SN Applied Sciences

, 1:1556 | Cite as

Optimization and parametric study of AC electroosmotic micropumping by response surface method

  • Amin FarzanehniaEmail author
  • Amin Taheri
Research Article
Part of the following topical collections:
  1. 3. Engineering (general)


Alternative current electroosmotic micropumps (ACEO) have gained a great interest in microfluidics research due to their low operating potential. In this paper, a parametric study is performed on various geometrical parameters of an ACEO micropump design. The problem is modeled and solved numerically using COMSOL Multiphysics. This micropump relies on the grouping of a periodic electrode array in terms of the applied voltage to create asymmetries in the electrode configuration. The pumping velocity is obtained and investigated for various geometry designs. In addition, the optimal parameters are obtained by employing the response surface methodology (RSM). The RSM could predict the effects of different factors affecting the performance of the system and their interactions. The most important parameters affecting the pumping velocity are identified. It is shown that the pumping velocity is controllable. Three various micropump designs with different cost to benefit ratios are presented. These designs are different due to their fabrication cost and the pumping velocity.


AC electroosmotic flow Lab-on-chip Micropump Optimization Response surface methodology (RSM) 

List of symbols


Electrode width (μm)


Inter-electrode gap (μm)


Microchannel height (μm)


Electrical potential (V)


Area-specific impedance (Ω m2)


Capacitance per unit area (F m−2)


Electrical potential at electrode (V)


Pressure (Pa)


Velocity field (m s−1)


Diffusion constant (m2 s−1)


Design parameter



Debye length (m)


Ratio of the diffuse layer of EDL potential drop to the total double layer potential drop


Fluid viscosity (Pa s)


Permittivity (F m−1)


Angular frequency (rad s−1)


Electrolyte conductivity (S m−1)


Polynomial coefficient



Outer edge of the EDL


Double layer


Applied voltage to the electrode



1 Introduction

ACEO is an effective technique for the manipulation of fluids and particles in microscale systems [1, 2, 3, 4, 5]. ACEO is defined as the flow generation in a microchannel at the surface of electrodes where low AC voltage is applied. It has proved to be a great potential to be employed for the fabrication of micropumps [6, 7]. Such micropumps benefit from low operating potential, use of no moving parts, minimum electrolysis, and consequent elimination of bubble formation and generation of new species over the electrodes [8].

ACEO was firstly introduced by Ramos et al. [9, 10]. Taking advantage of ACEO principles, Ajdari [11] developed an array of asymmetric electrodes for liquid pumping. Various designs of ACEO micropumps have been developed mainly employing asymmetric electrodes or asymmetric applied voltage [12, 13], transverse traveling waveforms [14, 15] and AC voltage with a DC bias [16, 17]. A detailed review of different designs of micropumps and their applications was presented in Ref. [18].

Loucaides et al. [13] proposed a novel ACEO micropump that works by the grouping of electrodes. The asymmetry generated in terms of applied voltage to produce a net flow in the microchannel. The method is useful or pumping using low voltages. They showed that by switching voltages the degree of asymmetry changes; therefore, a bidirectional flow is achieved. In another study, Loucaides et al. [19] investigated a configurable ACEO-based microdevice that showed dual functions of simultaneous pumping and mixing in microcapillaries. They examined various electrode excitation modes that can be applied to a single device. They achieved significant mixing efficiency; however, there is a trade-off between mixing performance and the pumping performance. Yoshida et al. [20] proposed a new ACEO micropump using a square pole—slit electrode array. They showed that the proposed method could generate a total flow by the slip velocity mainly on the slit electrodes. The COMSOL Multiphysics software was used to simulate the micropump and obtain the optimum parameters for maximizing the effective slip velocity. They also fabricated and experimentally studied the micopump. A maximum slip velocity of 1.6 mm s−1 was observed in their study.

Commonly, the goal in a pump design is to maximize the pumping rate [21] with minimized power consumption. To this end, few investigations on the optimization of ACEO-based have been performed. Ramos et al. [22] performed a theoretical and numerical analysis on ACEO micropumps with asymmetric pairs of microelectrodes. They obtained the pumping velocity by changing a geometrical parameter while keeping other parameters constant. Olesen et al. [23] used an unconstrained optimization method to optimize the geometry of a ACEO micropump with two asymmetric electrodes. However, they did not use an optimization method to reveal the interaction of various parameters. Surveying the literature suggests that a comprehensive optimization and parametric study on the ACEO micropumps to show the interaction of different parameters in the design of the micropump is rare. This type of ACEO micropump consists of planner electrodes, which has the advantage of simple and straightforward fabrication.

