Mechanical Engineering Laboratory,
National Research Laboratory of Metrology and the
Electrotechnical Laboratory (MEL/ETL)
Date Visited: September 28, 1993
Report Author: K.D. Wise
Dr. Kenichi Matsuno Director-General, Mechanical Engineering
Dr. Kunikatsu Takase Director, Intelligent Systems Div., ETL/AIST/MITI
Dr. Yoshihisa Tanimura Director, Mechanical Metrology Department, Natl Research Laboratory of Metrology, AIST/MITI
Ryutaro Maeda Sr. Researcher, Mechanical Engineering Laboratory, AIST/MITI
Dr. Yuichi Ishikawa Sr. Research Officer, Design Engineering Div., MEL/AIST/MITI
Dr. Tatsuo Arai Director, Autonomous Machinery Division, Robotics Dept., MEL/AIST/MITI
Prior to the meeting, Mr. Ryutaro Maeda introduced the schedule for the visit. He mentioned that the Mechanical Engineering Laboratory (MEL) had prepared answers to the questions about MEMS that were raised in preparation for the visit, and provided the JTEC panel with written copies of the answers. Copies of the questions and answers are attached, and include definitions of MEMS and of micromachines.
Dr. Kenichi Matsuno opened the meeting by giving an overview of work in MEL. He pointed out that the overall effort involves 254 personnel, including 207 professional scientists and engineers. This number has been slowly declining due to budgetary constraints, decreasing from 279 in 1989 through attrition. MITI's Micromachine Technology Project is one of eleven Designated Research Projects under the Industrial Science and Technology Frontier Program. These are funded at twice the rate of the Special Research Programs listed as well as a total of forty-six separate Ordinary Research Programs currently under way. The research topics under MEL span a very broad range, dealing with subjects such as energy generation (under the New Sunshine Program), underground space development, air pollution, robotics, properties of materials, bioengineering, and human factors. The total budget for MEL in 1993 is about ¥1.6 billion, of which the eleven programs under the IST Frontier effort are funded at about ¥262 million, or about 16.4 percent of the total for MEL.
Research efforts in micromachine technology in the MEL are focused on developing design and evaluation mechanisms for micromachines, particularly those having moving mechanisms. This involves efforts on complex microprocessing for 3-D structures, including complex micromachining, material modification, and wafer-level bonding. It also involves efforts on micromechanisms (mechanical properties of materials and microtribology) and on micromachine control structures. The micromachining efforts involved microgrinding to dimensions of 50 mm and combinations of microgrinding and electrochemical etching. The laboratory representatives also mentioned efforts to bond ceramics to metals and silicon to metals using ultrahigh vacuum technology, with ion cleaning of the surface followed by pressure-assisted bonding at room temperature. They are exploring the use of focused ion beam technology for microfabrication, and on the evaluation of the mechanical properties of the resulting microstructures. Goals are to modify the mechanical properties (e.g., elasticity) of the materials in favorable ways as well as their etching behavior in order to allow the creation of new structures. Piezoelectrically-actuated microgrippers have been realized for use in microassembly operations. Teleoperated devices have been constructed based on discrete piezoelectric drive elements capable of six degree-of-freedom operation (Arai, Larsonneur, and Yaya 1993; Arai and Stoughton 1992). Strain gauges are mounted on the individual elements and the actuation is keyed off of the sensor outputs (the actuators are slaved to the sensor drive signals). Panel members saw a demonstration of this system in the laboratory; it allowed the positioning of a needle tip with submicron accuracy and a range of tens of microns.
A brief review of work in the ETL was given by Dr. Kunikatsu Takase. Much of this work was aimed at merged robotic systems, where many small robots perform cooperative tasks to accomplish a larger effort. Specific topics included the development of technology for autonomous robot control, interactive man-machine interfaces, and distributed sensing systems, as well as the study of learning aspects of such systems (robot training and robots teaching robots). While robots are currently relatively large, the eventual devices are aimed at sizes comparable to a business card. The comment was made that if the devices were made much smaller than this, they would not be able to accomplish most real-world tasks, although for medical applications some structures would of course be still smaller. The applications in mind for much of this work appeared to be monitoring and small parts handling. There were also efforts on small micromachines. Stacked solid-state structures with 1 mm orifices containing electron emitters and grid overlays were shown with the intent of using these devices as end effectors for directed electron-discharge machining.
