Abstract
In this review, different types of active microvalves are discussed in relation to cell cultivation devices, namely; magnetic valves, manually pin valves, pneumatic valves, thermal valves, bistable valves and phase-change valves. All of these valves are made of polydimethylsiloxane, which is gas permeable. The mechanisms of action, fabrication and study results of each valve are mentioned. Eventually, the best suitable valve for cell cultivation devices is chosen. Valves are judged on ease of fabrication, integration, actuation, external energy sources, power consumption, costs and efficiency. Manually actuated pin valves seem the best option for cell cultivation.
1. Introduction
Nowadays microfluidic systems are very important in the biology, chemistry and pharmaceutical sciences, mainly due to their miniaturization [1]. Microfluidics is all about the control and the analysis of fluids on a micrometer scale system. Especially Lab-on-a-chip (LOC) devices are frequently used. In here, a lot of aspects of the biology or chemistry are integrated into one single chip. However, the forces which are important in such micrometer scale systems are different than the forces on macro scale systems, like surface tension, laminar flow, fluidic resistance, diffusion and surface area to volume ratio [2]. Microfluidics is frequently applied in life science. The main applications are macromolecular analysis (DNA analysis, enzyme assays and immunoassays) and cellular analysis (cytometry, cell-based assays and cell culturing). But microfluidics is also very useful from a pharmaceutical point of view in relation to personalized medicine (drug metabolism, efficacy and interactions) [1,2,3].
Microfluidic devices are all about miniaturization, which results in a decrease in consumption of samples and chemical reagents (which could be very expensive), less chemical reaction- and analysis time, more reliability, enhanced sensitivity, disposability and portability in relation to normal cell cultivation devices. Besides this, miniaturization enables parallel operation and less cross-contamination. In this way microfluidics changes the technology of biology and chemistry [1,2,4].
However, due to a lack of reliable components, the development and higher levels of integration was delayed. Micropumps, -valves and –mixers continued the progression of microfluidic devices. Especially the microvalves are important for an integrated system. Valves are mainly used for the regulation of the fluid flow inside the device by on/off switching of the valves. The valves can stop the fluid flow or they can switch the fluid to a specific location. Besides this they can seal sample fluids and they have to withhold the pressure inside the device (for example generated by the expansion of air due to a higher temperature). Other important features of microvalves have to be: compact, reproducible, low power consumption, no leakage and contamination, easy to open, stable, disposable, rapid response time, the valve material may not influence the process, small dead volume and integration has to be possible [1,4].
There are a lot of different valves, which can be divided into two groups based on their actuation: the passive- and the active microvalves [1]. Active microvalves use energy coming from an external device, passive valves don’t need external energy for operation, but they use energy from a potential made in the sample or device itself. Besides this, passive valves mainly limit the fluid flow to one direction and provide a temporary stop or delay of the flow. These valves cannot be closed or opened without the change of the geometry [2,3]. Because passive valves, like check valves, capillary valves and diffuser valves [1], have mostly other applications than cell cultivation, or because they are a part of a (micro)pump or they are used in centrifugal devices [5,6,7,8], only the different types of active valves are discussed in this review, in relation to cell cultivation devices. Differences and examples of several active valves are mentioned, together with their mechanism of action, fabrication and the study results (efficiency). Eventually, the best valve for cell cultivation will be chosen.
2. Microfluidic devices – polydimethylsiloxane (PDMS)
In most of the microfluidic devices polydimethylsiloxane (PDMS) is a common material used for flow delivery. Using soft lithography, PDMS can be easily molded into a complex circuit in which a fluid flow can be regulated. PDMS is cost-effective, flexible (viscoelactic), non-toxic, not chemically reactive, simple, bio-compatible, non-flammable and optical transparent. Besides this, PDMS is gas permeable for oxygen and carbon dioxide. In this case culturing cells have enough oxygen supply for proliferation and growth, and a buffer can be obtained so the physiological pH of 7.4 can be maintained. But PDMS is also permeable for water vapor. This can result in drying problems and a change in the medium osmolarity. Another feature of PDMS is its hydrophobicity. To facilitate growth and attachment of cells this hydrophobicity has to be reduced by for example UV treatment, oxygen plasma or coating with charged molecules or extracellular matrix proteins to make the surface more hydrophilic [9]. To make sure that the comparison of the different types of valves will be correctly interpreted and that the best valve for cell cultivation can be used, all of the studies below are performed with PDMS.
