UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

MEMS-enabled micro-electro-discharge machining (M³EDM) Alla Chaitanya, Chakravarty Reddy 2008

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

Item Metadata

Download

Media
24-ubc_2008_fall_alla_chaitanya_chakravarty_reddy.pdf [ 1.98MB ]
Metadata
JSON: 24-1.0066936.json
JSON-LD: 24-1.0066936-ld.json
RDF/XML (Pretty): 24-1.0066936-rdf.xml
RDF/JSON: 24-1.0066936-rdf.json
Turtle: 24-1.0066936-turtle.txt
N-Triples: 24-1.0066936-rdf-ntriples.txt
Original Record: 24-1.0066936-source.json
Full Text
24-1.0066936-fulltext.txt
Citation
24-1.0066936.ris

Full Text

MEMS-Enabled Micro-Electro-Discharge Machining EDM) 3 (M  by Alla Chaitanya Chakravarty Reddy B.Tech, Jawaharlal Nehru Technological University, 2004  A THESIS SUBMITTED [N PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Applied Science in The Faculty of Graduate Studies (Electrical and Computer Engineering)  The University of British Columbia (Vancouver) October 2008 © Alla Chaitanya Chakravarty Reddy 2008  Abstract A MEMS-based micro-electro-discharge machining technique that is enabled by the actuation of micromachined planar electrodes defined on the surfaces of the workpiece is developed that eliminates the need of numerical control machines. electrodes actuated by hydrodynamic force is developed.  First, the planar  The electrode structures are  defined by patterning l8-tm-thick copper foil laminated on the stainless steel workpiece through an intermediate photoresist layer and released by sacrificial etching of the resist layer. The planer electrodes are constructed to be single layer structures without particular features underneath. All the patterning and sacrificial etching steps are performed using dry-film photoresists towards achieving high scalability of the machining technique to large-area applications. A DC voltage of 80-140 V is applied between the electrode and the workpiece through a resistance-capacitance circuit that controls the pulse energy and timing of spark discharges.  The parasitic capacitance of the electrode structure is used to form a resistance  capacitance circuit for the generation of pulsed spark discharge between the electrode and the workpiece.  The suspended electrodes are actuated towards the workpiece using the  downflow of dielectric machining fluid, initiating and sustaining the machining process. Micromachining of stainless steel is experimentally demonstrated with the machining voltage of 90V and continuous flow of the fluid at the velocity of 3.4-3.9 m/s, providing removal depth of 20 Jtm. The experimental results of the electrode actuation match well with the theoretical estimations. Second, the planar electrodes are electrostatically actuated towards workpiece for machining.  In addition to the single-layer, this effort uses double-layer  structures defined on the bottom surface of the electrode to create custom designed patterns on the workpiece material. The suspended electrode is electrostatically actuated towards the wafer based on the pull-in, resulting in a breakdown, or spark discharge.  This instantly  lowers the gap voltage, releasing the electrode, and the gap value recovers as the capacitor is charged up through the resistor. Sequential pulses are produced through the self-regulated discharging-charging cycle.  Micromachining of the stainless-steel wafer is demonstrated  using the electrodes with single-layer and double-layer structures. The experimental results of the dynamic built-capacitance and mechanical behavior of the electrode devices are also analyzed.  Ii  Table of Contents  Abstract  ii  Table of Contents  iii  List of Tables  v  List of Figures  vi  Acknowledgements  ix  Statement of Co-Authorship  x  1  Introduction 1.1  Non-traditional Micromachining Processes  3  1.2  Micro-Electro-Discharge Machining  4  1.2.1  Principle of 1 iEDM Operation  6  1.2.2  Setup for tEDM  7  1.2.3  Applications Areas of 1 iEDM  9  1.3  1.4  2  1  Advanced jtEDM  10  1.3.1  .iEDM Batch-Mode 1  11  1.3.2  New Approach to .tEDM  13  Thesis Outline  13  References  14  EDM: MEMS-enabled micro-electro-discharge machining 3 M  17  2.1  Introduction  17  2.2  Machining Principle  18  2.3  Electrode Design and Actuation Mechanisms  20  2.4  Fabrication  22  2.5  Experimental Results  25  2.6  Analysis and Discussion  29  2.7  Conclusions  31  111  References 3  .  