Electromagnetic Remote Control and Down Scaling Advantages and Examples for MOEMS

Gilbert Reyne* a,b, Lionel Houlet b, Yoshifumi Takahashi b,c, Tarik Bourouina b,d, Hiroyuki Fujita b a LEG, Laboratoire d’Electrotechnique, INPG, Grenoble, France;

b LIMMS/CNRS-IIS, University of Tokyo; c ANRITSU.Corp., Japan; d IEF, Institut d’Electronique Fondamentale, Orsay, Paris

ABSTRACT Electromagnetic actuation is applied to MOEMS for industrial applications requiring large and long-range forces. Advantages and drawbacks are outlined while down scaling laws are discussed. Technological improvements and new available materials bring to much smaller dimensions, the limit between Electromagnetism and Electrostatics. Remote control and bi-stability are unique characteristics of electromagnetic actuation. The importance of remote control is stressed so as to allow easy tests and optimisation with no technological compatibility problem. These advantages are illustrated on three different electromagnetic MOEMS, developed in LIMMS, a Franco-Japanese research Laboratory based in Tokyo. The first is a resonant 1D magnetic scanner, the second is a magnetic bi-stable matrix array of optical micro-switches and the last is a remarkable application of the properties of thick Magnetostrictive thin layers to a 2D scanner.

Keywords: Microstructures, Micro-magnetic structures, Bi-stability, Permalloy Thin films, Magnetostriction, Magnetic Scanners, Optical Cross connect

1. INTRODUCTION Several MOEMS (Optical Electro-Mechanical Systems) such as Optical Micro-Scanners and Matrix Micro-Switches for Optical Cross-Connect have been developed and successfully actuated by electromagnetic forces in the frame of LIMMS [1], [2], and [3]. This Franco-Japanese laboratory is based in Japan. It depends both on the University of Tokyo and the French CNRS (French National Centre for Scientific Research) and aims at the development of Micro-Mechanical Systems.

For Industrial applications, due to the use of low cost optics, large Mirrors as well as large displacements or rotation angles are often required. This implies large air-gaps and significant forces, as large silicon structures are comparatively thicker and heavier. These requirements can be fulfilled with long-range magnetic forces. Moreover, such actuation can easily be made compatible with existing MOEMS. Thus, advantages of Silicon Technology (reliability, mass parallel production, High Frequencies…) can be made compatible with a simple, low cost and reliable actuation. This excitation can be either a small ‘classical’ electromagnetic circuit (which acts on a magnetic layer attached or deposited on the MOEMS structure) or it can be micro-coils realised on an associated wafer.

Starting with the running projects, and actual data, advantages and drawbacks of electromagnetic actuation for micro-systems are outlined and discussed. Due to new materials and technologies, made available for MEMS as well as favourable downscaling new elements, electromagnetic actuation presents noticeable interest for industrial MOEMS.

* Gilbert.Reyne@leg.ensieg.inpg.fr; phone 33 476 826377; fax 33 476 826300; http://www.ensieg.inpg.fr/Localweb/Recherche/Leg/Leg.htm; LEG, Laboratoire d’Electrotechnique de Grenoble, ENSIEG, BP46, F38402 SMH, Grenoble, CEDEX, France. reyne@fujita3.iis.u-tokyo.ac.jp; phone: 81 3 5452 6036, fax 81 3 5452 6088 LIMMS/CNRS-IIS, University of Tokyo, 3rd div., 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505 JAPAN

2. DISADVANTAGES OF ELECTROMAGNETIC ACTUATION FOR MICROSYSTEMS Electrostatics has been, for years, the targeted solution to actuate MEMS. No other solution seemed competitive or even able to provide a solution for micro-actuators. This is due to the many advantages of Electrostatics in the very small world, together with its easy compatibility with the processes of microelectronics. By outlining the advantages of Electrostatics in these dimensions, the drawbacks of Electromagnetism will directly appear.

