Microfluidics and Nanofluidics

, Volume 6, Issue 6, pp 823–833

A pneumatic micropump incorporated with a normally closed valve capable of generating a high pumping rate and a high back pressure

Authors

  • Yi-Ning Yang
    • Department of Engineering ScienceNational Cheng Kung University
  • Suz-Kai Hsiung
    • Department of Engineering ScienceNational Cheng Kung University
    • Department of Engineering ScienceNational Cheng Kung University
    • Medical Electronics and Device Technology CenterIndustrial Technology Research Institute
Research Paper

DOI: 10.1007/s10404-008-0356-7

Cite this article as:
Yang, Y., Hsiung, S. & Lee, G. Microfluid Nanofluid (2009) 6: 823. doi:10.1007/s10404-008-0356-7

Abstract

This study reports on a new pneumatic micropump integrated with a normally closed valve that is capable of generating a high pumping rate and a high back pressure. The micropump consists of a sample flow microchannel, three underlying pneumatic air chambers, resilient polydimethylsiloxane (PDMS) membrane structures and a normally closed valve. The normally closed valve of the micropump is a PDMS-based floating block structure located inside the sample flow microchannel, which is activated by hydraulic pressure created by the peristaltic motion of the PDMS membranes. The valve is used to effectively increase pumping rates and back pressures since it is utilized to prevent backflow. Experimental results indicate that a pumping rate as high as 900 μL/min at a driving frequency of 90 Hz and at an applied pressure of 20 psi (1.378 × 10Nt/m2) can be obtained. The back pressure on the micropump can be as high as 85 cm-H2O (8,610.5 Nt/m2) at the same operation conditions. The micropump is fabricated by soft lithography processes and can be easily integrated with other microfluidic devices. To demonstrate its capability to prevent cross contamination during chemical analysis applications, two micropumps and a V-shape channel are integrated to perform a titration of two chemical solutions, specifically sodium hydroxide (NaOH) and benzoic acid (C6H5COOH). Experimental data show that mixing with a pH value ranging from 2.8 to 12.3 can be successfully titrated. The development of this micropump can be a promising approach for further biomedical and chemical analysis applications.

Keywords

MicrofluidicsMEMSMicropumpNormally closed microvalve

List of symbols

CCD

charge-coupled device

C6H5COOH

benzoic acid

EMV

electromagnetic valve

MEMS

micro-electro-mechanical-system

NaOH

sodium hydroxide

PDMS

polydimethylsiloxane

pH

power of hydrogen ions

SEM

scanning electron microscope

UV

ultraviolet

μ-TAS

micro-total-analysis-system

1 Introduction

Micro-electro-mechanical-systems (MEMS) technology has been a promising approach in a variety of fields, including biotechnology, communications, optics and many others by integrating miniaturized mechanical elements, sensors, actuators and electronics on a single chip through micromachining techniques (Angell et al. 1983; Reyes et al. 2002). Among them, the applications of MEMS technology for genetic analysis, molecular biology, and analytical chemistry have attracted considerable interest recently. It makes possible the realization of a so-called micro-total-analysis-system (μ-TAS) (Auroux et al. 2002). A μ-TAS is usually capable of performing many critical sample pre-treatment and bio-process functions, including mixing, transportation, reaction, collection, separation and detection, on a single microfluidic device and in an entirely automated system (Sato et al. 2003; Raiteri et al. 2002). The major advantages of these microfluidic devices and systems are a significant reduction of the chemical reaction time and a lower consumption of expensive chemical reagents and samples, as well as the capability of automating the entire process. More importantly, the functionality and reliability of the microfluidic systems can be enhanced if additional microfluidic functions can be integrated on a single chip.

