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
Sponsor: Department of the Air Force
Technical POC: Robert W. Dutton
Develop and demonstrate fully-integrated, FEM-based prototype solver capabilities to model behavior and fabrication process-dependency of MEMS devices. This will include capabilities to consider multi-physics and materials dependencies as well as other process induced factors--geometry effects due to deposition/etching. The overall tool integration strategy will be to develop and test key components that overcome limits currently seen to be "bottlenecks" in commercial systems. Application-specific lumped modeling will be developed and used to guide the overall direction of the CAD efforts, including improved parameter extraction schemes that cross-link device performance with layout parameters that can in turn support design (and hopefully optimization) of practical MEMS devices (i.e., RF switch).
Approach:
Benchmarking and Parameter Extraction: During the last report we proposed to investigate weighting schemes used in computing residual errors, and noise sensitivity in optimization schemes for prototyping the RF switch (and its analysis using the C-V test structure). We found that automatic optimizers were unable to come up with the correct parameter values, even on tests done on synthetic (simulated) data. Thus we simulated the entire parameter space exhaustively and determined the precision of our extracted values as functions of the precision of our measurements.
In the process of performing detailed capacitance-voltage measurements of MUMPs clamped-clamped beam structures, a number of important details that influence modeling and characterization were identified. Measurements are greatly influenced by charge in the nitride and stiction. A "push-down" method (using a probe tip to push the beam down mechanically) for measuring the capacitance alleviates some of these problems. We have used the beams as charge measurement devices and fit the charge buildup to a quantitative model.
Another set of measurements indicate that the step up supports in the MUMPS process might be nonlinear---stiffer to upward rotations than to downward rotations. Also, stress gradients vary even for cantilever beams of the same length located near each other. Investigations are underway to characterize boundary conditions more carefully and to correlate the effect of sacrificial etching on stress gradients.
We initially planned to perform measurements in vacuum to eliminate influence of stiction due to humidity. The samples are now being prepared for wirebonding that will allow such measurements to be made later this year (samples to become available in August-September 1998).
Software Interoperability: Continued efforts on interoperability for process flow and solid model representation have focused on prototyping a system that exploits the "best-in-class" tools and interfaces currently available. We have demonstrated a prototype that includes:
In the area of specific tool developments, we have completed a new version of the level-set code, integrated in support of specific process modeling tools for MEMS (i.e., SPEEDIE), that is much faster and more robust. Combining this physical simulation (for deposition and etching) with solid modeling (i.e., SHAPES), realistic geometries for MEMS devices have been demonstrated [1]. This prototype is now being used to generate 3D geometries for RF switch structures that are proposed as canonical test structures for characterizing MEMS materials and processes.
In support of meshing for MEMS, we are now investigating the link between solid modeling (i.e., SHAPES) and automatic mesh generation---provided by the SCOREC (RPI) tool. To date we have used an interactive package (i.e., HyperMesh) that supports volume meshing for finite-element analysis (using ProPHLEX). This methodology is very similar to that used by several others in the DARPA Composite CAD community. While this approach provides for problem-specific meshing in the hands of an experienced analyst, it is a laborious task which is a source of large potential numerical error in the hands of a designer. This motivates the investigation of automatic mesh generation which is considered essential for complex geometry (e.g. geometry from SPEEDIE simulations or experimental data).
Difficulties/Problems:
Goals targeted for the next quarter:
Reference:
[1] N.M. Wilson, Z.K. Hsiau, R.W. Dutton and P.M. Pinsky, ``A Heterogeneous Environment for Computational Prototyping and Simulation Based Design of MEMS Devices,'' SISPAD '98, September 2-4, 1998.
Develop extraction schemes for materials properties using simplified MEMS test structures. In combination with the CAD tools efforts, the physical models (including microstructure) will be evaluated and modified as needed. The scope of material characterization efforts include both standard metal layers (possibly graded as well) and reliability issues such as plastic yield and fatigue. The extraction and validation steps for physical models is a key linking between the three sub-tasks.
Approach:
Fatigue Study: LabView software has been updated to perform experimental control and data acquisition for automated fatigue testing. Low cycle fatigue tests have been performed using 2 micron thick Al beams. These were carried out using a multiple step test method, in which each specimen was cycled with several blocks of constant maximum displacement amplitude. A load drop and its corresponding stabilized state were observed for each displacement block. Evidence of cyclic creep was observed; SEM images revealed fatigue markings on cyclically loaded samples but not on monotonically loaded beams.
Microtensile tests at different strain rates have been performed (0.0001/sec to 0.1/sec). As expected, no significant strain rate dependence of the yield strength has been observed. This indicates that increasing the cyclic test frequency up to 1000 Hz will most likely not result in a change in plastic flow mechanisms. Therefore testing in this frequency regime should yield results which are comparable to the low cycle fatigue results.