The aim of this study is to investigate the effects of different geometrical parameters on an ACEO micropump and to provide statistical analysis for the performance optimization of the electrode array in an ACEO micropump. A three-electrode micropump is considered in this study because this type of micropump benefits from enabling a bidirectional flow. Additionally, the micropump has the advantage of simple fabrication, configurable geometry, and versatility in terms of applied voltage. The geometrical parameters are optimized to achieve higher pumping velocities; therefore, lower voltages could be applied, and the design would be more energy-efficient and eliminates bubble formation and faradic reactions. In addition, in this study, the different electrode configurations of micropump are investigated and the response surface methodology (RSM) is used to obtain the optimum results. Three various optimum designs are suggested based on their fabrication cost and pumping velocities. The RSM is a collection of mathematical and statistical techniques used to explore the influence and interaction of several independent factors (i.e., the design parameters) on the design responses [24]. Therefore, the identified relation (the response surface) shows the effects of the design parameters using a limited number of controlled experiments/simulations and design parameters can be optimized to produce the most desirable design response [25, 26].

2 Theory and mathematical modeling

The electrical double layer (EDL) is formed when an electrolyte is in the vicinity of a charged solid surface. The solid surface can either be metal under a potential or a charge surface due to the difference in electrochemical potentials of the two phases [27]. In ACEO, applied AC signal over the electrode pair cause attraction of counter ions to electrode surface and produce a layer of induced charges. Therefore, the counter ions move with or against the tangential component of the electric field; consequently, the fluid motion is produced due to fluid viscosity [9, 10]. The ACEO fluid flow is the result of tangential electrical field on the induced charge within the diffuse double layer. The electrical field is produced using an array of electrodes, Fig. 1a. In this study, a design of a previously published paper on a micropump with three electrodes is investigated [13]. This micropump benefits from a simple fabrication, configurable geometry, and adjustable applied voltage. Figure 1b shows the 2D geometry of the micropump. The electrodes are assumed to be long and thin. Thus, they could be characterized by their widths labeled as \(W_{1}\), \(W_{2}\), and \(W_{3}\). The micropump working principle is based on the application of the asymmetric potential. The flow is generated utilizing the asymmetry in the electric field. In this system, the asymmetry is induced by applying the same voltage to two adjacent electrodes and applying anti-phase voltage to the other electrode. It is shown that whenever there is an asymmetry in the electrodes the electric and flow fields from the smaller electrode dominate and a net flow is produced toward the electrodes [28]. In this case, the flow is generated toward the two electrodes with the same voltage. A micropump employing the configuration shown in Fig. 1b would induce flow from left to right. However, when the electrodes 1 and 2 are grouped together the flow direction would be reversed. Thus, a bidirectional flow in the micropump could be achieved. In addition, changing the electrode grouping when the electrodes have different dimensions could modulate the degree of asymmetry.
Fig. 1

a A three-dimensional schematics of a rectangular planar array of electrodes implemented in a microchannel. b A two-dimensional view of the electrodes (not to scale)

Since the channel depth is considered much larger than its height, a two-dimensional model is employed. The model is established based on the Debye–Huckel double layer theory [29]. The mathematical model presented by the references [19, 22] is used here. Experimental studies show that for RMS voltage value higher than 1.5 Faradic reactions take place [15, 28, 30]. Therefore, the effects of Faradaic currents are not taken into consideration.

