- Email: [email protected]
Effect of micropillar electrode spacing on the performance of electrohydrodynamic micropumps
Journal of Electrostatics 68 (2010) 376e383
Contents lists available at ScienceDirect
Journal of Electrostatics journal homepage: www.elsevier.com/locate/elstat
Effect of micropillar electrode spacing on the performance of electrohydrodynamic micropumps P. Zangeneh Kazemi, P. Ravi Selvaganapathy, C.Y. Ching* Dept. of Mechanical Engineering, McMaster University, Hamilton, ON L8S4L7, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 September 2009 Accepted 21 May 2010 Available online 8 June 2010
Electrohydrodynamic (EHD) micropumps with three-dimensional 50 mm 50 mm micropillar electrodes were fabricated and tested in this study. Two basic electrode configurations were investigated: (i) micropillar emitter and collector electrodes (symmetric) and (ii) micropillar emitter and planar collector electrodes (asymmetric). The micropumps were fabricated by integrating chromium/gold planar electrodes with electroplated 3-D Nickel micropillars on a glass substrate with a 100 mm high PDMS microchannel. The effect of the spanwise micropillar spacing on the pump performance was determined. The pumps were tested using HFE-7100 as the working fluid for the maximum pressure generation under a no flow condition. The micropumps with the asymmetric electrode design generated a significantly higher pressure head than the corresponding micropumps with symmetric electrode configuration for the same applied voltage, with lower power consumption. A decrease in the spanwise spacing of the micropillar electrodes increased the pump performance for the symmetric configuration, while the performance decreased for the asymmetric configuration. Ó 2010 Elsevier B.V. All rights reserved.
Keywords: Micropumps Electrohydrodynamics Ion drag Microelectronics cooling Microfabrication
1. Introduction Microfluidics finds application in a number of diverse areas such as medical diagnostics, drug delivery, drug discovery, chemical synthesis, electronic cooling and environmental analysis. One of the critical components of any microfluidic device is the micropump whose design, functionality and suitability have to be carefully considered for specific applications. Among the different micropump concepts that have been developed, electrohydrodynamic (EHD) micropumps are especially suited for microfluidic electronic cooling applications because they have no moving parts, use very low power and are amenable to rapid smart control. In addition, they can be designed with very small form factors. In EHD pumping, the interaction between an applied electric field and a dielectric fluid results in an electrical body force on the fluid [1e4]. Specifically, an ion drag EHD micropump involves the interaction of an electric field with electric charges, dipoles or particles injected into a dielectric fluid [5]. This Coulombic force on the charged species is the main driving mechanism, where neutral molecules get dragged along with charged ions that are moving between the electrodes, thereby generating fluid flow. The induced electric body force due to the interaction between an applied electric field and a dielectric fluid is given by [6,7] * Corresponding author. Tel.: þ1 905 525 9140x24998; fax: þ1 905 572 7944. E-mail address: [email protected] (C.Y. Ching). 0304-3886/$ e see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.elstat.2010.05.008
1 1 v3 Fe ¼ re $E E2 $V3 þ V E2 $r$ vr T 2 2
(1)
where re is the charge density, 3 is the fluid permittivity, r is the fluid density and T is the fluid temperature. The three components of the electric body force on the right of Eqn (1) are: (i) the electrophoretic or Coulomb force, (ii) the dielectrophoretic force and (iii) the electrostrictive force, respectively. In single phase flows, the dielectrophoretic and electrostrictive forces are, in general, negligible compared to the Coulomb force. The Coulomb force is directly proportional to the electric field and electric charge density. The charge density and electric field are governed by Poisson’s equation and conservation of charge given by
V$E ¼
re 3
(2)
VE ¼ 0
(3)
vre þ V:J ¼ 0 vt
(4)
where the charge current density J is defined as the total charge flux due to bulk motion of fluid. The electric body force can be incorporated into the NaviereStokes equations as an additional body force to couple the effect of electric field and charge distribution on the fluid flow as
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
377
Fig. 1. Schematic of micropump with symmetric electrode configuration.
V$u ¼ 0
r
(5)
vu þ ðu$VÞu ¼ Vp þ r$g V$s þ Fe vt
(6)
To simulate the pumping phenomenon in EHD micropumps, equations (1)e(6) have to be solved simultaneously for charge density distribution, electric potential and fluid flow. Since the nature of charge injection at the emitter electrode in electrohydrodynamics is not fully understood, the proper boundary conditions for the charge density on the electrodes are unknown. Thus, it is not possible to obtain a solution without a model for the charge boundary conditions at the electrodes. Numerous studies have been conducted to characterize the performance of EHD micropumps with various operating conditions, working fluids and geometrical design parameters [8e16]. The first planar EHD micropump, fabricated from an array of gold electrodes on a glass substrate with ethyl alcohol as the working fluid, was developed by Ahn and Kim [11]. A simple one-mask process was used to fabricate the electrodes, with an electrode spacing of 100 mm and an electrode double spacing (distance between each pair of electrodes) of 200 mm. A maximum pressure head of 240 Pa and a flow rate of 50 mL/min were achieved. The pressure and flow rate were found to be proportional to the square of the applied voltage. Benites et al. [15] investigated the effect of channel height, electrode spacing and double spacing on back pressure and flow rate in an ion drag EHD micropump with planar electrodes deposited on the bottom wall of a microchannel using HFE-7100 as the working fluid. A flow rate up to 7.92 mL/min was
reported at an applied voltage of 500 V. More recently, Darabi and Rhodes [5] simulated a micropump with an array of interdigitated symmetric electrodes patterned along the top and bottom walls of a microchannel. The performance of the micropump with electrodes on both the top and bottom walls was considerably higher compared to a pump with electrodes only on the bottom wall. Heretofore, most development for EHD micropumps has been done with 2-D planar electrodes (i.e planar electrodes deposited on the channel wall). In this configuration, however, the electric field gradients are highest close to the channel wall and decreases rapidly away from the wall since the sharp edged corners of the electrodes are confined to the surface. Thus the charge distribution is mainly confined to the near-wall region. A better charge distribution within the pump domain can be achieved using 3-D micropillar electrodes, albeit at the complexity of the microfabrication. With 3-D micropillar electrodes, the total length of sharp edges and surface area of the electrode are increased substantially throughout the pump domain. As a result, the electric field gradient and the space charge density are expected to increase significantly, thereby increasing the Coulomb force within the pump domain and enhancing the EHD pump performance. The penalty, of course, is the increased surface area exposed to the flowing fluid which will increase the pressure drop. Recently, micropumps with 3-D micropillar electrodes were shown to have a better performance than the corresponding 2-D electrode geometry [17]. Surprisingly, a micropump with a 2-D asymmetric electrode geometry (where the width of the emitter was twice that of the collector) was found to have a better performance than a micropump with 3-D micropillar emitter and collector electrodes
Fig. 2. 2-D Cross section of micropump with different electrode configurations: (a) symmetric, (b) asymmetric electrode configuration.
