F
m
I
s
t
m
c
r
c
b
a
s
M
c


a
u
a
f
T
o
t
a
JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY 1, 013101
2009
1abrication of organic solar array for applications in
icroelectromechanical systems
Jason Lewis, Jian Zhang, and Xiaomei Jianga

Department of Physics, University of South Florida, Tampa, Florida 33620, USA

Received 11 August 2008; accepted 11 September 2008;
published online 6 November 2008
We have developed an innovative way to fabricate organic solar arrays for appli-
cation in dc power supplies for electrostatic microelectromechanical systems de-
vices. A solar array with 20 miniature cells interconnected in series was fabricated
and characterized. Photolithography was used to isolate the individual cells and
output contacts of the array, whereas the thermal-vacuum deposition is employed to
make the series connections of the array. With 1 mm2 for single cell and a total
device area of 2.2 cm2, the organic solar array based on bulk heterojunction struc-
ture of
-conjugated polymers and C60 derivative 6,6-phenyl C61 butyric acid
methyl ester produced an open-circuit voltage of 7.8 V and a short-circuit current
of 55 A under simulated air mass
AM 1.5 illumination with an intensity of
132 mW /cm2. The procedure described here has the full potential for use in future
fabrication of microarray with the size as small as 0.01 mm2. © 2009 American
Institute of Physics. DOI: 10.1063/1.2998825
. INTRODUCTION
Microelectromechanical systems
MEMS usually have their own requirements for power
upplies. It is desirable to have appropriate on-chip power source with the MEMS device, par-
icularly in cases of autonomous operations such as wireless communication, sensor network, and
icrorobotic systems. Previous solutions for such power supplies include magnetic field induced
urrent and voltage supplies,1 electrothermal microactuators based on dielectric loss heating,2
echargable lithium microbatteries,3 integrated thermopile structures,4 vibration-electric energy
onversion,5 and miniature fuel cells.6
Solar cell can also be a good option for such power sources since it is self-contained and can
e easily integrated with existing circuits of MEMS. Moreover, solar cell has the potential of
chieving the maximum size to power density ratio compared with other miniature power
ources.
7 There have been previous studies about on-chip solar cell arrays for applications in
EMS devices,7–11 and the majority of these works have been related to the silicon photovoltaic
ells.
Organic solar cells
OSC based on
-conjugated polymers e.g., poly-3-hexylthiophene
P3HT and fullerene derivatives e.g.,
6,6-phenyl C61 butyric acid methyl ester
PCBM have
ttracted attention over the past decades because they may provide a cost-effective route to wide
se of solar energy for electrical power generation.12–16 These organic semiconductors have the
dvantage of being chemically flexible for material modifications, as well as mechanically flexible
or the prospective of low-cost, large scale processing such as solution-cast on flexible substrates.
he world’s next generation of microelectronics may be dominated by “plastic electronics’” and
rganic solar cells are expected to play an important role in these future technologies.17
The photovoltaic process in OSC devices consists of four successive processes: light absorp-
ion, exciton dissociation, charge transport, and charge collection:
i Absorption of a photon
Author to whom correspondence should be addressed. Electronic mail: xjiang@cas.usf.edu.
1, 013101-1941-7012/2009/11
/013101/8/$25.00 © 2009 American Institute of Physics