3. Active microvalves
Active microvalves can be divided into three subgroups; the mechanical, non-mechanical and external active microvalves. This distribution is based on the actuation of the valves. Mechanical valves are flexible movable valves actuated by magnetic-, electric-, piezoelectric-, thermal-, bistable- or manually methods. Non-mechanical valves also have movable valves, but these valves are actuated by functionalized intelligent materials, for example electrochemical-, phase change- and rheological materials. Besides this, there are external active valves. These valves are actuated by an external system, such as modular or pneumatic [1]. For some types of active valves, other applications are frequently used (pumps) or the material is not PDMS (glass or SU-8) [10-14]. Therefore, electric-, piezoelectric-, rheological-, modular- and electrochemical valves are not discussed below, because comparison of the valves cannot be correctly interpreted.
3.1 Magnetic – mechanical valves
Magnetic actuated valves can open or close when a magnetic source surrounds the valve. Chen C Y. et al., (2011) introduced a permanent magnet for the closure of a microchannel. The device has five layers. The bottom layer is an iron plate, followed by a glass substrate, a PDMS microchannel, a spacer (polymethylmethacrylate, 1 mm diameter) and a permanent magnet (neodymium, 5 mm cube), see figure 1 (a). The microchannel of PDMS was fabricated using soft lithography.
Four 100 μm wide channels with different solutions (inlets, high pressure source of 2 psi for the fluid flow) come together in a 300 μm wide channel and then separate in four 100 μm wide channels (outlets). The mechanism of action is as follow: a spacer is pressed against the PDMS due to the attractive force between the permanent magnet and an iron plate. This results in a deformation of the PDMS and closure of the channel. This attractive force is critical for the function of the valve. The on/off state of the valve is required by removing or placing the magnet manually on the right location. Figure 1(b) illustrates this principle. To optimize the performance of the valve, the thickness of the spacer (important for the deformation of the PDMS layer and therefore for the closure of the channel) and the PDMS layer has to be as small as possible. But, a PDMS layer smaller than 1 mm results in difficulties during the bonding- and molding process. So, when the PDMS has a specific thickness, the spacer has to be thick enough to deform the PDMS layer. Also, the dimension of the spacer has to be larger than the dimension of the microchannel.
Figure 1: Magnetic actuated active microvalve. The five different layers of the device are visible in figure (a). Situation (a) is the open state; situation (b) is the closed state of the microchannel, where the spacer indirectly closes the channel via PDMS deformation due to the attractive force between the permanent magnet and the iron plate [15].
For the measurement of the electrical isolation and the time response, a straight channel (height and width 50 μm, length 30 mm) with a valve was used. The electrical isolation was measured by 100 mV and 20 ms test pulses. The resistance at the close- and open situation is calculated by the law of Ohm. Time responses were measured by continuously monitoring of the current response between two outlets at 100 mV. The degree of closing of the valve is revealed by the cross-sectional area of the channel. This is calculated, with respect to the different PDMS thicknesses and geometry of the channel, when the PDMS is deformed in different depths by the spacer. The process of closing is fast, depending on the channel geometry’s aspect ratio(width/height ratio), but not depending on the cross-sectional area of the channel. When the aspect ratio is larger, a shorter displacement of the spacer is required for channels which have the same cross-sectional area (faster valving process). Channels with the same aspect ratio act the same. But when the cross-sectional area is different, the degree of valve closing is identical. When the thickness of the PDMS layer increased, larger displacement of the spacer is needed for the same channel dimension, so a more powerful magnet is needed. The best thickness of the PDMS layer is less than 1.5 mm, because then a normal magnet can be used.
The close resistance was 10GΩ at 100 mV and 20 ms test pulse, for the opening it was 5 MΩ. This shows high electrical isolation. Switching time from open to close or vice versa is about 50-70 ms. Larger fluctuations are seen during closure of the valve than during opening due to the impact of the magnet on the elastic PDMS layer. The resistible pressure is approximately 200 kPa (could be increased by using a stronger magnet), this was determined by a pressure driven flow when the valve was closed. Pertubation vanishes in 200-300 ms.