378
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
Table 1 Micropumps design specifications. Micropump
des (mm)
deds (mm)
dp (mm)
Ne
H (mm)
W (mm)
S-P80-S120 S-P120-S80 S-P160-S80 A-P80-S80 A-P120-S80 A-P160-S80
120 80 80 80 80 80
240 160 160 160 160 160
80 120 160 80 120 160
100 100 100 100 100 100
100 100 100 100 100 100
5 5 5 5 5 5
[17,18]. However, the effect of the asymmetry on the flow rate, and the spanwise micropillar electrode spacing and the inter-electrode spacing were not reported. The objective of this study was to determine the effect of the spanwise micropillar electrode spacing and inter-electrode spacing on the performance of EHD micropumps. In particular two pump configurations were investigated: (i) symmetric, where both the emitter and collector electrodes were composed of micropillars and (ii) asymmetric, where the emitter electrode was composed of micropillars and the collector was a planar electrode. The microfabrication of the devices is described in the next section, followed by the presentation and discussion of the experimental results on the performance of the micropumps. Finally, the conclusions of the study are presented. 2. Microfabrication and experimental setup The micropillars (50 mm 50 mm square) were formed by electro depositing nickel on top of planar electrodes placed along the bottom wall of a microchannel as shown schematically in Fig. 1. A schematic of the symmetric and asymmetric configurations are shown in Fig. 2. The symmetric electrode configuration consisted of micropillar emitter and collector electrodes (Fig. 2a) whereas the asymmetric electrode configuration consisted of micropillar emitter electrodes and flat collector electrodes (Fig. 2b). Six different micropumps that were 5 mm wide and 100 mm in height were fabricated and tested with symmetric and asymmetric electrodes to investigate the effect of micropillar distance (i.e. the distance between the micropillars along the spanwise direction) and the inter-electrode spacing on the micropump performance. The geometric and design parameters for the micropumps are given in Table 1. The main design parameters for the micropumps are the inter-electrode spacing des, inter-electrode pair spacing deds,
spanwise micropillar distance dp, microchannel height H and width W, and number of electrode pairs Ne. To be consistent, the number of electrode pairs for all pump designs was kept constant at 100, and the length of the microchannel was set to cover all the electrode pairs. The electrodes were microfabricated using a two mask microfabrication process as illustrated in Fig. 3. First, a 2400 Å layer of gold was deposited on a glass substrate with a 100 Å layer of chromium as the seed layer in order to bind the thin film of gold to the glass substrate. A thick positive photoresist (AZ P4620 from MicroChem Corp) was patterned to form the mould for the micropillar electrodes. AZ P4620 was spun cast at 500 rpm for 30 s and ramped to 2000 rpm in 2 s. The sample was baked at 90 C and ramped to 120 C for 6 min. After exposing for 65 s at 7.2 mJ/s, the sample was developed in 1:3 AZ K400 to DI water for 8 min. Nickel was electroplated to fill the mould (Fig. 3b). The nickel-plating solution consisted of 200 g/l nickel sulfate, 5 g/l nickel chloride, 30 g/l boric acid, and 30 ml/l of wetting agent. The photoresist was subsequently removed by rinsing the sample with acetone. Subsequently, S-1808 photoresist was spun cast and patterned to delineate the interdigitated electrode base pattern. The gold and chromium layers were etched using commercial gold and chromium etchant. Finally, the photoresist was removed by rinsing the substrate with acetone and deionized water. A photograph of the top view of the electrodes, with SEM images of the micropillars is shown in Fig. 4. The microchannel was fabricated by casting polydimethylsiloxane (PDMS) on top of an SU-8 (MicroChem Corp.) mould which was patterned to delineate the microchannel structure. A silicon wafer was plasma oxidized for 1 min at 50 W in advance to improve adhesion properties. SU-8 100 was spun at 3000 rpm on a silicon wafer for 30 s to spread a 100 mm thick layer. The resist was soft baked for 10 min at 65 C, and for 30 min at 95 C. The sample was exposed for 90 s at 7.2 mJ/s using the microchannel negative mask. Following exposure, the sample was post baked for 1 min at 65 C, and for 10 min at 95 C. A 1:10 curing agent to PDMS mixture was used to cast the microchannels with a 100 mm height. The cured PDMS channels were peeled off from the SU-100 mould, and holes were punched on them in order to attach glass tubing for the inlet and outlet of the pump. Finally, the microchannels were adhered to the electrode glass substrate using a combination of plasma oxidization and liquid PDMS as glue. The PDMS microchannel was aligned on the electrode substrate relative to the electrodes. Both PDMS channel and glass wafer were plasma
Fig. 3. Schematic of microfabrication process for EHD micropump.
