BIOMICROFLUIDICS 4, 046501 共2010兲
`
`Electrical power free, low dead volume, pressure-driven
`pumping for microfluidic applications
`Mario Moscovici,1,2 Wei-Yin Chien,2 Mohamed Abdelgawad,1,3,a兲 and
`Yu Sun1,2,b兲
`1Department of Mechanical and Industrial Engineering, University of Toronto,
`5 King’s College Rd., Toronto, Ontario M5S 3G8, Canada
`2Institute of Biomaterials and Biomedical Engineering, University of Toronto,
`164 College St., Toronto, Ontario M5S 3G9, Canada
`3Department of Surgery (Urology), University of Toronto, 100 College Street, Toronto,
`Ontario M5G 1L5, Canada
`共Received 19 July 2010; accepted 21 September 2010; published online 13 October 2010兲
`
`This paper presents a simple-to-construct, low dead volume pump capable of gen-
`erating a wide range of positive and negative pressures for microfluidic applica-
`tions. The pump generates pressure or vacuum by changing the volume of air
`confined inside a syringe and is able to generate pressures between ⫺95 and
`+300 kPa with a resolution as high as 1 Pa. Different from syringe pumps and
`electrokinetic pumping, which are capable of controlling flow rates only, our pump
`can be used to generate constant flow rates or constant pressures, which are re-
`quired for certain applications such as the aspiration of biological cells for bio-
`physical characterization. Compared to syringe pumps, the new pump has almost
`zero dead volume and does not exhibit pulsatile flows. Additionally, the system
`does not require electrical power and is cost effective 共⬃$100兲. To demonstrate the
`capabilities of the pump, we used it to aspirate osteoblasts 共MC3T3-E1 cells兲 and to
`determine Young’s modulus of the cells, to generate a concentration gradient, and
`to produce variable-sized droplets
`in microchannels using hydrodynamic
`focusing. © 2010 American Institute of Physics. 关doi:10.1063/1.3499939兴
`
`I. INTRODUCTION
`
`Pumping is indispensable for microfluidic applications, such as culturing cells inside
`microchannels,1 separating a mixture of analytes,2 amplifying DNA fragments,3 sorting cells4 or
`isolating them,5 controlling drug delivery to individual cells,6 or synthesizing particles inside
`microchannels.7 Consequently, the development of pumps has received and will continue to re-
`ceive significant attention in microfluidic research.
`The various pumping methods in microfluidics can be classified into on-device pumps and
`external pumps. In the former category, micropumps are integrated in the microfluidic device and
`are fabricated with the device itself. This category includes piezoelectric, thermal, electrohydro-
`dynamic, electrostatic, rotary, and acoustic micropumps, among others.8 Although on-device
`pumps increase device portability, integrating pumping components 共e.g., diaphragms, actuators,
`valves, and heaters兲 on the device significantly increases fabrication complexity and device cost.
`Since most microfluidic devices are made to be disposable to reduce contamination across
`samples, on-device pumping is not always an attractive choice.
`The second category, external pumping, includes mainly electro-osmotic pumping and syringe
`
`a兲
`Currently with the Mechanical Engineering Department, Assiut University, Egypt.
`b兲
`Author to whom correspondence should be addressed. Department of Mechanical and Industrial Engineering, Institute of
`Biomaterials and Biomedical Engineering, University of Toronto, 5 King’s College Road, Toronto, ON M5S 3G8,
`Canada. Tel.: 416-946-0549. FAX: 416-978-7753. Electronic mail: sun@mie.utoronto.ca.
`
`1932-1058/2010/4共4兲/046501/10/$30.00
`
`4, 046501-1
`
`© 2010 American Institute of Physics
`
`1
`
`

`
`046501-2
`
`Moscovici et al.
`
`Biomicrofluidics 4, 046501 共2010兲
`
`pumps. Electro-osmotic pumping exploits ions accumulated in the electric double layer when a
`liquid comes into contact with channel walls. When an electric field is applied between two
`electrodes in the channel inlet and outlet reservoirs on the device, the ions in the electric double
`layer are attracted toward the opposite polarity electrode, hence inducing a net flow in the
`microchannel.9 Electro-osmotic pumping can induce and control the flow with no moving parts;
`however, it suffers from bubble generation due to electrolysis at the electrodes, which may cause
`channel blockage. Moreover, electro-osmotic flow requires high voltages 共e.g., kilovolts兲 to induce
`significant flow rates, which can result in significant Joule heating. Expensive high-voltage se-
`quencers are typically used for electro-osmotic pumping, increasing the complexity and capital
`cost associated with the setup.