The fabrication process of the valves is simple and no extra microfabrication is needed. Also, the location of the valves can be chosen randomly. Besides that, the valves can be easily integrated into a microfluidic system, they are inexpensive, feasible and the valves result in a flexible control of the fluid flow. The valve can be used as an on/off switch and no additional power source is required (just manual actuation). The device is disposable and portable and the magnet does not require an external power system to stay in a certain position. But there are some limitations: the valves are difficult to control accurately, the microchannel has to be made of soft material (otherwise, deformation is not possible), and the thickness of the PDMS is limited (the spacer has to be able to deform the PDMS). Due to the dimension of the magnet, the valves cannot be integrated intensively, and thus far the valves cannot be operated automatically [15].
3.2 Manually – mechanical valves
Brett M E. et al., (2011) introduced a simple valve (on chip) with a manually actuated pin, which can block the fluid flow directly in a microchannel. The pin valve is a with PDMS filled metal pin of one inch long, consisting of a PDMS tip on the end. This pin is inserted into a channel. The flow inside a channel is blocked when a seal is created due to pressing the PDMS against the substrate of the channel. Sealing of the channel prevents leakage at the valve port. By pulling the pin up, the seal can be broken and the fluid continues to flow (on/off flow control), figure 2 shows this mechanism.
Figure 2: Manually actuated pin valve. Situation (a) and (c) show the ON situation of the valve, (b) and (d) show the OFF situation. (a) image of the pin valve both blue and yellow can flow in the channel. (b) blue is blocked, only yellow can flow in the channel. (c) the pin valve is inserted in the channel, the PDMS on the tip of the pin is merged with the PDMS on the channel. (d) The pin valve is pushed through the channel, the PDMS on the tip is pushed against the bottom of the channel [16].
For the fabrication of the pin valves, stainless steel tubing was separated in sections of one inch. The ends of these sections were buffed (removing roughness and unevenness from the edges).
The pins were then put in a 96 well plate (bottom of the wells are pre-bored, and covered with tape). Now the well is filled with PDMS (degassed). After this, tape is placed on top of the 96 well plate, and a hole is bored in each well. The metal pins were put in the wells, and moved up and down to fill the pin with PDMS. After curing, the pins were removed from the wells. Excess PDMS was removed and all pins were cut in the same size, a punch was used to reduce the diameter in the PDMS, 2.3 mm (see figure 3). Using a Y-shaped channel (inlet- and outlet channels 4 mm length, 1 mm width, central channel 10 mm length, 1 mm width and 375 μm height) the pin valves can be validated. The valve is bored with a 2.05 mm needle, and dye was injected in one inlet of the device, the other inlet is filled with water. The flow rate of the syringe pumps was 10 μL/min. Pushing the valve down means blocking the channel with dye.
Figure 3: Fabrication scheme of metal pins, used as pin valve [16].
The breakthrough pressure (highest pressure the valves could withstand before the valves fail) was determined. The valves which where uneven cut at the bottom of the PDMS where excluded from the testing. The valves were tested several times and the pressure was released between each test. The amount of pressure each valve could withstand varied, overall it was similar for all conditions. There was no initial leakage and the sealing was complete. At the threshold the valve starts to leak and the highest pressure withstood cannot be reached anymore, the valve recovered slightly. Pressures between 2.7- and 17.5 psi could be reached with those valves. The mean pressure that is withstood in test one is 8.8 psi. With each usage this mean pressure decreased, some of the valves where not able to stop the flow anymore after one or two times they had been used, the reason could be that some of the valves where damaged or ripped during removal from the device, or the amount of pressure in the test before was too high. Pulling up the valves slightly could increase the breakthrough pressure at each test. The valves fabricated work efficiently, because most of the valves began to leak at 9 psi. The valves which could withstand a higher pressure are mostly better cut and better sealed. The usefulness of the pin valves was also demonstrated in a device for cell cultures. Here, an islet culture device was used. The valves were able to control the fluid flows during the whole experiment. No net vacuum was introduced during the opening.
