Lab-on-a-Chip Diagnostic Biosensor First Semester Report Fall Semester 2010

By Justin Grantham John Blatt Daniel Higley Andrew Armstrong Kelli Luginbuhl Kelly Nienburg

Prepared to partially fulfill the requirements for ECE401 Department of Electrical and Computer Engineering Colorado State University Fort Collins, Colorado 80523

Project advisors: Dr. Kevin Lear, Dr. Dave Kisker, Dr. Susan Hunter

1

Abstract The inability to diagnose tuberculosis quickly and accurately in developing countries has been pointed to as a possible source of the disparity of deaths due to tuberculosis and other diseases in developing countries as opposed to developed countries. The local evanescent array coupled waveguide biosensor chip is a potential combatant to this disparity due to its quick diagnosis time (2-5min). The LEAC chip optically senses the presence of antigens or antibodies in a test sample of blood. In helping the progress of the LEAC chip, the senior design team working on the project this year decided to focus primarily on automating the current LEAC chip measurement process for the fall semester. Before this semester, the standard method of probing photodetectors on the LEAC chip was to individually guide single probes to the metal pads of the LEAC chip. The senior design team decided to implement a probe card based system to quickly and more effectively collect data from the LEAC chip. This system uses a probe card to contact many pads on the LEAC chip at the same time. It also uses precision rotational and translational stages to guide the probe card, LEAC chip and a laser precisely together. Finally, this system uses an amplifier and switching circuit to interface with the probe card and LabVIEW software to automatically collect data from these components. At the end of the semester some data was collected, though there are still some modifications to be made before it will be working fully. The senior design team also made significant progress towards implementing a polydimethylsiloxane (PDMS) microfluidic network on top of the LEAC chip to carry a blood sample over the waveguides on top of the LEAC chip. A PDMS microfluidic network was not actually implemented this semester, but this task should be completed early next semester.

2

Table of Contents Abstract ......................................................................................................................................................... 2 Table of Figures ............................................................................................................................................. 4 Chapter I: Introduction and Motivation........................................................................................................ 5 Introduction .............................................................................................................................................. 5 Motivation................................................................................................................................................. 5 Chapter II: Background ................................................................................................................................. 6 Local Evanescent Array Coupled Waveguide Biosensor Basics ................................................................ 6 Summary of Previous Work ...................................................................................................................... 7 Chapter III: Project Goals and Requirements................................................................................................ 8 Chapter IV: Chip Fabrication ......................................................................................................................... 9 Fabrication Basics...................................................................................................................................... 9 Photolithography .................................................................................................................................. 9 Etching................................................................................................................................................. 10 Deposition ........................................................................................................................................... 10 LEAC Chip Fabrication ............................................................................................................................. 11 Chapter V: Probe Card Requirements and Selection .................................................................................. 13 Chapter VI: Probe Card, Sample, and Fiber Launch Motion Control .......................................................... 16 Overview ................................................................................................................................................. 16 Requirements and Considerations.......................................................................................................... 16 The Sample Mount .............................................................................................................................. 16 The Fiber Launch Mount ..................................................................................................................... 17 The Probe Card Mount........................................................................................................................ 17 Chapter VII: Addition of Micro-fluidic Channels using PDMS ..................................................................... 19 Chapter VIII: Test Biomolecule Selection .................................................................................................... 21 Chapter IX: Probe Card Electronic and Software Interface......................................................................... 21 Multiplexer Circuit .................................................................................................................................. 21 LabVIEW Technical Considerations ......................................................................................................... 23 Chapter X: Conclusions and Future Work ................................................................................................... 24 References .................................................................................................................................................. 25 Acknowledgements..................................................................................................................................... 25 Appendix A: Abbreviations ......................................................................................................................... 26 3

Appendix B: Budget Summary .................................................................................................................... 26 Appendix C: Guide to Probe Cards .............................................................................................................. 26 Appendix D: SU-8 Molds Fabrication Process ............................................................................................. 28 Appendix E: LEAC Chip Fabrication Flow .................................................................................................... 31 Appendix F: LEAC Fabrication Diagrams ..................................................................................................... 38 Appendix G: Gantt Chart for Fall Semester................................................................................................. 40

Table of Figures Figure 1: The LEAC waveguide biosensor (Courtesy of Rongjin Yan)............................................................ 7 Figure 2: 64 to 1 MUX Circuit with Amplifier (Courtesy of Matt Duwe) ....................................................... 8 Figure 3: Current LEAC Chip Mask Design (Top View) .................................................................................. 9 Figure 4: Positive and Negative Photoresists.............................................................................................. 10 Figure 6: Micro Reactive Ion Etching Machine ........................................................................................... 10 Figure 5: Electron Beam Evaporation Machine .......................................................................................... 10 Figure 7: PECVD Machine............................................................................................................................ 11 Figure 8: Mask of LEAC chip ........................................................................................................................ 11 Figure 9: Probe card received from Alpha Probes Incorporated ................................................................ 13 Figure 10: Close up view of the probes and epoxy rings ............................................................................ 14 Figure 11: The LEAC chip topology. Black dots indicate ground pads and red ovals enclose sets of pads 14 Figure 12: Block diagram of probing apparatus .......................................................................................... 16 Figure 13: Sample XYZ and rotation stage and Fiber Launch...................................................................... 17 Figure 14: Probe station with addition of the probe card .......................................................................... 18 Figure 15: Microscope image of probe card in contact with metal pads (Each pad is 105um Wide) ........ 19 Figure 16: PDMS outline ............................................................................................................................. 20 Figure 17: Original microchannel mask design ........................................................................................... 20 Figure 18: New microchannel mask design ................................................................................................ 20 Figure 20: MUX to probe card interface ..................................................................................................... 22 Figure 19: Interface Card to connect switching circuit to the probe card .................................................. 22 Figure 21: Edge Sensor Circuit .................................................................................................................... 23 Figure 22: Probe Card with Edge Connector Attached ............................................................................... 27

4

Chapter I: Introduction and Motivation Introduction A biosensor is a biologically sensitive device that is used to detect the presence of some biologically reactive material or analyte. A biosensor consists of three parts, a biologically sensitive material, a detector element and peripheral electronics to display the data. These three parts work together to sense the presence of the substance in question, the analyte. A common example is a blood glucose meter used by people with diabetes to monitor their blood sugar. In the case of the blood glucose meter, the analyte is the glucose concentration in the blood, an enzyme called glucose oxidase is the biologically sensitive material that reacts with the glucose in the blood and an electrode that senses currents resulting from this reaction, is the detector. If the biosensor is used to detect the presence of an illness and thus diagnose the illness, then it is a diagnostic biosensor. When detecting the presence of an illness, specific types of bio-molecules must be sensed. When a disease is present in the body of a host organism, the disease releases various chemicals into the body of the host organism. These chemicals resulting from the disease organism are called antigens. When the host organism senses the presence of these antigens, the host organism produces antibodies. These antibodies bind to the antigens neutralizing them and preparing them for transport out of the body of the host. Thus, when attempting to detect diseases with biosensors, the analyte could be an antigen generated by that disease organism or an antibody generated by the body of the host in response to the presence of the disease organism. Detecting either the antigen or the antibody is a good indication of the presence of the disease. Furthermore, if an antigen is the analyte then the corresponding antibody is a logical choice as the biologically sensitive material used to sense the analyte and vice versa. In the case of the LEAC waveguide diagnostic biosensor, the analyte is a Tuberculosis related antigen, the biologically sensitive material is the corresponding antibody, the waveguide and related features are the detector and peripheral electronics are used to interact with those detectors. Following in this report will be a discussion of the motivations for working on this project. Chapter 2 goes into further detail on the function and fabrication of the LEAC chip as well as prior work done with it. Chapter 3 will then include analysis of the project goals and requirements for this semester. Chapters 4 through 9 discuss the work done by this senior design team this semester which flows into plans for work next semester in Chapter 10 followed by several appendices that will be cited throughout the report.

