ON THE PERFORMANCE OF A MINIATURE ROTARY SHAFT PUMP (RSP) G. Pavesi – G. Ardizzon – A. Rossetti Department of Mechanical Engineering - University of Padova Via Venezia, 1, 35131 Padova – Italy Tel. +39 049 8276763 Fax. +39 049 8276785 Email [email protected]ABSTRACT This paper presents a new concept of designing and manufacturing of a micro rotary shaft pump (RSP). An impeller made by boring a hole in one end of the shaft and cutting slots in the side of the shaft at the bottom of the bored hole, such that the metal between the slots de- fines the impeller blades, is the key component of this micro pump. The impeller’s diameter measured 2.5 mm, with the volute width of 0.4 mm. The micro pump has been made of AISI 304, offering good biocompatibility and chemical resistance. Two impeller designs were tested over a range of operating conditions. Pump performance characteristics, including pressure rise up to 2.5 kPa, and flow rate up to 60 ml/min are pre- sented for several different pump speed. Furthermore, 3_D numerical simulations were performed to model the micropumps. Focus of the simulations was to study the blade height effects, and the volute influence. NOMENCLATURE C m meridional velocity component [m/s] g gravitational constant [m/s 2 ] h pump head [Pa] M Torque [Nm] n shaft speed [rpm] p pressure [Pa] Q flow rate [ml/min] U peripheral velocity [m/s] η Hy hydraulic efficiency [-] 2 2 m U C = ϕ flow coefficient [-] ρ density [kg/m 3 ] ω angular velocity [s -1 ] 2 2 U gh = ψ head coefficient [-] Subscripts 1 rotor blade leading edge 2 rotor blade trailing edge Superscripts ° total INTRODUCTION Micro-Electro-Mechanical Systems (MEMS) technology is one of the ultimate goals of today’s micro-fluidic research. For an effective regulation of small volume of fluids in micro-channels, the micro-pump is one of the most critical parts of this emerging research field (Nguyen, and Wereley, 2002). Micropumps can be classified into two categories, the non-mechanical and the mechanical, mainly attributed by the pumping mechanism (Nguyen, and Wereley, 2002, Nguyen, et al., 2002). The first category adds momentum to the fluid for pumping effect by converting another energy 1
Transcript
ON THE PERFORMANCE OF A MINIATURE ROTARY SHAFT PUMP (RSP)
G. Pavesi – G. Ardizzon – A. Rossetti
Department of Mechanical Engineering - University of Padova Via Venezia, 1, 35131 Padova – Italy
ABSTRACT This paper presents a new concept of designing and manufacturing of a micro rotary shaft
pump (RSP). An impeller made by boring a hole in one end of the shaft and cutting slots in the side of the shaft at the bottom of the bored hole, such that the metal between the slots de-fines the impeller blades, is the key component of this micro pump. The impeller’s diameter measured 2.5 mm, with the volute width of 0.4 mm. The micro pump has been made of AISI 304, offering good biocompatibility and chemical resistance.
Two impeller designs were tested over a range of operating conditions. Pump performance characteristics, including pressure rise up to 2.5 kPa, and flow rate up to 60 ml/min are pre-sented for several different pump speed.
Furthermore, 3_D numerical simulations were performed to model the micropumps. Focus of the simulations was to study the blade height effects, and the volute influence.
NOMENCLATURE Cm meridional velocity component [m/s] g gravitational constant [m/s2] h pump head [Pa] M Torque [Nm] n shaft speed [rpm] p pressure [Pa] Q flow rate [ml/min] U peripheral velocity [m/s] ηHy hydraulic efficiency [-]
Ugh=ψ head coefficient [-] Subscripts 1 rotor blade leading edge 2 rotor blade trailing edge
Superscripts ° total
INTRODUCTION Micro-Electro-Mechanical Systems (MEMS) technology is one of the ultimate goals of today’s
micro-fluidic research. For an effective regulation of small volume of fluids in micro-channels, the micro-pump is one of the most critical parts of this emerging research field (Nguyen, and Wereley, 2002). Micropumps can be classified into two categories, the non-mechanical and the mechanical, mainly attributed by the pumping mechanism (Nguyen, and Wereley, 2002, Nguyen, et al., 2002). The first category adds momentum to the fluid for pumping effect by converting another energy
1
form into the kinetic energy. These pumps utilize effects which are dominant in micro-scales including: electro-hydrodynamic (EHD) pump (Bart, et al., 1997, Mc Bride, et al., 1998, Richter et al., 1991) which pump dielectric liquids with extremely low electric conductivity; electro-kinetic [Manz, et al., 1994, Paul et al, 1998] which use the electro-osmotic and electrophoretic effects for molecular separation, magneto-hydrodynamics (MHD) (Jang, and Lee, 2000, Lemoff, and Lee, 2000, Huang, et al., 1999, Huang, et al., 2000), bubble pumps (Jun, 1996), and diffuser/nozzle pumps (Olsson, 1997) which can be used to pump any type of liquid.