One of the most critical issues for a microfluidic system is how to transport and precisely control a small amount of fluid. Micropumps capable of providing an appropriate pumping rate and a reasonable back pressure are usually inevitable requirements for a self-contained microfluidic system. With the advent of microfluidic technologies, miniaturized pumping devices utilizing various actuation mechanisms have being extensively investigated since 1980s. Briefly, these mechanical or membrane-based micropumps generally contain pressurized chambers and are bonded with flexible membranes. Deflections of the membranes compress the channel below and thus drive the flow. Various actuation methods for these membrane-based micropumps have been demonstrated in the literature, such as piezoelectric (Nguyen et al. 2004; Koch et al. 1998), thermopneumatic (Van de Pol et al. 1990), electrostatic (Ng et al. 2004; Bourouina et al. 1997), pneumatic (Jeong, and Konishi 2007) and electromagnetic (Zhang and Ahn 1996; Santra et al. 2002). Among them, pneumatic micropumps driven by injecting compressed air have attracted considerable interest. Using a soft lithography process involving thick-film micromachining and replica molding of elastomeric materials, pneumatic micropumps can be fabricated. For instance, Unger et al. first used an elastic material (PDMS) and standard soft-lithography processes to demonstrate a pneumatic micropump with a three-membrane layout controlled by three electromagnetic valves (EMV) (Unger et al. 2000). Furthermore, a peristaltic pneumatic micropump with an s-shape layout, which only requires one EMV, was reported by this research group (Wang and Lee 2006). The time-phased deflection of PDMS membranes along the microchannel length can generate a peristaltic effect which drives the fluid along the microfluidic channel such that it requires a simpler control circuit and less peripheral equipment.

To better control the fluid motion inside the microchannel, it is crucial to control the transport of different samples at different times without contamination. This typically requires a microvalve. The development of microvalves has progressed rapidly in recent years. They generally fall into two major categories, namely active (Neagu et al. 1997; Baldi et al. 2003) and passive valves (Yang et al. 1996; Paul and Terhaar 2000) depending on whether an external power supply is required. Microvalves could also be classified as two types, normally open or normally closed, as determined by their initial operating states. For example, one of the most common active valves in microfluidic systems is a pneumatic valve. When injecting compressed air into the chamber, the resulting deflection of the membranes can work as a valve to stop the flow passing through the microchannel (Wang and Lee 2005). Usually, another EMV is required to control the motion of the pneumatic microvalve. Alternatively, many passive microvalves have been reported in the literature including flap valves (Feng and Kim 2004), hydrophobic valves (Ahn et al. 2004), and ball valves (Yamahata et al. 2005). Among them, a floating block structure is commonly used as a normally closed valve by using selective bonding of membranes (Duffy et al. 1999; Hua et al. 2006). For example, Baek et al. used a controllable central block in a flow channel as a normally closed valve. It could be opened and closed easily by external air pressure (Baek et al. 2005). Similarly, Hosokawa and Maeda used three membranes selectively bonded with block structures to regulate three flow channels (Hosokawa and Maeda 2000). In general, the normally closed valve is especially useful for microfluidic systems since it can prevent cross-contamination of fluids, promote the performance of micropump and also consumes less power. However, passive valve still needs an individual controller to control it open.

In this study, a normally closed microvalve is integrated into a pneumatic micropump. The microvalve is opened depending on the driving of micropump. It only requires one EMV for flow control and transport. The resulting control system is then simplified and fewer EMVs are required. The micropump can avoid cross-contamination of fluids while it is not activated. Besides, any improvement in the pumping rate can be observed since the normally closed microvalve can efficiently stop the back flow. Moreover, the normally closed microvalve can also efficiently increase the back pressure, which is an important parameter for micropumps.

2 Materials and methods

2.1 Design

In this study, a normally closed microvalve is integrated into a pneumatic micropump, as schematically shown in Fig. 1. This prototype device consists of a sample flow microchannel, three underlying pneumatic air chambers, resilient PDMS membrane structures and a PDMS-based floating block structure as a normally closed valve. The PDMS-based membrane structure with a thickness of 150 μm can be deflected upwards sequentially to generate peristaltic actuation along a specific direction when compressed air is supplied. The sample flow is then transported forward inside the microchannel accordingly. In addition, to prevent backflow, a normally closed valve is integrated onto the micropump device. In this study, two different types of micropump designs, as shown in Fig. 1a, b, are investigated. The normally closed valve is placed at two different locations, one located between the last two air chambers (Type I) and another located at the terminus of the micropump structure (Type II). The cross-sectional views of the membrane activation sequence for these two types of micropumps are schematically shown in Fig. 1. The working principle of the micropumps is based on the deflections of PDMS membranes actuated by three air chambers underneath the sample flow microchannel to generate a peristaltic-like effect for driving the fluid forwards. The displacement of the membranes plays an important role in determining the pumping rate, which can be controlled by the applied air pressure and the driving frequency of EMV. The dimensions of each air chamber are 2,400 μm × 1,500 μm in length and width, respectively. As shown in Fig. 1a, b, the time-phased deflection of neighboring membranes induce a peristaltic effect which drives the fluid along the sample flow microchannel sequentially. Note that the three air chambers are connected by air channels. When the second PDMS membrane was activated to squeeze the liquid channel, a hydrodynamic pressure was built up to open the valve such that the fluid flow can be driven through the floating block (valve) as shown in Fig. 1a-3. As shown in Fig. 1b-4, when the third PDMS membrane was activated, the liquid was pushed to the left-hand side since the valve was still closed while the second PDMS was still deflected. With this approach, the flow direction of the sample fluid can be controlled by utilizing one EMV device without a complex control circuit.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig1_HTML.gif
Fig. 1

