Proceedings of PowerMEMS 2008+ microEMS 2008, Sendai, Japan, November 9-12, (2008)
A SUPER INK JET PRINTED ZINC-SILVER 3D MICROBATTERY
Christine C. Ho1, Kazuhiro Murata2, Daniel A. Steingart3, James W. Evans1, Paul K. Wright4
1
Department of Materials Science and Engineering, University of California, Berkeley, USA
2
Nanotechnology Research Institute, National Institute of Advanced Industrial Science and Technology,
Tsukuba, Japan
3
Department of Chemical Engineering, City College of New York, USA
4
Department of Mechanical Engineering, University of California, Berkeley, USA
A novel super ink jet printing (SIJP) system was used to fabricate 3D zinc-silver microbatteries directly on a
substrate. The SIJP provides a simple and flexible method to deposit interesting 2D and 3D structures of varying
morphologies without the waste and large energy inputs typical of standard microfabrication technologies. The
system deposited symmetric silver electrodes on glass substrates, and the battery self assembles during the first
charge. Using an aqueous electrolyte solution of KOH with dissolved ZnO, the SIJP printed structures showed
similar electrochemical behavior to batteries made of foil electrodes. For a sparse array of pillars (~2.5% footprint
area of each electrode pad occupied by pillars), a capacity increase of 60% for a given area was measured in
comparison to a cell with thin film electrodes. The cycle performance of the printed structures was poor and likely
due to the stress induced on the negative electrode upon initial zinc electrodeposition during the first charge cycle.
Keywords: 3D microbattery, Super ink jet printing, Zinc-Silver, Direct write technology
1. INTRODUCTION
The benefit of 3D architectures for laterally
constrained microbatteries (for devices with limited
footprint areas) is that increased electrode volume, and
therefore energy density, is achieved through
increasing the electrode lengths without degrading the
power density because short transport distances
between alternating electrodes are maintained [1].
Several 3D microbatteries and novel high aspect ratio
architectures have been demonstrated using a range of
methods, most of which entail multi-step processing to
pattern and/or remove templates backfilled with
electrode material [2]. As an alternative, direct ink
writing technologies such as ink jet printing are
garnering attention as flexible and low cost methods for
patterning materials onto a substrate. Unlike standard
microfabrication approaches where materials are
patterned through a series of depositions, lithographic
masking, and etch steps, direct write technologies
deposit only the necessary amounts of ink additively
onto a substrate at room temperature, therefore
minimizing their total environmental footprint and
processing energy intake. Several direct write tools
have already demonstrated the ability to deposit
multilayer structures and devices as well as 3D
architectures [3].
ink jet printing (SIJP) system with capabilities of
printing features of sub-micron dimensions with high
precision. This system has been demonstrated to eject
droplets with volumes as small as one femtoliter, three
orders of magnitude less than commercial ink jet
printers, resulting in sub-micron line widths. Due to the
reduced drop volume produced by the SIJP, the solvent
evaporates as the droplet travels towards the substrate
and is effectively dried upon impact, eliminating
difficulties with uneven drying and printed morphology
inconsistencies plaguing most other drop on demand
printers. Furthermore, this phenomenon allows high
aspect ratio 3D microstructures such as pillars, walls,
and cylinders to be rapidly printed with remarkable
accuracy through a rapid successive ejection of drops
[4].
Using the SIJP, 3D electrodes for zinc-silver alkaline
microbatteries can be fabricated directly onto a
substrate. Zinc-silver alkaline batteries are known to
have high specific energy and power densities. Unlike
lithium batteries, which are reactive to moisture and
oxygen and must be fabricated in dry boxes, processing
zinc-silver batteries is less difficult, and the materials
are considered recoverable if recycled. Using silver
nanopaste suspensions, the SIJP printed arrays of silver
pillars and these structures were tested as electrodes for
alkaline microbatteries. In this work we demonstrate
the assembly of a 3D alkaline microbattery using the
SIJP and investigate its electrochemical performance in
comparison to thin film electrodes.
Ink jet printing utilizes a “drop-on-demand” approach,
and typical inks consist of nanoparticle, colloidal, or
polymer active components suspended and/or dissolved
in a solvent of which the amount is tailored for a
desired rheology. In this work, we use a custom super
93
Proceedings of PowerMEMS 2008+ microEMS 2008, Sendai, Japan, November 9-12, (2008)
Figure 1. Assembly of a super ink jet printed 3D zincsilver microbattery.
Figure 2. (a) Super ink jet printed silver pillar
electrode (b) 3D image of pillars
2. EXPERIMENTAL
electrodes (Sigma Aldrich) were used to compare the
electrochemical performance of the Harima silver
nanopaste as well as determine an optimal electrolyte
composition.
