146 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 10, NO.

1, MARCH 2001

Microscale Materials Testing Using MEMS Actuators

M. A. Haque and M. T. A. Saif

Abstract—Small size scale and high resolutions in force and dis- during tests, the problems of fine force resolution and specimen
placement measurements make MEMS actuators appropriate for gripping can be approached by using a substrate layer (usu-
micromechanical testing. In this paper, for the first time, we present ally very compliant, and with known materials properties) along
methodologies for uniaxial tensile and cantilever bending testing
of both micrometer- and submicrometer-scale freestanding spec- with the actual film to be tested. This is demonstrated by [5],
imens using MEMS actuators. We also introduce dry fabrication in which aluminum films were tested with thickness from 60
processes for the specimens. The methodologies allow freestanding to 240 nm on polymer substrates. However, introduction of the
single or multilayered thin-film specimens to be fabricated sep- substrate complicates the experimental analysis because: 1) the
arately from the MEMS actuators. For the uniaxial tension test, microscale materials properties of the substrate itself may not
tensile forces are applied by lateral comb drive actuators capable
of generating a total load of 383 N at 40 V with resolutions on be known accurately and 2) the interface may influence the me-
the order of 3 nN. A similar actuator is used in the bending test, chanical behavior of the film. Therefore, it is desirable to test
with load resolution of 58 nN and spring constant of 0.78 N/m. freestanding thin films. This has been attempted by researchers
The tensile testing methodology is demonstrated with the testing of who designed experimental setups with larger specimen sizes
a 110-nm-thick freestanding aluminum specimen. The cantilever to cope with the coarser load resolutions. The tensilometer re-
bending experiment is performed on a 100-nm-thick aluminum
specimen. The experimental setups can be mounted in a SEM (and ported by Hoffman (1993) [7]is capable of generating 0.1-N
also in a TEM after modifications for tensile testing) for in situ force and was used to test 0.5 1.5 mm 150 m 100-nm
observation of materials behavior under different environmental aluminum films. Ruud et.al[10] used motor-driven micrometers
conditions. Remarkable strengthening is observed in all the spec- to produce elongation in freestanding films, then used a load cell
imens tested compared to their bulk counterparts in both tensile to read the force and laser spots diffracted from the gratings on
and bending experiments. Experimental results highlight the po-
tential of MEMS actuators as a new tool for materials research. the specimen surface to determine the strain. The force resolu-
[518] tion of their setup was 2 mN, and specimens could be tested with
Index Terms—Bending test, MEMS, tensile testing, thin films. 1 cm 3.3 0.013 mm 1.9 2.6 m dimensions. Read [9]
developed a piezoactuated tensile testing apparatus with force
and displacement resolutions of 200 N and 20 nm, respec-
I. INTRODUCTION tively, and demonstrated it on 700 200 1.2 m multilay-
ered film specimens. Piezoelectric actuators have been previ-
T HIN films at micrometer and submicrometer level are
prevalently used in MEMS. They experience intrinsic
loads developed during the deposition processes [13] and ex-
ously utilized by [16] and [3], who used load cell-laser inter-
ferometry and strain gauge-optical encoder assemblies, respec-
trinsic loads due to operational and environmental conditions of tively, to measure force and displacements. They tested polysil-
the devices. They may fail to maintain mechanical integrity, as icon structures with thickness of 3.5 and 2 m, respectively.
observed by cracking, delamination, and void/hillock formation Cantilever bending test is also a popular bulk testing method
under stresses [6]. Accurate prediction of thin-film materials re- and is equally difficult to implement on freestanding thin films.
sponse is a challenging problem because bulk testing methods, Since the bending stiffness of freestanding thin films is much
such as uniaxial tension test, are very difficult to apply directly smaller than their tensile stiffness, the force resolution of the
to thin films, and extrapolation of bulk materials properties to loading device must be high, and its spring constant must be
the microscale is not scientific and reliable [15]. The problem comparable to that of the specimen. The cantilever bending test
is further complicated by the fact that mechanical properties of was first applied to thin films by [17], who used a nanoindenter
thin films are significantly affected by the fabrication processes as the loading device. The thinnest freestanding film tested by
[1] and are very sensitive to the influences of interfaces and this method was a 0.87- m-thick gold film.
adjoining materials [8]. The motivation of this paper comes from the challenges in-
Uniaxial tensile test, a popular method in bulk testing, is dif- volved with micromechanical testing and the opportunities of-
ficult to perform on thin films because of the challenges in 1) fered by MEMS actuators. In this paper, we explore the effec-
generating small forces (on the order of micronewtons), 2) grip- tiveness of MEMS actuators for uniaxial tensile and cantilever
ping of the specimen, and 3) preventing bending force compo- bending testing of thin films. The flexibility in the design for
nent in the specimen. While it is difficult to ensure no bending total force generation, force resolution, and structural compli-
ance render MEMS actuators unique for the testing of microm-
Manuscript received January 4, 2000; revised November 30, 2000. This work eter- and submicrometer-level, freestanding, single- or multi-
was supported by the National Science Foundation under Career Grant ECS layer thin films, or even films deposited on substrates, if re-
97-34368. Subject Editor, R. T. Howe. quired. Their small size and thermal and vacuum compatibility
The authors are with the Department of Mechanical and Industrial Engi-
neering, University of Illinois at Urbana-Champaign, Urbana, IL 61801 USA. favor their application in experiments inside analytic chambers
Publisher Item Identifier S 1057-7157(01)01593-1. such as the SEM and TEM (with innovative stage design) under
1057–7157/01$10.00 © 2001 IEEE
HAQUE AND SAIF: MICROSCALE MATERIALS TESTING USING MEMS ACTUATORS 147