Schematic illustration of two different designs of the micropump and the corresponding sequence of membrane motions: a type-I micropump with the normally closed valve located between the last two air chambers, and b type-II micropump with the normally closed valve located at the terminus of the micropump structure

For pneumatic micropumps, the backflow of the sample fluid can seriously affect the pumping rate and the back pressure (Wang and Lee 2006). The design of the microvalve is useful since it can prevent cross-contamination of fluids and also increase the pumping rate. However, an additional control system is still needed to control the motion of the passive microvalve. In this study, a normally closed microvalve is integrated into the micropump to perform the functions as described above. The working principle of the normally closed valve is schematically illustrated in Fig. 2. It is composed of three major parts including a floating block structure inside the sample flow microchannel, thin membrane structures and an air chamber placed underneath the microchannel. It allows the generation of a membrane deflection such that the sample flow can pass through the valve device, as shown in Fig. 2a, b. The developed micropump can prevent the back flow during the fluid flow to increase the pumping rate and the back pressure. During the operating process of the pneumatic micropump, compressed air is first used to deflect the PDMS membranes such that a certain volume of fluid in the sample fluid channel can be pushed forward. When the air chambers are depressurized, the thin membrane structures recover to the original position. The sample fluid volume flows into and refills the space located above the membrane structure inside the microchannel. Therefore, a backflow is generated and can seriously influence the performance of the micropump. To overcome the backflow generation and to improve the performance of the micropump, a floating block structure is designed inside the sample flow microchannel. Note that the floating block structure is not bonded with the thin membrane structure and can be utilized to prevent the backflow generation such that the fluid can flow in only one direction (Fig. 2a). The floating block is 250 μm in length, 100 μm in depth and the width of liquid channel is 500 μm. The air chamber below the floating block is 500 μm × 1,500 μm in length and width. As shown in Fig. 2b, once the hydraulic pressure generated by the activation of the micropump is high enough, the non-bonded membrane can be deflected to allow the liquid to flow through the sample flow microchannel.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig2_HTML.gif
Fig. 2

Schematic illustration of the working principle of the normally closed microvalve. a A cross-sectional view of the normally closed valve showing that a sample fluid can be separated by the floating block structure. b The sample fluid can flow over the floating block when the micropump is activated

2.2 Fabrication

A simplified fabrication process for the micropump based on SU-8 lithography and a PDMS replication process are shown in Fig. 3 (Wang and Lee 2005). Initially, SU-8 negative thick photoresist (MicroChem, NANOTM SU-8, USA) was spun onto a silicon substrate at a spin-coating speed of 1,200 rpm for 1 min to achieve a thickness of 100 μm. Then a two-step soft baking process at 65°C for 10 min and then at 95°C for 30 min was performed. An exposure process using a total ultraviolet (UV) exposure dose of 600 mJ/cm2 was performed, followed by a two-step, post-exposure baking process initially at 65°C for 5 min and then at 95°C for 10 min. The baked wafer was subsequently developed using a SU-8 developer (MicroChem., USA) for 5 min. The developed wafer was then rinsed with isopropyl alcohol, followed by a nitrogen blow-dry.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig3_HTML.gif
Fig. 3

a Simplified fabrication process of the micropump based on SU-8 lithography and PDMS replication. The SEM images of the b SU-8 template and c PDMS inverse structure after replication