First, reduction scale laws usually show the advantage of down scaling electrostatic rather than electromagnetism [4], [5], [6], [7]. In the range of 2-10 micrometers, Pashen’s curve indicates an ever-increasing advantage [8]. The most efficient structures are obtained with a distance between the electrodes in the range of 1 to 10 µm. Hence, either the displacement is very limited (for variable capacitor type actuator) or clever ways must be used (horizontal or vertical comb drives, Scratch drive actuators….) to benefit from Pachen’s effect. On the opposite, down scaling assumptions for electromagnetism as, for instance, down scaling with a constant current density, would directly lead to an unfavourable and rapid decrease of the force with the dimension, than Electrostatics. Consequently, the large advantage of Electromagnetism at macro-scale would rapidly decrease with the dimension. Though, the limit dimension where Electrostatics becomes better than Electromagnetism is still unclear and depends on the assumptions, as discussed in part 3.

Another essential advantage of Electrostatics is that it is ‘naturally’ compatible with microelectronics technologies. Compatible in term of design and concept as just a sacrificial layer transforms a simple capacitor into an actuator. But what is even more important is that it is technologically compatible. Due to the heavy investments and large delays necessary to process Silicon, this compatibility also largely explains the emphasis on Electrostatics in the first years of the development of MEMS. Electrostatic actuation could be realised with the existing technologies and available machines, while no other type of actuation was achievable. Indeed, efficient electromagnetic actuation both implies conductive materials as copper and magnetic materials, which can be soft, hard or Magnetostrictive. And all this materials are to be available in thick films, typically in the range of 10 to 100 µm. Hence two disadvantages appear for Electromagnetism:

- The corresponding materials are to be developed in films for silicon technology with the appropriate thick technological processes and,

- Once the corresponding processes are available, the problem of pollution of apparatuses must be solved. This latter problem is now leading to the specific development of clean rooms dedicated to electromagnetic micro-systems either for research or industry. But this takes time. Once the appropriate machines are easily available, the arguments of the debate between Electrostatics and Electromagnetism are deeply changed.

The two main disadvantages developed above are either arguable (see part 3) or, on the way to be overcome. They are not intrinsic drawbacks of Electromagnetism but result from historical backgrounds. Though, Electromagnetism does present intrinsic drawbacks as developed bellow:

- Conductive materials with significant currents do present Joule losses that result in a pure loss of energy and corresponding thermal dissipation. The thermal problem is to be taken into consideration and it actually defines the limit of Electromagnetic MEMS in terms of achievable forces or performances.

- These losses also imply a significant power supply. Electrostatics requires a very low supply in terms or power with relatively high voltages (with all the related difficulties) and quite no currents. Electromagnetism requires the opposite: low voltages easily compatible with classical electronics but significant current and power.

- At least, for certain applications, the presence of magnetic field can be a problem in terms of ‘pollution’, human compatibility or electromagnetic compatibility.

3.ADVANTAGES OF ELECTROMAGNETIC ACTUATION AND DOWN SCALING As for industrial applications to MOEMS, due to low cost optics, large mirrors and displacements are required so as to deflect the light beam. Diameters of this light beam can be in the range of 5 mm. Though, for low cost production, reliability, high frequencies, optical Microsystems (MOEMS) are required. Corresponding requirements for actuation imply rather large and long-range forces so as to obtain rotations of a few tens of optical degrees or displacements in the range of 100 microns. The same dimensions characterise the output beam of most optical fibres. It results in an in-between to the macro-world, and 2-10 microns dimensions. This can be handled by electromagnetic actuation [9], [10], [11], [12]. Remote control, which is a unique characteristic of electromagnetism, allows actuating Microsystems with a silicon-based structure, which just reacts to an external magnetic field. This is a common characteristic of the three projects presented bellow. Hence, the system can be actuated without any supply to the silicon part and no resulting thermal losses. Then, the absence of electric connections is a positive point for packaging and even more for vacuum packaging. But not only, and the examples bellow, show that it can easily be adapted to an already existing silicon structure. It can then be used to test and optimise the behaviour in terms of dimensions, resonant frequencies, and robustness… Bi-stability can easily be achieved thanks to the use of magnets [13], [14], [15], [16], [17]. Magnets can either be included in the Microsystems or can be a large external magnet on which the microstructure is attached [18], [19], [20]. This is important for latching systems or for ‘memory effect’ and ‘security’ aspects (when keeping the position, even if the supply is deficient, is required). Moreover the possibility of latching leads to the fact that only a pulse of force is necessary to switch from one stable to another stable position. For such systems a pulse of current is sufficient and for evident thermal reasons, the possible current density achievable for such a pulse current, and so the forces, can be much higher than for a DC system [31]. Integration of the supply with micro-coils is also possible [21], [22], [23]. About down scaling laws: - As recalled in part 2, Electromagnetism for macro-systems is highly more efficient than electrostatics. - If one takes into account the real limits of electromagnetism in terms of current density and thermal dissipation of

losses, down scaling laws can be tremendously different. The same, if soft and/or hard magnetic materials are used when down scaling.

