Solar Energy Materials & Solar Cells 98 (2012) 267–272

Contents lists available at SciVerse ScienceDirect

Solar Energy Materials & Solar Cells

journal homepage: www.elsevier.com/locate/solmat

Silicon nanowire/organic hybrid solar cell with efficiency of 8.40%

Hong-Jhang Syu a, Shu-Chia Shiu a, Ching-Fuh Lin a,b,c,n
a
Graduate Institute of Photonics and Optoelectronics, National Taiwan University, No. 1, Sector 4, Roosevelt Road, Taipei 10617, Taiwan
b
Graduate Institute of Electronics Engineering, National Taiwan University, No. 1, Sector 4, Roosevelt Road, Taipei 10617, Taiwan
c
Department of Electrical Engineering, National Taiwan University, No. 1, Sector 4, Roosevelt Road, Taipei 10617, Taiwan

a r t i c l e i n f o abstract

Article history: In this work, we investigate in great detail the silicon nanowire (SiNW)/poly (3,4-ethylenedioxythio-
Received 7 July 2011 phene):poly(styrenesulfonate) (PEDOT:PSS) hybrid solar cells (SCs). Because of the light-trapping effect,
Received in revised form SiNW/PEDOT:PSS SCs absorb more light than planar Si/PEDOT:PSS SCs. Also, the light absorption
28 October 2011
increases with the length of the SiNWs. However, the SiNW’s length is not the only factor that
Accepted 6 November 2011
influences the SCs. Thus the SiNW/PEDOT:PSS SCs with the shortest wire length of 0.37 mm have the
Available online 30 November 2011
best performance in terms of the highest power conversion efficiency of 8.40%, the largest short circuit
Keywords: current density of 24.24 mA cm  2, and open circuit voltage of 0.532 V, compared with the SCs of other
Silicon nanowire wire lengths. The reasons are two-fold. First, long SiNWs tend to aggregate at the top portion, making
PEDOT:PSS
the infiltration of PEDOT:PSS difficult, so the coverage of PEDOT:PSS on the SiNWs is not complete.
Hybrid solar cell
Second, the increase of SiNW length greatly reduces the minority-carrier lifetime. Our investigation will
help develop SiNW SCs with improved performance.
Crown Copyright & 2011 Published by Elsevier B.V. All rights reserved.

1. Introduction to form the vertical n-type SiNW(n-SiNW) arrays and then

applied techniques of low-pressure chemical vapor deposition
Silicon nanowires (SiNWs) have attracted a lot of attention (LPCVD) and rapid thermal annealing (RTA) to synthesize a p-type
recently for many applications, such as diodes [1], transistors polycrystalline silicon sheath around n-SiNWs. Nevertheless, the
[2–5], biochemical sensors [6–8], solar cells [9–16], and others. PCE only slightly increases to 0.46% because the surface defects
For solar cells, a single SiNW radial p–n junction has been are made by wet etching and later sandwiched inside the p–n
developed. The junction area can be enlarged to produce more junction. Most of the carriers are trapped by depletion defects and
carriers than a planar junction in the same scale [9]. In practical cannot be transported out of the device [14].
applications, a bunch of SiNWs are more useful than a single Compared with Garnett and Yang’s process, we do not use
SiNW. The usage of SiNW arrays can be categorized into two vacuum procedures such as VLS depositing p-type materials on
types. The first one is for anti-reflection purposes. SiNW arrays SiNWs or conventional inorganic semiconductor manufacturing
are used to replace the conventional anti-reflection coating layer process of dopant-implantation and dopant-diffusion. Instead, a
[10,11]. The second one is SiNW-array core-sheath p–n junction low-cost solution process is used to coat a p-type-like conducting
solar cells [14–16]. In particular, the second type combines polymer poly(3,4-ethylenedioxythiophene): poly(styrenesulfo-
the advantages of the first type, that is, a large junction area, a nate) (PEDOT:PSS) onto the n-SiNWs to form the core-sheath
high anti-reflection ability, and a high light-trapping effect p–n junction. Because PEDOT:PSS is in water solution, it can
[10,13,17–21]. Therefore, lots of research has focused on the infiltrate into the space between SiNWs. The defects made by
SiNW array core-sheath p–n junction solar cells [14–16,21]. etching can then be passivated in this way, so the carriers will
Previously, Tsakalakos et al. used catalytic chemical vapor deposi- recombine less at the junction. In addition, PEDOT:PSS is a hole
tion (CVD) with vapor–liquid–solid (VLS) method to produce transport layer (HTL) that can deliver holes generated in the
SiNW array solar cells. However, the power conversion efficiency n-SiNWs and the n-type Si (n-Si) substrate out of the devices.
(PCE) is only 0.1% because the random orientation of nanowires Hence the electron-hole pairs can be separated. Moreover, vertical
grown by VLS method hinders the carrier’s transportation [12]. n-SiNW arrays can assist with photon trapping and decrease
Garnett and Yang used a metal-assisted chemical etching method photon reflection to improve photon absorption and carrier
generation [13,17–20]. The core-sheath junction also enlarge the
n
junction areas to improve the probability of carrier separation.
Corresponding author at: Institute of Photonics and Optoelectronics, National
Both the wet-etching step to form SiNWs and the spin-coating for
Taiwan University, Taipei, 10617 Taiwan. Tel.: þ886 2 33663540;
fax: þ 886 2 23642603. PEDOT:PSS are solution processes, so vacuum facilities are not
E-mail address: cfl[email protected] (C.-F. Lin). used, providing the benefits of low energy consumption, low cost,

