Autonomous Energy Supply for Electronic Grains and Wireless Sensors

Transcription

1 Autonomous Energy Supply for Electronic Grains and Wireless Sensors Frequenz Autonome Energieversorgung für drahtlose Sensoren und Mikrosysteme Abstract Effective power supply of autonomous electronic systems and sensor nodes is a major challenge and a limiting factor for the performance. In general two methods are possible to provide the power: store the needed amount of energy on the node or scavenge available ambient power at the node. As the system size decreases, designing a sufficient energy supply is getting more and more difficult. Starting with an overview of power supply technology for wireless systems, the development of micro fuel cells as future high energy density systems and the wafer-level-technology for secondary micro batteries developed at Fraunhofer IZM is described. Demonstration systems of one square centimetre area and 200 µm thickness deliver 40 mw and 5 to 20 mw with micro fuel cells and Li-polymer batteries respectively. Übersicht Die Energieversorgung von autonomen Elektroniksystemen oder Sensornetzwerken ist eine der entscheidenden Herausforderungen, die weitgehend die Leistungsparameter und damit mögliche Anwendungen bestimmt. Die drahtlose Informationsübertragung ist nur dann sinnvoll, wenn auch die Energieversorgung drahtlos erfolgt. Es muß also entweder ein ausreichend großer Energiespeicher vorhanden oder die Möglichkeit der Energiegewinnung aus der Umwelt gegeben sein. Während technische Lösungen für größere Systeme vorliegen, ergeben sich eine Vielzahl von Problemen bei deren Miniaturisierung (egrains). Ausgehend von einem Überblick zum Stand der Technik werden in dem Beitrag die Entwicklung von Mikrobrennstoffzellen als zukünftige Energiespeicher hoher Energiedichte und die Entwicklung einer Wafer-Level-Technologie für Sekundärbatterien beschrieben, die am Fraunhofer-Institut für Zuverlässigkeit und Mikrointegration durchgeführt werden. Mit 1 cm 2 großen 200 µm dicken Beispielsystemen sind Leistungen von 40 mw für PEM-Brennstoffzellen und 5 bis 20 mw für Li-Polymerbatterien erzielbar. By Robert Hahn* Für die Dokumentation Micro battery / micro fuel cell / power density / energy density / autonomous sensors / encapsulation / wafer level battery 1. Introduction Connecting power supply wires to a wireless communication device defeats the purpose of wireless communication. As the system size decreases, designing a sufficient energy supply is getting more and more difficult. Therefore, the power supply is usually the largest and most expensive component of the emerging wireless sensor nodes being proposed and designed. Furthermore, the power supply (usually a battery) is also the limiting factor on the lifetime of the system. If wireless sensor networks are to truly become ubiquitous, replacing batteries in every device is simply cost prohibitive. Using replaceable batteries in a system smaller than 1 cm 3 may also be difficult. The most important metrics for power supply technologies are power- and energy density as well as lifetime for energy storage devices and power density for harvesting devices. Fig. 1 gives an overview of the energy density of coin type primary batteries. The energy density of alkaline, silver oxide and lithium batteries does not differ as much as one would expect from the energy density values of the active materials since packaging contributes much to the volume of small batteries. Zinc-air batteries have the highest energy density of commercially available primary batteries but the lifetime is limited to few months which makes them unsuitable for most autonomous systems. Rechargeable batteries are used to store energy which is supplied for example from solar cells. Fig. 2 gives an overview of the energy density of secondary coin type Lithium and NiMH batteries. Power density of most coin type batteries is small, but recently some systems have been developed which show good high current behaviour. Fig. 3 shows the discharge characteristic of the VARTA high rate NiMH battery V6HR with diameter and height of 6.8 and 2.15 mm respectively. The nominal capacity of this battery is 6 mah and it can be discharged at 3CA [1]. In comparison to button cells Li-polymer batteries have a much higher power density because of the layered thin structure of electrodes and electrolyte but they are currently not available at sizes below ca. 1 cm 3. * Fraunhofer Institute for Reliability and Micro-Integration, Berlin Fig. 1: Volumetric energy density of coin type primary batteries as function of size (ZL: zinc-air, L: lithium: Ag: silver oxide, Alk: alkaline) Power harvesting of the environment to continously fill the electrical storage device of the system can be done by different ways and depends strongly on the environmental conditions at the location of the device. Some examples of energy conversion methods are summarized in Table 1. Practical applications are at the moment more or less limited to solar cells. Standard lighting conditions, that means the incident light at midday at a sunny day, yield 100 mw/cm 2. But under indoor conditions this value is reduced easily by a factor of twenty. Single crystal silicon solar cells exhibit efficiencies of 15% 20%. 2. Micro fuel cells As improvements in battery technology have so far been limited to energy density increases of only a few percent per annum, over the past few years many R&D activities have concentrated on al- 87

2 Fig. 2: Volumetric energy density of coin type secondary batteries as function of size Fig. 4: Volumetric and gravimetric energy density of fuel cells and batteries (* H2-FC stack: Fraunhofer 8W demonstration system with metal hydride hydrogen storage) Diffusion Layers (GDL) which are inserted between MEAs and bipolar plates to distribute the reactants uniformly. Fuel cells of the air-breathing type are using ambient air as an oxidant. PEM fuel cells operate with hydrogen. At the moment there is no hydrogen storage available which is suitable for miniature applications. For direct methanol fuel cells (DMFC) a better storage opportunity exists in form of methanol cartridges. 88 Fig. 3: Discharge curves of high rate NiMH button cell [1] Table 1: Examples of realized micro-power energy conversion modules for the use of ambient energy to power autonomous systems Energy Conversion module Power density Light radiation Photovoltaic modules mw/cm 2 [3] Thermal energy Thermoelectric devices μw/cm 3 [4] Mechanical energy: vibrations, air flow, pressure variations, movements of humans ternative forms of miniature power supply. One of the most promising candidates is micro fuel cells (FC) based on Polymer Electrolyte Membranes (PEM). Compared to batteries the environmental impact of fuel cells is much lower [1]. Fig. 4 gives an energy density comparison of fuel cells and batteries, showing the potential of this technology. Fuel cell principle Piezoelectric generators, rotary electromagnetic engines μw/cm 3 [5] Chemical energy Bio-fuel cells 175 μw/cm 2 [6, 7] Fuel cells operate on the same principle as batteries, electrochemically converting energy, but are open systems where the reactor size and configuration determine the energy and power output. Since a single fuel cell has an operation voltage of ca. 0.5 volts, a multitude of cells are needed which are typically assembled in stack configuration. A fuel cell stack can be subdivided into three constituent component groups: the Membrane Electrode Assemblies (MEAs) which fulfil the electrochemical function of the fuel cell, the bipolar plates (commonly consisting of graphite and carbon-filled polymers) which supply the MEAs with hydrogen and oxygen, providing cooling and conductive electronic paths, and the Gas Recently a number of papers have been published dealing with the development of MEMS based fuel cells [8 12]. There are two reasons for this development: 1. Using new technologies and designs it should be possible to significantly improve fuel cell performance when micro-scale phenomena are exploited. However, such benefits can only be realized if the fuel cell devices can be fabricated using available manufacturing techniques. They are in most cases adapted from semiconductor and micro systems technology. Some of these technologies may have the potential to solve problems which are critical in the conventional stack technology. 2. The majority of research on micro-scale fuel cells is also aimed at micro-power applications. Miniaturization of the conventional fuel cell stack technology is not possible down to these dimensions. At Fraunhofer IZM technologies for wafer level fabrication of planar PEM fuel cells between 1 mm 2 and approximately 1 cm 2 were developed. The investigations focused on pattering technologies for the fabrication of micro flow fields, design studies for integrated flow fields, material compatibility for fuel cells, patterning of membrane electrodes, serial interconnection of single cells in a planar arrangement, laminating and assembling processes. Although wafer technologies were applied, foil materials were used which allow low-cost fabrication in future production. Prototypes of self breathing PEM fuel cells with a size of 1 1 cm 2 and 200 μm thickness were fabricated. V/I curves were measured at a variety of ambient conditions. Fuel cells with 40 ma output current at 1.5 V (= 120 mw/cm 2, 25 C, 50 % RH) have been successfully demonstrated. Cell performance was validated under varying ambient conditions. Stable long term operation at 80 mw/cm 2 was achieved. The total performance of the micro fuel cells is in the same range of current and power density compared to the best conventional planar PEM fuel cells. At the same time this technology offers a high degree of miniaturization and the capability for mass production which is a clear success of our micro patterning approach. Fig. 5 shows a micro fuel cell based on foil technology. Ambient temperature and humidity have a significant influence on the fuel cell performance since power density depends on the water concentration inside the polymer membrane and the re-

3 Fig. 5: Flexible micro fuel cell demonstrator: size 1 cm 2, thickness 200 μm, 80 mw/cm 2 Fig. 7: Schematic micro fuel cell and hydrogen generation from chemical hydrides into an electronic system of size 1cm 3 moving of water which is produced by the reaction of oxygen and hydrogen at the cathode. Miniaturized systems must work passively and cannot be controlled by fans or pumps. The impact of ambient conditions on the current density at a voltage of 350 mv of the micro fuel cell is shown in Fig. 6. Performance degrades at high temperatures in combination with low humidity and at low temperatures. Since storage of gaseous hydrogen is not practical in small size applications we are trying to develop a micro chemical reactor which produces hydrogen on demand from NaBH 4 and water which should yield an energy density of 800 Wh/l. A possible integration of fuel cells and a storage container into an 1 cm 3 egrain is shown schematically in Fig Li-Polymer-Batteries and Wafer-Level Processing For systems which receive their power from ambient energy as shown in Table 1 or from fuel cells a temporary energy storage is needed in most cases, since time and height of the supplied power does not correspond with the instantaneous power demand of the device. The electrical energy can be stored in a secondary battery or in a capacitor. Ultracapacitors represent a compromise between batteries and standard capacitors. Their energy density is about one order of magnitude higher than standard capacitors and about 1 to 2 orders of magnitude lower than rechargeable batteries. Lithium polymer batteries have the highest energy density of all commercially available secondary batteries. They are in widespread use in portable electronics with capacities between 100 and ca mah. Table 2 gives an overview of technical energy densities. Due to the layered structure of thin foils, their power density is quite high. They can be easily adapted to the dimension of the device and are flexible in size as shown in Fig. 8 and 9. Miniaturization of these batteries down to sizes below ca. 1 cm 3 leads to a reduction of energy density, since the fraction of the battery package increases at the expense of active material. Fig. 10 shows the energy density of small Li-polymer batteries as function of battery thickness. There is a significant reduction of energy density at thickness below 1 mm because the thickness of the packaging foil becomes dominant. High energy density of small (mm) size batteries can only be maintained, if the volume fraction of the battery package is reduced. Neither the coin type metallic casing nor the Al-polymer foil package is suited for egrain batteries. Small geometrical dimensions, minor self discharge as well as a long live time, are additional needs for a secondary battery used to power an egrain. Furthermore, a fast charge rate and cost saving production technology is desirable. To reduce the size of the packaging and improve the handling and assembly of miniature batteries we established a wafer level process which combines foil processing of Li-batteries and Wafer-technologies for battery contacts and encapsulation. Table 2: Energy density of Li-polymer batteries Wh/l predicted for cylindrical prismatic Fig. 6: Impact of ambient conditions on the current density of the micro fuel cell at a voltage of 350 mv Single cell Battery pack with 6 mm cells 350 (notebook) 280 (cellular) 89

4 Fig. 8: Packaging flexibility of Li-polymer batteries Fig. 9: Bended Li-Polymer battery Work is also being done towards micro-primary batteries. Fraunhofer IKTS and IZM are investigating thick film batteries of AgO/Zn with an aqueous KOH electrolyte. Thick films are on the order of 100 μm, but overall thickness is minimized by use of current collector foils as outer package. While each cell is rated at 1.5 V, acceptable power outputs at small overall volumes are possible due to the low internal resistance by use of the alkaline electrolyte. One of the main issues here lies in maintaining a micro-fabricated structure that contains aqueous electrolyte. The packaging foil and the sealed seam used for battery encapsulation make it difficult to reduce the battery size of conventional Li-batteries. To achieve high energy density for small size batteries according to Table 2, new encapsulation strategies are necessary. An encapsulation based on parylene (condensation polymerization) deposition followed by a final metallization allows the production of secondary Li-batteries with a maximum relationship of active material to encapsulation material. The condensation polymerization of parylene and aluminum coating using vapor deposition are room-temperature processes, which are particularly suitable for batteries. A high number of batteries can be produced and encapsulated at the same time using wafer level technology. The arbitrary footprint of the battery allows the best utilization of the available space in small electronic systems like egrains. The thickness of a single battery assembly is reduced because the encapsulation layer is very thin. A satisfactory way to boost the deliverable capacity while avoiding a raise of the internal resistance is to create a battery stack. Table 3 shows the process sequence of the wafer level packaging process. The application and laminating of chip sized Li-battery layers on the substrate (step 3) has to be done under dry and inert (argon) ambient conditions. In future production this should be performed at the battery manufacturer. Then a gas tight containment system is needed to transfer the substrate to the semiconductor or thin film production facilities for final processing. The lamination of foils onto substrates has been performed in a glove-box for this investigations. A leak proof single wafer containment system has been developed which is shown in Fig. 11 to carry the wafer between glove box and the parylene deposition reactor. The tightness of the wafer box has been proven with the calcium thin film method. A cross sectional view of the wafer level battery is shown in Fig. 12. First investigations focused on parylene processing conditions and measurements of gas permeation rates using a mass spectrometer. At a pressure of 33 mbar, deposition rates are in the order of 1 μm per hour. Thickness deviations are below 10 %. Suffi- 90 Fig. 10: Energy density of small Li-polymer batteries as function of battery thickness for different suppliers Many other approaches have been described for the development of micro batteries. At Oak Ridge National Laboratory a process has been created by which a secondary thin film lithium battery can be deposited onto a chip [13]. The thickness of the entire battery is on the order of 10 s of μm and the areas studied are in the cm 2 range. This battery is a layered structure with alternating layers of Lithium Manganese Oxide (or Lithium Cobalt Oxide), Lithium Phosphate Oxynitride and Lithium metal. Maximum potential is 4.2 V with continuous/max current output on the order of 1 ma/cm 2 and 5 ma/cm 2 for the LiCoO 2 -Li based cell. All layers are fabricated by vacuum processing. Especially the deposition of the electrolyte layer is a highly sophisticated and time consuming process which leads to high production cost. Electrical cycling of the solid state battery leads to cyclic changes of crystallite volume and material tension. Therefore the total capacity is rather low since layer thickness cannot be increased above some micrometers. Table 3: Processing sequence of wafer level battery 1 Provide a substrate (holes for backside electrolyte fill may be fabricated in advance) 2 Base metallization of substrate as cathode current collector 3 Addition of the three battery layers (cathode, separator/ electrolyte, anode) by using laminating technology 4 Coating with parylene as the first passivating layer 5 Structuring the parylene coating and open contact pads 6 Second metallization (aluminum vapor deposition) in order to cover the system gas-tight ( depending on requirements, a sequence of multiple layers may be used) 7 Patterning the metal layer subtractively or by the use of shadow masks 9 Backside filling of electrolyte if necessary and formation of the battery (first cycle) 10 Finally close (solder) the filling holes with a local heating process

