Haptic Controlled MEMS Systems Seminar Report 2011-2012 Submitted in partial fulllment for the award of the Degree

of Bachelor of Technology in Electrical and Electronics By

Vivek Suresh Babu Thuravupala Univ Roll No. 65360 Under the guidance of Ms. Prathibha P.K. Mr. Jebin Francis

Department of Electrical and Electronics RAJAGIRI SCHOOL OF ENGINEERING AND TECHNOLOGY Rajagiri Valley, Cochin-682039 Kerala, India

CERTIFICATE

This is to certify that the report entitled Haptic Controlled MEMS Systems is a bonade record of the project done by Vivek Suresh Babu Thuravupala, of 7th semester Electrical and Electronics Engineering in partial fulllment of the requirements for the award of Degree of Bachelor of Technology in Electrical & Electronics Engineering of the Mahatma Gandhi University, Kottayam during the academic year 2011 2012.

Ms. Prathibha P.K. (guide) Asst Professor Dept. of Electrical & Electronics Engineering

Mr. Jebin Francis (guide) Lecturer Dept. of Electrical & Electronics Engineering

Place: Kakkanad Date: 19-12-2011

Prof. K R Varmah Professor & HOD Dept. of Electrical & Electronics Engineering

Abstract

Haptic technology is a form of feedback that takes advantage of a users sense of touch by applying forces, vibrations, and/or motions to the user. It allows the user to remotely perceive the feel of actually engaging in the actions he/she is emulating. This allows for remote perception of weight, force and other sensory experiences such as surface textures. When used in conjunction with control mechanics and remote operating systems, this allows for more reliable and natural operation which in turn reduces the chances of accidents and improves work quality. MEMS, which stands for microelectromechanical systems, refers to systems which operate on the micrometer scale. Generally, made up of components between 1 to 100 micrometers in size (i.e. 0.001 to 0.1 mm) and MEMS devices generally range in size from 20 micrometers (20 millionths of a meter) to a millimeter.

Acknowledgement

First and foremost I thank God almighty for giving me the physical and mental strength to complete the seminar successfully. At the successful completion of my seminar, I would like to lend my sincere thanks and profound sense of gratitude to the principal of Rajagiri School of Engineering and Technology, Dr J. Issac. I solicit sincere thanks to respected Head of the Department Prof. K R Varmah, for his constant encouragement and valuable suggestions. I would like to express my deep sense of gratitude to the seminar co-ordinators Mr. Jebin Francis and Ms. Prathibha P. K. for their whole hearted support. I extend my sincere thanks to my guides, Ms. Prathibha P. K. and Mr. Jebin Francis, for their kind co-operation and unfailing guidance throughout the completion of this seminar. Also I extend my cheerful thanks to all the sta members of Electrical and Electronics Department of my college for their helpful and timely suggestions. Finally, sincere thanks goes to my Parents and friends for their inspiration and prayer to complete this seminar.

Vivek Suresh Babu Thuravupala

ii

Contents
1 Overview 1.1 1.2 Haptics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 2 3 3 5 6 9 9

2 Haptics 2.1 2.2 2.3 General Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Haptic Rendering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 MEMS 3.1 3.2 3.3 General Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

MEMS Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Classications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 16

4 Haptic Controlled MEMS Systems 4.1 4.2

MEMS Microgripper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 MEMS Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4.2.1 4.2.2 Electrothermal Actuators . . . . . . . . . . . . . . . . . . . . . . . 17 Electrostatic Actuators . . . . . . . . . . . . . . . . . . . . . . . . 18

4.3 4.4

Haptic Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 23

5 Conclusion

iii

List of Figures
1.1 1.2 2.1 2.2 4.1 4.2 4.3 4.4 4.5 4.6 4.7 Example of a heightmap . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example of a Microelectromechanical System . . . . . . . . . . . . . . . . Augmented Reality Systems . . . . . . . . . . . . . . . . . . . . . . . . . . God-Object vs. Ruspini Rendering Algorithm . . . . . . . . . . . . . . . . Electrothermal Bent-beam Actuator Structure 1 2 5 7