The fabrication of this valve is simple and the time of fabrication is short. The valves could be made with low-cost and readily available materials, and they could be easily integrated into a lot of devices (prefabrication of the valves), without modifying the device. The actuation of the valves is simple, no external power is needed and no additional force is required to keep the valves closed. The valves could withstand high pressures for hours and multiple cycles and the dead volume is small. Besides this, the valves minimize the introduction of air bubbles in the device. But unfortunately no graded control is possible (just on/off) [16].
3.3 Thermal – mechanical valves
Gui L. et al., (2011) introduced a shape memory alloy (SMA) valve for controlling high aspect ratio channels (2.5 and 3.0) in the width direction. After SMA is deformed, it can return to its original shape. So, when heating the SMA wire, the wire shrinks. When cooling down to the original temperature the wire returns to its normal length. The SMA wire requires an electrical current for heating the wire by its joule heating. This makes the fabrication process simpler. Figure 4 shows a schematic view of the SMA valve. The black wire is the SMA wire (14 mm, 0.015 inch. diameter), the so called A-wire. This wire is coiled by an aluminum (grey) wire, the so called B-wire. The ends of the A-wire are connected to a 1 mm SMA wire (C-tip), and a 5 mm wire (D-handle). This whole structure lies in a PDMS layer with a microchannel. The D-handle is large compared to the C-tip, and is difficult to move. This serves to hold the whole wire in position due to a large resistance in the PDMS layer and only the C-tip can move. The B-wire can be electrically heated causing the A-wire to warm up due to joule heat. When the A-wire reaches a temperature higher than the transition temperature (70 ◦C), the wire shrinks in length and the C-tip comes closer to the D-handle en moves the adjacent channel. But, also the C-tip and D-handle will be warmed up and shrink, but the shrinkage of the C-tip is negligible (small length), the shrinkage of the D-handle does not affect the valve.
Figure 4: schematic view of the shape memory alloy (SMA) wire for thermal actuation of the valve for controlling the flow inside a microchannel [17].
Figure 5 shows the mechanism of this kind of valve in width direction. Situation (a) is the normal open valve. Here, the wire is placed above the channel and the C-tip is close to the channel wall.
When the B-wire is electrically heated, the A-wire shrinks and C-tip squeezes the channel and closes it. Situation (b) shows the normal closed valve. Here, the A-wire does not cover the channel, but the C-tip is close to the channel wall. A SU-8 block closes the channel. Heating en shrinking of the wire results in a stretch of the channel wall away from the block and thereby opening of the channel.
The SMA wire cools down when the electrical current is stopped and no joule heating is generated anymore.
Figure 5: Thermally actuated SMA valve. Situation (a) is the normal open state. Electrically heating of the grey aluminum wire results in a shrinking of the black SMA wire. The channel closes. Situation (b) is the normal closed valve. Heating the wire results in a stretch and opening of the channel [17].
Straight high aspect ratio channels were used with 150 μm height and 50-60 μm width. Devices were made using soft lithography. The B-wire was coiled around the A-wire (50 loops, no distance between the loops). The C-tip and D-handle were welded to the A-wire. To reduce the leakage, the C-tip is covered with SU-8, which helps to close the channel (smoother, softer and less sticky than SMA on the PDMS). Without SU-8 the valve has low reliability and repeatability due to the roughness of the surface. But it could deform the channel due to heating and cooling down repeatedly. Also, at the fabrication procedure for the normally open valve deformation of the channel can occur. To be sure that the channel will be completely closed or opened, the C-tip has to be at the same level as the bottom of the channel wall.
The valves were tested at a constant flow rate. The electrical current was 1.8 A in the beginning, 1.5 A when the actuation occurs (prevent overheating). Response time was measured for flow rates of 0.05-4.0 μL/min. For normally closed valves, an increase in flow rate results in a decrease in opening time and an increase in closing time (higher flow rates make closure more difficult than opening). The opening time is not influenced by the channel width, the closing time is more influenced, but the trend is similar. This closure time is influenced by the cooling time of the SMA wire, which is influenced by the PDMS layer and environment.
For normally closed- and open valves the response time for opening/closing of the channel decreases when the electrical current increases (increased heat and so increased shrinkage). At a normally open valve (1.8A electrical current) the working distance is shorter to open the flow than to close the flow. The opening time is therefore shorter than the closing time. But when the widths increase, the working distance is influenced, and also the opening- and closing time is increased. The valves shown in this design worked well.