2.1 Electrical field

Electrical potential in the microchannel obeys the Laplace’s equation:
$$\nabla^{2} \phi = 0$$
The equation is subjected to the following boundary condition at the electrode surfaces:
$$\sigma \varvec{n}\nabla \phi_{\text{ep}} = \frac{{\phi_{\text{ep}} - V_{\text{app}} }}{{Z_{\text{DL}} }}$$
where \(\phi_{\text{ep}}\) is the electrical potential on the outer edge of the EDL, \(V_{\text{app}}\) is the applied electrical potential to the electrodes, \(\sigma\) is the electrical conductivity of the fluid, \(Z_{\text{DL}} = \frac{1}{{i\omega C_{\text{DL}} }}\) is the impedance of the EDL, \(C_{\text{DL}} = \frac{\epsilon}{{\lambda_{\text{Debye}}}}\) is the capacitance of the double layer (given that the size of the compact layer is small), \(\omega\) is the frequency of the AC electric field, and \(\lambda_{\text{Debye}}\) represents the characteristic length of EDL.

The inlet and outlet of the domain are set to periodic boundary condition. The inter-electrode walls have Neumann boundary condition of \(\varvec{n} \cdot \nabla \cdot \phi = 0\), where \(\varvec{n}\) is the normal to plane vector.

2.2 Fluid flow

The governing equations for the fluid flow are the continuity equation and Navier–Stokes equations neglecting inertial term (Stroke flow).
$$\nabla \cdot \varvec{u} = 0$$
$$- \nabla P + \mu \nabla^{2} \varvec{u} = 0$$
where \(\mu\) is the fluid viscosity, \(\varvec{u}\) velocity field, and \(P\) is the fluid pressure. The associated boundary condition on the electrode surfaces (on the surface of EDL) is the tangential ACEO velocity derived from the Helmholtz–Smoluchowski velocity (Eq. (5)).
$$\varvec{u}_{\text{slip}} = - \frac{\epsilon}{4\mu } \varLambda \frac{\partial }{\partial x}\left| {\phi_{\text{ep}} - V_{\text{app}} } \right|^{2}$$
where the parameter \(\varLambda\) is defined as the ratio of the diffuse layer of EDL potential drop to the total double layer potential drop. The parameter \(V_{\text{app}}\) is the applied voltage to the electrode. The value of \(\varLambda\) is experimentally determined, that is considered as 0.25 for this study [13]. It should be noted that the applied voltage depends on the electrode groupings. The value of applied voltage could be either \(V_{0} \cos \left( {\omega t} \right)\) or \(- V_{0} \cos \left( {\omega t} \right)\) as shown in Fig. 1b.
The fluid is assumed Newtonian, incompressible with constant physical properties. Furthermore, the flow is limited to a steady, uniform flow in the microchannel. The properties of the electrolyte fluid used in this study are presented in Table 1.
Table 1

The properties of electrolyte material used in the numerical study [19]



Dynamic viscosity (μ)

\(1 \times 10^{ - 3} \,{\text{Pa}}\,{\text{s}}\)

Relative permittivity (ϵr)


Vacuum permittivity (ϵ0)

\(8.8542 \times 10^{ - 12} \,{\text{F}}\,{\text{m}}^{ - 1}\)

Debye length \((\lambda_{\text{Debye}} )\)

\(3 \times 10^{ - 8} \,{\text{m}}\)


\(1.23 \times 10^{ - 3} \,{\text{S}}\,{\text{m}}^{ - 1}\)

Diffusion constant (D)

\(10^{ - 10} \,{\text{m}}^{2} \,{\text{s}}^{ - 1}\)

Microchannel height

\(2 0 0\,\upmu{\text{m}}\)

The numerical model is established in COMSOL Multiphysics for simultaneous solving of the electrical field and fluid flow equations.

Non-dimensional frequency and slip velocity are employed to facilitate the explanation of the obtained results. Non-dimensional frequency is defined as \(\varOmega = {\raise0.7ex\hbox{${\omega \epsilon L}$} \!\mathord{\left/ {\vphantom {{\omega \epsilon L} {\sigma \lambda_{\text{Debye}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${\sigma \lambda_{\text{Debye}} }$}}\) where \(\omega\) is the angular frequency of AC voltage, \(\lambda_{\text{Debye}}\) is the Debye length, \(\sigma\) is the electrolyte conductivity, and \(L\) is the characteristic length of the micropump. Height of the channel is used as the characteristic length (200 μm). The non-dimensional slip velocity on the periodical length of the microelectrode array that indicates the pumping capacity of the micropump is defined as \(U_{\text{non}} = {\raise0.7ex\hbox{${4\mu Lu_{\text{slip}} }$} \!\mathord{\left/ {\vphantom {{4\mu Lu_{\text{slip}} } {\epsilon \left( {\Delta V} \right)^{2} \varLambda }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${\epsilon \left( {\Delta V} \right)^{2} \varLambda }$}}\) where \(\mu\) is the fluid viscosity, \(u_{\text{slip}}\) is the fluid velocity at the electrode surface, and \(\Delta V\) is the potential difference between the electrodes (\(\Delta V = 2V_{0}\)). This slip velocity is referred to as the pumping velocity in the present text as it is an indication of pumping capacity.