. . . . . . . . . . . . . . . 18

Electrothermal Hot/Cold Arm Actuator Structure . . . . . . . . . . . . . 18 Electrothermal Microgripper . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Electrostatic Microgripper . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Common Commercial Haptic Controllers . . . . . . . . . . . . . . . . . . . 20 FEA Analysis of a Microgripper . . . . . . . . . . . . . . . . . . . . . . . . 21 FEA Analysis of Asymmetrical Crash . . . . . . . . . . . . . . . . . . . . 21

iv

List of Abbreviations

1. MEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microelectromechanical Systems 2. DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deoxyribonucleic Acid 3. HAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High Aspect Ratio 4. HRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Haptic Rendering Algorithm 5. HIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Haptics Interface Point 6. IHIP . . . . . . . . . . Ideal Haptic Interface Point, Intermediate Haptic Interface Point 7. FEA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Finite Element Analysis

Chapter 1

Overview
By combining Haptic and MEMS technologies, its possible to extend the ecient workspace of a human being into the micrometer scale. Haptic technology consists of using sensors to convert sensory information at the end eector into electrical signals, transmitting this signal to the controller, and recreating the sensory data in such a way that allows the user to remotely perceive things like vibrations and surface textures.

Figure 1.1: Example of a heightmap Using haptic technology, we can recreate the sensation of touching the surface shown in Figure 1.1, using just the heightmap corresponding to the surface.

1.1

Haptics

Currently existing haptic controllers, sensors and rendering algorithms are useful for producing haptic feedback at a single point. Such devices are available on the commercial market, such as the Novint Falcon. At the same time, other haptic devices allow navigation of virtual environments with accompanied sensory feedback. Future improvements in haptic technology holds the promise of complete immersion in virtual reality environments.

1.2

MEMS

MEMS technology is currently capable of realizing sensors like accelerometers, gyroscopes and pressure sensors, and characterizes the ecient use of space and energy. Microelectromechanical Systems are basically microscale reproductions of macroscopic scale objects. These devices are capable of replacing their corresponding macroscopic scale components.

Figure 1.2: Example of a Microelectromechanical System

Chapter 2

Haptics
2.1 General Overview

Although the word haptics probably doesnt ring a bell for most people, the technology is actually all around us already, theyre just being marketed under proprietary brand names. A major example of this is the Force Feedback technology introduced by Logitech for their racing controllers, which provides vibrative feedback to the user when the vehicle crashed or hit a bump. In theory, the concept of haptics is simple enough. Consider a mechanical end eector thats capable of mirroring human actions exactly, however the human is unable to perceive the environment the eector is in. So rst, sensors are necessary to convert the environmental variables around the eector into electrical signals so they can be transmitted to the user. Next, the signals need to be converted back to the appropriate physical stimulus once it reaches the user. This conversion can be done using electric, pneumatic or hydraulic actuators. The complexity of haptics emerges from the fact that the signals cant be converted directly into a faithful reproduction of the environmental variable which produced it. Then a processor needs to be added to modify the signal according to a certain set of rules, and accordingly convey the signal to the actuators. The set of rules is known as the Haptic Rendering Algorithm, and is used to ensure a high delity reproduction of the original sensory input at the users end. Common examples of simple haptic devices in our everyday lives are: Mobile Phones