Dr. Yoshihisa Tanimura did not specifically describe the research activities in NRLM but laboratory representatives suggested discussing areas of mutual interest. There was considerable subsequent discussion of needs in MEMS and micromechanics. The laboratory representatives are interested in LIGA for applications in micromechanics, but felt that friction and microassembly of microcomponents were significant challenges. They felt that practical products with MEMS and micromachines are probably at least ten years out and did not seem very concerned about products within the next five years. They noted that their work was primarily focused on microactuators, which is primarily where the risks are: While useful sensors can certainly be realized, actuators are much less certain, and even though there are many possible drive mechanisms, they do not necessarily produce the desired effects and may be difficult to integrate. The researchers have a program in microinjection molding and are exploring its use in piezoelectric ceramic materials. On the subject of LIGA, they expressed concerns that this area was restricted by German patents and by additional work at Wisconsin in the United States. There was considerable discussion on this topic.
Panel members saw two projects in subsequent laboratory tours. The first involved work by Dr. Tatsuo Arai and others on the development of a dexterous micromanipulation system for use in microsurgery and microassembly. This was the piezoelectrically-driven structure mentioned above. Panel members first saw two macro versions of the system. The larger was several feet across and was aimed at construction applications (excavations). The smaller table-top device used feedback to accomplish tasks such as the insertion of a square peg in a square hole with very tight clearances. The microversion of this device used an end effector a few centimeters on a side positioning the tip of a needle a few centimeters long. They had not yet used the system for microassembly, but the overall feedback control was very nice and performed well.
The second project involved the use of ion implantation to modify materials. Specifically, the researchers used ion implantation to modify the elasticity of materials. They implanted 18 mm diameter wire to a depth of about 2 mm and studied its material properties. The shifts in Young's modulus were detectable; the effects on built-in stress and strain remain to be studied. The idea is to be able to use ion implantation to selectively tailor certain portions of a microstructure to produce flexible joints, where the structure itself was perhaps 2 mm thick. They showed one diagram of a silicon structure where ion implantation had been used with subsequent etching to give a void under an overhanging beam. The laboratory facilities were adequate for working with discrete hardware and advanced control systems. The JTEC panel did not see clean room facilities or any actual MEMS microfabrication.
The following questions were submitted to MEL and ETL prior to the panel's visit.
A. Advanced Materials and Process Technology
1. What materials and fabrication techniques are most likely to be used in production for MEMS? Will MEMS continue to be based primarily on silicon IC technology?
At present, Si-based technology plays a most important role in MEMS, but its role will decrease in the future. Expansion of available materials and process technology is [an] essential trend in MEMS or rather micromachine technology. We define "MEMS" as electromechanical systems fabricated by Si-based IC technology, and "micromachine" as [a] machine composed of microfunctional elements fabricated [with] various materials and process technology, which can do more complicated and precise work than conventional machines, for its more integrated structure.
2. What, in your opinion, is the most important new process or material needed for extending the capabilities of MEMS?
At this stage, this question is difficult to answer. In our project, the preceding four years' efforts will be focused on this issue. High-aspect-ratio fabrication processes, including etching and forming, seem important.
3. Are there plans to use X-ray lithography for micromechanics?
4. How attractive is the LIGA process for commercial use in MEMS at the present time? How attractive do you expect it to be in five years? Ten years?
Reported products by LIGA are not always considered to be fabricated by only LIGA, and to make matters worse, patent issues and secret guarantees are a hindrance to our access.
5. If you are pursuing LIGA and LIGA-like high-aspect-ratio structures, what materials are being investigated and why? To what extent do you feel that structures produced using high-aspect etching will be competitive with those formed by plating?
Piezoelectric ceramics microstructure for ultrasonic application [are under investigation]. The latter question is not clear enough to be answered.
6. What are the prospects for a low-temperature wafer-scale bonding process for MEMS? How low in temperature can we go? Do you feel metal-metal, silicon-glass, or other interfaces are most promising?
Surface-activated bonding and Si surface modified bonding are candidates for low-temperature bonding. Interfacing of Si/metal and Si/ceramic are attractive for application.