Silver nanopaste (Harima Chemicals, Inc.) was mixed
with n-tetradecane (Sigma Aldrich) to form ink with
suitable viscosity for printing. The printed silver
conductivity and porosity can be controlled by sinter
temperature and time. For high conductivity and dense
structures, we sintered the printed nanopaste at 250°C
for 1 hour. Cleaned glass slides (Fisher Scientific) were
used as substrates for the printed microbatteries. Two
arrays of 19 x 19 pillars (722 pillars total) were printed
on a 3mm2 silver electrode pad also printed by the
SIJP. The height to diameter ratios and cylinder
diameters of printed pillars can be easily controlled
with the SIJP through varying the residence time of the
printer tip at a location while printing and adjusting the
ejected drop size. Since lower aspect ratio electrodes
were easier to construct and characterize, they were
used as our initial test structures for electrochemical
testing. Pairs of electrode pads with and without pillars
were printed 5 mm apart from each other. The cell
configuration, though far from ideal, was beneficial for
in-situ visual observations during electrochemical
testing.
The structures were profiled using a Keyence VK9700 Confocal Laser Microscope. A Gamry Reference
600 Potentiostat/Galvanostat/ZRA was used to collect
cyclic voltammetry measurements. All cells were
constructed in a 2 electrode set-up and a scan rate of
10mV/s was used. Electrochemical constant current
experiments were conducted using a custom
potentiostat/galvnostat [5]. Any reported capacity,
energy, and power values will be normalized with areal
metrics since for our applications of interest, the area is
constrained rather than volume or weight.
3. RESULTS AND DISCUSSION
An SIJP printed silver electrode pad (3mm x 3mm)
with arrays of pillars is shown in Figure 2a. Figure 2b
displays a 3D image of a section of the electrode. From
the image, it was estimated that the pillars have an
average height of 40 µm and 10 µm diameters and were
located 100 µm apart.
As illustrated in Figure 1, a battery is constructed by
submerging a pair of silver structures into an aqueous
electrolyte of potassium hydroxide (KOH) (Sigma
Aldrich) with dissolved zinc oxide (ZnO) powder
(Strem Chemicals). The optimal molarity of KOH and
the concentration of ZnO dissolved were investigated.
A field is applied between the two electrodes, causing
the surface of the silver positive electrode to oxidize
while zinc electroplates from zinc ions in the
electrolyte onto the surface of the silver negative
electrode, assembling a zinc-silver microbattery in the
charged state. Control experiments with silver foil
Typical bulk silver zinc battery cells use electrolytes
with KOH molarities ranging from 4-10M depending
on the cell construction and desired performance.
Solution conductivity increases with increasing KOH
molarity as does ZnO solubility [6]. We optimized an
electrolyte recipe for 10M KOH through a series of
experiments sampling various ZnO concentrations from
0.1M-0.9M in 0.2M increments. We observed the
charge potential resulting from an application of a
constant current on silver foil pairs submerged in these
electrolytes (Figure 3). Below 0.5M and above 0.7 M,
94
Proceedings of PowerMEMS 2008+ microEMS 2008, Sendai, Japan, November 9-12, (2008)
Figure 3. Galvanostatic charge potentials of varying
ZnO concentrations in 10M KOH electrolyte.
Figure 4. Cyclic voltammogram comparing silver foil
electrodes to Harima silver
the cell potential was unable to surpass 1.7 V,
indicating that any zinc deposited on the negative silver
electrode likely dissolved back into the KOH solution.
The cell with 0.9M ZnO electrolyte also could not
charge beyond 1.7 V. With the 0.5 M (40.7 g/L) and
0.7 M (57 g/L) solutions, the cell was able to charge to
2.0 V and successfully discharged, however only in the
0.7 M solution was the cell able to cycle multiple
times. All subsequent experiments use an electrolyte of
10 M KOH solution with 57 g/L dissolved ZnO.
the reactions are irreversible. Over a series of six
cycles, the resulting voltammagrams in Figure 5 show a
decrease in overall magnitudes in all peaks except the
cathodic peak at 1.82, which increased in magnitude
upon cycling. This shift in magnitude suggests a
change in the stability of the silver oxides formed.
Cyclic voltammetry was used to find the potential
regimes and corresponding magnitudes of oxidation
and reduction reactions for Harima silver electrode
pairs. Strong anodic peaks at 1.73 V and 1.81 V were
visible. These peaks correspond to the formation of
monovalent and divalent oxides of silver at the positive
electrode. Upon reversing the scan direction
(decreasing potential), a slight cathodic peak is
detected at 1.82 V and a strong peak is seen at 1.5 V. A
strong reduction reaction is also seen at 1.3 V. This had
not been detected with foil electrodes (Figure 4), and is
likely due to unknown additives in the Harima silver
nanopaste. The magnitude of the current for the anodic
Upon applying a constant current to the battery, the
measured potential plateaus at certain voltages and this
invariance corresponds to the occurrence of a chemical
reaction, while the length of the plateau can be related
the amount of material reacted. Galvanostatic
experiments shown in Figure 6 verified the presence
and relative magnitudes of the same three cathodic
reduction potentials at 1.82V, 1.5V, and 1.3V seen
using cyclic voltammetry. Similar cycle behavior is
also seen as the length of the 1.5V plateau decreases
and the 1.8V plateau increases with cycle number. The
discharge capacity of the silver electrodes decreases
significantly within the first seven cycles to less than
20% of its initial capacity. For some microbatteries, the
negative electrode was even observed to delaminate
from the glass substrate during cycling. It has been
shown that the initial monolayers of deposited zinc
Figure 5. Cyclic voltammetry of a Harima silver
electrodes in 10M KOH and 57g/L ZnO over six cycles.