Mobile phones with interfaces which lack physical buttons face the diculty that the user cant know when the button is pressed except through a visual stumulus. Hence, a haptic feedback is required everytime such a button is engaged, in this case, a vibratory or auditory feedback is provided. Its common in full touchscreen phones for the device to produce a small vibration or a small tone to signal to the user that an input has been received. Human beings tend to be impatient, and without this feedback, they would continue to press the button until they saw some reaction, or they may inadervently press the button while doing something else. Aircraft Servomechanisms The ailerons and ns on modern aircrafts are actuated by servomechanisms, which mimic the actions of the pilot. However, if the pilot is unable to feel the resistance these surfaces encounter in dierent positions, then he is unable to maintain safe ying conditions at all times. For example, if these control surfaces are in inappropriate conditions, they encounter a large amount of air resistance. If the pilot is unable to perceive this resistance, then the unsafe condition is only noticed after the plane begins to stall. Providing haptic feedback at the pilots controls allow the pilot to perceive dangerous ying conditions before theyre felt. Teleoperation of Robotic Equipment Haptic feedback has also been implemented in excavators and digging machines. In cases where these machines must operate in a mixed environment, such as when excavating in an area where soft clay is mixed with hard rocks, the operators are able to manipulate the machines more eectively around the hard rocks when feedback is provided. This allow the operators to move around unseen obstacles and thus, increases productivity and decreases the possibility of damage to the equipment. Augmented Reality Simulators For training surgeons or pilots, its impractical to train using live specimen or real airplanes. In these cases, simulators and augmented reality systems are used. The physical objects that the surgen practices on, or the plane the pilot learns with, are nonexistent. They only exist as a virtual model, and the interactive forces 4

perceived are produced by haptic rendering algorithms. In the case of a surgeon learning to operate on a patient, this allows them to feel the dierence between cutting through bone or esh without endangering anyone. In the case of a pilot learning to y, this allows them to learn without endangering any lives, and also enables them to learn to handle dangerous scenarios without putting themselves in danger.

Figure 2.1: Augmented Reality Systems allow surgeons to practice dangerous procedures without endangering any lives.

2.2

Types of Feedback

Depending on the type of information thats gathered and reproduced, haptic feedback can be divided into two major categories: Kinesthetic and Tactile feedback. Kinesthetic Feedback Consider grasping an object, for example, an orange. If your eyes are closed, then there are certain details about the orange which can be perceived by touch alone. For example, one can immediately realize that the general shape of the orange is spherical and can even estimate its size. This information is gathered based on the joint angles and miscle lengts of each joint in the hand, which is then processed by the brain. Using this data, its position relative to the body can be determined as well. All this data is considered kinesthetic data. Tactile Feedback In addition to the kinesthetic data perceived, addition information is gathered by the ngertips, which have receptors for dierent stimuli such as pressure and 5

vibration. From these, we can determine the texture of the orange, how hard our grip is, and what temperature the surface of the orange is at. All this data is considered tactile data. For eective haptic recreation of an object, both tactile and kinesthetic feedback must be present. However, due to the complexity of the hardware required, usually only one type of feedback is provided. Or in the case of certain interface devices, tactile feedback is provided, but only at a single point on the virtual body.

2.3

Haptic Rendering

In a case where the physical object doesnt actually exist, then there are no sensors which can produce electrical signals to be processed. In this case, a virtual model of the object is required, and the forces felt by the user are created by a Haptic Rendering Algorithm(HRA). These algorithms use a 3-Dimensional model of the object being simulated to determine the point of contact between the user and the object, if any, and then produces the appropriate forces along the appropriate actuators to make the user feel like he/she is actually interacting with the object. Two widely used algorithms are the God-Object Algorithm and the Ruspini Renderer Algorithm. Both of these are very similar, however the latter corrects a major algorithmic aw in the rst can produce errors. God-Object Renderer The God-Object Renderer was introduced in 1995, and only requires the position of the end eector in space and the virtual model to determine the forces which need to be produced. Haptic interfaces function upon the principle of force equality, which directly follows Newtons Law that any force is accompanied by an equal and opposite force. Then, when a person, or a robot for that matter, touches an object, the object also pushes back with that same force. Then, by determining this force and applying it on the controller, its possible to emulate the eects of actually touching that object, even if the object itself is virtual. The renderer uses the current position of the end eector in space to create a Haptic Interface Point (HIP) in the model of the object. In addition to this,