Field-induced flow pumps have the advantage of requiring no moving parts, easing the fabrica-tion, sealing, and operation of the systems. Nor is there wear and fatigue problems caused by high pressure-drops across the check valves. But since the viscous force in micro-channels increases in the second order with the miniaturization, this pump category cannot deliver enough power in order to overcome its high fluidic impedance.
All mechanical pumps require an actuator, which converts electric energy into mechanical work. In most micropumps, a membrane is used as external or integrated actuator and it is set in motion by several different principles such as electrostatic (Zengerle et al., 1995), pneumatic (Branebjerg, et al., 1994, Rapp, et al., 1994), thermo-pneumatic (Zdeblick, et al., 1994, Jeong, and Yang, 2000), piezoelectric (Van Lintel, et al., 1988, Moroney et al., 1991), and electromagnetic (Zhang, and. Ahn, 1996). The biggest drawback of external actuators is their large size, which limits the miniaturization of the entire micro-pumps. The advantage is the relatively large force and displacement generated by external actuators. Despite their fast response time and good reliability, integrated electrostatic actuators cause small force and very small stroke.
Thermo-pneumatic actuators and bimetallic actuators require a large amount of thermal energy for their operation, and consequently, consume a lot of electric power. High temperature and complicated thermal management are further drawbacks of these actuator types. Electromagnetic actuators require an external magnetic field, which also restricts the pump size. Their large electric current causes thermal problems and high electric energy consumption.
Moreover, the flow produced by membrane pumps is pulsed rather than continuous. In contrast, dynamic pumps seem to have a wide variety of possible working fluids and applica-
tions. The performances of macro-scale centrifugal pumps are well known and established. How-ever, characteristic, and efficiency of centrifugal pumps change as the size of the pump are altered. Secondary flows, viscous forces, and the losses associated with them, become more significant as macro-scale impeller passage size decreases (Karniadakis, and Beskok, 2002). These effects can dominate the performance of the pump, but do not prevent the operation of the centrifugal micro and mini-pumps (Hainan, et al., 1997, Ahn, and Allen, 1995 and Blanchard, et al., 2006).
A significant disadvantage of the tested centrifugal micropump has been the gap distance be-tween the top of the blades, and the pump housing. The tip clearance is generally irrelevant on a macro-scale, oppose to on a micro-scale, or a millimetre-scale, where the gap between the blade tip and the pump housing can be about the same as the height of the impeller blades. Further, if the gap is too small or zero, the blades can be damaged by contacting the pump housing. Consistently, the overall hydraulic efficiency of the pump decreases due to the relevant leakage.
To eliminate this drawback the impellers under test have been integrated into the body of the shaft, instead of on the top of the shaft. By connecting the top of the blades to the shaft, the gap at the top of the blades has been reduced to zero. Pump performance has been characterized by the pressure rise, flow rate, and the hydraulic efficiency for a wide range of impeller speeds and impel-ler configurations. Tested impeller designs include two radial 4-blade arrangements with different diffusion angle and blade width.
The influence modification of the geometrical parameters, as well as of the angular velocity has been also studied numerically, in order to lay the base for a numerical optimization. The transient performances have been compared to the available experimental data.