`Syringe pumps are relatively cheaper compared to electro-osmotic pumping setups. However,
`they involve large dead volumes in the syringe and tubing. This large dead volume compromises
`the advantage of reduced sample and reagent consumption in microfluidic devices. In addition,
`syringe pumps are prone to producing pulsatile flows at low flow rates.10
`This paper reports a simple-to-construct system for pumping liquids in microchannels. The
`pump relies on pressure rather than forced displacement to drive liquids in microchannels. This
`technique was previously used in stop-flow lithography11 using external pressure sources and
`regulators. Pressure-driven pumping was also implemented using electrolysis gases produced
`on-chip,12,13 a pressurized gas in a microcavity,14 and capillary pressure inside microdroplets.15 In
`our pumping system, pressure is produced by changing the volume of air confined inside a
`syringe, which is simpler and more predictable compared to the aforementioned pressure-driven
`methods. Besides the capability of producing positive pressures, our method is also capable of
`producing vacuum, which is useful in some applications such as bubble elimination.16 Moreover,
`using pressure to drive liquids in microchannels results in almost zero dead volume and does not
`exhibit pulsatile flows, which are two limitations of conventional syringe pumps. To demonstrate
`the effectiveness and versatility of our proposed pumping system, we demonstrated its usage in
`several microfluidic applications, including micropipet aspiration of biological cells, droplet gen-
`eration in microchannels, and the generation of a concentration gradient.
`
`II. EXPERIMENTAL METHODS
`
`A. System setup
`The pump consists of a plastic syringe connected to a microfluidic device 共via 1/16 in. Tygon
`tubing兲 through an intermediate tank, which controls the initial system volume and thus changes
`pressure resolution and range 关Fig. 1共a兲兴. Different syringe sizes can be used to allow for course or
`fine tuning of the pressure. We used two 60 ml plastic syringes as a variable size tank to allow for
`a wide range of resolutions from the same setup by tuning the tank volume. However, larger or
`smaller containers can be used to further increase the resolution or pressure range, respectively.
`Syringes were mounted on an off-the-shelf threaded corner-clamp to facilitate fine changes to
`syringe volume. The clamp also helps hold the syringe plunger in its place against generated
`pressure/vacuum for steady pressure generation. Tubing was connected to syringes via barbed-luer
`connection adapters 共McMaster-Carr, Atlanta, GA兲 to allow for fast assembly/disassembly. A luer
`connection valve was added to the system to enable fast system venting. An optional pressure
`sensor 共D1-4V 10 INCH, $100, All Sensors, Morgan Hill, CA兲 was incorporated into the system
`to measure the actual generated pressures and compare it with theoretically predicted values.
`
`B. Pump operation
`
`To achieve certain pressure range for a particular application, first the tank volume is chosen
`according to the analysis in the results section. Generally speaking, a smaller tank volume corre-
`sponds to larger pressure range and vice versa. In the setup shown in Fig. 1共a兲, two 60 ml syringes
`are used as a tank to have the ability to change tank volume without the need of changing the tank
`itself. Second, a proper syringe size is chosen to achieve the required resolution, which depends on
`the size of the smallest division on the syringe relative to the tank volume 共the smaller the syringe
`
`2
`
`

`
`046501-3
`
`Simple and versatile microfluidic pump
`
`Biomicrofluidics 4, 046501 共2010兲
`
`FIG. 1. 共a兲 Picture of the pumping system with a pressure sensor included for system characterization. Two 60 ml syringes
`were used as a tank to allow for easy changes in tank volume. However, larger/smaller containers can be used as a tank to
`increase range/resolution, respectively. The pressure sensor 共powered using a standard 5 V adapter through a bread board兲
`is optional and was used to compare generated pressures with theoretical calculations. Sensor output span was 4 V 共linearly
`proportional to pressure兲 and was measured using a conventional multimeter. 共b兲 Working principle with variables used in
`Eq. 共2兲.