ct
o
w
n
F
w
T
t
s
s
c
t
r
s
e
a
i
e
c
w
c
c
c
r
t
u
t
M
a
o
w
p
w
u
I
i
t
s
p
u
0reates an exciton
bounded electron-hole pair;
ii the exciton diffusion to a region
for instance,
he interface of two different components, where exciton dissociation
or charge separation
ccurs;
iii finally, free charges move separately toward the anode
holes and cathode
electrons,
here
iv they are collected. Several parameters determine the performance of a solar cell,
amely, the open-circuit voltage
Voc, short-circuit current
Isc, and the so-called fill factor
FF.
F is calculated by
FF =
ImpVmp
IscVoc
,
1
here Imp and Vmp are the current and voltage operating points for maximum power, respectively.
he overall power conversion efficiency  is defined as
 =
ImpVmp
Pin
=
FF · IscVoc
Pin
.
2
Although the current power conversion efficiency
about 6% of OSC is still not high enough
o make it a practical solution for large scale commercial applications as general electric power
ources, it is promising to use OSC as a high-voltage power supply. The open-circuit voltage of
ingle junction OSC
Voc=0.6–0.7 eV is close to that of the single crystal Si or thin film poly-
rystalline Si.8 For many electrostatic MEMS, it is more critical to have high-voltage output
from
ens to hundreds of volts rather than high current or energy efficiency, with an operatable current
ange usually falling between nanoamperes to microamperes. According to the design criteria of
uch on-chip solar cells,8 OSC based on
-conjugated polymers and fullerene derivatives is an
xcellent choice. First of all, isolation of the solar cell array from the MEMS substrate is easy to
chieve, since OSC can be fabricated on any substrate including plastic. This also makes the
ntegration with microsensors and microactuators relatively effortless. Second, these polymers are
fficient light absorbers
with a typical absorption coefficient several orders higher than that of the
onventional semiconductors such as Si, meaning the active layer can be as thin as 100 nm,
hich makes it simple for series interconnection to produce high voltage. Third, Voc of a single
ell of these OSCs can be easily tuned as high as 0.87 V by chemical tailoring of both
onstituents.18 Fourth, the photoactive layer can be made through any solution processable fabri-
ation methods
i.e., spin-coating, spraying, and inkjet printing without the need of photolithog-
aphy, which is mandatory with silicon-related fabrication process. These OSC can be manufac-
ured on plastic substrates, making these cells lightweight, flexible, and very cost-effective. The
sual drawbacks of OSC
e.g., lower short-circuit current and power conversion efficiency are not
he major issues for using them as on-chip dc power sources, making OSC a perfect solution for
EMS inertia transducers, such as resonator, accelerometer, gyroscope, and pressure sensors.19
To the best of our knowledge, there has been no report of organic solar arrays based on P3HT
nd PCBM as MEMS power sources. Although there have been previous studies on large area
rganic solar modules.20–23 A small
2.2 cm2 photovoltaic minimodule having 20 cells in series
as reported in this article. The anode is made by patterning indium tin oxide
ITO on glass by
hotolithography, the cathode is made by thermal evaporation through a metal shadow mask,
hich simultaneously accomplishes the series connection of all cells. The active layer material
sed in our process is a blend of P3HT and PCBM, which forms a bulk heterojunction structure.
n this report, a detailed array fabrication process and the characterization of both single cell and
nterconnected solar array are present. In the end, a brief discussion will be given about the factors
hat could affect the output voltage and overall power efficiency, as well as several tentative
13101-2 J. Renewable Sustainable Energy 1, 013101 2009
olutions for short-circuit problems within the array. Our research has focused on the design of a
rocess to ensure full isolation of series connected cells, and this process has the full potential for
se in future fabrication of a microarray with a size as small as 0.01 mm2.

Is
b
2
c
A
T
p
m
F
I
p
t
0I. FABRICATION PROCESS
The first step was the design of the solar array. The ability to align the substrate with the
hadow mask by eye in the inert environment as well as other process parameters were considered
efore a final geometric design was made for the array. Figure 1 shows such an array consisting of
0 single cells. The top panel of Fig. 1 presents details of each cell. The whole fabrication process
onsists of four steps.
. Patterning of the anode
ITO coated glass substrates
85% transmittance, 5–15  /
 were purchased from Delta
IG. 1.
a Enlarged drawing of the anode, cathode, and sandwich structure of single cell with area of 1 mm2.
b
llustration of the interdigitated organic solar cell array consisted of 20 single cells. The bottom
light purple layer is
hotolithography-defined ITO anode, the middle
red layer is spin-coated P3HT:PCBM, and the top
light blue layer is
hermal deposited cathode by shallow mask technique.
13101-3 Organic solar array as MEMS power source J. Renewable Sustainable Energy 1, 013101 2009
echnology Inc. and cut into 1 in.1 in. pieces. The patterning of ITO is done via standard
hotolithography using a custom made photomask, as shown in Fig. 1
light purple. The photo-
ask was made by printing the desired pattern on one plastic transparency and taped onto a piece