A PDMS (Sylgard® 184, Dow Corning, USA) replication process was employed to fabricate the double-layer PDMS-based micropump by replicating the inverse images on the SU-8 templates (Wang and Lee 2005). Silicone elastomer and elastomer curing agent (Sylgard 184A and Sylgard 184B, Sil-More Industrial Ltd, USA) were first mixed in a specific ratio (10:1) and then poured onto the SU-8 microstructure mold. A vacuum pump was used to remove the bubbles formed during the mixing process to prevent the formation of air pockets in the microchannels. A PDMS layer was spin-coated on the SU-8 template and then cured at 100°C for 1 h. The PDMS inverse structure was then mechanically peeled off the template. Two layers of PDMS can be formed by using the similar method. Detailed information for fabricating pneumatic micropumps can be found in our previous work (Wang and Lee 2005). The inlet and outlet reservoirs were mechanically drilled afterwards. To bond all structures together, all PDMS layers and a glass plate were treated by oxygen plasma. Notably, a shelter was utilized to cover the blocks during the oxygen plasma process such that they were not bonded with any PDMS layer. With this approach, selective binding of the PDMS layers can be achieved and a floating PDMS block can be formed as a normally closed valve. Scanning electron microscope (SEM) images of the fabricated SU-8 template and PDMS replication for the PDMS-based floating block structure are shown in Fig. 3b, c, respectively.

2.3 Experimental setup

Figure 4a schematically illustrates the experimental setup for pumping rate testing. First, a commercial pipette (Nichipet EX, Nichiryo Inc., Japan) was utilized to provide a sample fluid with a volume of 1 mL to the sample inlet reservoir. The micropump was then activated to transport the sample fluid from the inlet reservoir to the outlet. The total time for pumping the 1 mL sample was measured to calculate the pumping rate.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig4_HTML.gif
Fig. 4

a Schematic illustration of the experimental setup for measuring pumping rate. b A custom-made controller to operate the micropump. c Schematic illustration of the back pressure measurement

To operate the micropumps, a custom-made controller (as shown in Fig. 4b) composed of an air compressor (MDR2-1A/11, Jun-Air Inc., Japan), EMVs (S070M-5BG-32, SMC Inc., Taiwan) and a programmable control system (ATMEGA8535, ATMEL Corp., USA) were used. A pressure gauge was also used in this custom-made controller for monitoring the supplied pressure to the actuators. The air tubing was directly connected to the pneumatic microchannel to activate the micropump. The reagents were stored in reservoirs prior to testing (see Fig. 10a). In addition, the EMVs were individually controlled by using the programmable control system for the filling and releasing processes of the compressed air. The injected air pressure and driving frequency of the EMV can be regulated by using the controller. The dimensions of the hand-held controller were measured to be 20 cm × 12 cm × 8 cm, respectively.

The pumping performance of the pneumatic micropump is mainly dependent on the movements of the pulsing membrane. Hence, a high-speed, charge-coupled device (CCD, MC1311, Mikrotron, Germany) and a microscope (TE300, Nikon, USA) were utilized to observe the motion of the pneumatic micropump at different pulsation frequencies. The frame rate of the image acquisition was set at 500 frames/s to capture the movement of the membrane pulsations. To measure the back pressure provided by the micropump, an experimental setup, as shown in Fig. 4c, was adopted (Sim et al. 2003). The height of the liquid at the outlet reservoir can be increased as the liquid was pumped continuously from the inlet reservoir. Notably, as the pressure difference (back pressure) between the inlet and outlet reservoir increased, the pumping rate decreased accordingly. By changing the frequency and air pressure using the control system, the pressure difference between the inlet and outlet tubing can be obtained such that the back pressure can be measured. A pH meter (Mettler Toledo, InLab423, Switzerland) was used to detect the pH value of the mixed sample for chemical titration applications.