7. To what extent will silicon fusion play a role in MEMS? Is the area coverage sufficiently high? How much of a real problem is the high bonding temperature?
Si fusion bonding is applicable to Si/Si bonding and SOI. The polishing and thinning processes after bonding are important. The process temperature is so high that its application is limited and difficult for bonding of dissimilar materials.
8. What polymeric materials are being explored as photomasks for high-aspect-ratio structures? Is the use of conformal coating processes practical?
These items are not direct targets of R&D of this project.
9. Polymers for thick photoresist applications are required by MEMS. Are Japanese photoresist suppliers responsive to this need?
Such resist [applications are] necessary, but IC industry also needs [them]. It seems to be better to wait for the development of IC.
10. Do you feel that nested/stacked wafer-level microstructures based on multiple bonding and/or etch-back operations will be feasible within the next five years? Are they important for MEMS?
They are very important to MEMS. The simplest one is expected to be feasible in five years.
11. It appears that precision injection molding could play a major role in MEMS. Would you comment on this, please.
This is already established for polymers. Application on metals and ceramics has high priority in R&D.
12. Are room-temperature superconductors being explored, and if so, what materials and processing techniques are being used?
In another project, superconductivity is [being] investigated, [so] there would be no necessity for R&D.
13. Do you expect major advances in micromachining curing the coming decade? Will photo-assist etching or new etch-stops emerge to play a major role? What other technology additions do you consider promising?
More precise and higher rates are expected for RIE; laser-assisted chemical etching is also expected.
14. What are the most important attributes you would insist on for a MEMS processing tool?
We must eliminate exclusiveness (dust impurities, etc.) of MEMS tools. Expansion of treatable material and flexibility of the tool must be pursued.
B. Sensors and Sensing Microstructures
1. For which sensed variables and types of sensors will MEMS technology have the greatest importance? What are the key advantages of MEMS technology for these sensors?
Medical use (chemical sensor), industrial use (gas sensor, acceleration sensor, gyro sensor), ultrasonic sensor, microforce sensor. Compact, lightweight, disposable.
2. Approximately when will the new sensors under development today based on MEMS technology first appear in the market place?
Various kinds of chemical sensors and ultrasonic sensors.
3. How can MEMS technology be used to improve existing sensors? What sensor types are likely to be improved the most?
Reducing size and weight, distributed intelligent. Pressure and tactile sensors may be refined.
4. What new sensors can only be developed using MEMS technology? How does MEMS make these new sensors feasible?
Two-dimensional array sensor, intelligent sensor, etc.
5. To what degree is self-testing likely to be possible with sensors? Will this be a major role for MEMS technology?
The problem of self-testing rather belongs to the problem of circuit-diagnosis than that of MEMS.
6. After pressure sensors and accelerometers, what is the next major sensor based on MEMS that will be mass-produced in high volume?
Medical use chemical sensors
7. To what degree are feedback readout schemes likely to be important in sensors? Will these feedback schemes involve MEMS?
We cannot understand the meaning of feedback readout schemes and this question.
8. What are the principal problems in the use of scanning surface probes (e.g. tunneling current) as an approach to high-sensitivity sensor readout? Is this approach likely to find wide application?
Making probes acute.
C. Microactuators and Actuation Mechanisms
1. Can designs using arrays of microactuators achieve large and useful forces? Are there other devices similar to the large optical projection displays that have been reported that can be realized using MEMS technology?
Arrays of microactuators can generate [large amounts of] power and are hopeful.
2. Considering their importance to MEMS, in what order of importance would you place the following microactuation mechanisms: shape-memory alloys, electromagnetics, electrostatics, thermal bimorphs, piezoelectric bimorphs, piezoelectrics, electrostriction devices, [and] phase-change devices? How many of these do you expect will find commercial applications in high-volume products?
We are going to study and develop all the microactuation mechanisms [that] are mentioned in the question. It is difficult to order the importance since the extent of importance [is] different for the application.
SMA, piezoelectric, and electromagnetic are hopeful in commercial use.
3. What are the most reasonable candidates for prime movers for microactuation?
The electrostatic force and electromagnetic force.
4. To what extent can sticking problems due to surface forces be suppressed in microactuators? Will these problems seriously constrain the practical application of microactuators based on narrow gaps?
So far, sticking is not a very serious problem, except for [the] rotational actuator. The dust problem is much more important.
