Documente Academic
Documente Profesional
Documente Cultură
School of Electronics Engineering, VIT University, Chennai Campus, Chennai 600127, India
School of Medical Science and Technology, Indian Institute of Technology, Kharagpur 721302, India
a r t i c l e i n f o
abstract
Keywords:
Galvanic corrosion
Multilayer metal stack
MEMS piezoresistive accelerometer
Selective metal deposition
Electroplating
Shadow mask technique
1. Introduction
When two or more different types of metals come into
contact in the presence of an electrolyte, a galvanic couple
is set up due to the difference of electrode potential of
various metals, which leads to galvanic corrosion of the
least noble metal [1]. For various technological applications, the use of two or more layers of different metal
lms is quite common [2,3]. One such example is a
MEMS based silicon piezoresistive accelerometer [3], where
three metal layers, i.e. aluminium (Al), chromium (Cr)
n
Corresponding author. Tel.: 91 44 3993 1274;
fax: 91 44 3993 2555.
E-mail address: a.ravishan@gmail.com (A. Ravi Sankar).
1369-8001/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mssp.2012.08.003
450
Al
Cr
Au
Cr etchant
Au etchant
210
140
130
620
62
240
lm in direct contact with the electrolytic solution whereas the contact area is the overlapping region of the metal
stack layers as schematically shown in Fig. 1(a). Thus in
the present study, samples were prepared with various
exposed areas of the individual lm along with a xed
overlapping region of different metal lms in binary metal
stack as shown in Fig. 1(b). A schematic of the top view
of the AlCrAu ternary metal stack is shown in Fig. 1(c)
with its cross-sectional views across AA0 and BB0 in
Fig. 1(d). In the ternary metal stack layer (AlCrAu)
experiment, cA is the intermetallic contact area of Au
layer at the top of the AlCr stack and eA is the exposed
area of the Al or Cr layer. In this structure, exposed areas
of the Cr and Al lm were kept equal. However, the
exposed areas of Cr layer and Al layer were placed
orthogonally to each other on the silicon surface as shown
in Fig. 1(c). For both binary and ternary metals stacks, ve
different samples with the area ratio cA =eA of 0.25, 0.5, 1,
1.5 and 2 were used for the etch rate study.
For accurate etch rate measurements, lithographically
dened edges are necessary. However, once a metal is
patterned, another metal cannot be deposited and subsequently patterned atop the already patterned metal, since
it may damage underneath metal lm edges due to the
galvanic effect. Hence, in the present study a novel
process step with the combination of lithography process
and selective metal deposition using shadow mask technique [5,6] was used for sample preparation. Silicon
wafers were used for shadow mask preparation by the
anisotropic wet etching technique because of the smooth
edges and wafer surface after anisotropic etching. In the
present study, initially the rst metal lm was deposited
and lithographically patterned. Subsequently the second
metal lm was selectively deposited using the silicon
shadow mask technique to achieve predened exposed
area and contact area. However, in some cases lithography and etching were performed to properly dene the
exposed area of the metal lm keeping the contact or
overlapping area fully protected by a photoresist to avoid
the galvanic corrosion of the underneath metal layer.
The process ow for sample preparation of binary
metal stack is given below with the schematics of each
individual step as shown in Fig. 2(a) to (d). An aluminium
lm of thickness 1 mm was thermally deposited atop a
Si/SiO2 layer (Fig. 2a) and lithography was performed
using a suitable mask (Fig. 2b). Subsequently a 50 nm
thick Cr lm was thermally deposited atop the existing Al
metal lm using a shadow mask as shown in Fig. 2(c).
Then lithography was performed to dene a sharp edge
for Cr lm as shown in Fig. 2(d). During this process the
underlying Al lm was covered by a photoresist. By
varying the dimensions of lithographic mask and shadow
mask, the contact area and exposed area of different metal
lms were varied and Fig. 1(b) shows the binary stack of
AlCr with different exposed areas eA of the lm to
be etched with constant contact area cA . Similarly,
sample for ternary metals stack of AlCrAu was prepared
with a thermally deposited 200 nm thick Au layer atop
the Cr layer. After preparing the samples, both the binary
and ternary metals stacks were dipped in different etching solutions at 25 1C for a xed time and subsequently
451
Fig. 1. Binary and ternary metal stacks. (a) top view of the AlCr binary stack, (b) a cross-sectional view of the binary metals stack across AA0 with a
varying exposed area of Al layer, (c) top view of the AlCrAu ternary metal stack and (d) cross-sectional views of the ternary metals stacks across AA0
and BB0 .
were
measured
using
the
surface
452
453
Fig. 6. Photograph of a fabricated accelerometer device with an electroplated gold layer and Al metal lines.
Acknowledgement
The authors would like to express their gratitude to
(Late) Prof. S. Kal for his valuable suggestions. The authors
acknowledge the members of Microelectronics and MEMS
Laboratory, IIT Kharagpur, for their help in the experiment.
References
[1] A. Peter, Physical Chemistry, 6th ed. W.H. Freeman and Company,
New York, 1997.
[2] H. Zeng, A. Wan, A.D. Feinerman, Nanotechnology 17 (2006) 31833188.
[3] A. Ravi Sankar, S.K. Lahiri, S. Das, Journal of Micromechanics
and Micro engineering 19 (2009). http://dx.doi.org/10.1088.09601317.19.2.025008.
[4] K.R. Williams, K. Gupta, M. Wasilik, Journal of Microelectromechanical Systems 12 (2008) 761778.
[5] A. Tixier, Y. Mita, J.P. Gouy, H. Fujita, Journal of Micromechanics and
Microengineering 10 (2000) 157162.
[6] J.M. Hong, J. Zou, Journal of Micromechanics and Microengineering
(2008) doi:10.1088.0960-1317.18.5.055002.
[7] N. Yazadi, K. Naja, A.S. Salian, Journal of Microelectromechanical
Systems 12 (2003) 479486.
[8] R.P. Van Kampen, R.F. Wolffenbutte, Sensors and Actuators A 64
(1998) 137150.