3 Results and discussion

3.1 Mesh study and validation

The computational domain is discretized using unstructured triangular elements. To ensure that the results are grid-independent, the non-dimensional pumping velocity in the outlet is numerically obtained using various element numbers. As shown in Fig. 2, the use of approximately 10,000 elements would lead to convergence and grid-independent results.
Fig. 2

Non-dimensional pumping velocity obtained using different numbers of elements to determine element number for grid-independent solution

To evaluate the accuracy of the numerical model, the obtained numerical results are compared with the results of Ramos et al. [22] and Loucaides et al. [13]. To this end, the non-dimensional pumping velocities of the micropump based on parameters values of \(W_{1} = 30, W_{2} = 60, W_{3 } = 30, G_{1} = 10, G_{2 } = 10, \,{\text{and}}\,G_{3 } = 100, H = 100\,(\upmu{\text{m}})\) are compared. As shown in Fig. 3, there is a remarkable agreement between the results. It can be seen that there is a bell-shaped relation between the frequency and the pumping velocity. Therefore, the velocity decreases at very low and high frequencies.
Fig. 3

Comparison of the accuracy of the numerical model with similar studies [13, 22]. The non-dimensional velocity with respect to non-dimensional frequency

Prior to performing the parametric study, an ACEO micropump with the geometrical parameters given in Table 2 is simulated as a baseline case [13]. The corresponding flow field and the streamlines along with the non-dimensional velocity over the electrode edges are demonstrated for the baseline case in Fig. 4a, b, respectively. The non-dimensional pumping velocity of the baseline case is obtained to be 0.019. The micropump operates by applying the same AC signal to the two adjacent electrodes while applying a signal with the same magnitude in anti-phase to the other electrode as in Fig. 1b. This creates asymmetry in terms of the applied voltage to produce a net flow rate. Another method of implementation is to apply an AC signal to one electrode group and earthing the other.
Table 2

Value of the electrodes parameters used for the baseline case [13]



Electrodes widths \(\left( {W_{1} = W_{2} = W_{3} } \right)\)

\(1 0 0\, (\upmu{\text{m)}}\)

Inter-electrodes gaps \(\left( {G_{1} = G_{2} } \right)\)

\(1 0\, (\upmu{\text{m)}}\)

Third inter-electrodes gap \(\left( {G_{3} } \right)\)

\(2 8 0\, (\upmu{\text{m)}}\)

Voltage \((V_{0} )\)

\(0. 5\, ( {\text{V)}}\)

Frequency \(\left( f \right)\)

\(2 5 0\, ( {\text{Hz)}}\)

Fig. 4

a Flow field of the baseline case once electrodes 2 and 3 are grouped together. Black lines show the streamlines and the red arrows indicate the direction of the flow field. b Flow velocity at the electrode edge for the baseline case. Solid red line shows slip velocity when electrodes 2 and 3 are grouped together. Dashed green line shows the grouped electrodes of 1 and 2

3.2 Parametric study

To investigate the pumping effect with respect to the geometry of electrodes, the electrodes widths, and inter-electrodes gaps \(W_{1}\), \(W_{2}\), \(W_{3}\), \(G_{1}\), \(G_{2}\), and \(G_{3}\) are examined in cogent ranges. The parameters are non-dimensionalized with respect to the height of microchannel. Figures 5, 6, 7, 8, 9 and 10 show the non-dimensional pumping velocity against the non-dimensional frequency for different non-dimensional electrodes widths and inter-electrode gaps.
Fig. 5

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(G_{1}\); b maximum non-dimensional velocity versus \(G_{1}\)

Fig. 6

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(G_{2}\); b maximum non-dimensional velocity versus \(G_{2}\)