5. What is the most promising candidate for a microrelay? Where are such devices most likely to be used?
The meaning of the first question cannot be understood. One of the answers is thought to be high current signal processing.
6. What are the main design issues in microfluidic systems? Where will such systems find their primary application?
The transport problem in capillary. The design method considering viscosity.
The blood-collecting equipments.
7. What designs are most likely to be adopted for practical microvalves and micropumps? What is holding up the practical realization of these devices?
The phase transformation excited by laser.
The miniaturization is difficult because of actuator size.
D. Sensor-Circuit Integration and System Partitioning
1. Will MEMS technology lead to complete microminiature "instruments on a chip?" For what types of instruments?
It is difficult for the present. An instrument only having sensing function may be possible.
2. Will MEMS technology combine sensors and actuators into complete control systems? If so, what types of control systems will be most affected?
Of course, MEMS technology combine sensors and actuators into complete control system. We cannot understand [the] latter question.
3. Are systems involving sensors, actuators, and embedded microcontrollers likely to be realized in monolithic form or will they be hybrid? What factors are driving for monolithic integration for what types of products?
We think [the] system will be used hybrid form. But if it is possible to make consuming power small and actuator by only silicon process, the system will be used monolithic form.
4. What levels of integration (transistor count) are we likely to see on MEMS chips by 1995? By the year 2000? Are full microprocessors needed or realistic?
The level of integration of MEMS does not correspond to that of IC.
5. To what extent will embedded microcontrollers be used in (possibly hybrid) integrated sensing nodes -- high-end devices only, or will they eventually become pervasive in even low-end products?
It is necessary to integrate low-end products if [the] sensing system is complex.
6. What are the prospects for adopting sensor bus standards, at least in specific industries (automotive, process control, HVAC)? How important is the evolution of such standards to the development of MEMS?
We presuppose that adopting sensor bus standards will be demanded primarily in the automotive industry. To avoid confusion, it is necessary to adopt standards in early stage. But R&D and the adapting standards should be parallel done.
7. Will sensor calibration continue to be done in hardware (laser-trimming or EPROM) or will it evolve to digital compensation in software? Is this an important issue in MEMS development?
It is better to do calibration in hardware, but [it] may be possible in software. For the present this problem has not yet been important in MEMS.
8. Will a few standard processes emerge to dominate sensor-actuator-circuit integration or will widely divergent processes continue to be the norm in MEMS? What issues will determine the answer to this question?
It is better that a few standard processes emerge. This is determined by designer's concept.
E. Advanced Packaging, Microassembly, and Testing Technology
1. What are the general directions for progress in packaging MEMS? What are the principal challenges here?
Packageless technique is anticipated.
2. What will be the application and impact of MEMS on the packaging of sensors?
The question is not specified to be answered.
3. How much of a problem is die separation for surface micromachined devices? Is this optimally performed after release?
At the moment, the problem has not yet been serious.
4. How important are the chip-level packaging schemes now under development to MEMS devices and systems? Is the increased process complexity worth it?
Chip level packaging is not so important in R&D, because the effort is early at this basic stage. For increasing complexity, it is not worthy commercially.
5. Are common packaging approaches across many types of devices feasible or will packaging continue to be very application specific?
Basically packaging is thought to be [a] versatile technology.
6. What viable techniques exist for coupling force from integrated microactuators while protecting the device from hostile environments? Can mechanical microactuators only be used inside hermetic packages?
Sealing and force transmission. There would be some mechanical actuators without [the necessity of] packaging.
7. To what degree should MEMS continue to focus on monolithic silicon microstructures and to what degree do you think it is really better suited to milliscale integration with components combined using microassembly techniques?
Sensing devices can be fabricated into monolithic [silicon microstructures], but the system with advanced functions must be fabricated by microassembly.
8. Are there microassembly techniques that can be sufficiently automated to make millielectromechanical devices in high volume at moderate or low cost? What are the generic barriers to such devices and techniques?
Automation is possible, in principle, but [is] commercially difficult for production diversity.
9. What techniques are available for testing MEMS structures after encapsulation/packaging? What happens when they are no longer viewable?
The system must be handled as a black box; viewability is favorable in [the] R&D stage.
