<%BANNER%>

Evaluation of a Microflowmeter and Development of a Microfluidic Valve

Permanent Link: http://ufdc.ufl.edu/UFE0024104/00001

Material Information

Title: Evaluation of a Microflowmeter and Development of a Microfluidic Valve
Physical Description: 1 online resource (75 p.)
Language: english
Creator: Pitchaimani, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Microelectronics was the technological revolution of the 20th century highlighting the power of miniaturization. Extension of the same into chemistry and life sciences is projected to be one of the next most significant scientific developments. Microfluidics implements this principle by reducing the size of a lab to that of a chip, aptly called 'Lab on a chip' (LOC). This downsizing has several predicted advantages including lower reagent consumption, faster analysis, higher separation resolution and detection sensitivity, and lower detection limit. For nearly the past 20 years many researchers have developed conceptual LOC devices to address various applications including DNA amplification in a chip. Several constituent elements are required to make a LOC system. Valves are one of the critical components which help control fluid flows in a microdevice. They are used for fluid containment, isolation and directing flows. Conventional MEMS (Micro-electro-mechanical system) based valves are made mostly on silicon whose integration with LOC tend to be problematic. Primary problems include optical clarity, difference in operational conditions, and suitability to gas rather than liquids. As a solution, a thermally actuated microfluidic valve is developed in this work for plastic microfluidic devices. The principle of operation is based on a thermally sensitive fluid, which expands to actuate a membrane that controls the fluid flow. The power input for thermal expansion is controlled using a printed circuit board (PCB) which is suited to mass production and large scale integration. The microfluidic valve was fabricated and its operation has been demonstrated. Its operating parameters including actuation times, range of operational power and repetitive operation have been studied. The temperature during operation was also measured. Two methods were used to test the valve operation. First method involves a flowmeter which was used to measure flow downstream. The flowmeter was evaluated for general microfluidic applications for this purpose. In the evaluation, the flowmeter was calibrated using direct method where the amount of water collected over a known period was weighed. It was also used to study the effect of stiffness of the tubing connecting the pump to the microfluidic device. Stiff tubing was found to maintain accurate flow rate. Existence of pulsations in flow was confirmed due to the use of syringe pump. Pulsations were dependent on tubing stiffness, presence of microfluidic device, flow rate and syringe size. The second method was based on measurement of the electric-current in a channel and it was used to study the local valve action. Fluorescence was also used to qualitatively show valve action. The valve was found to operate over a wide range of power from 36 mW to 80 mW with valve close and open response times in seconds. The local channel temperature during valve operation was found to be less that 45 degree centigrade. The valve can operate as many as 50 times when the hydrostatic head was 9.5 mm. The valve was reliable when the flow rate was as high as 300 nL/min (1.2 mm/s).
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Pitchaimani.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Fan, Zhonghui H.
Local: Co-adviser: Nishida, Toshikazu.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0024104:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024104/00001

Material Information

Title: Evaluation of a Microflowmeter and Development of a Microfluidic Valve
Physical Description: 1 online resource (75 p.)
Language: english
Creator: Pitchaimani, Karthik
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Microelectronics was the technological revolution of the 20th century highlighting the power of miniaturization. Extension of the same into chemistry and life sciences is projected to be one of the next most significant scientific developments. Microfluidics implements this principle by reducing the size of a lab to that of a chip, aptly called 'Lab on a chip' (LOC). This downsizing has several predicted advantages including lower reagent consumption, faster analysis, higher separation resolution and detection sensitivity, and lower detection limit. For nearly the past 20 years many researchers have developed conceptual LOC devices to address various applications including DNA amplification in a chip. Several constituent elements are required to make a LOC system. Valves are one of the critical components which help control fluid flows in a microdevice. They are used for fluid containment, isolation and directing flows. Conventional MEMS (Micro-electro-mechanical system) based valves are made mostly on silicon whose integration with LOC tend to be problematic. Primary problems include optical clarity, difference in operational conditions, and suitability to gas rather than liquids. As a solution, a thermally actuated microfluidic valve is developed in this work for plastic microfluidic devices. The principle of operation is based on a thermally sensitive fluid, which expands to actuate a membrane that controls the fluid flow. The power input for thermal expansion is controlled using a printed circuit board (PCB) which is suited to mass production and large scale integration. The microfluidic valve was fabricated and its operation has been demonstrated. Its operating parameters including actuation times, range of operational power and repetitive operation have been studied. The temperature during operation was also measured. Two methods were used to test the valve operation. First method involves a flowmeter which was used to measure flow downstream. The flowmeter was evaluated for general microfluidic applications for this purpose. In the evaluation, the flowmeter was calibrated using direct method where the amount of water collected over a known period was weighed. It was also used to study the effect of stiffness of the tubing connecting the pump to the microfluidic device. Stiff tubing was found to maintain accurate flow rate. Existence of pulsations in flow was confirmed due to the use of syringe pump. Pulsations were dependent on tubing stiffness, presence of microfluidic device, flow rate and syringe size. The second method was based on measurement of the electric-current in a channel and it was used to study the local valve action. Fluorescence was also used to qualitatively show valve action. The valve was found to operate over a wide range of power from 36 mW to 80 mW with valve close and open response times in seconds. The local channel temperature during valve operation was found to be less that 45 degree centigrade. The valve can operate as many as 50 times when the hydrostatic head was 9.5 mm. The valve was reliable when the flow rate was as high as 300 nL/min (1.2 mm/s).
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Karthik Pitchaimani.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Fan, Zhonghui H.
Local: Co-adviser: Nishida, Toshikazu.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0024104:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVALUATION OF A MICROFLOWMETER AND DEVELOPMENT OF A MICROFLUIDIC VALVE By KARTHIK PITCHAIMANI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Karthik Pitchaimani

PAGE 3

3 To my grandparents

PAGE 4

4 ACKNOWLEDGMENTS I would like to acknowledge Dr. Fan for guiding me through my curriculum. His training in research and presentation skills has been of immense value to me. I would like to thank Dr. Nishida for his regular feedback and ev aluation which helped me toward the right direction. I thank D r. Schmitz for serving on my supervisory committee. I would like to thank Brian Sapp for heater fabrication, Adam Winter and Joey Wilson for making the printed circuit board, and Daniel Olivero and Austin Gispanski for help in fabricating devices and topas film device picture For general thought provoking, brainstorming sessions and discussions I acknowledge Dr. Cham pak Das, Ruba Khnouf, Zheng Xia and Dr. Hong Chen. I gratefully acknowledge the support and cooperation from all past and present group member s in Dr.Fans Lab. Finally, I acknowledge the greatest support I received from my family They motivated me toward my goal to strive and achieve excellence.

PAGE 5

5 TABLE OF CONTENTS ACKNOWLEDGMENTS ........................................................................................................... 4 page LIST OF TABLES ...................................................................................................................... 7 LIST OF FIGURES .................................................................................................................... 8 ABSTRACT ............................................................................................................................... 9 CHAPTER 1 INTRODUCTION ............................................................................................................. 11 1.1 Microfluidics ............................................................................................................ 11 1.2 Microvalves .............................................................................................................. 12 1.3 Study Overview ........................................................................................................ 14 2 EVALUATION OF A MICROFLOWMETER AND ITS USE IN TESTING A MICROVALVE ................................................................................................................. 16 2.1 Introduction .............................................................................................................. 16 2.2 Experimental Section ................................................................................................ 17 2.2.1 Device Fabrication ........................................................................................ 17 2.2.2 Flow Rate Measurement ................................................................................ 18 2.3 Results and Discussion .............................................................................................. 19 2.3.1 Flow Meter Calibration ................................................................................. 19 2.3.2 Flow Measurement in Devices ....................................................................... 20 2.3.3 Flow Pulsations ............................................................................................. 21 2.3.4 Microvalve Testing Using Flowmeter ............................................................ 23 2.4 Conclusion ................................................................................................................ 25 3 FABRICATION AND CHARACTERIZATION OF A THERMALLY ACTUATED MICROFLUIDIC VALVE ................................................................................................. 34 3.1 Introduction .............................................................................................................. 34 3.2 Brief Review of Microvalves .................................................................................... 34 3.3 Principle of Operation ............................................................................................... 36 3.4 Experimental Section ................................................................................................ 36 3.4.1 Device Fabrication for Microvalve ................................................................ 36 3.4.2 Elastomeric Film ........................................................................................... 37 3.4.3 Polyester (PET) Film ..................................................................................... 37 3.4. 4 Microheaters ................................................................................................. 39 3.4.5 Thermal Fluid ................................................................................................ 40

PAGE 6

6 3.4.6 Valve Characterization .................................................................................. 41 3.4.6.1 Electriccurrent method ................................................................... 41 3.4.6.2 Temperature measurement ............................................................... 42 3.4.6.3 Fluorescence measurement .............................................................. 43 3.5 Results and Discussion .............................................................................................. 43 3.5. 1 Device Fabrication ........................................................................................ 43 3.5.1.1 Heater fabrication and integration .................................................... 44 3.5.1.2 Fluorinert encapsulation .................................................................. 44 3.5.2 Valve Characterization .................................................................................. 45 3.5.2.1 Cyclic valve operation ..................................................................... 45 3.5.2.2 Actuation time ................................................................................. 46 3.5.2.3 Operating temperature ..................................................................... 46 3. 5.2.4 Valve operation with liquid pressure: pumping and hydrostatic head ................................................................................................. 48 3.5.2.5 Fluorescence measurement .............................................................. 50 3.6 Conclusion ................................................................................................................ 51 4 CONCLUSIONS AND FUTURE WORK .......................................................................... 67 4.1 Microvalve Fabrication ............................................................................................. 67 4.2 Microvalve Operation ............................................................................................... 67 4.2.1 Evaluation Using a Flow Meter ..................................................................... 67 4.2.2 Electric Current Measurement ....................................................................... 68 4.3 Future Work .............................................................................................................. 69 LIST OF REFERENCES .......................................................................................................... 70 BIOGRAPHICAL S KETCH ..................................................................................................... 75

PAGE 7

7 LIST OF TABLES Table page 11 Comparison of different micromachined valve actuation mechanisms ............................ 15 21 The volume of syringes and parameters of syringe pump ............................................... 27 31 Elastomer film properties ............................................................................................... 53 32 Hydrostatic heads and flow rates .................................................................................... 53

PAGE 8

8 LIST OF FIGURES Figure page 21 E xperimental setup. ....................................................................................................... 28 22 Comparison of the flow rates measured by the direct met hod and the Sensirion flowmeter ...................................................................................................................... 28 23 Comparison of the flow rates measured by the Sensirion flowmeter in the presence and absence of a microfluidic device. ............................................................................ 29 24 Comparison of the flow rates measured by the Sensirion flowmeter in two setups using stiff or flexible tubing.. ......................................................................................... 30 25 Temporal profiles of flow rates measured by the Sensirion flowmeter in different setups.. .......................................................................................................................... 31 26 M icrofluidic valve.. ....................................................................................................... 32 27 Microvalve actuation data recorded by flowmeter showing the local flow dynamics. ..... 33 31 M icrofluidic valve concept based on thermal actuation principle.. ................................. 54 32 Multilayer microvalve device. ....................................................................................... 54 33 Electriccurrent method to study valve actuation.. .......................................................... 55 34 Labview GUI for data acquisition during valve testing. .................................................. 56 35 Valve actuation when 58 mW was applied to Au heaters. .............................................. 58 36 Valve actuation time as a function of the input power.. .................................................. 59 37 Temperature at different locations during valve actuation at 58 mW input heater power. ........................................................................................................................... 60 38 Temperature at different locations of a microvalve as a function of power. .................... 62 39 Valve actuation for a 300 nL/min flow with different heater input powers.. ................... 63 310 Electric current as a function of time when different hydrostatic heads were used.. ........ 64 311 Fluorescent images showing valve action. ...................................................................... 65 312 PCB connected to a microvalve device. ......................................................................... 66

