Professional Documents
Culture Documents
Sridhar Lola
A BSTRA CT
The two-dimensional nature of MEMS, and peculiarities of the microfabrication process oflèr numerous design challenges.
Batch fabrication demands nitnimal or no assembly. Micrornachining techniques limit the tYpe of mechanical structures that can
he built and thereb\ limit the type of mechanical t'unctions that can be realized at the niicro scale. Exploitation of elastic
deformation at the design stage. or use of mechanical compliance in design, leads to jointless. no-assembly monolithic
mechanical de'ices. rhe use of compliance seems to a prudent approach for designing micro mechanical devices. Besides, the
small scale and high aspect ratio of rnicromechanical structures makes them inherentl\ compliant. 1herefore. in MEMS. a
compliant design that needs no assembly is not merely a prudent choice, it is a necessit\.
Compliant mechanisms are monolithic mechanical structures that rely on elastic deformation to generate sophisticated
mechanical functions. This paper presents an overvie of mathematical procedures emplo\ ed for designing compliant
niechanisms. The paper outlines (a) topological synthesis — which involves generation of a functional design in the form of a
feasible topology starting from input output force motion specifcations. and (h) size arid shape opt rn i/at ion - to meet
performance requirements. Some examples of compliant MEMS are also presented.
1. INTRODUCTION
I raditionallv. engineered artifacts are designed to he strong and stiff Designs in nature, on the other hand. are strong hut not
necessarily stiff- the- are compliant '. Although nature thrives on use of compliance, the engineering '' orld has traditionall\
limited itself to rigid structures and rigid mechanisms. Compliant mechanisms are single-piece flexible structures that delier
the desired motion b undergoing elastic deformation as opposed to rigid bod\ motions of conventional mechanisms I Figure
I ). We introduce fully complaint mechanisms with distributed compliance for improved reliability and performance. and ease of
manufacture. Distributed compliant systems derive their flexibility due to topolog\ and shape of the material continuum rather
than concentrated tiexion at few regions as in plastic hinges. Compliant mechanisms are particularl\ suited liar applications
with small range of motions as their unitized construction ithout joints makes their manufacture extremel simple eliminating
assembly operations altogether. The Compliant Stapler shown in Figure 2 illustrates this paradigm o1 no—assemhl\
lurtherniore problems due to near. backlash, lubrication. and noise are alleviated.
Compliance in design leads to jointless. no-assembly. monolithic mechanical devices. Since it is often not very practical to
fabricate jointed micro-mechanisms, due to difficulties in micro-assembly. compliant mechanisms offer an alternative to obtain
micro-mechanical movements without requiring mechanical assembly. Although simple deformable structures such as beams
and diaphragms have performed adequately in many micro devices, more sophisticated micromechanical functions can be
realized by fully exploiting the preferred uses of elastic deformation. Figure 3 shows some examples of compliant
micromechanisms. Majority of microdevices and systems rely on mechanical compliance (flexible structures) to take advantage
of micromachining techniques and to meet batch production constraints. They often curl up because of the residual stresses
induced in the fabrication process. Their size makes them susceptible to elastic deformation under electrostatic forces - a
phenomenon not evidenced in the macro domain.
On one hand, the design of micromechanical system based on traditional rigid-body assumption yields designs that aredifficult
to manufacture and assemble and on the other hand the resulting designs negate the rigid body assumption by exhibiting
compliant behavior under the action of residual stresses. Therefore, in MEMS. micro mechanical design based on the
assumption that resulting mechanical structures will be compliant is not only correct but it also yields designs that meet the
micromachining constraints well.
Nature has realized the pivotal role that compliance plays at the realm of microorganisms, the level at which MEMS fit'.
Nearly 90 percent of living creatures are invertebrates and the percentage of invertebrates increases as we go down the dimension
scale where compliant structures reign.
ii :fl
(a) (b)
(c) (d)
Figure 3: Micro Compliant Mechanism (a) Crimping mechanism. (h) and (c) compliant grippers, and (d) micro four-bar mechanism
The first step in the design of a compliant mechanism is to establish a kinematically functional design that generates the desired
output motion when subjected to prescribed input forces. This is called topological synthesis. Although the size and shape of
individual elements can be optimized to a certain extent in this stage. local constraints such as stress and buckling constraints
cannot be imposed while the topology is being determined. Once a feasible topology is established, performance constraints can
46
In this section, we briefly explain systematic methods of design of compliant mechanisms starting from functional
specifications. First, we describe a method of deriving the topology (configuration) of a compliant mechanism given the desired
input forces and output displacements. We illustrate this procedure by two different methods of implementation (a) ground-truss
structure as initial guess and (b) homogenization method. Next we describe the size and shape optimization of compliant
mechanisms given the desired mechanical or geometric advantage, stress constraints, size constraints etc. All of the work
reported in this paper assumes a linear elastic model. Geometric non-linearities due to large deformation and the dynamic
characteristics are not taken into account.