2
MICROPUMP DESIGN Components were all ma-
chined using a computer numeric control machining centre (CNC). First, IGES formats of the compo-nents 3D files were generated. Then, the IGS files were used to generate machine codes. Special carbide tools were also prepared, modified, shaped and installed on the control machining centre. Each component was machined at high speed and low feed rate to assure accuracy while preventing damage
to thin profiles and sections.
pressure port Brushless motor
Coupling
Fig. 1 depicts the assembled micropump. There are three main components of the rotary shaft pump assembly: the impeller, the volute, and the bearings.
The AISI 304 stainless steel impeller has been realized using precision machining techniques. A lathe has been used to obtain the desired outside diameter of 2.5 mm, and to bore a hole of 1.0 mm in the end of the shaft. A countersink with an angle of 60° has also been made at the inlet of the shaft using a drill to reduce the entrance losses. Slots are then cut into the side of the shaft at the bottom of the bored hole using both, a milling machine and an indexing tool, and electrical-discharge proc-esses.
The metal between the slots act as the blades of the impeller, and the slots form the blade pas-sages. Two radial 4-blade impellers have been constructed for testing. The blades on the im-pellers are evenly spaced. The blade width of the
Impeller A was of 0.4 mm, with a diffusion angle of 60° (fig. 2). Whereas, the Impeller B was constructed with smaller width, 0.3 mm and diffusion angle of 40° (fig. 2).
The impeller has been mounted using two PTFE (Teflon) brasses, press fit in the volute housing, and located above and below the slots, as bearing. The bearings were designed to form a seal for the impeller, reducing the leakage to almost zero from the impeller outlet to the RSP inlet. The outer diameter of the bearings is 8 mm, and the inner diameter is 2.5 mm.
Surface forces significantly influence the flow from the impeller exit and within the volute. Therefore in the attempt to reduce viscous losses, the volute has been designed as an “open” one (Blanchard, 2005) to decrease the average fluid velocity, and velocity gradient. A V-shape channel from the
Fig. 1 Assembled Micropump
60°0.4
0.3
40°
Impeller slot
PTFE bearings
Impeller Impeller B
60°
2.5 1
40°
2.5 1
Fig. 2 Cutaway views of the two RSP
impellers 9°
0.4
O 2.5
O 2.5
O 4.5
O 6
Fig. 3 Cutaway views of the RSP
volute.
3
impeller to the outlet channel, as shown in figure 3 characterizes the tested volute. The volute and part of the outlet and inlet channels were machined from a Plexiglas stock using
a CNC milling machine (fig. 3). A height of 0.4 mm was used in all the current tests, as the height could not be less than the higher slots milled in the shaft.
The top volute housing has been design with the inlet and outlet channels as seen in fig. 3. The pressure ports have been connected to the channel passages, (fig. 1), and are just above the bearing at the inlet and about two millimetres from the volute outlet. The diameter of the pressure ports at the fluid channel interface is about 2 mm.
The volute and impeller wet surface roughness was of 0.2 μm after electropolishing.
TEST FACILITY The experiments were performed in an open facility, which was expressly designed to test the
micropump. A water reservoir has been connected to the pump inflow by a plastic tube with an in-ner diameter of 3 mm. The reservoir was filled continuously to a weir brim by an external pump, so that the water level change during operations could be considered negligible. The outlet was con-nected by a plastic tube (inner diameter of 3 mm) to a filling channel. The channel was mounted to the base of a linear slide, and its vertical position could be changed during the tests.
An externally mounted brushless DC motor with a maximum speed of 50 000 rpm, powered the impeller. The speed was maintained at a constant, with the accuracy of 1 rpm, for any variation in torque by the Hall effect sensors integrated in the brushless motor and controlled by a PC. The maximum testing speed has been limited, in these pilot surveys, at 20 000 rpm.
The pressure ports, shown in fig. 1, were connected to a DP-15 differential Validyne pressure sensor through two tubes. The pressure sensor diaphragm used in this transducer had a full range of 3.5 kPa and an accuracy of ±0.25% FS. A carrier demodulator has processed the output signal from the pressure sensor. The unit provided a Vdc output, suitable for recording by a device with a dy-namic range of 12-bit.
The brushless motor and the impeller were mounted to the base of a micrometric linear slide, while the volute and the housing were mounted to the shuttle of the same slide. The correct relative position of the impeller and the volute, were obtained moving the slide, as well as, using a micro-scope to aid the positioning of the impeller shaft.