Fig. 7

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(G_{3}\); b maximum non-dimensional velocity versus \(G_{3}\)

Fig. 8

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(W_{1}\); b maximum non-dimensional velocity versus \(W_{1}\)

Fig. 9

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(W_{2}\); b maximum non-dimensional velocity versus \(W_{2}\)

Fig. 10

a The non-dimensional pumping velocity versus the non-dimensional frequency for different values of \(W_{3}\); b maximum non-dimensional velocity versus \(W_{3}\)

Maximum pumping velocity increases once \(W_{2}\) and \(W_{3}\) are increased; however, by increasing the \(G_{1}\), \(G_{2}\), and \(G_{3}\) the maximum velocity decreases. Moreover, the increase of \(W_{1}\) value to 0.2 results in the initial increase of maximum velocity followed by its decline. For most cases, the maximum pumping velocity obtains at the non-dimensional frequency of 0.5. Figure 5a reveals that the maximum pumping velocity occurs at the non-dimensional frequency of 0.55. Furthermore, Figs. 6 and 7 indicate that the second and third gap (\(G_{2}\) and \(G_{3}\)) have a small effect on the pumping velocity.

Figure 8 shows that using \(W_{1} = 0.2\) results in the highest pumping velocity. It can be seen from Fig. 8 that there is a highly nonlinear relation between the parameter \(W_{1}\) and pumping velocity, making \(W_{1}\) an important parameter in the design of an ACEO micropump. Furthermore, Fig. 9 (a) shows that using \(W_{2} = 0.1\) would cause flow reversal at higher AC frequencies. Figure 10 also shows that \(W_{3}\) affects the pumping velocity significantly.

3.3 Statistical analysis

3.3.1 Overview of the response surface model

The response surface method is a polynomial regression model that expresses the relation between the obtained response and the independent variables [25]. The relation of the design variables with the response value is derived by fitting the numerical analysis results to a second-order polynomial (Eq. 6) [31].
$$y = \beta_{0} + \mathop \sum \limits_{p = 1}^{k} \beta_{p} x_{p} + \mathop \sum \limits_{p = 1}^{k} \beta_{pp} x_{p}^{2} + \mathop \sum \limits_{p = 1}^{k} \mathop \sum \limits_{q > i}^{k} \beta_{pq} x_{p} x_{q}$$
where y is the response value, \(\beta_{0} , \beta_{p} , \beta_{pp} ,\) and \(\beta_{pq}\) are the coefficients, \(x_{p}\) and \(x_{q}\) are the pth and qth design parameter, \(k\) is total number of parameters, respectively.
The first step of BBD is to code all the design parameters into − 1, 0, and 1 using Eq. 7. Therefore, in this design, each variable (factor) is set to three levels with the coded values of − 1 for low level, 0 for medium level and 1 for high level.
$$\bar{x} = \frac{{2\left( {x - x_{\hbox{min} } } \right)}}{{x_{\hbox{max} } - x_{\hbox{min} } }} - 1$$
where \(\bar{x}\) is the coded value of the parameter, \(x\) is the value of the original parameter, \(x_{\hbox{min} }\) and \(x_{\hbox{max} }\) are the minimum and maximum of the original value of the parameter, respectively.
Figure 11 shows the geometry of BBD with three design parameters of X1, X2, and X3 as an example. The design points are chosen at the edges and the central point, whereas the corners points are avoided to avoid extreme conditions [25]. The central point is repeated in order to comprise the errors produced by simulations.
Fig. 11

Geometry schematics of design points of three parameters using Box-Behnken Design

3.3.2 The response surface method implementation

The Box–Behnken Design (BBD) is used in this study to optimize the geometry and achieve maximum pumping velocity. Compared to other designs, the BBD requires fewer runs for determination of the response pattern. Six variables of \(W_{1}\) (A), \(W_{2}\) (B), \(W_{3}\) (C), \(G_{1}\) (D), \(G_{2}\) (E), and \(G_{3}\) (F) are the design factors, Table 3. The response value is the peak of pumping velocity. The number of runs for RSM in this study is 54.
Table 3

Variables and their corresponding levels used for BBD


Factor symbol


− 1

































All dimensions are in (\(\upmu{\text{m}}\))