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
379
Fig. 5. Variation of current with applied voltage for micropumps with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
Fig. 4. Top view of gold electrode base and SEM images showing micropillar array and single micropillar.
oxidized for 1 min at 50 W. Then a 1:3 curing agent to liquid PDMS mixture was poured on the microchannel border and baked at 150 C. The PDMS prepolymer on the microchannel boundary bonds the microchannel with the underlying glass layer. This prevents the leakage problems associated with PDMS microchannels with integrated electrodes. The micropumps were tested with HFE-7100 as the working fluid for the maximum pressure generation under a no flow condition. A Keithley 237 power supply with maximum output voltage and current of 1100 V and 10 mA was connected to the embedded emitter and collector electrode pads. Short glass tubings of 0.5 mm diameter were connected to the inlet and outlet of the micropump that functioned as the fluidic interconnections. Transparent plastic tubing of 1 mm diameter was subsequently connected to the glass tubings. For the static pressure generation tests, the plastic tubings were positioned vertically and the micropump was filled with HFE-7100 using a syringe so that the fluid level could be easily detected in both the inlet and outlet columns to ensure no entrapped bubbles. The fluid level in both the inlet and outlet columns was equalized before turning on the
Fig. 6. Variation of pressure generation with applied voltage for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
Fig. 7. Variation of pressure generation with current for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
380
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
0
500
1000
1500
2000
2500
Fig. 8. Variation of power consumption with pressure generation for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
power supply. The pressure head was determined at different applied voltages by measuring the difference in the fluid level in the inlet and outlet columns using a height gage with a measurement uncertainty of 0.1 mm. 3. Results and discussion The nomenclature used to describe the micropumps is given in Table 1, along with the key dimensions. The first letter is used to denote either the symmetric (S) or asymmetric (A) case, followed by the spanwise micropillar electrode spacing (P) and then the inter-electrode spacing (S). The current vs. voltage characteristics of two micropumps, S-P160-S80 with the symmetric electrode configuration and A-P160-S80 with the asymmetric electrode configuration, are plotted in Fig. 5. For both pumps the current increases exponentially with the applied voltage. Interestingly, the current for the asymmetric micropump is higher than that of the symmetric micropump and the difference becomes more significant as the applied voltage increases above 600 V. The discharge
characteristics of the two pumps are different since in one case the collector has planar electrodes compared to micropillar collector electrodes for the other. The pressure generation in the two pumps is plotted against voltage and discharge current in Figs. 6 and 7. For both pumps, the pressure increases exponentially with the applied voltage while it has a nearly linear relation to the current. However, the pressure generation of the asymmetric electrode configuration is significantly better than the symmetric configuration. For example, the maximum pressure head generated by S-P160-S80 is 1540 Pa at an applied voltage of 1100 V while the maximum pressure head generated with A-P160-S80 is 2240 Pa at an applied voltage of 900 V. The difference in the performance of the two pumps is very interesting, since this indicates that the asymmetry in the electrode geometry results in favorable electric field and charge distributions within the flow domain for pressure generation. In incompressible single phase pumping, the Coulomb force is the dominant electric body force, which is dependent on both the electric field and charge distribution within the domain. The discharge current of A-P160S80 is higher than that of S-P160-S80 which corroborates the higher pressure generation of this micropump. The asymmetric electrode configuration, however, generates a significantly higher pressure head compared to the symmetric electrode configuration at the same applied voltage which can not only be attributed to higher current density. The power consumption for both micropumps as a function of pressure generation is shown in Fig. 8. The power consumption increases nearly linearly with the pressure head generation. The A-P160-S80 has a lower power consumption compared to S-P160-S80 for any given pressure generation. This can be attributed to the better discharge characteristics of the asymmetric pump compared to the symmetric pump and lower pressure losses because of its electrode structural geometry. In order to better understand the differences between the performance of these micropumps, the electric field within the pump domain was simulated numerically in the absence of the fluid flow and electric charge density distribution using a finite element method on a commercial software package (COMSOL Multiphysics). The electric field in the y-direction (flow direction) in the x-z cross section of the microchannel for the micropumps with symmetric and asymmetric electrode configurations with
Fig. 9. Electric field in Y direction contour in cross section between micropillars normal to Y direction for micropump with symmetric electrode configuration. (Y is the flow direction).
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
381
Fig. 10. Electric field in Y direction contour in cross section between micropillars normal to Y direction for micropump with asymmetric electrode configuration. (Y is the flow direction).
a
b
Fig. 11. Variation of current with applied voltage for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillar distance of B 80 mm, , 120 mm, and 6 160 mm.
a
b
Fig. 12. Variation of pressure with applied voltage for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillars distance of B 80 mm, , 120 mm, and 6 160 mm.
382
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
a
a
b
b
Fig. 13. Variation of pressure head generation with micropillars distance for micropumps with (a) symmetric and (b) asymmetric electrodes with inter-electrode spacing of 80 mm at constant applied voltages of 6 700 V, B 800 V, and > 900 V.
inter-electrode spacing of 80 mm are shown in Figs. 9 and 10. The integration of the electric field over the whole domain in the flow direction is approximately the same for both electrode configurations. The electric field intensity between the micropillars contributes to the electric body force in the flow direction, but the collector micropillars act as obstacles to the fluid flow. Although in the steady state the net flow rate is zero, local flows and back flows are present in the microchannel. Therefore, the electrode geometry can result in internal loss and thus affect the pressure head. The micropump with the asymmetric electrode configuration has a more homogeneous electric field in the flow direction over the microchannel and the collectors provide no blockage to the flow. The effect of the spanwise micropillar spacing on the pump performance was investigated for the symmetric and asymmetric electrode configurations. The current-voltage characteristics for three different spanwise pillar distances of 80 mm, 120 mm, and 160 mm are shown in Fig. 11. The inter-electrode spacing was kept constant at 80 mm. The discharge characteristics of the micropump improves with a decrease of the micropillar spacing for the micropumps with symmetric electrode configuration. This is expected since introducing more micropillars per unit length in the spanwise direction would increase the discharge. Surprisingly, the trend is opposite for the micropumps with the asymmetric electrode configuration. In this case, the discharge current is approximately equal at applied voltages
Fig. 14. Variation of power consumption with pressure head for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillars distance of B 80 mm, , 120 mm, and 6 160 mm.