After assembly and positioning of the impeller shaft and housing, all the air was bled. The pump motor has been activated, and adjusted to produce the desired speed. When the steady state has been reached, the timer was then started, and water from the outlet filling channel was collected. Output signals related to pump rotational speed, pressure rise, electric power absorbed, and volumetric flow rate were then stored. The flow rate was determined by weighing the amount of water collected.
A first-order uncertainty analysis is per-formed using a constant odds combina-tion method, based on a 95% confidence level. The resulting uncertainty magni-tudes associated with experimentally measured pressure rise, flow rate, and rotational speed, are shown in table 1.
Table 1 - Uncertainties associated with experimental data
Variable Uncertainty h [Pa] 2.9 %
Q [ml/min] 2.5 % n [rpm] ± 1
NUMERICAL PROCEDURE In the present investigation, the two RPS impeller have been modelled, and performed by using
a commercial software. The models have been meshed with tetrahedral elements using higher grid densities near walls. A grid adaptation has been performed at impeller and volute inflows, so as to describe the step width variation of 0.05 mm with at least 5 elements. In addition to this meshing method the pump’s walls had divided into non-uniformly spaced elements using pave meshing scheme with a ratio 1.02 to assure a smooth mesh variation. The minimum mesh quality admitted
4
was 70%. All computations for the pump were performed using the unsteady model. The flow was as-
sumed laminar and incompressible, and no dissipative damping techniques have been used. The standard transient sliding interface approach was used to deal with the interface between stator/rotor blocks.
The time step utilized for simulating the transient process of the micro-screw pump was re-quired to be small to pick the changes over time in the flow. The time step for the explicit scheme was chosen according to an impeller rotation of one degree, resulting a Courant Number of approximately CFL=0.75. For the discretisation in time, a second order dual time stepping scheme was adopted. The maximum number of iterations for each time step had been set to 10, in order to give mass residues of 10-6, and momentum residues of 10-5.
The inlet boundary conditions consisted of prescribing mass flow rates that have been obtained from experimental data. The exit boundary condition consisted of prescribing the experimental exit pressure level with no-slip condition being imposed at the surfaces. To aid in interpretation and comparison of the unsteady simulation with the experimental data, the numerical results have been averaged out over the impeller rotation of 90° between 720° and 810°.
RESULTS The characteristics of head on flow rate for the two impeller are illustrated in the figs 4 and 5.
These experimental data were obtained using water as the working fluid, with rotational speeds be-tween 10000 and 20000 rpm. As the rotational speed was increased, the head and flow rate in-creased for both configurations.
Figs 4 and 5 show that by increasing the slopes width, an expected higher range of flow rate through the pump is seen, but greatly, affects the head. At lower shaft speed, data sets for both
impeller configurations exhibit completely different slopes, as well as different quan-titative trends. With increased rotation speed the discrepancies were reduced, al-though the impeller A attains greater head. The maximum head obtained at a rota-tional speed of 20000 rpm is approxi-mately h=2.55 kPa for impeller A configu-ration, whereas the impeller B is able to obtain only 2.1 kPa.
Figure 6 shows the non-dimensional tion of head versus flow rate for the
two radial impellers, and for different rota-tional speeds. The affinity laws are not applied at the Impeller A whereas the non-dimensional data curves for the Impeller B are sufficiently similar.
varia
0
500
1000
1500
2000
2500
0 10 20 30 40 50 60Q [ml/min]
h [Pa]10000 rpm12500 rpm15000 rpm20000 rpm
Fig. 4 Experimental pump A characteristics.
The maximum non-dimensional flow rate, obtained for Impeller B, is about φ=0.14 for all the rotational speed, while the higher flow coefficient for the Impeller A has been proved to be quite constant only for rotation speed grater than 15000 rpm (fig. 6).
The contradictory behaviour appears to be due to the different Reynolds number
range operating in the two pumps. Impeller B with the lower blade width was tested with a mean flow Reynolds number between 1000 and 3500 while Impeller A was tested with Reynolds number up to 2500.