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVALU ATION OF A MICROFLOWMETER AND DEVELOPMENT OF A MICROFLUIDIC VALVE By Karthik Pitchaimani December 2008 Chair: Hugh Fan Cochair: Toshikazu Nishida Major: Mechanical Engineering Microelectronics was the technological revolution of the 20th century highlig hting the power of miniaturization. Extension of the same into chemistry and life sciences is projected to be one of the next most significant scientific developments Microfluidics implements this principle by reducing the size of a lab to that of a chip, aptly called Lab on a chip (LOC). This downsizing has several predicted advantages including lower reagent consumption, faster analysis, higher separation resolution and detection sensitivity and lower detection limit. For nearly the past 20 years many researchers have developed conceptual LOC devices to address various applications including DNA amplification in a chip. Several constituent elements are required to make a LOC system. Valves are one of the critical components which help control fluid flow s in a micro device. They are used for f luid containment, isolation and directing flow s Conventional MEMS (Micro electro mechanical system) based valves are made mostly on s ilicon whose integration with LOC tend to be problematic Primar y problems include optical clarity, difference in operational conditions and suitability to gas rather than liquids

PAGE 10

10 As a solution, a thermally actuated microfluidic valve is developed in this work for plastic microfluidic devices The principle o f operation is based on a thermally sensitive fluid which expands to actuate a membrane that controls the fluid flow. The power input for thermal expansion is controlled using a printed circuit board ( PCB) which is suited to mass production and large scale integration. The micr ofluidic valve was fabricated and its operation has been demonstrated. Its operating parameters including actuation times, range of operational power and repetitive operation have been studied Th e temperature during operation wa s also measured. Two method s we re used to test the valve operation. First method involves a flowmeter which was used to me asure flow downstream. The flow meter was evaluated for general microfluidic applications for this purpose. In the evaluation, the flowmeter was calibrated using direct method where the amount of water collected over a known period was weighed It was also used to study the effect of stiffness of the tubing connecting the pump to the microfluidic device. Stiff tubing was found to maintain accurate flow rate Existe nce of p ulsations in flow was confirmed due to the use of syringe pump. Pulsations were dependent on tubing stiffness, presence of microfluidic device, flow rate and syringe size. The second method was based on measurement of the electriccurrent in a channel and it was used to study the lo cal valve action. Fluorescence wa s also used to qualitatively show valve action. The valve was found to operate over a wide range of power from 36 mW to 80 mW with valve close and open response times in seconds. The local channel temperature during valve operati on was found to be less that 45 degree centigrade The valve can operate as many as 50 times when the hydrostatic head was 9.5 mm. The valve was reliab le when the flow rate was as high as 300 nL/min (1.2 mm /s).

PAGE 11

11 CHAPTER 1 INTRODUCTION Technology has always played a critical role in changing the way of our lives. Invention of the wheel is an example of the earliest development to have made a considerable impact on the society. In the recent times, one such revolutionary technology is the science of microelectronics and microfabrication. Starting from the thought provoking speech by Richard Feynman1 and the invention of Integrated circuits2 this technology has come a long way in influencing our day to day activities. In the latter part of the 20th century microfabrication was adopted towards modeling of miniature phy sical systems (transducers) in the form of Micro electro mechanical systems (MEMS) .3, 4 The extension of MEMS and application of microfabrication to chemistry and life sciences was realized later on Notably the seminal paper by Manz5 brought the new application field into notice. The concept of miniaturizing l aboratory instruments has heralded a new stream of research with considerable promise. This science was termed as the mi niaturized total analyses syst em (TAS) where the concept of l ab on a chip was initiated .5, 6 1.1 Microfluidics Microfluidics is the science of manipulating fluids in micron scale .6 It is the technology behind realizing a TAS where fluid flows in microchannels are needed. The predicted advantages of miniaturization propel increasing number of researchers to work in this field .7 Some advantages are apparent such as point of care applications, clinical diagnostics, remot e monitoring and analysis due to its portability. Other advantages are owing to scaling which include reduction in reagent volume, shorter reaction times for biological and chemical applications resulting in smaller time constants, higher detection sensiti vity, lower detection limits and lower costs with mass production .5 7

PAGE 12

12 Several publication s demonstrate the feasibility of this principle8 11 in an atte mpt to fabricate a complete lab on a chip (LOC) incorporating sample introduction, pretreatment, analysis (reaction or separation) and finally detection. In addition, the high t hroughput predicted for LOC application have also been demonstrated in the form of l arge scale integrated microfluidic platforms .12 Pres ent progress in attaining a reliable TAS are reviewed in some reports1315 highlighting the achievements and further improvements needed to realize the potential promised by the technology. Initially, silicon was the primary substrate material used in the microfabricated analysis systems owing to the existing developed silicon fabrication technologies. However lack of optical transparency and the issues involved in fabrication made researchers pursue glass and recently polymers16, 17 such as polydimethylsiloxane (PDMS) ,1823 p olycarbonate ,24 polymethymethacrylate,25, 26 and c yclo olefin c opolymer27 as the alternative materials. Plastics offer the most lucrative prop erties in terms of ease of fabrication, lower cost for mass production and optical clarity. There are several components that are required to make a successful LOC platform. Valves, pumps, mixers are some of the most important ones to name. These component s are the building blocks of a LOC device needed to manipulate fluid samples. One of the main reasons why TAS has not been widely commercially implemented is the lack of reliable components which has proved to be a primary hindrance in realizing the pote ntial of LOC systems .28 This thesis will focus on the development of a novel microfluidic valve to address the needs of present LOC systems. 1.2 Microvalves Several attemp ts have been made to fabricate reliable microfluidic valve s ; however, there is still room for improvement in terms of performance. Conventional MEMS based microvalves

PAGE 13

13 have been developed as early as the soleno id valve fabricated by Terry et al.3 for his gas chromatographic air analyzer. The valve was based on a solenoid plunger actuating a nickel diaphragm that con trols the flow in conjunction with the valve seat. In general, majority of the valves were fabricated i n silicon substrate using the micromachining technolo gy. Their target applications were primarily MEMS based device integration. However, with the introd uction of TAS these principles were adapted to microfluidic platforms and several newer materials were also being increasingly used as a substrate for microvalve. Microvalves are primarily classified into active and passive depending upon the operational mechanism. Active microvalves need external stimulus in the form of electromagnetic,3 electrostatic ,29 piezoelectric,30, 31 pneumatic,19, 32 thermopneumatic,3339 and shape memory alloy40 (SMA) etc. In addition, there are several phase change microvalves making use of hydrogel ,41, 42 paraffin ,43, 44 polyethylene glycol.45 Passive microvalves make use of inherent characteristics of the system and are based on surface effects such as capillarity .46, 47 Each of these has its own advantages and limitations. A thorough review of the microvalves can be found in t he recent publication by Ahn et al.28. The traditional text by Shoji48 and Kovacs49 also summarizes these data. Table 1 1 compares the characteristics of these principles. The fastest actuation times are achievable through electrostatic microvalves H owever their low stroke and pressure coupled with very high voltage requirements limit their application. Piezoelectric actuators are fast with large force but prod uce very small strokes. The need of high voltage also restricts its use. A lthough pneumatic microvalves produce large stroke and pressure they are slow in response. Additionally, they need bulky accessories that affect device portability. One of the most w idely used valve actuation mechanism is thermal. At the small scale, this method has been attractive due to low thermal mass.49 It generates large pressure and

PAGE 14

14 deflection with moderatel y fast actuation times. H owever, it needs proper temperature control so that it will not affect the sample being analyzed. 1.3 Study Overview In order to ach ieve the true potential of a labon a chip in terms of high throughput a microvalve needs to be fabricated which is compatible with large scale integration. Adaptability to industrial fabrication is another critical factor. Keeping these applications in mind along with the various ideal characteristics of a microvalve49 a solution has been proposed and demonstrated in th is thesis. A thermally actuated microfluidic valve has been developed and it can be controlled with a printed circuit board (PCB). Supplying heat to thermally sensitive fluid results in its expansion This increase in volume has been used to deflect an elastomeric film into the microfluidic channel. Thermal expansion of the fluid was at a temperature well below the boiling point thus there is no phase change. This avoids excessive heating of samples. The thermal ly sensitive fluid was heated using a resistiv e heater fabricated using standard lithographic patterning procedures. The power supply to the heater can be controlled using a PCB making it adaptable to industrial applications A thermal ly actuated valve has been fabricated and the concept is demonstra ted in this work

PAGE 15

15 Table 1 1. Comparison of different micromachined valve actuation mechanisms28, 48 Electromagnetic Piezoelectric Pneumatic Electrostatic Thermopneumatic Stroke Large Small Large Small Large Response time Fast Fast Slow Very fast Medium Pressure or F orce Small Large Large Small Large Remarks Fabrication complexity Large voltage Bulky accessories Extremely high voltage Heat input

PAGE 16

16 CHAPTER 2 EVALUATION OF A MICROFLOWMETER AND I TS USE IN TESTING A MICROVALVE 2.1 Introduction Flow control and metering of liquid s in nanolit ers are often required in microfluidic devices. For devices using electroosmotic pumping, flow rates can be calculated from the applied electric field and osmotic mobility.8 For device s employing pneumatic pumps, flow rates can be estimated by the pumping rate.50 However, th ese methods do not enable us to detect flow variations in microchannel for example, due to the change in the surface charges in an electro osmotic flow or spatial restrictions in a pressure driven flow. Other possibilities that lead to f low rate changes include head loss51 due to the dead volume and geometry mismatch in macro to micro interfaces,52 flow variation due to the stretching of flexible tubing that connects pumps with devices, flow resistance caused by packed columns,50 and blockage resulting from microfluidic valve s.19, 53 For some applications, it is not critical to know exact flow rates; an approximate value from calculation or estimation suffices. However, knowing accurate flow rates is a necessity for other applications. F or instance the mixing ratio of a sample and a buf fer solution must be precisely measured and controlled so that the actual sample concentration analyzed in the device is accurate for quantification.9 An unaccounted change in the flow rate could lead to inaccurate reporting and documentatio n of the sample. Also when a microfluidic valve is present in a fluid network, flowmeters are required to monitor local flow rate s to detect if the flow is truly blocked without leakage.19, 53 The physical principles that can be used in flow meters have been reviewed by Nguyen.54 Many efforts described in this review are based on silicon based traditional MEMS (Micro electro mechanical systems) and developed primarily for gases. A few recent efforts have

PAGE 17

17 focused on mic roscale liquid flow measurement, including t hermal flow sens ors,53, 55 electro chemical velocimetry ,56 electronic current monitoring,57, 58 particle image velocimetry (PIV) ,59, 60 and l aser induced fluorescence photo bleaching anemometer.61 However, all of them either have difficulty in measuring instantaneous velocity or require specialized hardware that is not readily commercially available .61 One commercially availab le flowmeter that was developed for microfluidic applications is SLG 1430 from Sensirion AG ( Zurich Switzerland).62, 63 Two models (SLG 1430025 and SLG 1430150) are compa tible w ith most microfluidic devices. The operational flow rate for SLG 1430025 ranges from 50 nL/min to 1500 nL/min while that for SLG 1430 150 is 250 nL/min to 7000 nL/min a ccording to the manufacturer. The flowmeter is based on t hermal sensing and the sensors are fabricated using standard integrated circuit fabrication techn iques.62, 63 Since the flowmeter has not been extensively exploited in microfluidic research, SLG 1 430150 flowmeter was evaluated in this study and employed to measure and analyze the flow rate variations due to the presence of a microfluidic valve in a flow circuit. In the evaluation process, the effect of the stiffness of tubing used to connect a syr inge pump with a microfluidic device on the flow rate in the device was studied. Also, the flowmeter was used to confirm the existence of flow pulsation resulting from a syringe pump as suspected.64, 65 It was found that the degree of flow pulsation was reduced when flexible tu bing was used and/or when a microfluidic device was connected in the fluid network. It was also dependent on the size of syringe used in the pump. 2.2 Experimental Section 2.2.1 Device Fabrication Plastic microfluidic devices w ere fabricated using the procedure described previously.27 Briefly, a photomask pattern was designed using AutoCAD and then reproduced in glass using photolithography. By e lectroplating on the glass plate a nickel alloy mo ld was generated and

