IEEE J. Sel. Topics Quantum
Electron. 16, 1560–1564 (2010).
15. Zhou, Y. et al. Indium tin oxide-free and metal-free semitransparent organic
solar cells. Appl. Phys. Lett. 97, 153304 (2010).
16. Wang, X., Ng, G.-M., Ho, J.-W., Tam, H.-L. & Zhu, F. Efficient semitransparent
bulk-heterojunction organic photovoltaic cells with high-performance low
processing temperature indium–tin oxide top electrode. IEEE J. Sel. Topics
Quantum Electron. 16, 1685–1689
(2010).
17. Colsmann, A. et al. Efficient semi-transparent organic solar cells with good
transparency color perception and rendering properties. Adv. Energy Mater. 1,
599–603 (2011).
18. Lee, Y.-Y. et al. Top laminated graphene electrode in a semitransparent polymer
solar cell by simultaneous thermal annealing/releasing method. ACS Nano 5,
6564–6570 (2011).
19. Lewis, J. E., Lafalce, E., Toglia, P. & Jiang, X. Over 30% transparency large area
inverted organic solar array by spray. Solar Energy Mater. Solar Cells 95,
2816–2822 (2011).
20. Meiss, J., Holzmueller, F., Gresser, R., Leo, K. & Riede, M. Near-infrared absorbing
semitransparent organic solar cells. Appl. Phys. Lett. 99, 193307 (2011).
21. Peh, R. J., Lu, Y., Zhao, F., Lee, C.-L. K. & Kwan, W. L. Vacuum-free processed
transparent inverted organic solar cells with spray-coated PEDOT: PSS anode.
Solar Energy Mater. Solar Cells 95, 3579–3584 (2011).
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.276
ARTICLES
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 5
© 2013 Macmillan Publishers Limited. All rights reserved.