- For examples, down scaling laws for the interaction between magnets is exactly the same as for Electrostatics (and hard magnetic materials are to come soon to MEMS [18], [19], [20]). Hence, as a macro-system with magnets presents much higher volume forces, it will be the same for the corresponding micro-system. The volume density of forces for magnets increases as the inverse of the dimension ratio.

- If one considers down scaling, assuming thermal dissipation to be constant and proportional to the exchange surface, then, current density can increase as the inverse of the square ratio of the dimension. This brings down the limit of interest between Electromagnetism and Electrostatics to a much lower dimension.

- But, it is possible to go even further. Densities of a few thousand Ampere per square millimetre are currently reported and a DC current density of 10 kA/mm2 was measured in LIMMS on 3 micron-thick spiral micro-coils. This indicates that the current density can globally increase, roughly, in the inverse ratio of the dimension. This can be explained by the increase of the possible thermal losses due to silicon technology. Silicon is a good thermal-conductive material and the electric-conductive materials can be in direct contact and support higher temperatures. This results in an overall increase of the possible dissipated power and explains the very high current densities actually reported. With such a global down scaling law for the current density, the systems using currents, can experience down scaling laws, which are comparable to the ones with magnets or to the ones with Electrostatics.

Levels of magnetic induction and forces can, then, be similar to macroscopic ones and this usually leads to an increase of the force to volume ratio that can be obtained. Consequence is that such MAGMAS (MAGnetic Micro-Actuators and

SiPermalloy Pyrex

Gold Mirror (Chromium seed layer)

Systems) can, comparatively to the volumes and masses, produce more force than their macro-counterparts (that is already demonstrated in several silicon based micro-systems). If the electromagnetic micro-system developed takes benefit of these advantages, then the available forces are very large. In part 2, the advantage of the compatibility of Electrostatics with Electronics was stated and developed. But with the current passage of surface micro machining to volume micro machining (with typically the use of deep ICP-RIE), Electrostatics, now, also needs specific equipment. This 3D evolution is favourable to Electromagnetism. Indeed, both for thermal losses or saturation problems, thick films and 3D structures, are necessary. Hence, both types of actuation will require specific equipment. More and more, the advantage of Electrostatics in term of ‘easy available technology’ will decrease. All the above factors completely modify the downscaling laws and show that micro-technology and micro-systems are widely open for electromagnetic actuation. The three examples bellows are research performed in Fujita-lab, and LIMMS, University of Tokyo, by French and Japanese researchers together. They are direct illustrations of many points stated above.

4. REALISATION OF A MAGNETIC 1D SCANNER The research was performed in collaboration with an Anritsu Engineer, Yoshifumi Takahashi. The objective is to actuate a scanner to develop optical measuring systems (cf. Fig.1.). So as to reflect large size spots, required dimensions for the mirror were quite large, in the range of 0.5 cm2 or 50 mm2 (cf. Fig.2.).

Fig. 1. Principle of a 1D scanner: transform a light spot into a 1D scanning. Example of scanning obtained. The size of the mirror, combined with scanning optical angles of about 20 degrees is amazing requirements for a Microsystems in terms of actuation, frequency, amplitude range and quality factor. Hence, displacement is to reach some hundreds of micrometers (between 500 to 1000 µm).

Fig. 2. Cross-section of the scanner which rotates around a torsional through beam.

Scanning Micro-Mirror

Laser

Obtained scanning (Excitation magnétique 11.97kHz)

The electromagnetic actuation was manually added to the structure after its completion by silicon technology (cf. Fig.3.). The only modification is the gluing of a very small (1mm x 1mm x 100 µm) part of thin permalloy (easily available magnetic material in 100 µm thin sheets) (cf. Fig.4.). This was done manually but a parallel production is possible by electro-deposition.