lower than 600 V, but as the applied voltage increases beyond 600 V the micropump with a larger micropillar distance shows better discharge characteristics. At an applied voltage of 900 V, the current of the micropump with a micropillar distance of 160 mm is more than two times higher than the micropump with a micropillar distance of 80 mm. The reason for this is unclear at this time. The corresponding pressure generation is plotted as a function of the applied voltage for the two micropump geometries in Fig. 12. The pressure characteristics for both micropumps are similar to that of the current characteristic with respect to the micropillar spacing. To better distinguish the effect of the micropillar spacing, the pressure is plotted as a function of the micropillar spacing in Fig. 13. For the symmetric electrode configuration, the pressure increases with a decrease of the micropillar spacing, while for the asymmetric electrode configuration the pressure decreases with a decrease of the micropillar spacing. At an applied voltage of 800 V, the micropump with symmetric electrode configuration and micropillar distance of 80 mm generated a pressure head of 580 Pa, and this reduced to 215 Pa when the micropillar distance was increased to 160 mm. On the other hand, for the asymmetric electrode configuration, the micropump with micropillar distance of 160 mm and 80 mm generated a pressure head of 1320 Pa and 400 Pa, respectively. The power consumption for the micropumps is plotted against the pressure head for different micropillar spacings in Fig. 14. The power consumption decreases with a decrease of
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
micropillar distance for the micropumps with symmetric electrode configuration while it increases with a decrease of micropillar distance for the micropump with asymmetric electrode configuration distance, which is consistent with the current and pressure head generation characteristic. 4. Conclusions An experimental study was performed to investigate the effect of a micropillar electrode geometry on the performance of EHD micropumps. Six different micropumps with symmetric and asymmetric configurations with different micropillar spacing in the spanwise direction were microfabricated and tested for this study. For the symmetric case, both the emitter and collector electrode composed of micropillars, while for the asymmetric case, the emitter was composed of micropillars while the collector was a planar electrode. The micropumps were tested under a no net flow condition with HFE-7100 as the working fluid. The micropumps with the asymmetric electrode configuration generated a higher pressure and consumed lower power per unit pressure compared to the micropumps with the symmetric electrode configurations and same inter-electrode spacing. A maximum pressure head generation of 2240 Pa was achieved at an applied voltage of 900 V for the asymmetric configuration. Numerical simulations for the electric field within the pump domain suggest that the electric field vectors in the flow direction are approximately equal for both symmetric and asymmetric electrodes. Therefore, the higher pressure generation of asymmetric electrodes can be attributed to its better discharge characteristic and its lower pressure losses due to the electrode structural geometry. In the micropumps with symmetric electrode configuration, the pump performance increases with a decrease of the spanwise spacing of the micropillars, whereas for the asymmetric electrode configuration the pump performance decreases with an increase of the micropillar spanwise spacing.
383
References [1] O.M. Stuetzer, On-drag pressure generation. J. Appl. Phys. 30 (1959) 984e994. [2] W.F. Pickard, Ion-drag pumping: I Theory. J. Appl. Phys. 34 (1963) 246e250. [3] W.F. Pickard, Ion-drag pumping: II Experiments. J. Appl. Phys. 34 (1963) 251e258. [4] D.J. Laser, J.G. Santiago, A review of micropumps. J. Micromech. Microeng. 14 (2004) R35eR64. [5] J. Darabi, C. Rhodes, CFD modeling of an ion-drag micropump. Sens. Actuators A Phys. 127 (2006) 94e103. [6] W. Panofsky, M. Philips, Classical Electricity and Magnetism, second ed. Addison-Wesley, 1952. [7] J.S. Chang, A. Watson, Electromagnetic hydrodynamics. IEEE Trans. Dielectr. Electr. Insul. 1 (1994) 871e895. [8] V. Benetis. Experimental and Computational Investigation of Planar Ion Drag Micropump Geometrical Design Parameters, Ph.D. Thesis, Mechanical Engineering, University of Maryland, 2005. [9] J.E. Bryan, J.S. Yagoobi, Experimental study of ion-drag pumping using various working fluids. IEEE Trans. Electr. Insul. 2 (4) (1990) 647e655. [10] A. Richter, H. Sandmaier. An electrohydrodynamic micropump, IEEE Micro Electro Mechanical Systems. An Investigation of Micro Structures, Sensors, Actuators, Machines and Robots, New York, 1990. [11] S.H. Ahn, Y.K. Kim, Fabrication and experiment of a planar micro ion drag pump. Sens. Actuators A Phys. 70 (1e2) (1998) 1e5. [12] J. Darabi, H. Wang, Development of an electrohydrodynamic injection micropump and its potential application in pumping fluids in cryogenic cooling systems. J. Microelectromech. Syst. 14 (4) (2005) 747e755. [13] P. Foroughi, V. Benetis, M. Ohadi, Design, testing and optimization of a micropump for cryogenic spot cooling applications, 21st IEEE SEMI-THERM Symposium, 2005. [14] J. Darabi, M. Rada, M. Ohadi, J. Lawler, Design, fabrication, and testing of an electrohydrodynamic ion-drag micropump. J. Microelectromech. Syst. 11 (6) (2002) 684e690. [15] V. Benetis, A. Shooshtari, P. Foroughi, M. Ohadi, A source-integrated micropump for cooling of high heat flux electronics, 19th IEEE SEMI-THERM Symposium, 2003. [16] J.M. Crowley, G.S. Wright, Selecting a working fluid to increase the efficiency and flow-rate of an EHD pump. IEEE Trans. Ind. Appl. 26 (1) (1990) 42e49. [17] P.Z. Kazemi, P.R. Selvaganapathy, C.Y. Ching, Electrohydrodynamic micropumps with asymmetric electrode geometries for microscale electronics cooling. IEEE Trans. Dielectr. Electr. Insul. 16 (2) (2009) 483e488. [18] P.Z. Kazemi, P.R. Selvaganapathy, C.Y. Ching, Effect of electrode asymmetry on performance of electrohydrodynamic micropumps. J. Microelectromech. Syst. 18 (3) (2009) 547e554.