PAGE 18

18 then used to produc e plastic parts from cyclo olefin resins (Zeonor 1020R) using a Carver hydraulic press (Wabash, IN). A CNC milling machine was employ ed to trim the plastic parts into 1 x 3 in. substrate s and drill 2 mmdiameter holes at the channel ends. The substrate was then sealed with a 250m thick film (Zeonor 1020R ) from PLITEK ( Des Plaines, IN) using a Catena thermal laminator (Northbrook, IL). The holes at channel ends became wells for fluid connections. A picture of the device is shown in Figure 2 1 The device is 1.5 mm thick while all channel s in the microfluidic device are 74 m wide and 27 m deep The cross section of the channels is D shaped due to isotropic etching of the glass substrate that was used for fabricating the molding master.27 The length of the channel used in this work is 26 mm. 2.2.2 Flow R ate Measurement The e xperimental setup in Figure 2 1 was used for flow rate measurement. A UMP II syringe pump from World Precision Instruments, Inc. (Sarasota, FL.) was controlled by a Micro4 controller It was used with a 100l gastight Hamilton 1710LT syringe (Reno, Nevada) for the reagent delivery in all experiments except for one, in which a 250 L liquidtight SGE syringe was used. The pump controller was set to syringe type G a nd H for the 100l and 250l syringes, respectively, according to the manufacturers instructions. The syringe was connected to one well (t he inlet ) of the microfluidic device using a tub e (#1) and Upchurch Nanoport fittings ( Oak Harbour, WA) The outl et of the channel was connected via another tube (#2) to a Sensirion SLG1430150 flow meter (Zurich, Switzerland) The fluid out of the flowmeter was collected through a tube into a microcentrifuge tube. Two types of tubes with different stiffness were used for tube 1 One was an Upchurch PEEK tube with 100 m ID and 1.57 mm OD that will be referred to as a stiff tube in the text. The other is a Corning silicone tub e with 300 m ID and 0.64 mm OD referred to as a flexible tube in the text. The stiff tub e was expected to maintain the pumping pressure generated from

PAGE 19

19 the syringe pump while the flexible tube diminished the pumping pressure to a certain degree Tube 2 is an Upchurch PEEK capillary tub e with 100 m ID and 360 m OD. A stiff tube was always us ed for tube 2 so that the flow rate measured by the flowmeter truly reflects the value in the device. Purified water was used as the reagent except where specified otherwise. It was purified by a Barnstead Nanopure water system (Dubuque, Iowa). The flow r ate was set by the syringe pump ranging from 250 nL/min to 7000 nL/min which corresponds to 2.5 to 70 mm/s calculated using the cross section area of the channels in the device. It was then measured by the Sensirion flow meter and the data was acquired by a computer and graphical user interface (GUI) provided by the flowme ter manufacturer. The data readout frequency can be se lected from 1.56 Hz to 200 Hz. A high frequency (200 Hz) was used to collect the initial data for display ing fast changing flows whil e a low frequency (3.12 Hz) was employed to study microvalve actuat ion. The flow rate was further calibrated using a direct method,51 which was to measure the amount of water that accumulated in a container over a fixed time period. The collected water was weighed by a Fisher accu 124 analytical balance (Pittsburgh, PA) with an accuracy of 0.1 mg T he volumetric flow rate was calculated by the weight, time, and water den s ity at the ambient temperature. To establish a baseline, flow measurement was first carried out without a device. The syringe pump was directly connected to the flow meter and the flow rate was calibrated using the direct method 2.3 Results and Discussion 2.3.1 Flow M eter Calibration The operation range of the Sensirion flow meter is 250 nL/min to 7000 nL/min according to the manufacturer. As a result, the flow meter was calibrated over the same range using the direct method as described in the Exp erimental Section A water flow was generated by the syringe

PAGE 20

20 pump and the flow rate was measured by the flowmeter and the direct method ( weight measurement ) No microfluidic device was connected in the setup and stiff tubing was used for connections. Figur e 2 2 shows the obtained flow rate as a function of the pump ing rate of the syringe pump. The measured flow rate by direct method agreed very well with the pumping rate; t he slope of the calibration curve is 1.01. The measured flow rate by the Sensirion fl ow meter agreed well with the pumping rate at the lower range (t he slope of the calibration curve is 0.97 for values below 5 000 nL/min), but it started to deviate from the linear curve when the flow rate was larger than 5 000 nL/min The error bars indicate the standard deviation calculated from all data points collected over 1 5 minutes for each flow rate. The variation in the flowmeter measurement likely resulted from the flow pulsation as discussed later. 2.3.2 Flow Measurement in Devices The effect of p resence of a microfluidic device on the flow rate delivered by a syringe pump was first evaluated. The flow rate was measured using the flowmeter when it was directly connected to the syringe pump as in calibration experiments; the same measurements were t hen taken when the flowmeter was connected via microfluidic device to the syringe p ump as in Figure 21. Stiff tubing was used in both experiments. Figure 2 3 show s a comparison in the measured flow rates between these two setups The error bars in the pl ots are again the s tandard deviations. The results suggest that two groups of flow rates are the same within the experimental errors for these two setups in the presence or in the absence of the device, suggest ing that a microfluidic device has negligible effects on the flow rate delivered by the syringe pump. The effect of syringe pumping on flow rate using flexible tubing was next investigated and compared with the case for stiff tubing since stiff tube was expected t o maintain pumping pressure. The stiff PEEK tubing that connected the syringe p ump and the device in Figure 21

PAGE 21

21 was replaced with a flexible silicone tub e. Figure 2 4 show s a comparison in the measured flow rates between a setup with stiff tubing and the s ame setup with flexible tubing. A device was used in both setups. The measured flow rates were lower in the setup with flexible tubing than the one with stiff one, especially in the lower range of the flow rate (less than 5000 nL/min). T he slope of the c alibration curve is 0.96 for the setup with stiff tubing, which is similar to the Sensirion flowmeter calibration curve in Figure 2 2. However, it was reduced to 0.92 for the setup with flexible tubing, indicating lower flow rates in the devices. Larger standard deviations at the higher range of the flow rates make comparison between the two groups difficult. This result still indicates a flowmeter is not required if an approximate flow rate suffices. However, an external flowmeter becomes a necessity if one needs to know the accurate flow rate in a device when flexible tubing is used in connecting pumps with devices, especially when a complicated fluid network and the accessories have a possibility to diminish the pumping pressure. The pressure losses i n internal viscous flows, often called head loss, could result from channel/tube enlargements and contractions, inlets and exits, valves and other components.51 2.3.3 F low Pulsations Syringe pumps are extensively used as a pumping method in microfluidic applications. Pump is driven by the linear positive displacement of a syringe plunger resulting from the rotary motion of a stepper motor. The stepping motion is expect ed to generate flow pulsations.64, 65 Since the Sensirion flow meter is very sensitive with rapid response time, it was used to confirm the existence of flow pulsations and study the effects of a microfluidic device and tubing on flow pulsations. Figu re 2 5A shows the temporal profiles of flow rates measured by the flowmeter when it was directly connected to the syringe pump using stiff tubing. Significant flow pulsation was obvious as indicated by the sine wave embedded in white color. When the stiff tubing was

PAGE 22

22 replaced by fle xible one, flow pulsation was significantly reduced, presumably due to the damping effects of the tube flexibility on the pumping pressure. When the setup was connected with a device, flow pulsation was further reduced. The effects of the device on flow pulsation ar e also illustrated in Figure 2 5B in which a comparison was made between setups with and without a device. Stiff tubing was used in both cases. The results clearly suggest that connection of a microfluidic device with a syringe pump significantly reduced flow pulsations. The flexibility of the thin cover film of the plastic device likely played a role in diminishing pulsations. The results are in the agreement with a study on segmented flows in microfluidic devices.65 The effects of the flow rate (from 250 to 7000 nL/min ) on flow pulsation were also studied. Representative results at 1000 and 4000 nL/min are shown i n Figure 25B S ine waves in white color were added to the temporal profiles to show approximate pulse duration The average pulse duration over three periods was 31 s for 1000 nL/min while it was 7.6 s for 4000 nL/min. The duration of each pulse is abo ut 2 orders of magnitude longer than the period of the stepping motor in the syringe pump listed in Table 21. Longer pulse of the fluid oscillation was likely due to the damping effects of tubing and the device as discussed above. To confirm the pulse duration mismatch was not a part of the measurement system, two syringe sizes were used that would give different pulse duration due to the change in the stepping motor speed. The syringe with a volume of 100 L used in all experiments above was replaced w ith a syringe wit h a volume of 250L. Figure 25C shows the comparison of temporal profiles of flow rates between two setups using different syringes. The flowmeter was directly connected to the syringe pump without a microfluidic device. Stiff tubing was used in both setups and the flow rate was at 400 0 nL/min. S ine waves in white color were also added to the temporal

PAGE 23

23 profiles to show approximate pulse duration The average pulse duration over three periods was 19.0 s for the 250 L syringe, which was 2.5 times longer than the pulse duration of 7.6 s f or the 100 L syringe. This ratio is exactly the same as the expected ratio calculated from the parameters listed in Table 2 1. It was inferred from this result that the flowmeter accurately measures flow rates and the observations are true representation of flow pulsations. 2.3.4 Microvalve Testing U sing Flowmeter The valve testing was set up based on the flowmeter evaluation results discussed above. The concept of the valve is shown in Figure 2 6A which is, to some degree, similar to the elastomer valve s developed by Quake and co workers.19 Th e device consists of four layers, including substrate with fluid channels, elastomer, valve layer, and thin film. In Quakes valves, the valve layer contains control channels for pneumatic actuations. In contrast, thermosensitive materials was used in the present valves as indicated. When the heater was turned on, the thermosensitive materials expanded and deflected the elastomer to close the fluid channel as illustrated in Figure 2 6 A Details of microvalve device fabrication and operation are included in chapter 3. To demonstrate the valving m echanism the microvalve device was connected to the Sensirion flowmeter in a setup similar to Figu re 2 1. When the valve is closing, the cross sectional area of the fluid flow decreases, resulting in a chan ge in the flow rate. When the valve is c omplete ly closed, the elastomer is entirely deflect ed to the wall surface, thus the channel is blocked and the flow rate goes to zero Figure 26B shows a temporal profile of the flow rate recording from the Sensirion flo wmeter when the valve was open, closed and back to open. T he syringe pump was set at 300 nL/min and 400 mW power was supplied to the micro heater. The temporal profile in Figure 26B shows the dynamics of valve actuation Initially the flowmeter recorded the flow rate ( 300 nL/min ) set by the syringe pump. When the valve was