Motivation This project has the potential to lead to a final product with a very positive impact on society. The technology utilized provides significant benefits over current diagnostic practices and, when finalized, could lead to a product that would be used in developing nations to increase the rate of diagnosis and 5

treatment of Tuberculosis. Increased rates of diagnosis and treatment will lead in the long term to a decreased death rate due to Tuberculosis in the areas where the product is implemented. In developing nations such as many in Africa, the death rate due to Tuberculosis is significantly higher than that of developed nations. As a whole, the continent of Africa has over 11 times the deaths (normalized to population) due to Tuberculosis than the Americas (Citation Here). This disparity is a result of many factors including environment, lifestyle and access to healthcare facilities. The goal of this project is to combat the problem by increasing the access to healthcare in these nations. There are currently several groups working towards this goal in various ways, however their efforts are often impeded by the unavailability of inexpensive, transportable means of diagnosing tuberculosis. One highly transportable diagnostic tool is the Tuberculin skin test. In this test, a substance called Tuberculin is injected under the skin of an individual by a trained healthcare professional. The individual must then return 2-3 days later to have the healthcare professional examine the injection site to make a diagnosis. The problem is that the individual often doesn’t return to the clinic to be diagnosed and thus never receives treatment. Other diagnostic practices may suffer from this time delay between sample collection and diagnosis. In other cases, diagnosis is achieved more quickly, but there is a need for a full laboratory environment to perform diagnosis which is expensive. The LEAC diagnostic biosensor seeks to solve this problem by providing diagnosis in a matter of minutes in a cost effective way. The decreased cost compared to other systems and the increased speed of detection will increase the availability of treatment and have a positive impact on quality of life in these regions.

Chapter II: Background The local evanescent array coupled (LEAC) waveguide biosensor is a research project here at CSU that has been going on for quite a few years under Dr. Lear. It is currently headed by Phd student Rongjin Yan. He has numerous publications regarding the LEAC theory and application which can be easily found online[2][3]. In this section the physics of the LEAC chip will be discussed as well as a very brief summary of some of Rongjin Yan’s previous results.

Local Evanescent Array Coupled Waveguide Biosensor Basics The LEAC chip operates on the principle of total internal reflection (TIR) to optically sense biomolecules. A laser light is optically coupled into the silicon nitride waveguide structure at the edge of the chip. Since the index of refraction of the silicon nitride (n=1.8) is larger than the refractive index of the surrounding silicon dioxide (n=1.48) and air (n=1), there is a critical angle where the light entering will become trapped inside and will propagate down the waveguide. This is the same principle that is used

6

in fiber optic communication channels to send information with light.

Figure 1: The LEAC waveguide biosensor (Courtesy of Rongjin Yan)

The waveguide is seen in Figure 1 as the thin green layer on the top of the chip. Even though the light is totally internally reflected, there is a component of the electric field that is present on the outside of the waveguide. This component is known as the evanescent field and it is this field that we use to detect the biomolecules. When a biomolecule is attached to the top of the silicon nitride waveguide the increased index of refraction causes the evanescent field to shift upwards and couple into the biomolecule. This shift in power will reduce the field beneath the waveguide that is in contact with the polysilicon detectors. The evanescent field profile and the polysilicon detectors are shown in Figure 1. The field reduction on the polysilicon detectors increases the resistance of the detectors and thus decreases the current flowing through them for a given voltage bias. The voltage bias is applied to the ends of the detectors through metal vias also pictured in Figure 1. This change in current is how the biomolecule layer can be detected at any point along the waveguide where there are buried detectors that can be biased.

Summary of Previous Work This will be the first semester that a senior design team will be working in the LEAC project. Previously it has been worked on solely by graduate students and thus has a lot that can be expanded on by undergraduates. There currently exist two different chips that are referred to as LEAC chips. The first generation was a design that was on a corner of a chip that was manufactured by Avago. The most recent version of the chip however is fabricated in the CSU cleanroom. This new generation of LEAC chip allows for much more freedom of experimentation since new chips can be fabricated at will. Currently Rongjin has manual probe arms that he guides under a microscope to each individual pad to apply the bias. We wish to automate this system as described in Chapter III: Project Goals and Requirements. A switching circuit has also been designed already by graduate student Matt Duwe. This circuit consists of 9 separate 3-8 MUX’s arrayed together, allowing for the selection of only one of 64 channels. This selection is made by inputting 6 digital control lines into the multiplexers, choosing between the 26=64 channels. This chosen channel is then passed through a transimpedance amplifying stage with a gain of

7

500 million V/A. Further information may be found in [1]. The top view of the circuit can be seen below in Figure 2.

Figure 2: 64 to 1 MUX Circuit with Amplifier (Courtesy of Matt Duwe)

Chapter III: Project Goals and Requirements Throughout this semester and into next semester we as a group were hoping to achieve a real time sensing capability with the LEAC chip. As mentioned previously, right now the only way to measure each buried detector pad is to tediously move individual probes to each position. This makes it difficult and time consuming to obtain the information we need to analyze each experiment. In order to expedite this process we decided one of our main goals was to assemble and build a probe station that we could use to maneuver a probe card to the chip. Once the probe card is in contact with the chip we would then be able to easily toggle between all of the different detectors along the waveguide. In addition to this automation of the LEAC chip, we also desired to better implement the microfluidics that will be utilized on the chip. Right now antigens and antibodies are simply printed and rinsed off of the surface of the waveguide. This process also required that the biomolecules dry on the surface since the user had to individually probe each detector. Now with the automation process in operation, we will be able to flow biomolecules down a microfluidic channel on top of the waveguide and take a measurement instantly after the injection. We also wish to further explore different variations of the chip through the fabrication process. A mask has been made with a large number of variations in widths of features that we can test to find which dimensions and arrangements work the best. Once we implement the microfluidics and the probe station set up, we wish to further characterize the ability of the LEAC chip to sense the biomolecules. This includes achieving a given sensitivity (detect

8

whether a molecule is present) and specificity (whether we can tell biomolecules apart). This characterization portion of the project will be the main focus of the spring semester.

Chapter IV: Chip Fabrication Fabrication of the chip is done in the cleanroom located in the basement of the engineering building. We have recently just acquired all of the technology needed to manufacture the LEAC chip completely in our laboratory. In this section, a brief introduction to the fabrication processes and technology used in the manufacture of the LEAC chip will be introduced. A short walkthrough of the LEAC chip fabrication in particular will then be discussed. The full Fabrication flow for our chip may be found in Appendix E: LEAC Chip Fabrication Flow and Appendix F: LEAC Fabrication Diagrams.