Contents lists available at ScienceDirect
Journal of Electrostatics journal homepage: www.elsevier.com/locate/elstat
Effect of micropillar electrode spacing on the performance of electrohydrodynamic micropumps P. Zangeneh Kazemi, P. Ravi Selvaganapathy, C.Y. Ching* Dept. of Mechanical Engineering, McMaster University, Hamilton, ON L8S4L7, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 September 2009 Accepted 21 May 2010 Available online 8 June 2010
Electrohydrodynamic (EHD) micropumps with three-dimensional 50 mm 50 mm micropillar electrodes were fabricated and tested in this study. Two basic electrode configurations were investigated: (i) micropillar emitter and collector electrodes (symmetric) and (ii) micropillar emitter and planar collector electrodes (asymmetric). The micropumps were fabricated by integrating chromium/gold planar electrodes with electroplated 3-D Nickel micropillars on a glass substrate with a 100 mm high PDMS microchannel. The effect of the spanwise micropillar spacing on the pump performance was determined. The pumps were tested using HFE-7100 as the working fluid for the maximum pressure generation under a no flow condition. The micropumps with the asymmetric electrode design generated a significantly higher pressure head than the corresponding micropumps with symmetric electrode configuration for the same applied voltage, with lower power consumption. A decrease in the spanwise spacing of the micropillar electrodes increased the pump performance for the symmetric configuration, while the performance decreased for the asymmetric configuration. Ó 2010 Elsevier B.V. All rights reserved.
Keywords: Micropumps Electrohydrodynamics Ion drag Microelectronics cooling Microfabrication
1. Introduction Microfluidics finds application in a number of diverse areas such as medical diagnostics, drug delivery, drug discovery, chemical synthesis, electronic cooling and environmental analysis. One of the critical components of any microfluidic device is the micropump whose design, functionality and suitability have to be carefully considered for specific applications. Among the different micropump concepts that have been developed, electrohydrodynamic (EHD) micropumps are especially suited for microfluidic electronic cooling applications because they have no moving parts, use very low power and are amenable to rapid smart control. In addition, they can be designed with very small form factors. In EHD pumping, the interaction between an applied electric field and a dielectric fluid results in an electrical body force on the fluid [1e4]. Specifically, an ion drag EHD micropump involves the interaction of an electric field with electric charges, dipoles or particles injected into a dielectric fluid [5]. This Coulombic force on the charged species is the main driving mechanism, where neutral molecules get dragged along with charged ions that are moving between the electrodes, thereby generating fluid flow. The induced electric body force due to the interaction between an applied electric field and a dielectric fluid is given by [6,7] * Corresponding author. Tel.: þ1 905 525 9140x24998; fax: þ1 905 572 7944. E-mail address: [email protected] (C.Y. Ching). 0304-3886/$ e see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.elstat.2010.05.008
1 1 v3 Fe ¼ re $E E2 $V3 þ V E2 $r$ vr T 2 2
(1)
where re is the charge density, 3 is the fluid permittivity, r is the fluid density and T is the fluid temperature. The three components of the electric body force on the right of Eqn (1) are: (i) the electrophoretic or Coulomb force, (ii) the dielectrophoretic force and (iii) the electrostrictive force, respectively. In single phase flows, the dielectrophoretic and electrostrictive forces are, in general, negligible compared to the Coulomb force. The Coulomb force is directly proportional to the electric field and electric charge density. The charge density and electric field are governed by Poisson’s equation and conservation of charge given by
V$E ¼
re 3
(2)
VE ¼ 0
(3)
vre þ V:J ¼ 0 vt
(4)
where the charge current density J is defined as the total charge flux due to bulk motion of fluid. The electric body force can be incorporated into the NaviereStokes equations as an additional body force to couple the effect of electric field and charge distribution on the fluid flow as
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
377
Fig. 1. Schematic of micropump with symmetric electrode configuration.
V$u ¼ 0
r
(5)
vu þ ðu$VÞu ¼ Vp þ r$g V$s þ Fe vt
(6)
To simulate the pumping phenomenon in EHD micropumps, equations (1)e(6) have to be solved simultaneously for charge density distribution, electric potential and fluid flow. Since the nature of charge injection at the emitter electrode in electrohydrodynamics is not fully understood, the proper boundary conditions for the charge density on the electrodes are unknown. Thus, it is not possible to obtain a solution without a model for the charge boundary conditions at the electrodes. Numerous studies have been conducted to characterize the performance of EHD micropumps with various operating conditions, working fluids and geometrical design parameters [8e16]. The first planar EHD micropump, fabricated from an array of gold electrodes on a glass substrate with ethyl alcohol as the working fluid, was developed by Ahn and Kim [11]. A simple one-mask process was used to fabricate the electrodes, with an electrode spacing of 100 mm and an electrode double spacing (distance between each pair of electrodes) of 200 mm. A maximum pressure head of 240 Pa and a flow rate of 50 mL/min were achieved. The pressure and flow rate were found to be proportional to the square of the applied voltage. Benites et al. [15] investigated the effect of channel height, electrode spacing and double spacing on back pressure and flow rate in an ion drag EHD micropump with planar electrodes deposited on the bottom wall of a microchannel using HFE-7100 as the working fluid. A flow rate up to 7.92 mL/min was
reported at an applied voltage of 500 V. More recently, Darabi and Rhodes [5] simulated a micropump with an array of interdigitated symmetric electrodes patterned along the top and bottom walls of a microchannel. The performance of the micropump with electrodes on both the top and bottom walls was considerably higher compared to a pump with electrodes only on the bottom wall. Heretofore, most development for EHD micropumps has been done with 2-D planar electrodes (i.e planar electrodes deposited on the channel wall). In this configuration, however, the electric field gradients are highest close to the channel wall and decreases rapidly away from the wall since the sharp edged corners of the electrodes are confined to the surface. Thus the charge distribution is mainly confined to the near-wall region. A better charge distribution within the pump domain can be achieved using 3-D micropillar electrodes, albeit at the complexity of the microfabrication. With 3-D micropillar electrodes, the total length of sharp edges and surface area of the electrode are increased substantially throughout the pump domain. As a result, the electric field gradient and the space charge density are expected to increase significantly, thereby increasing the Coulomb force within the pump domain and enhancing the EHD pump performance. The penalty, of course, is the increased surface area exposed to the flowing fluid which will increase the pressure drop. Recently, micropumps with 3-D micropillar electrodes were shown to have a better performance than the corresponding 2-D electrode geometry [17]. Surprisingly, a micropump with a 2-D asymmetric electrode geometry (where the width of the emitter was twice that of the collector) was found to have a better performance than a micropump with 3-D micropillar emitter and collector electrodes
Fig. 2. 2-D Cross section of micropump with different electrode configurations: (a) symmetric, (b) asymmetric electrode configuration.