The majority of experiments (Hetsroni et al. (2005)) have concluded that transition from lami-nar to turbulent flow in smooth and rough circular micro-tubes occurred at Reynolds numbers about 2000 corresponding to those in macro-channels. Therefore, the Impeller A was investigated in fully developed laminar flow, at lower shaft speed, and in transition from laminar to turbulent flow at higher impeller speed. As opposes to, Impeller B was tested in transition from laminar to
turbulent flow for all the shaft speed. The head reflected the dependence of
the drag coefficient on the Reynolds num-ber and, at shut off, it decreased increasing the impeller speed whereas the slope of the non-dimensional data curves increased with the shaft velocity (fig. 6).
To enrich the experimental data analy-sis a numerical solution was developed. Figures 7 and 8 present the comparison between the experimental data and the numerical results for both pumps.
The preliminary results indicate meas-urable differences in behaviours. The nu-merical head of the prototype pump A is, generally, higher than those obtained ex-perimentally with an error increasing as the pressure head is increased. At low speeds, the experimental head is closer to the analytical predictions but still a little bit higher. For the Impeller B the numeri-cal results are, constantly overestimated.
Fig. 6 Non-dimensional characteristics of the pumps
This could be attributed to the afore-mentioned transition from laminar to tur-
ent effects, which were more pro-nounced at higher speed and for the im-peller B.
bul
O
-500
0
500
1000
1500
2000
2500
3000
0 10 20 30 40 50 6Q [ml/min]
h [P
a]
0
20000 rpm 20000 rpm10000 rpm 10000 rpm
NumericalExperimen
Fig. 7 Comparison of numerical and experimental Pump A characteristics The effect of blade design was particu-
larly important since it played a dominant role in the flow field inside the impeller and in the volute inflow (Figs 9 and 10), and therefore in determining the losses inside the pump.
The analysis of the mass-averaged to-tal pressure, inside the pump (figs 10 and 11), shows, that the impeller A was able to increase the head 25% more than the impeller B.
ne element for this head rise was the increase in Euler work due to the reduced relative velocity at the exit of the impeller.
-500
0
500
1000
1500
2000
2500
3000
0 10 20 30 40 5Q [ml/min]
h [P
a]
0
20000 rpm 20000 rpm10000 rpm 10000 rpm
NumericalExperimen
Fig.8 Comparison of numerical and experimental Pump B characteristics The losses due to the higher diffusion
6
angle and the pronounced wake, in the impeller A, were counterbalanced by the smaller velocity and boundary losses inside the blade channels.
In both impellers, the inflow transferred momentum to the passing fluid through viscous forces. This “pre-swirl” was significant on both the impeller, because the ratio of the in-flow length to diameter was large, and the circumferential wall velocity was greater than the average axial fluid velocity. The “pre-swirl” aided the overall pumping process by reducing the sudden accelera-tion of the fluid at the inner blade tip, and the flow separation near the leading edge of the impeller blade (fig. 9).
The impeller hydraulic efficiency ac-counted numerically by:
( )ω
−=η
°°
MQPP 12
pellerImHy (1)
and shown in figs 12 and 13, decreases as flow rate increases. This is tied to impeller performance at higher flow rates where the greater viscous stresses are present due to larger velocity gradients. The maximum value of hydraulic efficiency was 0.66, which was obtained with the impeller B at a flow rate of approximately 20 ml/min.
There was a difference of only 4 points in the magnitude of the hydraulic effi-ciency between the two impeller. This im-plies that both the impeller design and the flow rate were significant, but the flow rate had a more significant effect on the efficiency.
The lower blade width increased the mean velocity in the slots and the presence of a step at the volute inflow further com-plicated the flow field. It led to large ve-locity gradients and a ‘‘scraping motion’’ in the volute inflow. Therefore, the step at the volute inflow is a critical trade off be-tween, maximizing the cross flow, and the interaction of moving and fixed bounda-ries at the volute inflow. The solution of the flow field, near the intersection of moving and fixed boundaries at the volute inflow, showed that quite all the losses in the volute were located in the first milli-metre far from the impeller (fig. 10 and 11). This was due to the intense vorticity
Fig. 10 Mean total pressure behaviors in Pump A at 20000 rpm.
7
that had its source at impeller outlet (figs 14 and 15). The jet-like character of the volute intake flow, interacting with the wall, created large-scale rotating flow pattern (fig. 9). These flows appeared to become unstable and break down into three-dimensional turbulent motions. These vortices persisted up to approxi-mately 1 millimetre far from the impeller, but their intensity was so high that dissi-pated not less than 50% of the flow en-ergy.
