Supporting Information
Failure
Modes of Protection
Layers Produced by Atomic Layer
Deposition of Amorphous TiO
2
on GaAs Anodes
Pakpoom Buabthong,
1
Zachary
P.
Ifkovits,
2
Paul A. Kempler,
2
Yikai Chen,
1
Paul
D. Nunez,
2
Bruce S. Brunschwig,
3
Kimberly M. Papadantonakis,
2
and Nathan S. Lewis
2,3 *
1
Division
of
Engineering
and
Applied
Sciences,
California
Institute
of Technology,
Pasadena,
CA 91125
2
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena,
California 91125, USA.
3
Beckman Institute and Molecular Materials
Research Center, California
Institute
of Technology,
Pasadena, California
91125, USA.
*
Corresponding author:
nslewis@caltech.edu
S1
Electronic
Supplementary
Material
(ESI)
for
Energy
&
Environmental
Science.
This
journal
is
©
The
Royal
Society
of
Chemistry
2020
Estimate of Time to Etch Through GaAs Wafer
The time
to etch through the wafer
was estimated using Eq. S1, assuming
one
pinhole
orifice with a fixed diameter and isotropic spherical etching. The background
current density
(
J
before
= 0.05
mA cm
-2
)
was subtracted from the steady-state current density after the stepwise
increase (
J
after
= 0.19 mA cm
-2
). The
electrode area
was
A
el
= 0.068 cm
2
. The thickness of
the
GaAs wafer was
r
etch
=
350 μm
(ρ=5.32
g cm
-3
,
N
GaAs
= 145 g mol
-1
, the volume etched was
)
푉
푒푡푐ℎ
=
1
2
(
4
3
휋
푟
3
푒푡푐ℎ
)
푇
=
푉
푒푡푐ℎ
휌
퐺푎퐴푠
푀
퐺푎퐴푆
푁
푆푡
푁
퐴
(
퐽
푎푓푡푒푟
‒
퐽
푏푒푓표푟푒
)퐴
푒푙
=
(
1
2
(
4
3
∗
휋
∗
(
350 휇푚
)
3
)
)
∗
(
1
10000
푐푚
휇푚
)
3
∗
(
5.3176
푔
푐푚
3
)
∗
(
1
144.65
푚표푙
퐺푎퐴푠
푔
)
∗
(
6
푚표푙
푒
‒
1
푚표푙
퐺푎퐴푠
)
∗
(
6.022
∗
10
23
푒
‒
푚표푙
푒
‒
)
∗
(
1.602
∗
10
‒
19
퐶
푒
‒
)
∗
(
1000
푚퐴
푠
1
퐶
)
∗
(
1
(
0.19
‒
0.05
)
푐푚
2
푚퐴
)
∗
(
1
0.068
푐푚
2
)
∗
(
1
3600
ℎ푟
푠
)
=
55.64 ℎ푟
(Eq. S1)
Dissolution of TiO
2
ICP-MS could
not be utilized
to monitor analytically the formation of pinholes in
the
TiO
2
layer by
measuring the Ti concentration
of the
electrolyte. Assuming a pinhole
has
approximately a
100 nm radius, and the complete formation of a pinhole through the entire
layer
creates a perfectly
cylindrical void, the void would have a
volume
of
.
0.1
2
×
휋
×
0.1
=
0.003
휇푚
3
This is
an
equivalent
volume to
cm
3
. TiO
2
has a
density of 4.23 g cm
-3
, resulting in
3.14
×
10
‒
15
this representative
pinhole having lost a total mass of
TiO
2
of
g =
ng.
1.3
×
10
‒
14
1.3
×
10
‒
5
Titanium represents 60% of
the mass of TiO
2
, so the mass of titanium from one pinhole
is
approximately
ng. In a solution of 20 mL, this
is a concentration of
ng L
-1
, well
8
×
10
‒
6
4
×
10
‒
4
S2
below the minimum detection limit of ~1 ng L
-1
. Furthermore, the sample
is diluted with nitric
acid during sample preparation, reducing the concentration by an additional order of magnitude.
Challenges in Time-Series Scanning Electron
Microscopy (SEM)
Although a time-series SEM can
provide insights
on the evolution of a new pinhole, sample
preparation and
transport
can
alter
the
surface
chemistry and
morphology of samples.
Figure
S8a
shows current density as
a function of time of p
+
-GaAs/a-TiO
2
-2000x under operation with
a
stepwise increase suggesting a new pinhole
formation at 7.5 min. Figure
S8b
shows the
corresponding SEM
taken after the increase
in current density.
The sample
was taken out,
rinsed
with deionized
water, and blow-dried with
N
2
gas to
prevent further corrosion
from electrolyte
contact.
The SEM
shows possible
film
breakage from
the preparation
step
over the corrosion
pit. Subsequently, when
the sample
was put back
to
electrochemical operation, a substantial increase in the
corrosion
current density was observed
(form 0.7
mA cm
-2
to 1.4 mA
cm
-2
). The
increase
in current density is likely due to
an increase in
the surface area from
the
torn
a-TiO
2
film and
oxidized exposed GaAs which
is more
soluble in
1.0 M KOH(aq). To accurately monitor the evolution
of the pinhole,
an operando SEM
capability is needed to avoid changing
the surface morphology and chemical state during
the
electrochemical operation.
S3
Supporting Figures
Figure S1.
Current density
vs
potential (
J-E
) behavior of p
+
-GaAs samples coated
with a-TiO
2
-
2000x (2000
ALD cycles)
but without
Ni (GaAs/a
-TiO
2
-2000x). The
scan
rate was 40 mV s
-1
.
Two types of electrochemical behaviors were observed for the GaAs/a-TiO
2
-2000x samples: (i)
no apparent corrosion current and (ii)
substantial corrosion currents.
Figure S2.
Scanning-electron
micrographs
of
(a)
a-TiO
2
/Ti foil before Au
deposition
and
(b) a-
TiO
2
/Ti after Au deposition.
S4
Figure S3.
Current density
as a function of time
for p
+
- GaAs samples coated with
a-TiO
2
(2000
ALD cycles) but without Ni (GaAs/
a-TiO
2
-2000x) at (i) 0 h and (ii) 1 h. The scan rate
was
40 mV s
-1
.
Figure S4.
Schematic of the initial geometry used in COMSOL simulation
(out of scale).
Figure S5.
Micrograph of the p
+
-GaAs/TiO
2
-2000x after electrochemical testing
S5
Figure S6.
Area of under the peaks for Ti 2p (blue
circles), O 1s
(orange
triangles), Ga
3d (green
inverted triangles), and
As 3d (red squares)
signals as a
function of sputtering time on p
+
-
GaAs/TiO
2
-400x samples
measured (a)
before
testing, (b)
1 h after testing, and (c) 12 h after
testing.
Figure S7
. Image processing steps to determine the
density of plated Au
in pinholes (a) raw
scanning-electron micrograph (b) conversion to 8-bit
image and (c)
outlines of the gold sites
counted by the
software.
Figure S8.
(a) Current density as a
function of time of p
+
-GaAs/a-TiO
2
-2000x held at 1.13 V vs.
RHE in 1.0 M KOH(aq) (b) the corresponding scanning-electron
micrograph
taken
at ~10 min.
S6