Fabrication Basics The LEAC chip is fabricated using many common processes and techniques used in the semiconductor manufacturing industry. Among these are photolithography, etching, and deposition. Pictured in Figure 3 is one of the mask designs that are used in the fabrication of the chip. Using individual layers of this mask one at a time, all of the features can slowly be developed from a featureless silicon wafer.

Photolithography The technology to actually use the individual mask layers of the chip is called photolithography. Every sequence of the photolithography process starts by applying a layer of photoresist to the top surface of the chip. Photoresist is a liquid substance that is spun on to the chip using a vacuum chuck that holds the chip in place. Resist is placed on the top surface, and spun for a specified amount of time depending on how thick the user wishes the resist layer to be. Upon applying the resist to the surface, the chip is then placed under the mask aligner machine, where an individual mask layer is positioned over the chip. UV light is then exposed to the chip, chemically reacting with the photoresist depending on whether it is positive or negative. Upon development, positive photoresist will leave a mask of resist in the shape of the features that were prevented from being exposed to the UV light. Negative photoresist behaves in the opposite way, leaving a layer on the chip where the resist was exposed to the light, thus achieving a “negative” replication of the mask. A cross section diagram is shown of the difference in Figure 4. Figure 3: Current LEAC Chip Mask Design (Top View)

9

Figure 4: Positive and Negative Photoresists

This mask of photoresist is now used to shield certain parts of the chip so that etching and deposition can now be done on only the surfaces exposed. Etching Etching is the process of removing a specified amount of material from the surface of the chip. This is typically done by either a wet etch or a dry etch. Wet etching refers to dipping the whole chip in acid and removing material. Dry etching is performed by bombarding the surface with ions, which removes material at a specified rate. Dry etching is the cleanroom is done by using a micro reactive ion etching (micro-RIE) machine as pictured in Figure 6. The micro-RIE machine uses a strong radio frequency electric field to ionize a certain gas and create plasma, which reacts with the material on the chip, removing it. Wet and dry etching each has advantages and disadvantages. Both normally involve dangerous gases or chemicals, so caution must be taken when performing both processes. An advantage to dry etching is that its etching profile is normally more anisotropic. This means that the etching primarily acts in the vertical direction and will not etch sideways into material that is covered by photoresist. Wet etching on the other hand is normally very isotropic, meaning it etches in all directions. Figure 6: Micro Reactive Ion Etching Machine

Deposition Deposition is the process of adding a layer of material on to the surface of the chip. This can be a layer of metal, semiconductor, or even an insulator. Metal is deposited on the surface of the chip using a technology called electron beam evaporation physical vapor deposition. The machine that performs this is pictured in Figure 5. This involves placing the sample inside a vacuum chamber with a crucible of the desired metal. The metal is then heated to very high temperatures by shooting it with a beam of electrons, which then evaporates it. The evaporated material flies through the vacuum

Figure 5: Electron Beam Evaporation Machine

10

until it contacts the chip, leaving a layer of metal on top of the chip. To deposit layers of semiconductor or insulator, a technology called plasma enhanced chemical vapor deposition (PECVD) is used. In this process the sample is again placed inside of a vacuum. A specific mixture of gases is then pumped into the chamber and inductively turned into plasma. A chemical reaction then occurs on the surface of the chip which then builds up the layer of material. The PECVD machine is pictured below in Figure 7.

Figure 7: PECVD Machine

By cycling through these basic steps we are able to create all of the features needed on the LEAC chip.

LEAC Chip Fabrication The LEAC chip is fabricated using multiple runs of the processes describe above. The fabrication flow described here will correspond to the cross sectional view seen in Appendix F: LEAC Fabrication Diagrams. The following is just one of the fabrication flows possible that will produce a variation of the LEAC chip. The specific process depends on the available wafer substrate we have available to start with, as well as if we are experimenting with slightly different structures. This, however, is the most common as of now and chips using this flow have been successfully produced. Shown below in Figure 8 is the newest mask design that is used in the fabrication of the LEAC chip.

Figure 8: Mask of LEAC chip

11

The four different quadrants show the four main mask layers: Metal, Waveguide, Metal Via, and Buried Detectors. Within each quadrant are variations within each mask layer. These variations have different dimensions and features that will be tested on the LEAC chip. Refer back to Figure 3 for a top view of the chip with all 4 layers stacked on top of each other. Fabrication begins with a silicon wafer with 100nm silicon dioxide (SiO2) and 255nm silicon nitride (Si3N4) layers on top of it. The SiO2 layer is simply for insulating the polysilicon layer from the silicon substrate. The first step in fabrication is etching the polysilicon layer into the buried detector structures. This is done by using the buried detector mask (top right quadrant of Error! Reference source not ound.) to place a mask of photoresist on the top of the chip and etching it down, as described in Fabrication Basics. In this case the etching is performed by the micro-RIE machine, and thus is dry etched. Recently we have begun experimenting with a wet etch of KOH to remove the polysilicon as the dry etching is not very selective and at times has removed our insulating silicon dioxide layer. In this case, however, a mask of Si3N4 must be used as the photoresist cannot stand up to the KOH. Next the metal pads and wires are laid down. This is done by first applying a layer of negative photoresist and performing what is known as lift off. After the photoresist is applied and has been developed using the metal mask layer, aluminum is deposited using the e-beam evaporation machine. After, the whole sample is submerged in acetone with dissolves the photoresist and thus “lifting off” the metal lying on top of it. This leaves the aluminum only in the positions that we wish. After this a layer of insulating silicon dioxide must be deposited to separate the buried detectors from the waveguide. This is done by using the PECVD machine as described above, which will deposit silicon dioxide all over the top surface of the chip. Since a lot of insulation is desired we normally deposit at least 1um of SiO2 on the surface. Immediately following this insulator deposition a layer of Si3N4 is then applied directly on top of the SiO2, again using the PECVD machine. This layer of Si3N4 is then etched into the waveguide features by using positive photoresist (S1818) and the waveguide mask. All that is left after these steps is to gain access to the metal pads that are buried beneath the SiO2 layer. To achieve this we use the “via” mask to cover the whole chip with photoresist except for the locations with the metal pads. The SiO2 layer is so thick that dry etching would take way too long to perform so we use wet etching once again. For this step hydrogen fluoride (HF) is used. It eats through SiO2 very fast and thus only requires about an 8sec dip before the vias are opened all of the way through.