The volute hydraulic efficiency:
( )( )QPP2
VoluteVoluteHy =η
QPP
1
1Exit°°
°°
−
− (2)
decreases quite as a monotonic function of the flow rate (figs 12 and 13) by approxi-mately 50 points for the two pumps. It represented the main reason of the micro-pumps performance deterioration. Figures 12 and 13 shows pump hydraulic effi-ciency and flow-rate variations for the tested RSP.
CONCLUSIONS Rotating shaft micropumps are attrac-
tive because they are robust and easy to fabricate, capable of handling a wide vari-ety of fluids, and can operate with no valves allowing them to handle particle-laden fluids. This paper showed the feasi-bility of this solution investigating two micropump. Both configurations were studied experimentally and numerically.
The analysis presented focused the ef-fect of blade shape and width, and
eller volute interaction on pump performance. The similar characteristics for all impeller designs provided evidence that the trends of the data were influenced
by viscous forces and by impeller volute interaction mainly.
imp
Even though the performances were lower than we expected, the capacity of movement up to 60 ml/min and the pressure rise up to 2.5 kPa were meaningful. In future work, it will be necessary to work on different volute design, to further increase the head to obtain counter-pressure comparable with the micro check-valve solutions.
REFERENCES Ahn, C.H., and Allen, M.G., (1995). Fluid micropumps based on rotary magnetic actuators, Proceed-
Fig. 11 Mean total pressure behaviors in Pump B at 20000 rpm
-100
10203040506070
0 10 20 30 40 50 60Q [ml/min]
ηHy
Impeller Volute Pump
Fig. 12 Hydraulic efficiency in Pump A at 2000 rpm.
-100
10203040506070
0 10 20 30 40 50 60Q [ml/min]
ηHy
Impeller Volute Pump
Fig. 13 Hydraulic efficiency in Pump B at 20000 rpm.
8
ings of MEMS 1995, Jan.-Feb., pp. 408-418. Bart, S.F., Tavrow, L.S., Mehregany, M., (1997). Microfabricated electrohydrodynamic pumps. Sens.
Actuators A 21, pp. 193–197. Blanchard, D., Ligrani, P., and Gale, B., (2005). Performance and Development of a Miniature Rotary
Shaft Pump (RSP). Journal of Fluids Engineering, Vol. 127, pp. 752-760. Branebjerg, B. Fabius, P. Gravesen, (1994). Application of miniature analyzers: From microfluidic
components to mTAS, in: A. van den Berg, P. Bergfeld Eds. , Proceedings of Micro Total Analysis Systems Conference, Twente, Netherlands, Nov. 21–22, 1994, pp. 141–151.
Fig. 14 Vorticity in the volute mid-span plane, Pump A.
Fig. 15 Vorticity in the volute mid-span plane, Pump B
Q=13.63 ml/min Q=20.15 ml/min Q=11.20 ml/min
Q=11.20 ml/min
Q=29.95 ml/min
Q=13.63 ml/min Q=20.15 ml/min
Q=43.85 ml/min Q=58.45 ml/min
Q=29.95 ml/min Q=43.85 ml/min Q=58.45 ml/min
9
Hainan, C., Ahaoying, Z., Yong, L., Xiongying, Y., and Yihua, Y., (1997), A novel centrifugal miniature pump and its medical application. Proceedings of the 1997 International Symposium On Micro-mechatronics and Human Science, IEEE, Oct., pp. 115-117.
Hetsroni, G. ,Mosyak, A. ,Pogrebnyak, E. ,Yarin, L.P.(2005), Fluid flow in micro-channels. Inter-national Journal of Heat and Mass Transfer, vol. 48, pp. 1982–1998.
Huang, L., Wang, W., Murphy, M.C., (1999). A lumped-parameter model for a micro-pump based on the magnetohydrodynamic (MHD) principle. Proc. SPIE–Design Test Microfabric. MEMS MOEMS 3680, pp. 379– 387.
Huang, L., Wang, W., Murphy, M.C., Lian, K., Ling, Z.G., (2000). LIGA fabrication and test of a dc type magneto-hydrodynamic (MHD) micro-pump. Microsyst. Technol. 6, pp. 235–240.