another point known as the Ideal Haptic Interface Point is also created, which is always outsite the object, and will never cross into the volume of the object. In a situation where the user is navigating space outsite the object, both the HIP and IHIP coincide and the user feels no resistance to motion. However, if the user navigates into the volume of the object itself, then the HIP moves into the object and the IHIP remains at the surface, thereby a dierence in the positions of these two entities emerges. Now assume that these two points are connected by a spring. Then when their poisitions coincide, the HIP feels no force due to the spring. However, as their positions diverge, the spring stretches and the HIP feels a force directed towards the IHIP. This force is then realized using actuatars and applied onto the controller, which is then perceived by the user as resistive force, meaning that the user is now in contact with the object. Depending on the strength of the hypothetical spring, the hardness of the surface can vary in perception from soft to hard. Ruspini Renderer In 1997, an improvement of the God-Object Renderer was introduced, known as the Ruspini Renderer. When using the God-Object Renderer, if there are any holes in the mesh model used, then at certain points, the user may be able to poke through the object, and wont be able to get out until a hole in the mesh is encountered again. This is the main aw of the God-Object Renderer, and is corrected by the Ruspini Renderer. In the God-Object Renderer, the IHIP is

Figure 2.2: (a) Positioning of the HIP and IHIP using God-Object Renderer when the user is touching the object. (b) Flaw in the God-Object Renderer which allow the user to enter the object through holes in the model mesh. (c) Ruspini Rendering Algorithm uses a sphere instead of a point for the IHIP to overcome the aw of the God-Object Renderer.

represented by a point which always remains outside the surface of the object. Then if a hole in the mesh is encountered which is larger than the IHIP, it will pass though to the inside of the object. The Ruspini Renderer uses a sphere instead of a point for the IHIP, thereby even if it encounters a hole in the mesh which is smaller than the IHIP sphere, the IHIP will remain outside the object. Then by selecting a IHIP larger than the polygon size of the mesh, this aw can be eliminated completely. However, if the IHIP sphere is too large, then loss of resolution will occur and ne surface details will not be perceptible.

Chapter 3

MEMS
3.1 General Overview

MEMS stands for microelectromechanical systems and refers to any mechanical system with components raning between 1 to 100 micrometers in size and operating using electical energy. These machines can resemble replicas of macroscopic scale machinery, but since theyre operating at such small scales, theyre exempted from the constructs of classical physics, and are more aected by surface phenomenon such as electrostatics and wetting. Sinces these structure are so small, their volume to surface area is much smaller than usual, hence volume eects such as inertia and mass arent as important when considering their mechanics. There are a wide range of applications of MEMS in science and engineering elds such as: Medicine 1. MEMS pressure sensors used to monitor a patients vital signs, such as blood pressure and respiration. 2. Disposable MEMS sensors used to monitor blood pressure, providing a low cost alternative to the reusable sensors used before. 3. Wireless MEMS sensors which can be implanted into a human body. Measurements from these devices can be obtained by remote scanning. Communications 9

1. MEMS inductors and tunable capacitors provide signicantly improved performance compared to their integrated counterparts. 2. Provide improved performace through use of MEMS devices at a lower cost and space requirement. 3. Radio Frequency MEMS applications in resonators as mechanical lters for communication circuits. Inertial Sensing 1. MEMS accelerometers and gyroscopes are quickly displacing conventional gyroscopes and accelerometers in applications such as crash safety systems. 2. MEMS sensors such as those mentioned above are also being used in consumer devices such as smart phones. Biotechnology 1. Biochips for detection of hazardous chemicals, drug screening and detection of biological agents. 2. Microsystems for DNA amplication and identication 3. Micromachined Scanning Tunneling Microscopes

