AAWSAP DIRD, Biosensors and BioMEMS: A Survey of the Present Field, March 2010

UNCLASSIFIED//FQA QFFl&IAL t:l!H!! 8HL I Defense Intelligence Reference Document Acquisition Threat Support 31 March 2010 ICOD: 1 December 2009 DIA-08- 1003-020 Biosensors and BioMEMS: A Survey of the Present Field UNCLASSIFIED//FQR QFFIQlal.ls W&&i &P•Llt UNCLASSIFIED/ /POR: OPPl@IAL HS& 0Nk¥ Biosensors and BioMEMS: A Survey of the Present Field Prepared by: Acquisition Support Division (DW0-3) Defense Warning Office Directorate for Analysis Defense Intelligence Agency AAP Person 65 Author: Administrative Note COPYRIGHT WARNING: Further dissemination of the photographs in this publication is not authorized . This product is one in a series of advanced technology reports produced in FY 2009 under t he Defense Intelligence Agency, Defense Warning Office's Advanced Aerospace Weapon System Applications (AAWSA) Program. Comments or 1 uestions pertaining to this document should be addressed to!AAP Person 1 , AAWSA Program Manager, Defense Intelligence Agency, ATTN : CLAR/DWO-3, Bldg 6000, Washington, DC 20340-5 100. UNCLASSIFIED/ {FOR OFFICIO! !!ii Qllk¥ ii UNCLASSIFIED/ / P'OR: OP'P'l@IAL WS& 0Nk¥ Contents Introduction ........................................................................................................... vi What is BioMEMS? .................................................................................................. 1 Characteristics of BioMEMS ................................................................................ 2 Categories of BioMEMS ....................................................................................... 3 BioMEMS Micromachines ........................................................................................ 3 Brain-Implanted BioMEMS Micromachine Neuroelectrodes ................................ 4 Fluidic BioMEMS ..................................................................................................... 6 Drug-Delivery Pumps ......................................................................................... 7 What is a Biosensor? .............................................................................................. 9 BioMEMS Implantable Sensors ............................................................................. 10 MEMS Blood Glucose Sensors ............................................................................... 13 Commercial Blood Glucose Sensors .................................................................. 14 Enzyme-Based Biosensors................................................................................ 14 Thermopile Implantable Glucose Sensors ........................................................ 15 Neuroengineering by BioMEMS............................................................................. 20 Bioelectric Events............................................................................................. 20 Bionics and Neurointerfaces................................................................................. 22 Bioelectrodes and BioMEMS ............................................................................. 22 Microelectrode Array Fabrication ..................................................................... 24 Brain-Machine Interfaces ................................................................................. 25 Polymer BioMEMS Electrodes ........................................................................... 26 Retinal Devices Using BioMEMS........................................................................ 27 Microscale in Therapeutic Neurostimulation..................................................... 30 MEMS in Microfluidics ........................................................................................... 32 Lab Chips ......................................................................................................... 33 UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ iii UNCLASSIFIED/ / P'OR: OP'P'l@IAL WS& 0Nk¥ NASA Space Applications for Microfluidic Systems ........................................... 34 Microcantilever MEMs Sensors ............................................................................. 36 Conclusion ............................................................................................................ 39 Figures Figure 1. Varieties of BioMEMS and Sensors........................................................... 2 Figure 3. Schematic Diagram of a MEMS Motor Complete With a Method of Figure 7. An ISFET (Ion Sensitive Field Effect Transistor) Needle-Type pH Sensor Figure 9. Control of an Insulin-Delivery System by an Implantable Glucose Figure 2. Functional MEMS Micromachines Attached to Sensing Bioelectrodes....... 5 Converting Oscillatory Motion to Linear Motion ....................................... 5 Figure 4. Illustration of the Basic Physical Principle of the BioMEMS Motor ........... 6 Figure 5. Close-up Views of the Micro-Motor Gear .................................................. 6 Figure 6. An Insulin MEMS Pump............................................................................ 8 for Monitoring Tissue Physiologic Status ............................................... 11 Figure 8. A Fiber Optical Oxygen Sensor .............................................................. 12 Sensor ................................................................................................... 13 Figure 10. Glucose Oxidase .................................................................................. 14 Figure 11. An Experimental MEMS Blood Glucose Sensor ..................................... 15 Figure 12. Photomicrograph ................................................................................. 16 Figure 13. A MEMS Thermopile Glucose Sensor .................................................... 17 Figure 15. Thermopile Glucose Sensor (left) With Functional Illustration (right) Figure 16. Photograph and Illustration of the Thermopile Glucose Sensor in a Figure 18. A Neural Cell Showing Dendritic Inter-Connections That are Electrically Figure 14. Illustration of the Thermocouple Principle .......................................... 18 are Made Using Photolithographic Techniques .................................... 19 Catheter .............................................................................................. 19 Figure 17. A MEMS Biopotential Electrode System................................................ 20 Active With Other Cells........................................................................ 21 Figure 19. A Neuroprosthetic Interface ................................................................ 22 Figure 20. Scanning Electron Micrograph of a Brain Electrode Array Manufactured From Titanium and Produced by a Process of Electrodischarge Figure 21. Scanning Electron Micrograph of a 1141 Electrode Array Made to Be Figure 22. Polymer Based Cortical Penetrating Neuroelectrodes Made by Processes Figure 23. A MEMS Microelectrode Array Implanted Into the Cortex of a Rat Brain Figure 26. An Illustration of the Retinal Neuroprosthesis Created by the Boston Figure 27. Power Transfer by Magnetic Induction to an Implanted BioMEMS Retinal Figure 28. A New Generation of Implantable Neurostimulation Devices Can Pass Figure 29. Illustration of the Internal Construction of the Ultrasound Powered Machining ............................................................................................ 