PAGE 24

24 closing (i.e. the heater was turned on), t he elastomer was deflected into the channel and the reduction in the flow rate was recorded by the flow meter. A n increase in the flow rate before the reduction likely resulted from the elastomer deflection. The sudden elastomer motion displace d the fluid downstream, resulting in an impulse in the flow velocity. The temporal profile also consisted of a period of negative flow rates, which can b e explained by the back pressure. After the elastomer was completely deflect ed onto the channel wall surface and the flow rate decreased to zero the inertia of the flow carried the fluid further downstream and a back pressure was developed near the elas tomer This back pressure caused a flow reversal downstream resulting in a negative flow rate detected by the flow meter. Once th ese transient state s were completed the flow reached to the steady state. After the flow in the channel was blocked and cont inuous zero flow rate was recorded, the valve was opened (i.e., the heater was turned off). The elastomer quickly relaxed to its original shape and flow rate returned to the initial value. However, before reaching the initial flow value there was a slight surge in the flow rate recorded possibly caused by the pressure built up at the valve inlet when the valve was closed .66 Hence wh en the valve wa s opened a momen tary/temporal increase in flow was seen which stabilizes with time. This surge effect is c learly illustrated in Figure 27 which shows the valve action for an inlet water flow of 300 nL/min. The valve wa s allowed to stay closed for different times and the corresponding surge i n the flow rate after the valve o pen ing is shown. When the valve wa s kept closed for longer times, higher flow peaks are reached in Figure 27, supporting the argument that pressure builds up at the valve inlet due to the principle of syringe pump operation. Higher pressure built up results in higher magnitude of flow surge once val ve opens. The profile with 87 s valve close time uses 190 mW heater inpu t power and the other two (97 s and 183s ) use 66 mW. It can also be noted that for higher powers of valve actuation, here 190

PAGE 25

25 mW case, the in i tial flowrate increase (prior to valve closing) is hig her It also results in larger reverse flow downstream near the flowmeter once the valve is closed. F aster elastomer deflection result s in larger pumping action during valve closing and hence the corresponding negative backpressure near valve exit increases. T ransient characteristics including the sudden surge in the flow rate (before and after valve action) and the negative values were reproducible in many experiments that were performed. The case discussed above illustrates the utility of a flowmeter for studying valve actuation and investigating local flow dynamics. 2.4 Conclusion A Sensirion flowmeter was evaluated and used to demonstrate a valve actuation mechanism by measuring the real time flow rates. The flowmeter was calibrat ed using the direct met hod, which weigh ed the amount of fluid collected over a certain period of time. The general agreement between the flow rate set by the syringe pump and the flow rate measured by the flowmeter indicated the reliable accuracy of the flowmeter and that flowm eter is not required for simple setups To ensure the flow rate specified by a syringe pump (or those using similar pumping mechanisms) truly represents the actual flow rate in a microfluidic device, stiff tubing should be employed. E ffects of stiff or fle xible tubing on the flow rate in a microfluidic device was studied and it was found that the use of flexible tubing resulted in a reduction in the flow rate in the device because the pumping pressure was slightly lost due to the stretching of flexible tubi ng. In addition, flow pulsation was confirmed to exist when a syringe pump was used. It was found that the degree of flow pulsation is related to the volume of the syringe used in the pump, the stiffness of tubing, the flow rate and the presence of microfluidic device s in the fluid

PAGE 26

26 network. This finding is a significant factor to be considered while setting up the microvalve test station and also since syringe pumps are extensively used in microfluidic applications. The flow meter was then used to st udy the valve mechanism of a thermally actuated microvalve. The flow meter provided detailed flow dynamics associated with the valve act uat ion s One drawback of the Sensirion flowmeter is that it must be connected via tubing to a microfluidic device. It wo uld be ideal to have a flow sensor integrated in a device so that it can measure the flow rates at a variety of locations. For instance, flow rates in branched channels of a complicated fluid network must be locally measured in order to know the accurate f low rate in each channel, similarly, presence of valves in a microchannel would need accurate local flow monitoring so as to control the flow effectively. Devices with integrated flow sensors will be useful in applications where accurate flow controls are required

PAGE 27

27 Table 2 1. V olume of syringes and parameters of syringe pump Syringe volume 100 L syringe 250 L syringe Flow rate (nL/min) 1000 4000 4000 nL/step 5.315 5.315 13.19 Steps/min 188.2 752.5 303.0 Seconds/step 0.32 0.080 0.20

PAGE 28

28 Figure 21. Schematic of the experimental setu p for this work. A syringe pump was connected via tube 1 to a microfluidic device which was then connected through tube 2 to a Sensirion flow meter. The fluid flowing out of the flow meter was collected for the flow rate calibration Figure 22. Comparison of the flow rates measured by the direct method and the Sensirion flowmeter. The measured flow rates were plotted as a function of the pump ing rate of the syringe pump. No microfluidic device was connected b etween the syringe pump and the flowmeter. microfluidic device flowmeter fluid collector tube 2 tube 1 computer syringe pump syringe pump 0 2000 4000 6000 8000 0 2000 4000 6000 8000 Pumping Rate ( nL /min)Measured Flow Rate ( nL /min)direct method Sensiron flowmeter

PAGE 29

29 Figure 23. Comparison of the flow rates measured by the Sensirion flowmeter in the presence and absence of a microfluidic device. Stiff tubing was used to connect the syringe pump with the device or the flow meter. The measured flow rates were plotted as a function of the pumping rate of the syringe pump 0 2000 4000 6000 8000 0 2000 4000 6000 8000 Pumping Rate ( nL /min)Measured Flow Rate ( nL /min)without device with device

PAGE 30

30 Figure 24. Comparison of the flow rates measured by the Sensirion flowmeter in two setups using stiff or flexible tubing. The tubing was used to connect the syringe pump with the microfluid ic device as shown in Figure 2 1. The measured flow rates were plotted as a function of the pump ing rate of the syringe pump. 0 2000 4000 6000 8000 0 2000 4000 6000 8000 Pumping Rate ( nL /min)Measured Flow Rate ( nL /min)stiff tubing flexible tubing

PAGE 31

31 Figure 25. Temporal profiles of flow rates measured by the Sensirion flowmeter in different setups. Sine waves in white color were added to show flow pulsations. A) The effects of tubing stiffness on flow pulsations. The flow rate was 4000 nL/min for all of them; the top and middle curves were shifted up by 2000 and 1000 n L/min for clarity. B) The effects of a microfluidic device on flow pulsations. Stiff tubing was used in all setups. The flow rate was either 100 0 or 4000 nL/min as indicated. C) The effects of the syringe size on flow pulsations. Stiff tubing was used in both setups (without a microfluidic device). The flow rate was 4000 nL/min while the sizes of syringes were either 100 or 250 L as indicated. 30 40 50 60 70 3000 4000 5000 6000 7000 Time (s)Measured Flow Rate ( nL /min)flexible tubing with device stiff tubing, no device (a) flexible tubing, no device 20 30 40 50 60 70 500 1000 1500 2000 2500 3000 3500 4000 4500 Time (s)Measured Flow Rate ( nL /min) without device (b) with device 4000 nL /min 1000 nL /min A B

PAGE 32

32 Figure 25. Continued. Figure 26. Microfluidic valve. A) The valve is actuated by a microheater. The elastomer is deflected to the channel wall when thermo sensitive material expands. The drawing is not to scale. B) Temporal profile of the flow rate recorded by the Sensirion flowmeter, which was connected downstream to the device. Water w as supplied at 300 nL/min by a syringe pump and the valve was then closed, and back to open. 400 mW power was supplied to Minco heater. 30 40 50 60 70 2000 2500 3000 3500 4000 4500 5000 Time (s)Measured Flow Rate ( nL /min)250 L (c) 100 L (a)plate with channels elastomer thermal sensitive material plate with valves film to cover heater open closed heater flowmeter -400 0 400 800 0 400 800 1200 Time (s) Flow rate (nL/min) heater on heater off (b) C A B

PAGE 33

33 Figure 27. Microvalve actuation data recorded by flowmeter showing the local flow dynamics. 300 nL/min of water is s upplied by a pump. The valve is kept closed for different times resulting in proportional flow increase once the valve is opened due to pressure built up at valve inlet 190 mW power is supplied for the 87s econd(s) valve close test and 66 mW power is supp lied for 97 s and 183 s valve close test.

PAGE 34

34 CHAPTER 3 FABRICATION AND CHARACTERIZATION OF A THERMALLY ACTUATED MICROFLUIDIC VALVE 3.1 Introduction There are several advantages in microfluidics devices over conventional bench top instruments. A micro total analysis system (TAS) promises shorter analysis time, lower reagent consumption with low detection limits.5 Use of par allel processing enables higher throughput. In addition, plastic microfluidic devices are suitable for mass production lowering the device cost. Overall analysis cost is also reduced. These devices essentially incorporate several critical components such a s valves, pumps and mixers to achieve the desired fluidic manipulation necessary for an efficient process control. Inability to develop inexpensive, reliable and portable systems for these components has been a hindrance in the full fledged employment of m icrofluidic technologies in all the previously predicted application areas .6 Addressing the problem s this Chapter describes the fabrication and operation of a thermally actuated microvalve. The microvalve is designed to address the above mentioned drawbacks as an improvement over the existi ng valves reported in the literature. In addition, t his chapter describes an electriccurrent measurement method to study local valve operation. The valve operation was characterized in terms of actuation times, power and flow rates 3.2 Brief Review of Microvalves S everal microfluidic valves are reported in literature ( Chapter 1 ) Each of these valves is based on different operating principles, material combination and some targeted at specific applications. Earliest work on a complete Lab on a chip devi ce by Harrison et al .8 made use of electro osmosis to control sample flow as an EO val ve. Valves reported by Quake et al .19 and Grover et al .67 consist of pneumatic actuation wherein pneumatic pressure is used to deflect a m embrane into a channel. Both normally closed and normally open valves have been made this

PAGE 35

35 way. Beebe et al. have developed pH actuated hydrogel valves which makes use of the surrounding media composition or property such as pH to open or close a channel ba sed on volumetric expansion or contraction of the hydrogel.41 In situ polymerized pseudo gel valves have also been reported by Fan et al .68 which restricts sample flow. Thermally actuated valves for MEMS based applications have been developed by Tai et al .35 and a commercially successful valve based on thermal principle is reported by Redwood Inc.34 for refrigerant flow control. Whitesides et al.69 developed a simple PDMS membrane valve using quantified mechanical screw rotations to make the valve close. Several other valves have been reported and they are modified versions of the above basic principles. Although each of these valves demonstrates its functionality, there is still need for a robust, reliable, versatile microfluidic valve which has a wide operating pressure range with zero leakage, low fabrication and operation cost, fabrication ease, faster response times, low power consumption and can be integrated on a large scale. For example, Quake valves19 making use of multiple PDMS layers are not ideally suited for industrial m ass production and the pneumatic connections make the microfluidic device accessories not portable and cumbersome to operate. Valves based on EO flow have leakage problems and several MEMS based valves are based on silicon processing and are not compatible with microfluidic material requirements such as optical clarity for detection. As such, silicon based microfluidic devices are not suitable to realize the predicted advantages of microfluidics as a science of future with promise17. Hence a novel microfluidic valve has been proposed and developed which is compatible with industrial fabrication technologies suitable for mass production. The principle of operation for this normally open valve is based on thermal actuation of an elastomeric membrane.