3.2

MEMS Fabrication

The same methods used for fabrication of semiconductor devices cane be modied and used for fabrication of MEMS devices. The materials used for MEMS fabrication are: Silicon Silicon is already used for creating most integrated circuitry and it can also be adopted for creation of MEMS devices. Cheap availability of high quality silicon and ability to incorporate electronic functionality makes silicon a popular choice for fabrication. In addition, silicon is also an almost perfect Hookean material, so it produces almost no hysterisis when its exed, and hence almost no energy dissipation. This means a silicon device can have a service lifetime in the range of billions to trillions of cycles. 10

Polymers When materials are required in large volumes, synthetic polymers can be an alternative to silicon. These polymers can be produced in large quantities and with varying characteristics. Metals While silicon maybe the primary choice when it comes to MEMS fabrication, its also possible to use metals, as long as their operation is limited to within their limitations. Ceramics Ceramics can be used to produce sensors with piezoelectric properties which can have sensitivity to normal and shear forces. The types of ceramics used are generally nitrides of silicon, aluminium and titanium, allowing for a range of possible material properties. Using these materials, fabrication can be done in three methods, each of which adapts existing technology. The three methods are: Deposition The basic step in MEMS fabrication is the deposition of a thin lm of the base substrate. But just being able to deposit a lm isnt important, its the ability to control the thickness of the lm from a few nanometers to a 100 micrometers. The deposition process can be carried out in two methods: physical vapor desposition and chemical deposition. In pysical vapor deposition, the deposited material is acquired form a target and deposited onto the required surface. Sputtering is one method, in which an ion beam is used to liberate atoms from a target, and then move these to the desired substrate. Another method of physical vapor deposition is evaporation. Here the required atoms are evaporated from the target surface using hear or an electron beam in a vacuum. In chemical deposition, a controlled chemical reaction between the substrate and a gas is used to grow the required lm on the substrate surface. Two main processes are Low Pressure checmical vapor deposition abd Plasma Enchanced chemical 11

vapor deposition. In cases where oxide lms need to be produced, a technique known as thermal oxidation can be used, in which a silicon wafer is exposed to oxygen and steam to grow a thin lm of silicon dioxide. Patterning Patterning involves the transfer of a pattern onto a base material. The main method of doing this is known as Lithography. Lithography involves selectively exposing a photosensitive material to a radiation source. The photosensitive material then experiences some physical change when exposed to this radiation. This dierence in physical properties can then be used to our advantage. For example, consider a photosensitive lm which is exposed to light passed by a mask. The material is then shielded over an area which is identical to the shape of the mask, and the rest of the lm undergoes some change. Now this dierence in physical attributes can be used to remove the surrounding material, which then leaves us with the masked area, on a thin lm, revealing the underlying substrate. One type of lithography is Electron Beam Lithography(EBL). In EBL, a beam of electrons serves as the radiation source, and its scanned across a surface covered in a resist lm. The exposes areas of the resist is then removed, and very small structures can be produced in the resist using this method. Etching Etching is simliar to patterning, in that it uses a masking material. However, in patterning the substrate is on the surface. In etching, the surface acts as the mask and the base acts as the substrate. Ethcing can be broadly divided into two categories: dry etching and wet etching. In dry etching, the material to be etched is dissolved or removed using ractive ions or an etchant. Reactive ion echting is also called Plasma etching. In this process, the substrate is placed inside a reactor and several gases are introduced. The gases are then excited using an RF power source, and a plasma is formed. These ions have high enough energy levels to knock atoms out of the substrate. In addition, the ions also react with the surface. Hence, dry etching is composed of a chemical and a physical process. A proper balance between these two processes must be maintained to obtain the desired etching characteristics. 12

A subclass of dry etching is deep reactive ion etching, which involves al alternating composition of ion gases in the chamber. This method provides almost vertical sidewalls to depts of hundreds of micrometers. In wet etching, the substrate material is selectively removed by dipping it into a solution. A mask is used to determine the areas which are removed. Dierent classications of wet etching can be determined based on the speed of etching, and whether or not its dependent on direction. If the speed of etching is the same in all directions throughout the material, then its known as isotropic etching. In this case, long narrow holes in the mask will produce v-shaped grooves in the silicon, with atomically smooth surfaces if carried our correctly. Some crystal materials will have dierent etching rates depending on the crystallographic orientation in which the etching is carried out. This type of etching is known as anisotropic etching, and can be used to produce pyramid shaped holes in the substrate, with 54.7 walls. After a large number of devices have been prepared on a silicon wafer, the individual dies have to be separated. The parts of the wafer that are used and how theyre separated determine the manufacturing technology. The scale at which these fabrication methods are applied and the methods in which they are applied can be divided into the following categories: Bulk Micromachining In bulk micromachning, the entire volume of the substrate is used for MEMS fabrication. Usually, silicon wafers are used as the substrate, since its crystallographic structure allow for anisotrophic etching. This enables the productino of high regular structures, as the dissolution rates in dierent angles can vary by as much as a thousand times. The wet etching usually uses alkaline liquid solvents to dissolve silicon exposed by photolithography. Bulk micromachining starts with a silicon wafer or other substrates which is selectively etched, using photolithography to transfer a pattern from a mask to the surface. Like surface micromachining, bulk micromachining can be performed with wet or dry etches, although the most common etch in silicon is the anisotropic wet etch. This etch takes advantage of the fact that silicon has a crystal structure, 13

which means its atoms are all arranged periodically in lines and planes. Certain planes have weaker bonds and are more susceptible to etching. The etch results in pits that have angled walls, with the angle being a function of the crystal orientation of the substrate. This type of etching is inexpensive and is generally used in early, low-budget research. Surface Micomachining Unlike Bulk Micromachining which uses the entire substrate volume, surface micromachining uses successively deposited layers. A layer of polysilicon is used as a substrate layer, and is covered by a sacricial layer of silicon dioxide. This sacricial layer is used as a mask when fabricatino processes are applied. Once this layer is machined, the next layer is deposited and the process continues until the entire structure is machined. The sacricial layer is later removed leaving just the structural layers, which form the micromachine. For example, a suspended cantilever can be built by depositing and structuring a sacricial layer, which is then selectively removed at the locations where the future beams must be attached to the substrate. The structural layer is then deposited on top of the polymer and structured to dene the beams. Finally, the sacricial layer is removed to release the beams, using a selective etch process that will not damage the structural layer. So in Surface Micromachining, the properties of the substrate are not as important as in bulk micromachining, since the structures are built on top of the substrate and not inside it. In addition, this method can be applied to larger surfaces than silicon wafers, such as displays and thin lm solar cells. The deposition can even take place on non-rigid materials to allow for exible constructions. High aspect ratio silicon micromachining HAR micromachining is a method adapted specically for the machining of structures with a high aspect ratio, which in this case is considered the ratio of the width of an object to its height. A structure machined using this method can have thicknesses up to 100 m.

14

3.3

Classications

The fabricated devices can be classied by their use as: Sensors The most common MEMS sensors and transducers are pressure sensors, vibration sensors, gyroscopes and accelerometers. Due to the small scale of the MEMS architecture, sensors such as gyroscopes and accelerometers are not aected by inertia, so the forces on a displaced weight can be used to measure the motion of the object within which the sensor is embedded. Actuators Actuators are used in MEMS systems for microrobotic applications, such as microgrippers. These generally consist or a large array of linear actuators which together provide the required force to move the gripper. Depending on the structure, linear and curvilinear motion can be produced. Structures The sensors and actuators cant create a MEMS system on their own. In addition to these, structures are required to provide support for the sensors and actuators, necessary mounting points, and containment.

15

Chapter 4

Haptic Controlled MEMS Systems

Generally, MEMS systems are restricted to accelerometers, gyroscopes and other devices, other advances systems are built for testing. However, manipulation at the microscopic scale can be made possible using Haptics enabled MEMS systems. One basic example of this would be a microscopic realization of a robotic gripper. Robotic grippers are used widely today, ranging from serving as automated industrial robots to remote controlled arms. The design of these grippers is simple, featuring two ngers which move together to grasp an object. Realization of this arm on a microscopic scale allows for manipulation of microscopic objects ranging from cells to larger biological specimen, or MEMS structures, sensors and actuators.