3 Results and discussion

3.1 Pumping rate

Many operational parameters may influence the pumping rate of the micropump, including the driving frequency of the EMV, the applied pneumatic pressure, the number of membranes along the liquid channel, and the geometry of the membranes (Wang and Lee 2006). In this study, a micropump with three membranes with geometries of 2,400 μm × 1,500 μm of each was investigated. As mentioned previously, two types of micropumps with microvalves located at different locations were explored. For comparison, a micropump without a normally closed valve was also tested. Figure 5 shows the pumping rate of the micropump operating at an applied pneumatic pressure of 10 psi (a), and 20 psi (b). As expected, the pumping rate increases with increasing pneumatic pressure. It can be clearly seen that the pumping rate can be increased significantly when the normally closed microvalve is utilized. Besides, the pumping rate of the Type-II micropump is higher than that of the Type-I micropump. For micropumps with normally closed microvalves, the pumping rate also increases while increasing the driving frequency under a constant pneumatic pressure until it reaches a certain frequency. A decreasing pumping rate after this critical frequency is observed due to the fact that the maximum pumping rate at a constant applied pressure is limited by the release time of the compressed air (Huang et al. 2006). The air cannot be completely released from the air chamber and the pumping rate drops accordingly. The maximum pumping rate of the micropump without a normally closed valve is 385 μL/min, at a driving frequency of 76 Hz and a pneumatic pressure of 20 psi (1.378 × 10Nt/m2). The maximum pumping rate is 200 μL/min at a driving frequency of 62 Hz when the pneumatic pressure is decreased to 10 psi (6.890 × 10Nt/m2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig5_HTML.gif
Fig. 5

The relationship between pumping rate and the driving frequency of the EMV for different micropump designs at different applied air pressures: a 10 psi and b 20 psi

Interestingly, for the case of the micropump with a normally closed valve, the pumping rate keeps increasing with the EMV driving frequency (up to 90 Hz, limited by the experimental setup). As shown in Fig. 5a, at a driving frequency of 90 Hz and a pneumatic pressure of 10 psi, the pumping rate of the Type-I micropump with a normally closed valve, is 490 μL/min, which is about 2 times higher than the one provided by the micropumps without normally closed valves. For Type-II micropumps, it can be as high as 720 μL/min at the same operation conditions. If the applied air pressure is increased to 20 psi, the pumping rates for Type-I and Type-II micropumps are 565 and 900 μL/min at a driving frequency of 90 Hz. The experimental results show that the proposed micropump with a normally closed valve can significantly increase the pumping rate. It can be clearly seen that the pumping rate of the Type-I and Type-II micropump devices increased about twofolds and threefolds higher, respectively, than the one without the integrated valve.

Figure 6 shows a schematic illustration of the membrane motion sequence for Type-I and Type-II micropumps during the filling and releasing processes with compressed air. A cavity volume (V) can be generated by a single membrane during the filling or releasing process. When the single membrane is activated, the fluid can flow in two different directions with the same 1/2 cavity volume (V) due to the constant flow resistance of the microfluidic channel. It can be clearly seen that the total volume being pushed forward by the filling process for these two types of micropumps are both 2-cavity volumes (2 V, Fig. 6a-1 to a-4, 6b-1 to b-4). When the compress air is released, the relaxing membranes can also draw the fluid back and generate a backflow. With the normally closed microvalve as the check valve, the back flow can be stopped to drive the fluid in one direction with the full cavity volume. As shown in Fig. 6b-7, the floating block prevents the backflow from the membrane relaxation process, thus significantly improving the pumping rate. The total volume from the release process for Type-I and Type-II micropumps are one and three cavity volumes, respectively. Therefore, the entire volume of fluid pumped by the Type-I and Type-II micropumps are 3 and 5 cavity volumes, respectively. Experimental data shown in Fig. 5 also reveals that the pumping rate of the Type-II micropump is about 1.6 times higher than the one generated by the Type-I micropump. However, the effect of the normally closed microvalve may be affected by the different pressures on the both sides of the normally closed microvalve. Therefore, the normally closed microvalve may fail when the pressure of the back flow is greater than the 85 cm-H2O (8,611 Nt/m2) backpressure with 900 μL/min pumping rate.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig6_HTML.gif
Fig. 6

Schematic illustration of the membrane motion sequence for the two different designs during the compressed air filling and releasing processes: a Type-I and b Type II

Ideally, the fluid delivered by micropumps with/without a microvalve is 5-cavity volumes (Wang and Lee 2006). However, the back flow will significantly reduce the pumping rate for the practical application, especially in the rage of high driving frequencies. With the integration of the normally closed microvalve, experimental data showed that the pumping rate and back pressure can be significantly improved (see Fig. 5). In Fig. 5, we used two micropumps, one with a microvalve and another without a microvalve, to compare their performance. Apparently, the pumping rate has been significantly improved for the new design. Another function of the microvalve is to stop the back flow for the prevention of cross-contamination and the increasing of the backpressure, which was also confirmed by titration experiments. The improvement of the frequency response can be also observed from Fig. 5, while the maximum driving frequency can be as high as 90 Hz, which is much higher than the one without a microvalve.