378
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
Table 1 Micropumps design specifications. Micropump
des (mm)
deds (mm)
dp (mm)
Ne
H (mm)
W (mm)
S-P80-S120 S-P120-S80 S-P160-S80 A-P80-S80 A-P120-S80 A-P160-S80
120 80 80 80 80 80
240 160 160 160 160 160
80 120 160 80 120 160
100 100 100 100 100 100
100 100 100 100 100 100
5 5 5 5 5 5
[17,18]. However, the effect of the asymmetry on the flow rate, and the spanwise micropillar electrode spacing and the inter-electrode spacing were not reported. The objective of this study was to determine the effect of the spanwise micropillar electrode spacing and inter-electrode spacing on the performance of EHD micropumps. In particular two pump configurations were investigated: (i) symmetric, where both the emitter and collector electrodes were composed of micropillars and (ii) asymmetric, where the emitter electrode was composed of micropillars and the collector was a planar electrode. The microfabrication of the devices is described in the next section, followed by the presentation and discussion of the experimental results on the performance of the micropumps. Finally, the conclusions of the study are presented. 2. Microfabrication and experimental setup The micropillars (50 mm 50 mm square) were formed by electro depositing nickel on top of planar electrodes placed along the bottom wall of a microchannel as shown schematically in Fig. 1. A schematic of the symmetric and asymmetric configurations are shown in Fig. 2. The symmetric electrode configuration consisted of micropillar emitter and collector electrodes (Fig. 2a) whereas the asymmetric electrode configuration consisted of micropillar emitter electrodes and flat collector electrodes (Fig. 2b). Six different micropumps that were 5 mm wide and 100 mm in height were fabricated and tested with symmetric and asymmetric electrodes to investigate the effect of micropillar distance (i.e. the distance between the micropillars along the spanwise direction) and the inter-electrode spacing on the micropump performance. The geometric and design parameters for the micropumps are given in Table 1. The main design parameters for the micropumps are the inter-electrode spacing des, inter-electrode pair spacing deds,
spanwise micropillar distance dp, microchannel height H and width W, and number of electrode pairs Ne. To be consistent, the number of electrode pairs for all pump designs was kept constant at 100, and the length of the microchannel was set to cover all the electrode pairs. The electrodes were microfabricated using a two mask microfabrication process as illustrated in Fig. 3. First, a 2400 Å layer of gold was deposited on a glass substrate with a 100 Å layer of chromium as the seed layer in order to bind the thin film of gold to the glass substrate. A thick positive photoresist (AZ P4620 from MicroChem Corp) was patterned to form the mould for the micropillar electrodes. AZ P4620 was spun cast at 500 rpm for 30 s and ramped to 2000 rpm in 2 s. The sample was baked at 90 C and ramped to 120 C for 6 min. After exposing for 65 s at 7.2 mJ/s, the sample was developed in 1:3 AZ K400 to DI water for 8 min. Nickel was electroplated to fill the mould (Fig. 3b). The nickel-plating solution consisted of 200 g/l nickel sulfate, 5 g/l nickel chloride, 30 g/l boric acid, and 30 ml/l of wetting agent. The photoresist was subsequently removed by rinsing the sample with acetone. Subsequently, S-1808 photoresist was spun cast and patterned to delineate the interdigitated electrode base pattern. The gold and chromium layers were etched using commercial gold and chromium etchant. Finally, the photoresist was removed by rinsing the substrate with acetone and deionized water. A photograph of the top view of the electrodes, with SEM images of the micropillars is shown in Fig. 4. The microchannel was fabricated by casting polydimethylsiloxane (PDMS) on top of an SU-8 (MicroChem Corp.) mould which was patterned to delineate the microchannel structure. A silicon wafer was plasma oxidized for 1 min at 50 W in advance to improve adhesion properties. SU-8 100 was spun at 3000 rpm on a silicon wafer for 30 s to spread a 100 mm thick layer. The resist was soft baked for 10 min at 65 C, and for 30 min at 95 C. The sample was exposed for 90 s at 7.2 mJ/s using the microchannel negative mask. Following exposure, the sample was post baked for 1 min at 65 C, and for 10 min at 95 C. A 1:10 curing agent to PDMS mixture was used to cast the microchannels with a 100 mm height. The cured PDMS channels were peeled off from the SU-100 mould, and holes were punched on them in order to attach glass tubing for the inlet and outlet of the pump. Finally, the microchannels were adhered to the electrode glass substrate using a combination of plasma oxidization and liquid PDMS as glue. The PDMS microchannel was aligned on the electrode substrate relative to the electrodes. Both PDMS channel and glass wafer were plasma
Fig. 3. Schematic of microfabrication process for EHD micropump.