The statistical analysis is performed in terms of analysis of variance (ANOVA) in order to find significant model terms. The values of the partial probability (p value) less than 0.05 and high F value imply that the model factors are significant, whereas the p values higher than 0.1 indicate the statistically insignificant terms [31]. From Online Resource Table S1, it can be seen that the quadratic model used here has the p value of < 0.0001 indicating that the model response model is significant and there is only < 0.01% chance that the F value of the model is because of the random error. The model terms A, B, D, AD, A2, C2 (\(W_{1}\), \(W_{2}\), \(G_{1}\), \(W_{1} G_{1}\), \(W_{1}^{2}\), and \(W_{3}^{2}\)) are significant model terms. Therefore, the terms with the p value higher than 0.05 are removed from the model. The term A (\(W_{1}\)) with the smallest p value of less than 0.0001 and the largest F value of 54.48 verifies that the \(W_{1}\) has the most effect on the pumping capacity. The effects of six geometrical parameters and their interactions on the maximum of non-dimensional pumping velocity (Vpumping) are shown in Online Resource Figure S1. The 3D response surfaces are obtained when two variables are changed in the range of experiments and other variables are kept at their central level. It can be seen in Online Resource Figure S1 that the model variables are interdependent. Moreover, the results show that the maximum of non-dimensional pumping velocity is varied from 0 to 0.095 using different design parameters. The interactions between \(W_{1}\) and \(G_{1}\) affected the pumping velocity significantly (p < 0.05) (Online Resource Figure S1(b)).

The optimum design is achieved by RSM to produce the highest pumping velocity. Initially, the optimization objective is defined as maximizing pumping velocity and keeping all other geometrical variables in the range of experiments, Table 3. However, the achieved design is not preferable from the fabrication point of view, design 1 in Table 4. As the electrode and inter-electrode sizes reduce, the process of electrodes fabrication becomes very cost-intensive and complicated. Therefore, three different designs are suggested in Table 4, each with different pumping velocity to cost ratios. The design 1 in Table 4 could provide a maximum non-dimensional pumping velocity of approximately 0.1. The design 1 is the most cost-intensive design, whereas the design 3 is a cheap design where \(( {\text{G}}_{ 1} )\) is larger than 20 μm (similar to the gaps in micropump presented in Refs. [22, 32]). However, when using design 3 instead of 1 the pumping velocity reduces by about 9 times. Therefore, the design 2 where the \(G_{1}\) is relatively larger than the first design is presented where the pumping velocity is in between the two cases. The streamlines of this micropump design with the maximum pumping rate are shown in Fig. 12.
Table 4

Various optimum designs to maximize the pumping velocity in the micropump [geometrical parameters are non-dimensionalized with respect to the height of micropump in this case \(H = 200\, (\upmu{\text{m)}}\)]








Frequency \(\left( f \right)\)

Maximum pumping velocity




























Fig. 12

Flow field of the optimum case of the design 1. Black lines show the streamlines and red arrows indicate the direction of the flow field

3.4 Applications of the studied micropump

In this section, the applications and the fluid types and the range of pumping capacity of the presented micropump are discussed. In general, the ACEO micropumps work with fluids with a conductivity below 85 ms m−1 [7]. The velocity as shown in various cases in this study has a bell-shaped relation with frequency and there is an optimal frequency. The velocity also depends on the amplitude of the AC electric field and the electrolyte concentration [33]. The ACEO occurs in frequencies lower than 100 kHz [7, 34]. The concentration of the biological electrolyte is usually diluted to 10–100 mM buffers for ACEO micropumping [33]. Using the proposed micropump, it is possible to achieve the velocities up to 1 mm s−1. The studied ACEO micropump is used in low-voltage pumping in microfluidics. Furthermore, it can be used in flow control in microfluidic networks. Therefore, the direction of flow is controlled locally. The secondary flow generated is also used as a mixing enhancer.