`
`volume, the higher the resolution; see Sec. III for a complete analysis兲. Since resolution is also
`dependent on tank volume, a compromise between high range and high resolution exists. After the
`proper syringe and tank sizes are chosen, the syringe is mounted on the threaded clamps and is
`connected to the tank via 1/16 in. tubing as shown in Fig. 1共a兲. Rotating the knob where the
`syringe is mounted causes the volume in the system to change in small increments, ⌬V in Fig.
`1共b兲, which consequently changes the pressure of the system. If pressure needs to be measured, a
`
`3
`
`

`
`046501-4
`
`Moscovici et al.
`
`Biomicrofluidics 4, 046501 共2010兲
`
`pressure sensor can be connected to the system. Sensor output, in volts, is measured using a
`multimeter and converted into pressure according to sensor characteristics.
`
`C. Flow measurements
`
`For flow rate measurements, the syringe plunger position was set to generate certain vacuum
`pressure according to the analysis given in Sec. III. The actual generated vacuum was measured
`using the pressure sensor and was used to pump 1 ␮m diameter FITC fluorescent beads 共Bangs
`Laboratories, Fishers, IN兲 suspended in methanol 共␮=0.58 mPa S兲 inside a microchannel
`共140 ␮m wide, 25 ␮m high, and 30 mm long兲. The speed of the FITC beads was measured by
`increasing the exposure time to visualize streaks of the beads as they flow in the channel. Speed
`was calculated by dividing the streak length by the exposure time. Average flow rate was calcu-
`lated from the maximum speed at the channel centerline using the following equation17 for chan-
`nels with height to width ratio ␣⬍0.5:
`
`=冉 m + 1
`
`m
`
`冊冉 n + 1
`
`n
`
`冊冋1 −冉 y
`
`b
`
`u
`uav
`where u is liquid velocity at any location in the channel, uav is average liquid velocity, b is the
`channel height, a is the channel width, y and z are the two coordinates measured from channel
`centerline along channel height and width, respectively, and m and n are two numerical factors
`calculated according to
`
`冊n册冋1 −冉 z
`
`a
`
`冊m册,
`
`共1兲
`
`m = 1.7 + 0.5 ␣−1.4,
`
`再n = 2
`
`␣⬍ 1/3
`n = 2 + 0.3共␣− 1/3兲 ␣ⱖ 1/3.
`
`冎
`
`共2兲
`
`共3兲
`
`D. Micropipet aspiration of osteoblasts
`Conventional micropipet aspiration experiments were conducted on osteoblasts 共MC3T3-E1兲
`cells using the proposed pump to verify its capacity for high resolution vacuum generation. A
`borosilicate glass micropipet tip 共5 ␮m diameter兲 was held by a micromanipulator 共Sutter Instru-
`ment Co., CA, USA兲 mounted on an inverted phase-contrast microscope. The micropipet was
`connected 共via 1/16 in. tubing兲 to the pump where the tank volume used was 120 ml and a 1 ml
`syringe with visible divisions of 0.01 ml was used to generate the vacuum. The resolution gener-
`ated from the setup was 8 Pa with a range of 0–共⫺730兲 Pa.
`First, the micropipet tip was immersed inside a droplet of culture media with no cells for a few
`minutes to stabilize capillary rise into the pipet tip. A second droplet of culture medium with cells
`was added to the droplet after the system was stabilized. Then, the tip was positioned close to the
`surface of a target cell, and a small negative pressure 共20–50 Pa兲 was applied in order to immo-
`bilize the cell and to form a complete seal. From this reference state, subsequent suction pressures,
`in increments of 16 Pa, were then applied and images of the aspirated cell were captured from
`which the aspiration length and pipet diameter were measured. Young’s modulus values of the
`aspirated cells were determined by using common biomechanics models that approximate the cell
`as an elastic half-space solid.18,19
`
`E. Concentration gradient
`
`To generate a concentration gradient, red and green food dyes were loaded into the inlet
`reservoirs of a gradient generator device.20 A vacuum pressure of 15 kPa 共tank size=11 ml and
`syringe displacement=2 ml兲 was applied at the outlet, pulling the two dyes through the branches
`of the microchannel network and resulting in a seven-step concentration gradient at the outlet
`channel.