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
379
Fig. 5. Variation of current with applied voltage for micropumps with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
Fig. 4. Top view of gold electrode base and SEM images showing micropillar array and single micropillar.
oxidized for 1 min at 50 W. Then a 1:3 curing agent to liquid PDMS mixture was poured on the microchannel border and baked at 150 C. The PDMS prepolymer on the microchannel boundary bonds the microchannel with the underlying glass layer. This prevents the leakage problems associated with PDMS microchannels with integrated electrodes. The micropumps were tested with HFE-7100 as the working fluid for the maximum pressure generation under a no flow condition. A Keithley 237 power supply with maximum output voltage and current of 1100 V and 10 mA was connected to the embedded emitter and collector electrode pads. Short glass tubings of 0.5 mm diameter were connected to the inlet and outlet of the micropump that functioned as the fluidic interconnections. Transparent plastic tubing of 1 mm diameter was subsequently connected to the glass tubings. For the static pressure generation tests, the plastic tubings were positioned vertically and the micropump was filled with HFE-7100 using a syringe so that the fluid level could be easily detected in both the inlet and outlet columns to ensure no entrapped bubbles. The fluid level in both the inlet and outlet columns was equalized before turning on the
Fig. 6. Variation of pressure generation with applied voltage for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
Fig. 7. Variation of pressure generation with current for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
380
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
0
500
1000
1500
2000
2500
Fig. 8. Variation of power consumption with pressure generation for micropump with B symmetric (S-P160-S80) and 6 asymmetric (A-P160-S80) electrode configurations.
power supply. The pressure head was determined at different applied voltages by measuring the difference in the fluid level in the inlet and outlet columns using a height gage with a measurement uncertainty of 0.1 mm. 3. Results and discussion The nomenclature used to describe the micropumps is given in Table 1, along with the key dimensions. The first letter is used to denote either the symmetric (S) or asymmetric (A) case, followed by the spanwise micropillar electrode spacing (P) and then the inter-electrode spacing (S). The current vs. voltage characteristics of two micropumps, S-P160-S80 with the symmetric electrode configuration and A-P160-S80 with the asymmetric electrode configuration, are plotted in Fig. 5. For both pumps the current increases exponentially with the applied voltage. Interestingly, the current for the asymmetric micropump is higher than that of the symmetric micropump and the difference becomes more significant as the applied voltage increases above 600 V. The discharge
characteristics of the two pumps are different since in one case the collector has planar electrodes compared to micropillar collector electrodes for the other. The pressure generation in the two pumps is plotted against voltage and discharge current in Figs. 6 and 7. For both pumps, the pressure increases exponentially with the applied voltage while it has a nearly linear relation to the current. However, the pressure generation of the asymmetric electrode configuration is significantly better than the symmetric configuration. For example, the maximum pressure head generated by S-P160-S80 is 1540 Pa at an applied voltage of 1100 V while the maximum pressure head generated with A-P160-S80 is 2240 Pa at an applied voltage of 900 V. The difference in the performance of the two pumps is very interesting, since this indicates that the asymmetry in the electrode geometry results in favorable electric field and charge distributions within the flow domain for pressure generation. In incompressible single phase pumping, the Coulomb force is the dominant electric body force, which is dependent on both the electric field and charge distribution within the domain. The discharge current of A-P160S80 is higher than that of S-P160-S80 which corroborates the higher pressure generation of this micropump. The asymmetric electrode configuration, however, generates a significantly higher pressure head compared to the symmetric electrode configuration at the same applied voltage which can not only be attributed to higher current density. The power consumption for both micropumps as a function of pressure generation is shown in Fig. 8. The power consumption increases nearly linearly with the pressure head generation. The A-P160-S80 has a lower power consumption compared to S-P160-S80 for any given pressure generation. This can be attributed to the better discharge characteristics of the asymmetric pump compared to the symmetric pump and lower pressure losses because of its electrode structural geometry. In order to better understand the differences between the performance of these micropumps, the electric field within the pump domain was simulated numerically in the absence of the fluid flow and electric charge density distribution using a finite element method on a commercial software package (COMSOL Multiphysics). The electric field in the y-direction (flow direction) in the x-z cross section of the microchannel for the micropumps with symmetric and asymmetric electrode configurations with
Fig. 9. Electric field in Y direction contour in cross section between micropillars normal to Y direction for micropump with symmetric electrode configuration. (Y is the flow direction).
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
381
Fig. 10. Electric field in Y direction contour in cross section between micropillars normal to Y direction for micropump with asymmetric electrode configuration. (Y is the flow direction).
a
b
Fig. 11. Variation of current with applied voltage for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillar distance of B 80 mm, , 120 mm, and 6 160 mm.
a
b
Fig. 12. Variation of pressure with applied voltage for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillars distance of B 80 mm, , 120 mm, and 6 160 mm.
382
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
a
a
b
b
Fig. 13. Variation of pressure head generation with micropillars distance for micropumps with (a) symmetric and (b) asymmetric electrodes with inter-electrode spacing of 80 mm at constant applied voltages of 6 700 V, B 800 V, and > 900 V.
inter-electrode spacing of 80 mm are shown in Figs. 9 and 10. The integration of the electric field over the whole domain in the flow direction is approximately the same for both electrode configurations. The electric field intensity between the micropillars contributes to the electric body force in the flow direction, but the collector micropillars act as obstacles to the fluid flow. Although in the steady state the net flow rate is zero, local flows and back flows are present in the microchannel. Therefore, the electrode geometry can result in internal loss and thus affect the pressure head. The micropump with the asymmetric electrode configuration has a more homogeneous electric field in the flow direction over the microchannel and the collectors provide no blockage to the flow. The effect of the spanwise micropillar spacing on the pump performance was investigated for the symmetric and asymmetric electrode configurations. The current-voltage characteristics for three different spanwise pillar distances of 80 mm, 120 mm, and 160 mm are shown in Fig. 11. The inter-electrode spacing was kept constant at 80 mm. The discharge characteristics of the micropump improves with a decrease of the micropillar spacing for the micropumps with symmetric electrode configuration. This is expected since introducing more micropillars per unit length in the spanwise direction would increase the discharge. Surprisingly, the trend is opposite for the micropumps with the asymmetric electrode configuration. In this case, the discharge current is approximately equal at applied voltages
Fig. 14. Variation of power consumption with pressure head for micropumps with (a) symmetric and (b) asymmetric electrodes with micropillars distance of B 80 mm, , 120 mm, and 6 160 mm.