4 Conclusion

This paper presents the parametric study of an ACEO micropump using finite element method. The proposed micropump works with the asymmetries in electrodes dimensions and their applied voltage. The electrode widths and inter-electrode gaps are varied in reasonable ranges, and their effects on the pumping velocity are studied. In addition, the response surface methodology (RSM) is employed to show the effects of the independent variables and their interactions on the output variable (non-dimensional pumping velocity). Moreover, the most important parameters affecting the pumping velocity are found. The micropump is optimized using RSM to enhance the pumping velocity, and the optimized design parameters have been identified. Three optimized designs are suggested based on the fabrication cost and pumping output. The design with the highest output and highest fabrication cost has a maximal non-dimensional pumping velocity of 0.1. Based on the finding of this study, it can be concluded that an optimized geometry of micropump is required to achieve fast micro pumping.



We kindly acknowledge that some parts of the numerical study were performed on the High Performance Computing Center of Ferdowsi University of Mashhad.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

42452_2019_1605_MOESM1_ESM.pdf (1.2 mb)
Supplementary material 1 (PDF 1212 kb)


  1. 1.
    Jie W, Islam N, Meng L (2006) High sensitivity particle detection by biased AC electroosmotic trapping on cantilever. In: 19th IEEE international conference on micro electro mechanical systemsGoogle Scholar
  2. 2.
    Liu Z et al (2015) Particle focusing by AC electroosmosis with additional axial flow. Microfluid Nanofluid 18(5):1115–1129CrossRefGoogle Scholar
  3. 3.
    Loucaides NG, Ramos A, Georghiou GE (2011) Dielectrophoretic and AC electroosmotic trapping of DNA: numerical simulation incorporating fluid dynamics and steric particle effects. J Electrostat 69(2):111–118CrossRefGoogle Scholar
  4. 4.
    Mirzajani H et al (2016) Design and characterization of a passive, disposable wireless AC-electroosmotic lab-on-a-film for particle and fluid manipulation. Sens Actuators B Chem 235:330–342CrossRefGoogle Scholar
  5. 5.
    Hilber W et al (2009) Density-dependent particle separation in microchannels using 3D AC-driven electro-osmotic pumps. Sens Actuators A Phys 156(1):115–120CrossRefGoogle Scholar
  6. 6.
    Yang H et al (2009) AC electrokinetic pumping on symmetric electrode arrays. Microfluid Nanofluid 7(6):767CrossRefGoogle Scholar
  7. 7.
    Wu J (2010) In situ nanoparticle focusing within microfluidics. In: Kumar CS (ed) Microfluidic devices in nanotechnology: fundamental concepts. Wiley, Hoboken, p 295CrossRefGoogle Scholar
  8. 8.
    Chen J-L, Shih W-H, Hsieh W-H (2010) Three-dimensional non-linear AC electro-osmotic flow induced by a face-to-face, asymmetric pair of planar electrodes. Microfluid Nanofluid 9(2–3):579–584CrossRefGoogle Scholar
  9. 9.
    Ramos A et al (1998) Ac electrokinetics: a review of forces in microelectrode structures. J Phys D Appl Phys 31(18):2338CrossRefGoogle Scholar
  10. 10.
    Ramos A et al (1999) AC electric-field-induced fluid flow in microelectrodes. J Colloid Interface Sci 217(2):420–422CrossRefGoogle Scholar
  11. 11.
    Ajdari A (2000) Pumping liquids using asymmetric electrode arrays. Phys Rev E 61(1):R45–R48CrossRefGoogle Scholar
  12. 12.
    Studer V et al (2004) An integrated AC electrokinetic pump in a microfluidic loop for fast and tunable flow control. Analyst 129(10):944–949CrossRefGoogle Scholar
  13. 13.
    Loucaides N, Ramos A, Georghiou GE (2007) Novel systems for configurable AC electroosmotic pumping. Microfluid Nanofluid 3(6):709–714CrossRefGoogle Scholar
  14. 14.
    Ramos A et al (2005) Pumping of liquids with traveling-wave electroosmosis. J Appl Phys 97(8):084906CrossRefGoogle Scholar
  15. 15.