Given, a single-force-input and a single-displacement-output design specifications, first, we formulate an objective function that
captures the need for (a) compliance to undergo desired deformation (kinematic requirement), and (b) stiffness to resist external
loads (structural requirement) once the mechanism assumes the desired configuration. Second, we adapt a formal structural
optimization technique to synthesize a form, which is an optimal topology, shape and size of a compliant mechanism that
performs the intended function [1]. Three different implementation schemes have been developed — (I) ground truss structure
approach , (ii) continuum approach based on theory ofhomogenization2'5 , and (iii) an array ofbeam elements6. We will briefly
describe the first two.
To satisfy both the kinematic and structural requirements in compliant mechanism synthesis, a two-part problem is posed in
terms of potential energies. The first part, the "mechanism design", is where the kinematic requirements are met by
maximizing the deflection at a specified point along a specified direction. This is achieved by applying a fictitious force at the
point of interest, B, along the direction of the desired output deflection, _. This "dummy load" is denoted by lB and as
shown below for a general design domain subject to an applied force, fA at the point A and some specified boundary
conditions. Maximizing the deflection at the point B in the direction offB is equivalent to maximizing the mutual potential
energy, VBTK UA, where UA i5 the deflection field due tofA, VB is the deflection field due to lB and K is the global stiffness
matrix. The constraints are two equilibrium equations, one due to the applied load, and one due to the dummy load.
Combined Problem
" fB=resistance
maxi
r mutua' energy 1
[ strain energy j
I = max
'BK UA
tIB K UB
max (s,BTKUA) mm (UBTKUB) subjectto KUA =fA
subject to K 1A fA subject to K 14B —fB K VB fB
KV8=J8 —
KUB fB
total resource constraint
lower and upper bounds
The second part of the two-part problem is the structure design, where the structural requirements are met by maximizing the
stiffness. Here the point A is considered fixed, and the resistance of the workpiece is accounted for by applying the force lB at
point B in the opposite direction. Maximizing the stiffness is equivalent to minimizing the strain energy, uBT K UB, where UB
47
In this method, the prescribed design domain (this is the area within which the mechanism should fit) is first divided into a
number of nodes. Each node is connected to every' other node by a truss element. This serves as an initial guess. Certain nodes
are "fixed" to imply the points where the mechanism is anchored to the substrate. The cross sectional area of each truss
elements serve as the design variables with specified upper and lower bounds. The resource constraint provides less material
than the available space. The objective then is to distribute the material in a way that maximizes the objective function. During
the optimization process. those truss elements whose cross sectional area reaches the lower bound are removed (deemed
unnecessary) leaving only a network of truss elements whose area reached the upper hound. This defines the topology of the
compliant mechanism.
The results of the automated synthesis method are illustrated by the following example of a compliant gripper mechanism. The
design problem represents a half-view, where the dashed line represents the design domain, and the nodal constraints (boundary
conditions) are as indicated. The design specifications are that the applied force. F. cause the motion. D. at the indicated
location, which 'ill allow the device to grip some object at that point.
The initial guess is a full ground structure with a uniform distribution of cross-sectional areas. When the algorithm converges.
the solution consists of truss members whose design variable reached (or is close to) the upper bound. The truss members
whose design variable reached the lower bound constraint are eliminated. The optimized solution and corresponding finite
element model are shown below, where the undeformed shape is denoted by the dashed lines and the deformed shape is denoted
by the solid lines. Compliant grippers based on this design were built out of nylon using the Stratasys 3D Modeler.
- __ - 1-
Design Problem Initial Guess
In Homogenization methods the design domain is parametrized using cellular microstructure (Figure 5(a)) and the optimization
method arrives at optimal values for the parameters of each cell (ab, and q). Thus, there will be 3n variables in the
optimization problem if there are n elements (cells) in the domain. If the optimized hole dimensions a and b reach their limit
values. i.e.. the length and width of the cell, a hole will be generated: if a and b are zero, then a solid cell is created and
intermediate values give rise to porous regions. This gives the method the ability to generate any topology, shape and size that
are optimal for given problem specifications (which are applied forces, desired output displacements and the amount of material
to be distributed in a prescribed design domain).