lower than 600 V, but as the applied voltage increases beyond 600 V the micropump with a larger micropillar distance shows better discharge characteristics. At an applied voltage of 900 V, the current of the micropump with a micropillar distance of 160 mm is more than two times higher than the micropump with a micropillar distance of 80 mm. The reason for this is unclear at this time. The corresponding pressure generation is plotted as a function of the applied voltage for the two micropump geometries in Fig. 12. The pressure characteristics for both micropumps are similar to that of the current characteristic with respect to the micropillar spacing. To better distinguish the effect of the micropillar spacing, the pressure is plotted as a function of the micropillar spacing in Fig. 13. For the symmetric electrode configuration, the pressure increases with a decrease of the micropillar spacing, while for the asymmetric electrode configuration the pressure decreases with a decrease of the micropillar spacing. At an applied voltage of 800 V, the micropump with symmetric electrode configuration and micropillar distance of 80 mm generated a pressure head of 580 Pa, and this reduced to 215 Pa when the micropillar distance was increased to 160 mm. On the other hand, for the asymmetric electrode configuration, the micropump with micropillar distance of 160 mm and 80 mm generated a pressure head of 1320 Pa and 400 Pa, respectively. The power consumption for the micropumps is plotted against the pressure head for different micropillar spacings in Fig. 14. The power consumption decreases with a decrease of
P. Zangeneh Kazemi et al. / Journal of Electrostatics 68 (2010) 376e383
micropillar distance for the micropumps with symmetric electrode configuration while it increases with a decrease of micropillar distance for the micropump with asymmetric electrode configuration distance, which is consistent with the current and pressure head generation characteristic. 4. Conclusions An experimental study was performed to investigate the effect of a micropillar electrode geometry on the performance of EHD micropumps. Six different micropumps with symmetric and asymmetric configurations with different micropillar spacing in the spanwise direction were microfabricated and tested for this study. For the symmetric case, both the emitter and collector electrode composed of micropillars, while for the asymmetric case, the emitter was composed of micropillars while the collector was a planar electrode. The micropumps were tested under a no net flow condition with HFE-7100 as the working fluid. The micropumps with the asymmetric electrode configuration generated a higher pressure and consumed lower power per unit pressure compared to the micropumps with the symmetric electrode configurations and same inter-electrode spacing. A maximum pressure head generation of 2240 Pa was achieved at an applied voltage of 900 V for the asymmetric configuration. Numerical simulations for the electric field within the pump domain suggest that the electric field vectors in the flow direction are approximately equal for both symmetric and asymmetric electrodes. Therefore, the higher pressure generation of asymmetric electrodes can be attributed to its better discharge characteristic and its lower pressure losses due to the electrode structural geometry. In the micropumps with symmetric electrode configuration, the pump performance increases with a decrease of the spanwise spacing of the micropillars, whereas for the asymmetric electrode configuration the pump performance decreases with an increase of the micropillar spanwise spacing.
383
References [1] O.M. Stuetzer, On-drag pressure generation. J. Appl. Phys. 30 (1959) 984e994. [2] W.F. Pickard, Ion-drag pumping: I Theory. J. Appl. Phys. 34 (1963) 246e250. [3] W.F. Pickard, Ion-drag pumping: II Experiments. J. Appl. Phys. 34 (1963) 251e258. [4] D.J. Laser, J.G. Santiago, A review of micropumps. J. Micromech. Microeng. 14 (2004) R35eR64. [5] J. Darabi, C. Rhodes, CFD modeling of an ion-drag micropump. Sens. Actuators A Phys. 127 (2006) 94e103. [6] W. Panofsky, M. Philips, Classical Electricity and Magnetism, second ed. Addison-Wesley, 1952. [7] J.S. Chang, A. Watson, Electromagnetic hydrodynamics. IEEE Trans. Dielectr. Electr. Insul. 1 (1994) 871e895. [8] V. Benetis. Experimental and Computational Investigation of Planar Ion Drag Micropump Geometrical Design Parameters, Ph.D. Thesis, Mechanical Engineering, University of Maryland, 2005. [9] J.E. Bryan, J.S. Yagoobi, Experimental study of ion-drag pumping using various working fluids. IEEE Trans. Electr. Insul. 2 (4) (1990) 647e655. [10] A. Richter, H. Sandmaier. An electrohydrodynamic micropump, IEEE Micro Electro Mechanical Systems. An Investigation of Micro Structures, Sensors, Actuators, Machines and Robots, New York, 1990. [11] S.H. Ahn, Y.K. Kim, Fabrication and experiment of a planar micro ion drag pump. Sens. Actuators A Phys. 70 (1e2) (1998) 1e5. [12] J. Darabi, H. Wang, Development of an electrohydrodynamic injection micropump and its potential application in pumping fluids in cryogenic cooling systems. J. Microelectromech. Syst. 14 (4) (2005) 747e755. [13] P. Foroughi, V. Benetis, M. Ohadi, Design, testing and optimization of a micropump for cryogenic spot cooling applications, 21st IEEE SEMI-THERM Symposium, 2005. [14] J. Darabi, M. Rada, M. Ohadi, J. Lawler, Design, fabrication, and testing of an electrohydrodynamic ion-drag micropump. J. Microelectromech. Syst. 11 (6) (2002) 684e690. [15] V. Benetis, A. Shooshtari, P. Foroughi, M. Ohadi, A source-integrated micropump for cooling of high heat flux electronics, 19th IEEE SEMI-THERM Symposium, 2003. [16] J.M. Crowley, G.S. Wright, Selecting a working fluid to increase the efficiency and flow-rate of an EHD pump. IEEE Trans. Ind. Appl. 26 (1) (1990) 42e49. [17] P.Z. Kazemi, P.R. Selvaganapathy, C.Y. Ching, Electrohydrodynamic micropumps with asymmetric electrode geometries for microscale electronics cooling. IEEE Trans. Dielectr. Electr. Insul. 16 (2) (2009) 483e488. [18] P.Z. Kazemi, P.R. Selvaganapathy, C.Y. Ching, Effect of electrode asymmetry on performance of electrohydrodynamic micropumps. J. Microelectromech. Syst. 18 (3) (2009) 547e554.