High-speed camera images were then utilized to investigate the peristaltic-like membrane motion and the activation of the normally closed valve. The EMV frequency and applied air pressure were 90 Hz and 20 psi, respectively. The membrane operating sequence of the Type-I micropump during the filling and releasing process with compressed air is shown in Fig. 7a. The sample flow can be transported from the right-hand side to the left-hand side by utilizing time-phased membrane motion. The membrane structure was deflected as the compressed air fills the chamber and relaxes as the air is released from the air chamber, in sequence. The three air chambers are defined as #1, #2, and #3 in both designs. Apparently, the motion sequence of the Type-I design is a sequential activation from membrane #1 to #3 (Fig. 7a-1 to a-5). As shown in Fig. 7a-2, a-3, the normally closed valve can be activated as the compressed air filled air chambers #1 and #2 to deflect the membrane structure. Then the sample flow can pass through the floating block structure. The release process occurs when compressed air first flows into air chamber #1, then #2, and finally chamber #3 (as shown in Fig. 7a-6 to a-8). Figure 7a-7, a-8 shows that the floating block structure can be employed to stop backflow generated by the release of air from chamber #3. Similarly, the membrane working sequence of the Type-II design is represented in Fig. 7b. Correspondingly, the motion sequence of the Type-II design is operated in order from air chambers #1 to #3 (Fig. 7b-1 to b-5). The normally closed valve can be activated as the compressed air finally filled air chamber #3 to deflect the entire membrane structure. Then the fluid can flow over the floating block structure. The release process for the compressed air is performed by sequentially filling air chamber #1 to #3 (as shown in Fig. 7b-6 to b-8). Figure 7b-7, b-8 shows that the floating block structure can successfully break the continuity of the fluid stream and prevent backflow generation, while the sample fluid can be drawn back from the outlet reservoir when air chamber #3 releases.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig7_HTML.gif
Fig. 7

High-speed camera images of the peristaltic-like membrane motion and the normally closed valve activation for the two different micropump designs: a Type-I and b Type-II

3.2 Back pressure

Figure 8 shows the relationship between the back pressure and the driving frequency of the EMV at two applied pressures of 10 psi (a) and 20 psi (b), respectively. It can be clearly seen that the Type-II micropump has a higher back pressure than the Type-I design. Prior to micropump activation, the normally closed valve was provided with a hydrostatic pressure of 12 cm-H2O (1,215.6 Nt/m2), which was measured by filling the water into the tubing inserted into the outlet reservoir. Apparently, the back pressure between the inlet and outlet reservoirs is associated with the applied pressure and the driving frequency of EMV. It can be clearly observed that the back pressure increases with the increasing applied pressure. At a driving frequency of 90 Hz, the maximum back pressure for Type-I and Type-II micropumps can be as high as 60 cm-H2O (6,078 Nt/m2) and 68 cm-H2O (6,888 Nt/m2) at 10 psi, 73 cm-H2O (7,395 Nt/m2) and 84 cm-H2O (8,509 Nt/m2) at 20 psi, respectively. This high back pressure is generated by the integration of the normally closed valve. When compared to micropumps without normally closed valves (Wang and Lee 2006; Huang al. 2006), this value is about tenfold higher.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig8_HTML.gif
Fig. 8

The relationship between the back pressure and the driving frequency of the EMV at two applied pressures: a 10 psi and b 20 psi

The relationship between the pumping rates (at 90 Hz) and the back pressures at driving pressures of 10 psi (a) and 20 psi (b), respectively are shown in Fig. 9. Theoretically, the hydrodynamic pressure generated by the micropump (ΔP) is the product of the pumping rate (Q) and the fluidic resistance (Rf), which can be described as follows.
$$ R_{f} = \frac{12\eta \cdot L}{{wh^{3} }} $$
(1)
$$ \Updelta P = R_{f} \cdot Q $$
(2)
where w, L, h, and η are the channel width, the channel length, the channel depth, and the fluid viscosity, respectively. Since the hydrodynamic pressure will be fully balanced by the back pressure between the inlet and outlet reservoirs at a zero pumping rate. Therefore, the higher the back pressure is, the lower the pumping rate can be generated. The similar results have also been reported in the previous paper (Sim et al. 2003). Initially, the back pressure from the Type-I micropump is zero while the pumping rate is 490 μL/min for the 10-psi case. For Type-II micropumps, the pumping rate is 720 μL/min when the back pressure is zero. Once the back pressure increases, the pumping rate decreases accordingly. The maximum back pressure for the Type-I and Type-II micropumps are 0.85 psi and 0.97 psi, respectively, where the pumping rate becomes zero. Similarly, when increasing the applied pressure to 20 psi, the maximum back pressure for the Type-I and Type-II micropumps can be as high as 1.04 and 1.19 psi as the pumping rate drops to zero. However, many factors may influence the maximum back pressure value, including the viscosity of the liquid, surface tension, and surface roughness. The developed normally closed valve and the micropump devices have been operated for over four weeks without failure. Since PDMS is a resilient polymer material, the performance of the developed microfluidic devices was satisfactory as long as the applied air pressure is well regulated.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig9_HTML.gif
Fig. 9