4.1

MEMS Microgripper

At the micrometer scale, the gripper begins to closely resemble a pair of tweezers. To accomplish its goal of manipulating objects eectively on the micrometer scale, it needs to be able to manipulate objects without damaging them, so it must have precise control over the forces it exerts. In cases where the gripper is manipulating biological specimen, it must also be able to operate in a liquid media, and move the cells as desired without causing any damage. In order to achieve these goals, its desirable for the operator to be able to feel the same forces felt by the gripper when handling the objects.

16

4.2

MEMS Actuators

The actuators used in the microgripper can be divided into two broad categories: Electrothermal and Electrostatic. The former uses dierences in expansion due to temperature to produce motion, while the latter uses electrostatic forces to produce motion.

4.2.1

Electrothermal Actuators

The main principle of operation for electrothermal actuators is based on the thermal expansion of elastic exures or beams. Consider the simple linear actuator shown in Figure 4.1. The four lateral beams connecting the movable shuttle to the xed bases will expand as temperature increases. So then, by connecting a voltage source as shown in the gure, between the two base points, a small current will ow through the lateral beams. This ow will produce a heating, the magnitude of which can be controlled by controlling the current, by varying the applied voltage. As these lateral beams heat up, they will expand, which will cause the moving shuttle to displace in the direction indicated by the arrow in the gure. As the voltage is now decrease, the current ow decreases and the temperature of the beams also decreases, causing them to shrink back to their normal size. This will cause the movable shuttle to displace in the direction opposite to the one shown in the gure. This type of actuator is known as a Bent-beam, or V-shape, actuator due to its shape. Another type of electrothermal actuator is the Hot/cold arm, or U-shaped, actuator. An example of a U-shaped actuator is shown in Figure 4.2. In the U-type actuator, a radial motion is produced due to the dierent temperatures in the hot and cold arms of the actuator. As shown in the gure, the thicker of the two arms is called the cold arm. Since the heating is produced by the ow of current, the heating eect will be lesser for a thicker arm. Hence, even though the current through the two arms are the same, the temperature of the hot arm will be much greater than the cold arm, and so will the change in length. Since the hot arm expands more than the cold arm, the end of the arm moves in the direction indicated in the gure. This type of actuator can be used to form a gripper as shown in Figure 4.3.

17

Figure 4.1: Electrothermal Bent-beam Actuator Structure.

Figure 4.2: Electrothermal Hot/Cold Arm Actuator Structure

4.2.2

Electrostatic Actuators

Another type of MEMS actuator is known as a Electrostatic Actuator, which uses the electrostatic forces between two plates to produce motion. This type of actuator has intersecting plates with protruding teeth, as shown in Figure 4.4. One of these plates

18

Figure 4.3: Electrothermal Microgripper using two U-shaped actuators is anchored, while the other is movale. As the voltage applied between these two plates is increased, the attractive forces between them also increases, and eventually causes the movable plate to move closer to the anchored plate. When a large number of these plates with interlocking teeth are connected in unison, the force due to a small voltage can be amplied enough to produce linear motion in the gripper. Since this mechanism can only serve to move the plates closer together, and cant be used to move the plates apart, an elastic mechanism must be included at the mounting point for the movable nger to ensure that when the applied voltage is reduced, the movable plate will pull back away from the anchored plate.

