of 10
Copyright WILEY-VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2016.
Supporting Information
for
Adv. Optical Mater.,
DOI: 10.1002/adom.201600252
Effectively Transparent Front Contacts for Optoelectronic
Devices
Rebecca Saive
, Aleca M. Borsuk
, Hal S. Emmer
, Colton R.
Bukowsky
, John V. Lloyd
, Sisir Yalamanchili
, and
Harry A.
Atwater
*
Supplemental:
Effectively transparent front contacts for
optoelectronic devices
Rebecca Saive
1
, Aleca M. Borsuk
1
, Hal S. Emmer
1
, Colton R. Bukowsky
1
, John V. Lloyd
1
, Sisir Yalamanchili
1
,
and Harry A. Atwater
1*
1
Department of Applied Ph
ysics and Material Science,
California Institute of Technology, Pasadena, CA 91125, USA
S1: Optical simulation
s
and measurements of the wavelength and angle dependence
Figure S1
.1
:
a
) Measured wavel
ength dependent reflection and b
) angle dependent
reflection of areas on the solar cell with flat lines, triangular
cross
-
section lines and only the antireflection coating.
Angle and wavelength dependent reflection measurements were performed in a
spectrophotometer equipped with an integ
rating sphere. The spectrophotometer instrument uses chopped monochromated light from a
supercontinuum laser and a silicon photodiode detector, as described elsewhere
(M.D. Kelzenberg, PhD thesis, 2010)
. The axis of angular
rotation was aligned parallel to
the contact finger lines or triangular contacts
(compare angle
α
in Fig. S1.2)
.
In contrast to the spatially resolved
photocurrent measurements presented in the main manuscript in Fig. 2
the illumination spot size used in these measurements is large (~ 200
μm), and averages over regions with good fidelity in the fabri
cated triangular cross
-
section contact st
ructure, along with regions con
taining
imperfections.
Figure S1.2:
Ray optical (Lighttools) simulation of the wavelength and angle dependence of the transmission through a grid of free
-
standing a), c) flat
silver
lines and b), d) triangular
cross
-
section silver lines. In
a) and b) the angle was varied perpend
icular to the line (angle α on the right side), in c) and
d) the angle was varied parallel to the contact lines (angle β on the right side).
Decreased transparency in the short wavelength regime can be attributed
to losses in the silver.
S2
Optical simulation of triangular cross
-
section line patterns with different pe
riods
Figure S2
:
S
imulation of the transparency of
free
-
standing
, solid silver
triangular cross
-
section lines depending on the wavelength and
on the period of the line pattern. The simulation
was performed by
two
-
dimensional
rigorous
coupled wave analys
is (RCWA) using RSofts’s
DiffractMOD in order to account for wave
optical effects occurring particularly for smaller
periods. Periods between 3
μ
m and 20
μ
m and
wavelengths between 280 nm and 1107 nm
were simulated. The triangular lines were 2.5
μ
m wide an
d 7.0
μ
m high. Thus, e.g. a period
of 6.0
μ
m corresponds to a gap of 3.5
μ
m. It
can be seen that the transparency remains
almost 100 % for periods down to around 5
μ
m. Decreased transparency in the short
wavelength regime can be attributed to losses
in the
silver.
S3
:
Comparison of different photonic approaches to improved front contact transparency
Title
Authors
Paper
Scheme
Transparency
Equivalent
current
*
Sheet
resistance
Effectively
transparent
front contacts
for
optoelectronic
devices
Rebecca S
aive,
Aleca M. Borsuk,
Hal S. Emmer,
Colton R.
Bukowsky, John
V. Lloyd, Sisir
Yalamanchili, and
Harry A. Atwater
This
paper
99.9 %
41.2
1
mA/cm
2
4.8 Ω/sq
24.7% Record
Efficiency HIT
Solar Cell on
Thin Silicon
Wafer
Mikio Taguchi,
Ayumu Yano,
Satoshi To
hoda,
Kenta
Matsuyama, Yuya
Nakamura,
Takeshi
Nishiwaki,
Kazunori Fujita,
and Eiji
Maruyama
IEEE
Journal of
Photovolt
aics 4, 96
-
99 (2014)
~ 96.6 %
1
(only grid
fingers,
deduced from
geometry and
current)
39.5
mA/cm
2
~ 5 Ω/sq
1
Hybrid Metal
Semiconductor
Nanostructure
for Ultrahigh
Optical
Absorption and
Low Electrical
Resistance at
Optoelectronic
Interfaces
Vijay K.