Fig. 1. Schematic of the tensile testing experimental setup. The actuator moves
left upon actuaton. The dashed lines indicate the cleaves in the substrates to
make the actuator grip and specimen hang freely on air. The set of verniers are
cofabricated with the actuator are used to measure its axial displacement.

different environmental conditions. In the next sections, we de- Fig. 2. Experimental setup. The actuator grip and the tensile test specimen are
scribe the experimental setups for tensile and bending testing to aligned by the precision positioning translation and rotational stages. This is
illustrate the potentials of MEMS actuators as mentioned above. done under an optical microscope.

II. UNIAXIAL TENSILE TESTING METHODOLOGY

These constants can be expressed as a unified constant . After
We have developed a MEMS-based tensile testing method-
fabrication of the actuator, the actual dimensions of and vary
ology to test freestanding films with thickness ranging from
from the design values. They also vary from actuator to actu-
nanometers to micrometers. An electrostatic comb drive
ator fabricated in the same wafer, rendering the constant and
actuator generates the tensile force. The actuator is capable of
the spring constant unknown. The calibration beam shown in
gripping the specimen and has a self-calibration mechanism.
Fig. 1 is used to determine these unknown parameters. It is a
The specimen is fabricated separately from the actuator and is
slender beam of single crystal silicon with known dimensions
designed to mesh with the actuator grip. We now present the
and material properties. One end of this beam is attached to the
methodology in detail.
moving parts of the actuator, while the other one is fixed to an
Fig. 1 shows the schematic diagram of the tensile testing
anchor. The actuator generates a compressive force on the beam
experimental setup. The actuator (shown on the left) has a
and buckles it as the load exceeds a critical value that de-
C-shaped grip, a calibration beam made of single crystal
pends on the beam geometry and its elastic modulus. The lateral
silicon, and banks of fixed and movable combs. It spans an
buckling displacement at different voltages, along with the di-
area of 3 7 mm and is 20 m deep. The backbone and
mensions of the beam, can be used to find the value of and the
the moving combs of the actuator are supported by structural
spring constant of the actuator [11].
beams. These beams are supported by anchors on the wafer
In a typical experiment, the wafer pieces containing the ac-
and contribute to the actuators spring constant. The wafer is cut
tuator and the specimen are mounted on positioning stages, as
through the vertical dashed line so that the grip hangs freely in
shown in Fig. 2. The gripping end of the specimen is then posi-
air. The specimen (shown in the right) has a freestanding gauge
tioned inside the actuator grip with the stages. A rotational stage
length section and a geometry that conforms to the actuator
is used to align the actuator displacement direction to the ten-
grip. The actuator and the specimen are mounted on precision
sile axis of the specimen. Initially, a small gap is left between the
motion stages to bring the specimen inside the actuator grip.
gripping edge of the specimen and the actuator grip. This gap,
This is done under an optical microscope. Once the specimen
denoted by , is about 1–2 m and prevents any preloading
and actuator are aligned to each other, the whole experimental
of the specimen. Voltage is then applied on the actuator. Upon
setup can be mounted inside a SEM to conduct in situ tests.
actuation, the calibration beam absorbs all the force generated
The complete setup, shown in Fig. 2, occupies a space of 4 in
until its critical load is reached, after which the specimen
1.5 in 1.5 in, including the positioning stages. The size of
and the actuator springs start to share the generated displace-
the stages can be reduced further.
ment and force. For an axial displacement of the
The force generated by a lateral electrostatic comb drive ac-
actuator, the force balance is given by
tuator is given by [14]

(1) (2)

where where
number of moving and fixed comb pairs; applied voltage;
permittivity constant; calibration parameter defined in (1);
height of the combs; spring constant of the actuator;
lateral gap between the fixed and moving combs. length of the calibration beam.
148 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 10, NO. 1, MARCH 2001

Fig. 3. Schematic of specimen fabrication process.