os
s
t
f
o
e
g
B
p
d
s
p
C


t
5
P
m
P
F
s
w
c
0f 5 in.5 in. glass. Positive photoresist
Shipley 1813 is spun-coat onto the ITO side of the
ubstrate at a rate of 4500 rpm for 45 s, creating a layer with thickness of about 1.5 m. The
ubstrate is then soft baked on a 90 °C hotplate for 90 s, followed by a 3 s exposure to UV light
hrough the photomask, and the substrate is then developed in Shipley MF319 for about 1 min,
ollowed by a hard bake at 150 °C for 10 min. Etching of the ITO was done in a mixed solution
f HCl and HNO3. The patterned ITO substrate then undergoes sonification cleaning in trichloro-
thylene, acetone, and isopropanol at 50 °C for 20 min each, followed by drying with N2. The
lass substrate now has the desired pattern of ITO, which acts as the anode part of the solar array.
. Creation of the shadow mask
A 1 in.1 in. piece of stainless steel was patterned following a similar photolithography
rocedure described above. Etching of the photoresist coated stainless steel was done using a
iluted ferric chloride
FeCl3 solution 25% in deionized
DI water for 2 h. The patterned
hadow mask
Fig. 1, light blue was rinsed by DI water and sonification in acetone and isopro-
anol at 50 °C for 20 min.
. Formation of the photoactive layer
The original aqueous poly
3,4ethylenedioxythiophene:poly
styrenesulfonate
PEDOT:PSS
Baytron 500 obtained from H. C. Starck was diluted and filtered three times, then filtered out
hrough a 0.45 m filter. The solution is then spun coat on top of the patterned ITO at a rate of
000 rpm for 90 s after which the substrate is then heated up to 120 °C for 100 min. P3HT and
IG. 2. The fabrication process of miniature solar cell array. Start from
1 a clean ITO on glass substrate, followed by
2
pin-coating photoresistance,
3 development of desired pattern by photolithography,
4 etching off the unwanted ITO,
5
ashing off the photoresistance,
6 spin-coating active layer
P3HT:PCBM,
7 clean off excessive material,
8 deposit
athode via shadow mask.
13101-4 J. Renewable Sustainable Energy 1, 013101 2009
CBM were purchased from American Dye Source Inc. The active layer solution is made by
ixing P3HT and PCBM with a weight ratio of 1:1 in chloroform, then spun-coat on top of the
EDOT:PSS coated substrate at a rate of 700–800 rpm for 90 s. This provides a thickness of

2
c
3
D
p
fi
c
D
1
i
s
F
v
r
o
000 nm. The excess film is then wiped off in order to allow the aluminum to make the series
onnections required for the device
Fig. 1. The sample is then allowed to dry for a minimum of
h in vacuum before thermal evaporation of the cathode.
. Deposition of the cathode
In order for the device to function as a series array, the patterned shadow mask must be
recisely aligned to the ITO substrate
see Fig. 1. With the alignment done the substrate is then
xed onto the chuck and loaded into the deposition chamber. Aluminum was chosen for the
athode due to its desirable work function
for collection of electrons and cost-effectiveness.
eposition of aluminum was done under high vacuum
10−7 torr, with a final thickness of
00 nm. Device fabrication is completed with a final annealing on a hotplate at 110 °C for 5 min
IG. 3. Upper panel: schematic of a single organic solar cell with bulk heterojunction structure. Lower panel: current-
oltage characteristics of single cell made with P3HT:PCBM mixed with weight ratio of 1:1 under simulated AM1.5G,
adiation at 132.6 mW /cm2. The active layer was spun-coat on patterned ITO substrate at 800 rpm, with a final thickness
f about 200 nm. Post-device thermal annealing at 120 °C for 5 min was done before the I-V measurements.
13101-5 Organic solar array as MEMS power source J. Renewable Sustainable Energy 1, 013101 2009
n the glove box, prior to the I -V measurements.
Figure 2 illustrates the fabrication process for the organic solar array. The active layer is
pun-coat from a chloroform solution of P3HT:PCBM blend with a weight ratio of 1:1.

It
o
p
o
p
p
l
8
F
o
f
s
0II. EXPERIMENTAL RESULTS
In order to examine the solar array fabrication procedure described above and find the opera-
ional parameters for various processes, we fabricated test OSC in a simpler geometry consisting
f four single cells, each has an active area of 4 mm2
Fig. 3 upper left panel. The upper right
anel of Fig. 3 shows the side view of each cell in bulk heterojunction structure. Preliminary
ptimization was performed in terms of spin rate and thermal annealing conditions. The best
erformed single cell was fabricated with an active layer thickness of 200 nm followed by a
ost-device thermal annealing at 120 °C for 5 min.
IG. 4.
a A digital picture of the organic solar array with 20 miniature cells in series,
b current-voltage curve of an
rganic solar array with nine functioning cells measured at simulated AM1.5G with radiation of 132.6 mW /cm2. The
abrication parameters are the same as single cell
in Fig. 3. The inset shows array Voc as a function of number of cells in
eries. An output voltage of 7.8 V was achieved with 18 cells in series.
13101-6 J. Renewable Sustainable Energy 1, 013101 2009
The current-voltage
I -V characterization of the solar cells was performed on a solar simu-
ator consisting of a xenon arc lamp
Oriel 66485 and an air mass
AM 1.5 global filter
Oriel
1094 with irradiation of 132.6 mW /cm2. No spectral mismatch with the standard solar spectrum