23 Inserted Into the Surface of the Human Brain .................................... 25 of Thin Film Deposition and RF Etching ............................................... 26 ............................................................................................................ 27 Figure 24. A Representation of the Retinal Prosthesis ......................................... 28 Figure 25. A BioMEMS Fabricated 4x4 Microelectrode Retinal Array..................... 28 Implant Project ................................................................................... 29 Prosthesis ........................................................................................... 30 Through the Lumen of a Syringe Needle.............................................. 31 Neurostimulator Developed at ASU ..................................................... 32 UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ iv UNCLASSIFIED/ /P"OR: OP"P"l@IAL HS& 91'11.¥ Figure 30. Microfluidics in Glass........................................................................... 32 Figure 31. Commercial Lab-Chip Devices That Allow Analysis for Biomarkers of Figure 32. Design of an Automated Cell Culture System for NASA Space Applications Incorporating Microfluidics and Integrated Optical Various Diseases Using On-Chip Electrophoresis in Microscale........... 34 Components ........................................................................................ 35 Figure 33. NASA Space Application of Microfluidic Cell Culture System Made by Figure 34. Principle of a Microcantilever That Bends When It is Loaded With an Figure 35. Scanning Electron Micrograph Showing a Cantilever Beam With a Single Micromachining................................................................................... 36 Adherent Mass .................................................................................... 37 Vaccinia Virus Particle ......................................................................... 38 Figure 36. Antibodies Attached to a Cantilever Array Create Different Specificities to Substances in Blood That are Indicative of Cancer.......................... 38 UNCLASSIFIED/ (FOR OFFICIO! 1155 01!1.¥ V UNCLASSIFIED/ /P'OR: OP'P'l@IAL WS& 0Nk¥ Biosensors and BioMEMS: A Survey of the Present Field Introduction Biomedical sensors and implantable devices employing micro­ electromechanical technology will be important to the future of biomedicine. Once just science fiction, the notion of a bionic man is now becoming a reality through the application of microscale technologies. Today it is not unusual to have a friend or family member who has an implantable device, such as a pacemaker, defibrillator, cochlear implant, biosensor, neurostimulator, or insulin pump. We have only seen the beginning of this technology development, and today's state-of-the-art implantable devices will be seen as crude and cumbersome tomorrow. With the rapid pace of development, it is probable that within the next two decades we will have as many medical treatments based on microdevices as there are pharmaceutical solutions today. This paper examines some of the most recent advances in the field, as well as the technologies that are likely to appear in the next few years. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ vi UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ What is BioM EMS? BioMEMS stands for biomedical micro-_glectro-mechanical ~ystems. It is a name applied to biological and medical devices that are created using advanced fabrication processes that allow the devices to be very small relative to comparable devices produced by traditional techniques. BioMEMS devices can also exploit the microscale to provide new functions that are not practical or possible in large-scale devices. The name applies to an exceptionally wide variety of engineered devices that derive from electrical, mechanical, chemical, and molecular engineering. The name distinguishes these from nanoMEMS which are submicron in scale such as carbon nanotube structures. Recently BioMEMS has become something of a misnomer as many of the latest technologies are being designed and developed based on nanoscale technologies which are many times smaller than microscale technologies. While current devices are manufactured mostly on the microscale, many of the functioning parts and the materials they operate on are at the nanoscale level. NanoMEMS for biomedical applications are mostly carbon-based materials that have emerged as prime materials because of their favorable mechanical and electrical properties. Carbon-based nanostructures such as graphene exhibit a high Young's modulus (stiffness), high strength, low density, low friction and large surface area. The low friction of a carbon nanotube allows production of practically frictionless bearings and has thus been a huge motivation towards applications such as nanomotors. Carbon nanostructures are much stronger than steel, which allows carbon-based materials to meet high-stress demands in biomedical applications such as weight-bearing prosthetics (like hip-joint or bone replacements), where other materials would fail. The field of BioMEMS encompasses micro devices that are often but not exclusively made by the same photolithographic techniques used to make computer chips. Their applications include neuroprosthetics, sensors and actuators, and microchemistry systems. A microchemistry system, often called a lab on a chip, can analyze chemical properties of a very small quantity of material such as a tiny blood sample. Advanced systems can perform several tests on the sample at one time. There are also drug-delivery systems, miniature hearing aids, artificial retinas, DNA analysis systems, cancer diagnostics, and an amazing variety of devices which support the function of the human body. Figure 1 shows some devices that were developed by the faculty of Biomedical Engineering at Arizona State University. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 1 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Varieties of BioMEMs and Sensors --- - -· ·-•· - - . I . . I . - -· - -· ---· Figure 1. Varietie.s of BioMEMS and Sensors. These BioMEMS devices were made at Arizona State University (ASU). They represent many of the areas of BioMEMS research. CHARACTERISTICS OF BIOMEMS Conceiving and designing BioMEMS devices requires a different perspective of the physical world. These devices can operate on principles such as capillary force, van der Waals forces, and electric-field forces. These forces become relatively strong when the size scale reaches very small dimensions. In the realm of the very small, the force of gravity is far less important than electrical charge and viscosity. From the perspective of microscale devices, engineers need to think about accomplishing tasks on an extremely subtle scale but with an exceptionally effective result. BioMEMS employs engineering sciences in ways that are more than just scaling down familiar devices used in biomedical applications. Rather, employment of new structures and new materials including polymers and biological components is necessary. In addition we need to be concerned about things like biocompatibility (the effect of the device on body tissue) as well as the degradation effects of tissues and body fluids on the device itself. Development of a new product that is targeted for implantation into the human body is particularly expensive due to federal regulation of medical devices. Devices must be shown to be effective for their intended use, and above all devices must be proved safe. Bringing a new device to market is not unlike the development of a new drug. In order UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 2 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ to gain federal approval there may be biocompatibility issues to resolve, clinical trials to perform, and Food and Drug Administration (FDA) requirements to satisfy. CATEGORIES OF BIOMEMS We divide the BioMEMS field into several subcategories. Some devices are targeted to be implanted into the human body or are applied to the body. Many of these are being developed as alternatives to drugs which therapeutically treat illness or disease. There is also activity in using microscale stimulation devices to treat neurological disorders such as epilepsy and Parkinson's disease and to treat neurological injuries such as stroke or trauma to the spinal cord. Another promising field of research is directed towards integration of the human brain with microelectrical components. For example implants are being developed which allow volitional (thought) control of machines. Most of this effort is targeted towards rehabilitation of individuals who are quadriplegics in order to allow them some control over their environment. Then there are devices being developed for clinical applications such as rapid blood analysis, and there are those targeted for benchtop biological research. In this review, the focus is on specific microscale and MEMS-based devices that are used in medicine and biology. These include: • Micromachines that interface to brain electrodes. • Blood glucose sensors. • Microfluidics. • Neural Interfaces. • Neurostimulators. • Microbeam sensors. Several of these technologies have been the subject of research by the present author and by colleagues at Arizona State University. BioMEMS Micromachines Micromachines made by photolithography are among the most complex and sophisticated of all MEMS devices. They have seen some application in various forms of biomedical devices where motion must be achieved with an implantable device. Motion within the human body by a device is difficult since it implies sliding surfaces, a need for electrical power, and long term reliability and stability. Biomedical implants are usually introduced by surgery and so once implanted cannot be easily removed for servicing. UNCLASSIFIED/ (FOR OFFICIO! P!Sli Qllk¥ 3 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Only a few BioMEMS applications are well known for implanted devices and these are used for positioning and repositioning of sensors. A specific example is application to brain-electrode systems. BRAIN-IMPLANTED BIOMEMS MICROMACHINE NEUROELECTRODES There are applications in biomedicine and research where very small electrodes are implanted into the brain. These microelectrodes are used in the cortex (surface layer) to detect electrical activity associated with the volitional desire to move some part of the body. Electrical activity is recorded as a very small change in voltage within the tissue. Near the top of the human head and about 2 cm beneath the scalp lie the parts of the brain where nerve cells (neurons) are found which control the muscles of the body. Specific locations in the cortex are associated with specific parts of the body. When a person moves a limb, there can be detected a corresponding electrical activity of these neurons. Neuroscientists have recorded these signals and developed a kind of map of the brain that defines what brain cells actuate certain muscles of the body. An interesting phenomenon occurs in which some of these brain cells become active even when there is intent to move a limb but no actual movement occurs. In a healthy person the intent to move can be detected by measurement of microvolt signals from the brain cells about 120 milliseconds before any muscle movement occurs. In a person who has lost a limb or has become paralyzed, the intent to move can still be detected in the brain even though there is no limb movement. Thus in principle there is an ability to electrically record from the brain and determine a person's intent to move. Monitoring of the brain is accomplished by using an array of implanted microelectrodes whose signals anticipate movement of specific limbs. These signals can in turn be used to control machines. This is the idea behind advanced devices that allow quadriplegics to interact with their environment. Signals from implanted electrodes are used by computers to control robotic actuators. A problem recording the signals occurs because the body naturally tends to encapsulate the electrodes with scar tissue, meaning the electrodes lose electrical contact with the neurons. A solution to this problem incorporates a very small electromechanical actuator attached to the electrode. This device allows the implanted electrodes to change their position occasionally in order to continue to monitor neural events. By moving the thin rod-like electrode up or down a small distance after several months of implantation, the useful lifetime of the electrode array can be greatly extended. The BioMEMS device described here is meant to be implanted under the skull and on top of the brain cortex. It allows fine adjustment of less than a millimeter in order to make sure the electrode system, once implanted, is able to contact the desired brain cells even after the electrode has been encapsulated by scar tissue over time. Figure 2 shows a photomicrograph of a gear-driven micromachine that was made through collaboration between ASU and Sandia National Laboratories. The device is driven by an electrostatic vibrating comb motor . Under a microscope it can be seen that the combs move in an oscillatory fashion at about 40 Hz when energized. Electrostatic forces between two blade-like combs a few tens of microns apart are operative with about 15 volts as the electric field source. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 4 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 8NI.¥ The Neural Probe chip enables precise bi-directional positioning of the microelectrodes in the brain with a step resolution in the order of 8.8 µm. The thermal microactuators allow for a movement of the microelectrodes of up to 5 mm in either direction making it suitable for positioning microelectrodes in deep structures of a rodent brain. The rest of the mechanism converts an oscillatory motion to a linear motion through a ratchet-type configuration. Figure 3 illustrates the basic unit where there are two comb­ like arrays on either side. These generate the actual force. They are connected to a cross piece that converts oscillatory motion in concert with a spring to a pulsatile linear motion of the vertical shaft. This connects to some gears which do a mechanical transformation. Figure 2. Functional MEMS Micromachines Attached to Sensing Bioelectrodes. The gears have features as small as 50 microns, less than the size of a period on a printed page. (Courtesy of Dr. J. Muthuswamy, Bioengineering Department, ASU) , fon•11, ~---...~ Figure 3. Schematic Diagram of a MEMS Motor Complete With a Method of Converting Oscillatory Motion to Linear Motion. (Courtesy of Dr. J. Muthuswamy, Bioengineering Department, ASU) UNCLASSIFIED/ {FOR OFFICIO! !Pfli ODIi.¥ 5 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Figure 4 illustrates the basic physics where the force is generated. A voltage V, is applied to the horizontal sliding shaft. There arises an electrical field (shown as the curved arrow) that tends to pull the shaft into the cavity between the two outer combs. This is the basic motive force of the device. The force generated by each comb is very small so there are constructed many similar structures that form a comb-array. All of the forces act in parallel. (1) V (3) Figure 4. Illustration of the Basic Physical Principle of the BioMEMS Motor (Muthuswamy et al.) The device is fabricated by using Sandia's Ultraplanar Multi-level MEMS Technology (SUMMiTV) process, a 5-layer polysilicon micromachining technology. The layers are chemically sensitive to different processes and can be selectively removed by etching with different reagents. For example silicon can be removed using an etchant which preferentially removes silicon only along certain crystalline planes. Other layers can be undermined through the use of acids. The components of the micromachine are patterned on the silicon using photoresist masks. A mask is a covering which prevents parts of the silicon from being etched away. Figure 5 gives an idea of the sophistication of this MEMS fabrication technique. The end result is that the long rod at the right of the picture in Figure 3 moves up and down with a speed of about 1-2 mm per second. The speed can be adjusted through the number of teeth designed on the reduction gears. Single unit recordings were obtained from the somatosensory cortex of adult rats over a period of three days demonstrating the feasibility of this technology. This device has been implanted in rats and its development supported by the neuroprostheses program of the US National Institutes of Health. Fluidic BioMEMS The application of microscale fabrication techniques has allowed the manipulation of very small volumes of fluids on the nanoliter and even picoliter scales. This allows 6 UNCLASSIFIED/ (FOR OFFICIO! P!Sli Qllk¥ Figure s. Close-Up Views of the Micro-Motor Gears (Sandia National Labs) UNCLASSIFIED//P'OR: OP'Pl@IAL WS& 0NI.¥ chemical processes such as separations, reactions, and analysis to be conducted with very small amounts of sample. According to a forecast by the Nexus Task Force, the market for BioMEMS is expected to reach $18 Billion in 2005. The commercial success of these devices and the technical potential of other BioMEMS has driven research in a number of areas. As a result, over the past few years, fluidic BioMEMS devices have become the largest and most diverse applications of MEMS devices. Fluidic BioMEMS now include: • Drug-delivering neuroprobes. • Biosensors (general). • Bioreactors. • Cell-handling devices. • Drug-delivery devices. • Micro-chromatography systems. • Microfluidics. • Molecular detection/handling. • Neural interface devices. • Optical/retinal sensing. • Surgical devices. • Tissue-handling devices. There are obviously application overlaps within these devices and some are integrated with others to create system-level or multi-sensing devices. Applications run the gamut of the imagination including identification of bacterial or viral agents, drug testing, home testing, environmental safety and security, and drug-delivery technologies. DRUG-DELIVERY PUMPS Drug delivery to the human body for medical therapeutic purposes has long been by swallowing substances or by injection. However this approach creates rapid rises and falls in drug concentration in the blood stream because the dose is typically introduced all at one time. The blood titer of a drug may thus peak at undesirably high levels in an effort to sustain the drug action over a longer period of time. Treatment of many medical conditions, such as cancer by chemotherapeutic drugs would best be accomplished by a steady and long sustained infusion of the drug into the blood. This would keep the tumor-fighting drug at some optimal concentration and so make it maximally effective. High levels of chemotherapeutic drug levels in the blood are toxic and levels that are below some threshold are ineffective. UNCLASSIFIED/ (FOR OFFICIO! P!Sli QDII.¥ 7 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Methods of creating a more sustained blood concentration of a given drug are desirable to increase drug effectiveness. Similarly there are other medical applications for controlled-release drug-delivery systems, such as that for insulin, where it is desirable to maintain a more sustained dose over a period of hours. There has been the application of MEMS technologies in the creation of better ways to deliver drugs to the human body by way of small implantable reservoirs that can slowly release their drug over a prolonged period of time. Figure 6 shows a MEMS-based insulin-delivery pump for the treatment of diabetes. This is perhaps one of the largest applications of BioMEMS in medicine. Subcutaneous drug delivery Debiotech p iezoelec ric in sulin pump S itzerland http://lllww.tfebiotech.com/dobiotoch.h • MEMS technology using Si and glass biocompatibility • Precise control of nl volumes • Prevent under/overdosing and detect occlusions, other problems • Small size: 114th the size of existing monitors • Proprietary hermetic packaging • In he U.S. alone, 60 million people are affec e by diabe es • 15% of worldwide heal h spending goes award rea ing diabe es. • 450,000 people now wear ransportable insulin pumps worldwide. sources: Debiotech ebsite· Advanced Packaging Figure 6. An Insulin MEMS Pump. An implantable drug-delivery device. Diabetes is a disorder in which glucose (blood sugar) is not properly taken into the cells of the body. This process is normally mediated by the hormone insulin, which is produced by the islet cells of the pancreas. Insulin circulates in the blood to actuate a receptor on tissue cells that causes them to uptake glucose from the blood stream. If there is too little insulin (as in Type 1 diabetes) the cells will not uptake glucose from the blood, the cells will starve, and blood sugar levels become too high from unabsorbed glucose. Alternatively, cells may not respond properly to normal insulin levels (as in Type 2 diabetes) and require artificially higher levels of insulin to uptake glucose. This is called insulin resistance. In either case, blood insulin levels can be increased by using insulin injection devices. Brittle diabetes is a condition where blood glucose levels fluctuate wildly with eating or fasting causing hyperglycemia (too high a blood glucose-more than about 100 mg percent) or too low (hypoglycemia-less than about 60 mg percent). Brittle diabetics UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 8 UNCLASSIFIED//P'OR: OP'Pl@IAL WS& 0NI.