12

Chapter V: Probe Card Requirements and Selection This section describes the requirements for the probe card ordered this semester, our reasoning for ordering a probe card and decisions on parameters of the probe card. If further information is desired on the general aspects of probe cards, Appendix C: Guide to Probe Cards, serves as an introductory guide to probe cards. To make the process of testing the LEAC chip much quicker the team decided to implement a probe card to be able to probe many of the LEAC chip's photodetectors simultaneously. This probe card was purchased and implemented this semester. Before the implementation of this probe card, making measurements with the LEAC chip entailed probing individual pads with individual probes and measuring the current across a photodetector for a specific voltage with each iteration. After the implementation of the probe card, probing the first 32 photodetectors connected to any one waveguide on the LEAC chip simply required setting down the probe card once and using a system of multiplexers, amplifiers and LabVIEW software to systematically measure the currents running through all individual photodetectors for a given applied voltage. The LabVIEW probe card interface is described in Chapter IX: Probe Card Electronic and Software Interface. The motion control system necessary to precisely align the probe card with the LEAC chip is described in Chapter VI: Probe Card, Sample, and Fiber Launch Motion Control. Figure 9: Probe card received from Alpha Probes Incorporated

The requirements for the probe card were fairly standard since we do not have a high-frequency, high-power, high-temperature, or low-temperature system. Also, the LEAC chip has aluminum pads to be contacted (as opposed to gold pads) since high contact resistance is not a large problem. Therefore, 13

we chose to use standard tungsten probes. Since the LEAC chip requires a moderately high probe density (probe pitch, or distance between adjacent probes, of 105 mils) and does not need any of the benefits blade cards provide, we chose to use an Epoxy-ring probe card topology. We chose to order our probe card from Alpha Probes Incorporated, which is based in Colorado Springs, Colorado because they provided us with the cheapest quote ($235 for Alpha Probes versus $500 for Accuprobe) out of the vendors we contacted and they are relatively local. Figure 9 is an image of the probe card we received. Figure 10 is a zoomed in view of the probes and epoxy-ring configuration on the probe card. Essentially, our probe card consists of a large PCB (4.5” width, 9 3/8” length, 0.062” thickness) which has a hole to accommodate probes on one side and metallic fingers to accommodate connection of the probes to external electronic circuits on the other. The probes are very precisely aligned on the bottom of an aluminum ring using epoxy. The metallic fingers are accessed via a standard edge card connector. Our probe card has a total of 34 probes and 2 edge sensors. We chose to use this many probes because, it allows us to probe all of the probes for a given waveguide Figure 11: The LEAC chip topology. Black dots indicate ground pads simultaneously well avoiding the significant and red ovals enclose sets of pads cost of added probes. Thirty-two of these probes are used to contact the biasing side of photodetectors and the other two are used to contact ground pads for the photodetectors. All of the photodetectors for a single waveguide share a common ground. The reason for having two ground probes was because the ground pad is on a different side of the pads, depending on which section of the LEAC chip is being probed. Figure 11 shows this. The edge sensors were placed in line with the rest of the probes but at a strategic distance away to avoid damaging any features on the chip. The edge sensors are used to determine when contact is made. Our utilization of them is further described in Chapter IX: Probe Card Electronic and Software Interface. We had the choice between insulating and noninsulating edge sensors when we ordered the card. We went with the non-insulating Figure 10: Close up view of the probes and epoxy rings

14

for two reasons; they were cheaper ($15 versus $35)[1] and our chip has an insulating layer on the top surface and would not be effected by the live circuit. An application specific problem we had during the design process was that we will have a ~4mm tall obstacle in the center of our chip which had to be avoided. This obstacle was the

Polydimethylsiloxane (PDMS) which is further described in Chapter VII: Addition of Micro-fluidic Channels using PDMS. The approximate location of PDMS is the orange rectangle in Figure 11. In order to avoid the possibility of any part of the probe card contacting the PDMS, we increased the height of the epoxy-ring holding our probes from the standard 80 mils to 250 mils. We also left out the half of the epoxy-ring (white substance in Figure 10) that would be interior to the LEAC chip when probing. Another possibility we considered was cutting the PCB of the probe card completely off halfway through the hole accommodating probes. This option was discarded when our vendor informed as that this might damage the structural integrity of the card and possibly warp it. We also considered using probes with longer probe tip lengths, but our vendor informed us that this would cause the probes to be much too fragile and that we would easily break them[2]. Because our application for this probe card is necessarily very low noise, we thought it would be good to place the amplifier circuit for signals exiting photodetectors on the probe card itself. Unfortunately, our vendor informed us that this would add thousands of dollars to the cost and we could not afford such an increase so we were forced to keep the circuit external and make sure we had effective shielding for the probe card electronic interface. Ordering and receiving the probe card took much longer than we originally anticipated and was a major factor for us not making our originally stated goals. We had planned on receiving the probe card by the 20th of October but did not actually receive it until the 22nd of November. The major contributing factors to this delay were difficulty in determining exactly how we wanted the probe card configured (due to a lack of readily available information) and difficulty in communicating with the vendor exactly what we needed for our application.

15

Chapter VI: Probe Card, Sample, and Fiber Launch Motion Control Overview Figure 12 above shows a block diagram of the Probing Apparatus used in our project. To take a measurement from a sample of the LEAC chip, one must first bring probes into contact with metal pads on the chip. To achieve the goal of automated measurement, a row of 32 probes must be simultaneously aligned to 32 corresponding metal pads on the sample. Once the probes are in contact, laser light is then coupled into a waveguide on the LEAC chip. This process is very sensitive and thus

Figure 12: Block diagram of probing apparatus

requires the use of an extremely precise piezo-electrically controlled XYZ stage called a fiber launch. Once everything is in place, 32 measurements are taken very quickly through the switching circuit with the help of the LabVIEW interface. This set up is fairly complex, especially compared to a single probe measuring system. This results in a longer set up time per sample. However, even if this set-up requires four times longer getting ready, the number of measurements that can be taken without modifying the setup (unlimited) more than compensates for this disadvantage.

Requirements and Considerations There are a number of things that needed to be taken into consideration throughout the design of the probe station. With so many movable stages we had to thoroughly plan how every piece would translate together and in the order they would do so. The Sample Mount The pads on the sample must be precisely aligned to the probes on the probe card. This is achieved through the use of an XYZ stage combined with a rotation stage. With these degrees of freedom, the probe card can be arbitrarily fixed in position and then the sample moved precisely into a position just below it. Planarity must also be considered so that some probes do not contact the sample before the 16

others. To guarantee planarity, the metal bracket that connects the XYZ and rotation stages must be fabricated very precisely. For this reason we had the metal bracket fabricated to very strict tolerances using a digitally controlled mill in a machine shop. It was also designed to accommodate the previously mentioned rotation stage and a second stage that allows a planarity adjustment if we think we need it. The Fiber Launch Mount The control of the fiber must be very precise to allow light coupling into the waveguide. To achieve this we took advantage of a piezo-controlled x-y-z translation stage that we found in the lab. This allows us to precisely align the fiber to within micrometers of where we need it to acheive the most optical power into the waveguide. Considerations must be made to eliminate any tall features on this stage to allow clearance for the probe card to move down to the pads. We were able to extend the fiber only in the direction toward the sample by utilizing glass slides as a sort of “plank” to support it as seen in Figure 13. The Probe Card Mount The probe card must be affixed to something that allows very precise Z-motion in order to bring the probes into contact with the sample. To achieve such precise tolerances for something as large as the probe card appeared to be a difficult task until we came upon the mounting station as seen in Figure 14. These metal slats proved to have a very precise z-motion capability (2mil per full revolution of the handle) which was very much in line with our 1.5mil over travel specification on the probe card. The probe card however has exposed wires beneath it so it could not simply be setting on top of the metal slats. In order to achieve isolation between the items we had glass custom cut to sit beneath the probe card. This also proved to be beneficial in allowing clearance for the edge connector to interface with the probe card.

