UBC Undergraduate Research

Development of micropump using thermally activated hydrogels Vohradsky, Honza; Fu, Jun Wei; Edgcumbe, Philip 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52966-Vohradsky_Honza_et_al_APSC_479_2011.pdf [ 2.4MB ]
Metadata
JSON: 52966-1.0074469.json
JSON-LD: 52966-1.0074469-ld.json
RDF/XML (Pretty): 52966-1.0074469-rdf.xml
RDF/JSON: 52966-1.0074469-rdf.json
Turtle: 52966-1.0074469-turtle.txt
N-Triples: 52966-1.0074469-rdf-ntriples.txt
Original Record: 52966-1.0074469-source.json
Full Text
52966-1.0074469-fulltext.txt
Citation
52966-1.0074469.ris

Full Text

Rm. 3112 2332 Main Mall Vancouver, BC V6T 1Z4 Dear Dr. Boris Stoeber, We are enclosing the report “Development of micropump using thermally activated hydrogels”. The report was written by Honza Vohradsky, Jun Wei Fu and Philip Edgcumbe as a requirement for the UBC Engineering Physics APSC 479 report. The goal of this project is to design and fabricate a thermally activated peristaltic micropump in a monolithic multi-layer polydimethylsiloxane (PDMS) device. Designing a new type of micropump in the field of microfluidics is important because microfluidics is a fast growing field with applications in everything from micro-PCR to digital microfluidics for early cancer detection. Project objective #1, eliminate leakage, and objective #3, production of portable microfluidic device were completed. Project objective #2, microvalves with response time of <3 seconds incorporated into a micropump was not completed. In early January, 2011 we discovered that replacing 15% Pluronic with 17% Pluronic and spinning on PDMS at 8000rpm allowed for good gel formation and elimination of leakage on the device. In the next few weeks we will test the spin-on technique with 17% Pluronic and hope to complete objective #2 shortly. Sincerely, Honza Vohradsky, Jun Wei Fu and Philip Edgcumbe Enclosure  Executive Summary The aim of this project is to design and fabricate a thermally activated peristaltic micropump in a monolithic multi-layer polydimethylsiloxane (PDMS) device. The advantage of our micropump over the conventional pressure activated micropumps is that our design only requires one pressure-control valve whereas the conventional approach requires one pressure-control valve for each valve. The goal of the project is to incorporate the newly designed micropump into a fully portable microfluidic device that can be used to pump fluid and cells through the chip. The goal is to design thermally activated valves with a response time of <3 seconds and micropumps that can pump fluid at 0.1 nL/sec. Designing a new type of micropump in the field of microfluidics is important because microfluidics is a fast growing field with applications in everything from micro-PCR to digital microfluidics for early cancer detection. The equipment and resources of this project fall into three distinct sub-categories. They are: Design, fabrication and testing. For device design we used a computer with 2D AutoCAD drawing capabilities and a printer for transparencies with 10um resolution is required. For fabrication, we used a cleanroom with wet bench, spinner, hot plate, UV light, gold for evaporation, gold evaporator, PDMS mixing facilities, oxygen plasma, plasma bonding and lab space is required. For testing, we used an inverted microscope, pressure source and fluorescent particles. The project is sponsored by Dr. Boris Stoeber and he has agreed to provide the resources and equipment that we need to make this project a success. This is an exciting project which has the potential to offer a significant new tool to the field of microfluidics. Project objective #1, eliminate leakage, and objective #3, production of portable microfluidic device were completed. Project objective #2, microvalves with response time of <3 seconds incorporated into a micropump was not completed. Leakage was eliminated by spinning on uncured PDMS onto a glass slide, curing the PDMS and then bonding the glass slide to a PDMS multi-layer device. A portable carriage with pressure sources, microcontroller and power supply was designed and built for carrying our microfluidic device and a program developed for the microcontroller to operate the microfluidic pump. The development of the micropump was not completed because of inconsistent Pluronic gel behavior and it took us until January, 2011 to eliminate leakage in our device. All team members are committed to continuing the project at a collective work rate of 10 hours per week until the micropump works on the portable device. Key recommendations are to make and test more microfluidic devices with spin-on PDMS and characterize device response time, pumping rate and pressure. Further, we will test the adhesion promoter GE SS412 for improving the spin-on PDMS bonding to glass.  Contents 1  Introduction ............................................................................................................................. 7 1.1  2  1.1.1  Technical Background .............................................................................................. 7  1.1.2  State of the art technology - A comparison ............................................................ 8  1.1.3  Alternative strategies .............................................................................................. 9  1.1.4  Results from previous experimental work ............................................................ 11  1.1.5  Project Sponsor ..................................................................................................... 11  1.1.6  Power consumption calculation ............................................................................ 11  Discussion .............................................................................................................................. 13 2.1  Design ............................................................................................................................ 13  2.1.1  Macroscopic Portable Device ................................................................................ 13  2.1.2  Fluid Supply Systems ............................................................................................. 13  2.1.3  Heater Power and Controller ................................................................................ 14  2.1.4  Series 4 Design ....................................................................................................... 16  2.2  3  Background and Motivation ............................................................................................ 7  Testing ........................................................................................................................... 20  2.2.1  Overview ................................................................................................................ 20  2.2.2  Pluronic Rheology .................................................................................................. 20  2.2.3  Plasma Bonding ..................................................................................................... 22  2.2.4  Parylene Bonding ................................................................................................... 23  2.2.5  RTD Tests ............................................................................................................... 25  2.2.6  Electrolysis Test ..................................................................................................... 28  Project Deliverables ............................................................................................................... 31 3.1  Deliverables ................................................................................................................... 31  3.2 As presented in our Project Charter (Appendix B - Recipe for 10 um SPR220-7.0 Mold for 4-inch Si Wafers ................................................................................................................... 31 3.2.1  Deliverable 1: Microvalve with 3 second response time ...................................... 33  3.2.2  Deliverable 2: Portable device............................................................................... 33  3.2.3  Deliverable 3: Peristaltic pump ............................................................................. 33  3.3  Financial Summary ........................................................................................................ 33  3.3.1  Macro Components ............................................................................................... 33  3.3.2  Micro Components ................................................................................................ 34  2  3.4 4  Ongoing Commitments by Team Members .................................................................. 34  Conclusion ............................................................................................................................. 35 4.1  Important Results .......................................................................................................... 35  4.2  Project Review ............................................................................................................... 35  4.2.1  Gel formation of Pluronic could easily be predicted, achieved and reproduced .. 35  4.2.2  There would be no electrolysis in the device. ....................................................... 36  4.2.3 We were confident that we had appropriate steps and contingency plans to quickly eliminate leakage in our device. ............................................................................... 37 4.2.4 PDMS spun onto a glass slide at high rpm (>1000rpm) would have extremely high porosity due to stretching of the polymer. ........................................................................... 38 5  Recommendations................................................................................................................. 39 5.1  Specific Recommendations ........................................................................................... 39  5.2  General Recommendations ........................................................................................... 39  References ..................................................................................................................................... 40 6  Appendices ............................................................................................................................ 42 6.1  Appendix A - AZ 5214E Photoresist Datasheet.............................................................. 42  6.2  Appendix B - Recipe for 10 um SPR220-7.0 Mold for 4-inch Si Wafers ......................... 48  6.3  Appendix C - Project Charter (No signatures) ............................................................... 50  6.4  Appendix D - Team Member's Time Contributions ...................................................... 53  6.5  Appendix E - Arduino Code for Controlling One Cycle of Valve Operation ................... 54  3  6.9  Appendix I - Series 4 Lithography Masks ..................................................................... 130  6.10  Appendix J - Cleanroom Wafer Fabrication Record .................................................... 132  6.11  Appendix K - Glass Slide Record .................................................................................. 138  6.12  Appendix L - Macro Assembly ..................................................................................... 144  4  Figure 1 - Block Diagram of macroscopic portable device ............................................................ 13 Figure 2 - Macroscopic portable device ........................................................................................ 13 Figure 3 - Round-tipped pogo pin. Source: Sparkfun <http://www.sparkfun.com/products/9173> ............................................................................... 14 Figure 4 - Circuit diagram, npn transistor amplifier circuit ........................................................... 15 Figure 5 – Heater designs; a) Series 3; b) Series 4 ......................................................................... 17 Figure 6 – Heating area details; a) Series 3; b) Series 4 ................................................................ 17 Figure 7 – Control channel heating area details; a) Series 3C; b) Series 3A; c) Series 4B. The orange box indicates the nominal area of flow. ............................................................................ 18 Figure 8 - Time lapse of gel formation and dissolution. Heater voltage 13V was held constant with pressure varied. a) 1psi, no flow; b) 3psi, no flow; c) 5psi, no flow; d) 13 psi with flow. .... 19 Figure 9 - Control channel; a) Series 3B; b) Series 4B. The orange and green boxes indicate the diffusion barrier and the valve actuation areas, respectively. ...................................................... 19 Figure 10 – Control channel heating area details; a) Series 3C; b) Series 3A; c) Series 4B. The orange box indicates the nominal area of flow. ............................................................................ 20 Figure 11 - Viscosity vs temperature for 3 tested batches of Pluronic ......................................... 21 Figure 12 - Press used to apply 1MPa pressure during parylene thermal bonding. ..................... 24 Figure 13 – RTD resistance vs heater voltage for the bare and covered heater cases ................. 26 Figure 14 - Experimental apparatus, hotplate test ....................................................................... 27 Figure 15 - Resistance vs. temperature for the bare, covered dry channel, and covered wet channel .......................................................................................................................................... 27 Figure 16 – Schematic of the electrolysis test apparatus; a) crossed leads; b) parallel leads; c) conventional connection ............................................................................................................... 29 Figure 17 – Electrolysis seen when the device is connected as in Figure 16 (a); a) before the application of 4V; b) t=2s after power applied; c) t=7s d) t=14s. Recorded using 10x lens. ........ 30 Figure 18 – Dirty PECVD Plasma Chamber .................................................................................... 38 Figure 19 – Team member’s time contributions vs. time ............................................................. 53  5  Table 1 - Properties of Pluronic batches tested by Rheology ....................................................... 21 Table 2 - Macro Component Breakdown ...................................................................................... 33 Table 3- Micro Component Breakdown ........................................................................................ 34  6  1 Introduction 1.1 Background and Motivation We propose to develop a thermally activated micropump and implement the technology in a portable microfluidic device. The thermally activated micropump will be a new kind of tool in the field of microfluidics. Microfluidics is a fast growing field with many interesting applications in biology and beyond. 1.1.1 Technical Background 1.1.1.1 An introduction to microfluidics The field of microfluidics involves the manipulation of small (10-9 to 10-18 liters) amounts of fluids with channels that are tens to hundreds of micrometers across. The early microfluidic devices were developed to use very small quantities of samples and reagents and to do low cost analysis of chemical solutions. Microfluidics owes its ongoing popularity not only to its size but also due to the behavior of fluidics at the microscale. Namely, microfluidics operates in a regime of low Reynolds number (i.e.: the ratio of momentum of a fluid to its viscosity is low) which allows for laminar flow. Further, factors like fluidic resistance, surface tension and energy dissipation starts to dominate the system.  1.1.1.2 Paradigm shift: Application of polydimethysiloxane (PDMS) for multilayer microfluidic chips and valves Microfluidics, and its associated ability to take advantage of the behavior of solutions at the microscale, was first explored in detail in the 1990s. Manz et al.'s paper written in 1992, is an example of one of the first articles in the field1. However, the era of rapid and affordable microfluidic device prototyping only came of its own in 2000. In 2000, Unger et al published an article titled: "Monolithic Microfabricated Valves and Pumps by Multilayer Soft Lithography"2. In this article, Unger et al showed the world how to use Polydimethysiloxane (PDMS) to make simple multilayer valves. Unger's valves are simple to make and simple to operate. The multilayer devices consist of a fluidic channel and control channel. The control channel and the fluidic channel are on top of each other and where they cross they are separated by a thin member (approx. 10μm). The crossing point between the control and fluidic channel is where a valve is. To activate the valve, pressure is applied in the control channel the membrane deflects into the fluidic channel. The membrane is easily deflected because PDMS, the elastomer used to make  1  A. Manz, et al., "PLANAR CHIPS TECHNOLOGY FOR MINIATURIZATION AND INTEGRATION OF SEPARATION TECHNIQUES INTO MONITORING SYSTEMS - CAPILLARY ELECTROPHORESIS ON A CHIP," Journal of Chromatography, vol. 593, pp. 253-258, 1992. 