`
`4
`
`

`
`046501-5
`
`Simple and versatile microfluidic pump
`
`Biomicrofluidics 4, 046501 共2010兲
`
`F. Droplet generation
`Water droplets 共aqueous phase兲 were generated in a continuous phase of silicone oil 共DMS-
`T01, viscosity⫽1 cSt., Gelest, Inc., Morrisville, PA兲 using a hydrodynamic focusing device similar
`to one reported previously.21 First, the whole device was filled with silicone oil, and a pressure of
`20 kPa 共Vtotal=60 ml and plunger displacement=−10 ml兲 was applied on the two side reservoirs
`to generate oil flow in the downstream channel. Water was then added to the center channel
`reservoir, and a pressure of 100 kPa 共Vtotal=60 ml and plunger displacement=−30 ml兲 was
`therein applied to generate a stream of water along the center of the downstream channel. Tween
`20 共Sigma-Aldrich, Oakville, ON兲 can be added to the water stream as a surfactant, at low
`concentrations 共e.g., 1.75% v/v兲, to facilitate breaking of the water-oil interface. Oil flow rate in
`the side channels was then increased by increasing the applied pressure on oil reservoirs to 114
`kPa 共Vtotal=60 ml and plunger displacement=−32 ml兲 to pinch the water stream into discrete
`droplets. Smallest droplet size 共see results section兲 was achievable with a pressure of 300 kPa
`共Vtotal=60 ml and plunger displacement=−45 ml兲 applied on the oil reservoirs.
`
`III. RESULTS AND DISCUSSION
`
`Any change in the volume of a confined gas produces a change in its pressure according to the
`equation of state. Assuming an isothermal process, the change in pressure of an ideal gas is
`inversely proportional to its volume according to
`
`PV = C,
`
`共4兲
`where P is the absolute pressure, V is the volume of the confined gas, and C is a constant that can
`be calculated from the initial pressure and volume values. Assuming that the initial volume of our
`system is Vtotal, which comprises the volume of the tank, the tubing, and initial volume on the
`syringe 共Vtotal= Vtank+Vtubing+ Vsyringe,initial兲, and that the initial pressure is atmospheric pressure
`共Patm兲, then the gauge pressure of the system after a volume change 共⌬V兲 can be calculated from
`
`Pgauge = − Patm冉
`
`冊,
`
`⌬V
`Vtotal + ⌬V
`
`共5兲
`
`where ⌬V is positive for increase in volume.
`The generated pressure from the proposed pump agreed quite well with Eq. 共5兲 关Fig. 2共a兲兴,
`with a maximum error less than 6% and always in the direction of increasing the absolute value of
`the generated pressure. This error was possibly produced by the increase in air temperature due to
`the heat generated from the friction between the syringe plunger and its barrel, which deviates
`from the isothermal process assumption of Eq. 共4兲. This was clearly demonstrated when we tested
`the system hysteresis, which showed a total increase in the system pressure when the plunger
`returned to its original starting point 共supplementary material, Fig. S1兲.22 Heat generation and
`corresponding pressure error can be mitigated by choosing a proper tank or syringe size to allow
`for generating the required pressure with shorter plunger displacement. Nonetheless, the changes
`in air temperature due to plunger friction had negligible effects on the generated pressure at typical
`pressure ranges required for microfluidic applications.
`Generated pressures were highly reproducible 关largest standard deviation in Fig. 2共a兲 is ⫾600
`Pa兴, suggesting that the integration of a pressure sensor into the system is not crucial since the
`required pressures can be generated with high accuracy 共less than 6% error兲 by calculating the
`corresponding values of Vtotal and ⌬V. The system can also support applications requiring instan-
`taneous changes in pressure 共e.g., micropipet aspiration兲 with both the resolution and pressure
`range adjustable by choosing the right tank volume and syringe size 共which specifies the smallest
`achievable volume change ⌬V兲. With a 1 mlsyringe 共0.01 ml divisions兲 and a tank volume of 1 l
`共a water bottle兲, a resolution of 1 Pa is achieved.