Fiber extension

Figure 13: Sample XYZ and rotation stage and Fiber Launch

17

Notice above in Figure 13 the glass slides extending the fiber to the center of the rotation stage. This is desirable for the alignment of the fiber to the LEAC chip. In addition, the sample is the tallest feature in this setup, guaranteeing clearance for the probe card. Figure 14 below shows the probe station setup with the addition of the probe card. Notice how the entire sample and fiber apparatus has been rotated 90 degrees to accommodate the probe card.

Probe Card

Metal slat of probe station

Figure 14: Probe station with addition of the probe card

18

Figure 15: Microscope image of probe card in contact with metal pads (Each pad is 105um Wide)

Figure 15 above shows the contact of the probe card tips with the metal pads on the surface of the LEAC chip. The blurry probe structures on the right are the edge sensors which are further discussed in Chapter V: Probe Card Requirements and Selection and Chapter IX: Probe Card Electronic and Software Interface.

Chapter VII: Addition of Micro-fluidic Channels using PDMS One of the primary goals of this design project is to integrate microfluidic channels onto the chip. The integration of these channels will enable the implementation of a real-time detection system using LabVIEW. The initial setbacks with this portion of the design were limitations in fabrication needs. The lab did not possess a vacuum oven or the polymer media needed to fabricate the channels. The polymer used to create microchannels is polydimethylsiloxane (PDMS). Having also acquired training in the clean room for fabricating the SU-8 mold, from which the PDMS is cast, fabrication of the microfluidics is on track. The clean room process for making the SU-8 molds can be found in Appendix D: SU-8 Molds Fabrication Process. Next semester we hope to optimize the channels and to test surface treatments and new designs that promote binding and mixing of the fluid to bring the biomolecules to the surface. We also intend to optimize the positioning of the PDMS piece for integration and test leakage, different PDMS to surfactant ratios for varied flexibility of the polymers, and different curing techniques that alter the level of reversible binding for chip reuse.

19

In addition to learning the methods needed for fabrication this semester, it was found that the PDMS needs cut to within 500 microns to a specific size so as not to cover the sensors on the chip. However, we also want the PDMS to be big enough to cover the first few sensors, which are the most accurate and have the strongest signal, are still utilized. These requirements yield a PDMS piece about 0.9 mm2. The group came up with several ideas including the assembly of a metal dish of the exact size, which the PDMS would be poured into, or redesigning the mask with cutting guides. Figure 16: PDMS outline

The mold is typically placed in the bottom of a Petri dish and the PDMS then fills the dish and is cut out around the features with a razor blade. By designing and creating a dish of the exact dimensions needed for overlay on the chip, the final PDMS piece would be the exact size and less material would be wasted in excess pouring. The PDMS can be as thin as 0.5 mm and as thick at 5 mm with a pouring precision of about  0.3 mm. This thickness will be affirmed next semester when the size of the pumping reservoir is determined. The team decided to try a mask redesign with cutting guides first. Unfortunately, when opening the old mask designs to add cutting features on the LASI program, it was seen that the original design is too big for the size guidelines and the microfluidic features themselves are already outside the bounds needed to keep the pads exposed. The design needs to be completely revamped to reduce the channels and reservoirs and enable implementation of razor blade guides outlining the desired size for the PDMS square. Figure 16 shows an approximate PDMS perimeter, positioned correctly over the chip.

Figure 17: Original microchannel mask design

In order to cut out the entire design, the PDMS square would cover the entire chip. After reducing the size of the reservoirs (circles) and adjusting the channel positioning, the design of Figure 18 was created, with a cutting guide border. This design still needs to be run by the advisors and we hope to schedule a meeting with Dr. Lynn to discuss whether the new parameters, particularly channel thickness and reservoir size, will be appropriate for fabrication. Once all of those issues have been addressed, the new mask will be ordered, a new mold processed in the clean room, and the properly-sized PDMS pieces fabricated and ready for integration studies.

Figure 18: New microchannel mask design

20

Chapter VIII: Test Biomolecule Selection In order to make a diagnosis our system is dependent on the occurrence of antibody/antigen bonding. Antibodies are a part of the immune system produced by a person in order to bind to and thus, inactivate foreign objects. These foreign particles can be a virus or bacteria and are the antigens. Their structure is basically universal, consisting of a heavy chain and a light chain, however there is a section on every antibody called the variably region. This region is very specific to the type of antigen the antibody will bind to. This specificity is what allows us to test for a disease in someone’s blood. Once a probe molecule, an antigen or antibody, is printed onto the chip or on the microfluidic channel, if the sample has the corresponding disease, then the preprinted biomolecule will bind to the target in the sample. For initial testing of the system, bovine serum albumin (BSA) will be used. BSA is an accessible and relatively cheap biomolecule. Using fluorescently labeled BSA will allow us to verify that the biomolecules are patterned correctly. Eventually the use of a tuberculosis strain will be used but currently it is more cost effective to use a cheaper biomolecule for the testing out the system. One of the further off goals for this project is to have a lab-on-a-chip with various different channels to test multiple biomolecules. These different biomolecules could be different forms of tuberculosis that would specify the test more or could test for other diseases effecting a similar population. Size is also an important factor in determining which biomolecules to test for in this system. Both malaria and hepatitis B are diseases that could show interest to look into testing in the future as both affect similar populations as tuberculosis and are two of the most common and deadly diseases. They both also have similar sizes to tuberculosis biomolecules thus would be likely candidates to work within our system.

Chapter IX: Probe Card Electronic and Software Interface Multiplexer Circuit In order to efficiently measure all 32 of the pads connected to the buried detectors a multiplexing circuit was designed by Matt Duwe (See Summary of Previous Work). This circuit was capable of using six binary input channels from a computer to select and let pass through only one of the sixty-four separate inputs. The circuit is then followed by a transimpedance amplifier and an additional voltage amplifier. These two amplifying stages provide a gain of 500 million times. This is needed since the buried detector resistances are so high and thus for a reasonable bias we would only be seeing currents in the range of 10-100 nA.

21

While this circuit was a great start, it needed a lot of peripheral connections made with it. This including all of the input channels going from the circuit board and interfacing with the probe card. Deciding how to achieve this connection proved to be quite a bit more complex than was originally thought. The probe card size we ordered require a 70 conductor edge connector, which itself is not a very common item. Due to this an edge connector could not be found that would interface directly with a ribbon cable, thus an interface card had to be made that would allow the conduction paths to successfully interface with the Figure 19: Interface Card to connect switching circuit to the probe card two connectors (Figure 19). The fully assembled circuit can be seen in Figure 20 .

Figure 20: MUX to probe card interface

As mentioned in the previous sections, the probe card contains edge sensor pins that we could interface with to tell whether we have made contact with the chip. The Accuprobe website describes the edge sensors as being normally closed (in contact with each other) and when contact is made by the probes the circuit is broken. In order to take advantage of this system a very simple circuit was designed that would indicate whether the edges sensors have come in contact. The edge sensor circuit design is shown in Figure 21.

22

Figure 21: Edge Sensor Circuit

It can be seen from the circuit diagram that if either of the edge sensors come in contact with the surface of the chip it will break the connection. This break will stop current flow which will shut off the corresponding LED. Thus, when the LED turns off we can tell whether that edge of the probe card has come in contact.