2 M. A. Unger, et al., "Monolithic microfabricated valves and pumps by multilayer soft lithography," Science, vol. 288, pp. 113-116, 2000.  7  the device is a soft material with Young’s modulus of 750 kPa3. The pressure of the various valves is controlled outside of the chip by an array of solenoid valves.  1.1.2 State of the art technology - A comparison 1.1.2.1 Current microfluidic valving strategies Since flow control is an integral part of microfluidic devices, many research groups have proposed alternative designs for valves and pumps. For example, materials with large thermal expansion coefficients have been used to open and close4 and hydrogels have been used to develop pH-sensitive microvalves5 or other thermally sensitive hydrogels have been heated up by a laser6. Each of the proposed valve and pump designs has drawbacks. The large thermal coefficient valves have poor response time, the pH sensitive hydrogels have very limited applications and the laser activation of thermally sensitive valves introduces added complexity and cost to the system.  1.1.2.2 Our proposal - a novel microfluidic valving strategy Our design uses Pluronic F127, a triblock copolymer, which when heated, has a phase transition from low viscosity to a soft, high viscosity cubic crystalline gel phase. We heat the Pluronic with an on-chip heater that is made of gold and selectively deposited by evaporation. Gel formation controls the multilayer valve by determining whether or not the external pressure reaches and deflects the valve. The advantage of our design is that it promises a fast response time and it is simple to fabricate and operate. 7 We will use the thermally activated micropump to build a lowpower portable microfluidic device capable of moving cells in a loop in the microfluidic chip.  1.1.2.3 Currently available portable microfluidics technologies Microfluidics is a rapidly developing field where applications are being found for the technology at a rapid pace. Currently, there are many commercial implementations of benchtop microfluidic systems capable of carrying out complex experiments at significant time and cost savings. One area of particular promise is being able to deliver medical diagnostic tests in the field. This has two key advantages: firstly, results of the test can be reported immediately, and 3  J. C. Lotters, et al., "The mechanical properties of the rubber elastic polymer polydimethylsiloxane for sensor applications," Journal of Micromechanics and Microengineering, vol. 7, pp. 145-147, 1997. 4 K. Pitchaimani, et al., "Manufacturable plastic microfluidic valves using thermal actuation," Lab Chip, vol. 9, pp. 3082-7, 2009. 5 D. J. Beebe, et al., "Functional hydrogel structures for autonomous flow control inside microfluidic channels," Nature, vol. 404, pp. 588-+, 2000. 6 K. Tashiro, et al., Micro flow switches using thermal gelation of methyl cellulose for biomolecules handling. Berlin: Springer-Verlag Berlin, 2001. 7 B. Stoeber, et al., "Flow control in microdevices using thermally responsive triblock copolymers," Journal of Microelectromechanical Systems, vol. 14, pp. 207-213, 2005.  8  secondly, the cost of doing such a test is dramatically reduced without the need to equip and staff a wet lab. To our knowledge, only electrokinetics have been demonstrated using a portable (i.e. one that can be unplugged from the wall) device.8 The company closest to bringing a product to market may be Micronics Microfluidics in the form of immunoassay and immunohematology systems. However, there is no mention of either system or their technologies in the academic literature. 1.1.3 Alternative strategies 1.1.3.1 Choosing polydimethysiloxane (PDMS) over glass for microfluidic chip We had to choose between using PDMS or glass for our device. The five main differences between PDMS and glass are: Fabrication process, bonding, softness of material, heat transfer coefficient and heat capacity. For PDMS the fabrication process involves making a photo-resist via UV exposure that once fabricated can be used repeatedly for multiple PDMS chips. PDMS can be bonded to glass very easily via oxygen plasma treatment. PDMS is very soft, it has a Young’s modulus of 750 kPa. Finally, it has a heat transfer coefficient of 0.21 w m-1K-1 and heat capacity of 1500 w m-2 K-1 (Niu, ZQ; Chen, WY; Shao, SY, et al). For glass, the fabrication process is more involved. Fabrication includes etching glass using hydrofluoric acid (HF) and anodic bond of the silicon to glass. This process uses HF, a dangerous chemical, is more expensive, and no mold is created. Glass is very hard so any valve would have had to be built by putting a PDMS membrane between two glass layers. Glass has a heat transfer coefficient of 0.75 w m-1K-1 and heat capacity of 834.61 w m-2 K-1 (Niu, ZQ; Chen, WY; Shao, SY, et al). The lower heat capacity of the glass is attractive because our heaters will inevitably heat the surrounding channel walls. A low heat capacity assures that heating the surrounding walls does not interfere too much with the heating or cooling of the solution in the microfluidic chip. Further, unlike PDMS channels, glass channels do not expand under pressure. Expanding channels extend response time because more fluid needs to be cycled in and out of the valve for each cycle. Even though glass has a lower heat capacity and non-expanding fluidic channels, we chose to use PDMS. The advantage of cheaper, simpler and faster fabrication process for PDMS more than outweighs the advantages that glass has to offer. 1.1.3.2 Chip fabrication considerations Another key consideration was the method of fabrication for the fluidic chip. To date, the fabrication procedure we have used is multilayer soft lithography. This process begins with the design of a mask using Solidworks. This is then printed on transparency sheets using laser printers to form a mask. This mask is then used to selectively etch a silicon wafer with UV  8  D. Erickson, et al., "A miniaturized high-voltage integrated power supply for portable microfluidic applications," Lab on a Chip, vol. 4, pp. 87-90, 2004.  9  exposure. Using a typical office printer, feature sizes of 250μm are achievable; by using high resolution 20 000 dpi printers, feature sizes of up to 10μm are possible.9  The key advantage of this method is that has well characterized, repeatable results in terms of channel dimensions. Drawbacks of it are the need for ~20min of cleanroom time and the turnover time for shipment of high-resolution transparency prints from a US supplier (CAD Art Services). Recently, novel methods of prototyping soft fluidic chips have been examined, the most notable of which being the use of Shrinky Dink prestressed thermoplastic sheets. This material is used for children's toys whereby designs can be drawn onto the sheets and then heated to induce shrinking of the material. In fabricating microfluidic devices, Shrinky Dinks can be used either to make a mould for the PDMS or be used to make the fluidic chip itself.10,11 Yet another mould making method is to transfer the layout from a laser printed piece of glossy paper to brass plate to selectively shield desired areas from an etching agent. 12  All of these methods are capable of giving feature sizes acceptable to the needs of this project with the benefit of significantly reduced cost, iteration time, and eliminating cleanroom time. However, since we have no firsthand experience with these methods and are not expecting the need for a large number of design iterations, we have opted to continue with the soft lithography process.  1.1.3.3 Connecting macro to micro The gold heater traces are embedded between the PDMS and glass, with only small pads exposed to which leads are attached to provide power. Currently, the leads are attached by soldering, which frequently causes problems with jumpers since the pad spacing is quite tight. Our solution was to re-design our heater pads to fit the standard protoboard spacing and use pogo-pins attached to a protoboard which are perfectly aligned with our heater pads. This design modification streamlined our testing procedues.  9  A. Singhal, et al., “Microfluidic Measurement of Antibody-Antigen Binding Kintetics From LowAbundance Samples and Single Cells. 10 C. J. Easley, et al., "Rapid and inexpensive fabrication of polymeric microfluidic devices via toner transfer masking," Lab on a Chip, vol. 9, pp. 1119-1127, 2009. 11 C. S. Chen, et al., "Shrinky-Dink microfluidics: 3D polystyrene chips," Lab on a Chip, vol. 8, pp. 622-624, 2008. 12 A. Grimes, et al., "Shrinky-Dink microfluidics: rapid generation of deep and rounded patterns," Lab on a Chip, vol. 8, pp. 170-172, 2008.  10  1.1.3.4 Pressure source considerations To build a portable device, a pressure source to drive the control fluid independent of central lab compressed air is required. For this application, several configurations of camp stove fuel reservoirs were considered, along with a custom solution. Camp stove fuel reservoirs offer fluid capacity ranging from 325 - 975mL and an integrated pump. The fluidic chip is expected to require 1.54x10^-9 m^3/s, giving an expected run time of 2.44 days, which is extremely oversized. Instead, the Stoeber lab has a small, custom-designed reservoir which was modified to retain pressure and allow for a pressure relief valve and gauge.  1.1.4 Results from previous experimental work This is an important project for our sponsor, Dr. Boris Stoeber, because he has worked extensively with Pluronic and is keen to see it's applications in microfluidics. In 2005, Dr. Stoeber published an article showing that Pluronic could be used to stop flow in glass channel. In 2010, Dr. Stoeber published an article with graduate student Mr. Vahid Bazargan which is a proof-ofconcept article for thermally activated microvalves.13This project is a natural extension to that publication because we propose to make a micropump using a similar concept and show how the technology can be applied on a portable device. Critically, we propose to improve the microvalves response time from 20 seconds to 3 seconds and to implement a micropump. We have made significant changes to Mr. Bazargan's design to improve our response time. Philip worked with Dr. Stoeber during the summer to design and develop the micropump. However, the design and fabrication steps proved to be more difficult than anticipated and as of the end of the summer, Philip had sent in two design iterations and fabricated several devices. He was able to form a Pluronic gel in the microfluidic chips and could see valve deflection when a large pressure was applied to the system. However, leakage of PDMS between the gold, glass and PDMS remains an ongoing problem by the end of the summer Philip had not successfully deflected the membrane of the valve via gel formation. At the end of the summer the project was far from done. There is still a lot of work to do to solve the leakage problem, to optimize and characterize the design and to build a portable microfluidic device. 1.1.5 Project Sponsor The sponsor for this project is Dr. Boris Stoeber, a Professor at UBC that is cross-appointed in electrical and mechanical engineering.  1.1.6 Power consumption calculation Power Drain: Arduino = 25mA Pressure Transducer = 20mW, Ip = (20mW)/(5V) = 4mA 13  V. Bazargan, et al., “Flow Control Using a Thermally Actuated Microfluidic Relay Valve”. Journal of Microelectromechanical Systems, 2010.  11  Heaters = (15mW) x 6heaters = 90mW, Ih = (90mW) / (5V) = 18mA Total Current = 25mA + 4mA + 18mA = 47mA Battery Charge = 565mAh = 2034C Time to drain the battery = (2034 C) / (47*10^-3A) = 43300s = 12h It will take 12 hours of continuous operation to drain the battery.  12  2 Discussion 2.1 Design 2.1.1 Macroscopic Portable Device One of our key objectives is was the design of a portable device that allowed us to actuate the micropump without any fixed air or electric utilities. To achieve this, we designed a pressure/fluid storage reservoir with power provided from 2 9V batteries controlled by an Arduino Mini.  Heater Power and Controller Pressure Charger  Fluidic Channel Fluid Reservoir Fluidic Device  Pressure Charger  Control Channel Fluid Reservoir  Figure 1 - Block Diagram of macroscopic portable device  Figure 2 - Macroscopic portable device 2.1.2 Fluid Supply Systems The pressure charger and reservoirs for the fluidic and control channels are identical systems designed to supply fluid at constant pressure. The reservoirs are pressurized at the beginning of the experiment; since its capacity is much greater than the volume of fluid delivered, the pressure at the beginning and end can be assumed to be constant.  13  The reservoir was furnished with a Schrader valve to allow for flexibility in the initial source of pressure. In the lab, this can be provided from compressed air utilities while in the field, a hand bike pump can be used. 2.1.3 Heater Power and Controller In the past, the heaters have been connected to the fluidic device by soldering wires to heater pads on the glass slide. To improve this, we have made 2 main improvements:   Pogo pins    Arduino-controlled heating  2.1.3.1 Pogo Pins Previously, to power the heaters, wires needed to be soldered by hand to heater pads. This meant that the pads had to be relatively large and there was a high chance of shorting pads together during soldering. Because of how the gold is deposited onto the glass slides, removing and resoldering wires will often damage the pad, rendering the heater unusable. A pogo pin is a spring-loaded connector commonly used for attaching electronics equipment temporarily to produced systems for testing and initial programming. The spring-loaded action allows the connector to make solid, repeatable contact with the pad and allows for differences in height due to manufacturing variation. Using pogo pins as opposed to allows for a higher density of electrical interconnects to the device, potentially allowing a greater number of heaters and sensors to be implemented in the same envelope; whereas previously, the pad density was governed by the pad area needed for hand soldering, we now are able to use standard protoboard spacing of 0.1” for ease of fabrication. In the future, it will be possible to reduce the pad pitch to slightly larger than the pogo pin tip diameter.  Figure 3 - Round-tipped pogo pin. Source: Sparkfun <http://www.sparkfun.com/products/9173> 2.1.3.2 Arduino-Controlled Heating The Arduino Pro Mini microcontroller has 14 GPIO pins, of which 6 have PWM implemented. The use of a microcontroller to control heating levels allows for greater precision of heating, as well as the ability to experiment with different heating curves. To boost the available current, a basic npn transistor amplifier circuit was used.  14  Figure 4 - Circuit diagram, npn transistor amplifier circuit The code used to control one cycle of valve operation is shown in Appendix E. The code is quite simple, as all that it needs to do is turn heaters on and off in the proper sequence. It uses two PWM outputs, one for each heater. The voltage to be sent to the heater through the PWM output needs to be high enough to cause Pluronic gel formation but not as high as to cause electrolysis. Therefore the values sent to the PWM outputs will change depending on the heaters and the powers necessary to form gel formation. Theoretically, the power necessary to cause gel formation should be constant and as such the output of the PWM should be determined based only on the resistance of the heater. However, in reality the power delivered to the Pluronic depends on the heat transfer properties of the heaters and other factors. Therefore the power necessary to turn the heater on needs to be determined experimentally using a DC power supply. This experimental power is equal to the average power output of the PWM, which in turn depends on the duty cycle set in the code. The average power of the PWM output is: (  )  Where nPWM is the value (between 0 and 255) sent to the PWM output in the code, V is the voltage output of the Arduino after amplification and R is the resistance of the heater. This power needs to be equal to the power PDC necessary to cause gel formation which is experimentally determined from voltage as:  Where VDC is the power supply voltage necessary to cause gel formation and R is the resistance of the heater. Setting these equal and solving for the PWM value gives:  15  nPWM is equal to the variable int PWMvalveOn in the code of Appendix E. Finally, it should be noted that the value VPWM is not the 5V voltage output of the Arduino board but rather the 18V signal after amplification, i.e. VPWM = 18V. In addition, the time delays between the different parts of the valve cycle need to be determined and set in the code. These time delays need to be determined experimentally when the response time of the valve is determined and they need to be as low as possible to minimize the valve’s response time. The time delays are likely the same for each valve. 2.1.4  Series 4 Design  2.1.4.1 Goals The Series 4 design is a general revision of the heater and control channel layers. Compared to the Series 3 design, the main goals were:   Protoboard compatibility    Integrated temperature sensing    Greater fluid path length    Larger valve actuation area    Removal of other channels and diffusion barriers  2.1.4.2 Protoboard Compatability To eliminate the problems associated with soldering wires to the heater pads, we opted to connect the heaters to power sources and other devices using pogo pins soldered to standard pitch (0.1”) protoboards. Since no soldering of wires would be needed, the pad size can be much smaller, allowing for greater density of electrical interconnects; whereas the Series 3 heater pads required 61mm2 for 4 interconnects, we were able to put 6 pads in a 74mm2 envelope.  