48
Force
Figure 5 ( C): Micro compliant crimping mechanism fabricated by F3ulk-mieromachining process. Note the simiIarit of the top-half of
the mechanism with the topology generated in the last design iteration shown in Figure 5(h
Once the topology is established, the next logical step is to perform a size and shape optimization. In order to produce
practical compliant mechanism designs. the following design criteria must be addressed.
(i) required kinematic motion (both magnitude and direction). (ii) required stiffness to an external load. (iii) design space. (iv)
materials properties. (v) stress limitations, (vi) buckling instabilities, (vi) dynamic considerations, and (v) weight limitations.
49
I Work performed at the output is measured by applying an external resistive load which opposes the desired direction of the
output port on the body. The external load can be considered to be a worst case load (on an average load) arising from the
environment of the mechanism.
As shown in Figure 6 below, these boundary conditions are applied in two separate stages. First, the external force is applied to
the body while the input is held fixed. Second, the input is actuated a fmite distance with the external load applied. Triangular
regions in Figure below at the input and output illustrate the energy absorbed due to loading and flexure. The shaded areas at
both ports represent the reciprocal work or a fixed kinematic relationship between the input and the output.
max TIe.r.yiclency(Si, k)
subject to:
Uoit2
= MAdesired or GAdesIred
ipil U1?1
U0uf2*Uperp
Fe5 < F'crit
i Tmax
.1
Volume Resourcem
The results ofthe size and shape optimization based on the energy formulation are presented in Figure 7. Note that the
geometry, size and shape of the optimized mechanism have changed but the topology remains in tact. The energy efficiency of
the fmal design is 95.5%. The mechanism amplifies the input motion by a factor of two (GA = 1:2). By connecting a series of
such motion amplification mechanisms, we have developed a compliant mechanism, called MEMS Multiplier, that magnifies
the input motion by 20 times. This design is currently being fabricated by Sandia National Labs for possible use in their micro
weapon-lock system.
50
reprOcai raIatorhi
(equa' reas
disp*rricri
3
E 2 MPa
:\ T
4.2:j
•, 1.99:1
GA,j= 1:2.09 .
2.4 V-Beam Suspensions: A novel suspension mechanism described here illustrates yet another application of compliant
MEMS. The two main components of a linear microactuator are, (i) electrostatic drive system and, (ii) elastic suspension
system. Mechanical design aspects of an elastic suspension system play a major role in controlling the performance
characteristics of a linear microactuator. The role of the elastic suspension system in microactuators is to provide an elastic
restoration force and to guide a planar shuttle in a rectilinear motion. For best performance. an elastic suspension system should
satisfy the following requirements:
(i) rectilinear motion guidance - the suspension system should have very low stiffness in the direction of motion and very
high stiffness in the other two perpendicular directions,
(ii) linear response characteristic - the suspension system should not introduce any geometric non-linearities. and
(iii) minimum residual stresses -the design should he robust against fabrication-induced residual stresses.
51
A compliant mechanism exhibits the behavior of both a mechanism and a structure, and can be considered as a synergistic
combination of the two. Thus, the design and analysis of compliant mechanisms can be treated in two wavs-considering their
abilirv to transform energy and transmit motion/force, they can be treated purely as mechanisms, or considering their ability to
undergo controlled deformations, they can be treated purely as a structures. Accordingly. two different approaches have been
developed for synthesis of compliant mechanisms, one based predominantly on the lines of traditional rigid-link mechanism
design and the other based predominantly on the lines of traditional structural design.
In a related field of research, Griffis and Duffv', and Pigoski and Duft' have presented modeling and inverse force analysis of a
system of springs. Although, these works discussed systems of springs from a different perspective, the mechanisms, however.
have interesting correlation to the class of compliant ,nechanis,ns under discussion. In a more recent work that relates to
compliant mechanisms directly, Yin ci al. '' have presented the inverse force analysis of a four-bar mechanism with compliant
input and output links.
The treatment of the synthesis and analysis methods in the above-cited literature was largely based on the lines of traditional
kinematics of rigid-link mechanism. Moreover, in all of the above-mentioned studies, the compliant mechanisms considered
had either highly localized compliance. i.e.. flexural pivots, to simulate conventional revolute joints, or flexible links with at
least one rigid moving link.
52
4.0 Conclusions
Complaint mechanisms play an important role in the design of micro mechanical structures for MEMS applications. These
monolithic mechanical structures can be designed to perform complex mechanical functions and fabricated within the constraints
of present day micromachining processes. Based on linear elastic models, we have developed methods of synthesis of
compliant mechanisms to meet kinematic and static stiffness requirements. Our future work includes non-linearities due to large
deformation, and dynamic aspects of micromechanical structures. The design methods developed to-date can generate micro
mechanism designs for a variety of applications including, motion/force amplification, static shape change, and multiple
input/output force-displacements.