0927-0248/$ - see front matter Crown Copyright & 2011 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.solmat.2011.11.003
268 H.-J. Syu et al. / Solar Energy Materials & Solar Cells 98 (2012) 267–272

and easy production. Here the effect of SiNWs’ length on the The reactions occur repeatedly. Because of the regular crystal
device’s performance and the physical reasons are investigated. orientation of crystalline Si, reaction rate is much faster in the
Also, the comparison of our devices with other core-sheath p–n direction of [100] [23]. Therefore, SiNW arrays are formed on the
junction solar cells will be discussed. The results reveal that the wafer surface with their direction normal to the surface (100). In
device fabricated by SiNW arrays lower than 0.5 mm exhibits the addition, before etching, the wafer surface was treated in a
best performance with a PCE of 8.40%, short circuit current particular condition to obtain a large density of SiNWs [24].
density (Jsc) of 24.24 mA cm  2, and open circuit voltage (Voc) of Etching time was controlled at 3, 10, and 20 min to form SiNWs
0.532 V. with lengths of 0.37, 2.15, and 5.59 mm. The nanowire length is
linearly proportional to the etching time. After etching, we soaked
the SiNWs with Si substrate in the concentrated nitric acid and
2. Experimental diluted HF, successively, to eliminate residual Ag and SiO2.
After the SiNWs were fabricated, Ti and Ag were deposited on the
SiNW/PEDOT:PSS solar cells with different nanowire lengths back of SiNWs to be a cathode contact using an e-gun evaporator.
were fabricated by the following steps. Single-side polished Before combining the SiNWs and PEDOT:PSS, the SiNWs were
and phosphor-doped crystalline Si (100) wafers with resistivity placed in the environment with humidity of 60% at room tempera-
of 1–10 O cm were used. N-type SiNW arrays were produced by ture for 5 h to form a thin oxide layer that can improve the
metal assisted chemical wet etching [22]. The etchant contained hydrophilicity of the Si surface, so the PEDOT:PSS can easily adhere
silver ions (Ag þ , 0.023 M) and fluoric acid (HF, 5.6 M). As the n-Si to SiNWs. Although silicon dioxide is an insulator, carriers can still
substrate was immersed into the etching solution, Ag þ and Si had penetrate due to the quantum tunneling effect as long as it is very
a redox reaction. Ag þ was reduced to Ag and deposited on the Si thin. According to the prior reports, devices with p–i–n structure
surface to form nanoparticles. At the same time, Si near the Ag perform better than those with a regular p–n junction [9]. The
nanoparticles was oxidized and became silicon dioxide (SiO2), PEDOT:PSS solution (CleviosTM P, HC V4 Stark GmbH, Leverkusen,
located between Ag and Si. SiO2 was etched by HF right away. Germany) was not directly deposited on the Si substrate with
Because the electronegativity of Ag (1.9) is greater than Si (1.8), SiNWs. Instead, it was first spin-coated on the ITO/glass (Indium
the electrons of Si near Ag nanoparticles distribute towards Ag. tin oxide/glass). Then the SiNWs were stuck into PEDOT:PSS on the
Thus, the Ag surface has partially negative charges to attract Ag þ . ITO/glass. Because the SiNWs already had good hydrophilicity,
In the etching solution, the silver ions gather together and attract PEDOT:PSS easily filled the space between each SiNWs to form the
electrons, so Ag nanoparticles form. Meanwhile, Si underneath core-sheath heterojunction due to the capillarity. Finally, the devices
the Ag is oxidized into SiO2, which was etched by HF immediately. were annealed in the condition of 140 1C for 10 min on the hot plate
The chemical reaction can be referred to in Eqs. (1)–(3). in nitrogen ambient. The SiNW/PEDOT:PSS solar cells were formed.
A schematic of the device can be referred seen to Fig. 1.
Ag þ þ e -Ag ð1Þ
Before the device was tested, the cell area was delimited by
measuring the substrate area because nanowires exist on the entire
Si þ2H2 O-SiO2 þ4H þ þ 4e ð2Þ silicon substrate, which was used for the solar cell. The cell
performance and J–V curves were measured using a Keithley 2400
SiO2 þ 4HF-H2 SiF6 þ2H2 O ð3Þ
in the condition of darkness and 1 sun AM 1.5 G 100 mW cm  2
illuminations, which was simulated by a solar simulator, SUN 2000,
Abet Technologies, Inc., and calibrated by a NREL certified mono-
crystalline silicon reference cell (PVM-236). Incident photon conver-
sion efficiency (IPCE) was measured by QE-R-3015, Enli Technology
Co., Ltd. The lamp was calibrated by a mono-crystalline silicon
reference cell provided by Newport Corporation. All of light incident
on the cells through glass, ITO, PEDOT:PSS, and SiNWs, successively.
The ultraviolet–visible–infrared (UV–vis–IR) spectrophotometer
(Lambda 35, Perkin Elmer Inc.) and an integrated sphere were used
to evaluate the reflectance of the SiNWs. Also, field emission
scanning electron microscopy (SEM) (ERA-8800FE, Elionix Inc.) was
used to closely examine the SiNWs and PEDOT:PSS-covered SiNWs.