PAGE 36

36 3.3 Principle of Operation Fabricated microvalve is based on the principle of thermal actuation. When heat is supplied to an incompressible fluid there is volumetric expansion of the fluid due to its temperature rise. This change in volume is used to defle ct an elastomeric film into a microfluidic channel to be controlled. The elastomer selected should be of low stiffness and high elasticity so that it can be controlled without excessive heat input and can be used repeatedly. Heat input to the thermally exp anding fluid (actuating fluid) is supplied through micro fabricated heaters. The power supply to these heaters is controlled using printed circuit board (PCB). The concept of using PCB controlled valve actuation will enable large scale integration of these valves for high t hroughput applications. Figure 3 1 shows the schematic of the multilayer device assembly illustrating the concept of microvalve operation. The nomenclature assigned to each layer is as follows. The microfluidic channel layer is termed ch annel layer. The elastomeric film is termed as elastomer film or film and the cavity where the thermal fluid is encapsulated is called FC40 well, named after the commercial grade of the thermal fluid used, and the plastic layer in which it is filled i s called valve layer. The plastic thin film with the heater pattern next to the valve layer is called the heater layer. 3. 4 Experimental Section 3.4.1 Device Fabrication for Microvalve Plastic m icrofluidic device s for microvalve were fabricated, as des cribed in the previous section, by injection molding of cyclo olefin resin (Zeonor 1020 R) pellets using a Carver hydraulic press A CNC milling machine was employ ed to trim the plastic parts into 1 x 3 in. substrate s and drill 2 mm diameter holes at the ch annel ends. The holes at channel ends became wells for fluid connections. The device was 1.5 mm thick while all channel s in the microfluidic device were 110 m wide and 45 m deep The lengths of channel used for the valve action

PAGE 37

37 study were 25 mm and 47 mm. The cross section of the channels was D shaped due to isotropic etching of the glass substrate that was used for fabricating the molding master. The substrate was then covered with the elastomeric film and subsequent layers as described in the followin g sections. 3.4.2 Elastomeric Film The function of elastomeric film is to deflect itself into the channel during expansion of thermal fluid. The factors in choosing the film materials were based on Modulus of elasticity Elastomer/membrane thickness Bonding with COP Zeonor 1020R Based on the above factors, following three membrane materials were used for microvalve device fabrication in this study: Polydimethylsiloxane (PDMS) Polyester (PET) film 4 mil (100 m) thick topas 8007 The method of fabrication based on PET film is described in detail in the following sections. 3.4.3 Polyester (PET) Film In this section, microvalve device fabrication using PET film is described. The microfluidic substrate (channel layer) milled as described above was cleaned first For cleaning, the device was submerged in acetone (NF/FCC/EP, F.W. 58.08) and let to sit on a New Brunswick Scientific Shaker (Edison, NJ) for 5 minutes. It was then rinsed in running stream of DI water and dried overnight in the laminar fume hood. Doubl e coated pressure sensitive adhesive (PSA) tape 9019 was obtained from 3M (Minneapolis, MN) in the form of a sample roll. The PSA tape has a Polyester (PET) base film of 13 m thickness coated on both sides with 8.4 m thick adhesive 300. Thus the total fi lm thickness is 29.8 m. Adhesive 300 is an acrylic

PAGE 38

38 based medium firm adhesive suited for low surface energy (LSE) plastics like cyclo olefin polymers. In order to transfer the PET film over channel layer the kraft liner from one side of adhesive 300 was r emoved. Hand held rubber Squeegee was used to manually laminate the PET film over the device. The film was laminated from one end to the other longitudinally to minimize air entrapment and thus maximize bonding area. This process was carried out in the cle an air environment of the laminar hood. The squeegee was firmly used over the other side of the film on the liner for 3 4 times to ensure firm contact with the channel layer. The two layer assembly was then left for 24 hours to allow the bond strength to i ncrease. Next, the layer of microvalve to hold the thermal fluid was integrated with the previously made two layers. This third layer called the valve layer was made of 250 m thick Zeonor 1020R Cyclo olefin polymer film from PLITEK (Des Plaines, IN). T he film was cut into 1 x 3 inch size and has 2 mm holes at locations where valve actuation is desired. The hole/cavity was for filling the thermal fluid. The valve layer was cleaned with the same method as the channel layer using acetone and DI water rinse and left to dry overnight in a laminar hood. The clean and dry valve layer was then assembled with the previously assembled two layers of channel layer and PET film. For the same, the remaining liner on the PET film (opposite to channel side) is peeled of f followed by valve layer assembly. The valve layer was visually aligned such that the microfluidic channels runs below the centre of thermal fluid cavity and laminated over the PET film with exposed adhesive 300. Lamination was carried out using a J type Gundlach rubber roller (Belleville, IL) with firm manual pressure to ensure maximum surface contact and minimum air entrapment. Roller was run over the three layer device assembly for 3 4 times. Assembly was performed in laminar hood to ensure no dirt/dust or foreign particles affect interlayer bonding. The assembled device was then left in the laminar hood for at least 2 hours to

PAGE 39

39 increase the adhesive bond strength. The thermal cavity, referred to as FC40 well, was temporarily sealed using a Costar thermow ell sealer polyethylene tape from Corning (Corning, NY) to prevent dust from entering. The cover was removed before the next layer is bonded to the device. The three layer assembly was then bonded with a 250 m thick Zeonor 1020R film to seal the FC40 wel l. The cavity was subsequently filled with Fluorinert FC40 from 3M (Minneapolis, MN) using vacuum. 3. 4.4 Microheaters Resistive Microheaters were fabricated by sputter deposition of gold on plastic films. Heaters are serpentine Au resistors with specific thickness. The serpentine heater was designed in AutoCAD and the pattern was printed on a Transparency using a high resolution printer (5080 dpi from Pageworks, Cambridge, MA). The transparency was used as a photomask to photolithographically pattern the g old film deposited on plastic. 250 m thick Zeonor 1020R film was cut to 4 x 2 inch sizes and rinsed with acetone and DI water and dried overnight in a laminar hood. Kurt Lesker CMS 18 Multi Target Sputter Deposition tool (Pittsburgh, PA) was used to depos it 1000 thick Au film on the cleaned 1020R 250 m film. The deposition rate was 5 /sec and the time for deposition is 200 sec. After deposition of the Au film, positive photoresist Microposit S1813 from Shipley (Marlborough, MA) was spun on the Au coati ng in two steps. In the first step, the spinner was set at 1000 rpm for 15 sec followed by 5000 rpm for 50 sec. The Laurell spinner WS 400A 6NPP/Lite (North Wales, PA) was set at an acceleration of #15 (1560 rpm/sec) and the corresponding thickness of PR w as 1.8 m. The PR was then prebaked at room temperature for 24 hours. Karl Suss MA6 Contact Mask Aligner (Germany) was used to pattern the PR through the transparency mask. The PR was exposed at 365 nm for 20 seconds. The exposed PR was then developed usin g AZ 300 MIF for about 40 to 50 seconds followed by DI water rinse and Nitrogen stream dry. The developed Au pattern was then

PAGE 40

40 hardbaked at room temperature for 24 hours. Gold etchant TFA from Transene Company Inc. (Danvers, MA) was used to etch the pattern ed Au. The film with Au was submerged in the etched solution for 40 seconds to obtain the serpentine Au resistive heaters. The etched Au pattern on the plastic film was finally rinsed in acetone followed by DI water. It is then left in laminar hood to dry overnight or dried using a Nitrogen stream. The resistive heater pattern has provision for a hole to be drilled so that the thermal fluid Fluorinert FC40 can be filled. The hole was drilled in the heater film using a 100 m diamter drill bit MD 0102FL fro m Cutting edge technologies (Plymouth, MA). The CNC milling machine was used to drill the hole and the alignment was done visually. The heater is kept faceup while drilling so that the Au side of the heater film has smoother finish. Once the holes were dr illed the heater film was cut along the outline of the heater. It was then rinsed with DI water and dried using filtered dry air or Nitrogen stream. Alternately, it can be left to dry in laminar hood. The dried heater film with the patterned Au heater and the filling hole was then assembled over valve layer in the previously made three layer microvalve device. The heater layer was solvent bonded on top of the FC40 well using decalin70 from Fischer scientific (Fair Lawn, NJ). The heater layer was stamped over 0.1 mL of decalin placed on a 1 x 3 in. microscope slide and transferred on top of the FC40 well. Alignment was done visually such that the serpentine heater pattern was within the FC40 well cavity. Uniform hand pressure was sufficient to ensure enough surface contact with minimal air bubbles. The assembly was then heated at 60oC for 20 min. to complete the solvent bonding. 3. 4.5 Thermal Fluid Once the heater layer was bonded, the FC40 well in valve layer is then filled with Fluorinert FC40. 10 ml of Fluorinert was taken in a 200 ml Pyrex glass beaker and the microvalve device was immersed in the liquid such that the filling hole was about 1 2 mm below

PAGE 41

41 the liquid meniscus level. This assembly was pl aced in a vacuum chamber which was at a 20 in. Hg vacuum pressure. The assembly was subjected to vacuum for 30 seconds followed by vacuum release and then again for 4 minutes followed by vacuum release. The filled cavity was then sealed with a drop of RBC2001 e poxy (Warwick, RI) or conductive epoxy from Chemtronics (Kennesaw, GA) if needed for fl uorescence experiments. The epoxy was cured at room temperature overnight. The filled cavity co uld also be sealed by solvent bonding another layer of 100 m thick Zeonor 1020R using decali n and letting it cure at room temperature overnight. Upchurch Nanoport fitting (Oak Harbour, WA) was then connected to the inlet of the microfluidic channel with valve on it usi ng RBC2001 epoxy. Heater connecting wires were bonded on the heater traces using conductive epoxy fr om Chemtronics and let to cure overnight. Figure 3-2 shows the multilayer microvalve device. 3.4.6 Valve Characterization 3.4.6.1 Electric-current method Valve testing for actuation and proof of c oncept were performed by measuring channel conductivity71. A Microfluidic channel with a valve wa s filled with either 1M or 0.1M NaCl solution and the change in current through it was measured as a function of time for a specific applied power to heater. Platinum electrodes, 0.3 mm in diameter, 99.9% purity from Alfa Aesar (Ward Hill, MA) were immersed into the inlet and outlet wells of the channel to supply a constant voltage. The current through this circ uit was measured using an external series resistance of 56K The setup for principle of Electric-Current method based on channel conductivity measurement is shown in Figure 3-3. A Labview program was written to automate the data measurement and analysis process. Multifunctional DAQ card NI-PCI 6229 was procured from National Instruments (Austin, TX) along with SCC68 68 pin connector block and SCCTC thermocouple signal processing module.

PAGE 42

42 The DAQ card and the connector block were used to measure voltage and also to supply DC voltage for electric-current method of valve ac tuation. The signal conditioning module was also used to measure temperature through the ther mocouple attached near FC40well during valve action. Data acquisition was set at 10 Hz sampling. Figure 3-4 shows the Labview user interface for the data acquisition. A supply voltage of 8V DC was used for current-method of measurement. For a 25 mm long channel the resis tivity of a 0.1M salt so lution was determined experimentally to be 1.899 m corresponding to a cha nnel resistance of 13 M The connected series resistance of 56 K is 0.4% of the channel resistance. Supply voltage to heater is given by DC power supply from Agile nt (Santa Clara, CA). The setup discussed above is for valve charac terization when hydrostatic head was used as a pumping mechanism. In order to test valve for higher inlet pressures either a syringe pump or a hydrostatic head column was used. A UMP-II syringe pump from World Precision Instruments, Inc. (Sarasota, FL.) was controlled by a Mi cro4 controller. When using a syringe pump an Upchurch T fitting P-712 was used to connect the Pt electrode to the tubing. An Upchurch PEEK tube with 100 m ID and 1.57 mm OD was used to connect the syringe to the device. A 100-l gastight Hamilton 1710LT syringe (Reno, Nevada) was used for the reagent delivery. The reagents used in the experiments were 1M or 0.1M NaCl solution. 3.4.6.2 Temperature measurement Type K thermocouples (TC) from Omega (Stamford, CT) were used to measure temperature in the valve area during valve actuation. A 76 m diameter 5SC-TT-K-40-36 TC was used to measure temperature outside the h eater film on the valve assembly. The TC was insulated with four 1x1 cm kapton tapes. To calib rate the temperature insi de the channel, Chal 002, 50 m diameter, K type TC was inserted into a channel and FC 40 during valve action.