A
s
V
t
t
m
d
1
s
e
m
t
s
w
w
c
s
i
m
s
n
t
d
a
I
a
t
f
p
u
0M 1.5
with an intensity of 132 mW /cm2 was corrected in the I -V characterization. The best of
uch single devices has a short-circuit current density Jsc=12.7 mA /cm2, open-circuit voltage
oc=0.60 V, FF=0.43, and a power conversion efficiency of 2.45%
Fig. 3 lower panel. While
his modest efficiency certainly has room to be improved by nanomorphology manipulations,15,16
he main purpose herein is to find the right parameters for each fabrication process.
Using these appropriate parameters, organic solar cell array based on the same photoactive
aterial
P3HT and PCBM blend used above was fabricated according to the fabrication proce-
ure described in Sec. II. The interconnected series consists of 20 cells each with active area of
mm2 on a 1 in.2 ITO substrate. A picture of such an array is shown in Fig. 4
a. Figure 4
b
hows the I -V curve of the best performed array
array 1 in Table I. Though extra caution and
fforts have been made to avoid short circuits among individual cells, the alignment of shadow
ask with the ITO anodes inside the glove box turned out to be very challenging, especially when
he active layer is thin
200 nm. Not-so-perfect alignment resulted in “shadow effect,” which
meared out the contact to neighboring cells, causing unintentional lateral connection.
In this preliminary work, we also tried to increase the active layer thickness to see how it
ould help with short circuits of individual cells. Table I gives a summary of three solar arrays
ith different active layer thicknesses. It can be seen that, with thicker film, a smaller number of
ells was short circuited. The inset of Fig. 4
b plots the array Voc versus the number of cells in
eries, and a linear relation is shown; for a total of 18 minicells, the measured Voc is 7.8 V.
Although the overall device performance is less impressive, and the poor FF might be due to
ncreased lateral collection, causing the increase of series resistance
Rs of the solar array.11 The
ore important point is the capability to obtain larger Voc in terms of the application for dc power
upply. Our prefactory results demonstrate the potential to easily tune the output voltage by the
umber of cells in series. Further improvement of the array performance is ongoing to determine
he optimization of active layer thickness and nanomorphology, as well as to reduce Rs of the array
evice by means of thermal annealing and modifying the contact properties between active layer
nd the electrodes.
V. CONCLUSION
In conclusion, a miniature organic solar array was designed, fabricated, and characterized for
pplication in MEMS device power supplies. The photoactive layer was formed by spin coating a
hin film of
-conjugated polymer P3HT and fullerene derivative PCBM blend mixed in chloro-
TABLE I. Summary of device parameters for three organic solar cell arrays containing different numbers of
cells in series. The current voltage characteristics in dark and under simulated solar AM1.5 with an intensity of
132.6 mW /cm2 are present. Each cell has an active area of 1 mm2. The power conversion efficiency
 was
calculated using Eq.
2 in text.
Array
name
Active layer
thickness

nm
Number of
cells in series
Voc

V
Isc

mA
Jsc

mA /cm2 FF


%
Array 1 203 9 5.2 0.0545 0.605 0.32 0.76
Array 2 202 15 7.0 0.0245 0.163 0.17 0.15
Array 3 232 18 7.8 0.0135 0.075 0.13 0.06
13101-7 Organic solar array as MEMS power source J. Renewable Sustainable Energy 1, 013101 2009
orm. The electrodes were patterned by photolithography and thermal evaporation through a
atterned shadow mask. An output voltage of 7.8 V and short-circuit current as large as 55 A
nder simulated solar AM1.5 illumination were achieved with the small array device
2.2 cm2.