3. Results

3.1. Cell performance

Fig. 1. Lateral schematic diagram of SiNW/PEDOT:PSS solar cells with light Table 1 and Fig. 2 show the device parameters, including the
illuminating direction. cell areas, and current density-to-voltage (J–V) curves under 1 sun

Table 1
Solar cells performance of each SiNW devices and planar Si devices.

Jsc (mA cm  2) Voc (V) PCE (%) FF (%) Rs (O cm2) Rsh (O cm2) Cell area (mm2)

Planar Si/PEDOT:PSS 11.14 0.511 2.12 37.24 25.2 121 7.83

0.37-mm-SiNW/PEDOT:PSS 24.24 0.532 8.40 65.14 2.95 461 7.42
2.15-mm-SiNW/PEDOT:PSS 20.06 0.474 6.00 63.10 3.97 885 5.94
5.59-mm-SiNW/PEDOT:PSS 13.06 0.435 3.76 66.18 4.25 694 9.50
 H.-J. Syu et al. / Solar Energy Materials & Solar Cells 98 (2012) 267–272 269

Fig. 2. Current density-to-voltage curves of each SiNW devices and planar Si devices under the illuminance condition of 1 sun AM 1.5 G. (Left) Curves around fourth
quarter. (Right) Curves in the range of  2 V to 2 V.

Fig. 3. Incident photon conversion efficiencies. (Left) External quantum efficiencies. (Right) Internal quantum efficiencies.