16  (a)  (b)  Figure 5 – Heater designs; a) Series 3; b) Series 4  2.1.4.3 Integrated Temperature Sensing Since Pluronic has a small temperature range in which it forms a gel. Both above and below this range, the fluid has identical flow properties, making it difficult to tell if the fluid is above or below the gelation temperature. To measure the temperature, a RTD was implemented by interdigitating two heater systems. In this way, one can be used as a heater and the other as an RTD. An added advantage of this design is redundancy; if using the RTD is not necessary, two independent heaters are available in each heating area.  (a)  (b)  Figure 6 – Heating area details; a) Series 3; b) Series 4 2.1.4.4 Greater Fluid Path Length From our experiments on the Series 3 devices, we believe there is a strong relationship between total wall area heated and the ability for the formed gel to prevent flow. In particular, design 3B, which uses dense columns in the heating area instead of zig zag channels, did not appear to hold a gel at all. For this reason, we wanted to increase the total fluid path length under the heating area. 2.1.4.5 Larger Valve Actuation Area The volumetric flow rate from a peristaltic pump is dependent on the volume of fluid it is able to displace. To increase our pumping capacity, the fluidic channel was tripled through the pumping zone, requiring the valve area to be stretched.  17  (a)  (b)  (c)  Figure 7 – Control channel heating area details; a) Series 3C; b) Series 3A; c) Series 4B. The orange box indicates the nominal area of flow.  (a)  (b)  18  (c)  (d)  Figure 8 - Time lapse of gel formation and dissolution. Heater voltage 13V was held constant with pressure varied. a) 1psi, no flow; b) 3psi, no flow; c) 5psi, no flow; d) 13 psi with flow. 2.1.4.6 Removal of Other Channels and Diffusion Barriers Series 3 includes a coolant flow channel and a diffusion prevention channel beside the primary flow path. In characterizing these devices, we saw that the fluid would leak along heater traces into these side channels. Removing these would simplify the overall design and may increase reliability.  (a)  (b)  Figure 9 - Control channel; a) Series 3B; b) Series 4B. The orange and green boxes indicate the diffusion barrier and the valve actuation areas, respectively. 2.1.4.7 Greater Fluid Path Length From our experiments on the Series 3 devices, we believe there is a strong relationship between total wall area heated and the ability for the formed gel to prevent flow. In particular, design 3B, which uses dense columns in the heating area instead of zig zag channels, did not appear to hold a gel at all. For this reason, we wanted to increase the total fluid path length under the heating area.  19  (a)  (b)  (c)  Figure 10 – Control channel heating area details; a) Series 3C; b) Series 3A; c) Series 4B. The orange box indicates the nominal area of flow.  2.2 Testing 2.2.1 Overview In the context of the deliverables we hoped to achieve, several experiments were designed to address individual issues. Presented chronologically, these tests were:   Plasma Bonding:    Parylene Bonding:    Pluronic Rheology: Since we were unable to form Pluronic gel as expected in the fluidic devices, the quality of the Pluronic itself became suspect. To verify the viscometric properties, rheology was carried out on several samples.    RTD: A key feature of the Series 4 heater design is the addition of an integral RTD. To verify that it functions as predicted and produces repeatable data, the resistance of several samples of this device was tested under known temperature conditions.    Electrolysis: Although the formation of bubbles had been seen within the control channel in the past, this had been attributed to heating the fluid past its boiling temperature. To investigate if electrolysis was possible and how to detect it, an experiment was designed to rule out other sources of bubbles and check for this effect.  2.2.2  Pluronic Rheology  2.2.2.1 Introduction  20  When we had difficulty getting consistent results for Pluronic gelation in the device, we first attempted to check the Pluronic response by putting it in an oven and on a hotplate. For these experiments, we varied the temperature from ambient to 65 degrees and did not observe gelation at any temperature in this range. As this is far higher than the published gel temperature, this result was very unexpected. These experiments were consistent between over 6 batches of Pluronic made with the following parameters varied:   Water type: distilled and deionized    Fluorescent tracer particles: in the mix and not  To undertake a more careful analysis the thermal properties of Pluronic, we used a rheometer to check if the viscosity profile is as expected. 2.2.2.2 Methods We tested 3 samples using an Anton Paar MRC series rheometer using the cone and plate geometry. Their preparation parameters are as shown in Table 1. Batch Number 2 11 12 1  2  Target Pluronic % by wt 15 15 15  Fluorescent Particles [g]1 0.114 0 0.0345  Water [g] 8.37932 4.2603 4.2187  Pluronic [g] 1.5037 0.7497 0.7848  Pluronic % by wt. 14.96 15.57  The fluorescent particles are provided to us in a concentrated, 0.02% solution with water. This sample was prepared with DI water; subsequent samples were all prepared with distilled water.  Table 1 - Properties of Pluronic batches tested by Rheology 2.2.2.3 Results The results are presented in Figure 11. 3000  Viscosity [Pa·s]  2500 2000 1500  12  1000  11 2  500 0 31  32  33  34  35  36  37  Temperature [˚C]  Figure 11 - Viscosity vs temperature for 3 tested batches of Pluronic  21  2.2.2.4 Analysis The erratic viscosity measurements above the gel temperature can be attributed to the fluid being a two-phase system whereby there are areas of gel and liquid, depending on localized temperature differences. This effect also explains why we do not see the fluid return back to a liquid at high temperatures as expected. 2.2.2.5 Conclusion The lower bound on the gelation temperature varies between 32-34˚C. We do not see a temperature at which the gel re-liquefies; this is believed to be since the phase there may be a two-phase (gel and liquid) system present between the cone and plate whereby there are areas of liquid and gel. This means that due to the experimental apparatus, all the data above the lower gelation temperature should be treated as suspect. 2.2.3  Plasma Bonding  2.2.3.1 Introduction Oxygen plasma bonding is a very popular technique in the field of microfluidics. It is used to create a bond between PDMS to PDMS or PDMS to glass. Oxygen plasma improves adhesiveness by cleaning the surface of contaminants and introducing reactive chemical groups. In PDMS, the -O-Si(CH3)2- group is converted to a silanol group (-OH) which changes the PDMS surface chemistry from hydrophobic to hydrophilic and allows for Si-O-Si bonds between PDMS to PDMS surfaces or PDMS to glass surfaces. We use the direct bonding via oxygen plasma both for bonding our multi-layer devices to glass and for bonding our multi-layer devices to the 10um of PDMS that is spun onto the glass surface. 2.2.3.2 Methods The following parameters were varied:   Power: 15, 30 watts    Time: 10, 15, 20 seconds    Weight: 0, 4, 8 pounds  The following parameters were not varied:   Gas composition: 100% oxygen    Chamber pressure: 500mtorr  2.2.3.3 Conclusions There was no apparent link between the variation of the plasma parameters and the quality of the bond. The best 30 watts, 15 seconds and 4 pounds, although even this would yield approximately 30% success.  22  2.2.4  Parylene Bonding  2.2.4.1 Introduction / Motivation Parylene bonding is an alternate method of bonding PDMS to PDMS or PDMS to glass. As mentioned above, one of the issues we were having with the device was leakage of Pluronic along the gold heater traces at the glass-PDMS interface. Sometimes the bond between the PDMS and the glass failed completely and the PDMS was partially coming off the glass, rendering the chip unusable. The method that we have been using to bond the PDMS to glass was plasma bonding, a standard bonding method widely used in microfluidics. This method works very well for bonding PDMS to plain glass but the gold heater traces have a significant negative effect on the bond strength (though they are only 100nm in height). To resolve this issue, we have explored an alternate bonding method, parylene bonding. This method is not as widely used as diffusion or plasma bonding so there is less available literature on the subject. Parylene is the name given to a family of polyparaxylylene polymers that are typically used as moisture and dielectric barriers. As such, it is often coated on printed circuit boards and, due to its biocompatibility, also on medical devices. This biocompatibility makes parylene a potentially useful material in the fabrication BioMEMS devices. Parylene is typically deposited on the desired surface by chemical vapour deposition, which also results in the parylene polymerization. There are 3 different types of parylene and the one that we were using for the bonding test is Parylene-C. Should the parylene bonding method prove to be successful, it would make a useful new bonding method for the UBC MEMS group. 2.2.4.2 Method First of all, the two surfaces to be bonded are coated with a thin layer of parylene. The coating is done in a parylene coating chamber where parylene is evaporated, directed towards the chamber containing the glass slides and PDMS to be coated, and then the vapour deposits on everything inside the chamber with a uniform thickness. We investigated the bonding process based on available literature and used parameters that other researchers have reported as having produced positive results. The resulting bonding method is as follows. After the coating is finished, the two layers are pressed together in a high temperature environment for 30 minutes. The idea is that during this time the parylene is heated above its glass transition temperature and as a result the chains on the two surfaces interlink, resulting in a bond. The bonding process needs to be done in a vacuum oven as parylene has a tendency to oxidize at high temperatures. Since the Stoeber lab does not possess a vacuum oven, we performed the experiment together with Kevin Heyries, a postdoc in Dr. Hansen’s lab and used the Hansen lab’s vacuum oven and cleanroom. In addition, the plates need to be pressed together quite strongly. The published papers recommend pressures of 0.5MPa - 16MPa. To minimize damage to the chip and the microfluidic channels, we used a pressure at the lower end of the range, 1MPa. We designed and built a  23  thermal press held together by 5 bolts which could be adjusted by a torque wrench to achieve the desired pressure. The temperature of the oven and the baking time were also parameters that needed to be set. According to Noh, Moon et al.14 the ideal temperature is between 160°C and 200°C. In fact, the results were the same within experimental uncertainty for this range of temperatures. Therefore we used a temperature of 160°C. The baking time does not have an effect on the bond strength, as long as it is greater than about 10 minutes. Therefore we baked the chips for 30minutes, then turned off the oven and waited for it to cool down to below 90°C (glass transition temperature of parylene).  Figure 12 - Press used to apply 1MPa pressure during parylene thermal bonding. 2.2.4.3 Results After the parylene was pulled from the oven, it was allowed to cool to room temperature. Then we placed it under the inverted microscope and attempted to run water through it. However, we have observed complete channel collapse as though the channels did not even exist and the water did not flow through the parylene bonded channels. At a pressure of 30psi the bonding failed. Thus parylene bonding has proved to be a method unsuitable for our purposes, as it resulted in a weak bond and complete channel collapse. The channels collapsed due to the high pressure that was applied to the PDMS during the thermal bonding. After all, applying 1MPa of pressure to a small channel made of an elastic material is very likely to squeeze it and bond the top of the channel to the bottom of the channel. Therefore to eliminate channel collapse, we need to lower the applied pressure.  14  H. S. Noh, et al., "Wafer bonding using microwave heating of parylene intermediate layers," Journal of Micromechanics and Microengineering, vol. 14, pp. 625-631, Apr 2004.  24  After the channel failed, we tried peeling the PDMS off the glass. It was quite easy to peel it off near the edges but the PDMS was strongly bonded to the glass in the center. The reason is that when the elastic PDMS was pressed during the thermal bonding, the pressure was the highest in the center of the chip and decreased towards the outside due to elastic strain in the material. This suggests that pressure has a significant effect on the bonding properties, with high pressure being necessary to bond properly. Therefore to increase the bonding strength we need to increase the applied pressure, contradictory to the channel collapse requirement. 2.2.4.4 Conclusion Parylene bonding is not a method that can be used in our project and thus was abandoned. 2.2.5  RTD Tests  2.2.5.1 Introduction The series 4 glass slide heating area comprised of two independent heater systems. These were interdigitated, introducing redundancy as well as the provision to use one as a resistance temperature detector (RTD). The relationship between resistance an temperature is governed by the equation (  )  Where: T = temperature of the device T0 = ambient temperature R = measured resistance R0 = resistance at ambient α = temperature coefficient of resistance, a material property In general, α is typically found in a table. In our case, since the heater is a combination of chromium and gold, α is not well-defined. Thus, to find α, we needed to find a relationship between R, T0 and R0. To achieve this, we used several tests. 2.2.5.2 Methods, Results 2.2.5.2.1 Test 1: Heating one set, detecting with the other In this experiment, we attached one of the heaters to a power supply and HP 34401A DMM. We then varied the power to the heater and measured the change in resistance seen by the RTD. This was done on two slides, one bare (i.e. no PDMS) and one covered with 5psi 15% Pluronic flow. The data from this test is presented in Figure 13 shows that the results were quite different for these two cases. To reduce experimental uncertainty, the experiment was modified to remove variables.  25  0.6  Resistance [kΩ]  0.5 0.4 0.3  Bare Heater  0.2  Covered Channel  0.1 0 -0.1  0  2  4 6 Voltage [V]  8  Figure 13 – RTD resistance vs heater voltage for the bare and covered heater cases 2.2.5.2.2 Test 2: Hotplate heating Instead of heating the slide with the on-slide heaters, we used a Fisher Isotemp hotplate. This hotplate had an integral temperature display but its resolution was ±5˚C and was an element temperature, rather than a hotplate surface temperature. To determine the temperature more accurately, we used the Omega HH23 thermocouple reader and J-type thermocouple measuring the slide temperature. The hotplate temperature was then varied and the thermocouple temperature and RTD resistance were measured for three cases:   Bare heater    Covered heater, no fluid    Covered heater, fluid flowing at 5psi.  The apparatus is shown in Figure 14; the results are shown in Figure 15.  26  Figure 14 - Experimental apparatus, hotplate test 0.7 Resistance [kΩ]  0.6 0.5 0.4  Wet channel  0.3  Bare Heater  0.2  Dry channel  0.1 0 0  20  40  60  80  Temperature [˚C]  Figure 15 - Resistance vs. temperature for the bare, covered dry channel, and covered wet channel  2.2.5.3 Analysis 2.2.5.4 Analysis The goal of the data analysis is to come up with a relationship that allows us to determine the temperature of any RTD we will be using in the future based on its resistance. The relationship between resistance and temperature for and RTD is: ( )  (  (  ))  27  Where R0 is the resistance at the temperature T0 and α is a material property at the temperature T0. As is obvious from the above equation, the relation between resistance and temperature is linear, in agreement with our experimental results. Thus a linear relation has been fitted to the data. The results are: Dry channel:  R(T) = 0.0011960T + 0.3724058 *Ω+  Wet Channel: R(T) = 0.0013129T + 0.3784029 *Ω+ Bare Heater:  R(T) = 0.0016149T + 0.5567747 *Ω+  As mentioned above, R0 and α need to be specified at a certain temperature. For simplicity, this temperature T0 will be 20°C, as most tabulated results are at this temperature. The resulting relationship is: Dry channel:  R(T) = 0.4116058Ω * ( 1 + 0.004761838/°C * ( T - 20°C) )  Wet Channel: R(T) = 0.4046609Ω* ( 1 + 0.003244444/°C * ( T - 20°C) ) Bare Heater:  R(T) = 0.5890727Ω* ( 1 + 0.002741427/°C * ( T - 20°C) )  The resistances at 20°C, R0 are quite variable due to variations that result from the manufacturing of the devices. The evaporation and lift-off of the RTD gold traces is not always reproducible and sometimes even results in unusable RTD’s. However, for our purposes it is not necessary to have exactly the same resistance in all the RTD’s that we produce because we can always measure the 20°C resistance in order to be able to take temperature measurements with our RTD device. The more important constant is α, also known as the temperature coefficient of resistance. Theoretically, α should be a material property and should not be affected by the state of the channel (dry/wet/bare). However, according to our results α varies in between experiments. Therefore in order to use the RTD to measure temperature we need to conduct more hotplate experiments and determine the average value of α. As we improve our experimental technique, we may be getting more and more consistent results. The tabulated value for the temperature coefficient of resistance of gold is 0.003715 /°C which is quite close to the values that we found. However our RTD is not made entirely out of gold; rather it is a 50nm layer of chromium and a 60nm layer of gold. 2.2.6  Electrolysis Test  2.2.6.1 Introduction The unusual results of RTD Test 1 caused us to ask if electrolysis was causing the strange results. Furthermore, throughout the characterizations of devices dating back even to the series 3 fluidic chips, when we have applied in excess of 10V, we have seen the formation of bubbles. Since  28  electrolysis had never been seen in past work by Bazargan et. al15, it was not initially suspected and instead, the effect was attributed to overheating of the fluid to the point of boiling . In theory, electrolysis would happen whenever a potential of 1.23V is created in water. In our experiments, we typically explored voltages between 0-9V and did not believe any electrolysis occurred. 