Figure 4.4: Electrostatic Microgripper using an array of electrostatic actuators

19

4.3

Haptic Controllers

Haptic Controllers are the actual devices which are used to control the MEMS systems and where the user feels the feedback. They give a force-feedback to the user in response to computer generated values. This makes haptic controllers perfect for interacting with micrometer scale objects. Using this, a person can assemble machinery using a MEMS controller, sort cells and perform micrometer pick-and-place operations. In order to to able to manipulate objects eectively, a haptic controller requires two types of control: Mesoscale and Microscale. Mesoscale control means having the ability to control the position of the entire positioning system itself, so an object can be moved over large distances. Microscale control means having the ability to control the position of the object very precisely in a small area, without moving the actual actuator itself. Both types of control are necessary to move a microscale object from one poisition to another eectively. An actuator with only mesoscale control abilities will be unable to accurately move objects on the microscale, while a controller with only microscale control abilities will be restricted to working in a small area.

Figure 4.5: Commercial Haptic Controllers: (Right) The Novint Falkon and (Left) The Phantom Omni Common haptic controllers on the market now are the Novint Falkon, which is a 3-axis 3-Dimensional haptic controller, and the Phantom Omni, a 6-degrees of freedom haptic controller which applies a rendering algorithm at its tip.

4.4

Finite Element Analysis

Although we approximate the MEMS structures as rigid bodies, in actual pratice, these are susceptible to bend when placed under a certain amount of pressure. So, in order

20

to ensure that the device will remain operational under the required conditions, FEA techniques are applied to a simulated model of the device before fabrication takes place. FEA analysis basically uses a 3D mesh representing the device, along with parameters which describe the material such as its resistivity, Youngs modulus, thermal conductivity etc., to generate a simulated model. This model can then tested under dierent conditions to determine the areas where the device is most stressed and how the device will react in specic conditions.

Figure 4.6: FEA Analysis of a Microgripper showing input voltage vs. displacement in a simulation

Figure 4.7: FEA Analysis of Asymmetrical Crash showing dierence in vertex density based on the areas of interest

21

FEA is used widely in the industry to simulate interactions which would be too costly or impractical to repeatedly test in reality. For example consider the asymmetric collision test performed on a car shown in Figure 4.7. In any simulation, the amount of unnecessary processing must be reduced to increase ecieny. So when carrying out FEA analysis, the mesh is specically generated so that the areas of interest will have a more detailed mesh while the remaining areas will have less detail. In the case of the car, the area which is experiencing the crash has much more vertices than the back of the car which will most likely be unaected by the crash. Similarly, when testing on a MEMS device, the density of vertices will be distributed dierently for each test, as shown in Figure 4.6.

22

Chapter 5

Conclusion
Further concentration developing the integration of Haptics into MEMS will allow for groundbreaking discoveries such as miniature factories. In a lecture by Richard P. Feynman given in 1959, he talks about the obsession with making things larger, and how making things smaller should be given more consideration. Now more than half a decade later, we still havent reached his vision of micrometer scale factories and robots. However, with further development, this technology holds the promise to provide novel solutions to some of our most pressing problems.

23

Bibliography
[1] Feynmans Talk. Aug. 2011. url: http://www.zyvex.com/nanotech/feynman. html. [2] Finite Element Method. Aug. 2011. url: http : / / en . wikipedia . org / wiki / Finite_element_method. [3] Haptic Technology. Aug. 2011. url: http://en.wikipedia.org/wiki/Haptic_ technology. [4] William Harris. How Haptic Technology Works. Aug. 2011. url: http : / / electronics . howstuffworks . com / gadgets / other - gadgets / haptic technology.htm.

[5] Microelectromechanical Systems. Aug. 2011. url: http : / / en . wikipedia . org / wiki/Microelectromechanical_systems. [6] [7] [8] H3D Tutorials. Haptic Rendering Algorithms. Aug. 2011. url: https://sites. google.com/site/h3dtutorials/haptic_rendering_algorithms. Ashwin Vijayasai. Haptic controlled three-axis MEMS gripper system. Oct. 2010. url: http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=5614847. Ashwin Vijayasai. Haptic Controlled X-Y-Z MEMS Gripper System. Aug. 2010. url: http://dspace.lib.ttu.edu/etd/handle/2346/ETD-TTU-2010-08-848.

24