Contractor: Stanford University
Agreement Number: F30602-96-2-0308
Project Number: HJ1500-3221-0599
DARPA Order Number: N-7-0504/AO E117/12
Contract Period: September 13, 1996 - September 12, 1999
Date: February 9, 1998
Sponsor: Department of the Air Force
Technical POC: Robert W. Dutton Report
Prepared by: Robert W. Dutton
APPROACH:
a) enhanced gridding capabilities compatible with VLSI TCAD.
b) object-oriented design of layered access to model definition that
allows fast prototyping of new and realistic physical and constitutive
models.
c) enhanced element technology for discretization of multi-physics
system of equations.
d) fully-integrated MEMS fabrication and device simulation with detailed
material and stress analyses monitored by the test MEMS structure
discussed below.
e) development of lumped models for specific MEMS applications of
interest to DARPA.
PROGRESS:
Important multi-physics interactions and numerical requirements that need to be addressed in dynamic response simulation of MEMS systems are a key focus of prototype simulation development efforts. A simulation methodology has been devised (to be completed in stages) and outlined in the last quarterly report. Current progress in selected aspects of the multi-stage approach are now highlighted:
Last quarter we reported progress in demonstration of a versatile framework to test MEMS requirements for solving coupled systems of equations using ProPHLEX (from COMCO) as a prototyping vehicle. The solution of coupled elastostatic-electrostatic equations was demonstrated based on a relaxation scheme and ongoing work is targeted to exploit hp-adaptivity using gradient based error predictors. A key challenge is initial capture of MEMS structures in a flexible way that supports geometric transformation/simplification.
Major progress has been made to improve solid modeling and automation of mesh generation tools in support of the simplified process specification of MEMS, as suggested above. SHAPES was integrated as the solid modeler in a prototype tool that automates translation of process specification (i.e. process flow input and CIF) into 3D models. VRML is under serious investigation as a portable geometry format. For example, translators to VRML that work for two different commercial geometry modelers--ACIS and SHAPES--have been demonstrated at Stanford. This makes possible web-based (and remote) geometry creation of MEMS. A sample VLSI SRAM device is now accessible at http://www-tcad.stanford.edu/tcad/SRAM/
A further important aspect of this work is the potential impact in support of interoperability. As indicated above, we are actively exploring the definition of a standard process flow using a simple spreadsheet (i.e.Tanner-like) approach. Moreover, the definition of a minimal representation of solid geometry base on simplexes, using VRML as a de facto standard, shows great promise and growing interest among the other groups in the Composite CAD project. Among the commercial CAD community, AnisE (anisotropic etch) from IntelliSense can export structures as VRML files--an important "existence proof" for utility and acceptance on the industrial side.
In the area of development of an inverse modeling technique for MEMS parameter extraction based on C-V measurements, 2D fully-coupled electro-mechanical simulations of clamped-clamped beams using ABAQUS (with the electrostatic force applied as a user-defined load) have been performed. Simulations include: contacts, the effect of compliant supports, and post-buckling. This has made possible the study of the entire displacement-voltage characteristics instead of simply studying the pull-in voltage alone.
This demonstration of computational modeling in support of the C-V test structure analysis is sufficiently accurate for planar MEMS structures, as well as much faster (i.e. 3 orders of magnitude faster) than generalized 3D MEMS simulators, that it allows for the use of the simulator within an optimizer to extract device parameters.
b) Demonstrate meshing capability for MEMS utilizing the link between the solid modeler (SHAPES by XOX) and automatic mesh generation (SCOREC by RPI). Push the capabilities of solid modeling further in terms of providing realistic details through the incorporation of 3D process simulation results (level-set) and geometric manipulations of 2D simulations that are critical for deposition and etching steps.
c) Evaluate the confidence intervals of C-V parameters that we extract using optimization techniques. Investigate weighting schemes used in computing residual errors, and noise sensitivity in optimization schemes.
d) Compare to capacitance-voltage measurements of MUMPs clamped-clamped beam structures.
e) Perform measurements in vacuum to eliminate influence of stiction due to humidity. Explore possibility of using this as a technique to measure stiction.
Develop extraction schemes for materials properties using MEMS test structures, in combination with the CAD tools developed above, to support predictive modeling. The scope of material characterization includes the use of multilayer and graded structures, and reliability issues such as plastic yield and fatigue. This closed-loop process of test structures, simulation and extraction/validation of models is the key link between groups in CAD/Material Science/MEMS developers.
APPROACH:
a) mechanical characterization of MEMS materials with dependencies on
processing and microstructure.
b) reliability study including fracture and fatigue.
c) feedback to the constitutive models in the FEM-based solver in Task 1.