LabVIEW Technical Considerations One of the main goals for this semester was to increase the speed of taking data by writing a LabVIEW program that would automatically switch between the channels, take multiple data points, and figure out the standard deviation of the data points. This would hopefully reduce the time to take data from over an hour to just minutes. The first iteration of this program had the ability for multiple channels to be selected, although they had to be in series (For example, testing channels 5 through 12 is possible, but only testing channels 3, 7 and 24 is not.) A further addition gave the ability to look at a reference channel after every other channel, comparing a known value every time to make sure there is no drift in the system that would skew later results. There was also a residence time feature added, which is a waiting period between when the Hytek iUSBDAQ would switch channels and when it would start to record data points. The limitations of this version of the program only allow it to take one point on each channel, and output the data and channel number into one column to excel. Further testing still needs to take place, as well as an update to the graphical user interface to make it more user friendly so that anyone from the team can easily run test.

23

Chapter X: Conclusions and Future Work This fall semester has proven to be very successful. We have successfully worked together as a team to fabricate the LEAC chips, design the PDMS channels, and partially automate the measurements. The LEAC chip has been successfully probed and measurement s taken through a computer interface. We are prepared to fabricate and begin tests using the new PDMS channel designs. Throughout this semester we have also learned a lot about working in research laboratories with a large variety of equipment and how to manage our time throughout the semester (See Appendix G: Gantt Chart for Fall Semester). Future work for the spring semester will consist of a few parts. The LabVIEW switching circuit still needs to be finalized and tested thoroughly on old LEAC samples to be sure we are getting reliable data. We also are ready to fabricate the newly designed PDMS microfluidic channels that we will use to locally apply biomolecules to the surface of the chip. Lastly, we will characterize the ability of the LEAC chip to sense different types of biomolecules and obtain sensitivity and selectivity baseline and try to improve upon it if needed.

24

References [1] A. Nott, private communication, October-November 2010 [2+ Accuprobe, “Probe Card Tutorial,” 2003, http://www.accuprobe.com/Downloads/Probe%20Card%20Tutorial.pdf [3]Compiled by A. D. McNaught and A. Wilkinson. (1997) “IUPAC. Compendium of Chemical Terminology, (the “Gold Book”). (2nd ed.) . *On-Line]. Available: http://goldbook.iupac.org [12/6/10] [4] M. Duwe, “Low Noise Selecting Circuit for Local Evanescent Array Coupled Biosensor” (Senior Honors Thesis), unpublished. [5] R. Yan (2010, January 10). Review of label free optical biosensor. [6] R. Yan, Mestas, S., Safaisini, R., & Lear, K. (2009, August 7). Label-free silicon photonic biosensor system with integrated detector array. Lab on a Chip, 9(15), 2163-2168.

Acknowledgements We would like to acknowledge a few individuals that have considerably helped us throughout this semester. It is with their guidance that we have been able to get so far on this project this semester. Our advisors Dr. Kevin Lear, Dr. Dave Kisker, and Dr. Susan Hunter have been invaluable guides with so much combined experience in research. PhD student Rongjin Yan has also been a great guide and has been very patient in recommending goals and helping us get adjusted to the research lab environment. Graduate students Tim Erickson and Iris Yi have also been a great help in the fabrication of the LEAC chip and without them we could not have done it. Andy Nott at Alpha Probes has also been very patient and has been willing to guide us for quite some time in the design of our probe card. Lastly we would like to thank two professors on campus, Dr. Lynn in the Chemical Engineering Department and Dr. Henry in the Chemistry Department, both were very generous with both their equipment and materials. Also a postdoctoral researcher in Dr. Henry’s lab, Jason Emory, was generous enough to provide training on the PDMS casting process and also expressed the potential that our group would be able to use the lab’s fluorescence microscope for future imaging needs as well.

25

Appendix A: Abbreviations E-Beam PVD: Electron Beam Physical Vapor Deposition LEAC: Local evanescent array coupled LED: Light Emitting Diode MUX: Multiplexer PDMS: Polydimethylsiloxane PECVD: Plasma enhanced chemical vapor deposition S1818: Shipley 1818 positive photoresist Via: A vertical “passage” connecting different layers of a silicon chip

Appendix B: Budget Summary Item Probe Card Switching Circuit Supplies Probe Station Total

Cost $305.00 $20.00 $50.00 $20.00 $395.00

Spent From Rongjin Yan's Research Budget Personal Rongjin Yan's Research Budget Personal

Personal expenses may be reimbursed through the $50 per semester per student ($300 for the whole year) at the end of the academic year. Rongjin Yan’s research budget may be further used at his or Dr. Lear’s discretion. There are no guarantees of future funding through the research budget.

Appendix C: Guide to Probe Cards This appendix is intended to serve as an introduction to probe cards and the terminology associated with using them. Chapter V: Probe Card Requirements and Selection outlines the requirements and selection of the probe card used with the LEAC chip. Probe cards are commonly used to manually and automatically probe test wafers in industry. The probe card is essentially a printed circuit board (PCB) with a circular hole which accommodates probes. The PCB part of the probe card generally has traces from these probes to metallic fingers on one edge of the card. If making a custom probe card, the PCB can be layed out using PCB layout software such that discrete components or entire circuits can be soldered directly to the probe card. The metallic fingers which are electrically connected to the probes of the probe card can be accessed through a card edge connector. Figure 22 shows the probe card received to be used in conjunction with the LEAC chip with a card edge connector attached. Figure 9 shows a more zoomed in view of this probe card. 26

Figure 22: Probe Card with Edge Connector Attached

There are two main types of probe cards. The traditional blade probe card has superior mechanical stability and a higher integrity signal path than the more recently developed epoxyring probe card. To make a blade probe card, ceramic blades are inserted into lands on the probe card and soldered down. These blades have probes sticking off them which are bent into the required shape for the component to be probed. Because of their superior mechanical stability, blade probe cards are less likely to need re-planarization adjustments. Blade probe cards are inferior to epoxy-ring probe cards when it comes to probe densities and probe counts (blade probe cards generally have a probe count less than 88 where as epoxy-ring probe cards can have probe counts greater than 2000) [2]. The epoxy ring probe card technology used to be more expensive, but with recent advances in techniques and technology, has become around the same price as the blade probe card technology. The epoxy-ring probe card topology has now overtaken the blade probe card topology in terms of popularity. An epoxy-ring probe card can be used for most applications except those which deal with high-frequencies (> 3 GHz) or very low leakage currents (< 1 pA/V) [2]. The most commonly used materials to make probes out of are tungsten, tungsten-rhenium and beryllium-copper. Tungsten is the most commonly used and cheapest of these materials. It provides good probe stability and probe life for most applications. Oxide crystals can become trapped in Tungsten probes periodically due to their fibrous nature [2]. This means tungsten probes will require occasional cleaning to maintain acceptable contact resistance. Tungstenrhenium probes are very similar to tungsten probes except they do not have the fibrous nature of tungsten and thus, trap oxide crystals within them less readily. Beryllium-copper probes are used when low contact resistance or high current is required. These probes wear faster because beryllium-copper is softer than tungsten or tungsten-rhenium. Typical epoxy-ring probes have a beam and taper which is where the probe comes out of the epoxy-ring and begins to taper towards its tip size. At the junction of the taper of the probe and 27

the probe tip, the probe is bent so it points downward at the point to be contacted. The part of the probe which is pointed downward is called the probe tip. An example probe beam and taper lengths is a few hundred mils and a probe tip length of around ten mils. Probe tips can be increased in length with the consequence that they become more fragile [1]. Probes should be sized to be around 30-40% of the size of the bond pad to which they are to connect [2]. Edge sensors are sets of two or three probes which form a closed circuit between two probes when they are not in contact with the device being probed and form an open circuit when they are in contact with the device to be probed. Edge sensors are commonly used in industry to automatically probe test wafers (the machine doing the probing can tell when contact is made). For people that are experienced with using probe cards, edge sensors will likely not be necessary, but we found that having edge sensors helped us tremendously in determining when contact was made and determining if there were any planarity errors between the probes. Edge sensors can be designed to have an isolated tip (3 probe configuration) to avoid interfering with the device under test in case it contacts a conductive part of the device under tests [2].