AM 1.5 G illumination. The 0.37 mm-SiNW/PEDOT:PSS solar cell

has the highest PCE of 8.40%, highest Jsc of 24.24 mA cm  2, and
highest Voc of 0.532 V. When the SiNWs’ length extends to
5.59 mm, PCE reduces from 8.40% to 3.76%, Jsc decays from
24.24 mA cm  2 to 13.06 mA cm  2, Voc decreases from 0.532 V
to 0.435 V, and Rs increases from 2.95 O cm2 to 4.25 O cm2. Only
fill factor (FF) stays in the range between 63% and 67%.
The planar Si/PEDOT:PSS solar cell was also fabricated for
comparison and is shown in Table 1 and Fig. 2. The native oxide
had already been removed by the buffered HF (BHF) before the
planar n-Si wafers were attached to the PEDOT:PSS coated
ITO/glass. The PCE is 2.12% and Jsc is 11.14 mA cm  2. Both of
them performed much worse than the other devices with SiNWs.
In addition, the FF of 37.24% is just a half, and Rs of 25.2 O cm2 is
Fig. 4. Reflectance of each SiNWs in the wavelength ranging from 400 to 1100 nm.
about 5.93 times to 8.54 times of Rs for nanowire devices.
Although Rsh of the planar device is the worst, the corresponding
Voc of 0.511 V is not the worst. It is near the Voc of the solar cell wavelength of 580 nm. The peaks of both EQE and IQE are red-
with a 0.37 mm wire length. All of the device parameters were shifted when the SiNWs’ length increases.
derived from the J–V curves.
3.3. Silicon nanowire arrays reflectance

3.2. Incident photon conversion efficiency (IPCE) Because the device with long SiNWs has performed worse than
the one with short SiNWs, it is worth examining the reasons. At
To further understand the device’s characteristics, the incident first, the surface reflection was investigated. Fig. 4 shows the
photon conversion efficiency (IPCE), including external quantum reflectance versus the wavelength for Si wafers with SiNWs of
efficiency (EQE) and internal quantum efficiency (IQE), was different lengths. The reflectance of a planar wafer is also
measured and shown in Fig. 3. IQE is obtained from EQE divided measured and plotted for comparison. In the wavelength range
by the transmittance of ITO/glass. Both EQE and IQE reveal that of 400 nm to 1000 nm, SiNW arrays reduce the reflectance by
photons can be absorbed and transformed to a current in a wide about 26% to 65%. The reflectance in the wavelength ranging from
wavelength range from 400 nm to 1000 nm. The 0.37 mm-SiNW/ 400 nm to 1000 nm is under 10%, 4%, and 3% for wire lengths of
PEDOT:PSS solar cell has the best EQE and IQE, compared with 0.37 mm, 2.15 mm, and 5.59 mm, respectively. It is not surprising
solar cells with different lengths of SiNWs. This is consistent with that the reflectance of Si wafers with SiNWs is less than that of
the previously measured PCE. The highest EQE of 63% occurs at a the planar Si wafer. In addition, the reflectance decreases as the
wavelength of 510 nm, while the highest IQE is 76%, occurring at a wire length increases. The reason can be easily explained by the
270 H.-J. Syu et al. / Solar Energy Materials & Solar Cells 98 (2012) 267–272

Fig. 5. (a–c) 451 tilted side-view of SEM images of SiNWs covered by PEDOT:PSS, with wire length of 0.37 mm, 2.15 mm, and 5.59 mm, respectively. All of them are
as-detached from ITO/glass. (d–f) Top-view of SiNW arrays with wire length of 0.37 mm, 2.15 m, and 5.59 mm, respectively.