PAGE 43

43 3. 4.6.3 Fl uorescence measurement Rhodamine B (0.1 mM 20 mM carbonate buffer) was used to study valve action by measuring its fluorescence during valve action. The microvalve device was placed in the sample stage of the inverted microscope (IX51, Olympus, America I nc., Melville, NY), which was equipped with a 75W Xenon lamp (U LH75X) A 10x objective lens (Olympus ) with a numerical aperture of 0.3 and a working distance of 10 mm was used. The light passes through an excitation filter (HQ480/40, Chroma Technology, Ro ckingham, VT), a beam splitter, and an emission filter (HQ630/60m, Chroma Technology, for Rhodamine B (special blue filter set)). The channel is filled with the fluorescent solution and images were acquired by scientificgrade CCD digital camera C4742 8012AG from Hamamatsu (Bridgewater, NJ). The camera exposure rate is set at 21ms. Image data acquisition and analysis was done using Wasabi software from Hamamatsu. In certain cases, the channel walls were coated to reduce residual fluorescent dye adhesion to walls and surfaces. The coating was performed as follows: channels were first rinsed with DI water or 1x tris HCl tricine buffer. They were then filled with 2% HEC 1% BSA, 1x tris HCl tricine buffer and let to stay for 30 to 60 min, followed by DI water or buffer rinse. 3.5 Results and Discussion 3.5.1 Device Fabrication The first challenge in developing the microvalve was to fabricate the multilayer device wherein different layers had to be bonded to form a cohesive unit. Different elastomer materials w ere tried including PDMS, PET film and 4 mil cyclo olefin copolymer topas film. PDMS elastomer film, although satisfactory for bonding channel layer with elastomer, was not sufficient to retain the thermal fluid. The possible reason was improper adhesion b etween the PDMS elastomer film and the valve layer. 4 mil topas film, although, enabled fabrication of a complete microvalve device, was not an ideal elastomer material choice owing to its higher thickness of

PAGE 44

44 100 m compared to channel cross section, ve ry low elastic elongation and high brittleness. Table 3-1 shows the comparative modulus of elas ticity of the three membrane materials studied in this work. The successful microvalve devi ce was achieved using PET film which showed good bonding with Zeonor 1020 R layers (channel a nd valve layer) owing to its low-surfaceenergy (LSE) plastic bonding propert ies. Polyester film has very good solvent resistance and the PSA maintains its properties over a wide range of temperature. It can be operated up to 121oC for few hours and can be used up to 82oC for weeks. Manual squeegee PET application was sufficient to bond layers with minimal air entrapment for operation of microvalve. 3.5.1.1 Heater fabrication and integration The sputter deposition of 5/sec for 200 sec wa s intended to give a 1000 thick Au heater film. The mean resistance for a sample size of 36 resis tive patterns was 222 with a standard deviation of 12 Solvent bonding of the heater layer over the valve-layer using decalin was irreversible and was found to retain its bond during repeated valve actuation cycles. The temperature of the thermal fluid during valve ope ration was dependent on the power input to the heater. The maximum power that could be suppl ied to the heater was limited by the Tg (glass transition temperature) of the heater layer (250 m plastic film ), however, power less than the maximum allowed was sufficient for valve actuation. 3.5.1.2 Fluorinert encapsulation Fluorinert encapsulation is the filling of thermal fluid in the FC40 well. Vacuum filling was the method used for encapsulation. Based on visual inspection th e size of air bubble (diameter) as seen in Figure 3-2 from heater top is reproducibly at around 25% of the FC40-well diameter which corresponds to 6% of cavity vo lume. The air bubble is assumed to be a cylinder with height equal to valve layer thickness i.e., 250 m. The PET film was not permeable to FC40 as observed for devices which retained FC40 for months. FC40 was stable and did not have any

PAGE 45

45 adverse reaction with the PET film, Zeonor plastic and sealing epoxy. Sealing of the filling hole using epoxy or solvent bonding another plastic film was successful as it retained the thermal fluid over the cou rse of experiments. Solvent bonding method of sealing the hole although adds to the device complexity but helps to maintain the optical clarity of the valve region. This in turn helps monitor FC40 well for Fluorinert expansion, leakage and local fluid flow during valve close or open state. Additional details are mentioned in the next section. 3.5.2 Valve Characterization 3.5.2.1 Cyclic valve operation The electric current method was used for most of the valve characterization experiments. Here, current thr ough the channel was measured and the degree of valve action governs the magni tude of this current. Figure 35 shows the cyclic valve actuation for the microfluidic valve with PET elastomer film. Au heaters were used to heat the fluid. A power of 58 mW was supplied to the heater and the corresponding rise in temperature was measured across the heater film on the exposed side. The microfluidic device was supplied with a 9.5 mm hydrostatic head of 1 M NaCl salt solution. It can be seen that the channel condu ctivity reduces with increase in temperature which agrees with the physics of the microvalve actuation. As the temperature of the thermal fluid increases the membrane deflects into the channel. This results in a reduction in local cross sectional area ther eby increasing the channel resistance. Further, the increase in channel resistance causes the reduction in the current and the corresponding voltage drop across the series resistance. This voltage drop was measured and recorded by Labview program using the NI data acquisition system. Cyclic valve actuation is thus demonstrated for a specific pressure input to fluid. It also shows the repeatability of valve operation. Consecutive cycles show a reduction in the actuation time which can be explained based on t he thermal diffusivity of FC40

PAGE 46

46 since consequent cycles will need less power to heat the thermal fluid to the actuation temperature. 3.5.2.2 Actuation time The valve actuation time is defined as the time when the heater power is switched on to the time the voltage reduces to a value less than 3100 V and holds to that value for ~5 sec. Figure 36 shows the variation of this valve actuation time for different input power to heater. The driving pressure for the 1 M NaCl to flow was supplied by the hydros tatic head of 9.5 mm. Figure 35 and Figure 36 are recordings for the same microvalve device. It is seen that the valve can be operated over a wide range of inp ut power to heater. Depending on the input power, the valve actuation time varies. With increase in power, initially there is a considerable reduction in the actuation time as can be expected since higher power results in higher temperatures in shorter times. However, beyond a power value of 50 mW the actuation time reaches a plateau probably because a saturation heat input value has been reached and beyond which the actuation time or the membrane deflection time constant does not vary significantly. For each power value, except for the lowest power, 6 to 8 valve tests were conducted to obtain the error bars. Error bars are at one standard deviation. For the 36 mW power input, of the three tests, only one showed complete actuation remaining two had partial v alve action in an hour. 3.5.2.3 Operating temperature Being a thermally actuated microfluidic valve, one of the most critical operating parameters is the operation temperature. It is important to characterize the operating temperature at various locations in the microvalve device. The temperature of thermal fluid in the FC40 well determines the actual operating parameter. The temperature of the fluid in the microchannel is another important data of interest to end user. Since enzymes denature at a temperatu re in excess of 50oC72 it is very important for the microvalve to ope rate within this cut off value. Figure 37A

PAGE 47

47 shows the temperature data measured for thermocouple in the FC40 well. Two cycles of 2 min heater operatio n are shown. The temperature measured in FC40 well is compared with the temperature measured on the exposed surface of the heater film. The outside thermocouple has a stack of four 1 cm x 1cm kapton tape insulation. The temperature in FC40 well is higher b ecause of direct con tact with the heater. Figure 3 7B compares temperature on the exposed surface of the heater layer with the temperature in microchannel. The thermocouple in the channel was placed at a distance of 0.2 mm from the FC40 well edge. The temp erature in channel is lower than outside temperature due to insulation provided by the elastomer, radial heat conduction across the device and heat loss to the ambient. The locations of thermocouple in the device microvalve are illustrated in Figure 3 7C. Thermal characterization of the microvalve involves a plot of temperature as a function of power. Figure 38 shows the relationship between temperature and power supplied. The error bars are the standard deviations. For the channel temperature measurement, the thermocouple was placed at a distance of 0.8 mm and 2 mm from the FC40 well edge. As explained above, the temperature in the FC40 well is higher than that in the channels. The maximum operational power is constrained such that the maximum temperature in the microvalve device is less than the lowest Tg of the device components. Since the PET film can withstand temperatures up to 121oC the restricting parameter is the Tg of the 1020R Zeonor film which is 105oC. The power was limited to 80 mW for the 1000 Au heater. Above this power distortion in the heater pattern could be seen, on repeated uses, in the form of wrinkles which are indicative of excessive thermal stresses in the Au plastic interface. The channel temperature corresponding to the maximu m operating pow er was found to be less than 45oC. Thus, the power required for valve operation and the corresponding heater and component temperatures were found to be within the

PAGE 48

48 critical limits of device integrity without affecting end user applications. The minimum temperature difference required to actuate the valve based on th ermal fluid expansion, (Equation 3-1), was found to be 8.7oC. The temperature measured by thermocouple in FC40 well shows that actuation is achieved for temperature difference ranging from 25oC to 55oC. The probable reasons for difference include heat losses th rough conduction to the substrate material, convection of heat from the exposed device surf ace and possible deflection of heater film. The presence of an air bubble approximately covering 6% of the thermal fluid well volume also contributes to this difference. The bubble size was approximately less than 500 m in diameter corresponding to 6% of FC40 well volume. Howe ver, the theoretical temperature difference needed corresponding to this reduction in FC40 volume was 9.3oC. Hence other factors play a greater role. The relationship between the volumetric expans ion and temperature can be described by Vd T dV 1, (3-1) where is 0.0012 cm3/ (cm3 oC) the volumetric coefficient of thermal expansion of FC40. dV and dT are the change in volume and temperature. The change in volume is equal to channel volume in valve region equal to 8.2e-6 cm3. V is the total volume of FC40 well 785.4e-6 cm3. Corresponding minimum dT is 8.7oC. 3.5.2.4 Valve operation with liquid pressure: pumping and hydrostatic head From the practical application perspective, the valve actuation is tested for syringe pumped flow and with hydrostatic head. The syringe pump was used to supply flow at a constant flow rate or a liquid column at the inlet was used for hydrostatic head. Figure 3-9A and B show a comparative valve action plot for a 300 nL/min (1.2 mm/sec flow velocity) pumped flow at two different heater input po wers. 1M NaCl solution is used as the reagent. With increase in input

PAGE 49

49 power, the actuation time for valve to close decr eases and the time for valve to open increases. The results are in agreement with the physics; with increase in heater input power, the temperature at which valve actuates is reached faster. However it will take longer time to cool down and open the valve due to th e thermal diffusivity of the FC40 thermal fluid. Thus there should be an optimum valve operation power depending on the time constants desired. Addressing the above claim, Figure 3-9C summari zes the actuation times of the valve to open and close as a function of heater input power. Syringe pump is a constant flow source as a result during valve action pressure will rise locally near the valve region.66 There are two outcomes of this fa ct. Firstly, such an increase in pressure locally is not controlled hence the pres sures at times were found to delaminate the well region elastomer film. As a cons equence, this was found to a ffect the device performance. Secondly, to monitor this pressure rise and to completely characterize valve using a syringe pump a pressure sensor needs to be connected upstream between the syringe pump and valve to monitor the pressure in microf luidic channel during operation. As an alternate method to characterize valve action with positive ap plied fluid pressure hydrostatic head was used. Hydrosta tic head was designed here to be a constant pressure source as a result uncontrolled pressure rise at the valve region was a voided. The flow rate through the microfluidic channel corresponding to the head was calculated using the Hagen-Poiseuille flow (Equation 3-2)73 which considers the head lo ss due to viscous effect of the fluid flow. The flow rate Q can be calculate d using the equation: L PD Qh 1284 (3-2) where A is the cross sectional area 4.1e-5 cm2 and P is the perimeter 1.8e-2 cm. The dynamic viscosity is represented by 8.9e -7 Pa-s for 0.1M NaCl and L is the channel length 4.7 cm. The