PROGRESS:
Microtensile testing has been performed on the first set of 2 micron thick aluminum fatigue test structures. Maximum loads of 117 MPa and strains at fracture of 3.5% have been observed. Drawbacks in the fatigue test structure design have been alleviated in a second generation mask set. With this new mask set backside alignment is made easier and the width of the Si support which is needed for sample mounting has been decreased to minimize vibrations during cutting. The first (manual) large-strain fatigue tests on the piezo-driven testing equipment have been performed with the improved fatigue test samples. We are in the process of automating the tests, developing customized scripts and hardware for use with LabView.
If successful with the new test structures and measurement approach, it will be possible to apply controlled, arbitrary tensile strains to thin metal beams and to monitor the mechanical response of the beams by measuring the load. Achieving controllable strain rates over a large number of load cycles is an additional advantage of this work. The extraction of these characteristics has not been achieved previously with other instruments or extraction techniques. With these new experiments, we will be able to assess whether MEMS beams survive under high cycle loading simply because they are loaded and unloaded quickly (essentially outpacing the plasticity mechanisms), or if there is something more intrinsic to the scale and microstructure. Finally, if there are scale and microstructure issues (as we expect), we will be well positioned to characterize and understand them.
Initial work on design of test structures for high cycle, displacement controlled, massively parallel testing of thin metal films with electrostatic actuation and in-situ sensing of failure has been started. These structures will complement the single-beam tests by providing statistical failure data, but under less well controlled conditions.
Work on the in-situ stress monitoring device has resumed. Our initial design includes a deflecting membrane, capacitive deflection electrodes, and a tunneling tip for high-precision membrane displacement measurements. This design is based on principles exploited by a MEMS pressure sensor developed by Professor Tom Kenny (Stanford). He has agreed to provide fabricated structures that we will modify for proof-of-concept ex-situ and in-situ trials. Modeling shows that the modified devices may be suitable for detection of stresses from approximately +/- 100 MPa. We have also identified key issues that must be addressed in development of the new stress measurement device: thermal noise and drift, on-wafer feedback circuitry, and compatibility with deposition equipment.
b) Begin modification of sensor structures and develop packaging scheme.
c) Conduct initial ex-situ trials of stress measurement device.
Design and objective implementation of MEMS structures for in-situ materials parameter extraction (mentioned above), which will be applicable to MEMS devices of interest to DARPA. The experimental measurements made with these structures will facilitate the critical evaluation of models, parameters and overall simulation accuracy for MEMS devices. The geometry and process flow design will support not only the application side (such as micromachined RF switches) but also the underlying need to quantify and understand MEMS device limits resulting from materials and process dependences.
APPROACH:
a) design and implementation of in-situ test structures.
b) design and implementation of example prototype MEMS device.
c) measurement, parameter extraction, modeling and testing.
PROGRESS:
Fabrication of the test structure array continues to progress although slowly due to equipment availability. The first attempt at releasing the structures was not successful and we are currently reviewing the fabrication process to determine the cause. There were unexpected polyimide materials and process dependences for the release layer that have required re-work. During the next few weeks we expect to overcome these difficulties. The related challenge is the measurement and extraction of relevant data. Fortunately, the requested measurement equipment procurement has now been approved and the supplier will provide immediate access to equipment in the interim. Ultimately, the test structures and extracted data will provide essential feedback to the modeling efforts discussed above.
In the process of preparing for measurements on the Stanford test structures (as well as those that are being fabricated using the MUMPS process), capacitance-voltage characteristics have been simulated. The results compare well to initial measured characteristics and give further insight about possible physical mechanisms that relate to materials and processing conditions. A model for stiction is now being developed that will account for an additional sharp transition along the release path that is not due to solid mechanics and electrostatics only. Interferometric measurements will provide more direct data for parameter extraction. We intend to quantify the reliability of the parameters we can extract from both displacement-voltage data and capacitance-voltage data.
With test structures available, we can proceed with the process that we have anticipated. This will include the visual indication of stress condition with the buckling structures. We will also be able to generate SEMS of structures that have a measurable displacement. Finally, we intend to measure the properties of the structures we can actuate to correlate the results from each method. The status of the interferometer system is still pending and prevents us from pursuing another, unique, measurement avenue.
b) Complete fabrication of test array and begin measurement.
c) Procure and install interferometric measurement system.
d) Make interferometric measurements of displacement as a function of applied voltage of MUMPs structures.
Active dialog between the ACADEMIA project and commercial suppliers within Composite CAD is focusing on three main aspects that will directly leverage technology transition: