DOI QR코드

DOI QR Code

Bulk Micromachined Vibration Driven Electromagnetic Energy Harvesters for Self-sustainable Wireless Sensor Node Applications

  • Received : 2012.11.21
  • Accepted : 2013.05.27
  • Published : 2013.11.01

Abstract

In this paper, two different electromagnetic energy harvesters using bulk micromachined silicon spiral springs and Polydimethylsiloxane (PDMS) packaging technique have been fabricated, characterized, and compared to generate electrical energy from ultra-low ambient vibrations under 0.3g. The proposed energy harvesters were comprised of a highly miniaturized Neodymium Iron Boron (NdFeB) magnet, silicon spiral spring, multi-turned copper coil, and PDMS housing in order to improve the electrical output powers and reduce their sizes/volumes. When an external vibration moves directly the magnet mounted as a seismic mass at the center of the spiral spring, the mechanical energy of the moving mass is transformed to electrical energy through the 183 turns of solenoid copper coils. The silicon spiral springs were applied to generate high electrical output power by maximizing the deflection of the movable mass at the low level vibrations. The fabricated energy harvesters using these two different spiral springs exhibited the resonant frequencies of 36Hz and 63Hz and the optimal load resistances of $99{\Omega}$ and $55{\Omega}$, respectively. In particular, the energy harvester using the spiral spring with two links exhibited much better linearity characteristics than the one with four links. It generated $29.02{\mu}W$ of output power and 107.3mV of load voltage at the vibration acceleration of 0.3g. It also exhibited power density and normalized power density of $48.37{\mu}W{\cdot}cm-3$ and $537.41{\mu}W{\cdot}cm-3{\cdot}g-2$, respectively. The total volume of the fabricated energy harvesters was $1cm{\times}1cm{\times}0.6cm$ (height).

Keywords

1. Introduction

Energy harvesting technologies that scavenge ambient natural energies such as light, heat, and mechanical vibrations, and then convert them to an electrically usable power have been widely investigated for ubiquitous sensor nodes and low-power consumed microsystem applications. In particular, micro-fabricated energy harvesters to scavenge low level ambient vibrations have been attracted because the ambient vibration source is abundant and easy to apply via a simple mechanical coupling without limitations of area and environment. These ambient vibrations, which are generated by car engine, micro oven, person tapping, HVAC vents in building and CD on laptop, exhibited vibration frequencies and magnitude of accelerations less than 120Hz and 1g, respectively [1].

Vibration-driven energy harvesters can be classified as piezoelectric, electrostatic, or electromagnetic devices due to their energy converting principles [2-4]. The electromagnetic device provides simple geometry, no external power source, relatively high output power, and simple fabrication process [5]. An electromagnetic energy harvester was reported with a samarium-cobalt (SmCo) permanent magnet, planar gold (Au) coil, and polyimide membrane. It generated 0.3μW of output power at the resonant frequency of 4MHz and the optimal load resistance of 39Ω [6-7]. Recently, several electromagnetic energy harvesters were reported using cantilever structures and mass-spring-damper systems [8-17]. These devices were comprised of single steel beam, coils, and magnets which were attached to the end of the beam. Although they provide small size/volume (under 0.15cm3), low load voltages and high operating frequencies were still left to be solved. For addressing these issues, a device with fourmagnet configuration and tungsten mass was presented [9]. Cantilever array based devices were also reported to control the output voltages and powers by adjusting the number of single cantilever [10-12]. In particular, several cantilevers connected in series with different lengths and resonant frequencies were applied for fabricating the energy harvester with wide operating bandwidth [11]. However, these harvesters provide low output power due to the limited windings of micro-coils which were fabricated on each single cantilever.

In this paper, highly miniaturized electromagnetic energy harvesters for scavenging ultra-low level ambient vibrations were developed by using silicon bulk micromachining and PDMS packaging techniques. The fabricated devices were comprised of a miniaturized permanent magnet, silicon spiral spring, solenoid copper coil, and PDMS housing. The magnet was highly miniaturized by milling a permanent neodymium-ironboron (NdFeB) magnet and the multi-turned solenoid coil geometry was applied to minimize the scaling effect and DC resistance. To reduce the nonlinear characteristics and resonant frequency of mechanical spring, and increase the deflection of the movable mass, two different silicon spiral geometries were applied and compared to find out an optimal geometry for the proposed electromagnetic energy harvesting devices.