PAGE 50

50 hydrostatic pressure difference across the channel is given by P in Pa as shown in Table 3-2. Hydraulic diameter, Dh can be defined by: P A Dh2 (3-3) Figure 3-10 shows a plot of channel conductiv ity as a function of time for values of pressure heads. The heads are given in mm of 0.1 M NaCl solution. The power is maintained constant at a value of 55 mW. It is observed th at the actuation time for valve to close increases with increase in inlet pressure. Longer time was needed for higher heat accumulation for valve to actuate and counter the increase in inlet pressure. Table 3-2 lists the pressure and flow rate values corresponding to the hydrosta tic heads for a 47 mm long microfluidic channel used for the valve test. The density was measured using Fischer accu-124 analytical balance (with an accuracy of 0.1 mg) to be 990 Kg/m3 which is similar to the value of 1001 Kg/m3 calculated based on the data published in the NIST scientific and technical database for 25oC. 3.5.2.5 Fluorescence measurement Rhodamine B (0.1 mM in 20 mM carbonate buffer) was used as the fluorescent dye for the valve action imaging experiments. The valve area under the FC40 well was imaged using the CCD camera controlled by Wasabi software GUI. The channel was surface coated and filled with Rhodamine B prior to valve action. The coating described in the experimental section helps reduce the residual adhesion of dye to channe l wall surfaces. Figure 3-11 shows fluorescence images of the valve area before and after valve ac tuation. The reduction in in tensity of dye in the FC40 well region indicates valve action. On supplyi ng a power of 68 mW to heater, the PET film deflects into the microfluidic channel and disp laces the fluorescent dye thus resulting in an intensity reduction. The higher intensity near th e longitudinal channel edge is probably due to displacement of PSA during actua tion. The PSA on PET film would be displaced towards the

PAGE 51

51 pivot point o f elastomer deflection which is the channel edge. Fluorescent imaging thus demonstrates the concept of valve operation. 3.6 Conclusion This chapter reports the fabrication and testing of a thermally actuated microvalve for a plastic microfluidic device. M ultilayer device fabrication involves bonding several layers together. The elastomeric films elastic properties and its ability to be bonded to the microfluidic channel were found to be critical for successful device fabrication. The device has been fabri cated and the concept was demonstrated in terms of valve actuation using an electriccurrent method. The valve actuation times were found to be less than 10 sec, however, consistent repeatability needs to be improved. The operation temperature is within the maximum allowable temperature the components of microvalve device can withstand. Valve actuation was achieved over a range of heat er input powers from 36 mW to 80 mW. For consistent actuation higher than 45 mW is desirable. The maximum number of times the valve could be used varied depending on the pressure source used. For a input head of 9.5 mm the valve was found to operate for 50 actuations. The valve was also integrated to a printed circuit board (PCB) (Figure 312) to control the power supply to t he heater thus setting the stage for next level of device fabrication addressing large scale integration. The PET film used for valve test demonstration has pressure sensitive adhesive on either side. The adhesive was found to exhibit strong bonding with p lastic and was sufficient to hold the device intact enabling valve testing. However, it was found that on repeated valve actuations the glue would accumulate near the channel edge which is a hindrance to realize large number of cyclic valve operation. Also presence of glue makes it necessary for the channel surface to be pretreated if it is to be used for real life sample manipulation.

PAGE 52

52 The valve has been shown to operate with syringe pumped flows or under hydrostatic inlet heads. The valve relaxation time or the time for valve to switch off once the heat input is stopped was found to be dependent on heater power and for subsequent cycles dependent on the inlet flow rate. Presence of adhesive on elastomer sometimes resulted in sticking effect causing hystere sis in valve operation. The electriccurrent method of measurement gives a very high response time for local valve actuation. However, it has an inherent disadvantage since current will flow through any conducting path. Although, the elastomer, adhesive and plastic were found to be not conductive when tested separately, even a thin layer of conductive solution would shunt the electric circuit. Thus controlling current leakage path was difficult in some cases. Also, the electriccurrent method does not gi ve the actual flow rate of the reagent. However, the methods ability to give fast, local action outweighs its negatives. More so, similar current m ethod have been used previously .40, 71 In summary, the microfluidic valve demonstrated here is the first step towards realizing a reliable valve whose fabrication will suite mass production manufacturing technology and large scale integration.

PAGE 53

53 Table 3 1. Elastomer film properties Elastomer film Modulus of elasticity, Kpa Tensile strength, Kpa Elongation % Solvent resistant Operating temperature range oC Poly dimethyl siloxane (PDMS) 750 74 6.2 x 10 6 100 Swells with nonpolar organic fluids 75 55 to 200 Polyester (2.8 to 3.1) x 106 5.5 to 7.5 x 106 50 150 Medium to Low solvent resistance 40 to 82 (long term). 40 to 121 (short term) Cyclo olefin copolymer Topas 8007 2.6 x 10 6 63 x 10 6 10 Attacked by nonpolar organic solvents Heat deflection temperature = 75 Table 3 2. Hydrostatic heads and flow rates Hydrostatic head in mm Pressure in Pa Flow rate in nL/min 40 388 890 70 680 1577 100 971 2225 135 1311 3004 215 2087 4748 The inlet pressure and flow rate corresponding to the applied hydrostatic heads are tabulated. The length of microfluidic channel used is 47 mm.

PAGE 54

54 Figure 31. Schematic of the microfluidic valve concept based on thermal actuation principle. A) Valve open B) Valve closed. Heat supplied to an expanding thermal fluid causes an elastomer film to deflect into the channel and regulate the flow. Figure 32. Multilayer microvalve device. Inset: FC40 well filled with Fluorinert FC40. The serpentine Au resistive heaters are also seen with the sealed filling hole. Section X X shows the actual cross section of the multilayer device made using Lapping films with g rain sizes 5 m, 3 m, 1 m and 0.1 m obtained from Thorlabs (Newton ,NJ) A B

PAGE 55

55 Figure 33. Electriccurrent method to study valve actuation. A) Valve is in open position. The channel is filled with a conductive solution which allows curr ent to flow through it. B) When power is supplied to the heater, the valve goes to closed position blocking the channel. The corresponding current through the channel, thus, goes to zero. C) Actual experimental setup showing the microvalve device connected to a power supply. Pt electrodes are used for electriccurrent measurement. B A C

PAGE 56

56 Figure 34. Labview GUI for data acquisition during valve testing. A) Shows temperature and voltage data for 10 samples over a second. B) Real time temperature and voltage data acquired. C) Numerical values of data acquired. D) Block diagram of the Labview program A B C

PAGE 57

57 Figure 34. Continued D

PAGE 58

58 Figure 35. Valve actuation when 58 mW was applied to Au heaters. A hydrostatic head of 9.5 mm drives the flow. The channel conductivity and the temperature corresponding to resistive heater are shown in the plot as a function of time.

PAGE 59

59 Figure 36. Valve actuation time as a function of the input power. Error bars indicate one standard deviation and a hydrostatic head of 9.5 mm drives the flow of 1M NaCl salt solution.

PAGE 60

60 Figure 37. Temperature at different locations during valve actuation at 58 mW input heater power. A) Temperature of FC40 is measured using a K type thermocouple. The outside t emperature is used as a comparative reference. B) Temperature measured in the microfluidic channel at a distance of 0.2 mm from well edge. C) Schematic showing thermocouple location in the device A

PAGE 61

61 Figure 37. Continued. B C

PAGE 62

62 Figure 38. Temperature at different locations of a microvalve as a function of power.

PAGE 63

63 Figure 39. Valve actuation for a 300 nL/min flow with different heater input powers. A) Current through the microfluidic channel. B) Temperature on the expose d side of a heater film across the FC40 well. Higher powers result in higher temperature in shorter times and thus faster valve action, however, slower valve release to its initial off state. C) Valve time constants as a function of power. C B A

PAGE 64

64 Figure 310. Electric current as a function of time when different hydrostatic heads were used. The pressure heads are in mm of 0.1M NaCl. The power was kept constant at 55 mW for all the tests. Rolling time averaging ,over 0.3 sec (every 3 consecutive da ta points), was used to smoothen the curve. A) Channel conductivity. B) Temperature corresponding to input power. B A

PAGE 65

65 Figure 311. Fluorescent images showing valve action. Images are captured in the FC40 well showing the valve open to close d action. The reduction in fluorescence intensity shows the elastomer deflection in the valve area of microfluidic channel resulting in displacement of fluorescent dye. A) Filled FC40 well over the microfluidic channel. B) Valve open state with fluore scent dye filled in channel. C) Valve closed state wherein the PET film deflects into channel and displaces the dye. A B C

PAGE 66

66 Figure 312. The PCB connected to a microvalve device

PAGE 67

67 CHAPTER 4 CONCLUSIONS AND FUTURE WORK A thermally actuated microfluidic valve has been fabricated and its operation has been demonstrated. The principle of the valve is based on thermal expansion of a fluid that deflects an elastomeric film into a microfluidic channel. Heat is supplied by a li thographically patterned heater. The power input to the microvalve can be controlled by a PCB. 4.1 Microvalve Fabrication The microvalve has been fabricated with an aim to extend the concept to large scale integration. The valve is made of polymer s and th e bulk of the material used was cyclo olefin polymer with the elastomeric film being polyester. All plastic assembly helps to achieve lower device cost and faster fabrication times. The channel layer was injection molded and the remaining layers are assemb led onto that layer by lamination making it suitable for industrial mass production. 4.2 Microvalve Operation 4.2.1 Evaluation U sing a Flow M eter The microvalve was tested using two methods one is to use a micro flowmeter while the other is electrical current measurement The conclusions using each is summarized below. The first method used a commercial micro flow meter. The flow meter was connected downstream to measure the flow rate during v alve action. Prior to valve testing, the flowmeter was evaluated for accuracy based on weight measurement. Tests were also performed to determine the effect of stiff and flexible connecting tubing as well as the effects of the microfluidic channel. In add ition, the pulsations produced by the syringe pump were studied. The results show that stiff tubing should be used for applications that require accurate flow control. The microflow meter was required for flow evaluation When connected downstream the flow meter showed the valve

PAGE 68

68 actuation dynamics. The flow meter showed the flow rate go to zero when the valve was closed and back to the or iginal value when it was open. During the operation the pumping action due to membrane deflection was also observed, indi cated by a hump or increase in flow rate. The flowmeter also showed that a longer valve close time with constant syringe pump supply resulted in higher pressure built up at the valve entrance. This was indicated by an increase in flow rate beyond the syrin ge pump rate once the valve is opened The valve was tested reliably for a 300 nL/min (1.2 mm/s) flow supplied by a syringe pump. Although the flowmeter gives the flow rate variation during valve action it had certain limitations. The flowmeter could not b e connected near the valve region; as a result the actual local flow changes could not be monitored. Connecting the flow meter downstream showed flow variations which were due to flow reversals downstream and the pumping action. Hence a new method was de veloped to measure local valve actuation. 4.2.2 Electric C urrent M easurement The channel conductivity was measured to study the local effect of valve action s This method also demonstrate d repetitive valve operation. It was found that the valve openand c l ose cycles were in coherence with the heater input, which was proportional to the temperature in the valve. The valve could be operated over a wide range of power from 36 mW to 80 mW. The pow er controls the actuation times, typically a few seconds. The tem perature in a channel during operati on was found to be less than 45oC which is within the maximum temperature allowed before enzymes denature.72 The valve was reliably tested at a flowrate of 300 nL/min (1.2 mm/s). Valve actuation w as also demonstrated for an inlet pressure of 2000 Pa corresponding to 215 mm hydrostatic head. A valve can operate as many as 50 times Electric current measurement was advantageous since instantaneous change in the cross section due to elastomer deflecti on could be monitored. The downstream effects seen in the flow

PAGE 69

69 meter method were overcome. However, any leakage current s through conductive parts could result in an erroneous measurement Interfacial layer of thin conductive fluid could also result in erro rs The valve operation was also validated by fluorescence measurement in the actual valve region. Fluorescent solution filled in the channel was displaced due to valve action. The valve was also shown to be operable when it is connected with a PCB. 4.3 F uture Work Our study demonstrated the concept of a thermally actuated valve. Since the elastomer film used in the device has uniform adhesive coat on all areas channel region is partially exposed to the membrane glue. The presence of glue results in hysteresis due to sticking effect. Hence, the adhesive should be removed locally in the future to improve the valve suitability to biological applications. The p resence of glue can contaminate samples. With respect to channel geometry, the aspect ratio of the channel can be made less stringent for e.g. 10:119 rather than presently used 2.4:1. The single device fabrication process should be extended to an array of valves and all of them should be controlled using a PCB. The valve testing can also be improved if a pressure sensor is used in conjunction wi th a flowmeter. A pressure sensor connected upstream will enable monitor the pressure rise in the valve inlet during operation. It will also help define the maximum operation pressure66 that the valve can withstand without delamination. In addition, a flow meter that can be integrated in the microchannel will enable local flow rate measurement.