Appendix D: SU-8 Molds Fabrication Process Fabrication Process for SU-8 Molds 1. Clean a Si wafer (approximately 2”×2.5”) with Acetone, Methanol and DI water; blow dry with N2, bake on the hot plate @120 °C for at least 5 min; 2. Pour an even layer of PR 2035 on the cleaned Si wafer (make sure the first attempt covers the center area); 3. Make the PR reach the edge of the chip by tilting it slightly (note that wrinkles are BAD!); 4. Spin on PR with the speed shown in Fig 4.3 (ramp_1 = 100 rpm/sec determines how even the layer would be, ramp_2 = 300 rpm/sec determines the final layer thickness);

28

4500

4000

4000

4000 Spin Speed /rpm

3500 3000 2500

Ramp_2

2000 1500

600

1000

600

500 0 Ramp_1 0 0

0 20

40

60

80

100

Spin Time /sec

Fig. 4.3 Spin speed to achieve an even SU-8 layer 5. Soft bake @60-65 °C for 2-5 min (it’ll even defects our, if any); then @90-95 °C for 15 min (until the PR is hard); let sample cool down to room temperature; 6. Expose with UV-34 filter for 190 mJ/cm2; 7. Post bake @60-65 °C for 1 min, then @90-95 °C for 10 min, then ramp down to 60-65 °C for 2-3 min; let sample cool down to room temperature; if the exposure is adequate, the outline of the features will show up after 1min of 90-95 °C baking; 8. Use PGMEA(Propylene Glycol Methyl Ether Acetate) as developer, develop until unexposed area is clear (usually take 5-8 min); 9. PGMEA goes to solvent waste bottle immediately after use; 10. Before use, the mold can be cleaned by rinsing it with Methanol, then blow dry with N2. And the PDMS channel fabrication process is: 1. Take a plastic container (with lid) and put it on the scale, then turn it on (this would zero the scale reading); 2. Measure PDMS and its curing agent to a weight ratio of 10:1; 3. Stir slowly with a disposal stirrer to mix well; 4. Degas the mixture carefully until clear; 5. Pour adequate amount of the mixture into a mold loaded tray, and make the mold as parallel to the fluid (gel) level as possible; 6. Degas again; 7. Put the tray in the oven and bake @85 °C for 2 hours, or @65 °C for 4 hours to reduce its rigidity; 8. Use a Methanol cleaned razor blade to cut the cured PDMS and gently peel the desired area off; (for short-time storage, put the PDMS on a piece of cleaned Si wafer);

29

9. Always throw away the first piece peeled off from a mold, it could only offer rough channel edges; 10. If done, seal the container and put it in the fridge (it could only last a few days, even in the fridge); if not, go to step 5 to make more pieces of PDMS.

30

Appendix E: LEAC Chip Fabrication Flow LEAC sensor fabrication flow 1a) Buried detector fabrication for dry etching a) etch the whole sample down to the desired thickness, poly SI etching rate is ~1.3nm/sec, not very accurate, varies everytime, need filmmatrix to confirm MicroRIE: O2 Plasma, 40sccm, 40 W, 30secs Place chip into the chamber with glass slips around it CF4+8%O2, 40sccm, 50 W, starting pressure 108mT, time = d/1.3nm b) BD PR mask Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin S1818 (positive), 6000rpm, 30secs Softbake: 100C, 2mins Remove edge bead with razor Exposure: total 191.4mJ, Right: 13.2mj, 14.5secs. Left: 10mj 19secs. Develop: AZ400K, 17secs or more till all red color is gone. Avoid over-etching. Do not hesitate to rinse sample with DI water and inspect sample under microscope during development process. Postbake (reflow): 120C, 1min c) Etching MicroRIE: [O2 Plasma, 40sccm, 40 W, 30secs] Place chip into the chamber with glass slips around it CF4+8%O2, 40sccm, 50 W, starting pressure 108mT, time = d/1.3nm d) Wash off PR with acetone, methanol and water

1.5) SiO2 layer deposition and planarization a) Deposit 300-350nm thick SiO2 b) Planarization with BD features Take the AZ2070 PR out from refrigerator >1 hr before fabrication (never use AZ2070 older than 1 day outside) Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin AZ2070 (negative), 500-(10secs)-4000rpm, total 50secs Softbake: 100C, 2mins Remove edge bead with razor, critical step, make sure the edge bead is gone 31

Soak in water, 2mins Exposure: Left: 10mj, 30secs?? (underdevelop it to get “enlarged” feature) Postbake (harden): 110C, 100secs Wet etching: 10-15 secs?? O2 plasma clean, 40sccm, 40W, 1min c) Measured it using Alpha step to make sure the feature is as expected

1b) Buried detector fabrication using KOH etching a) Deposit 10-15 nm of SiNx on sample surface using PECVE. Total deposition time is one minute. Follow PECVD instructions carefully. b) After deposition, acetone, methanol, water rinse sample. c) Make BD PR mask Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin S1818 (positive), 6000rpm, 30secs Softbake: 100C, 2mins Remove edge bead with razor Exposure: total 191.4mJ, Right: 13.2mj, 14.5secs. Left: 10mj 19secs. Develop: AZ400K, 17secs or more till all red color is gone. Avoid over-etching. Do not hesitate to rinse sample with DI water and inspect sample under microscope during development process. Postbake (reflow): 120C, 1min d) Remove SiNx layer not covering buried detector. Micro RIE etch for 30 seconds using CF4 Plasma at 25 sccm and 25 W. e) Selectively remove SiNx covering buried detector. Heat 85% H2PO4 to 110 degrees C, by setting hot plate to 180 degrees C. Place sample in acid bath for 2 minutes. Wear safety goggles/mask and let others know that hot acid is on hot plate. f) Remove sample from acid bath and rinse with DI water for 30 seconds to remove any remaining acid.