`Although the proposed pump generates constant pressures rather than constant flow rates,
`which are more commonly used in microfluidic applications, desired flow rates can still be accu-
`
`5
`
`

`
`046501-6
`
`Moscovici et al.
`
`Biomicrofluidics 4, 046501 共2010兲
`
`FIG. 2. Pump performance. 共a兲 Vacuum pressure range of ⫺35 kPa with a resolution of 0.9 kPa. Error bars appear inside
`data markers and present ⫾1 standard deviation. 共b兲 Flow rate inside a microchannel vs applied pressure. Blue marks
`represent theoretically calculated flow rates, according to Eq. 共6兲, from vacuum pressures generated by induced volume
`changes on the syringe, according to Eq. 共5兲. Red marks are actual flow rates measured inside the microchannel. Horizontal
`error bars represent ⫾1 standard deviation in the generated pressure, whereas vertical error bars represents ⫾1 standard
`deviation in measured flow rate.
`
`rately and easily generated. In microfluidic devices, where laminar flow is dominant, flow rates are
`linearly proportional to the applied pressures according to Eq. 共6兲 共Ref. 23兲,
`
`4
`␲Dh
`2C␮L
`where Q is the flow rate, ⌬P is the pressure difference between microchannel ends, Dh is the
`hydraulic diameter of the microchannel connecting the two reservoirs, ␮is the liquid viscosity, L
`is the channel length, and C is a constant depending on the channel cross section 共C=64 for
`circular channels, and C=78 for channels we tested兲.23 This means that any desired flow rate can
`
`共6兲
`
`Q =
`
`⌬P,
`
`6
`
`

`
`046501-7
`
`Simple and versatile microfluidic pump
`
`Biomicrofluidics 4, 046501 共2010兲
`
`TABLE I. Comparison between the performance of our system and other conventional micropumping techniques. Data for
`other pumping techniques were obtained from Refs. 33 and 8.
`
`Type
`
`Vibrating
`diaphragm
`
`Peristaltic
`
`Fluid displacement
`
`Rotary
`Electroosmotic
`Syringe pumpb
`Our system
`
`Actuation method
`
`Electromagnetic,
`piezoelectric,
`magnetic, etc.
`Thermopneumatic,
`piezoelectric
`Magnetic fluid,
`gas permeation,
`phase change
`Magnetic or viscous
`dc or ac voltage
`dc voltage
`Manual
`
`Pressure range
`共kPa兲a
`
`Maximum
`resolution
`共Pa兲a
`
`Maximum
`flow rate
`共␮l/min mm2兲 a
`
`Voltage needed,
`electrical power
`共V, W min/ml兲a
`
`920
`
`Not reported
`
`1446
`
`400, 0.938
`
`3.5
`
`2.5
`
`Not reported
`
`Not reported
`
`2.14
`
`0.92
`
`100, 10.9
`
`3.4, 2961.5
`
`Not reported
`8
`Not reported
`10 000
`⬃207 共30 psi兲 Not reported
`⫺95 to 300
`1
`
`601.2
`191.0
`35 ml/h
`4260c
`
`6, 7
`6000, 514.7
`Standard power outlet
`No electrical power needed
`
`aValues presented are the largest reported with the pumping method.
`bFlow rate independent of cross section of channel 共up to specification limit due to resistance from smaller microfluidic
`channels兲. Maximum flow rate and pressure depend on syringe size. Number shown is for 60 cm3 syringe. Numbers are
`from New Era Pump Systems INC 共http://www.syringepump.com兲.
`cNumber shown is for a channel of 10 cm length. Flow rate varies with pressure and channel geometry by Eq. 共6兲.
`
`be translated to a pressure difference, which can be directly generated using the proposed pump
`关Fig. 2共b兲兴. An advantage of using constant pressures to generate desired flow rates is the absence
`of the pulsating flow phenomenon as in the use of syringe pumps10 since flow is generated due to
`a constant pressure rather than a frictional plunger movement driven by a stepper motor as in
`syringe pumps.