After the actuator is displaced by an amount of , gripping of beam is a magnified form of the axial displacement of the
the specimen starts, and tensile force is applied on the specimen. actuator. Therefore, the displacement resolution depends on the
The net force on the specimen is given by length of the calibration beam, as well as on the accuracy of
reading of its lateral displacements. This is given by
for (3)
(6)
For each value of applied voltage, the corresponding lateral
The force and displacement resolutions of the actuator used
displacement of the calibration beam is measured with the
in this study are given later in this section. We now briefly de-
help of cofabricated vernier scales (shown later in Fig. 5) that
scribe the design and fabrication issues for the specimen and the
can read a change of with 0.3 m resolution under an optical
microscope. The corresponding of the actuator is obtained actuator.
from [11] A. Design and Fabrication of the Specimen
The wet processing of freestanding films involves large sur-
(4)
face forces, and hence films with submicrometer thickness are
difficult to fabricate. Also, the specimen may be preloaded by
Once the axial displacement is obtained, the force on the
these forces before the actual test. To avoid this problem, a dry
specimen is then computed using (3). It is important to note that
process for specimen fabrication [4] was introduced.
elongation of the specimen may not necessarily be restricted to
Fig. 3 shows the schematic of the fabrication process. The
the gauge length section. Therefore, the actual elongation in the
process steps are as follows.
specimen needs to be measured from the recorded digital images
of the gauge length section under different applied forces. This 1) Photoresist AZ 5214 spun at 4000 rpm and soft-baked for
was done using a software for dimension measurement, which 35 s at 100 C;
has a pixel-to-pixel resolution of 90 nm. Resizing the acquired 2) Hard contact exposure with 65 mJ;
images before taking length measurements can further enhance 3) Postexposure bake for 45 s at 120 C;
this. 4) Flood exposure bake with 2 mJ/cm energy flux for 10 s.
The force resolution of the actuator is obtained by differenti- 5) Development with 1:3 diluted AZ 351 developer for 45
ating (1) s. At this stage, a negative pattern of the specimen is im-
printed on the wafer with liftoff profile;
(5) 6) Evaporation of aluminum (more than one metal can be
evaporated for multilayer specimens) on the pattern up to
which implies that the resolution changes with the applied desired thickness (100 nm in this study);
voltage and depends on the resolution of the voltage source. 7) Liftoff of the photoresist;
The displacement resolution is obtained by differentiating (4), 8) The specimen is then annealed at 200 C for 1 h. To pre-
where we see that the lateral displacement of the calibration vent folding of the specimen at the gripping edges, we
HAQUE AND SAIF: MICROSCALE MATERIALS TESTING USING MEMS ACTUATORS 149

Fig. 4. Schematic of a freestanding tensile test specimen. The wafer is cleaved so that the gauge length section hangs freely on air. AB and CD are the gripping
edges. The specimen is reinforced with a top layer of photoresist except for the gauge length section.

Fig. 6. A 110-nm-thick freestanding aluminum sample being gripped and

stretched by the actuator.
Fig. 5. The actuator showing the main combs, additional combs, calibration
2
beam, and the grip (50 ). Inset: vernier scale attached to the calibration beam. B. Design and Fabrication of the Actuator
A lateral comb drive actuator with 3150 combs was designed
reinforced the gripping end by patterning a layer of pho- to generate a total force of 382 N at 40 V. In addition, there
toresist on the specimen excluding the gauge length sec- are 280 combs that can be separately activated for fine force
tion. We then evaporate 10 nm of aluminum onto the spec- resolution. The maximum allowable axial displacement of the
imen, although this is not a required step in the fabrication actuator was 10 m. Fig. 5 shows part of the actuator with the
process; main and the additional combs and the calibration beam with the
9) Reactive ion etching for 30 min at 50 mtorr pressure, 20 vernier scale. It was fabricated by a single crystal reactive ion
sccm SF at 60 W to release the specimen. etching and metallization process [12]. The calibration beam is
Fig. 4 shows the schematic of a tensile testing specimen. The 517 m long, 1 m wide, and 20 m deep. The calibrated value
freestanding film is reinforced (everywhere except the gauge of is 0.228 N/V for the 3150 pairs of combs. The spring
length section) by a layer of photoresist layer at its top. Dis- constant is 4.1 N/m.
tributed tensile load is applied by the actuator grip on the edges The main set of 3150 comb pairs has a force resolution of
AB and CD. 1.368 N at 30 V for 0.1-V increments. The additional set of
150 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 10, NO. 1, MARCH 2001