enhanced light-trapping effect caused by the increasing wire The first one is based on the relation between Voc and Rsh, shown
length. However, this cannot explain the better performance of in Eq. (4)
the device with a shorter wire length than that with a longer     
nkT Jph V oc
SiNWs. The reason needs to be further investigated and will be V oc ¼ ln  þ1 ð4Þ
q J0 J 0 Rsh
discussed later. It is also noted that the reflectance increases pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
evidently for the wavelength approaching 1100 nm. This is Jph ¼ qgðxÞðL
wherepffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi p þ Ln þ WÞ and J 0 ¼ qni 2 ½ð1=Na Þ Dn =tn0 þ
different from the result reported by Huang et al. [18] and will ð1=N d Þ Dp =tp0 . n is ideality factor; k is Boltzmann constant; T
be discussed later, as well. is operation temperature; q is an electron charge; Jph is photon
generated current density; J0 is reverse saturation current den-
3.4. Scanning electron microscopy images of PEDOT:PSS covered sity; g(x) is generation rate of electron-hole pairs; Lp and Ln are
SiNWs minority-carrier diffusion lengths in n-type region and p-type
region, respectively; W is carrier diffusion length in depletion
To investigate the mechanisms that influence the device region; ni is intrinsic concentration; Na and Nd are dopant
performance, SEM images of the SiNW arrays were taken, as concentrations in p-type region and n-type region, respectively;
shown in Fig. 5. The upper row of Fig. 5 (Fig. 5(a–c)) shows SiNW Dp and Dn are diffusion coefficients of holes and electrons,
arrays coated with PEDOT:PSS, while the lower row (Fig. 5(d–f)) respectively; tp0 and tn0 are minority-carrier lifetimes of holes
shows the top view of SiNW arrays without PEDOT:PSS. Figures and electrons, respectively. Therefore Voc varies not only with Rsh,
from left to right column correspond to the SiNWs with the but also with n, Jph, and J0. As a result, the worst Rsh does not
lengths of 0.37 mm, 2.15 mm, and 5.59 mm, respectively. correspond to the worst Voc because the planar cell and nanowire
Fig. 5(a) shows that 0.37 mm-SiNW arrays can be covered by cell have different n, Jph, and J0, which result from different Lp,Dp,
PEDOT:PSS thoroughly, while 2.15 mm-SiNW arrays shown in g(x), W, tp0. The other reason is because of minority carrier
Fig. 5(b) are only covered at the top portion and the places near lifetime. From the previous report [24], planar Si has minority
it. In contrast, the 5.59 mm-SiNW arrays shown in Fig. 5(c) are carrier lifetime about twice of SiNWs less than 1 mm. High
covered just by a very thin PEDOT:PSS. In addition, long SiNWs minority carrier lifetime diminishes the probability of carrier
tend to form bundles at the top portion and diminish the gap recombination. Therefore planar device has high Voc, though high
between nanowires, as evidenced by the comparison among reflectance reduces the photo current.
Fig. 5(d–f). On the other hand, it is also expected that the long SiNWs will
have more light-trapping effect than the short SiNWs. The
reflectance measurement does confirm this speculation because
4. Discussion the reflectance decreases with the wire’s length, as shown in
Fig. 4. However, this is not consistent with the devices’ perfor-
The performance of devices with SiNWs is generally better mance because the device with the shortest wire length has much
than that of the device with only a planar surface. The reason is better performance than the one with the longest length. As a
straightforward. Because the SiNWs reduce the light reflection, matter of fact, Jsc, Voc, PCE, Rs, and IPCE become worse with the
these devices absorb more light. For example, the Jsc of the wire’s length. Only the FF remains approximately at the same
0.37 mm-SiNW/PEDOT:PSS solar cell is 24.24 mA cm  2, which is range. The reason is two-fold, as discussed below.
more than twice of the Jsc (11.14 mA cm  2) of the planar device. First, the coverage of PEDOT:PSS on the SiNWs is not thorough
Also, SiNWs enlarge the junction area about 7.00 to 9.35 times of as the wire is long. According to Fig. 5, long SiNWs tend to form
the area of planar cell. From Fig. 5(d), the widths of nanostruc- bundles at the top portion. This phenomenon had also been
tures are ranging from 20 nm to 50 nm, and the interspaces observed before [16]. As the top portion aggregates to become
between nanostructures have an average value around 68 nm. bundles, the gap between SiNWs shrinks, so PEDOT:PSS gels can
Because contact resistance is inversely proportional to the area hardly get in. In addition, the viscosity of PEDOT:PSS used here is
for charges passing through, SiNW arrays reduce Rs from the value CleviosTM P HC V4. Its viscosity ranges from 100 to 350 mPa s,
of planar device of 25.2 O cm2 to 2.95 O cm2 of 0.37 mm-SiNW which is much higher than that of CleviosTM P used previously by
device. In addition, the Voc of planar cell is not the worst and near Shiu et al., ranging from 60 to 100 mPa s [16]. The larger viscosity
the value of the 0.37 mm-SiNW device. It is owing to two reasons. makes it even harder for the PEDOT:PSS to infiltrate into the space
 H.-J. Syu et al. / Solar Energy Materials & Solar Cells 98 (2012) 267–272 271