Fig. 3. Finite Element Modelling with FLUX2D, Whole structure, Magnetic induction and related forced on the attached permalloy

part. (Www.cedrat.com)

Fig. 4. Scanners with the 100�P SHUPDOOR\ SDUW DGGHG� %RWWRP� 6WUXFWXUH SDWWHUQHG E\ ,&3 5,( WR GHFUHDVH WKH mass and increase the resonating frequency.

Then, a small, classical electromagnet, which easily fits in 1cm3 assures the field remote excitation. Hence, the total structure is a remote controlled MAGMAS (MAGnetic Micro-Actuator and Systems) with no power supplied to the Microsystems and remote excitation. Due to different objectives in terms of frequency and amplitude, different scanners were realised, optimised and tested. This could be done easily, thanks to that simple electromagnetic actuation. The range was from 1 kHz, 0,5 cm2 and 20 optical scanning degrees with a Q factor superior to 1000, to a frequency of 20 kHz for a 2x4 mm mirror and angles of only 1 to 5 degrees. For some scanners, the high Q factor required could only be achieved with vacuum operation. (Vacuum assured by the packaging).

Mirror matrix

Optic fibres

Micro-coils connections

5.REALISATION OF AN ARRAY OF OPTICAL MICROSWITCH WITH ELECTROMAGNETIC BISTABILITY. [1], [2], [24]

An array of volume micro-machined vertical mirrors on suspensions allows optical switching of fibres (cf. Fig.5.). The whole bulk structure obtained by ICP-RIE and KOH uses crystalline directions for the 45 degrees mirrors and V-grooves. Hence, due to perfect alignment and good quality mirror, attenuation between input and output is very low (0.5 dB). The out-of-the-plane displacement required is 100 microns. Similar technology was used as for the scanner but a magnet added for latching. A small electromagnet allowed easy testing, and optimisation and integration of a network of micro-coils is on the way (cf. Fig.6. and Fig.7.).

Fig. 5. Principle of the electromagnetic actuation of the optical matrix switches

Fig. 6. Mask for an 8x8 array of micro-coils and principle of association of the two wafers for actuation

Fig. 7. Matrix array of micro-coils and details of SU8 moulds.

B

Fig. 8.. FEM Magneto-mechanical Modelling with ANSYS and FLUX2D

Bi-stability was modelled and tested (cf. Fig.8.). Switching in just a few milliseconds means the use of a pulse current (cf. Fig.9.). Then, higher current densities can be reached, resulting in much higher pulse forces (the force is proportional to the square of induction).

B

Fig. 9. Principle of bi-stability

The straight line is the mechanical force due to the springs (here the cantilever). The middle hyperbolic curve corresponds to the magnetic force for the magnet only. The system presents two stable positions (one mechanical and one magnetic) when the two curves present two intersections. Then the two stable positions are point A and B. Switching is then either obtained by adding (top curve) or subtracting) bottom curve) to the induction created by the magnet, an induction due to the coils (pulse of current). The same principle applies to the micro-coils and matrix array, but is to be cleverly adapted.

6. REALISATION OF A 2D MAGNETOSTRICTIVE SCANNER [3]

In collaboration with Nissan, a 2D scanner to be used for automobile obstacle detection has been developed. It relies on the unique feature of magnetostrictive pre-oriented layers to simultaneously excite bending and torsion in a resonant bimorph. The actuation is added, after the completion of the structure by sputter deposition of a multi-layer giant Magnetostrictive material of a few micrometers thick. This material was deposited by E. Quandt and A. Ludwig in CAESAR, Germany [25], [26], [27]. Application of such materials is potentially very large [28], [29], [30]. Due again to Optics, the mirror size is 50 mm2 and scanning angles of a few degrees were achieved. An external AC field excites the structure. Adding a DC magnetic field with a magnet allows to easily modify the scanning window. The low field answer of the Magnetostrictive layer (5 to 10 mT), allows to actuate them with the use of micro-coils (cf. Fig.10, Fig.11, Fig.12 and Fig.13.).