Fig. 7. Load-displacement plot for the 110-nm-thick aluminum film. Upper curves are for loading and unloading of both specimen and actuator. Lower curve
shows load-displacement characteristics of the actuator only.

280 combs can provide 3-nN force resolution for an applied evidence of the difference of materials behavior in the macro-
voltage of 1 V with a voltage source with 0.1 V resolution. The scopic and microscopic scale. The aluminum film survived a
displacement resolution of the actuator is obtained from (6) and tensile stress of roughly 616 MPa, which is higher than previ-
is 58 nm after 2 m of axial displacement. ously obtained values of about 180 MPa for freestanding films
(Hoffman, 1993) [7]and 280 MPa [5] for 110- and 120-nm-thick
aluminum films with polymer substrates, respectively. This be-
III. TENSILE TESTING RESULTS
havior is significantly different from that the bulk pure alu-
To demonstrate the proposed methodology, we tested a spec- minum with tensile strength of 55 MPa [2]. Conclusive results
imen 2.3 m wide with 10 m gauge length and thickness of 110 and fundamental understanding of the behavior of micrometer-
nm made of 99.99% pure evaporated aluminum. The gripping and submicrometer-scale materials will require further studies
end of the specimen was reinforced with a 1.3- m-thick layer under the SEM or the TEM and are beyond the scope of this
of photoresist, as shown in Fig. 4. Fig. 6 shows the specimen’s paper.
being gripped and stretched by the actuator.
The force-displacement plots for loading and unloading of
IV. MICROCANTILEVER BENDING TESTING METHODOLOGY
the aluminum film are given by the two upper curves in Fig. 8.
The loads shown here are the amounts generated by the actu- The experimental setup for the microcantilever bending test
ator as given by (1) and not just the loads on the specimen. consists of the microcantilever specimen, the MEMS actuator,
The displacements shown in these curves include that of the and precision motion positioning stages to properly align the
actuator, as well as the gauge length and the gripping end of actuator with the specimen. We describe the setup components
the specimen. Direct measurement of elongation in the gauge and the experimental procedure below.
length section could not be performed under the optical mi- 1) Specimen Preparation: The specimen used in this study
croscope; hence the results are not expressed in terms of en- is a microcantilever aluminum beam of 99.99% purity. It is fab-
gineering strain. Due to intrinsic stress, the film was curled up- ricated separately from the actuator. The beam is 11.3 m long,
wards after release from the substrate. Upon gripping and subse- 2.1 m wide, and 100 nm thick. The fabrication process is ba-
quent loading, it was first straightened, and hence it showed little sically the same as described for the tensile testing specimen in
resistance against displacement during this stage. After that, the an earlier section.
specimen started to show resistance against deformation, as in- 2) MEMS Actuator: The actuator used in the microcan-
dicated by the change in the slopes of the load-displacement tilever bending test has the same operating principles as
curve. Upon unloading, the specimen followed a different path, described previously. It has a probe that applies a force on the
which indicates the occurrence of yielding in the specimen. The cantilever beam specimen, as shown in Fig. 8. The actuator is
lower curve in Fig. 7 shows the load displacement characteris- 10 m deep, spans an area of 2 mm 3 mm, and has a total
tics of the actuator structure alone. of 660 combs. The value of the actuator is 2.92 10
The results obtained in this study substantiate the feasibility N/V . The actuator spring constant is 0.78 N/m. This value is
of the proposed methodology and go beyond that. They provide comparable to the elastic spring constant of the specimen (0.38
HAQUE AND SAIF: MICROSCALE MATERIALS TESTING USING MEMS ACTUATORS 151

Fig. 8. SEM micrograph of the MEMS actuator and the microcantilever specimen after alignment. The actuator and the specimen are mounted on precision
motion stages and then brought to the above configuration. The actuator probe motion is in the negative y direction only, applying bending load on the cantilever
beam (barely visible at this magnification) at a constant lever arm. Right: schematic of the experimental setup.