`Since the tubing and syringe in the proposed pump are not filled with the liquid being
`pumped, the system have zero dead volume, which is defined herein as the volume of liquid left
`over in the syringe and tubing. At the beginning of an experiment, the amount of liquid desired to
`be pumped is sucked into the free end of the tubing before it is connected to a microfluidic device.
`All liquid aspirated into the tubing can be pumped into the microfluidic device.
`Since liquid volumes pumped in microfluidic devices are typically of the order of microliters,
`the displacement of the liquid will not realistically change the volume of the air trapped inside the
`system; thus, the pressure applied will remain constant throughout the pumping process. None-
`theless, the pump can still deliver large liquid volumes for relevant applications by properly
`choosing tank volume and syringe size. For example, using a 500 ml tank 共e.g., a small water
`bottle兲 and a 60 ml syringe, liquid volumes as large as 5 ml can be discharged with a change in
`pressure less than 10%.
`The flow rate generated in microchannels using the new pump followed Eq. 共6兲 satisfactorily
`with a maximum error of 11% 关Fig. 2共b兲兴. This error in flow rate combines both the error in the
`generated pressure discussed above and the error in flow measurement, which was performed by
`tracing fluorescent beads flowing inside the channel. As expected, flow rate was steady throughout
`the pumping process and did not exhibit any pulsatile behavior.
`When compared to existing micropumping methods, our pump demonstrated competitive
`performance in terms of pressure range, resolution, and flow rate as summarized in Table I. Being
`electrical power free and being of low cost are additional advantages of the system.
`
`A. Micropipet aspiration
`
`Micropipet aspiration is a common technique used for mechanical characterization of single
`cells, which could be used as a biomarker for the onset of some diseases such as malaria, sickle
`
`7
`
`

`
`046501-8
`
`Moscovici et al.
`
`Biomicrofluidics 4, 046501 共2010兲
`
`FIG. 3. 共a兲 Video frames 共top to bottom兲 depicting micropipet aspiration and release of an MC3T3-E1 cell. 共b兲 Variation
`of the cell aspiration length inside the micropipette against applied vacuum pressure. Error bars are ⫾1 standard deviation
`共n=10 cells兲. 共c兲 Generating a concentration gradient using two food dyes 共red and green兲. 共i兲 A negative pressure of 15
`kPa was applied at the outlet 共marked with a dashed box兲. 共ii兲 Zoomed-in view of the output channel showing the gradient
`generated. 共d兲 Generation of water droplets in a continuous phase of silicone oil inside a cross-shaped microchannel. Values
`of applied pressure on each inlet branch are indicated on the three panels.
`
`cell anemia, and cancer.19 In micropipet aspiration, a micropipet 共few microns in diameter兲 is
`positioned adjacent to the cell, and a light vacuum pressure 共in the order of hundreds of pascals兲
`is applied such that the cell is aspirated into the micropipet. The elongation of the cell into the
`pipet due to the suction pressure is used, along with the pressure applied, to determine the Young’s
`modulus of the cell.19,24
`We used the proposed pump to aspirate osteoblasts 共MC3T3-E1 cells兲 into micropipets to
`calculate its Young’s modulus 关Fig. 3共a兲兴. We used tank and syringe volumes of 120 and 1 ml,
`respectively, which allowed us to produce a pressure range of 730 Pa with a resolution of 8 Pa.
`This high resolution permitted fine measurement of the aspiration length under different pressures
`without lysing the cells 关Fig. 3共b兲兴. The relationship between aspiration length and applied vacuum
`pressure was linear, as predicted by the half-space model.18 The Young’s Modulus of the osteoblast
`cells was calculated to be 555⫾183 Pa 共n=10 cells兲, which is in agreement with previously
`reported values.25
`
`B. Generation of a concentration gradient
`
`Generating concentration gradients is important for studying many phenomena such as
`chemotaxis in cell biology and nucleation and growth of crystals in surface chemistry.20 The
`dominance of laminar flow in microfluidic devices made it relatively easy to generate concentra-
`tion gradients of different shape.26 We used the new pump to generate a concentration gradient
`inside a 1300 ␮m wide channel 关Fig. 3共c兲兴. Instead of applying two equal pressures on the two
`
`8
`
`

`
`046501-9
`
`Simple and versatile microfluidic pump
`
`Biomicrofluidics 4, 046501 共2010兲
`
`reservoirs holding the dyes, we applied a vacuum of 15 kPa 共total volume=10 ml and syringe
`displacement=2 ml兲 at the device outlet, which generated equal flow rates for both dyes since the
`hydrodynamic resistances of the device two branches were the same due to device symmetry.27
`Applying vacuum to induce the required flow has the advantage of using one pump to deliver
`different reagents instead of dedicating one pump for each reagent.