The response time of normally closed valves are less influenced by the channel with, so this valve could be used for channels with a larger width. Materials could influence the response time. A higher actuation current can require rapid response times, but is limited, because of too much heat for the PDMS and the liquid [17].
3.4 Pneumatic – external valves
Hulme S E. et al., (2009) created prefabricated pneumatic valves. These valves are first fabricated and they are integrated into a device when they are needed. Pneumatic actuation of microvalves is all about the actuation of valves with pressurized gas. Prefabrication of valves results in uniform operation of the valves due to standardized procedures; the valves will all need the same amount of pressure to close a channel. Figure 6 (a) shows the prefabrication of pneumatic valves. Here, a PDMS structure of two layers is used, to support the valves. In one layer (width of 2 mm) holes are made with a diameter of 3.5 mm. Then a second layer is introduced (2 mm width) above the first layer. The two layers are exposed to oxidizing plasma and sealed together with covalent bonds. In the second layer (on top) holes are made with a diameter of 1.5 mm above the larger holes in the first layer. Polyethylene tubing is used for the air inlet. The tubing has a diameter of 1.5 mm, the tips of the tubing are melted (diameter is 3 mm). This tubing is threaded through the PDMS structure, starting at the larger holes. At a certain point the tubing stay in the PDMS structure and cannot leave the structure, due to a 3 mm diameter tubing and a 1.5 mm hole. Then photocurable polyurethane (epoxy, NOA 81) is added around the interface between PDMS and the tubing. UV light requires that NOA 81 becomes hard. The tubing is now glued in the PDMS structure and the structure is airtight sealed. After this a third PDMS layer (1 mm width) is added under the PDMS structure and the valve is cut out of the structure. Figure 6 (c) shows the incorporation of the prefabricated valve into the device. Masters are used, which are made with photolithography, to make features in the SU-8. Then the prefabricated valve is manually pressed onto a SU-8 channel on the master, coated with PDMS. This PDMS results in a balanced valve on top of the SU-8 channel and PDMS prevents bubbles between the master and the valve. When the valve is positioned in the right place, uncured PDMS was put on the master, so the valve is molded with the device. Then the device was plasma oxidized and sealed to a glass substrate. Connecting the tubing to pressurized air makes actuation of the valve possible. By adding pressurized air (regulation of the gauge pressure), the channel closes (see figure 6 (c) (6)).
Figure 6: (a) prefabrication of the pneumatic valve, (b) a pneumatic valve when the prefabrication is finished, (c) the way in which a pneumatic valve is incorporated into a device [18].
The density of the channels in a device is limited by the size of the valves. The minimum valve-to-valve (channel-to-channel) separation is determined by the footprint. For these pneumatic valves the footprint is 3.5 x 3.5 mm. Only small numbers of valves can be integrated into a device. Each valve was embedded in one inlet of a Y-shaped device to test the valves. This device has two inlets (width 50 μm), and one outlet (50 μm width), with two laminar flows of equal width. One inlet contains dye. The valves were tested using channels of different height. Pneumatic valves could close channels with a height of 10 μm or less. To see how the valve performs, a valve was closed and the decrease in width of the laminar flow was monitored, which is proportional to the decrease in flow rate. The performance of a single valve (after several times of opening and closure) is compared to that of different valve, fabricated with a standardized procedure. To close the channel completely, a pressure of 45 psi (310 kPa) is required. The membrane that closes the channel has an area of 3500 x 50 μm. Here, the membrane deforms more along the axis of the channel.
The transition from an open- to closed valve was continuous, so the valves can be used as a regulator of the fluid flow. When the applied pressure increases, the fraction of the original flow width decreases, so also the flow rate decreases.
After several cycles of opening and closing of the channel (up to 20 cycles), a single valve performed the same at different pressures (PDMS membrane is 200 μm thick). So the single valve measured shows reproducible performances. Also different valves with the same procedure show equal performances during 7 cycles at different pressures (PDMS membrane is 150 μm thick). So the different valves are also reproducible.