Narasimhan,
Thomas M.
Hymel, Ruby A.
Lai, and Yi Cui
ACS nano
9
, 10590
-
10597
(2015)
97 %
absorption,
with 4.2 %
parasitic
absorption
within
metal
92.8 %
transmission
38.40
mA/cm
2
Below 20
Ω/sq
Cloaked
contact grids
on solar cells
by coordinate
transformation
s: designs and
prototypes
Martin F.
Schumann,
Samuel
Wiesendanger,
Jan Christoph
Goldschmidt,
Benedikt Bläsi,
Karsten Bittkau,
Ulrich W.
Paetzold,
Alexander
Sprafke, Ralf B.
Wehrspohn,
Carsten
Rockstuhl, and
Martin Wegener
Optica
2
,
850
-
853
(2015)
T
ransparency
not explicitly
reported
.
---
Not
reported
but this
geometry
should
allow for
low sheet
resistance.
Catoptric
electrodes:
transparent
metal
electrodes
using shaped
surfaces
Piet
er G. Kik
Optics
letters
39
,
5114
-
5117
(2014)
84 %
(Simulation)
34.68
mA/cm
2
Not
reported
but this
geometry
should
allow for
low sheet
resistance.
Transparent
metallic fractal
electrodes for
semiconductor
devices
Farzaneh
Afshinmanesh,
Alberto G. Curto,
Kaveh M.
Milaninia, Niek F.
van Hulst, and
Mark L.
Brongersma
Nano
letters
14
,
5068
-
5074
(2014)
No average
transparency
reported.
---
No sheet
resistance
reported.
Performance
enhancement
of metal
nanowire
transparent
conducting
electrodes by
mesoscal
e
metal wires
Po
-
Chun Hsu,
Shuang Wang,
Hui Wu, Vijay K.
Narasimhan,
Desheng Kong
Hye Ryoung Lee
and Yi Cui
Nature
communic
ations
4
(2013)
92 %
3
8.10
mA/cm
2
0.36
Ω/sq
Transparent
Conducting
Silver
Nanowire
Networks
Jorik van de
Groep, Pierpaolo
Spinelli, and
Albert Polman
Nano
Letters
12
, 3138
-
3144
(2012)
91 %
37.67
mA/cm
2
6.5 Ω/sq
*
The equivalent circuit is calculated at the example of the 24.7 % efficiency
Panasonic
HIT cell (
IEEE Journal of Photovoltaics
4
, 96
-
99
(2014)
). The grid fingers are replaced by using a contact structure with the reported transparency, the busbars ar
e kept the same.
Furthermore, a current of 1.21
mA/cm
2
is subtracted in order to ac
count for parasitic absorption within the amorphous silicon.
For our contact structure we also performed a realistic ray tracing simulation that takes the front texture of silicon in rea
l solar cells
and
losses in the amorp
hous silicon into account
.
Note
,
that
it is not clear for all structures
reported here
how integration with structured
silicon would be achieved.
1
These values are not explicitly given in the paper but deduced from the optical and electrical properties by comparison with
similar
solar c
ells.
S4
: Additional spatially resolved
laser beam induced
photocurrent measurement
The left figure shows a spatially resolved
laser beam induced
photoc
urrent measurement of an area on
the very same solar cell u
sed for
the measurement in Fig
2
. In this area an approximately 100 nm silver layer covers the flat lines as a result from the angular
evaporation.
T
he silver was evaporated from the right side so that right of the line there is a silver hill while left of the line is only
little or
no
silver.
From the profile shown on the right side, it can be seen that there is a non
-
negligible photocurrent ind
uced between the lines
although this area was covered with around 100 nm of silver. The signal goes down to zero at the position of the silver h
ills showing that
this is not a measurement artifact but that 100 nm of silver transmit a significant amount of light.
S5: Electrical properties of the line grid