Jang, J., Lee, S.S., (2000). Theoretical and experimental study of MHD micro-pump. Sens. Actuators A 80, pp. 84–89.
Jeong, O.C., Yang, S.S., (2000). Fabrication and test of a thermo-pneumatic micro-pump with a corru-gated p+ diaphragm. Sens. Actuators A 83, pp. 249–255.
Jun, T.K., (1996), in: Microscale pumping with traversing bubbles in microchannels, Solid-State Sensor and Actuator Workshop, Hilton Head, SC, p. 144.
Karniadakis, G.E., Beskok, A., (2002). Micro Flows, Fundamentals and Simulation, Springer, New York, pp. 1-31.
Lemoff, A.V., Lee, A.P., (2000). An ac magnetohydrodynamic micro-pump. Sens. Actuators B 63, pp. 178–185.
Manz, C.S., Effenhauser, N., Burggraf, D.J., Harrison, K., Seiler, K., (1994). Flurri, Electroosmotic pumping and electrophoretic separations for miniaturized chemical analysis systems, Journal of Micromechanics and Microengineering 4, pp. 257–265.
Mc Bride, S.E., Moroney, R.M., Chiang, W., (1998). Electrohydrodynamicpumps for high-density mi-crofluidic arrays, in: D.J. Harrison, A. Van Den Berg Eds. , Proceedings of Micro Total Analysis Systems Conference, Banff, Canada, Oct. 13–16, pp. 45–48.
Moroney, R.M., White, R.M., Howe, R.T., (1991). Ultrasonically induced microtransport, in: Proceed-ings of the IEEE 1991 Micro Electro Mechanical Systems Workshop MEMS ’91 , Nara, Japan, Jan. 30–Feb. 2, pp. 277–282.
Nguyen, N. T., Huang, X.Y., Toh, K.C., (2002). MEMS–micropumps: a review, ASME Trans. J. Fluids Eng. 124, pp. 384–392.
Nguyen, N.T., Wereley, S.T., (2002). Fundamentals and Applications of Microfluidics, first ed., Artech House, Boston, MA.
Olsson, A., Enoksson, P., Stemme, G., Stemme, E., (1997). Micromachined flat-walled valveless diffuser pumps. J. Microelectromech. Syst. 6 (2), pp. 161–166.
Paul, P.H., Arnold, D.W., Rakestraw, D.J., (1998). Electrokinetic generation of high pressures using po-rous microstructure, in: D.J. Harrison, A. Van Den Berg Eds. , Proceedings of Micro Total Analysis Systems Conference, Banff, Canada, Oct. 13–16, , pp. 49–52.
Rapp, R., Schomburg, W.K., Mass, D., Schulz, J., Stark, W., (1994). LIGA micro-pump for gases and liquids. Sens. Actuators A 40, pp. 57–61.
Richter, A., Plettner, A., Hofmann, K.A., Sandmaier, H., (1991). A micromachined electrohydrodynamic (EHD) pump. Sens. Actuators A 29, pp. 159–168.
Van Lintel, H.T.G., Van De Pol, F.C.M., Bouwstra, S., (1988), A piezoelectric micro pump based on micromachining of silicon, Sensors and Actuators 15, p. 153.
Zdeblick, M.J., Anderson, R., Jankowski, J., Kline-Schoder, B., Christel, L., Miles, R., Weber, W., (1994), in: Thermopneumatically Actuated Microvalves and Integrated Electro-Fluidic Cir-cuits,Solid-State Sensor and Actuator Workshop, Hilton Head, SC, June 13–16, pp. 251–255.
Zengerle, R., Kluge, S., Richter, M., Richter, A., (1995). A bidirectional silicon micropump, in: Pro-ceedings of the IEEE 1995 Micro Electro Mechanical Systems Workshop MEMS ’95 , Amsterdam, Netherlands, Jan. 29–Feb. 2, IEEE, pp. 19–24.
Zhang, W., Ahn, C.H., (1996), in: A Bi-Directional Magnetic Micropump on a Silicon Wafer,Solid-State Sensor and Actuator Workshop, Hilton Head, SC, p. 94.