The fabrication is simple and flexible and provides easy integration of valves into a device, no special equipment or advanced techniques are required. When another valve is needed, not the whole device has to be made new. The operation of the valves is uniform due to prefabrication and standardized procedures and the valves show reproducibility. The valves can be fabricated quickly with low costs. Multiple modes of actuation can be embedded for different applications and the valves can be controlled precise. But to maintain the valve is a closed state, a continuous apply of pressurized gas is needed [18].
3.5 Bistable – mechanical valves
Chen A. et al., (2014) introduced bistable microvalves. These valves are quite unique because the valves remain stable in an open or closed position without power. Only for the transition between the two states power is needed. The transition is achieved through pneumatic actuation (actuation of valves with vacuum). Here, the device is manually operatable enabled by a vacuum pneumatic network. A simple manual operation can direct open or close the valve. No internal electrodes or external systems to maintain the two states are required. This design, operation and fabrication are simple, less power is needed (less costs) and integration into devices is easy. The portability is increased, but the fluid control stays high. Figure 7 shows a schematic view of the components in the bistable valve. The device consists of three layers PDMS. The middle layer (a moving diaphragm membrane of 200 μm) was plasma oxidized to the other two layers (1.0-1.5 μm). Laser-etching techniques were used to make the different structures in the PDMS layers (membrane cantilevers, cavities, a vacuum channel, chambers, venting channel, output etc.). The center of the chambers has a pillar structure. The chambers in the top layer can be pressed manually. One chamber serves as the activation of vacuum, the other serves as releasing the pressure in the device. When a chamber is pressed, the pillar deforms the underlying membrane into the cavity. This membrane also separates the inlet channel (bottom layer) from the chamber, acting like a pneumatic check-valve. The vacuum channel is connected to a vacuum source. In this study three different sources are used; one with a aspiration pressure of 85 kPa, one with 30 kPa, and one with a negative pressure of 94 kPa. A venting channel is connected to the environment of the atmosphere. Other channels in the device connect the two chambers. The output channel controls the pressure of the vacuum or atmosphere.
Figure 7: the components of a bistable pneumatic microvalve, which can be controlled manually. Pressing the chambers results in the pressure of the pillar onto the membrane, which deforms in the cavity. Vacuum channel is connected to a vacuum source, the venting channel is connected to the atmosphere. The output channel controls the pressure of both [19].
Figure 8 shows the mechanism of the bistable valves. The device consists of two chambers; the pressure release chamber (PRC) and the vacuum activation chamber (VAC). The VAC is adjacent to the vacuum channel and when this chamber is manually pressed, the membrane is stretched and vacuum enters. A normally closed valve opens. The vacuum state is activated. This state remains because the pressure between the atmosphere and VAC is different. When pressing the PRC, the pressure inside is directed to the atmosphere. The pressure gradient in the VAC is removed and the connection to the vacuum channel is closed. The atmospheric state is introduced and maintained by equalizing the external- and internal pressure. The device is now in its original closed state.
Figure 8: Bistable microvalve. Pressing the VAC chambers result in a deformation of the membrane and thereby activation of the vacuum (vacuum flows in) (b/c). Pressing PRC results in an open atmosphere channel. The vacuum is released (d) [19].
No degradation of performance and integrity is showed after 200 cycles of transition between vacuum- and atmospheric state. An increase in the fluid pressure (2.5-10.0 kPa) results in a nearly linear increase in the vacuum pressure needed for opening of the valve (2.0-6.0 kPa).
The minimum valve-activation force is nearly constant (7 mN) when the fluid pressure is around 5-10 kPa. A fluid pressure <5 kPa will not open the valve immediately and completely at a weaker vacuum level. There is a constant leakage of vacuum through the PDMS. The vacuum syringe loses pressure at a rate of 0.0133 kPa per second, when it is connected to the device (0.0040 kPa per second without device). From 94 kPa to 5 kPa, the device will last approximately 2 hours. These valves improve the portability and a high degree of controlling the fluid maintains. This kind of valves is used in point-of-care devices (immunoassays and blood-typing assays) in low resource settings. [19].