2.2.6.2 Methods To determine at what voltage electrolysis will occur, a power supply was attached to the heater set. Since the magenta side is separate from the blue one, any current would pass through the fluid, allowing for electrolysis. To investigate if the location of electrolysis in the heating zone varies, 3 geometries were tested as shown in Figure 16. For all experiments, the power was varied from 0-6V with 15% Pluronic containing fluorescent beads being pushed through the control channel at 5psi.  (a)  (b)  (c)  Figure 16 – Schematic of the electrolysis test apparatus; a) crossed leads; b) parallel leads; c) conventional connection 2.2.6.3 Results In both apparatus setups (a) and (b), the electrolysis would occur at random locations within the heating zone. Throughout the entire experiment, the current flow would remain constant at 0.045mA. Electrolysis was seen as small bubbles approximately 5 microns in diameter being formed in the control channel at a heater trace. These would first appear at 2.89V. As the voltage was increased, the bubbles would appear more in more locations within the heating zone and with greater diameters. Only above 4V are the bubbles large and widespread enough to be seen under UV illumination. From 2.89-4V, the only way to see the electrolysis is under visible light, which is not typically when characterizing fluidic devices.  15  V. Bazargan, et al., “Flow Control Using a Thermally Actuated Microfluidic Relay Valve”. Journal of Microelectromechanical Systems, 2010.  29  (a)  (b)  (c)  (d)  Figure 17 – Electrolysis seen when the device is connected as in Figure 16 (a); a) before the application of 4V; b) t=2s after power applied; c) t=7s d) t=14s. Recorded using 10x lens. 2.2.6.4 Conclusions Throughout our device characterizations, we have applied voltages far above the electrolysis threshold. Because this effect is not visible under UV light until 4V, we have considered voltages below this level to be safe; however, it is not apparent that the application of greater than 2.89V will cause electrolysis. Since this is not high enough to cause reach the gelation temperature of 15% Pluronic, the heater will need to be electrically isolated or redesigned with lower resistance to keep the applied voltage low.  30  3 Project Deliverables 3.1 Deliverables 3.2 As presented in our Project Charter (Appendix B - Recipe for 10 um SPR220-7.0 Mold for 4-inch Si Wafers Preparation Steps Wafer is cleaned with Acetone, then Methanol, then Isopropanol Alcohol, and then gently blown with N2 gas. Wafer is baked for dehydration for 20 minutes at 200°C, then cooled to room temperature for 10 min. Spin Coating The wafer is centered on the spinner chuck and vacuum sealed. HMDS (hexamethyldisilazan) drops are placed on the wafer using a dropper until 50% of the wafer is covered. HMDS is spun at 3500 rpm for 35 seconds. Let the wafer sit for 1 minute on the spinner before pouring photoresist. SPR220-7.0 photoresist is poured over the wafer straight from the bottle, covering roughly half the area. The bottle should be cleaned thoroughly with a clean wipe before and after pouring. The wafer is spun for 5 sec at 500 rpm and 40 sec at 1500 rpm. The wafer is let sit on the spinner for 1-2 minutes. Photoresist Soft bake The wafer is gradually warmed to 90 C, by using a layer of aluminum foil or wipe on the hotplate. After 1 minute transfer it directly to the 90 C hotplate, let it sit there for 2 min, then transfer to another hotplate at 115 C, let it sit for 3 min. The wafer is slowly cooled to room temperature for 10 min UV-Light Exposure The wafer is loaded into the Canon PLA-501F double-side 100mm mask aligner. The printed mask on a transparent sheet is attached to a thick glass plate and is loaded as the photomask into the mask aligner. UV-light is exposed to the layer for 70 sec. The wafer is let sit on the mask aligner for 1 min. The wafer is cooled down at the room temperature 21°C for 30 min for dehydration. Photoresist Develop An MF 319 bath and a DI-water bath are prepared. The wafer is placed into the MF-319 for 5 min and visually checked. If the developments looks completed, the wafer is placed into the water bath for 1 min and is rinsed with DI water and dried with a N2 gun. The pattern is checked under the microscope and especially the corners and posts are examined for complete development. The developing process can be repeated if additional development is needed.  31  The thickness of the pattern then is measured using the Wyko NT1100 interferometer. Reflow Process The wafer is placed on the hot plate for 2 min at 90°C. The hotplate is set for 140 C and ramps up for 5 minutes, then the hotplate is shut off. The wafer remains on the hotplate for 5 minutes for gradual cooling, then taken off and allowed to cool to room temperature. The shape and the thickness of the pattern then are measured using Wyko NT1100 optical.  32  Appendix C), the following deliverables were agreed upon at the start of the project: 1. A hydrogel actuated microvalve that has a response time of less than 3 seconds. 2. A portable device that allows the use of the fluidic chip independent of fixed power and air systems. 3. A peristaltic pump composed of hydrogel actuated valves capable of moving a cell in a loop. 3.2.1 Deliverable 1: Microvalve with 3 second response time The design of the new Series 4 fluidic chips was completed. Due to ongoing fabrication and Pluronic issues, the performance has not been characterized. At present, we are able to form gel using the PDMS-coated chips, but these devices clog rapidly to allow for sustained characterization. 3.2.2 Deliverable 2: Portable device The hardware for this device is complete. The code that drives the PWM heater controls has been written but due to the unavailability of functioning fluidic chips, it has not been validated. 3.2.3 Deliverable 3: Peristaltic pump Due to ongoing fabrication and Pluronic issues, we have not been able to actuate valves with any degree of reliability. The requirement for 2 adjacent valves to function predictably has not been seen, causing us to not achieve this deliverable.  3.3 Financial Summary 3.3.1 Macro Components The cost of the macroscopic components is as follows: Description  Qty  Vendor  Cost Per  Total Cost  Purchased by  Funded by  Arduino Mini RB-Ard-02  1  Robotshop.ca  28.16  28.16  Project Lab  Stoeber Lab  Pogo Pins  20  Sparkfun.com  0.95  19  Project Lab  Stoeber Lab  9V Battery  1  London Drugs  6.49  6.49  Project Lab  Stoeber Lab  Pressure Reservoir  1  Stoeber  0  0  Stoeber Lab  Stoeber Lab  Pressure Gauge  2  McMaster  10.70  21.40  Project Lab  Stoeber Lab  Table 2 - Macro Component Breakdown  33  3.3.2 Micro Components Description Qty Vendor  Cost Per  Total Cost  Purchased by  Funded by  Cleanroom time  12  UBC  45  540  Stoeber Lab  Stoeber Lab  Transparency  1  CAD/CAS Art Services  120  120  Stoeber Lab  Stoeber Lab  Gold Evaporation  3  UBC  60  180  Stoeber Lab  Stoeber Lab  Table 3- Micro Component Breakdown  3.4 Ongoing Commitments by Team Members The team will continue to collectively put in a sum of 10 hours per week in an effort to complete the items as outlined in Section 5.1.  34  4 Conclusion 4.1 Important Results Below is a list of important discoveries made during the project. 1. Parylene-parylene bonding does not work for PDMS microfluidic devices because the microfluidic devices collapse under the pressure applied to the two PDMS pieces to activate parylene bonding between them. 2. Diffusion of water through the PDMS, and resulting increase in Pluronic concentration in the microfluidic device is not an issue. We never saw the gel stop flowing of its own accord. 3. Spin-coating our glass slides with uncured PDMS at 8000rpm results in a 10um PDMS thickness and eliminates leakage. 4. We can form gel in our devices with 17% Pluronic. 15% Pluronic and below are not good candidates for gel formation. 5. It is possible to achieve heaters with 80nm thickness. Previously, Vahid used heaters with 250nm and there was concern that imperfections would have a dominant role below 250nm and make the heater non-fucntional..  4.2 Project Review We completed two of our three objectives. We complete objective #1, eliminate leakage and objective #3, production of portable microfluidic device. We did not complete objective #2, microvalves with response time of <3 seconds incorporated into a micropump.. We did not complete objective #2 because there were many unforeseen challenges in the fabrication and gel activation process. Four assumptions we made from the outset of the project were: 1. Gel formation could easily be predicted, achieved and reproduced. 2. There would be no electrolysis in the device. 3. We were confident that we had appropriate steps and contingency plans to quickly eliminate leakage in our device. 4. PDMS spun onto a glass slide at high rpm (>1000rpm) would have extremely high porosity due to stretching of the polymer. 4.2.1 Gel formation of Pluronic could easily be predicted, achieved and reproduced We had inconsistent Pluronic behaviour throughout our experiments at both the macroscopic and microscopic level. Pluronic is supposed to reproducibly form a gel at a concentration dependent gel point. We calculated approximately how much power we needed from our heaters to heat up the fluid in our microchannels beyond the gel point. However, the Pluronic seemed to only sometime form a gel despite application of identical parameters. The following three paragraphs describe a situation in which the gel formation seemed to be lost over the period of two hours. In early October we used design iteration 3 and 15% Pluronic to show that after we initially put Pluronic in our devices that we could get formation in both the snake design and dense column  35  designs. We observed that the snake design had better gel formation and pressure holding abilities and concluded that the more surface area the better the gel holding capability of the device. We tested the snake design to 13psi and the gel held for the entire test, the dense columns failed at 7psi. When the snake design device was first connected and the control channel pressure inlet pumped at 10psi we observed gel formation and complete stoppage of flow within one second of turning on the heater at 490mW of power. Within five seconds of turning off the heater there was full liquification of the gel and fluid flow had recovered to its previous high velocity. The device was left in place with Pluronic flowing through the device while other experiments were performed. After two hours had gone by we tried the same experiment with identical pressure and power parameters and could not get gel formation. It was only at a pressure of one psi that we could get gel formation again. We still do not understand why the gel behaviour changed over the course of the two hour experiment. Pluronic from the same bottle was continually cycled through the device and the resistance of the heater and power applied was the same. Inconsistent Pluronic behaviour was also observed at the macroscopic level as well. For example, when we initially placed 15% Pluronic (batch #2) in an oven we observed that from 35C to 45C the Pluronic was a solid gel, we could flip the Pluronic bottle upside down and the Pluronic stayed in place. At 50C the Pluronic became a liquid again. The Pluronic was kept in a tightly sealed bottle overnight and the following day when the same bottle was placed at 40C the Pluronic did not respond as before. It became more viscous but it still continually flowed to the bottom of the bottle. We made more Pluronic mixtures but none of them had the same response we originally observed. When we could not reproducibly get gel formation by heating the Pluronic in the microfluidic channels or by heating the big bottle of Pluronic in the lab oven we started to suspect the Pluronic itself so we proceeded to do viscometry measurements with a viscometer. The viscometer persuasively showed that the 15% Pluronic had a sharp viscosity increase between 32C and 34C. We still cannot explain the inconsistent Pluronic behaviour in which we had gel formation and then a few hours later with identical parameters could not reproduce the same gel formation. Further, we cannot explain why the viscometry shows a sudden spike in viscosity and why the Pluronic in the microfluidic device did not have the same behaviour. The most likely explanation is that the range of gel formation for 15% Pluronic is small and we consistently overheat the Pluronic past the gel formation point. However, this is an unsatisfactory explanation because we have carefully tested the full range of power that can be applied to the heaters. 4.2.2 There would be no electrolysis in the device. Bazargan et al. undertook a similar project in 2008 which included creating a valve with two heaters in a microfluidic device. He heated the Pluronic by applying 25 mW of power (4.29 volts  36  and 5.83mA) across a 250nm thick gold heater. Vahid did not observe electrolysis so we did not expect to have electrolysis. However, we did observe electrolysis and the electrolysis may have contributed to our poor gel performance. 4.2.3  We were confident that we had appropriate steps and contingency plans to quickly eliminate leakage in our device. We could not eliminate leakage in our device by either parylene-parylene bonding or varying plasma treatment parameters for PDMS to glass bonding. Our final solution is to spin on uncured PDMS on a glass slide at 8000 rpm which gave a thickness of 10um and a PDMS to PDMS bonding surface that can hold pressure up to 20 psi and does not have any leakage. We did not consider spinning on PDMS in our initial project proposal because Vahid told us that PDMS spun on at (>1000rpm) would be porous and prone to extensive diffusion. In section 3.4 of our project proposal we proposed two new techniques by which we could eliminate leakage. 1) Collaboration with Dr. Eric Lagally to implement Dr. Lagally’s recently published peptide bonding technique16. We planned to treat the PDMS with (3-aminopropyl)-trimethoxysilane (APTMS; 97%) and the glass and gold with 10% TMS-EDTA. The treatment results in peptide bond formation between the PDMS and glass and PDMS and gold. Philip did one round of unsuccessful tests in late August and we did not choose to pursue this option. 2) Collaboration with Dr. Carl Hansen to implement parylene-parylene bonding. This technique was pursued by Honza and it was declared unsuccessful on November 19th, 2010. We predicted that if either of the two techniques did not work we could further optimize the direct glass to PDMS bonding by reducing the gold heater thickness, modifying the plasma treatment time, modifying the bake time and adjusting the weights we placed on the PDMS after bonding. However, despite an exhaustive round of testing with the PECVD plasma machine we could not get consistent PDMS to glass bonds. Sometimes we had too much collapse and other times we had leakage everywhere. On December 13th, 2010 Mario Beaudoin, UBC Cleanroom Manager, sent an email with attached picture to the UBC cleanroom noting that the inside of the PECVD machine was very dirty. We were never taught how to clean the internal chamber of the PECVD and most of our tests were done before December 13th, 2010 so the dirtiness of the PECVD may have contributed to our poor bonding results.  37  Figure 18 – Dirty PECVD Plasma Chamber 4.2.4  PDMS spun onto a glass slide at high rpm (>1000rpm) would have extremely high porosity due to stretching of the polymer. Vahid told us that PDMS spun on at (>1000rpm) would be porous and prone to extensive diffusion. Thus, we did not consider spinning on PDMS onto our glass slide until early January, 2011 when it was obvious that parylene-parylene bonding and plasma treatment bonding parameter modification had not worked.  