Fig. 10. Large size Magnetostrictive scanner with distance field command

Fig. 11. Modification of scanning window shape by adding a DC field

Fig. 12. Thick electroplated micro-coils and associated mask

Fig. 13. Principe of the magnetic excitation of the Magnetostrictive layer with two opposite-winded micro-coils

Metallic mirror

Stiff mirror support Piezoresistive detectors

Magnetostrictive material

Si

Si

SiO2

Fig. 14. Silicon structure of the 2D scanner

Fig. 15. Examples of realised scanners with different dimensions for optimisation and Packaging.

7.CONCLUSION Three examples of MAGMAS applied to MOEMS, developed in LIMMS by a group of Japanese and French researchers have been presented. These examples show the good adaptation of Electromagnetism to industrial requirements, both in terms of performances (large dimensions, large and long-range forces…) and in term of easy adaptation to silicon processes for easy tests and optimisation or low cost actuation.

MStr. Material

Furthermore, the many possibilities in term of magnetic materials (soft, hard or Magnetostrictive materials…) have been illustrated. The possibility of latching is demonstrated. More important, a general presentation of the advantages and drawbacks of Electromagnetism versus Electrostatics for micro-systems shows that Electromagnetism is becoming more and more competitive in the either smaller dimension. This is due to different factors as the trend towards volume micro machining, the availability of new thick films magnetic materials and the modification of the Down Scaling laws due to the presence of Magnets or the possible spectacular increase in the current density. Hence, one can expect much more reports on MAGMAS (Magnetic Micro-Actuators and Systems) in the years ahead.

ACKNOWLEDGMENTS This project is performed in the framework of LIMMS, Laboratory for Integrated Micro-Mechatronic Systems, located at Tokyo. It is a joint Franco-Japanese laboratory between the French Centre of Scientific Research, CNRS and the Institute of Industrial Science of the University of Tokyo. The author would like to thank Fujita laboratory members. The 1D scanner could be achieved in collaboration with Y. Takahashi. The matrix switch project also implies the French researchers P. Helin, T, Bourouina and L. Houlet. The 2D Magnetostrictive scanner results also from common research together with A. Garnier, T. Bourouina, E. Lebrasseur and A. Debray.

REFERENCES 1. HELIN P., BOUROUINA T., FUJITA H., MAEKOBA H., CUGAT O., REYNE G., Self-Aligned Vertical Mirrors

and V-Grooves for Magnetic Micro Optical Matrix Switch, NMT’2000, Vol.1, No1, p. 55-87 (33 pages). “Microcapteurs et Microsystemes Integres “

2. HELIN P., MITA M., FUJITA H., Self-aligned vertical mirrors and v-grooves applied to a self-latching matrix switch for optical networks, Proc.IEEE MEMS 2000, 23-27 Jan. 2000, Miyazaki, Japan.

3. BOUROUINA T., LEBRASSEUR E., REYNE G., FUJITA H., MASUZAWA T., MURO H.,. LUDWIG A, QUANDT E., MURO H., TAKAHIKO O., ASAOKA A., A Novel Optical Scanner with Integrated Two-Dimensional Magnetostrictive Actuation and Two-dimensional Piezoresistive Detection, Proc. Transducers’01, Munich, Germany, June 10-14 (2001) 1328-1331

4. FLUITMAN J., GUCKEL H., Micro actuator principles, Actuator 96 Proc. pp. 23-28 5. MADOU M., Fundamentals of microfabrication, ISBN: 0-8493-9451-1, CRC Press LCC, 1997, 589 p. 6. HAMEYER K., BELMANS R., Design of very small Electromagnetic and Electrostatic Micro Actuators,

Actuators 98 pp. 580-583 7. WOOD D., BURDESS J., HARRIS A., Actuators and their mechanism in microengineering, Eng.Science and

Educ.J. 98 pp. 19-27 8. PASCHEN F., Über die zum Funkenübergang in Luft, Wasserstoff und Kohlensäure bei verschiedene Drucken

erforderliche Potentialdifferenze, Annalen der Physik 37 (1889), pp. 69-96 9. GUCKEL H., FISCHER K., STIERS E., Closed loop controlled, large throw, magnetic linear microactuator with

1000 /spl mu/m structural height,. MEMS 98. Proc., The 11th Annual International Workshop, pp. 414-418 10. CUGAT O., BASROUR S., DIVOUX C., MOUNAIX P., REYNE G., Deformable Magnetic Mirror for Adaptive