`
`C. Droplet generation
`
`Multiphase flow in microchannels offers a new set of advantages besides those seen in single
`phase flow, such as faster mixing, enhanced heat and mass transfer, and reduced dispersion.28
`Additionally, droplet generation in microfluidic devices has been used for single cell analysis,4,29
`electrophoretic separation,30 and DNA analysis.31 To further demonstrate the capabilities of the
`pump, we used it to generate water droplets in a continuous phase of silicone oil through hydro-
`dynamic focusing of a stream of water in a cross-shaped microchannel 关Fig. 3共d兲兴. Droplet size in
`our experiments was easily controlled by changing the pressure applied on 共and hence the flow
`rate of兲 the oil phase as seen in panels ii and iii of Fig. 3共d兲. Using the same hydrodynamic
`focusing device, we were able to focus a stream of a miscible dye 共supplementary material, Fig.
`S2兲,22 which is useful in various microfluidic applications such as selective delivery of small
`molecules into biological cells6 and low-voltage electroporation of cells.32
`
`IV. CONCLUSION
`
`The pressure-driven pumping system presented in this paper does not require electrical power.
`It costs less than $50 without a pressure sensor or $150 with a pressure sensor and can be built
`within a few minutes from off-the-shelf components. The pump is capable of accurately generat-
`ing a wide range of positive and negative pressures with high resolutions for microfluidic appli-
`cations. The use of constant pressures to generate flows in microfluidic devices, instead of dis-
`placement pumps 共e.g., syringe pumps兲, has the advantages of zero dead volumes, smooth
`continuous flow rates, and being compatible with pressure-sensitive applications such as micropi-
`pet aspiration and mechanical stimulation of cells. The capacity of the pump was demonstrated by
`performing micropipet aspiration of osteoblasts to determine their Young’s modulus values, real-
`izing a concentration gradient, and generating microdroplets of water in a continuous oil phase.
`
`ACKNOWLEDGMENTS
`
`We thank Jason Li for culturing the MC3T3-E1 cells, Di Xue for helping with flow measure-
`ment experiments, and Chris Moraes for supplying the fluorescence beads. We also thank Chris
`Moraes, Professor Axel Gunther, and Dr. Edmond Young 共from Professor David Beebe’s labora-
`tory at the University of Wisconsin at Madison兲 for helpful discussions.
`
`1 E. W. K. Young and D. J. Beebe, Chem. Soc. Rev. 39, 1036 共2010兲.
`2 D. Wu, J. Qin, and B. Lin,J. Chromatogr. A 1184, 542 共2008兲.
`3 A. T. Woolley, D. Hadley, P. Landre, A. J. DeMello, R. A. Mathies, and M. A. Northrup, Anal. Chem. 68, 4081 共1996兲.
`4 J. F. Edd, D. Di Carlo, K. J. Humphry, S. Köster, D. Irimia, D. A. Weitz, and M. Toner, Lab Chip 8, 1262 共2008兲.
`5 S. Nagrath, L. V. Sequist, S. Maheswaran, D. W. Bell, D. Irimia, L. Ulkus, M. R. Smith, E. L. Kwak, S. Digumarthy, A.
`Muzikansky, P. Ryan, U. J. Balis, R. G. Tompkins, D. A. Haber, and M. Toner, Nature 共London兲 450, 1235共2007兲.
`6 F. Wang, H. Wang, J. Wang, H. Y. Wang, P. L. Rummel, S. V. Garimella, and C. Lu, Biotechnol. Bioeng. 100, 150
`共2008兲.
`7 S. A. Khan, A. Gunther, M. A. Schmidt, and K. F. Jensen, Langmuir 20, 8604 共2004兲.