38  5 Recommendations The recommendation section is broken into two sub-sections. The first sub-section is the specific recommendations that are directly related to the ongoing project goal of developing a micropump. The second sub-section is more general and includes retrospective recommendations about work flow and other project management strategies.  5.1 Specific Recommendations 1. Increase PDMS spin-on speed. Uncured PDMS is currently spun onto the glass slide at 8000rpm which results in a 10um thickness. 8000rpm is the upper bound of the spin speed of the spinner that is available to us in Dr. Stoeber’s lab. Finding a spinner capable of faster spin speeds will allow us to further reduce the PDMS thickness and increase response time. 2. Test adhesion promoter GE SS4120 in hopes of increasing the pressure that the triple layer PDMS devices can withstand. 3. Measure response time of individual valves. 4. Measure pumping rate and pumping pressure of final micropump. 5. Explore the discrepancy in the viscometry results between the 15% Pluronic that do and do not have fluorescent beads. 6. Repeat viscometry measurements for 17% Pluronic and compare to 15% Pluronic results.  5.2 General Recommendations 1. Recreate Vahid’s results. Use the identical microfluidic and heater design and identical project parameters to recreate Vahid’s results. If we had done this at the outset of the project we would quickly have identified that we needed 17% Pluronic instead of 15% Pluronic. Recreating Vahid’s device will help to establish whether our ongoing poor results are due to a poor design or something that is wrong with the Pluronic solution. 2. Optimization and large scale production of multi-layer microfluidic chips. We should have made three identical control channel wafers and three polyeurathane molds so that we could make 12 microfluidic multi-layer chips at once. We only made microfluidic chips in batches of 4 which proved to be very time consuming.  39  References [1] D. J. Beebe, et al., "Functional hydrogel structures for autonomous flow control inside microfluidic channels," Nature, vol. 404, pp. 588-+, 2000. [2] C. S. Chen, et al., "Shrinky-Dink microfluidics: 3D polystyrene chips," Lab on a Chip, vol. 8, pp. 622-624, 2008. [3] C. J. Easley, et al., "Rapid and inexpensive fabrication of polymeric microfluidic devices via toner transfer masking," Lab on a Chip, vol. 9, pp. 1119-1127, 2009. [4] D. Erickson, et al., "A miniaturized high-voltage integrated power supply for portable microfluidic applications," Lab on a Chip, vol. 4, pp. 87-90, 2004. [5] A. Grimes, et al., "Shrinky-Dink microfluidics: rapid generation of deep and rounded patterns," Lab on a Chip, vol. 8, pp. 170-172, 2008. [6] S. W. Lee and S. S. Lee, "Shrinkage ratio of PDMS and its alignment method for the wafer level process," Microsystem Technologies-Micro-and Nanosystems-Information Storage and Processing Systems, vol. 14, pp. 205-208, 2008. [7] J. C. Lotters, et al., "The mechanical properties of the rubber elastic polymer polydimethylsiloxane for sensor applications," Journal of Micromechanics and Microengineering, vol. 7, pp. 145-147, 1997. [8] A. Manz, et al., "PLANAR CHIPS TECHNOLOGY FOR MINIATURIZATION AND INTEGRATION OF SEPARATION TECHNIQUES INTO MONITORING SYSTEMS - CAPILLARY ELECTROPHORESIS ON A CHIP," Journal of Chromatography, vol. 593, pp. 253-258, 1992. [9] Z. Q. Niu, et al., "DNA amplification on a PDMS-glass hybrid microchip," Journal of Micromechanics and Microengineering, vol. 16, pp. 425-433, 2006. [10] H. S. Noh, et al., "Wafer bonding using microwave heating of parylene intermediate layers," Journal of Micromechanics and Microengineering, vol. 14, pp. 625-631, Apr 2004. [11] E. Ouellet, et al., "Novel carboxyl-amine bonding methods for poly(dimethylsiloxane)-based devices," Langmuir, vol. 26, pp. 11609-14, 2010. [12] K. Pitchaimani, et al., "Manufacturable plastic microfluidic valves using thermal actuation," Lab Chip, vol. 9, pp. 3082-7, 2009. [13] X. T. Qiu, et al., "Localized Parylene-C bonding with reactive multilayer foils," Journal of Physics D-Applied Physics, vol. 42, 2009. [14] B. Stoeber, et al., "Flow control in microdevices using thermally responsive triblock copolymers," Journal of Microelectromechanical Systems, vol. 14, pp. 207-213, 2005.  40  [15] K. Tashiro, et al., Micro flow switches using thermal gelation of methyl cellulose for biomolecules handling. Berlin: Springer-Verlag Berlin, 2001. [16] M. A. Unger, et al., "Monolithic microfabricated valves and pumps by multilayer soft lithography," Science, vol. 288, pp. 113-116, 2000. [17] A. Singhal, et al., “Microfluidic Measurement of Antibody-Antigen Binding Kintetics From Low-Abundance Samples and Single Cells. [18+ V. Bazargan. “Micro Flow Control Using Thermally Responsive Polymer Solutions”. A thesis submitted in partial fulfillment of the requirements for the degree of master of applied science in the Faculty of Grad Studies at UBC. 2008  41  6 Appendices 6.1 Appendix A - AZ 5214E Photoresist Datasheet  42  AZ 5214 E Image Reversal Photoresist  43  GENERAL INFORMATION This special photoresist is intended for lift-off-techniques which call for a negative wall profile. Although they are positive photoresists (and may even be used in that way) comprised of a novolak resin and naphthoquinone diazide as photoactive compound (PAC) they are capable of image reversal (IR) resulting in a negative pattern of the mask. In fact AZ 5214E is almost exclusively used in the IR-mode. The image reversal capability is obtained by a special crosslinking agent in the resist formulation which becomes active at temperatures above 110°C and - what is even more important - only in exposed areas of the resist. The crosslinking agent together with exposed PAC leads to an almost insoluble (in developer) and no longer light sensitive substance, while the unexposed areas still behave like a normal unexposed positive photoresist. After a flood exposure (no mask required) this areas are dissolved in standard developer for positive photoresist, the crosslinked areas remain. The overall result is a negative image of the mask pattern. As everybody knows a positive photoresist profile has a positive slope of 75 - 85° depending on the process conditions and the performance of the exposure equipment (only submicron-resists get close to 90°). This is mainly due to the absorption of the PAC which attenuates the light when penetrating through the resist layer (so called bulk effect). The result is a higher dissolution rate at the top and a lower rate at the bottom of the resist. When AZ 5214E is processed in the IR-mode this is reversed as higher exposed areas will be crosslinked to a higher degree than those with lower dose, dissolution rates accordingly. The final result will be a negative wall profile ideally suited for lift-off. The most critical parameter of the IR-process is reversal-bake temperature, once optimised it must be kept constant within ± 1°C to maintain a consistent process. This temperature also has to be optimised individually. In any case it will fall within the range from 115 to 125°C. If IR-temperature is chosen too high (>130°C) the resist will thermally crosslink also in the unexposed areas, giving no pattern. To find out the suitable temperature following procedure is suggested: Coat and prebake a few substrates with resist. Without exposing them to UV-light subject them to different reversalbake temperatures, i.e. 115°, 120°, 125° and 130°C. Now apply a flood exposure of > 200mJ/cm² and afterwards immerse them into a standard developer make up, i.e. AZ 351B, 1:4 diluted, or AZ 726 MIF for 1 minute. From a part of the substrates the resist will be removed, another part (those exposed to a too high temperature) will remain with the resist thermally crosslinked on it. Optimum RB-temperature now is 5° to 10°C below the temperature where crosslinking starts. The flood exposure is absolutely uncritical as long as sufficient energy is applied to make the unexposed areas soluble. 200 mJ/cm² is a good choice, but 150 - 500 mJ/cm² will have no major influence on the performance. Finally it should be noted that the imagewise exposure energy is lower than with normal positive processes, generally only half of that. So a good rule of thumb is: compared to a standard positive resist process, imagewise exposure dose should be half of that, flood exposure energy double of that for AZ 5214E IR-processing. Once understanding and being familiar with this IR-procedure it is quite simple to set up a different process for liftoff. A T-shaped profile can be achieved by the following process sequence: The prebaked AZ 5214E photoresist is flood exposed (no mask) with a small amount of UV energy, just to generate some exposed PAC at the surface. Now the reversal-bake is performed to partially crosslink this top areas. By this treatment a top layer with a lowered dissolution rate compared to the bulk material is generated. After this the resist is treated like a normal positive photoresist (imagewise exposure and development) to generate a positive image! Due to the lower dissolution rate in the top layer a T-shaped profile with overhanging lips will be the result.  44  AZ 5214E 28.3 24.0 0.76 methoxy-propyl acetate (PGMEA) 0.50 310 - 420 nm striation free 0.1  Solids content [%] Viscosity [cSt at 25°C] Absorptivity [l/g*cm] at 377nm Solvent Max. water content [%] Spectral sensitivity Coating characteristic Filtration [µm absolute]  PHYSICAL and CHEMICAL PROPERTIES  spin speed [rpm]  2000  3000  4000  5000  6000  AZ 5214E  1.98  1.62  1.40  1.25  1.14  FILM THICKNESS [µm] as FUNCTION of SPIN SPEED (characteristically)  Dilution and edge bead removal Prebake Exposure Reversal bake Flood exposure Development Postbake Removal  AZ EBR Solvent 110°C, 50", hotplate broadband and monochromatic h- and i-line 120°C, 2 min., hotplate (most critical step) > 200 mJ/cm² (uncritical) AZ 351B, 1:4 (tank, spray) or AZ 726 (puddle) 120°C, 50s hotplate (optional) AZ 100 Remover, conc.  PROCESSING GUIDELINES  HANDLING ADVISES Consult the Material Safety Data Sheets provided by us or your local agent! This AZ Photoresists are made up with our patented safer solvent PGMEA. They are flammable liquids and should be kept away from oxidants, sparks and open flames. Protect from light and heat and store in sealed original containers between 0°C and 25°C, exceeding this range to -5°C or +30°C for 24 hours does not adversely affect the properties. Shelf life is limited and depends on the resist series. The expiration date is printed on the label of every bottle below the batch number and coded as [year/month/day]. AZ Photoresists are compatible with most commercially available wafer processing equipment. Recommended materials include PTFE, stainless steel and high-density poly-ethylene and -propylene.  45  The information contained herein is, to the best of our knowledge, true and accurate, but all recommendations are made without guarantee because the conditions of use are beyond our control. There is no implied warranty of merchantability or fitness for purpose of the product or products described here. In submitting this information, no liability is assumed or license or other rights expressed or implied given with respect to any existing or pending patent, patent application, or trademarks. The observance of all regulations and patents is the responsibility of the user. AZ, the AZ logo, BARLi , Aquatar and Kallista are registered trademarks of Clariant AG.  Clariant GmbH Business Unit Electronic Materials Rheingaustrasse 190 D-65203 Wiesbaden Germany Tel. +49 (611) 962-6867 Fax +49 (611) 962-9207 Clariant Corporation Business Unit Electronic Materials 70 Meister Avenue Somerville, NJ 08876-1252 USA Tel. +1 (908) 429-3500 Fax +1 (908) 429-3631 Clariant (Japan) K.K. Business Unit Electronic Materials 9F Bunkyo Green Court Center 2-28-8 Honkomagome Bunkyo-Ku Tokyo 113, Japan Tel. +81 (3) 5977-7973 Fax +81 (3) 5977-7894 Clariant Industries Ltd. Business Unit Electronic Materials 84-7, Chungdam-dong, Kangnam-ku  46  Seoul Republic of Korea Tel. +82 (2) 510-8000/8442 Fax +82 (2) 514-5918  47  6.2 Appendix B - Recipe for 10 um SPR220-7.0 Mold for 4-inch Si Wafers Preparation Steps Wafer is cleaned with Acetone, then Methanol, then Isopropanol Alcohol, and then gently blown with N2 gas. Wafer is baked for dehydration for 20 minutes at 200°C, then cooled to room temperature for 10 min. Spin Coating The wafer is centered on the spinner chuck and vacuum sealed. HMDS (hexamethyldisilazan) drops are placed on the wafer using a dropper until 50% of the wafer is covered. HMDS is spun at 3500 rpm for 35 seconds. Let the wafer sit for 1 minute on the spinner before pouring photoresist. SPR220-7.0 photoresist is poured over the wafer straight from the bottle, covering roughly half the area. The bottle should be cleaned thoroughly with a clean wipe before and after pouring. The wafer is spun for 5 sec at 500 rpm and 40 sec at 1500 rpm. The wafer is let sit on the spinner for 1-2 minutes. Photoresist Soft bake The wafer is gradually warmed to 90 C, by using a layer of aluminum foil or wipe on the hotplate. After 1 minute transfer it directly to the 90 C hotplate, let it sit there for 2 min, then transfer to another hotplate at 115 C, let it sit for 3 min. The wafer is slowly cooled to room temperature for 10 min UV-Light Exposure The wafer is loaded into the Canon PLA-501F double-side 100mm mask aligner. The printed mask on a transparent sheet is attached to a thick glass plate and is loaded as the photomask into the mask aligner. UV-light is exposed to the layer for 70 sec. The wafer is let sit on the mask aligner for 1 min. The wafer is cooled down at the room temperature 21°C for 30 min for dehydration. Photoresist Develop An MF 319 bath and a DI-water bath are prepared. The wafer is placed into the MF-319 for 5 min and visually checked. If the developments looks completed, the wafer is placed into the water bath for 1 min and is rinsed with DI water and dried with a N2 gun. The pattern is checked under the microscope and especially the corners and posts are examined for complete development. The developing process can be repeated if additional development is needed. The thickness of the pattern then is measured using the Wyko NT1100 interferometer. Reflow Process The wafer is placed on the hot plate for 2 min at 90°C.  48  The hotplate is set for 140 C and ramps up for 5 minutes, then the hotplate is shut off. The wafer remains on the hotplate for 5 minutes for gradual cooling, then taken off and allowed to cool to room temperature. The shape and the thickness of the pattern then are measured using Wyko NT1100 optical.  49  6.3 Appendix C - Project Charter (No signatures) Project Charter - APSC 459/479, Engineering Physics Project Lab Project Number, Title: 1071, A Micropump Using Thermally Activating Hydrogels Project Summary: The aim of this project is to design and fabricate a thermally activated peristaltic micropump in a monolithic multi-layer polydimethylsiloxane (PDMS) device. This project will incorporate the newly designed micropump into a fully portable microfluidic device that can be used to pump fluid and cells through the chip. The goal is to design thermally activated valves with a response time of <3 seconds and micropumps that can pump fluid at 0.1 nL/sec. Start Date: September 27, 2010  End Date: December 14, 2010  Statement of Deliverables: o A hydrogel actuated microvalve that has a response time of less than 3 seconds. o A portable device that allows the use of the fluidic chip independent of fixed power and air systems. o A peristaltic pump composed of hydrogel actuated valves capable of moving a cell in a loop. Criteria for Success: o A hydrogel actuated microvalve that has a response time of less than 3 seconds. o A portable device that allows the use of the fluidic chip independent of fixed power and air systems. o A peristaltic pump composed of hydrogel actuated valves capable of moving a cell in a loop. o Team members become familiar with state of the art fabrication, bonding and microfluidic device characterization Initial Budget Estimate and Source of Funds: Total expected costs: $928. Source of Funds: Dr. Stoeber and Granting Agencies. For details, see proposal.  50  Project Scope - Activities in Scope o Design, fabrication, characterization of all microfluidic devices supporting portable systems.  Activities out of Scope o None.  Assumptions and Anticipated Risks (for detailed analysis, see proposal) o Peptide bonding and parylene bonding does not result in a leak-free bonding o Eric Lagally Unavailable o PIV does not work o Peristaltic valve kills cells o Water jet cutter is taken out of service o Heating is not consistent across heater o Cannot get thermally activated micropump to work.  Stakeholders: Project Sponsor: Boris Stoeber Team Members: Philip Edgcumbe, Jun Wei Fu, Jan Vohradsky Project Lab: Jon Nakane, Chris Waltham, Bernhard Zender Advisor: Vahid Bazargan Communication and Meeting Schedule: The Team Members will provide written weekly reports on Monday. The Team Members will meet with the Project Sponsor weekly on Tuesday. Other communication with stakeholders and resources by email and telephone as needed.  Other Issues: None.  Project Charter Sign-Off  Name/Date Project Sponsor  Name/Date Project Sponsor  51  Name/Date Team Member 1  Name/Date Team Member 2  Name/Date Team Member 3  Name/Date Project Lab  52  6.4 Appendix D - Team Member's Time Contributions 500 450 400  Hours  350 300  250 200 150 100  50 0 1  2  3  4  5  6  7  8  9  10  11  12  13  Week  Presented in Figure 19 is a graph of the team’s time contributions to the project per week. The cumulative hours contributed is compared to the required 10 hours per week per team member. 