PAGE 70

70 LIST OF REFERENCES 1. R. P. Feynman, Microelectromechanical Systems, Journal of 1992, 1 6066. 2. J. S. Kilby, Proceedings of the IEEE 2000, 88 109 111. 3. S. C. Terry, J. H. Jerman and J. B. Angell, IEEE Trans. Electron Devices 1979, 26 18801886. 4. K. E. Petersen, Proceedings of the IEEE 1982, 70 420457. 5. A. Manz, N. Graber and H. M. Widmer, Sens. Actuator B Chem. 1990, 1, 244248. 6. G. M. Whitesides, Nature 2006, 442 368 373. 7. W. M. Kurt Petersen, Gregory Kovacs, Allen Northrup, Lee Christel, and Farzad Pourahmadi, in IVD Technol. Mag. 1998, pp. 4349. 8. D. J. Harrison, K. Fluri, K. Seiler, Z. H. Fan, C. S. Effenhaus er and A. Manz, Science, 1993, 261 895897. 9. 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, 1998, 282, 484487. 10. A. M. Skelley, J. R. Scherer, A. D. Aubrey, W. H. Grover, R. H. C. Ivester, P. Ehrenfreund, F. J. Grunthaner, J. L. Bada and R. A. Mathies, Proc. Natl. Acad. Sci. U. S. A. 2005, 102 10411046. 11. J. S. Marcus, W. F. Anderson and S. R. Quake, Anal. Chem 2006, 78 956958. 12. T. Thorsen, S. J. Maerkl and S. R. Quake, Science, 2002, 298, 580584. 13. J. West, M. Becker, S. Tombrink and A. Manz, Anal. Chem. 2008, 80 44034419. 14. S. Haeberle and R. Zengerle, Lab Chip, 2007, 7 10941110. 15. D. B. Wei bel and G. M. Whitesides, Curr. Opin. Chem. Biol. 2006, 10, 584591. 16. H. Becker and C. Gartner, Anal. Bioanal. Chem. 2008, 390 89111. 17. A. de Mello, Lab Chip, 2002, 2 31N 36N. 18. J. C. McDonald and G. M. Whitesides, Accounts Chem. Res. 2002, 35 491499. 19. M. A. Unger, H.P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Science, 2000, 288 113116.

PAGE 71

71 20. J. M. K. Ng, I. Gitlin, A. D. Stroock and G. M. Whitesides, Electrophoresis 2002, 23, 34613473. 21. H. K. Wu, T. W. Odom, D. T. Chiu and G. M. Whitesides, Journal of the American Chemical Society, 2003, 125, 554559. 22. S. R. Quake and A. Scherer, Science, 2000, 290, 15361540. 23. S. K. Sia and G. M. Whitesides, Electrophoresis 2003, 24 35633576. 24. Y. Liu, D. Ganser, A. Schneider, R. Liu, P. Grodzinski and N. Kroutchinina, Anal. Chem. 2001, 73, 41964201. 25. M. Hupert, W. Guy, S. Llopis, H. Shadpour, S. Rani, D. Nikitopoulos and S. Soper, Microfluid. Nanofluid. 2007, 3, 1 11. 26. Q. Mei, Z. Xia, F. Xu, S. A. Soper and Z. H. Fan, Anal. Chem. 2008, 80 60456050. 27. C. K. Fredrickson, Z. Xia, C. Das, R. Ferguson, F. T. Tavares and Z. H. Fan, J. Microelectromech. Syst. 2006, 15, 10601068. 28. K. W. Oh and C. H. Ahn, J. Micromech. Microeng. 2006, 16, R13 R39. 29. B. Bae, J. Han, R. I. Masel and M. A. Shannon, J. Microelectromech. Syst. 2007, 16, 14611471. 30. I. Fazal and M. C. Elwenspoek, J. Micromech. Microeng. 2007, 17, 23662379. 31. H. Q. Li, D. C. Roberts, J. L. Steyn, K. T. Turner, O. Y aglioglu, N. W. Hagood, S. M. Spearing and M. A. Schmidt, Sensors and Actuators A: Physical.Micromechanics section of Sensors and Actuators, based on contributions revised from the Technical Digest of the 16th IEEE International conference on Micro Electro mechanical Systems (MEMS 2003) 2004, 111 5156. 32. W. H. Grover, A. M. Skelley, C. N. Liu, E. T. Lagally and R. A. Mathies, Sensors and Actuators B: Chemical 2003, 89, 315323. 33. M. L. A. H. Zdeblick, CA), The Board of Trustees of the Leland Stanford Junior University (Stanford, CA), United States, 1989. 34. A. K. Henning, J. Fitch, D. Hopkins, L. Lilly, R. Faeth, E. Falsken and M. Zdeblick, in 1997 International Conference on Solid State Sensors and Actuators, 1997. TRANSDUCERS '97 Chicago. 1997, pp. 825828 vol.822. 35. X. Yang, C. Grosjean, Y.C. Tai and C. M. Ho, Sensors and Actuators A: Physical.Tenth IEEE International Workshop on Micro Electro Mechanical Systems 1998, 64, 101108. 36. X. Yang, C. Grosjean and Y. C. Tai, Microelectromechanical Systems, Journal of 1999, 8, 393402.

PAGE 72

72 37. C. A. Rich and K. D. Wise, Microelectromechanical Systems, Journal of 2003, 12 201 208. 38. J. H. Kim, K. H. Na, C. J. Kang, D. Jeon and Y. S. Kim, Microelectronic Engineering.Micro and Nano Engineering 2003, 2004, 7374 864 869. 39. H. Takao, K. Miyamura, H. Ebi, M. Ashiki, K. Sawada and M. Ishida, Sensors and Actuators A: Physical 2005, 119 468 475. 40. S. Vyawahare, S. Sitaula, S. Martin, D. Adalian and A. Scherer, Lab Chip 2008, 8 1530 1535. 41. D. J. Beebe, J. S. Moore, J. M. Bauer, Q. Yu, R. H. Liu, C. Devadoss and B.H. Jo, Nature 2000, 404 588 590. 42. D. T. Eddington and D. J. Beebe, Microelectromechanical Systems, Journal of 2004, 13 586593. 43. E. T. Carlen and C. H. Mastrangelo, Micro electromechanical Systems, Journal of 2002, 11, 408420. 44. R. H. Liu, J. Bonanno, J. Yang, R. Lenigk and P. Grodzinski, Sensors and Actuators B: Chemical 2004, 98 328 336. 45. G. V. Kaigala, V. N. Hoang and C. J. Backhouse, Lab Chip, 2008, 8, 10711078. 46. C. H. Ahn, J. W. Choi, G. Beaucage, J. H. Nevin, J. B. Lee, A. Puntambekar and J. Y. Lee, Proceedings of the IEEE 2004, 92, 154173. 47. M. Yamada and M. Seki, Anal. Chem. 2004, 76, 895 899. 48. S.Shoji., in Microsystem technology in chemi stry and life sciences ed. H. E. B. Andreas Manz, Holger Becker, Springer, 1999, p. 252. 49. G. T. Kovacs, Micromachined Transducers SourceBook McGraw Hill Higher Education, 1998. 50. L. A. Legendre, J. M. Bienvenue, M. G. Roper, J. P. Ferrance and J. P. Landers, Anal Chem 2006, 78 1444 1451. 51. R. W. Fox and A. T. McDonald, Introduction to fluid mechanics J. Wiley, New York, 1998. 52. C. K. Fredrickson and Z. H. Fan, Lab Chip, 2004, 4, 526533. 53. N. T. Nguyen, A. H. Meng, J. Black and R. M. White, Sensors and Actuators aPhysical 2000, 79 115121. 54. N. T. Nguyen, Flow Measurement and Instrumentation, 1997, 8 7 16.

PAGE 73

73 55. M. Ashauer, H. Glosch, F. Hedrich, N. Hey, H. Sandmaier and W. Lang, Sensors and Actuators aPhysical 1999, 73 7 13. 56. E. Kj eang, B. Roesch, J. McKechnie, D. A. Harrington, N. Djilali and D. Sinton, Microfluidics and Nanofluidics 2007, 3 403 416. 57. X. H. Huang, M. J. Gordon and R. N. Zare, Analytical Chemistry 1988, 60 18371838. 58. D. K. Kim, A. Majumdar and S. J. Kim, Sensors and Actuators aPhysical 2007, 136, 8089. 59. J. G. Santiago, S. T. Wereley, C. D. Meinhart, D. J. Beebe and R. J. Adrian, Experiments in Fluids 1998, 25, 316319. 60. C. Helmbrecht, R. Niessner and C. Haisch, Analytical Chemistry 2007, 79, 70977103. 61. G. R. Wang, I. Sas, H. Jiang, W. P. Janzen and C. N. Hodge, Electrophoresis 2008, 29 12531263. 62. U. Kanne and C. Sauvain, Medical Device Technology 2006, June 63. U. Kanne, T. Preiswerk and G. Bohm, Lc Gc Europe 2003, 16, 65 66. 64. S. Kar, S. McWhorter, S. M. Ford and S. A. Soper, Analyst 1998, 123, 14351441. 65. J. M. Kohler and P. A. Gross, Microfluidics and Nanofluidics 2007, 3, 653663. 66. G. F. Chen, F. Svec and D. R. Knapp, Lab Chip, 2008, 8 11981204. 67. W. H. Grover, M. G. von Muhlen and S. R. Manalis, Lab Chip 2008, 8 913918. 68. C. Das, C. K. Fredrickson, Z. Xia and Z. H. Fan, Sens. Actuator A Phys. 2007, 134 271277. 69. D. B. Weibel, M. Kruithof, S. Potenta, S. K. Sia, A. Lee and G. M. Whitesides, Anal. Chem. 2 005, 77, 47264733. 70. T. I. Wallow, A. M. Morales, B. A. Simmons, M. C. Hunter, K. L. Krafcik, L. A. Domeier, S. M. Sickafoose, K. D. Patel and A. Gardea, Lab Chip 2007, 7 1825 1831. 71. X. Huang, M. J. Gordon and R. N. Zare, Anal. Chem. 1988, 60 1837 1838. 72. D. L. Purich, Contemporary enzyme kinetics and mechanism Academic Press, San Diego, 1996. 73. J. M. C. Yungas A.engel, Fluid mechanics : fundamentals and applications McGraw HillHigher Education, Boston, 2006.

PAGE 74

74 74. M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Science, 2000, 288 113116. 75. J. N. Lee, C. Park and G. M. Whitesides, Anal. Chem. 2003, 75 6544 6554.

PAGE 75

75 BIOGRAPHICAL SKETCH Karthik was born and brought up in Nagpur, India. He earned his Bachelor of Engineering degree in mechanical engineer ing in June 2004 from National Institute of T echnology Karnataka Surathkal. In August 2006, h e enrolled at the University of Florida to pursue a Master of Science degree in mechanical engineering.