`8 B. D. Iverson and S. V. Garimella, Microfluid. Nanofluid. 5, 145 共2008兲.
`9 S. Arulanandam and D. Li, Colloids Surf., A 161, 89共2000兲.
`10 R. Yokokawa, T. Saika, T. Nakayama, H. Fujita, and S. Konishi, Lab Chip 6, 1062 共2006兲.
`11 D. Dendukuri, S. S. Gu, D. C. Pregibon, T. A. Hatton, and P. S. Doyle, Lab Chip 7, 818 共2007兲.
`12 H. V. Fuentes and A. T. Woolley, Lab Chip 7, 1524 共2007兲.
`13 J. W. Munyan, H. V. Fuentes, M. Draper, R. T. Kelly, and A. T. Woolley, Lab Chip 3, 217 共2003兲.
`14 C. C. Hong, J. W. Choi, and C. H. Ahn, J. Micromech. Microeng. 17, 410 共2007兲.
`15 G. M. Walker and D. J. Beebe, Lab Chip 2, 131共2002兲.
`16 A. M. Skelley and J. Voldman, Lab Chip 8, 1733 共2008兲.
`17 R. K. Shah and A. L. London, Laminar Flow Forced Convection in Ducts; A Source Book for Compact Heat Exchanger
`Analytical Data 共Academic, New York, 1978兲.
`
`9
`
`

`
`046501-10 Moscovici et al.
`
`Biomicrofluidics 4, 046501 共2010兲
`
`18 D. P. Theret, M. J. Levesque, M. Sato, R. M. Nerem, and L. T. Wheeler, J. Biomech. Eng. 110, 190 共1988兲.
`19 C. T. Lim, E. H. Zhou, and S. T. Quek, J. Biomech. 39, 195 共2006兲.
`20 N. L. Jeon, S. K. W. Dertinger, D. T. Chiu, I. S. Choi, A. D. Stroock, and G. M. Whitesides, Langmuir 16, 8311 共2000兲.
`21 H. Kim, D. Luo, D. Link, D. A. Weitz, M. Marquez, and Z. Cheng, Appl. Phys. Lett. 91, 133106 共2007兲.
`22 See supplementary material at http://dx.doi.org/10.1063/1.3499939 for measurments of the pump hysteresis and results
`of a hydrodynamic focusing experiment using the proposed pump.
`23 B. R. Munson, D. F. Young, and T. H. Okiishi, Fundamentals of Fluid Mechanics, 4th ed. 共Wiley, New York, 2002兲.
`24 R. M. Hochmuth, J. Biomech. 33, 15共2000兲.
`25 E. Takai, K. D. Costa, A. Shaheen, C. T. Hung, and X. E. Guo, Ann. Biomed. Eng. 33, 963 共2005兲.
`26 S. K. W. Dertinger, D. T. Chiu, J. Noo Li, and G. M. Whitesides, Anal. Chem. 73, 1240 共2001兲.
`27 T. Stiles, R. Fallon, T. Vestad, J. Oakey, D. W. M. Marr, J. Squier, and R. Jimenez, Microfluid. Nanofluid. 1, 280 共2005兲.
`28 A. Günther and K. F. Jensen, Lab Chip 6, 1487 共2006兲.
`29 M. Y. He, J. S. Edgar, G. D. M. Jeffries, R. M. Lorenz, J. P. Shelby, and D. T. Chiu, Anal. Chem. 77, 1539 共2005兲.
`30 J. S. Edgar, C. P. Pabbati, R. M. Lorenz, M. He, G. S. Fiorini, and D. T. Chiu, Anal. Chem. 78, 6948 共2006兲.
`31 M. A. Burns, B. N. Johnson, S. N. Brahmasandra, K. Handique, J. R. Webster, M. Krishnan, T. S. Sammarco, P. M. Man,
`D. Jones, D. Heldsinger, C. H. Mastrangelo, and D. T. Burke, Science 282, 484共1998兲.
`32 T. Zhu, C. Luo, J. Huang, C. Xiong, Q. Ouyang, and J. Fang, Biomed. Microdevices 12, 35共2010兲.
`33 D. J. Laser and J. G. Santiago, J. Micromech. Microeng. 14, R35共2004兲.
`
`10