500 450 400  Hours  350 300 250 200 150 100 50 0 1  2  3  4  5  6  7  8  9  10  11  12  13  Week  Figure 19 – Team member’s time contributions vs. time  53  6.9 Appendix I - Series 4 Lithography Masks  131  PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCT  PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCT  ODUCED BY AN AUTODESK EDUCATIONAL PRODUC  DUCED BY AN AUTODESK EDUCATIONAL PRODU  6.10 Appendix J - Cleanroom Wafer Fabrication Record  133  Time at Date start of Photore (dd/m process sist m) (24:00)  spr 2202 7.0  spr 2203 7.0 spr 2204 7.0 spr 2205 7.0  SPR 220HMDS spin Time that 7.0 spin (length of wafer is left (length of time @ on spinner time @ speedafter HMDS speedrpm) spin (min) rpm) (spin program mistake)~1 20 @ 1500  Time that wafer is left on spinner after SPR spin Soft bake step 1: length of time (min) @ temperatur e (celcius) Expo sure time (sec)  Developme nt time Developme (time when nt Time glass slide until first or wafer feature are removed Feature readily from height apparent developer) (microns) Comments  70  70  1:30:00 11:00:00  15:00:00  8 - this developed and let off more photoresist faster  3  8  2.5  3  Allign vacuu m needl e press ure  75  2:30:00  Time cooling before exposure (min)  60  Soft bake step 2: length of time @ temper ature celcius  -Made spinner programming mistake so wafer with SPR spun for ~ 6 120secs instead of 40 secs.  10 yes -  60  70  40 @1500  10 yes -  Questions  I fabricated this design with Patrick and we both used the same program… yet his features were 8 microns. The only difference is that I developed my wafer four hours after exposure. I didn’t think this would be an issue because Sarah had told me that some groups wait an entire day before doing development.  40 @ 1500 40 @ 1500  10 yes -  6:00:00  - This was the 2nd wafer in a batch of 3 that Partrick and I made. - First wafer had bubbles form immediately when we placed it on the hot plate. - Put aluminum foil on hot plate to try to increase heat uniformity of hot plate surface. One or two bubbles formed when we placed the wafer on the hot plate. However, we were quite pleased with the wafer; - Each time after pressing mask load and allign we checked to see if the glass slide with the mask on it was held in place. We found that it was not been held in place. However, Patrick said that previously the glass slide was not held in place for him and his mold had still worked out succesfully so we went ahead with the process. - Because we had spun the SPR more slowly (1300 rpm instead of 1500 rpm) we decided to expose my wafer for 75 seconds instead of - Maybe bubbles are from heat 70 seconds. I was very surprised at how well the features were shock? Should we perhaps defined. I could see the pattern on the wafer very easily - far better than transfer the wafers to a hot plate I would have liked. that is at 60 degrees before - Once we had pressed allign we moved the glass slide that was sittingtransfering it to a 90 degree plate? on the mask aligner around to test the vaccuum and even noted that wAllison does something like this. could pivot the glass slide on what appeared to be the wafer. This - Bad idea to soft bake wafers on might be partly why part of the wafer (0.5 cm on the outside part of the aluminum and not increase wafer for 30% of the circumference) did not develop very well. temperature or duration of baking. - Full development took about 6 minutes. - Future plan is to put wafer on - Observation with microscope showed the same problem with the aluminum foil for 30 secs and then columns (30 microns across instead of 10 microns across and poorly remove the aluminum foil and start defined edges) the timer for the regular baking process.  3@ 115 3@ 115 3@ 115  1:00:00  2 2 @ 90  2 2 @ 90  60  8:00:00 35 @ 3500  5 @ 500 40 2 @ 1300 5 @ 500 40 2 @ 1150 5 @ 500 40 2 @ 1100  20  2 1 @ 115  2 2 @ 90  7:30:00 35 @ 3500 yes and the plate was actuall y sucke d onto the 10 top  7:30:00 35 @ 3500  10  7:30:00  spr 2206 7.0 23/06 spr 2207 7.0 24/06 spr 2208 7.0 24/06  3@115  spr 2209 7.0 25/06 Shipley 18-13 25/06  1800  Profile height with alpha profilometer  10  2.5 min  exposure complete at 8:34pm. Note: this heater recipe did not work.  I will let the wafer sit for 90 secs after HMDS and Shipley 18-13 spinnin before moving to the next step. Both HMDS and photo-resist will get a slow spin-on time.  I followed the exact recipe that Jonas gave me. spun HMDS on for 15 secs at 3500 rpm and let it sit for 10 mins. This is because I made a mistake in the programming of the spinner. I went ahead and baked to make the HMDS less dangerous. Let spr sit for 7 min while I waited for the hot plate to cool. I didn't put on enough photoresist so this wafer was essentially waster. A critical part was not covered. I did not expose this wafer.  H5  H4  H3  13  20 5 secs  2.5 min  H2  20 5 secs  - the only reason I had soft bake of 3 @ 90 instead of 2 @ 90 has not changed. - I was doing wide channel fabrication (design C and design F) - I found that the columns in design had the columns but design C on the other end of the wafer did not have the 10 micron columns develop. - I found that the glass slide was suctioned very strongly to the mask aligner today. This is likely because Patrick had cleaned the glass plat 9.5 before we used it today.  2 3 @ 90  H2  Shipley 18-13 #####  2 1 @ 115  2 1 @ 115  no becau se I place d glass slide on dumm y wafer and the mask directl y on top of the dumm y 10 wafer. no ditto 10 above  5 @ 500 40 4 @ 1500  H3  Shipley 18-13 #####  5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500  7:30:00 35 @ 3500  H4 SPR 220-7.0 #####  5 @ 200 5@ 500 1800 35 @ 3500 5 @ 200 5@ 500 1800 35 @ 3500  H5  2  4  3  5  6  8  7  9  Shipley 18-13 #####  Shipley 18-13 #####  Shipley 18-13 #####  Shipley 18-13  Shipley 18-13  H10  H12  H11  H9  H6  5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500  Shipley 18-13 #####  5 @ 200 5@ 500 1600 35 @ 3500 5 @ 200 5@ 500 1600 35 @ 3500 5 @ 200 5@ 500 1600 35 @ 3500 5 @ 200 5@ 500 1600 35 @ 3500 5 @ 200 5@ 500 1600 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 2 40 @ 2500  Shipley 18-13 #####  Shipley 18-13  Shipley 18-13  5 @ 200 5@ 500 35 @ 3500  H8  H14  H13  Shipley 18-13  H7  H15  Shipley 18-13  Shipley 18-13  5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500 5 @ 200 5@ 500 2 40 @ 2500  H17  Shipley 18-13  Shipley 18-13  5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500 5 @ 200 5@ 500 35 @ 3500  H16  H19  H18  2 1 @ 115  no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115 no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above no ditto 10 above  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  2 1 @ 115  H6  2.5 min  H8  H7  glass slides H6-H10 were under-developed. I suspected that they were dirty but it turns out that I was just looking at under-developed heater. I left them over the weekend and exposed them to UV light so it was too late to do anything. Vahid showed me how he could still scratch the heater electrode surface (where there should have been only glass) and thus knew that it was under-developed. I tried putting two of the heaters back in developer but the entire plate was completely cleaned off. Note.  20 5 secs  2.5 min  2.5 min  H10  H9  20 5 secs  2.5 min  H12  H11  2.5 min  2.5 min  2.5 min  20 5 secs  20 5 secs  H19  H18  H17  H14  H15  H13 - sat on hot plate on paper towel for about 90 seconds and then 30 seconds on the hot plate. Soft bake was longer than usual. Slow to develop - I might have spun it at 3500 rpm instead of 2500 rpm. Not sure. When I went to adjust the program for H16 HDMS I found it already at 3500 rpm instead of the 2500rpm I expect. I could have already changed it back. But I doubt it.  H16  2.5 min  2.5 min  - Started spinning it while I was still in the program. It reach 2500 rpm very quickly (about 500 rpm/sec) and then stayed around 2500 rpm for about 5 secs before I stopped it and properly started the program.  2.5 min  20 5 secs  2.5 min  2.5 min  20 5 secs  20 5 secs  2.5 min  2.5 min  - wide channel C and F (only wide channel F can be used because C is missing some of its columns) - snake design C,D,E and F - good, I used this wafer a lot and then I broke it in half. Now only design C and E are still useable.  - accidentally put heater at 110 instead of 115 for 2nd step of soft bake  - accidentally put heater at 110 instead of 115 for 2nd step of soft bake - cleaned - Heater C - snake design - I spun HMDS at 2500 and then again at 3500 (because the first time 2 was too slow, a mistake) - Heater B - snake design - Heater C - snake design: exposed on july 27 at 1:38pm - right most heater doesn't have connection - gold evaporated I mixed up H23 and H29 because their indelible labels washed off in the acetone bath. One of H23 and H29 had a strange development where after the 3rd time of rinsing and sitting in acetone the gold - Heater F - wide channel - no good, heater elements still connected became orange and the heating together elements were much less clear. - Heater A - snake design - good - Gold evaporated - Put in acetone bath on July 27th. Left in bath for 24 hours, no good results. - Heater F - wide channel - good - Gold evaporated - Heater E - snake design - good - Gold evaporated onto this one - First one to develop (evening of July 27th... I didn't have a good scratching tool yet so I destroyed all of the heater connections) - Heater B - snake design - good - Gold evaporated - Put in acetone at 1:10pm on July 29 I mixed up H23 and H29 because their indelible labels washed off in the acetone bath. One of H23 and H29 had a strange development where after the 3rd time of rinsing and sitting in acetone the gold - Heater C - snake design - new S1813 and did not clean slide - good became orange and the heating - Gold evaporated elements were much less clear.  h28  H27  H26  H25  H24  H23  H22  H21  H20  20 5 secs  - HMDS sat on wafer for 10 min instead of the usual 1 min because I found the SPR 220-7.0 lip was covered in dried up very carefully. PDMS really stuck to this wafer.  20 5 secs  20 5 secs  20 5 secs  20 5 secs  20 5 secs  Gold element: 0.3 microns  20 5 secs  65  65  65  2  2  3@110  65  10@500 2 40@1500  10@500 10 40@1500 10@500 2 40@1500 10@500 2 40@1500  2 2@90  2  10@500 23:00:00 40 @ 3500  10@500 19:00:00 40 @ 3500 10@500 19:00:00 40 @ 3500 10@500 23:00:00 40 @ 3500  10@500 2 40@1500  7:30:00  10@500 17:00:00 40 @ 3500  3@110  65  SPR 14 220-7.0 19/07  2 2@90  6:30:00  10@500 2 40@1500  65  10@500 17:00:00 40 @ 3500  2  SPR 13 220-7.0 15/07  SPR 10 220-7.0 15/07 SPR 11 220-7.0 15/07 SPR 12 220-7.0 15/07  saw needl e, no vacuu m on 5 plate saw needl e, no vacuu m on 5 plate  SPR 15 220-7.0 19/07  3:45:00  2:30:00  (3 to 7 mins)  7:00:00  (3 to 7 mins)  3:00:00  10:00:00  20 5 secs  H21 Shipley 18-13 26/07  10@500 11:30:00 40 @ 3500  20 5 secs  H22  Shipley 18-13 27/07  10 yes  25  (3 to 7 mins)  H23  10 yes  10@500 40@1500 ~2 10@500 40@1500 ~2  yes  20  (3 to 7 mins)  10@500 40@1500 ~2 10@500 40@1500 ~2  1@115  10@500 14:00:00 40 @ 3500 10@500 14:00:00 40 @ 3500  1@115  yes  20  20  10@500 11:30:00 40 @ 3500 10@500 11:30:00 40 @ 3500  25  20  1@115  Shipley 18-13 26/07 Shipley 18-13 26/07 10@500 40@1500 ~2  1@115  yes  yes  Shipley 18-13 27/07 Shipley 18-13 27/07  yes  yes  H20  10@500 14:00:00 40 @ 3500  10@500 40@1500 ~2  1@115  1@115  H25  1@115  1@115  H24  10@500 40@2000 ~2  10@500 40@1500 ~2  25  10@500 11:30:00 40 @ 3500  yes  10@500 11:30:00 40 @ 3500  1@115  Shipley 18-13 27/07  10@500 40@2000 ~2  Shipley 18-13 27/07  10@500 14:00:00 40 @ 3500  H27  Shipley 18-13 27/07  H26  h28  11  10  13  12  14  15  h29  spr 2207.0 30/07 Shipley 18-13 30/07 spr 2207.0 30/07  30/07 30/07 SPR 220-7.0 30/07  Shipley 18-13 30/07  Shipley 18-13 29-07 Shipley 18-13 29-07  spr2207.0 29-07  spr2207.0 29-07  Shipley 18-13 27/07  10@500 15:00:00 40 @ 3500 10@500 15:00:00 40 @ 3500 10@500 15:00:00 40 @ 3500 10@500 18:30:00 40 @ 3500 10@500 18:30:00 40 @ 3500 10@500 18:30:00 40 @ 3500 10@500 18:30:00 40 @ 3500  10@500 15:00:00 40 @ 3500 10@500 15:00:00 40 @ 3500 10@500 15:00:00 40 @ 3500  12:30:00 12:30:00 10@500 12:30:00 40 @ 3500  10@500 12:30:00 40 @ 3500  10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500  10@500 17:30:00 40 @ 3500  10@500 17:30:00 40 @ 3500  10@500 14:00:00 40 @ 3500  10@500 1 40@4000 10@500 1 40@4000  10@500 1 40@4000 10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500  10@500 1 40@4000 10@500 1 40@1500 10@500 1 40@4000  10@500 1 40@4000  10@500 1 40@1500  10@500 40@1500 10@500 40@1500  10@500 40@4000  10@500 40@4000  10@500 40@2000 ~2  1 2@115  1 1@115  1 2@115  1 2@115  1 1@115  1 1@115  1 1@115  1 1@115  1 1@115  1 1@115  1 2@115  1 2@115  1 1@115  1 2@115  1  1  1 1@115  1 1@115  1 1@115  2@115 (not very sure might have been 1 1@115) 2@115 (not very sure might have been 1 1@115)  60 min  60 min  60 min  60 min  60 min  60 min  60 min  60 min  60 min  60 min  10+  10+  10+  10+  10+  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes  yes yes  yes  yes  yes  yes  yes  yes yes  50  50  50  50  50  50  53 10s  25  25 10s  25 10s  25  25  50 55 (acci dent)  25  50  50  22  28  28  60  60  25  <2  <2  <2  <2  <2  <2  2:05:00  2:00:00  1:40:00  1:40:00  2  2  1:50:00  2:00:00  4:40:00  4:40:00  - only waited 10 secs before removing glass slide from mask alligner - 40 mins bw exposure and development - top 3 are good, bottom three are merged... passable but far from ideal  - dirty specs after spinning - started development 40 mins after exposure - all good except for top right heater - bake time questionable - a bit over 2 min - started development 40 mins after exposure - all good - spun on HMDS at 4000rpm by mistake - waited 35 mins after exposure before development - bottom left heater = good - centre left heater = ok - centre right heater = good - top three heaters are all merged together  - snake design E - heater snake design D - quick check = good - heater snake design C - - quick check = good - heater L5 - quick check = good - heater wide channel A - quick check = ok to good - heater L5 - exposure of 53 was an accident. It should have been 50 - quick check = good - heater wide channel A - quick check = ok to good  - snake design B  - snake design C - very good  - Wide channel F - Snake design A - Ok  - Heater C - snake design - new S1813 and did not clean slide - good - evaporated gold onto this but I placed it upside down in the evaporator so the gold was evaporated onto the wrong side so I can't use it.  H30  ##### 10@500 18:30:00 40 @ 3500 10@500 18:30:00 40 @ 3500  1 2@115  60 min  50  (3 to 7 mins)  H31  #####  1 2@115  60 min  50  1@115  H33  spr 2207.0 Shipley 18-13 Shipley 18-13 Shipley 18-13 Shipley 18-13 Shipley 18-13 Shipley 18-13  ##### 5 secs @500rpm 40 secs @ 1 1500rpm 10@500 1 40@4000 10@500 1 40@4000 10@500 1 40@4000  1 2@115  60 min  h37  h35 h36  h38 h39 h40  h41 h42 h43 H44 H45 H46  h52  h55 h56  30/07 30/07 ##### ##### #####  1 1@115  50 50  50  7  2:30:00  - spun on spr at 2000rpm by mistake - two of three of top heaters are merged together so I can't use them. Bottom heaters are good. - started development 16 mins after exposure  - dirty, can't use - spun on HMDS at 4000rpm by mistake - dirty can't use  - dirty specs after spinning  9.6 - Design 3A,3B,3C and 3D  - wide channel A - good results - Wide channel F - good - Note: for H38-H43 I was really careful to make sure to wait 10 mins for slides to cool after I dehydrated them after washing, I waited 10 mins after soft bake and 20 mins after exposure. All of the slides (H38-H43) worked well. I'm hesitant to attribute the success to the extended waiting times. Of H30-H37, H35 and H36 were mixed up. I suspect that H30 and H31 got 1@115 instead of 2@115 and. That leaves H32 (terrible - s1813), H33 (good -s1813), H34 (terrible - s1813), H37 (good shipley 220-7.0). I can't explain why H32 and H34 did not work out... maybe it was the wait times.  - Heater A wide channel - over-developed or over-exposed. There seems to be more heater than their should be and the corners are not very sharp. - I planned to do soft bake of 2@115 but I might have done 1@115 out of habit... this might explain the poor results. - Heater A wide channel - did not use the 10 secs on cloth before soft bake for this glass slide (accident) - terrible features, can barely make out the elements, bubbly - Heater B snake design - ok to good result - wide channel A - very very blurry features, cannot even see elements. Eric Ouelette (student in Lagally lab) suggested it was because the mask and photoresist on the glass slide (4cm x 7.5cm) were not making contact. However, he agreed that I was setting up the glass slide on the mask alligner correctly. He also suggested that I use MF-24A instead of MF319 for development. Vahid said that MF-24A and MF-319 are quite similar and that it was not important. H35 and H36 got mixed up... I exposed one of the slides twice and the other not at all. I only realized the mistake once I put the slide I had not exposed in the developer and nothing happened.  Note for H30-H33: I cleaned the glass slides with acetone, IPA, water, N2 blowing, 5 mins at 110, 10 mins cooling. But they still looked quite dirty. - Heater E snake design - over-developed or over-exposed. There seems to be more heater than their should be and the corners are not very sharp - I planned to do soft bake of 2@115 but I might have done 1@115 out of habit... this might explain the poor results.  