Optics, Technological Aspects, Sensors and Actuators A: Physical, Volume 89, Issues 1-2, 20 March 2001, Pages 1-9

11. ROGGE B., SCHULZ J., MOHR J., THOMMES A., Magnetic Microactuators fabricated by the LIGA-technique for Large Displacements or Large Forces, Actuator 96 pp. 112-115

12. NAKAZAWA H., WATANABE Y., MORITA O., EDO M., The two-dimensional micro conveyer, Principles and fabrication process of the actuator, Proc. Transducers, pp.33-36, 1997

13. RUAN M., SHEN J., WHEELER C.B., Latching Micro Magnetic Relays with Multistrip Permalloy Cantilevers, IEEE MEMS 2000, 23-27 Jan. 2000, Miyazaki, Japan.

14. TABAT N., KLEIN J., GUCKEL H., Single flux-path bidirectional linear actuators,. TRANSDUCERS '97, Vol. 2 , 1997, pp. 789 –792

15. MAEKOBA H., HELIN P., REYNE G., BOUROUINA T., FUJITA H., Modeling and Optimization of Bi-stable Optical Switch, Proc.of MSM2K, San Diego, March 00

16. YI Y.W., LIU C., Magnetic actuation of hinged microstructures, IEEE Journal of Microelectromechanical Systems, Vol.8, No.1, 1999

17. SHINOZAWA Y., ABE T., KONDO T., A proportional microvalve using a bi-stable magnetic actuator, IEEE 1997 0-7803-3744-1/97 pp 223-237

18. LINETSKY Y.L., KORNILOV N.V., Structure and magnetic properties of sputtered NdFeB alloys, J. Mater. Eng. Perform.4–95. Pp.188-195

19. LAGORCE L.K., Brand O., On magnetic microactuators based on polymer magnets, IEEE, Journal of Microelectromechanical Systems, Vol. 8, No. 1, pp.2-9, 1999

20. DUTOIT B.M., BESSE P.A., BLANCHARD H., GUERIN L., POPOVIC, High performance micromachined Sm2Co17 polymer bonded magnets, Sensors and Actuators 77, A, 1999, pp. 178-182

21. WATANABE Y., EDO M., NAKAZAWA H., YONEZAWA E., A new fabrication process of a planar coil using photosensitive polyimide and electroplating, Sensors & Actuators A54, pp 733-738, 1996

22. KLEIN J., GUCKEL H., High winding density microcoils for magnetic actuators, Microsystem Technologies 4 (1998) pp 172-175

23. CARDOT F.,GOBET F., BOGDANSKI M., RUDOLF F., Fabrication of a magnetic transducer composed of a high-density array of microelectromagnets with on-chip electronics, Sens. Actuators A 43 (1994), pp. 11-16

24. MILLER R and al., An electromagnetic MEMS 2*2 fiber optic bypass switch, Proc. Transducers’97, pp 89-92, 1997.

25. QUANDT E., LUDWIG A., Giant magnetostrictive multilayers, J. Appl. Phys. Vol.85 (8) (1999) 6232-6237 26. LUDWIG A., QUANDT E., Giant magnetostrictive thin films for applications in microelectromechanical systems,

J. Appl. Phys. Vol.87 (9) (May 2000), pp. 4691-4695 27. BETZ J., Torsion-Based, Drift-Free Magnetostrictive Microactuator, Proc. Actuators 96, Ed. Axon (Bremen, G.),

1996, pp. 283-286. 28. GARNIER A., BOUROUINA T., FUJITA H., HIRAMOTO T., ORSIER E., PEUZIN J.C., Magnetic actuation of

bending and torsional vibrations for 2D optical-scanner application, Sensors and Actuators 84, Part.A, pp. 156-160, 2000

29. BODY C., REYNE G., MEUNIER G., Modelling of magnetostrictive thin films - Application to a micromembrane, Journal de Physique III, 7 p67-85 (1997)

30. CLAEYSSEN F., LHERMET N., LE LETTY R., BOUCHILLOUX P., Actuators, Transducers and Motors Based on Giant Magnetostrictive Materials, Journal of Alloys and Compounds 248, 1997, pp. 61-73.

31. MACKAY K., BONFIM M., GIVORD D., 50T Pulsed Magnetic Fields in Microcoils, J.of Applied Physics, 2000, Vol.87,No4