Figure 6. Constant current experiments showing a
decrease in battery capacity with increasing cycle
95
Proceedings of PowerMEMS 2008+ microEMS 2008, Sendai, Japan, November 9-12, (2008)
4. CONCLUSION
Alkaline
zinc-silver
microbatteries
were
fabricated using a novel super ink jet printing system.
This provides a simple process to construct 3D
architectures of a variety of morphologies with
minimum processing steps. An electrolyte of 10M
KOH with 57g/L dissolved ZnO was shown to promote
good electrochemical performance when used with the
printed silver electrodes. The SIJP pillared electrodes
were able to demonstrate an increase in capacity
compared to a thin film of similar footprint area, and it
is believed that printing 3D microbatteries with higher
density electrode arrays of larger aspect ratio pillars
with the SIJP will lead to even larger capacity
improvements for a given area.
Figure 7. Galvanostatic discharge curves for pillared.
A discharge current of 1.1 mA/cm2 was used.
5. ACKNOWLEDGEMENTS
form an alloy with the silver electrode followed by
subsequent bulk plating of zinc [7]. As a result, the
initial alloy formation is a large source of stress to the
electrode. Gross analysis of the cathode showed little
cycle fatigue, but further characterization is required.
Future work should include investigating various
charging rates and algorithms that encourage the
formation of stable silver-zinc alloys upon initial
charge, as well as observing how electrolyte
composition might affect this initial deposition and
cycle life.
We would like to thank the East Asia and Pacific
Summer Institutes Program sponsored by the National
Science Foundation and the Japanese Society for the
Promotion of Science as well as the California Energy
Commission (Contract 500-01-43) for their financial
support. We also would like to express our appreciation
to Miyuki Ooyama and Jay Keist for their
contributions.
6. REFERENCES [1] J. Long, B. Dunn, D. Rolison, and H. White,
"Three-dimensional battery architectures,"
Chemical Reviews, vol. 104, p. 4463-4492, Oct
2004.
[2] F. Chamran, Y. Yeh, H. Min, B. Dunn, and C.Kim,
"Fabrication of High-Aspect-Ratio Electrode
Arrays for Three-Dimensional Microbatteries,"
Microelectromechanical Systems, Jan 2007, p. 844852.
[3] J. Lewis, "Direct writing in three dimensions,"
Materials Today, vol. 7, no. 7-8, 2004, p. 32-39.
[4] K. Murata, "Direct Fabrication of Super-Fine
Wiring and Bumping by Using Inkjet process,"
Polymers and Adhesives in Microelectronics and
Photonics, Jan 2007, p. 293-296.
[5] D. Steingart, A. Redfern, C. Ho, J. Evans, and P.
Wright, "Jonny Galvo: A Small, Low Cost
Wireless Galvanostat," ECS Transactions, vol. 1,
2006.
[6] C. Zhang, J. Wang, L. Zhang, Q. Zhang, and C.
Cao, "The Behavior of the Amalgamated Zinc
Electrode in Supersaturated Alkaline Zincate
Solutions," JECS, vol. 148, no. 7, p. E310-E312,
Jun 2001.
[7] G. Adzic, J. McBreen, and M. Chu, "Adsorption
and Alloy Formation of Zinc Layers on Silver,"
Journal of The Electrochemical Society, vol. 128,
no. 8, Jan 1981, p. 1691-1697.
In Figure 7 the galvanostatic discharge performance
of a microbattery with SIJP printed pillar electrodes is
compared to a cell constructed with SIJP printed
electrode pads with the same footprint area of 3mm2.
For a discharge current of 0.1 mA, the pillared
electrodes show a ~60% increase in areal capacity due
to a calculated volume increase of 50% and surface
area increase of 20%. This confirms the notion that
capacity indeed scales with volume.
For our initial investigations, the electrode array
configuration used in this experiment was sparse: only
2.5% of the electrode pad footprint area was occupied
with pillars, and each had diameter to height aspect
ratio of approximately 1:4. For future experiments, the
SIJP can easily print denser arrays of higher aspect
ratio pillars, and we can expect even higher capacities
for a given area. The power behavior of the pillar
electrodes was not investigated since the electrode
configuration was not ideal. In the future, printing
arrays of interdigitated configurations of positive and
negative electrode pillars will both enhance areal
capacity while maintaining small transport distances
between electrodes. Current work is also concentrated
on developing a solid-state polymer electrolyte to
eliminate the need to hermetically package any liquid
components in the battery.
96