between long SiNWs to form a thorough junction, leading to the forth of light. Each time the light bounces the interface between
decrease of the junction surface. This situation also makes Rs to the air and the SiNW, some portion of the light is absorbed, so the
increase with nanowire length. Second, according to the previous multiple bouncing of light between SiNWs provides a greater
report by Shiu et al. [24], the minority carrier lifetime depends on probability of light absorption. However, as the wavelength
the length of the SiNWs. Because longer SiNWs have more surface approaches 1100 nm, the indirect band gap of Si reduces light
area, they have more surface states. As a result, they have a absorption, although the light still experiences multiple bouncing.
shorter minority carrier lifetime and carrier recombination hap- When the photon energy is less than the silicon band gap of about
pens more easily [24]. Therefore Jsc, Voc,and PCE decay when 1.12 eV, corresponding to the light wavelength of longer than
SiNWs are long. The minority carrier lifetime may also explain the 1100 nm, light cannot be absorbed, so a significant amount of
result reported by Garnett and Yang [14] for which Jsc is light is bounced out of the SiNW arrays, leading to the increased
4.28 mA cm  2, Voc is 0.29 V, PCE is 0.46%, and FF is 33%. The reflection. In comparison, the nanowires in Huang et al.’s work
much lower Jsc and PCE, compared to the results shown here, is have a gradual change in the size of the cross-section, which
probably because the SiNWs presented by Garnett and Yang have results in a gradual change of the reflective index and the
a long length of 18 mm. On the other hand, though Rs increases nanowires function more like the anti-reflection coating.
with the nanowire length, the FF still approximately remains in
the same range. According to the previous report [25], in the
condition of finite Rs and Rsh, FF can be expressed as Eq. (5) 5. Conclusions
 