5 Table 4: Comparison of PEM micro fuel cell with NaBH 4 hydrogen generator and Li-polymer secondary battery in 1 cm cubic size. 0.5 mm housing thickness was assumed. Frequenz Micro fuel cell with NaBH 4 hydrogen generation Li-polymer battery, wafer level encapsulation 1. 1 cm 1 cm 200 μm Voltage (0.5V) 1.5 V (at 40 mw) V Power 40 mw 5 20 mw Energy 0 3,5 mwh 2. 1 cm 1 cm 1 cm Voltage (0.5V) 1.5 V 7.5 V V Power 40mW 200 mw* mw Energy 500 mwh 150 mwh References Fig. 11: Gas tight containment system for Wafer-Level-Battery fabrication Fig. 12: Cross sectional view of Wafer Level Battery cient adhesion on Cu, Al, Si and FR4 material has been achieved with a variety of surface preparations. The moisture vapor transmission of parylenes is better than most polymers while the gas permeation of O 2 and N 2 is comparable to epoxides. Since most of the battery is covered with thin film metal, only some square μm of parylene at the current feed through are effective for gas permeation and a nearly hermetic package can be achieved. Summary Power supply of wireless sensor nodes and egrains is a real challenge. Storing the energy needed for long operation periods at the egrain is conflicting with miniaturization. Size reductions may be achieved, when micro fuel cells are available with higher energy density compared to primary batteries which are the well established sources at the moment. For widespread use and miniaturization of the systems, the power harvesting of ambient energy or use of alternative power sources is essential. In this context a small size secondary energy storage is needed which can be easily integrated. We considered a combination of standard Lipolymer battery processing and wafer technologies to reduce battery dimensions to chip size and keep fabrication cost low. In the first stage of this project we developed equipment and processes for the merge of battery and IC fabrication. Still a great effort is needed to make the wafer level battery a reality. Table 4 gives an overview of energy and power densities obtainable in 1 cm 3 size systems. [1] Ilic, D.; Holl, K.; Pytlik, E.; Birke, P.; Haug, P.; Wöhrle, T.; Perner, A.; Birke-Salam, F.: VARTA Microbattery Smart Batteries Solution, Minatec 2003, Sept.2003, Grenoble. [2] Hahn, R.; Müller, J.: Future Power Supplies for Portable Electronics and Their Environmental Issues, International Congress ELECTRON- ICS GOES GREEN (EGG 2000), September 11 13, Berlin, Germany pp [3] Hahn, R.: Autarke Systeme mit Piezogeneratoren, Mikrosolarmodulen und flexiblen Speichern. 2. GMM-Workshop Energieautarke Sensorik, 6., 7. Juni 2002, Dresden. [4] Böttner, H.: Thermoelektrische Wandler, Stand der Technik. 2. GMM- Workshop Energieautarke Sensorik, 6., 7. Juni 2002, Dresden. [5] Roundy, S. R.: Energy Scavenging for Wireless Sensor Nodes with a Focus on Vibration to Electricity Conversion. Ph.D. Thesis, University of California, Berkeley CA, May [6] O Neill, H.; Woodward, J.: Construction of a Bio-Hydrogen Fuel Cell, DARPA Advanced Energy Technologies Energy Harvesting Program, Arlington, VA, April [7] Palmore, G. T.: Biofuel Cells. DARPA Advanced Energy Technologies Energy Harvesting Program, Arlington, VA, April [8] Miller, M. A.: MEMS Manufacturing Technologies as Applied to the Development of Embedded Micro Fuel-Cells. in NCMS Fall Workshop Series Dearborn, MI. [9] Meyers, J. P.; Maynard, H. L.: Design Considerations for Miniaturized Pem Fuel Cells. Journal of Power Sources 109 (2002) 76 88, 14 January [10] Seo, Y.-H.; Cho, Y.-H.: A Miniature Direct Methanol Fuel Cell Using Platinum Sputtered Microcolumn Electrodes with Limited Amount Of Fuel. 16. Int. IEEE Congress MEMS-2003, Kyoto. [11] Schmitz, A.; Tranitz, M.; Wagner, S.; Hahn, R.; Hebling, C.: Planar self-breathing fuel cells. Journal of power sources 5213 (2003) [12] Hahn, R.; Krumm, M.; Reichl, H.: Thermal management of portable micro fuel cell stacks. Proceedings of the 19. IEEE Semiconductor Thermal Measurement and Management Symposium SEMI-THERM, San Jose, CA USA, March 11 13, 2003, pp [13] Bates, J.; Dudney, N. et al.: Thin-film lithium and lithium-ion batteries. Solid State Ionics 135 (2000) 35. [14] Hahn, R.: Batterie, insbesondere Mikrobatterie, und deren Herstellung mit Hilfe von Wafer-Level-Technology. Deutsche Patentanmeldung DE Dr. R. Hahn Fraunhofer Institute for Reliability and Micro-Integration Gustav-Meyer-Allee 25 D Berlin Germany hahn@izm.fhg.de (Received on February 2, 2004) 91