N/m), which implies that the actuator springs are soft enough
to capture the actual response of the beam to the bending load,
and not just the data noise. The force resolution of the actuator
is 58 nN at 10 V, assuming increments of 0.1 V.
After fabrication, the silicon substrate is cleaved such that
the probe freely hangs in air. Precision motion stages were then
used to align the actuator probe with the specimen, as shown
in Fig. 8. At this stage, the lever arm of the cantilever beam is
chosen to be 4.5 m by positioning the actuator probe to the
desired location along the length of the beam. Fig. 8 shows the
SEM micrograph of the actuator with the probe after they are
aligned. The specimen is barely visible at this magnification.
After alignment, an initial gap of about 1 m was left between
the specimen and actuator probe. Upon actuation, the probe ap-
proaches the specimen and applies force on it. It is important to Fig. 9. Load-deflection profile of a 100-nm-thick aluminum microcantilever
note that the position of the actuator is fixed with respect to the beam. Deflection measured at a lever arm of 4.5 m.
specimen, so that the force is always applied to the specimen at
a fixed lever arm. The contact point between the probe and the probe and the specimen and does not represent the true mate-
specimen moves along the length of the specimen during defor- rials behavior. This is evident from the materials response in the
mation. rest of the loading and unloading history. This instability set-
For each value of the applied voltage, corresponding force tles down and the beam deflects linearly with increasing load
on the specimen was calculated by determining the total force up to 0.62 N (point C), when it just starts to show nonlinear
generated by the actuator and subtracting the restoring force. response again. Upon unloading, the specimen behaves linearly
The deflections at the probe tip for each value of the applied (from point D to F) but shows some permanent deflection, indi-
load were acquired in form of digital images. cating that plastic deformation has taken place. This permanent
deflection was accounted for in determining the load-deflection
profile for the next loading cycle. In this cycle, the specimen
V. MICROCANTILEVER BENDING EXPERIMENTAL RESULTS
showed linear elastic behavior until 0.79 N (point G) and then
The microcantilever specimen was loaded and unloaded in deviated from linearity. The load at point G is larger than that at
two cycles. Fig. 9 shows the load-deflection profiles for these point C, which reflects the strain hardening that occurred in the
loading cycles. In this figure, several data points have been la- specimen. Once again, the specimen unloaded in a linear way
beled with alphabets, which will be used to describe the results and showed more plastic deformation when the probe was com-
and analysis in this section. In the first cycle, loading was in- pletely withdrawn from the specimen.
creased up to 0.75 N (point D in Fig. 9). The initial nonlin- The stress on the specimen at the beam support is about 880
earity (point A to B) is due to the unstable contact between the MPa for the loading point C where the film begins to yield. This
152 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 10, NO. 1, MARCH 2001