AP
N
f
1
1
1
1
1
1
1
1
1
1
2
2
2
2
0CKNOWLEDGMENTS
The authors are grateful for the financial support from USF Grant No. GFMMD03, ACS
etroleum Research Fund
PRF 47107-G10 and the U.S. Department of Army USAMRMC Grant
o. W81XWH-07-1-0708. We would also like to acknowledge Robert Tufts and Richard Everly
or their help with USF NNRC facilities and training.
1H. Matsuki, M. Yamaguchi, T. Watanabe, and K. Murakami, J. Appl. Phys. 64, 5859
1988.
2B. Rashidian and M. G. Allen, in Proceedings of the 6th IEEE Workshop on Micro Electro Mechanical Systems, Fort
Lauderdale, FL, February 1993
IEEE, Piscataway, NJ, 1993, p. 24.
3 J. B. Bates, G. R. Gruzalski, and C. F. Luck, in Proceedings of the 6th IEEE Workshop on Micro Electro Mechanical
Systems, Fort Lauderdale, FL, February 1993
IEEE, Piscataway, NJ, 1993, p. 82.
4Z. Olgun, O. Akar, H. Kulah, and T. Akin, in Proceedings of the Transducers 97, International Conference on Solid-State
Sensors and Actuators
IEEE, New York, 1997, p. 1263.
5S. Meninger, J. O. Mur-Miranda, R. Amirtharajah, A. Chandrakasan, and J. Lang, in Proceedings of the 1999 Interna-
tional Symposium on Low Power Electronics and Design
IEEE, New York, 1999, p. 48.
6A. Jansen, S. van Leeuwen, and A. Stevels, in Proceedings of the 2000 IEEE International Symposium on Electronics
and the Environment
IEEE, Piscataway, NJ, 2000, p. 155.
7P. Ortega, S. Bermejo, and L. Castañer, IEEE Trans. Compon., Packag. Manuf. Technol., Part B 24, 169
2001.
8 J. B. Lee, Z. Chen, M. G. Allen, A. Rohatgi, and R. Arya, J. Microelectromech. Syst. 4, 102
1995.
9S. Bermejo and L. Castañer, in Proceedings of the 19th European Photovoltaic Solar Energy Conference and Exhibition,
Paris, France, 2CV.2.69, 2004, p. 135.
0M. A. Kroon, R. A. C. M. M. van Swaaij, and J. W. Metselaar, in Proceedings of the 16th European Photovoltaic Solar
Energy Conference
James & James
Science Publishers Ltd., London, 2000, p. 486.
1S. Bermejo, P. Ortega, and L. Castañer, Prog. Photovoltaics 13, 617
2005.
2G. Yu, J. Gao, J. C. Hummelen, F. Wudl, and A. J. Heeger, Science 270, 1789
1995.
3S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz, and J. C. Hummelen, Appl. Phys. Lett. 78, 841

2001.
4M. M. Wienk, J. M. Kroon, W. J. H. Verhees, J. Knol, J. C. Hummelen, P. A. van Hal, and R. A. J. Janssen, Angew.
Chem., Int. Ed. 42, 3371
2003.
5W. Ma, C. Yang, X. Gong, K. Lee, and A. J. Heeger, Adv. Funct. Mater. 15, 1617
2005.
6M. Reyes-Reyes, K. Kim, and D. L. Carroll, Appl. Phys. Lett. 87, 083506
2005.
7G. Dennler and N. S. Sariciftci, Proc. IEEE 93, 1429
2005.
8F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon, and J. C. Hummelen, Org. Lett. 9,
551
2007.
9N. V. Lavrik, M. J. Sepaniak, and P. G. Datskos, Rev. Sci. Instrum. 75, 2229
2004.
0S. Yoo, W. J. Potscavage, B. Domercq, J. Kim, J. Holt, and B. Kippelena, Appl. Phys. Lett. 89, 233516
2006.
1F. C. Krebs, H. Spanggard, T. Kjær, M. Biancardo, and J. Alstrup, Mater. Sci. Eng., B 138, 106
2007.
2C. Lungenschmied, G. Dennler, H. Neugebauer, S. N. Sariciftci, M. Glatthaar, T. Meyer, and A. Meyer, Sol. Energy
Mater. Sol. Cells 91, 379
2007.
3M. Niggemann, B. Zimmermann, J. Haschke, M. Glatthaar, and A. Gombert, Thin Solid Films 516, 7181
2008.
13101-8 J. Renewable Sustainable Energy 1, 013101 2009