F. MEMS Design Techniques, Applications, and Infrastructure
1. The term "MEMS" has many meanings. Could you tell us your interpretation?
We think that MEMS means microelectromechanical systems manufactured by IC processing technology and mainly made of Si.
2. Is there a MEMS technology driver equivalent to the DRAM in the IC industry? If so, what is it?
An integrated sensing and actuating device for automobiles, process controls and so on.
3. Is the integrated-circuit industry the principal application driver for MEMS? If so, what are the alternative drivers during the next decade, if any?
The integrated-circuit industry cannot be recognized [as] the principal application driver for MEMS, though it has seeds for MEMS. In the near future, the principal application drivers [will be] the automobile and medical industries.
4. In what sensor application areas do you see MEMS technology having the greatest importance? (Examples might be: automotive, medical, robotics, consumer products, etc.) What are the key advantages of MEMS technology for these applications?
[MEMS technology will be important in the areas of] automobiles, medical instruments, process controllers, etc. The advantages of MEMS are that they have small size, intelligence, disposability, low cost, etc.
5. What will be the application and impact of MEMS on the interfacing of sensors to the environment?
Since MEMS have small size, not only [can the] number of sensing points be increased, but also they can approach the target point.
6. Looking ahead five years, what new MEMS-based sensors and sensor applications do you anticipate?
7. In what time frame do you anticipate MEMS technology having the most impact on sensor products (for example, three years, five years, ten years, etc.).
8. What is the prognosis for MEMS foundries? Are they needed? What technologies would need to be present in a MEMS foundry?
What are "MEMS foundries?"
9. What is the state of the art in MEMS reliability? Where are the principal problems?
Now MEMS are on the R&D, so they probably have poor reliability. The R&D of estimation technology is important.
10. What is the "state of the art" in MEMS designability? How important are CAD tools for MEMS? Do you have an active CAD effort for MEMS in your organization?
There is a CAD system developed by modifying a CAD system for IC design and IC process simulator, but there is not a CAD system for micromechanism or microstructure. MEMS CAD is not investigated, but [the] design concept of micromechanism is studied.
11. What priority would you place on the creation of a central MEMS database of material and design parameters? Does your organization have an active project in this area?
We place high priority on the creation of a MEMS database, but we do not have enough researchers to execute any active project.
12. How much funding is being directed into MEMS research and development in your organization? What percentage is this amount of the total R&D budget?
13. In your view, is enough funding available to support MEMS research in Japan? In the world? Are the principal bottlenecks to more rapid progress money or ideas?
In Japan, [MEMS R&D] should be funded three times as much as it is now. In the USA, [many] more companies should execute [MEMS R&D]. Money occupies 70 percent of the principal bottlenecks to more rapid progress.
14. What portion of your R&D funding for MEMS comes from internal, government or other sources? Has increased government funding had a major impact in your organization?
Government funding has a major impact on the development of microactuators, because it has a high risk for companies. They develop microsensors from funds on hand because sensors can be commercialized in the near future.
15. What percentage of your research is directed toward specific products? Toward basic research?
In the first half of the national R&D project, funds for basic research occupy the [larger] part.
16. What emerging technologies do you see as having the largest influence on MEMS and why (i.e., LIGA, superconductors, submicron ICs, piezoelectrics, thin-film magnetics, etc.)?
LIGA and LIGA-like high-aspect-ratio processing.
17. What are the principal barriers to the success of MEMS?
Finding various applications in many fields (for example: DRAM for IC technology).
Molecular machines (nanomachines) and improvement of various manufacturing technology.
19. Are patents a key to commercial success or are base technology skills more critical?
There are both species of patent.
20. The first generation of university graduates who specialized in MEMS are now finishing their degrees. Are employment opportunities for these students plentiful in Japan?
Because of [the] recession, employment opportunities are not plentiful for such students, but there are not enough researchers in this field.
Arai, T., R. Larsonneur, and Y.M. Yaya. 1993. "Basic Motion of a Micro Hand Module." Proc. 1993 JSME Int. Conf. on Adv. Mechatronics. Pp. 92-97.
Arai, T., and R. Stoughton. 1992. "Micro Hand Module Using Parallel Link Mechanism." Proc. Japan-USA Symposium on Flexible Automation. ASME Book No. 10338A. Pp. 163-157.