value is 49 times larger than the bulk yield stress of pure alu- [7] G. T. Mearini and R. W. Hoffman, “Tensile properties of aluminum/alu-
minum (55 MPa). The experimental results once again confirm mina multi-layered thin films,” J. Electron. Mater., vol. 22, no. 6, pp.
623–629, 1993.
the strengthening effect in thin films and incite further experi- [8] W. D. Nix, “Mechanical properties of thin films,” Metallurgical Trans.,
mentation and analysis that are beyond the scope of this paper. vol. 20A, p. 2217, 1989.
[9] D. T. Read, “Piezo-actuated microtensile test apparatus,” J. Testing
Eval., vol. 26, no. 3, pp. 255–259, 1998.
VI. CONCLUSION [10] J. A. Ruud, D. Josell, and F. Spaepen, “A new method for tensile testing
of thin films,” J. Mater. Res., vol. 8, no. 1, pp. 112–117, 1993.
We have demonstrated the potentials of MEMS actuators on [11] M. T. A. Saif and N. C. MacDonald, “Measurement of forces and spring
micromechanical testing by performing uniaxial tensile test and constants of microinstruments,” Rev. Sci. Instrum., vol. 69, no. 3, pp.
microcantilever bending test on freestanding thin films in the 1410–1422, Mar. 1998.
[12] K. A. Shaw, Z. L. Zhang, and N. C. MacDonald, “SCREAM I: A
micro- submicrometer-scale using MEMS devices. The attrac- single mask single-crystal silicon, reactive ion etching process for
tive features of the proposed methodologies are: micro-electro-mechanical structures,” Sensors Actuators A, vol. 40,
1994.
1) flexibility in design for actuator force and displacement [13] F. Spaepen and A. L. Shull, “Mechanical properties of thin films
resolution; and multilayers,” Current Opinion Solid State Mater. Sci., vol. 1, pp.
2) easy dry fabrication of freestanding thin-film specimens 679–683, 1996.
[14] W. C. Tang, C. H. Nguyen, and R. T. Howe, “Laterally driven polysil-
with thickness ranging from micrometers to nanometers; icon resonant microstructures,” Sensors Actuators A, vol. 20, pp. 25–32,
3) very small overall setup size encouraging in situ observa- 1989.
tion of materials behavior in analytical chambers such as [15] R. P. Vinci and J. J. Vlassak, “Mechanical behavior of thin films,” in
Annu. Rev. Mater. Sci., vol. 26, 1996.
the SEM and TEM. [16] B. Yuan and W. N. Sharpe, “Mechanical testing of polysilicon thin
Two tensile tests were carried out on polymer (photoresist) and films,” Exper. Mech., pp. 32–35, Mar./Apr. 1997.
[17] T. P. Weihs, S. Hong, J. C. Bravman, and W. D. Nix, “Mechanical deflec-
aluminum films of thickness 1.3 m and 110 nm, respectively. tion of cantilever microbeams: A new technique for testing the mechan-
The results show significant difference in materials behavior be- ical properties of thin films,” J. Mater. Res., vol. 3, no. 5, pp. 931–942,
tween the bulk and microscales in uniaxial tension. Similar re- 1988.
sults are obtained from the microcantilever beam bending test on
a 100-nm-thick aluminum film, where the yield stress is about
880 MPa, which is about 49 times the bulk yield stress for com- M. A. Haque received the B.S. and M.S. degrees in
mercially pure aluminum. mechanical and industrial engineering, respectively,
from Bangladesh University of Engineering & Tech-
nology, Bangladesh, in 1993 and 1995, respectively.
ACKNOWLEDGMENT He received the M.B.A. degree from the University of
Alberta, Canada. He is currently pursuing the Ph.D.
The authors acknowledge C. Sager for his help during actu- degree at the Department of Mechanical & Industrial
ator fabrication. The actuators were fabricated in the Cornell Engineering, University of Illinois at Urbana Cham-
NanoFabrication Facility at Cornell University. The specimens paign.
His research interests include micro/nanoscale ma-
were fabricated in the Frederick Seitz Materials Research Lab- terials science, design and fabrication of MEMS for
oratory at the University of Illinois at Urbana-Champaign. microinstrumentation, and engineering optimization.

REFERENCES
[1] F. R. Brotzen, “Mechanical testing of thin films,” Int. Mater. Rev., vol. M. T. A. Saif received the B.S. degree in structural
39, no. 1, pp. 24–45, 1994. civil engineering from Bangladesh University of
[2] W. D. Callister, Materials Science and Engineering: An Introduction, Engineering & Technology, Bangladesh, in 1984,
4th ed. New York: Wiley, 1997. the M.S. degree in structural civil engineering from
[3] S. Greek, F. Ericson, S. Johansson, and J. A. Schweitz, “Micromechan- Washington State University, Pullman, and the Ph.D.
ical tensile testing,” in Proc. Mater. Res. Soc. Symp., vol. 436, 1997, pp. degree in theoretical and applied mechanics from
227–232. Cornell University, Ithaca, NY, in 1993.
[4] M. A. Haque and M. T. A. Saif, “Investigation of micro-scale materials He was a Postdoctoral Associate in electrical engi-
behavior with MEMS,” in Proc. Int. Mechanical Engineering Congress neering at Cornell University. He is currently an As-
and Exposition, Nashville, TN, 1999. sistant Professor in the Department of Mechanical &
[5] Y. S. Kang and P. S. Ho, “Thickness dependent mechanical behavior Industrial Engineering, University of Illinois at Ur-
of submicron aluminum films,” J. Electr. Mater., vol. 26, no. 7, pp. bana-Champaign. His research interests include design, analysis, and fabrica-
805–813, 1997. tion of MEMS for submicrometer-scale materials studies and noninvasive in-
[6] M. Madou, Fundamentals of Microfabrication. Boca Raton, FL: CRC vestigations of living cells; nonlinear dynamics of MEMS, and bistable MEMS
Press, 1997. sensors.