ðvoc þ 0:7Þ FF 0 ð1r s Þ
FF ¼ FF 0 ð1r s Þ 1 ð5Þ In conclusion, SiNWs/PEDOT:PSS solar cells synthesized by
voc r sh
SiNWs less than 0.5 mm through solution process have the highest
where FF 0 ¼ ððvoc lnðvoc þ0:72ÞÞ=ðvoc þ 1ÞÞ, voc ¼ ðV oc Þ=ðnkT=qÞ, PCE of 8.40%, Jsc of 24.24 mA cm  2, and Voc of 0.532 V, compared
r s ¼ ðRs Þ=ðV oc =J sc Þ and r sh ¼ ðRsh Þ=ðV oc =Jsc Þ. n is ideality factor; k is with SiNWs/PEDOT:PSS solar cells of other wire lengths. Although
Boltzmann constant; T is operation temperature; q is one electron shorter SiNWs have higher reflectance, they do not aggregate and
charge. FF is not only determined by Rs, but also determined by n, form bundles, letting PEDOT:PSS easily infiltrate into the space
Voc, Jsc, and Rsh. Therefore, the variation of FF with the wire length between SiNWs and thoroughly form the SiNW/PEDOT:PSS junc-
is not the same as Rs. tion. On the other hand, the long SiNWs’ distribution tend to
It is also worth investigating the differences between the aggregate and so when the PEDOT:PSS with a large viscosity is
similar work reported previously by Shiu et al. [16] and the current used the SiNW/PEDOT:PSS junction cannot form thoroughly. In
one. In Shiu et al.’s work, the 2.78 mm-SiNW/PEDOT:PSS solar cell addition, the long SiNWs have a smaller minority carrier lifetime,
has Jsc of 19.28 mA cm  2, Voc of 0.47 V, PCE of 5.09%, and FF of 61%, giving rise to a higher recombination rate, so few carriers can be
while the planar Si/PEDOT:PSS solar cell has Jsc of 1.27 mA cm  2, collected externally. Our investigation suggests that long SiNWs
Voc of 0.34 V, PCE of 0.08%, and FF of 18%. The 0.37 mm-SiNW/ have better light absorption, but coverage of PEDOT:PSS over the
PEDOT:PSS solar cell here has better performance than the Shiu SiNWs needs to be improved and proper passivation on the
et al.’s nanowire devices. The reason is the same as the discussion surface of SiNWs is necessary for better performance of SiNWs/
before. Although the 0.37 mm-SiNW arrays have higher reflectance PEDOT:PSS solar cells.
than the 2.78 mm-SiNW arrays, both of them are thoroughly
covered by PEDOT:PSS [16]. Because the PEDOT:PSS used here is
CleviosTM P HC V4 with a conductivity of 200 S cm  1, which is Acknowledgment
much higher than that of the PEDOT:PSS (CleviosTM P with
conductivity of 10 S cm  1) used by Shiu et al. [16]. Therefore, the We acknowledge the support of the National Science Council,
0.37 mm-SiNW/PEDOT:PSS solar cell has a better performance. Taiwan under contract number of NSC 98–3114-E-002-001.
Similarly, the higher conductivity of CleviosTM P HC V4 also makes
the planar Si/PEDOT:PSS solar cells of this work perform better References
than the one reported in Shiu et al. [16].
The 2.15 mm-SiNW/PEDOT:PSS solar cell of this work can be
[1] K. Peng, Z. Huang, J. Zhu, Fabrication of large-area silicon nanowire p–n
closely compared with the 2.78 mm-SiNW/PEDOT:PSS solar cell by junction diode arrays, Advanced Materials 16 (2004) 73–76.
Shiu et al. [16]. Both of them have the SiNWs being created using [2] Y. Cui, Z. Zhong, D. Wang, W.U. Wang, C.M. Lieber, High performance silicon
the same etching concentration. Both of them have similar perfor- nanowire field effect transistors, Nano Letters 3 (2003) 149–152.
[3] J. Goldberger, A.I. Hochbaum, R. Fan, P. Yang, Silicon vertically integrated
mance, only the PCE of the 2.15 mm-SiNW device is slightly larger nanowire field effect transistors, Nano Letters 6 (2006) 973–977.
than that of 2.78 mm-SiNW device. The reason can also be [4] N. Singh, A. Agarwal, L.K. Bera, T.Y. Liow, R. Yang, S.C. Rustagi, C.H. Tung,
explained by the higher conductivity of the PEDOT:PSS used here. R. Kumar, G.Q. Lo, N. Balasubramanian, D.-L. Kwong, High-performance fully
depleted silicon nanowire (diameter r 5 nm) gate-all-around cmos devices,
Because the CleviosTM P HC V4 does not fill in the space between IEEE Electron Device Letters 27 (2006) 383–386.
the SiNWs of the 2.15 mm-SiNW array, the high conductivity [5] V. Schmidt, H. Riel, S. Senz, S. Karg, W. Riess, U. Gc-sele, Realization of a silicon
cannot give the 2.15 mm-SiNW/PEDOT:PSS solar cell the full benefit nanowire vertical surround-gate field-effect transistor, Small 2 (2006) 85–88.
[6] F. Patolsky, G. Zheng, C.M. Lieber, Fabrication of silicon nanowire devices for
and thus only leads to a slight improvement of performance. In ultrasensitive, label-free, real-time detection of biological and chemical
other words, if the PEDOT:PSS can infiltrate thoroughly into the species, Nature Protocols 1 (2006) 1711–1724.
space between SiNWs, better performance should be expected. [7] Z. Gao, A. Agarwal, A.D. Trigg, N. Singh, C. Fang, C.-H. Tung, Y. Fan,
K.D. Buddharaju, J. Kong, Silicon nanowire arrays for label-free detection of
It is also noted that the reflectance measured here differs from dna, Analytical Chemistry 79 (2007) 3291–3297.
that reported by Y.-H. Huang et al. [18]. Fig. 4 shows that the [8] M.C. McAlpine, H. Ahmad, D. Wang, J.R. Heath, Highly ordered nanowire
reflectance of SiNW arrays increases rapidly for wavelength arrays on plastic substrates for ultrasensitive flexible chemical sensors,
Nature Materials 6 (2007) 379–384.
approaching 1100 nm. However, the report by Huang et al.
[9] B. Tian, X. Zheng, T.J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang, C.M. Lieber,
reveals that the reflectance is low even if wavelength is near Coaxial silicon nanowires as solar cells and nanoelectronic power sources,
1100 nm. The reason is because the cross-section of our SiNWs Nature 449 (2007) 885–890.
does not vary obviously from the wire tip to the wire bottom. [10] K. Peng, Y. Xu, Y. Wu, Y. Yan, S.-T. Lee, J. Zhu, Aligned single-crystalline Si
nanowire arrays for photovoltaic applications, Small 1 (2005) 1062–1067.
Thus, there is no gradual change of the reflective index. The [11] T. Stelzner, M. Pietsch, G. Andrä, F. Falk, E. Ose, S. Christiansen, Silicon
reduction of reflection is mainly caused by bouncing back and nanowire-based solar cells, Nanotechnology 12 (2008) 295203-1–295203-4.
272 H.-J. Syu et al. / Solar Energy Materials & Solar Cells 98 (2012) 267–272