After several months of testing on the "proof of concept" tensile test tool, we are finalizing design for a specialized fatigue rig. Orders for components have been placed. Testing on the old tool will continue until the new rig is fully available (Q4).
In keeping with the planned schedule, we have begun to investigate possible constitutive models for static and cyclic creep. This effort will also cross-couple with the CAD modeling task to consider implications to more accurate (detailed) modeling of failure mechanisms. Data from Q3 and Q4 testing will be used to extract model parameters.
In-situ Stress Measurement: Sample IR detectors have been fabricated and are now being diced at IC Sensors. The IR sensors will be used for initial tests of the stress measurement concept over a limited range of stress and temperature.
Plans for fabrication of modified (first-generation) in-situ stress sensors have been made. During the coming quarter we will begin building our devices at CIS using selected elements from the IR sensor masks. A variety of sample geometries will be included. For example, we will be able to vary the gap between the deflection electrodes (by controlling the etching process) to select possible stress measurement ranges.
A packaging scheme has been developed in which the sensor die will be wirebonded in a standard IC carrier suitable for use inside the deposition chamber. Feedback circuitry will be external. Approximately four conductors will be needed to electrically connect the device to voltage supplies and readout apparatus from outside the deposition chamber. Detailed issues are still under study.
Difficulties/Problems:
[2] G. Cornella, R. Vinci, J. Bravman, "Observations of Fatigue in Al Microbeams for MEMS Applications," in preparation for MRS Spring Symposium, April 1998.
Design and implementation of MEMS structures, including materials parameter extraction, will be used to test the CAD and physical models based on MEMS devices of interest to DARPA. The experimental measurements made with these structures will facilitate the critical evaluation of models, physical parameters and overall simulation accuracy of CAD for MEMS devices. The test vehicles from this work also support canonical benchmarking of both new materials for MEMS and accuracy of Composite CAD in MEMS applications, specifically the RF switch.
Approach:
Progress:
The processing challenges in the test structure fabrication have caused us to carefully consider critical steps in the process that induce thermal gradients. Thin films of Al that have been deposited and patterned with the Testarosa mask set, have exhibited extreme deformations in their released test structures. The large scale deformations of the released thin film structures seem to have been caused by the heating and ion bombardment that the wafers are subjected to during the release step of the process. However, one wafer was completed with test devices that could be tested sufficiently to demonstrate "proof of concept" for Testarosa. With samples from this wafer, we have been able to extract experimental data of the deflection of doubly-supported beams as a function of applied voltage. These same test patterns will to be used with alternative metal structures for MEMS--Aluminum is certainly far from ideal--and progress in this area is discussed below.
Modest changes in the methods and materials that are used to fabricate the thin film test structures are also being investigated. The polyimide sacrificial layer will continue to be used for now, but other materials are being considered. Our first modification is based on previous work done at Stanford, where it was shown that the structure can be strengthened using multi- layered thin films instead of a single layer.
This process uses 4 layers, beginning with a sputtered Si film, that seals the polyimide surface. Following this layer, the metal film to be analyzed is deposited. An additional sputtered Si layer is added to protect this metal layer and a final structural support layer of thick Al is deposited and patterned to enhance the overall structural integrity of the test devices. The second approach to be tried is a two layered film structure that uses Ti/W and silicon nitride. The silicon nitride has been studied at Stanford for many years and it's mechanical properties are well understood. The problem with silicon nitride is that it is an insulator and most of the test structures require a conductive film. To work around this a thin metal film is first deposited to make contact with the rest of the circuit and the silicon nitride is laid down on top. Both of these approaches are currently under way and results will be reported next time.
Difficulties/Problems:
Our first insight into the issues plaguing the Testarosa process flow is that the materials are very sensitive to temperature excursions, particularly for aluminum. To improve the overall process flow, inspite of the known limitations of the Al layers, several process changes are now being considered. Modifications in the plasma dry etch release process are being tested that would minimize the affect this step has on the film and enable us to get an unambiguous measure of the Al mechanical properties. Experiments are being carried out at reduced RF power levels to determine if we can mitigate the damage effects in this way. A grounded screen may also be placed above the chips during the release process to block ions that induce damage and heating. Finally, we will test pulsing of the plasma periodically between etching cycles to allow the samples to cool and minimize thermal excusions.
Goals targeted for the next quarter:
In the area of test structures focused at the application pull side, we have had ongoing discussions with both Captain Rob Reid (AFB, Hanscomb) and the Hughes Malibu Research Center. In the case of the Air Force, they are interested in millimeter-wave applications-- phase-shifters and switches in support of complete RF systems. The Hughes group is looking at GaAs-based design of advanced antenna concepts. They have expertise in and equipment for high frequency testing and are looking for collaborations that can provide help and prototype access to silicon-based MEMS devices that would be suitable for possible flip-chip integration with GaAs.