3.6 Phase change – non-mechanical valves
Yang B. et al., (2007) reported a novel latchable phase-change valve which is energy-efficient. Only switching between the open/closed state requires energy. Paraffin wax is used, which transforms from solid to liquid to solid. Heating results in the melting of paraffin which can result in valve switching. Here, the paraffin is reusable and no fluid contamination can occur (sealing of paraffin). The valves described are flexible and robust, but external heaters and pneumatic pressure sources are used, which influenced the switching time response, energy consumption and size negatively. Integrated systems can reduce these features. Figure 9 shows the composition of the valve. This valve has three layers. The bottom layer is a glass substrate (0.5 mm thick), the middle layer is a PDMS layer (30 μm thick) with a fluid channel (with an inlet and an outlet, the channel is 12 μm high, 200 μm width, 6 mm long), and the top layer is a PDMS layer with a chamber for the paraffin wax (5.5 mm thick, the chamber: 300 μm width, 350 μm height, 2 mm long, 0.21 nL volume) , consisting of an inlet for pneumatic control. Soft lithography is used to make the device from PDMS, which is flexible and facilitates the membrane between the channel and chamber to deform easy under the required pressure. The pneumatic control ports and fluid inlet/outlet were connected to the device using epoxy glue. The chamber (paraffin) and the channel (fluid) are separated by a thin flexible membrane, no paraffin can reach the fluid even at high pressure. This membrane is the active component in the valve to open/close the valve. The paraffin wax is the actuation material and a phase change of the paraffin (low melting point of is 44-46◦C) is needed for valve switching. The channel itself is rounded to make sure that the valve will close completely.
The paraffin is injected in the chamber while the whole device is put in a water bad (65◦C). To get the paraffin in the chamber, two pressure ports are used, connected to the chamber. The first port guided the melted paraffin in the chamber, the second port vented air out while the paraffin is injected. When the chamber was completely filled with the wax, the last port was sealed permanent. The first port is used as pressure control when the valve is switched between the two states.
Figure 9: Schematic view of the composition of the phase-change valve. Situation (a) shows the composition along the fluid channel, situation (b) shows the composition across the channel [20].
Figure 10 shows the mechanism of this kind of valve. Heating the paraffin results in melting of the wax. A pressure, which is applied at the pneumatic control port is transferred through the paraffin to the membrane. The membrane deforms, which results in a narrowed channel.
When the pressure is high enough, the channel will be closed completely.
As the paraffin cools down (solid phase), the valve is locked in a closed state without other energy/pressure needed to get the valve in a special position. When melting the paraffin again, the elastic force of the deformed membrane forced the valve to go back to its original state. This happens passively (without an external pressure). As the paraffin cools down, the valve is again locked, this time in the open state.
Figure 10: Phase-change microvalve. When heating the paraffin (yellow) a pneumatic pressure is transferred to the membrane, closing the channel. Reheating the paraffin results in reformation of the membrane due to elastic forces, the channel opens [20].
Flow rates are measured with a flowmeter (resolution 1 μL/min). Heating and melting of the paraffin happened using an isothermal water bath (63◦C), cooling down of the wax happens passively at 25◦C (room temperature) with a heat sink under the device.
The thermal energy required to heat the device from 25 to 63◦C is 58.9 J (not optimized method of heating). For melting paraffin 40 mJ is enough. The flow rate in an open valve is 27 μL/min. A closed valve has a flow rate of <1μL/min (under detection limit). The inlet pressure is 40 kPa. From open to close state required 60 seconds. The pneumatic control pressure was 70 kPa. From closed to open state required 100 seconds (pneumatic pressure is stronger than the passive elastic force). The heating time is relatively long, because the whole device has to be warmed up. The cooling down time of the paraffin was about 3-5 minutes (depends on the heat transfer of the device). A heat sink reduces the time for cooling down. The switching time depends on the heating time. When the valve was open the flow rate is linear with the driving pressure (constant channel cross-section and flow resistance). When the valve is closed the flow rate is zero up to driving pressures of 35 kPa. Further increase of the pressure results in an increase in leakage (strength of the wax is not sufficient to keep the valve closed). At a pressure higher than 70 kPa, the flow rates of the closed- and open valves are nearly equal. When the inlet flow pressure is higher than 35 kPa, the valve cannot be closed passively anymore (critical latching pressure). The elastic restoring pressure is 28 kPa at the closed state, this is not influenced by the driving pressure at the inlet [20].
...(download the rest of the essay above)