h34  spr  10@500 1 40@4000  1 2@115  60 min  H32  H47  spr 5 secs @500rpm 35 secs 9:45:00 @3500rpm 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500  1 2@115  30/07  h48  10@500 17:30:00 40 @ 4000  10@500 1 40@4000  1 2@115  h57  yes  2  spr 22016 7.0 ##### 16/8/2 h50 s1813 010 16/8/2 h51 s1813 010 spr 220- 16/8/2 7.0 010  h49  spr 220- 16/8/2 7.0 010  10@500 17:30:00 40 @ 3500  10@500 1 40@4000 10@500 1 40@4000 10@500 1 40@4000 10@500 1 40@2000  60 min  65 1min  spr 220- 16/8/2 7.0 010  10@500 17:30:00 40 @ 4000 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500  1 2@115  20 yes  h53  16/8/2 010 16/8/2 010 16/8/2 010 16/8/2 010  10@500 1 40@4000  60 min  h54  spr 2207.0 spr 2207.0 spr 2207.0 spr 2207.0  10@500 17:30:00 40 @ 3500  3@115  h58 spr 220- 16/8/2 7.0 010  1 2@90  h59  h29  spr 2207.0 #####  spr 2207.0 #####  spr 22019 7.0 #####  spr 22018 7.0 #####  spr 22017 7.0 #####  10@500 21:00:00 40 @ 3500  10@500 21:00:00 40 @ 3500  10@500 21:00:00 40 @ 3500  10@500 21:00:00 40 @ 3500  5 secs @500rpm 40 secs 17:10:00 @3500rpm  5 secs @500rpm 40 secs 17:10:00 @3500rpm  5 secs @500rpm 40 secs 16:10:00 @3500rpm  10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500 10@500 17:30:00 40 @ 3500  10@500 1 40@4000  10@500 1 40@4000  10@500 1 40@4000  10@500 1 40@4000  10@500 1 40@4000  10@500 1 40@4000  5 secs @500rpm 40 secs @ 1 1500rpm  5 secs @500rpm 40 secs @ 1 1500rpm  5 secs @500rpm 40 secs @ 15 1500rpm  10@500 1 40@4000 10@500 1 40@4000 10@500 1 40@4000 10@500 1 40@4000  1 2@115  1 2@115  1 2@115  1 2@115  1 2@115  1 2@115  1 2@115  1 2@115  1 2@90  1 2@90  1 2@90  1 2@115  1 2@115  1 2@115  1 2@115  3@115  3@115  3@115  10+  10+  60 min  60 min  60 min  60 min  yes  yes  yes  yes  50  50  50  50  1:40:00  1:40:00  1:40:00  2:00:00  16/8/2 010 16/8/2 010 16/8/2 010 16/8/2 010  H63  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  1 2@115  spr 220h60 7.0 spr 220h61 7.0 spr 220h62 7.0 spr 220h62b 7.0  H64  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  1 2@115  10+  10+  - top left 2 heaters are good the rest all have some merging making them less than ideal - 40 mins sitting out between exposure and development - all good - bottom three are good - one of three of top one are good - bottom three are good - one of three of top one are good  - turned on UV light at 4:20pm - had long delay between HMDS and SPR 220-7.0 application because couldn't open the spr container and spent a lot of time cleaning the lid. - when I took wafer out of spinner after spinning on spr it looked really good and smooth. I put the wafer onto 1 sheet of paper towel that I had just placed on the heater surface. In the first four seconds on the paper towel on the surface four bubbles appeared on the wafer. Also, I noticed that the reflection of light off the wafer also changed while it wa been heated. Instead of of clean reflection all the way across it started to have islands (some large some small) of reflection. - at 4:42pm I took wafer off 115C and left it to cool. - let sit for 7 minutes between exposure and development - looked good under the microscope after exposure. Wei Wei noted a few dark streaks that almost went all the way across the channels. Philip is not sure what it is but feels that it is a small imperfection and that the fabrication process is good so we should continue with the other two. - at 6:20pm started reflow. Put 4 time folded napkin flat on 90C hot plate immediately before placing wafer on napkin. Let sit for 1 min. Removed napkin and left on hot plate until 90C for 5 min. Ramped temp up to 140C over a period of 5 min with wafer on hot plate. Left on hot plate at 140C for 5 mins then turned off heater and left wafer on heater for 10 minutes.  2  1  55  2  65  - two spots after spinning - three more spots appeared during heating - saw a spec of dust on it when we were about to do exposure... not sure where the dust came from.  55  2  -23 min after exposure - no blemishes, medium dirty - H4L - 45 min after exposure - quite clean but blemishes at 3,6 and 8 - H4S Didn't take the slide off the paper towl on the oven for 2 minutes. Bake it for an extra 1 min to compensate. - wiped down transparency before exposure - generally clean. Slight blemish on heater 1. - I rinsed with DI water for extra long time. - 8 min after exposure - H4W - accidentally left it on uv for 5 extra secs - 16 min after exposure - extra rinse. generally clean. 5 is slightly suspect. - All heater conduct. Heater 4,5,6,9 and 12 did not have good lift-off. 8 has small part that didn't come off for lift-off. I will use this slide for plasma bonding and place pdms on heaters 7-12. - H4L - very clean, very good. Slight blemish at 11. I rinsed this one with water after development for an extra long time. - 45 min after exposure -Developed before exposure. washed with DI water, then exposed. - 14 mins after exposure - extra rinse. generally clean. No blemishes. - All heater conduct and have resistance of approx. 300 ohm. Heaters 6,8,9,11 and 12 didn't have good lift-off - 7 min after exposure - extra rinse. generally clean. Number 12 has hair lying across it. - All heater traces conduct except for #12.  1 6min  60  2  -H4L  65  no ditto above no ditto above no ditto above  55  2  - wei wei did this process on his own with Philip's supervision - perfect wafer. No specs after spinning or after soft bake. - this wafer ended up been a write-off because we put the glass slide down with the transparency facing up instead of sitting directly on top of the wafer. This meant that the exposure pattern was way off. We had to throw out this wafer. - H4S - accidentally exposed for 55 secs. I meant to expose for 50s - after exposure and development the slide looks good. There is one imperfection at heater 11. Contact might be broken. - defined new convention for heaters. There are 12 heaters per slide and we number them in rows of 3 going from left to right. - H4L - two specs after spinning. - - generally good after development. Slightly concern that the glass and photoresists seem quite dirty. Lots of 5 micron specs. Only heate 11 has a piece of spec big enough that makes me think it might not form a full heater.  no ditto above  55  2  55  no ditto above  55  7um  no ditto above  55  7um  no ditto above no ditto above no ditto above  2  7  H65  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  6min  H66  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  65 40sec  H67  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  no can see vacuu m is on (main vacuu m displa 30 y) no can see that vacuu m is on thoug 25 h no can see that vacuu m is on thoug 12 h  H68  spr 2207.0 #####  10@500 21:00:00 40 @ 3500  2 min  H69  spr 2207.0 #####  55 20 secs  H70  spr 2207.0 #####  no ditto 20 above  H71  10+  H72  10  n/a  n/a  2  n/a  3@115 n/a  65  n/a  1 2@90  no,  n/a  5 secs @500rpm 40 secs @ 1 1500rpm  3@115 10+  5 secs @500rpm 40 secs 22:00:00 @3500rpm  1 2@90  spr 22020 7.0 #####  5 secs @500rpm 40 secs @ 2 1500rpm  5 secs @500rpm 40 secs 22:00:00 @3500rpm  10 10.2um  spr 220- 27/12/ 21 7.0 2010  2  - Wafer came from box of wafers in bioMEMS lab. The box was no longer sealed and left out in bioMEMS so there are specs of dust on the wafer. We used the cleanroom microscope and found a waver with what appeared to be no dust on it. We scanned the entire wafer with the 5x objective and saw nothing. After pouring and spinning on the PDMS and placing the wafer on several layers of paper towel on the 90C heater we found that at least 50 bubbles appeared on the wafer. This must be because of the dirtiness of the wafer because wafers 17, 18 and 19 did not have this problem and they come from the box of wafers that were kept in the cleanroom. - As per Ben Mustin's suggestion I cleaned the wafer with acetone, ipa and methanol, blow dried it and let it sit for 45 minutes on the hot plate at 200C and then cool down at room temp at 200C. Ben suggested I l it sit for 10 minutes but I had a long conversation with Sultan so the wafers sat for 45 minutes on the hot plate. - No specs appeared during heating, good final result. - As per Ben Mustin's suggestion I cleaned the wafer with acetone, ipa and methanol, blow dried it and let it sit for 45 minutes on the hot plate at 200C and then cool down at room temp at 200C. - Only shook developing try every three minutes. This may explain why the two left fluidic channels do not have well defined edges. - this is a reject because photoresist didn't cover entire glass slide. but went ahead to see how exposure works - happy with developer results. Design looks very good. heater trace #2 has one part with slightly rough edges but should still work fine. Maybe slightly overdeveloped - valleys for heater traces are a little fat. - happy with developer results. Design looks very good. heater trace #2 has one part with slightly rough edges but should still work fine. Sharp features, sharper then H74. This is strange because both had 25 second exposure and both were in same developer bath for same length of time +/- 5 secs - more spots on this one - i suspect the spots come from the permanent marker dissolving. I could see the spots in the solution. I should use fine tip marker to minimize this problem.  55  25  2  2  2  1  - all good after development - generally good except for: 4,5, 8,12. All the bad ones have dirt on them. Maybe this is because I didn't use a new developer. I re-used the developer bath from H81, H82 - batch 2  - all good after development  - all good. - I feel this is the best glass slide i have. I was agitating it the whole time but i only let if sit in developer for 1 minute instead of two. The features are much more sharp and the gold traces will be spaced farther apart from each other. I stopped developing a little before the last of the developer came off.  - might have some water with photoresist because i cleaned the pipette bulb with water after it got some photoresist in it and when I next used it (on this heater) some water came out. - happy with results. All good - photoresist does not cover glass slide entirely so won't use. One of the heaters also has something on it which spans three traces. - good. I find that it is slightly overdeveloped. I worry that gold traces are not spaced far enough apart.  55  - all good - batch 3 - all good, cleanest glass slide by far. This is interesting because I did not use permanent marker. I suspect that permanent marker contributes to a dirty glass slide. Surprisingly, 1-6 looks slightly overdeveloped whereas 7-12 looks really good. It's hard to understand how one half of a slide can be over developed. - batch 4 - all good, cleanest glass slide by far. This is interesting because I did not use permanent marker. I suspect that permanent marker contributes to a dirty glass slide - batch 4  - generally good except for: 8 (but its a minor blemish), and 6 (minor blemish) - still go ahead and develop. - batch 3 - generally good except for: 6,7,8,10. All the bad ones have dirt on them. Maybe this is because I didn't use a new developer. I re-used the developer bath from H81, H82 - batch 2  55  3 2um  65  1.5  no,  1 1@115  3@115 10+  1 1@115  1 2@90  5 secs @500rpm 40 secs @ 2 1500rpm 10@500 1 40@1500  25  10@500 1 40@1500  25  5 secs @500rpm 40 secs 22:00:00 @3500rpm 10@500 1600 40 @ 3500  no  10@500 1600 40 @ 3500  no  spr 220- 27/12/ 22 7.0 2010 29/12/ 2010 10+  29/12/ 2010  s1813  1.5  s1813  25  H73  no  H74  1 1@115  29/12/ 2010  10@500 1 40@1500  s1813  10@500 1600 40 @ 3500  H75  1 1@115  s1813 2  H76  2  10+  25  10+  no  s1813  1 2@115  10+  1 1@115  H77  10@500 1600 40 @ 3500 10@500 1600 40 @ 3500 10@500 1600 40 @ 3500 10@500 1600 40 @ 3500  1 2@115  2  29/12/ 2010 29/12/ 2010 29/12/ 2010 29/12/ 2010  10@500 1 40@4000  10@500 1 40@1500  1 2@115  25  s1813  10@500 1600 40 @ 3500  10@500 1 40@4000  10+  25  s1813  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  10+  no  H78  29/12/ 2010  10@500 21:00:00 40 @ 3500  1 2@115  no  H79  spr 2207.0 #####  s1813  10@500 21:00:00 40 @ 3500  1 2@115  1 1@115  H80  spr 2207.0 #####  10@500 1 40@4000  1 1@115  H81  spr 2207.0 #####  10@500 1 40@4000  10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500 10@500 1 40@1500  H82  10@500 21:00:00 40 @ 3500  1 1@115  H83  10@500 21:00:00 40 @ 3500  10+  2  spr 2207.0 #####  10+  2  spr 2207.0 #####  1 2@115  10+  55  H84  1 2@115  55  H85 10@500 1 40@4000  1 2@115  2  10@500 21:00:00 40 @ 3500  10@500 1 40@4000  55  spr 2207.0 #####  10@500 1 40@4000  no no ditto above no ditto above no ditto above no ditto above no ditto above no ditto above  H86  10@500 21:00:00 40 @ 3500  2  10@500 21:00:00 40 @ 3500  2  spr 2207.0 #####  55  spr 2207.0 #####  55  H87  no ditto above no ditto above  H88  6.11 Appendix K - Glass Slide Record  139  12B  SD-D  PDMS design  D  D  Single (S) or double (D) layer PDMS - if D include the thin layer spin speed  22/7/2010  20/7/2010  CT: 48 hours in the oven B  B  Post-bonding treatment. Date that glass Cooking time (CT) Plasma cleaning and PDMS bonded and Weight (W) on treatment. BioMEMS together device (B) or cleanroom (C)  Heater #  9 10  SD-B  2 3 4 5 6 7 8  11  D  3C and 3D.  D- 3000rpm  D -3000rpm  17/8/2010  W: ?  W: cleanroom 20/8/2010 phone (5min) W: two scrap aluminum blocks in bioMEMS CT: 4 hours in 23/8/2010 bioMEMS oven W: 8lbs brick CT: 14 hours in BioMEMS 65deg C 2000 3/10/2010 oven W: 8lbs brick CT: 14 hours in BioMEMS 65deg C 2000 3/10/2010 oven - pressed freshly plasma treated pdms with backend of tweezers quite hard. This likely contributed to channel collapse. The glass slide was cracked when it was removed from the oven. This chip either had an 8lb or 3lb weight on it. I (Philip) am 90% sure it was 8 8/10/2010 lb.  C - direct placement Vahid's recipe  C - direct placement Vahid's recipe  C - direct placement Vahid's recipe  B - direct placement (between 30secs and 1 min) - 45 seconds of plasma  C - direct placement Vahid's recipe  C - direct placement Vahid's recipe  C: Vahid recipe. Glass aligning (GA), wait time (WT) #1: No GA, no WT #2: No GA, yes WT #3: Yes GA, yes WT #4: Yes GA, yes WT  B  SD-A 13 14 15  D - ? rpm  22  3A and 3B  W:Cleanroom 10/8/2010 phone 12/8/2010  Two good snake designs cut into two pieces each  16 Vahid's old heater 17 18 19  23 H62  Snake Design C  3A (ok to bad allignment) and 3B (good allignment) D - 2700rpm  24 H41  3D  20 H46 21  25 blank glass slide  3B and 3C  D- 2000  26 H53  Pressure where leakage occurs across gold (psi)  10-12psi  13psi  10  Pressure where leakage occurs between PDMS and glass (psi)  25+  Comments  - very dirty surface => long strands of hair. - right valve has channel collapse and the membrane bonding to the glass slides. - cannot use this chip. - I suspect that there are two slide 9s  - I placed the pdms on the activated glass surface and then picked it up to reallign it. - Only the right valve has a connected heater and it leaks in many places. It would take several minutes to fill actual channels in the valve because there is so much leakage. - Right valve has gold heater that is very poorly defined and has hair on the slide too... but for some reason there is no leakage to 20 psi. Small leakage started around 20 psi.  This PDMS was suspect because it did not cure properly.  - Looks dirty. This is frustrating because I used tape to remove dust from PDMS and I used compressed air to remove dust from the heaters before plasma treatment so I expected the bond to be clean. Brought PDMS slides from cleanroom to BioMEMS lab for ~5min before brick was placed on. Heater was not used previously because of gold specks on surface. Tested it and leakage did not start until 13psi over very thick gold surface of 13psi.  This is the good slide with the new design that is my best hope yet in terms of flow stoppage. There is leaking along the gold both between the interior channels and side channel (that is there for allignment puropose) and the inside channel and the coolant channel. .  Leakage and very dirty wit lots of hair on it. When I was hole punching the device I leaned over the exposed devices and maybe 20 that’s when some hair fell down?  Did not see any leakage. Highest pressure we went to was 21 psi.  - Wei Wei mixed the PDMS for this device. M40 - slide 26 movie was created for design 3C that shows that dense columns do not provide good holding point for gel. - Design 3C, Snake Design, has serious membrane collapse. The main channel leaving the pressure outlet is 90% collapsed and the solution needs to go a long way along the collapsed channels before getting into the heater element of the device. Design 3C also has a glass crack that runs across it. - Design 3B has dense columns and we showed it is not very good at holding a gel.  32 blank glass slide  31 blank glass slide  30 blank glass slide  29 blank glass slide  28 blank glass slide  27 blank glass slide  SD-F  SD-C  SD-E  3D  3B  3C  3A  PDMS design  single layer, 10:1  single layer, 10:1  single layer, 10:1  single layer, 10:1  single layer, 10:1  single layer, 10:1  single layer, 10:1  single layer, 10:1  Single (S) or double (D) layer PDMS - if D include the thin layer spin speed  23/10/2010  23/10/2010  23/10/2010  23/10/2010  21/10/2010  21/10/2010  21/10/2010  21/10/2010  No finger pressing, bond and then weight only. W: 8lb CT: 24 hours in bioMEMS  No finger pressing, bond and then weight only. W: 0lb CT: 24 hours in bioMEMS  No finger pressing, bond and then weight only. W: 8lb CT: 24 hours in bioMEMS No finger pressing, bond and then weight only. W: 3lb CT: 24 hours in bioMEMS  5 mins at room temp then. W: 8lb brick, CT: 24 hours in bioMEMS 65C oven  10 mins at room temp then. W: 0lb, CT: 24 hours in bioMEMS 65C oven  10 mins at room temp then. W: 3lb metal block, CT: 24 hours in bioMEMS 65C oven  10 mins at room temp then. W: 8lb brick, CT: 24 hours in bioMEMS 65C oven  C - direct placement power: 30, pressure: 500, temp:25, O2:100, time:10secs No gold on chip  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  C - direct placement power: 30, pressure: 500, temp:25, O2:100, time:10secs No gold on chip  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  Post-bonding treatment. Date that glass Cooking time (CT) Plasma cleaning and PDMS bonded and Weight (W) on treatment. BioMEMS together device (B) or cleanroom (C)  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs No gold on chip  Pressure where leakage occurs across gold (psi)  33 blank glass slide  SD-D  Heater #  34 blank glass slide  Pressure where leakage occurs between PDMS and glass (psi) Comments Never saw leakage. Tested until 30 psi. The - I gently touched the entire PDMS surface with my finger until I collapsed pdms could see the plasma bond starting to occur. I did not exert more stayed collapse so than half a pound of force with the flat of my finger so the chip felt there wasn't much at most the pressure equivalent to a 3 lb weight. room for flow. - Full channel collapse after 24 hr in oven. Never saw - I gently touched the entire PDMS surface with my finger until I leakage. Tested could see the plasma bond starting to occur. I did not exert more until 30 psi. Very than half a pound of force with the flat of my finger so the chip felt fast flow because at most the pressure equivalent to a 3 lb weight. no channel - Very little channel collapse after 24 hr in oven (at first I thought collapse in the there was none). A tiny bit of collapse on 1/6 of chip (one side of fluidic channel. one of the valves) Maybe saw leakage at 25psi. Difficult to tell because flow resistance is so low that the fluorescent dye - I gently touched the entire PDMS surface with my finger until I reservoir is used could see the plasma bond starting to occur. I did not exert more up at 30 psi in only than half a pound of force with the flat of my finger so the chip felt a few seconds. at most the pressure equivalent to a 3 lb weight. Generally, good - Only collapse that occured after 24 hr in oven was in the coolant leakage protection. channels. - So much channel collapse that I can not get any flow in the chip. I didn't see any leakage... but that's of limited - I gently touched the entire PDMS surface with my finger until I value because no could see the plasma bond starting to occur. I did not exert more fluid left the than half a pound of force with the flat of my finger so the chip felt pressure inlet at most the pressure equivalent to a 3 lb weight. area. - - Full channel collapse after 24 hr in oven. - For slides 31-34 I cleaned with acetone, IPA and water and 10min at 115C and 10 min at room temp before bonding. I placed the PMDS on glass as soon as they came out of the PECVD and I could see the bond occuring. I did not put any pressure on the PDMS immediately after placing the PDMS on the glass with my finer like i did previously. I instead put the respective weights onto the PDMS. This took about two minutes after bonding because the weights were in the gowning room and I had to place the pdms on the glass first. - Collapse everywhere. I could only get flow in one of the three valves because there was so much collapse in the pressure inlet. In the channel that I did get flow it was only temporary because the pressure outlet was blocked. There was no leakage up until 20 psi which is where I stopped testing. - Tested to 20 psi and saw no leakage.  - Tested to 20 psi and saw no leakage. - Collapse everywhere - Tested to 20 psi and saw no leakage in top left valve (M71). - Tested and saw major leakage from one part of channel to another at 5 psi in bottom right valve (M72) - Collapse in many places - Tested to 20 psi on left (M73) and right channels and saw no leakage. There is significant collapse at pressure outputs (near bottom) so flow can not leave the channels. - Collapse in many places  3C and 3D  PDMS design  single layer  D- 3000rpm  Single (S) or double (D) layer PDMS - if D include the thin layer spin speed  31/10/2010  31/10/2010  25/10/2010  no finger pressing c: 8lb/2 CT: 24 hrs in bioMEMS  No finger pressing, bond and then weight only. W: 8lb/2 (I had the brick lying on both chips) CT: 24 hours in bioMEMS light finger pressing c: 3lb CT: 24 hrs in bioMEMS  C - direct placement power: 30, pressure: 500, temp:25, O2:100, time:10secs - no gold  C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs  Post-bonding treatment. Date that glass Cooking time (CT) Plasma cleaning and PDMS bonded and Weight (W) on treatment. BioMEMS together device (B) or cleanroom (C)  Heater #  36 blank glass slide  single layer  20/11/2010  - No finger pressing. W: 4lb CT: 36hrs  - no gold  C - direct placement power: 15, pressure: 500, temp:25, O2:100, time:10secs - no gold  Bonded immediately, no finger pressing, 0.5 lb weight on pdms C - direct placement from 2min-20min Vahid's recipe: power: at room temp. 30, pressure: 500, Then 45 min in temp:25, O2:100, time: oven. 20secs n/a C - direct placement - No finger Vahid's recipe: power: pressing. 30, pressure: 500, W: 8lb temp:25, O2:100, time: CT: 36hrs 20secs C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs  parylene-parylene  Pressure where leakage occurs across gold (psi) - Failure at 5 psi for all three channels on chip on the "top" part of the glass slide. ie: nearest to L5S - - Failure at 5 psi for all three channels on chip on the bottom of the chip.  37 blank glass slide  35 H52  38 blank glass slide  no finger pressing c: 8lb/2 CT: 24 hrs in 31/10/2010 bioMEMS 11/7/2010  multi-layer  21/11/2010  parylene-parylene  single layer 3000 spin  39  40  super thin multilayer  21/11/2010  4C  42 blank glass  super thin multilayer  41 blank glass  43 blank glass  - Tested to 30 psi via T junction test and saw no leakage - Slide 3D : Tested to 30 psi via T junction test and saw no leakage - Slide 3B: Tested to 20 psi via flow technique and saw no leakage - Slide 3A: Tested to 30 psi via T junction test and saw no leakage - Slide 3C: Tested to 30 psi via T junction test and saw no leakage (Note: I did see leakage in the center valve but I have not reported it as leakage because it was simply a piece of hair or dust that had fallen on the device and causing the leakage.  - Tested to 10 psi and saw no leakage  Pressure where leakage occurs between PDMS and glass (psi)  -complete collapse on both pdms chips. At 30 psi, could not push any solution into the chip. - complete collapse on 11/12 pressure inlets so that at 30 psi we could not push any solution into the chip. One pressure inlet was not totally collapsed so we could push solution into the chip. However, there was lots of leakage. It flowed out of the channels almost as if the channel boundaries weren't there. - This was a very thin multi-layer device. I tested it at 5psi and found that most of the fluid came out at the pressure inlet because the seal was not very good for such a thin PDMS chip. - Note: I found hair on the chip. I cleaned glass slide with aceton, ipa and water and i took pdms directly off wafer so I didn't think I needed to tape it. I suspect the hair fell onto the PDMS. I think it is worthwhile to tape the PDMS so that it is protected up until treatment and cleaned just before the bonding when the tape is removed.  - Slide 3A: Collapse in right valve (defined wrt the numbering 3A on top left of chip) at pressure inlet and outlet. Bottom is very badly collapsed, only allows two streams of 20 microns across through. Centre valve had no collapse. After about 10 secds at 30 psi all of the columns in the bottom part of the centre valve had their adhesion fail so that I could not longer see the columns when the pressure was on (note, I probably went over psi because the pressure valve was about three full turnns of the adjuster past the max display). On left valve, top and bottom, the columns failed at 30 psi after a few seconds. For the left valve I was careful not to exceed 30 psi. - Slide 3C: No collapse anywhere.  - After I saw 5 psi failure on the gold heater I set up a T junction to try and see how much pressure the PDMS to glass boundary could hold. However, I was only to go up to 12 psi before the path along the gold from the control channel to coolant channel started to leak. - I confirmed that the valves are working. I applied 10 psi at pressure inlet and could see the valves moving up and down. - a few places with collapse. But not much to speak of.  Comments  - this slide has two multi-layer chips on it. 3B and 3D. 3D is flawless, it has no collapse and holds pressure to 30 psi. 3B was cut too small so the pressure outlets are not fully enclosed so solution can leak out from the part of the pressure output which has an interface with the outside air. 3B was only test to 20 psi on one channel and it did not leak. There is some collapse in the coolant channel and in the column areas near the collapsed pressure outlets.  - no collapse anywhere - glass slide cracked when I was placing weight on top of it so I could only test two of the three valves.  n/a  - pdms for this chip was cured at 65C for 24+ hours before plasma bonding - Collapse on two areas before zigzags in control channel everything else is fine.  - pdms for this chip was cured at 65C for 24+ hours before plasma bonding - minor collapse in coolant channel and side of inlet - but generally good.  two multi-layer chips  super thin multilayer  24/11/2010  21/11/2010  - No finger pressing. W: 4lb CT: 26hrs  - No finger pressing. W: 4lb CT: 26hrs  - No finger pressing. W: 0.5lb CT: 36hrs  C - direct placement power: 30, pressure: 500, temp:40, O2:100, time:15secs  C - direct placement power: 30, pressure: 500, temp:40, O2:100, time:15secs  Single (S) or double (D) layer PDMS - if D include the thin layer spin speed  44 blank glass  one double-layer chip  24/11/2010  C - direct placement power: 30, pressure: 500, temp:25, O2:100, time:15secs  PDMS design  45 H4L heater  4C  one double-layer chip  - No finger pressing. W: 4lb  Heater #  46 blank glass slide  4D  30/11/2010  no apparent leakage, although no high pressure testing was done on this device.  Post-bonding treatment. Date that glass Cooking time (CT) Plasma cleaning Pressure where and PDMS bonded and Weight (W) on treatment. BioMEMS leakage occurs together device (B) or cleanroom (C) across gold (psi) C - direct placement - No finger Vahid's recipe: power: pressing. 30, pressure: 500, W: 0.5lb temp:25, O2:100, time: 21/11/2010 CT: 36hrs 20secs C - direct placement Vahid's recipe: power: 30, pressure: 500, temp:25, O2:100, time: 20secs  47 H4X heater  two double-layer chip  B - 40seconds  C - direct placement power: 30, pressure: 500, temp:25, O2:100, time:15secs  4B, 4D  Vahid's old device.  D - PDMS made in late nov and pulled off wafer #18 on dec 27 28/12/2010  no finger pressing W: 4lb CT: ___  48 H4X heater 49 blank glass slide 50 blank glass slide  52  4A  D  51 blank glass slide  53 H70  4A, 4B  30/12/2010  54 H63  55 H69  Pressure where leakage occurs between PDMS and glass (psi) Comments  - pdms for this chip was cured at 65C for 24+ hours before plasma bonding - no significant collapse.  - pdms for this chip was only cured at 65C for 4 hours before plasma bonding - extensive channel collapse. Could barely push any fluid through. This might be because the PDMS only had a four hour bake time. - acetone, ipa and water clean and 115C for 5 mins before plasma - Pdms prepared on Nov. 21 and only had 4 hours in oven for curing on Nov. 21st. It set out at room temp from Nov. 21st to Nov. 24th. no failure up to - valved closes around 12 psi. Minor collapse but generally happy 30psi. with bonding results. - - acetone, ipa and water clean and 115C for 5 mins before plasma -- Pdms prepared on Nov. 21 and only had 4 hours in oven for curing on Nov. 21st. It set out at room temp from Nov. 21st to Nov. 24th. no failure until 12 - 2/3 of the PDMS control layer was left on the wafer when it was psi and later had pulled off so the device only had one functioning valve. very bad failure at - valve closes around 12 psi. Minor collapse but generally happy 10 psi during with bonding results. flushing of device.. - see video M94 - note: i did not plasma treat to clean inside of pecvd like jonas suggested before bonding. need to remember to do that in future. - Valves 1-3 were were not properly bonded, causing a widespread rupture when fluid was pushed in. - Valve 4 exhibited significant collapse. - The 7-9 valves on this chip appeared to be viable, but the PDMS covered the set of pads closest to the device. - Valve 10 was partially collapsed; valve12 were completely collapsed such that fluid could not be pumped through. - spun at 3500 rpm 1 min after mixing - acetone, ethyl alcohol, water, hot plate clean - spun at 7000 rpm 5 min after mixing - air clean only - spun at 3500 repm 8 min after mixing - air clean only Vahid used this device 2 years ago for his research. We want to see if we can recreate the same results that he got. - Height of gold traces is 400nm In one of the channels the bonding failed at Philip writes on Jan. 9: Only info I could find was in work record 30psi. The next D366 which says: "Tried pumping water through the channel. First channel survived channel has leakage all the way to the outside. 30psi and the rest The middle channel seems OK when I am running water through it. of the channels There is some collapse but I am able to pump water through the were not tested. channel." The PDMS on the completed chip had to be trimmed. This probably reduced the bonding strength of the spun-on PDMS-glass bond. All of the channels - On Dec 30 Philip spun on a thin layer of PDMS at 8000 rpm. on the device - 3/6 blocked and the heaters are bad. subsequently - 9/12 toast failed badly at 25- - Heater 5 hairy but flow is good. 30psi. - Heater 8/11 hairy. Flow good. Jan.4: - cleaned slide with acetone, IPA and water and put glass slide at 115C for 10 minutes then let it cool. - Prepared 10:1PDMS with 3:30 mixing and 2:30 defoaming and spun on at 8000rpm for 1 min. - Put in 65C oven at 14:50. Jan.5: Thin membrane was ruined so I poured 50 grams of PDMS onto the slide and plan to pull off all of the PDMS after the PDMS has cured in the oven. I'll have to spin on new 8000rpm PDMS onto this glass slide  - Vahid gave us this slide  Heater #  PDMS design  Single (S) or double (D) layer PDMS - if D include the thin layer spin speed  A single multilayer valve from the Jan. 5th wafer #21 fabrication.  56 H72  small blank glass 57 slide  Double layer, spin 4A (covers valve speed of 2000 7-12) and 4D rpm, 80 minutes (covers valve 1-6). before allignment,  B - 40 seconds. I cleaned both surfaces with tape before plasma bonding.  B - 40 seconds. Put 250 mL beaker on it and left it in the over for 10 minutes after plasma treatment. I cleaned the PDMS with tape and did not clean the glass slide at all.  Post-bonding treatment. Date that glass Cooking time (CT) Plasma cleaning and PDMS bonded and Weight (W) on treatment. BioMEMS together device (B) or cleanroom (C)  W: 450 gram weight over both PDMS chips (so 225 grams per chip) CT: 20 1/7/2011 minutes.  1/9/2011  58 H4W  triple layer 2000rpm  1/9/2011  4B on both triple layer 2000rpm  59 H4W  60 H4W  Pressure where leakage occurs across gold (psi)  Pressure where leakage occurs between PDMS and glass (psi) - Valve 2-5 leaked at 10 psi - Valve 1-4 leaked at 5 psi after I lifted needle out of PDMS and that caused the PDMS to come up. The needle was really stuck in so I don't think this means there was bad PDMS-glass bonding.  - Fluorescent dye started leaking out of channels before I'd even turned on the pressure. It seemed like the capillary pressure was pulling the fluorescent dye forward and out. When I took the needle out of the chip the PDMS almost came off the glass. This was an EXTREMELY bad bond. There is no way I could have tested the multi-layer valve with this. - The heaters on this device were extremely hairy.  - Jan.5: I cleaned heater 72 with acetone, ipa and water and 115C for 5 minutes and poured 10:1 PDMS (3:30 mixing and 2:30 defoaming) onto it and spun it at 8000 rpm and placed it in the oven. Tomorrow it will be ready to bond to the last two of the multilayer PDMS chips I made today.  Comments  1psi  - heaters on 2/5 and 8/11 are bad. all other valves were ruptured in experimenting. - heater 1 jumpered - heater 3 has broken trace - instant rupture when 5psi applied to 7,8 and 9. 1,2,3,10,11 and 12 were ruptured from surgery to clear pads. - pushing fluid from side 5, I could not get it through the area 5 zig zags at 15psi. attempts to massage the chip failed. The device ruptured at 20psi. - pushing from side 4, the flow is sluggish and needs >10psi. there is leakage to the fluidic layer such that the fluid prefers to fill the fluidic channel instead of entering the area 1 zig zags. - 9/12 faulty pdms and cut off before bonding. - holes not punched in control channel for 1-6. - device was a complete failure from the start.  6.12 Appendix L - Macro Assembly  145  50  430.87  45.72  71.12  159.61  247.61  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52966.1-0074469/manifest

Comment

Related Items