[12] L. Tsakalakos, J. Balch, J. Fronheiser, B. Korevaar, O. Sulima, J. Rand, [19] L. Tsakalakos, J. Balch, J. Fronheiser, M.-Y. Shih, S.F. LeBoeuf,
Silicon nanowire solar cells, Applied Physics Letters 91 (2007) 233117- M. Pietrzykowski, P.J. Codella, B.A. Korevaar, O. Sulima, J. Rand, A.D. Kumar,
1–233117-3. U. Rapol, Strong broadband optical absorption in silicon nanowire films,
[13] V. Sivakov, G. Andra, A. Gawlik, A. Berger, J. Plentz, F. Falk, S.H. Christiansen, Journal of Nanophotonics 1 (2007) 013552-1–013552-10.
Silicon nanowire-based solar cells on glass: synthesis, optical properties, and [20] O.L. Muskens, J.G. Rivas, R.E. Algra, E.P.A.M. Bakkers, A. Lagendijk, Design of
cell parameters, Nano Letters 9 (2009) 1549–1554. light scattering in nanowire materials for photovoltaic applications, Nano
[14] E. Garnett, P. Yang, Silicon nanowire radial p–n junction solar cells, Journal of Letters 8 (2008) 2638–2642.
the American Chemical Society 130 (2008) 9224–9225. [21] B.M. Kayes, H.A. Atwater, N.S. Lewis, Comparison of the device physics
[15] E. Garnett, P. Yang, Light trapping in silicon nanowire solar cells, Nano Letters principles of planar and radial p-n junction nanorod solar cells, Journal of
10 (2010) 1082–1087. Applied Physics 97 (2005) 114302-1–114302-11.
[16] S.-C. Shiu, J.-J. Chao, S.-C. Hung, C.-L. Yeh, C.-F. Lin, Morphology dependence of [22] K. Peng, J. Hu, Y. Yan, Y. Wu, H. Fang, Y. Xu, S. Lee, J. Zhu, Fabrication of single-
crystalline silicon nanowires by scratching a silicon surface with catalytic
silicon nanowire/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) het-
metal particles, Advanced Functional Materials 16 (2006) 387–394.
erojunction solar cells, Chemistry of Materials 22 (2010) 3108–3113.
[23] C.-Y. Chen, C.-S. Wu, C.-J. Chou, T.-J. Yen, Morphological control of single-
[17] L. Hu, G. Chen, Analysis of optical absorption in silicon nanowire arrays for
crystalline silicon nanowire arrays near room temperature, Advanced Mate-
photovoltaic applications, Nano Letters 7 (2007) 3249–3252.
rials 20 (2008) 3811–3815.
[18] Y.-F. Huang, S. Chattopadhyay, Y.-J. Jen, C.-Y. Peng, T.A. Liu, Y.-K. Hsu, C.-
[24] S.-C. Shiu, S.-B. Lin, S.-C. Hung, C.-F. Lin, Influence of pre-surface treatment on
L. Pan, H.-C. Lo, C.-H. Hsu, Y.-H. Chang, C.-S. Lee, K.-H. Chen, L.-C. Chen, the morphology of silicon nanowires fabricated by metal-assisted etching,
Improved broadband and quasi-omnidirectional anti-reflection properties Applied Surface Science 257 (2011) 1829–1834.
with biomimetic silicon nanostructures, Nature Nanotechnology 2 (2007) [25] M.A. Green, Solar cells operating principles, technology and system applica-
770–774. tions, Prentice-Hall, New Jersey, 1982.