5. 0 Acknowledgements
The author gratefully acknowledge the financial support from the National Science Foundation and the Air Force Office of
Scientific Research to carry this research. This research was carried out in close collaboration with Professor Noboro Kikuchi
and is based on doctoral dissertations of G.K. Ananthasuresh, Mary Frecker, Laxman Saggere, and Joel Hetrick at the
University of Michigan.
6.0 References
I. Ananthasuresh, G. K., and Kota, S., Designing Compliant Mechanisms, ASME Mechanical Engineering, November
1995, pp.93-96.
2. Ananthasuresh G.K., Saggere L., (Faculty Advisor S. Kota), "A Single-piece Compliant Stapler" Won First place at the
ASME Mechanisms Conference Design Competition, Minneapolis, September 1994.
3 . Ananthasuresh G.K., Kota S., Role of Compliance in the Design of MEMS, Proceedings of the 1996 ASME Design
Technical Conferences - Design for Manufacturability Conference, August, Irvine CA.
4. Frecker, M. I., Ananthasuresh, G. K., Nishiwaki, N., Kikuchi, N., and Kota, S., 1997: "Topological Synthesis of
Compliant Mechanisms Using Multi-Criteria Optimization". Journal of Mechanical Design, Transactions of the ASME,
Vol. 1 19 No. 2, June 1997, pp. 238-245.
5. Bendsoe, M.P., and Kikuchi, N., 1988, "Generating Optimal Topologies in Structural Design Using a Homogenization
Method," Computer Methods in Applied Mechanics and Engineering, Vol. 71, pp 197-224, 1988.
6. Joo , J., Kota S., Kikuchi N., Topological Synthesis of Compliant Mechanisms using Beam Elements, submitted for
review, Journal of Materials and Mechanics.
7. Hetrick J., and Kota S., Size and Shape Optimization of Compliant Mechanisms, Proc. of the 1998 ASME Design
Technical Conferences, Atlanta, September.
8. Saggere L, Kota S., Crary S., " A New Suspension for Micro Linear Actuators", 1994 ASME Winter Annual Meeting -
Symposium on MEMS.
9. Kota S., Ananthasuresh, G.K., Crary S.B., Wise, K.D., Design and Fabrication of Microelectromechanical Systems,
Journal of Mechanical Design, Vol. 16, No. 4, pp 108 1-1088.
53
12. Griffis, M., and Duffy, J., 1991, "Comparing Structures of Stiffness Matrices Using Invariants," Proc. of the VII CISM-
IFToMM
Symposium, pp.85-92.
13 . Pigoski, T.M., and Duffy, J., 1995, "An Inverse Force Analysis of a Planar Two-Spring System, Journal of Mechanical
Design, Vol. 1 17, Dec. 1995, pp 548-553.
14. Yin, J.P., Matthew, G.K., and Duffy, J., 1996, "The Inverse Force Analysis of a Four-bar Mechanism with Compliant
Input and Output Links," Proc. (CD-ROM) ofthe 1996 ASME Design Engineering Technical Conferences and Computers
in Engineering Conference, August 12-22, Irvine, CA, Paper No. 96-DETC/MECH-1 141.
15. Ananthasuresh, G. K., Kota, S., and Gianchandani, Y., 1993: "Systematic Synthesis of Microcompliant Mechanisms -
Preliminary Results'. Proceedings ofthe Third National Conference on Applied Mechanisms and Robotics, Nov. 8-10,
1993, Cincinnati, Ohio, Vol. 2, Paper 82.
16. Ananthasuresh, G. K., Kota, S., and Gianchandani, Y., 1994a: 'A Methodical Approach to the Synthesis of Micro
Compliant Mechanisms". Technical Digest, Solid-State Sensor and Actuator Workshop, June 13-16, 1994, Hilton Head
Island, South Carolina, pp. 189-192.
17 Frecker, MI., 1997, "Optimal Design of Compliant Mechanisms,"Ph.D. Dissertation, University of Michigan (Co-
Advisors: N. Kikuchi and S. Kota).
18. Sigmund, 0., 1997,"On the Design of Compliant Mechanisms Using Topology Optimization," Mechanics of Structures
and Mechanics, Vol. 25, No. 4,pp. 493-524, 1997.
19. Larsen U.D., Sigmund, 0., and Bouwstra , S., 1997 " Design and Fabrication of Compliant Micromechanisms and
Structures with Negative Poisson's Ratio," Proc. of the Ninth IEEE International Workshop on Micro-Electromechanical
Systems, San Diego, CA pp. 365-392.
54