 

2. Design Theory

2.1 Basic theory

Fig. 1. shows a mass-spring-damper system modeling of proposed ambient vibration-driven elecrtomagnetic energy harvester. It is comprised of damping coefficients, and , spring constant, k, and seismic mass, m [9]. When the housing is vibrated with a displacement, y(t), the displacement of mass, z(t), is relatively generated by the spring attached to the mass. The electrical damping coefficient, be and mechanical damping coefficient, bm, are defined as the transduction mechanism from mechanical to electrical energy conversion and unwanted parastic loss such as air damping. From the above analytical modeling, the following equation of motion can be derived as

Fig. 1.Schematic diagram of mass-spring damper system model for proposed vibration driven electromagnetic vibration energy harvester.

Assuming the harvester is driven by a harmonic base excitation, y(t)=Ysin(𝜔t), the average power dissipated within the damper can be obtained by including the mechanical and electrical damping ratio as follows [9].

where m is the mass, ζe is the electrically generated damping ratio given by ζe=be/2m𝜔n, 𝜔 is the input vibration frequency, 𝜔n is the natural frequency of massspring- damper system (𝜔n =√(k/m)), ζT is total damping ratio (ζT = ζe + ζm), and Y is the displacement of input vibration.

If 𝜔n is equal to 𝜔, the maximum power will be dissipated within the generator and Eq. (2) can be simplified as

where A is the acceleration magnitude of input vibration (A=𝜔2Y).

Thus, the average power delivered to the electrical domain can be expressed as Eq. (4) and the actual power in the load which excludes loss within the coil, can be expressed as a function of the coil and load resistance shown in Eq. (5):

where Pavg-load is the average power in the load, Rload and Rcoil are the resistances of load and coil, respectively. These resistances can also influence the electrical damping ratio and it can be simplified at low frequencies where the inductive impedance is much lower than the resistive impedances as shown in Eq. (6)

where B is the magnetic flux density to which it is subjetced , Nl is the effective coil length, and Lcoil is the coil inductance in the proposed electromagnetic harvester.

The Eqs. (5) and (6) show that be can be adjusted by Rload and the maximum power in the load can be obtained by matching bm and be. Thus, the optimum load resistance, Rload to achieve the maximum load power, PLmax can be calculated as folllows

2.2 Design of proposed electromagnetic energy harvesters

The key features of energy harvester used for microsystem applications are its size and energy conversion efficiency. However, it is difficult to reduce the size of electromagnetic energy harvester because scaled-down coil and magnet can have adverse effects on its output voltage and power performances [2, 18]. Therefore, optimal geometry, fabrication methodology, and core materials should be carefully considered to minimize the scaling effect.

Fig. 2 shows a cross-sectional view of the proposed electromagnetic energy harvester. In order to achieve a large amount of electrical power, the bulk magnet was applied as an inertial mass at the center of the spiral spring to increase the magnetic flux density and decrease the resonant frequency of mass-spring system. The NdFeB was selected for the bulk magnet due to the high magnetic flux density of approximately 1.2T and the ease of miniaturization. It has a cylindrical shape with a volume of π mm2 x 2mm (height) and the density of 7.4g/cm2, coercivity of 859kA/m, and relative permeability of 1.1, respectively.

Fig. 2.Cross-sectional view of proposed vibration driven electromagnetic vibration energy harvester.

As shown in Fig. 3, the silicon spiral spring was designed with two different geometries for its optimization. The spiral spring was selected to maximize the output power at very low resonant frequency and acceleration. The silicon has much higher yield strength than gallium arsenide (GaAs), copper (Cu), and polymeric material such as Parylene C, PDMS, and SU-8. Moreover, it has many advantages such as low mechanical loss, low material cost, and processing compatibility with CMOS rectifying circuitry [13]. Also, the bulk micromachined silicon spring has merits in terms of mass production and cost reduction through batch fabrication. In order to scavenge low vibration, the designated spring constant for mass-spring systems comprised of Si spiral springs and NdFeB magents were obtained from calculation, k=𝜔02m, and FEA simulation was performed to confirm. The designated spring constants were calculated through Eq. (9) and FEA simulation to obtain designated low spring constant. These spiral springs were designed to have the same size/volume of 8mm x 8mm x 40μm and line width/spacing of 300/100μm. Type A (two linked spring, each line length of 16.7mm) and Type B (four linked spring, each line length of 12.2mm) springs have approximately 32Hz, 65Hz, 2.09N/m, and 8.44N/m, respectively.

where n is the number of spring links, E is Youngth modulus of spring material, b, w, and l are the thickness, width, and length of spring. The detailed design parameters were described in Table 1.

Fig. 3.3D modeling of proposed silicon spiral springs: (a) 2 linked spiral spring (Type A) and (b) 4 linked spiral spring (Type B).

Table 1.Geometry parameters and material properties of proposed vibration driven electromagnetic energy harvesters.

Also, the solenoid coil was designed with 183 turns and low DC resistance of 13.5Ω to maximize the output power. It has the coil diameter of 3mm and height of 2mm. Its small size and low coil resistance are more independent from the scale effect than the electroplated planar copper coil. The magnetic flux densities were also estimated using the FEA method. The result was approximately 0.1T on coil. According to Eq. (5), the output power of 110 μW can be generated to 100Ω of load resistance from their resonant frequency with 0.4g of acceleration and the mechanical damping ratio of 0.0082 was applied for calculation. The mechanical damping ratio will be discussed in the following section.

 

3. Fabrication

The fabrication process was highly simplified for low cost production using two photolithography masks, silicon bulk micromachining, and PDMS housing. As shown in Fig. 4, the spiral shaped spring was firstly formed using photolithography and deep silicon reactive-ion etching (RIE) process. Silicon nitride was then deposited on all the surface of silicon wafer and the backside of wafer was selectively patterned and etched away using RIE plasma and KOH etching solution [19]. All the silicon nitride film was removed for releasing the silicon spiral springs and each sample was individually diced. Fig. 5 (a) and (b) show photomicrographs of the fabricated silicon spiral springs with two links and four links. Their sizes and volumes were 8mm x 8mm x 0.04mm (height).

Fig. 4.Fabrication sequences of silicon spiral spring and assembly of permanent magnet.

A highly miniaturized bulk NdFeB magnet was mounted at the center of spiral spring as a seismic mass. In addition, PDMS packaging substrate was fabricated using Teflon mold and the multi-turned solenoid copper coil was then inserted into the inner hole of the packaging substrate. Finally, the silicon spiral spring with a magnet was bonded with the PDMS packaging substrate with a coil. The fabricated energy harvesters have total mass of about 4.5g. Fig. 5 (c) shows the photomicrograph of the fabricated energy harvester mounted on an evaluation PCB board for test. When a vibration is applied to the device, the NdFeB magnet moves vertically at the inner part of solenoid copper coil and the changing magnetic flux induces a voltage via Faraday’s law of induction. Its total volume is 1cm x 1cm x 0.6cm (height).

Fig. 5.Photomicrographs of (a) 2 linked spiral spring (Type A), (b) 4 linked spiral spring (Type B), and micro-fabricated EMPG device mounted on an evaluation PCB board (c).

 

4. Experimental Results and Discussion

The fabricated electromagnetic energy harvesters with two different spring geometries were measured, analyzed, and compared using the experimental setup shown in Fig. 6. As shown in Fig. 6, an accelerometer, vibration exciter, oscilloscope, power amplifier, and function generator were utilized. The power amplifier and accelerometer were used to control the magnitude of acceleration and test frequency was controlled using the function generator. The oscilloscope was connected to the coil of the fabricated energy harvester in order to measure the open circuit voltage and the load voltage at its terminal.

Fig. 6.Experimental set up with vibration exciter and accelerometer for measuring the output voltage of the fabricated electromagnetic energy harvester.

Firstly, the open circuit voltage was measured to find the resonant frequency of device as shown in Fig. 7. The type A and type B energy harvesters with 183 turns of coil exhibited 68.65mV and 29.95mV of open circuit voltage at about 36Hz and 61Hz of vibration frequency with the acceleration of 0.04g, respectively. The measured resonant frequencies were slightly different with the calculated and simulated ones. This discrepancy might be caused by the fabrication and assembly tolerance of spring and magnet. At the fabrication process of deep RIE (DRIE) and KOH wet etching for forming the silicon spiral spring, maximum 5% of thickness tolerance was measured on the same silicon wafer. Since their open circuit voltage curves were not perfectly symmetric shape as shown in Fig. 7, the mechanical damping ratios and Q-factors were approximately measured using the logarithmic decrement method shown in Fig. 8. Fig. 8 (a) and (b) illustrate the decay of open circuit voltage in case of turning off the input vibration amplifier during eight and four seconds, respectively. The mechanical damping ratio of each energy harvester was calculated using the following Eq. [8]:

where n is a number of cycles, and V1 and V1+n are the voltage amplitudes. In addition, through the following Eq. (11), the unloaded quality factor of Qm was calculated.

Fig. 7.Measured open circuit voltages of the fabricated electromagnetic energy harvester with 2 linked spiral spring (Type A) or 4 linked spiral spring (Type B).

Fig. 8.Decay plots using logarithmic decrement method of open circuit voltage at 0.04g for the calculation of Q-factor and mechanical damping factor: (a) Type A and (b) Type B.

The calculated unloaded Q-factor and mechanical damping ratio were approximately 348 and 0.0014 at Type A, and 277 and 0.0018 at Type B devices, respectively. Since the air damping was increased by the number of links of spring, Type B device exhibited lower quality factor and higher mechanical damping than Type A.

Fig. 9 (a) and (b) show how the output power and load voltage vary with the load resistance at the acceleration of 0.3g. As shown in Fig. 9, the Type A and B devices exhibited the maximum powers at the load resistance of 99Ω and 55Ω, respectively. The optimum load resistance was much larger than the coil resistance. This discrepancy between the coil and optimum load resistance was caused by damping constant of bm as shown in Eq. (7). Through these results, we could ascertain that optimum load resistance is inversely proportional to the mechanical damping ratio.

Fig. 9.The output power (a) and load voltage (b) of the fabricated energy harvesters by varying the load resistances.

As shown in Fig. 10, the Type A device had the almost same resonant frequency as the input acceleration. However, the resonant frequencies of Type B device were not agreed with the input accelerations. These phenomena are related to the presence of nonlinear effect due to the spring stiffening. At the higher displacements of spring, because the force-deflection characteristics are dominated by tensile stress, the nonlinear characteristic is increased at the higher acceleration [7].

Fig. 10.The load voltages of the fabricated energy harvesters at various accelerations ranged from 0.04g to 0.3g.

Figs. 11(a) and (b) show the load voltage and output power measured at the accelerations ranged from 0.04g to 0.3g. Type A device exhibited 107.3mV of load voltage, 29.02μW of output power, and 48.37μW/cm3 of power density at 0.3g. On the other hand, Type B device exhibited 73.17 mV of load voltage, 24.2μW of output power, and 40.33μW/cm3 of power density at 0.3g. The theoretically calculated results were well agreed to the experimental ones, as shown in Fig.11. As discussed in Eq. (8), since angular frequency and mechanical damping ratio are inversely proportional to the output power, the output power of Type A device was higher than one of Type B at the same volume and input acceleration. Table 2 shows the comparison of key features of the fabricated energy harvesters and previously reported works. Glynne-Jones et al. reported the highest power density by using discrete approach and Koukharenko et al. presented the highest power density by using MEMS based energy harvester from a very high vibration frequency. However, the discrete approach has disadvantage in productivity and the ambient vibration has a frequency below 200 Hz. The proposed EMPGs have the highest normalized power density of 537μW/cm3·g2 in the MEMS based vibration energy harvesters with low vibration by combining the MEMS based silicon spiral spring, discrete magnet, and multi-turned coils.

Fig. 11.Comparison of measured and numerically calculated output voltages (a) and powers (b) of the fabricated energy harvesters at their resonant frequencies and various accelerations.

Table 2.Comparison of key performance features of the fabricated devices with others.

These results show that the fabricated devices are useful for powering the wireless sensor nodes without batteries in intelligent buildings (temperature, humidity, and air quality monitoring), transportation (wheel flats and baring degradation monitoring in rail), environmental (air pollution and earthquake monitoring), and industrial (structure, process, and equipment condition monitoring) applications by increasing the volumes of energy harvesters. Since the ambient vibrations from many household appliances and objects in and around buildings are reported to have vibration frequencies on the order of 100 Hz and accelerations ranging from 0.05 to 5 m/s2 [1]. Furthermore, Gilbert et al. reported that the typical average power consumptions for ZigBee based wireless network node applications were around 3 mW [20]. For doing successfully commercialization, the harvesting devices should provide sufficient power, ease of installation, high reliability, adequate life time, and low production cost.

 

5. Conclusion

In this paper, ambient vibration-driven electromagnetic energy harvesters were designed, fabricated, and evaluated. The fabricated devices were comprised of bulkmicromachined silicon spiral springs, low loss copper solenoid coils, and miniaturized NdFeB magnets. For low cost and mass production, the fabrication and packaging process were highly simplified. The NdFeB magnet and multi-turned copper coil was highly effective to increase the output voltage and power of device. Micro-fabricated silicon spiral springs were also effective to generate the electrical power from low level ambient vibration sources. In comparison of the fabricated harvesters with two different spring geometries, the device with two links of spring exhibited better performance characteristics than the one with four links of spring at low ambient vibrations. As the number of links increased, the mechanical damping was increased and the quality factor was decreased. At the higher accelerations, the mechanical damping and resonant frequency were rapidly increased and non-linear. As expected, the fabricated devices exhibited excellent power density at low vibration under 0.3g. The nonlinear effect of mechanical and electrical damping to the input acceleration will be analyzed in near future.

References

  1. S. Roundy, P.K. Wright, J. Rabaey, "A study of low level vibrations as a power source for wireless sensor nodes," Jour. of Computer Communications, Vol. 26, Issue 11, pp. 1131-1144, July. 2003. https://doi.org/10.1016/S0140-3664(02)00248-7
  2. D. P Arnold, "Review of microscale magnetic power generation," IEEE Trans. Magnetics, Vol. 43, Issue. 11, pp. 3940-3951, Nov. 2007. https://doi.org/10.1109/TMAG.2007.906150
  3. S. P. Beeby, M.J. Tudor and N.M. White, "Energy harvesting vibration sources for microsystems applications," Jour. of Meas. Sci. Technol., Vol. 17, No. 12, pp. R175-R195, Dec. 2006. https://doi.org/10.1088/0957-0233/17/12/R01
  4. Hudak, S. Nicholas, "Small-scale energy harvesting through thermoelectric, vibration, and radiofrequency power conversion," Jour. of Applied Physics, Vol. 103, Issue. 10, pp. 101301-101301-24, May. 2008. https://doi.org/10.1063/1.2918987
  5. Soliman, M.S.M., "Electromagnetic MEMS Based micro-power generator," IEEE Sym. on Industrial Electronics, Vol. 4, pp. 2747-2753-23, July. 2006
  6. C. B. Williams, R.B. Yates, "Analysis of a microelectric generator for microsystems," Jour. of Sensors and Actuators A: Physical, Vol. 52, Issue 1-3, pp. 8-11, Mar.-Apr. 1996. https://doi.org/10.1016/0924-4247(96)80118-X
  7. C. Shearwood, R.B. Yates, "Development of an electromagnetic micro-generator," Electron. Lett., Vol. 33, pp. 1883-1884, 1997. https://doi.org/10.1049/el:19971262
  8. P. Glynne-Jones, M.J. Tudor, "An electromagnetic, vibration-powered generator for intelligent sensor systems," Jour. of Sensors and Actuators A: Physical, Vol. 110, Issues. 1-3, pp. 344-349, Feb. 2004. https://doi.org/10.1016/j.sna.2003.09.045
  9. S. P. Beeby, R.N. Torah, M.J. Tudor, P. Glynne-Jones, T.O'Donnell, C.R. Saha and S. Roy, "A micro electromagnetic generator for vibration energy harvesting," Jour. of Micromechanics and Microengineering, Vol. 17, No. 17, pp. 1257-1265, July. 2007. https://doi.org/10.1088/0960-1317/17/7/007
  10. Makoto Mizuno and D.G. Chetwynd, "Investigation of a resonance microgenerator," Jour. of Micromechanics and Microengineering, Vol. 13, No. 2, pp. 209-216, Mar. 2003. https://doi.org/10.1088/0960-1317/13/2/307
  11. Ibrahim Sari, Tuna Balkan, Haluk Kulah, "An electromagnetic micro power generator for wideband environmental vibrations," Jour. of Sensors and Actuators A: Physical, Vol. 145-146, pp. 405-413, July-Aug. 2008. https://doi.org/10.1016/j.sna.2007.11.021
  12. Ibrahim Sari, tuna Balkan and Haluk Kulah, "An electromagnetic micro energy harvester based on an array of parylene cantilevers," Jour. of Micromechanics and Microengineering, Vol. 19, No. 10, pp. 105023, Oct. 2009. https://doi.org/10.1088/0960-1317/19/10/105023
  13. N. S. awaja, D.K. Sood, Thurai vinay, "Modeling and simulation of a flat spring for use in an electromagnetic microgenerator," Proc. SPIE, pp. 361-372, Vol. 5649, Feb. 2005.
  14. M. El-hami, P. Glynne-Jones, N. M. White, M. Hill, S. Beeby, E. James, A. D. Brown, and J. N. Ross, "Design and fabrication of a new vibration-based electromechanical power generator," Sens. Actuators A, Phys., Vol. 92, No. 1-3, pp. 335-342, Aug. 2001. https://doi.org/10.1016/S0924-4247(01)00569-6
  15. E. Koukharenko, S. P. Beeby, M. J. Tudor, N. M. White, T. O'Donnell, C. Saha, S. Kulkarni, and S. Roy, "Microelectromechanical systems vibration powered electromagnetic generator for wireless sensor applications," Microsyst. Technol., Vol. 12, No. 10-11, pp. 1071-1077, Sep. 2006. https://doi.org/10.1007/s00542-006-0137-8
  16. S. Kulkarni, S. Roy, T. O'Donnell, S. P. Beeby, and M. J. Tudor, "Vibration based electromagnetic micropower generator on silicon," J. Appl. Phys., Vol. 99, No. 8, pp. 08P511, Apr. 2006. https://doi.org/10.1063/1.2176089
  17. C. T. Pan, Y. M. Hwang, H. L. Hu, and H. C. Liu, "Fabrication and analysis of a magnetic self-power microgenerator," J. Magn. Magn. Mater., Vol. 304, No. 1, pp. e394-e396, Sep. 2006. https://doi.org/10.1016/j.jmmm.2006.01.202
  18. T. O'donnel, C. Saha, S. Beeby, J. Tudor, "Scaling effects for electromagnetic vibrational power generators," Jour. of Microsystem Technologies, Vol. 13, Issue. 11-12, pp. 1637-1645, July. 2007. https://doi.org/10.1007/s00542-006-0363-0
  19. K. Chang, S. Kang, K. Park, and H. Kim, "Pieoelectrically actuated micro corner cube retroreflector for free-space optical communication applications," Journal of Electrical Eng. & Tech., Vol. 5, Issue. 2, pp. 337-341, Nov. 2010. https://doi.org/10.5370/JEET.2010.5.2.337
  20. J. M. Gilbert and F. Balouchi, "Comparison of energy harvesting systems for wireless sensor networks," Int. J. Autom. Comput., Vol. 5, Issue. 4, pp. 334-347, Oct. 2008. https://doi.org/10.1007/s11633-008-0334-2