    García-Sánchez P, Loucaides NG, Ramos A (2017) Pumping of electrolytes by electrical forces induced on the diffusion layer: a weakly nonlinear analysis. Phys Rev E 95(2):022802CrossRefGoogle Scholar
  16. 16.
    Islam N, Reyna J (2012) Bi-directional flow induced by an AC electroosmotic micropump with DC voltage bias. Electrophoresis 33(7):1191–1197CrossRefGoogle Scholar
  17. 17.
    Islam N, Askari D (2013) Performance improvement of an AC electroosmotic micropump by hydrophobic surface modification. Microfluid Nanofluid 14(3–4):627–635CrossRefGoogle Scholar
  18. 18.
    Wang Y-N, Fu L-M (2018) Micropumps and biomedical applications—a review. Microelectron Eng 195:121–138CrossRefGoogle Scholar
  19. 19.
    Loucaides N, Ramos A, Georghiou GE (2012) Configurable AC electroosmotic pumping and mixing. Microelectron Eng 90:47–50CrossRefGoogle Scholar
  20. 20.
    Yoshida K et al (2017) A study on an AC electroosmotic micropump using a square pole—slit electrode array. Sens Actuators, A 265:152–160CrossRefGoogle Scholar
  21. 21.
    Yuan Q, Yang K, Wu J (2014) Optimization of planar interdigitated microelectrode array for biofluid transport by AC electrothermal effect. Microfluid Nanofluid 16(1):167–178CrossRefGoogle Scholar
  22. 22.
    Ramos A et al (2003) Pumping of liquids with ac voltages applied to asymmetric pairs of microelectrodes. Phys Rev E 67(5):056302CrossRefGoogle Scholar
  23. 23.
    Olesen LH, Bruus H, Ajdari A (2006) AC electrokinetic micropumps: the effect of geometrical confinement, Faradaic current injection, and nonlinear surface capacitance. Phys Rev E 73(5):056313CrossRefGoogle Scholar
  24. 24.
    Liu X et al (2011) Optimization of slurry components for a copper chemical mechanical polishing at low down pressure using response surface methodology. Microelectron Eng 88(1):99–104CrossRefGoogle Scholar
  25. 25.
    Zhang S et al (2018) Response-surface-model-based system sizing for Nearly/Net zero energy buildings under uncertainty. Appl Energy 228:1020–1031CrossRefGoogle Scholar
  26. 26.
    Ma S et al (2018) Ultrasound-induced changes in structural and physicochemical properties of β-lactoglobulin. Food Sci Nutr 6(4):1053–1064CrossRefGoogle Scholar
  27. 27.
    Salari A, Thompson M (2018) Recent advances in AC electrokinetic sample enrichment techniques for biosensor development. Sens Actuators B Chem 255:3601–3615CrossRefGoogle Scholar
  28. 28.
    Chen J-L, Shih W-H, Hsieh W-H (2013) AC electro-osmotic micromixer using a face-to-face, asymmetric pair of planar electrodes. Sens Actuators B Chem 188:11–21CrossRefGoogle Scholar
  29. 29.
    Green NG et al (2002) Fluid flow induced by nonuniform ac electric fields in electrolytes on microelectrodes. III. Observation of streamlines and numerical simulation. Phys Rev E 66(2):026305MathSciNetCrossRefGoogle Scholar
  30. 30.
    Červenka P, Přibyl M, Šnita D (2009) Numerical study on AC electroosmosis in microfluidic channels. Microelectron Eng 86(4–6):1333–1336CrossRefGoogle Scholar
  31. 31.
    Perez JVD et al (2017) Response surface methodology as a powerful tool to optimize the synthesis of polymer-based graphene oxide nanocomposites for simultaneous removal of cationic and anionic heavy metal contaminants. RSC Adv 7(30):18480–18490CrossRefGoogle Scholar
  32. 32.
    Green NG, Ramos A, Morgan H (2000) AC electrokinetics: a survey of sub-micrometre particle dynamics. J Phys D Appl Phys 33(6):632CrossRefGoogle Scholar
  33. 33.
    Pribyl M, Snita D, Marek M (2008) Multiphysical modeling of DC and AC electroosmosis in micro-and nanosystems. In: Petrone G, Cammarata G (eds) Modelling and simulation. IntechOpen, RijekaGoogle Scholar
  34. 34.
    Morgan H, Green NG (2003) AC electrokinetics: colloids and nanoparticles. Research Studies Press, PhiladelphiaGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Department of EnergyPolitecnico di MilanoMilanItaly
  2. 2.Department of Mechanical EngineeringFerdowsi University of MashhadMashhadIran

Personalised recommendations