The relationship between the back pressures and pumping rates when operated at a driving frequency of 90 Hz and at different driving pressures of a 10 psi and b 20 psi

3.3 Titration

In addition to a single micropump, a device comprising two Type-II micropumps and a V-shaped microchannel was developed, as shown in Fig. 10. By using a 2-way, 3-position EMV, two different samples can be controlled and injected into the reservoir separately. The two micropumps were individually connected to the EMV by two air tubing, and the switch of the EMV can control the direction of the compressed air to activate one of the two micropumps. Accordingly, not only does this simplify the control scheme, but it also reduces the size of the peripheral apparatus and the total power consumption of the device. To demonstrate the capability of the developed micropump for basic chemical analysis applications, a simple experiment regarding the titration of 0.1 N benzoic acid (C6H5COOH) with 0.1 N sodium hydroxide (NaOH) was performed by using a device comprising of two Type-II micropumps and a V-shaped microchannel, as shown in Fig. 10a. A pH meter was utilized to measure the pH values of the resulting solution. First, two different samples, 0.1 N NaOH (pH = 12.89, 100 μL) and 0.1 N C6H5COOH (pH = 3.02, 100 μL), respectively, were injected into the separate reservoirs (on the right-hand side). 100-μL 0.1 N C6H5COOH can be first transported with 900 μL/min into the reaction reservoir (on the left-hand side) by controlling the direction of compressed air through the EMV. The other sample, 0.1 N NaOH, can be transported with 900 μL/min and 10 μL each time into the reaction reservoir by switching the direction of the air inlet of the EMV. Therefore, the whole process can be controlled by just using a single EMV. Then the pH value of the resultant samples can be measured by the pH meter. The titration curve of the pH value in the reaction reservoir is shown in Fig. 10b. Experimental data show that a mixture with a pH value ranging from 2.8 to 12.3 can be successfully titrated. Additionally, the pH values of the sample in each sample reservoir remains at the original value, indicating that no backflow occurs into the reservoirs thanks to the normally closed microvalve which block the sample flows successfully. The proposed micropump can transport samples precisely, and successfully avoid contamination of samples.
https://static-content.springer.com/image/art%3A10.1007%2Fs10404-008-0356-7/MediaObjects/10404_2008_356_Fig10_HTML.gif
Fig. 10

a A new microfluidic device by combining two micropumps and a V-shape channel for the titration test. b The titration curve of the resulting pH values inside the outlet reservoir

4 Conclusion

This study reports a new pneumatic pump integrated with a normally closed valve. The normally closed microvalve can be activated by the hydrodynamic pressure which is generated by the pneumatic micropump itself, without an additional control system. When external compressed air was applied to the pneumatic pump, the time phase among the membranes generated a peristaltic deflection to compress the liquid channel and then drive liquid forward. The normally closed valve can efficiently stop any backflow, increasing the pumping rate and back pressure as well. The location of the normally closed valve affects the performance of the micropump. When the normally closed valve was placed at the end, the pumping rate and back pressure was further improved. Experimental data showed that the pumping rate was mainly determined by the applied pressure and the driving frequency of the EMV. From the titration test, it was found that this micropump can transport two different samples by using one EMV and can prevent contamination of sample reservoirs. The development of this micropump can be crucial for a future micro-total-analysis-system and other biomedical applications.

Acknowledgments

The authors gratefully acknowledge the financial support provided to this study by the National Science Council of Taiwan (NSC 96-2120-M-006-008).

Copyright information

© Springer-Verlag 2008