¥ have a significant problem in controlling their blood sugar and sometimes insulin doses are not adjusted finely enough or frequently enough to maintain the blood glucose at healthy levels. This form of diabetes is responsible for fainting (ketoacidosis) and other serious symptoms such as poor blood circulation to the limbs. Poor circulation can result in diabetic ulcers and may sometimes necessitate amputation of the limb. Insulin pumps are often worn by brittle diabetics because a slow infusion of insulin works better in stabilizing blood glucose levels rather than periodic injections. A belt­ worn insulin-delivery system looks something like an old-style audio player. These systems are typically controlled by a small screw-type pump powered by batteries. The reservoir carries several milliliters of insulin for dispensing over an extended period of time. The injection needle is connected (underneath the clothes) to a catheter and then to the pump. The systems are reasonably effective but are cumbersome and require the needle to continuously reside subcutaneously in the abdomen. A MEMS implanted insulin pump is a less cumbersome and perhaps more convenient means of slowly infusing insulin at a programmed rate. The implanted device is refilled periodically by introducing a needle through the skin and tissue to a septum in the device. The MEMS insulin pump shown in Figure 6 is surgically placed under the skin. It has a rubber septum on the top for filling with insulin. This implementation has a piezoelectric element that moves in response to electrical charge. When actuated by a timer the element oscillates, creating a pressure inside the device that dispenses insulin. The device is programmable for dispensing at various rates. Although the implantable MEMS pump was developed for diabetes, the pump has application to the slow measured delivery of many other drugs including 5-fluoruracil used for cancer therapy and theophyllne for treatment of asthma. What is a Biosensor? A primary application of BioMEMS is in the creation of sensors for blood chemistry and other biophysical parameters of the human body. We can define a Biosensor as: • A sensor whose application is primarily in the measurement of quantities within a biological system, such as chemical, electrical, and physical parameters. • A sensor incorporating a biological component (enzymes, living cells, antibodies) typically used to measure chemical concentration. This definition does not require that the sensor be deployed within a biological system. The two definitions which are somewhat different have their origin with different influential investigators who wrote textbooks in the early days of this field. The latter definition is prevalent in Europe. A biosensor is normally constructed by immobilizing a biologically active material (such as an enzyme) onto an electrical sensor that measures a fundamental physical quantity like electrical current, voltage, mechanical strain, temperature, or frequency. The specific sensor material is chosen because it reacts to a desired measurand (such as UNCLASSIFIED/ {FOR OFFICIO! !PEii ODIi.¥ 9 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ concentration of glucose) and thus causes a change in one of these fundamental quantities. Electrical sensors can be inherently small and so we have a combination device with biology and electrical sensing in a small and compact form. When compared to much larger bench top machines, microscale configurations of sensors usually have a superior performance because they are less prone to various interferences such as power-line noise and they have much shorter diffusion distances for sensed molecules. Some physical quantities such as micro-degrees of temperature change are much easier to measure over small distances. BioMEMS Implantable Sensors Biosensors placed inside the body, or in-vivo, measure biological parameters such as blood pH, oxygen, carbon dioxide, and blood glucose. (Blood pH is a measure of the acidity of the blood.) These parameters are the most important in medicine since these are all independent blood chemistries that give a moment-to-moment insight into the physiological state of a living being. Since the early days of the space program NASA has been interested in ways of noninvasively monitoring these blood parameters in astronauts for the instant assessment of their physiological condition. Until recently, blood withdrawal was the only accurate and reliable way to obtain such information. Indwelling sensors that use needle penetrations are now available but still have trouble with accuracy and longevity. The military has also been interested in the assessment of the state of readiness of a soldier which is reflected in his blood chemistry. An exhausted solider will show a highly acidic blood pH (less than about 7.3). Remote electronic readout of biosensor information of a soldier to a central command center is presently the stuff of science fiction movies but reflects real desires of the military. Figure 7 shows a photo of a pH sensor (which measures concentration of hydrogen ions). It is designed as a needle for tissue insertion or placement at the end of a catheter for introduction into the blood stream. The pointed tip is inserted into the medium to be measured, and the electrical signal and power supply are connected to the device at the gold film contacts at the base. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 10 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Vgs 1: reference 2: gate oxide 3: insulating resin b) 4: channnel S: source D: drain B: bulk Vgs a) V r Vdr MOSFET ISFET Figure 7. An ISFET (Ion Sensitive Field Effect Transistor) Needle-Type pH Sensor for Monitoring Tissue Physiologic Status (http://www.ee.seikei .ac.jp/~seiichi/lecture/Biomedical/09/09-biosensor.html) The sensor element is an ion sensitive field effect transistor (ISFET). This device is micromachined from silicon by using standard semiconductor photolithography processes. The actual active sensor surface is near the tip of the device where there is a thin film of silicon nitride covering the gate region of the transistor. The gate is sensitive to very small changes in electric field, and these in turn cause relatively large changes (or gain) in the current flow through the transistor. At the center in Figure 7 is a standard transistor diagram showing the gate and its relation to the other contacts. On the right is shown the pH sensitive ISFET system that measures an electric field change across the nitride membrane. The membrane is placed directly in contact with the liquid to be measured. It is not known the exact process that occurs at the membrane that gives rise to electric field shifts on the gate. Presumably hydrogen ions reversibly adsorb onto the interface between the nitride membrane and the solution causing local electric field changes on the gate. This device is not meant for permanent implantation but rather short term applications in research or medical surgery where the sensor is used over just a few hours. Proteins and other biological molecules adhere to the gate region of the ISFET and cause a slow UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 11 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ drift and offset in the reading and so shifting the calibration. Depending on its exact construction some versions however can be effective for as much as a few days. Figure 8 shows the latest in microdevice oxygen sensors. (This particular version is not specifically a MEMS device since it is not made by photolithography and it is neither .§lectrical nor mechanical in nature.) It represents however the trend in oxygen sensors towards optical measurement techniques that are evolving towards optical systems based on MEMS. These sensors can be implemented in small sizes with optical fibers that transmit and receive light from the sensor surface at the tip. Variations on this basic principle are used for indwelling blood oxygen catheters during surgery. fiber op cs silicone cover Figure 8. A Fiber Optical Oxygen Sensor Showing Two Optical Fibers Mated to Plastic Prisms That are Coated With an Oxygen-Sensor Coating, Usually a Fluorescent Ruthenium Compound (http://www.imtek.de/content/projekte_en. php?ls= 11) The oxygen sensors respond to changes in absorption of oxygen into the special coating at the tip, or in some versions, they respond to fluorescence of a thin oxygen-sensitive coating placed on the prism surface. The secret to good performance lies in the proprietary selection of the specific chemical species selected for the coating. Blue light is usually transmitted down the optical fiber and the coating glows orange in proportion to the amount of oxygen present. The light from the glow is conducted back toward the supporting instrumentation system. The intensity of the orange light then indicates oxygen levels in the solution. The particular sensor shown in Figure 8 uses two optical fibers. The one on the left side is not sensitive to oxygen, but rather acts as a reference system that is used to compensate for changes temperature, ambient light, and other system variables. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 12 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 8NI.¥ So far, despite decades of work, no continuous monitoring sensor that is self contained and implantable has been successful for long term usage. The human body is an exceptionally hostile environment for foreign materials. Even our most advanced biosensors fail over a period of days or weeks if continuously exposed to the body environment. This failure is mostly due to the sensor chemistry wearing out in one way or another or attack of the sensor interface by the body's immune system. Failure is not generally attributed to the materials of which the sensor is made. MEMS Blood Glucose Sensors Diabetes is an enormous world problem. A significant fraction of the annual world health care expenditures can be traced back to diabetes, and its cousin, obesity. Measurement of blood glucose on a regular basis, usually 3-4 times a day, allows a diabetic to regulate his diet and insulin dose to achieve normal glucose levels. Perhaps one of the most needed biosensors is that for an in-vivo blood glucose sensor. These could be used to automatically control of the output of an insulin-delivery pump. This would constitute a major improvement in the treatment of diabetes by automatically using blood glucose concentration feedback to stabilize the blood glucose levels with the exact level of needed insulin. This kind of system is sometimes called an artificial endocrine pancreas since it mimics the normal function of the pancreas in regulating insulin release. There are several attempts at using MEMS devices to produce indwelling glucose sensors. These are presently on the market, but currently the sensors have relatively short lifetimes, are disposable, and require replacement every few days. Figure 9 shows this basic idea. An implantable sensor produces an electrical output that reports the blood glucose concentration. Its signal is transmitted to a receiver and then processed by a computer to drive a belt-worn pump. The creation of such a sensor has been a daunting problem from more than 30 years. The most recent attempts have been in the use of micromachining to produce optical devices of very small size and to produce wireless transmitting electrochemical sensor devices small enough to be injected into the body• Figure 9. Control of an Insulin-Delivery System by an Implantable Glucose Sensor (Medtronic Inc.) Coincident with the development of sensors has been the need to develop very small wireless telemetry systems that transmit the sensor data to outside the body. UNCLASSIFIED/ {FOR OFFICIO! !Pfli ODIi.¥ 13 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ COMMERCIAL BLOOD GLUCOSE SENSORS The present method of blood glucose monitoring depends on a needle puncture of the skin to withdraw a drop of blood to place on a color changing test strip. The test strip is read by a small handheld reader. Glucose test strips do not have to necessarily be small, but by making the sensor very small the amount of blood required for the test is reduced. The sensors often are made by techniques of photolithography or in some cases by microscale screen printing in order to achieve reproducibility. The skin-puncture test is painful and time consuming, and thus noninvasiveness is the key desired characteristic of glucose sensors. Research is being directed at a noninvasive glucose sensor that is accurate enough to work external to the body and through the skin. Short term (a few days) wearable needle glucose sensors are available from major companies like Medtronic Inc. but fall short of the convenience of a noninvasive sensor. ENZYME-BASED BIOSENSORS The key component in most biosensors is a reactive chemistry on the sensor surface. The sensor chemistry is chosen to give it specificity to only one analyte (such as glucose). The concentration of the analyte is determined by a sensor that can directly measure the analyte reaction products reacting with the sensor surface. For example sensor chemistries to measure glucose are often based on glucose oxidase enzyme which promotes a chemical reaction at the sensor surface. Glucose oxidase enzyme complex structure is seen in Figure 10. Glucose oxidase catalyzes the reaction: glucose + 0 2 (glucose oxidase) ➔ gluconolactone + H2O2 + heat (79 kJ/mole) This reaction of glucose with oxygen (from the air) occurs in the presence of glucose oxidase enzyme. The enzyme itself is a catalyst to the reaction and so is not consumed. Rather it presents favorable conditions and ability to transfer electrons on its molecular structure for glucose and oxygen to come together to react. Figure 10 shows the enzyme structure. Typically in a sensor the enzyme is a large molecule and can be trapped in a porous gel and thus is not able to diffuse away from the sensor surface. Glucose and oxygen, being small molecules, can diffuse through the gel to the enzyme whereby the reaction occurs and the reaction products will diffuse away. The enzyme is unconsumed and the reaction process is continuous as long as glucose is present. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ Figure 10. Glucose Oxidase Enzymes Like Glucose Oxidase are Large Folded Molecules That Act as Catalysts for a Chemical Reaction (http://www. innovations- report.de/bilder _neu/17279_gluc.jpg) 14 UNCLASSIFIED//P'OR: OP'Pl@IAL WS& 0NI.¥ An enzyme-gel is applied in a thin layer over a electrically biased metal electrode. An electrical current flows through the electrode when the hydrogen peroxide produced by the glucose reaction is decomposed and causes an electrical current to flow. If there is a fixed concentration of glucose present, a corresponding amount of peroxide will be produced, and this can be measured by the electrical current flow. The amount of peroxide measured is an indicator of the glucose concentration. The actual amount of analyte consumed by a small BioMEMS sensor is exceedingly small and has no significant effect on the local concentration. Figure 11 shows a MEMS implementation of a blood glucose sensor using this kind of approach (Advanced Biosensors Inc). The fine needle geometry is generally meant for insertion into tissue. This device is comprised of multiple independent small sensors and is actually a system of components that adds signal processing and interface electronics to allow its communication with insulin pumps for automated delivery of insulin based on patient need. Advanced BioSensors blood glucose sensor - Flip chip assembly ASICs, power sources and circui componen s Separa e sensors, moun wi h electrical connec ions, add he biocompa ible polymer, assemble he sensor pa ch. The sensor based on glucose oxidase reaction wi h blood plasma from capillaries in he dermis. - biochemical coa ing - sensor in place 3 - 7 days - Docking par for elec ronics assembly wi h amplifi er, ADC, wireless transceiver, power source - encryp ed digi al da a to wristwa ch-sized morn or/recorder module In fu ure combine the CG S wi h an insulin pump long x 200 u Ide o produce a closed-loop sys em-an artificial pancreas. . • - ' • •· -~-s . . - • - - - 2- Figure 11. An Experimental MEMS Blood Glucose Sensor Glucose sensors based on electrical measurement of hydrogen peroxide products have been found to have the significant problem of a drifting baseline over time. It is difficult to maintain calibration of the sensors, and the electrode is affected by proteins and other substances in the blood stream. Although this sensing approach has had some success, its longevity in the body is much less than desired; sensors based on this approach decay alter a few days of use. THERMOPILE IMPLANTABLE GLUCOSE SENSORS Towe et al. at Arizona State University have been working on an improved approach to blood glucose sensing. It depends on the fact that the glucose oxidase reaction generates a small amount of heat as indicated in the above chemical reaction. The UNCLASSIFIED/ (EAR OFFICIO! !Pfli 0111.¥ 15 UNCLASSIFIED//P"OR: OP"P"l@IAL WSli 01'11.¥ released chemical energy warms the surrounding environment to a degree that depends on the energy released in the exothermic reaction, which is characteristic of the particular chemistry and the amount of reactants. The heating is then dependent on the glucose concentration. Figure 12 shows these devices. They are made by MEMS techniques involving the •--· processes of using patterning by exposure to light through a mask (photolithography), metal deposition by evaporation, and sputter etching. The result is the fabrication of sensors in large arrays with many identical sensors on a single substrate. The figure is a ·--­ photomicrograph of a part of a wafer production run with nine identical devices. Each one of the devices is less than 1 mm in size and would be cut apart and used independently. The ·--- ·--· darker rectangular areas at the top of ·•- each device are the electrical bonding pads where wires attach to the device to readout its electrical response. Figure 13 shows the basic process in the Figure 12_Photomicrograph 3 x 3 Array of sensor operation. The sensor is planar Thermoelectric Glucose Sensors Constructed by and composed of a layered structure Towe et al. at ASU consisting of an enzyme gel and a thermoelectric temperature sensor. Glucose from the local medium naturally diffuses to the gel and the temperature sensor underlies a thin film of enzyme gel that contains the glucose oxidase. A second enzyme called catalase is used to secondarily break down hydrogen peroxide produced by the first reaction. UNCLASSIFIED/ (FOR OFFICJA 1 1155 01!1.¥ 16 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Figure 13. A MEMS Thermopile Glucose Sensor (Towe et al., Arizona State University) Heat energy released in the gel is measured by the thermopile and is characterized in terms of kiloJoules (kJ) per mole of glucose consumed. At concentrations of glucose that are common in the blood stream, the temperature rise is very small, on the order of a hundredth of a degree (10 millidegrees). Measuring this temperature change in the presence of ambient body heat or room­ temperature changes of a few degrees is a daunting task, but can be done. The measurement of the small reaction temperature in the face of possible ambient temperature shifts is accomplished by the use of a differential temperature measurement system. This sensor employs what is known as the Seebeck Effect whereby dissimilar metals heated at their junction produce a voltage that is proportional to temperature. This concept is shown in Figure 14. This principle is then implemented as an array of metal junctions that when connected in series produce a higher output in order to detect smaller temperature changes. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 17 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Figure 14. Illustration of the Thermocouple Principle (Guilbeau and Towe, ASU) Figure 15 on the left panel shows a photomicrograph of a thermopile temperature sensor that is made from the metals antimony and bismuth. These metals are vacuum deposited as alternating thin films using a masking process to form horizontal lines of conductive metal. Each pair of lines has a sensing junction and a reference junction. When they are heated on their sensing junction with respect to their reference end (that is exposed to the ambient temperature), they generate a small electrical voltage. Each junction is only a few tens of microns in size so it is possible to place dozens of junctions in series in order to increase the voltage signal. Figure 15 on the right panel shows that there are two banks of thermoelectric sensors with the central common junctions supporting an enzyme gel. Thermopiles are thin film temperature sensors and are differential in nature in that they measure temperature differences, and are thus sensitive enough to detect exceedingly small temperature rises. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 18 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 8NI.¥ Thin Mylar Substrate Figure 15. Thermopile Glucose Sensor (left) With Functional Illustration (right) are Made Using Photolithographic Techniques. This larger device is about 8 mm in length but has approximately 50 micron minimum geometries. (Towe et al. ) At ASU the glucose sensor has been constructed on a thin mylar plastic substrate and so can be curved into a cylinder to form a small tube. This tube is then sensitive to glucose concentrations on its outside exposure to the blood stream. Figure 16 shows a photograph and an illustration of a glucose sensor implemented in the form of a catheter. The electrical output signal is routed by wires down the length of the catheter for a remote readout. The device was implanted in a pig over a short duration and the decrease in blood glucose concentrations in response to administration of insulin were recorded and are shown in the right hand panel of Figure 16. 10 min Figure 16. Photograph and Illustration of the Thermopile Glucose Sensor in a Catheter. Its configuration is shown (left) and its response in detecting changes in blood glucose in a pig as a result of insulin administration to t he animal. (ASU resea rch.) Towe et al. This glucose sensor approach has been found to have limitations in stability when introduced into tissue or blood. Primarily the problems, as with many sensors, arise from the system chemistry and not as much from the electronic portion of the sensor. The glucose oxidase enzyme slowly decays over time and thus the sensor sensitivity and calibration drifts making it eventually unusable. These problems have no easy solution. Investigators have been working with various forms of blood chemistry sensors for forty years or more. Microfabricated UNCLASSIFIED/ {FOR OFFICIO! !Pfli ODIi.¥ 19 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 8NI.¥ thermoelectric sensors have improved reproducibility of the sensors, but there has only been limited success in applications of implantation into the body. Neuroengineering by BioMEMS BioMEMS types of devices, as we understand them today, were used first in neuroscience. There has been a long history of the study of the electrical nature of body tissue dating back to the days of Galvani and Volta. Frog nerves were found to be electrically stimulatable and the first recordings of bioelectrical events were accomplished very early with the invention of the string galvanometer. More recently, pointed wires inserted into muscle, nerve and brain have given way to MEMS electrode systems that are made on silicon supporting substrates and the processes of photolithography used to define electrical current pathways. Figure 17 shows a modern electrode system for detecting and recording electrical signals from living things. Each of the square regions is an exposed film of platinum while the thinner conductors that contact the pads are insulated by a thin layer of a plastic and conduct the detected electrical signals to a connector system (not shown). Figure 17. A MEMS Biopotential Electrode System. It is approximately 100 microns in width. Each square pad is an electrically sensitive region. (Wikipedia) BIOELECTRIC EVENTS The living processes of biology can in some ways be likened to that of a battery. The metabolism of life produces the charging of the battery while it discharges through a myriad of bioelectrical events that produce movement, thought, and cognition. The transfer of electrical charge in the form of ions, mostly sodium, potassium and chloride is initiated by bioelectrically excitable cells of the nerves, muscles, and neurons of the brain. Collectively they constitute the wiring of the body. These ions result from the salts that are part of the composition of living things that have evolved from the sea. The movements of these ions constitutes the flow of an electrical current in tissues and are called bioelectrical currents. These currents give rise to electrical potentials that can be detected from the skin or by biopotential electrodes inserted into organs or muscle. UNCLASSIFIED/ {FOR OFFICIO! !Pfli ODIi.¥ 20 UNCLASSIFIED//P'OR: OP'Pl@IAL WS& 0NI.¥ Bioelectrical events are generated in tissues when there is an electrochemical change in the membranes of specialized cells of the brain and nervous systems. Neuroelectrical devices interface to the brain and nervous systems through electrodes that touch excitable cell membrane and convert ionic current flows to electron flow in a wire. Conversely current flows in a wire are converted by electrodes to ionic flows in tissue. The function of the entire human body is under control of the brain and the nervous system. So the use of electrodes of various types as a method of monitoring and controlling the function of the nervous system through bioelectrical currents and electrodes is potentially a very powerful method of treating in therapeutic ways. Figure 18 suggests the idea of an excitable cell of the brain having many tendrils or dendrites that extend outward to interconnect with other cells. Essentially these cells are the relay points of the bioelectrical wiring of the human body. Figure 18. A Neural Cell Showing Dendritic Inter-Connections That are Electrically Active With Other Cells (artist's illustration) An electrical interface to body tissues typically occurs through the use of microelectrode systems. These electrodes are often on the order of tens of microns to millimeter-order in sizes depending on their function. For example electrodes that simply measure bioelectrical activity of the nervous system can resemble needles insulated along their length having micron order tips both because they carry only small electrical currents and also because they need to be placed in very specific places and they contact just a single or few cells that control specific functions. UNCLASSIFIED/ {FOR OFFICIO! !PEii ODIi.¥ 21 UNCLASSIFIED//P'OR: OP'P'l@IAL WS& 0Nk¥ Bionics and Neurointerfaces Neuroprosthetics (also called neural prosthetics) is a discipline related to neuroscience and biomedical engineering concerned with developing neural prostheses. Neural prostheses are devices that attempt to substitute for a motor, sensory or cognitive modality that might have been damaged as a result of an injury or a disease. Bionics and neurointerfacing are relatively new concepts. A neurointerface is a combination of software and hardware that translates electrode signals to something which is understood by an electronic system. The term bionics is used when we connect organic matter with something artificial (human-made). The design and development of neural interface electrodes to the brain or peripheral nerve has been going on for more than twenty years. Figure 19 shows a mockup of a neuroprosthetic Figure 19. A Neuroprosthetic Interface. Adapted to the needs of a quadriplegic for control of his environments electrode system for the human brain (Duke University) that is meant to tap into brain signals and then through the use of a computer allow a quadriplegic to control his environment. BIOELECTRODES AND BIOMEMS Electrodes that interface to the brain or nervous system are the most important components of neuroprosthetic systems. There are basically of two classes of electrodes. One class does recording of bioelectric events and the other provides current to stimulate bioelectrical responses. Both classes of electrodes need to be designed to be very robust for long term survival in the hostile environment of the human body. For this reason only the noble metals of platinum, iridium, gold, and to some extent the resistant metal tungsten are used for electrodes. Nearly all other metals would corrode in the warm saline environment of living tissues. Electronic exchange reactions occur at the electrode surfaces in both recording and stimulation modes. Many bioelectrodes are nothing more than fine wires insulated down their length and then exposed at a needle tip where an electrical contact exists to tissue. These are usually used in research for trying to understand brain function. Sometimes a single electrode performs both recording as well as stimulation but more frequently the electrode size and material is optimized to do one function or the other. UNCLASSIFIED/ {FOR OFFICIO! !PEii QDlk¥ 22 UNCLASSIFIED//P"OR: OP"P"l@IAL WSli 01'11.¥ Electrode arrays are sometimes used, such as in the case of deep brain stimulation for late-stage Parkinson's disease, which is characterized by tremors of hands. There is a "brain pacemaker" that sends electrical impulses to specific parts of the patient's brain via permanently inserted electrodes. This can stop the tremors for reasons that are not exactly known and for which the best location in a given patient to place the electrode is also not known beforehand. Thus introducing multiple electrodes and stimulating each one in-turn until the best result is found allows a greater degree of possible effectiveness and therapeutic result. MEMS techniques are being employed in this application to permit a greater number of contacts and to reduce the size of the lead wire on the electrodes. MEMS multichannel electrode stimulating and recording electrode systems are also used in brain interface applications to record or stimulate the activity of many neural circuits. Such electrodes allow ability to record or stimulate complex muscle movements associated with the limb such as in walking or in grasping objects with the hands. Figure 20 shows an electrode system looking much like a bed of nails. Figure 20. Scanning Electron Micrograph of a Brain Electrode Array Manufactured From Titanium and Produced by a Process of Electrodischarge Machining ("Electrical Discharge Machining and Chemical Etching", Fofonoff* , Martel, Hatsopoulos, Donoghue, Hunter) All muscles of the body are activated