32

1c) Buried detector fabrication for Si substrate / SOI planar detector a) Deposit 250 nm of SiO2 layer using PECVD b) BD PR mask, negative tone Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature Spin 2070 PR 4000rpm, 50secs Softbake: 100C, 2mins Remove edge bead with razor, critical step, make sure the edge bead is gone Exposure: Left: 10mj, 45 seconds Postbake (harden): 110C, 100secs Develop: AZ300MIF, ~30secs and then examine under microscope and then further develop till all PR is gone O2 plasma clean, 40sccm, 40W, 1min c) Etch in buffered HF (1:6) for 10 seconds. Examine under microscope to make sure Si02 removed. If possible, use filmetrics. d) Wash off PR with acetone, methanol and water

2) Metal pads fabrication a) metal pads PR mask Take the AZ2070 PR out from refrigerator >1 hr before fabrication (never use AZ2070 older than 1 day outside) Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin AZ2070 (negative), 4000rpm, total 50secs Softbake: 100C, 2mins Remove edge bead with razor, critical step, make sure the edge bead is gone Exposure: Left: 10mj, 45 seconds (overdevelop it to get “shrink” feature) Postbake (harden): 110C, 100secs Develop: AZ300MIF, ~30secs and then examine under microscope and then further develop till all PR is gone O2 plasma clean, 40sccm, 40W, 1min

33

b) E-Beam Deposition : follow E-Beam instruction. 300 nm aluminum. c) Wash off PR with Acetone (airbrush), methanol and water

3) Deposit SiO2 and SiNx (See Operating Procedure). Deposition rate is roughly 15 nm/minute for both SiO2 and SiNx. We need 1.5 microns. 100-200 nm of SiNx deposition depending on desired waveguide modes.

PECVD Operating Procedure

Version 1.0

This document will provide detailed instruction for depositing thin films of silicon Nitride (SiNx) and silicon dioxide (SiO2) with the (Plasma-enhanced chemical vapor deposition) PECVD located in the white room. Deposition 1. Turn on the main circuit breaker (located on the bottom left of the equipment), press the “electric safety test” button and if the system passes the safety test, the red light will lit; 2. Flip on remaining switches: pump, module supply, RF generation and heater, as well as the power strip on the left side of the PECVD; 3. Turn on the chiller from the main switch on the back and press “On/OFF” on the front panel until temperature is display on the front; 4. Turn on the “MAIN ON” switch and “HEAT ON” switch on the temperature controller to increase the plate temperature to 250°C; 5. On the CPU-500 unite, press “ROTARY PUMP” then “PUMP DOWN”, make sure the yellow LEDs are lit; 6. Go to the service corridor and manually turn on the pump by holding the “CYCLE” switch to the left for 5 seconds, pump will be noisy when it starts; 7. Make sure the toggle switch on the top of the equipment is on “CVD” side; 8. Press “Left”, “MAN” to start a manual run; 9. At this time, turn on the vacuum gauge controller and monitor the pressure indicator (#1) on the vacuum gauge controller, the pressure should be ~150mTorr within 1 minute; 10. After pumping down, by pressing the numbers and “ENTER”, input “min=0, sec=0” for process time and “min=20, sec=0” for purge time, choose “P2” for purge gas to start a 34

dummy run to let the system enter an idle mode, in which “VENT or RESTART” is prompt on the screen; 11. Press “VENT” and wait 20 seconds until the gas flow can be heard near the CVD chamber, then press “VENT” again to turn it off; 12. Simultaneously press both green “UP” buttons (located in the middle of the two chambers) and lift the CVD chamber lid; 13. Load the sample(s) with metal tweezers, make sure the sleeves of your suit or the gloves are not touching the 250°C plate; 14. Close the lid by simultaneously pressing both red “DOWN” buttons; 15. Press “RESTART”; 16. Press “Left”, “MAN” to start another manual run; 17. After pumping down, input “min=20, sec=0” for process time (long enough to adjust the gas flow and chamber pressure, can be changed during the process) and “min=5, sec=0” for purge time, choose process gases (G1, G2, etc) and “P2” for purge gas; 18. Make sure all gas controller had the flow “off”; 19. Press “RUN” and make sure the countdown timer is running; 20. Turn on the process gas flow one by one, adjust the flow rate to desired values; 21. Adjust the dial that controls the opening of the valve between the chamber and the pump, until desired chamber pressure is achieved; 22. Turn on the RF matching network (ATX-600) and the RF generator (RFX-600), check the RF power set point and adjust it if necessary; 23. Press “RF ON” on the control unit CPU-500; 24. Press “TIMERESET” and input the new process time calculated based on the “total film thickness / deposition rate”; 25. Turn on the RF power on the RF generator by pressing “RF ON” on RFX-600; 26. Check the color of the gas plasma through the windows on the chamber; 27. Fill out process log sheet; 28. The timer will turn off the RF and start the purge process once the timer goes to 0; 29. After the process is down, repeat step #11-15 to take out the sample(s); 30. If this is not the end of the process go back to step #16; 31. If this is the end of the process, press “PUMP DOWN” to turn off the pump; 32. Then turn off all switches, power strip and the chiller. Cleaning routine A. Electrode cleaning If flake particles are seen in the chamber or on the shower head, and this is very common when microns or more of film has been deposited in the chamber: 1. Make sure the plate temperature is below 80°C; 2. Spray methanol on a clean wipe and wipe the electrode plate and the shower head; 3. Wet another clean wipe with DI water and wipe again to remove the residual methanol. 35

B. Flush A flush run should be done at the end of the day when all process is done, to remove the gases in the pipes and flow sensors: 1. When system is under idle mode, press “RESTART”; 2. Press “Left”, “FLUSH” (instead of “MAN”); 3. Use “min=20, sec=0” for process time; 4. Turn off the system when the flush is done. C. O2+CF4 clean An O2+CF4 clean should be done after 5-6 process: 1. Use “O2+CF4” (pre-mixed gas) as process gas and “min=20, sec=0” for process time; 2. Use 200W RF power; 3. Tick on the “clean” column in the logbook to indicate a cleaning process is done.

4) Waveguide fabrication b) Waceguide PR mask Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin S1818 (positive), 6000rpm, 30secs Softbake: 100C, 2mins Remove edge bead with razor Exposure: Left: 10mj 19secs??. Develop: AZ400K, 17secs or more till red color is gone Postbake (reflow): 120C, 1min c) Etching MicroRIE: [O2 Plasma, 40sccm, 40 W, 30secs] Place chip into the chamber with glass slips around it CF4+8%O2, 40sccm, 40 W, etching rate 75nm/min, 1.25nm/sec, 85secs for 100nm d) Wash off PR with acetone, methanol and water

36

5) Open via a) Via PR mask Take the AZ2070 PR out from refrigerator >1 hr before fabrication (never use AZ2070 older than 1 day outside) Clean the sample with acetone, methanol and water Prebake, 120C, 2mins, let it cool to room temperature [apply PMDS, 3000rpm, 30secs] Spin AZ2070 (negative), 500-(10secs)-4000rpm, total 50secs Softbake: 100C, 2mins Remove edge bead with razor, critical step, make sure the edge bead is gone Soak in water, 2mins Exposure: Right: 14mj, 30secs (not overdeveloped) Postbake (harden): 110C, 100secs Develop: AZ300MIF, ~30secs and then examine under microscope and then further develop till all PR is gone O2 plasma clean, 40sccm, 40W, 1min b) Wet etching Make HF(48%):H2O = 1:1 (24% HF) Dip the sample into HF for 8secs (plastic tweezers) c) MircoRIE [O2 Plasma, 40sccm, 40 W, 30secs] Place chip into the chamber with glass slips around it CF4+8%O2, 40sccm, 40 W, 5mins d) Wash off PR with Acetone (airbrush), methanol and water

37

Appendix F: LEAC Fabrication Diagrams

38

39

Appendix G: Gantt Chart for Fall Semester

40

41

42

43