EDM) by electrostatic actuation 3 MEMS-based micro-electro-discharge machining (M of machining electrodes on the workpiece  4  32  34  3.1  Introduction  34  3.2  Machining Principle  35  3.3  Device Design  37  3.4  Fabrication  40  3.5  Experimental Results  45  3.6  Analysis and Discussion  49  3.7  Conclusions  52  References  53  Conclusions  55  References  57  iv  List of Tables 1.1  Non-traditional micromachining processes  V  5  List of Figures  1.1  Mechanism of tEDM  6  1.2  iEDM tests Setup for 1  7  1.3  A transistor-type pulse generation circuit  8  1.4  3D geometry by iEDM using single electrode  9  1.5  An SEM image of a machined sample  9  1.6  Wire electro discharge grinding  10  1.7  Batch-mode j.tEDM using single pulse generation circuit  12  1.8  Batch-mode jiEDM pulse timing circuits (a) electrodes connected to a single pulse generation circuit; (b) electrodes connected to individual pulse generation circuits  12  2.1  Cross sectional view of MEMS-based .1 tEDM and its process steps  2.2  Sample design of the iEDM devices with (a) fixed-fixed and (b) cantilever  .  .  configurations 2.3  22  Two dry-film processes developed for the fabrication of the movable electrode devices on the workpiece  2.4  19  23  An SEM image of (a) a fixed-fixed electrode and (b) a cantilever electrode both with the layouts shown in Figure 2.2, and (c) an optical image of the fabricated devices  2.5  24  2 piece of sacrificial dry-film photoresist with patterned electrode A 6x6-cm devices (the 3” wafer underneath the resist film was placed for dimensional comparison with the film)  25  2.6  A set used for the characterization of the electrode actuation and pEDM tests  2.7  Built-in capacitance vs. fluid flow velocity measured with a fabricated device with the design shown in Figure 2.2a  2.8  (a) Measured pulses of discharge current at the voltage of 90 V with an inset  vi  26  of single pulse close-up, and (b) an optical image of spark light captured through electrode’s holes 2.9  27  Micromachined result obtained with a cantilever electrode: (a) an SEM image and (b) optically measured geometry of the machined structured (electrode removed after machining); (c) a top view at one of the holes of an electrode that was stuck to the workpiece during the machining, showing a circular surface of the workpiece  2.10  through  the  hole  with  a  discharge  gap  of  -  10  jim  Theoretical deflections vs. fluid flow velocity for the two electrodes in Figure 2.2 calculated using the experimental conditions of the fluidic set-up used (Figure 2.6)  3.1  30  Cross sectional view of the MEMS-based jiEDM based on electrostatic actuation and its steps; (b) dynamic behavior of discharge voltage and current corresponding to the steps  3.2  A sample layout of the jiEDM device for (a) torsional design with the single-layer scheme;  3.3  36  (b)  custom  design  structures  double-layer  scheme  Two processes developed for electrode fabrication with (a) single-layer structures; (b) double-layer structures  3.4  with  41  (a) An optical image of the fabricated devices on a stainless-steel substrate; (b) an SEM image of a sample device fabricated for the single-layer design in Figure 3.2a  3.5  43  Fabrication results for the double-layer design: (a) an SEM image of the electroplated copper on the copper supporting electrode; (b) an optical image of the fabricated device with the design in Figure 3.2b; (c) an SEM image of the close-up  44  3.6  A set-up for device characterizationjiEDM tests  45  3.7  (a) Measured current pulses of micro spark discharge; (b) single pulse close-up;  vii  (c) an optical image of the micro sparks captured through electrode’s holes  3.8  (a) An SEM image of the jiEDMed stainless-steel substrate (electrode removed after machining) with inset of close-up; (b) optically measured geometry of the machined structures.  Using  WykoTM  NTI 100 profiler, batch structures are  measured to have the depth of- 20 jim 3.9  47  Vertical displacement of the electrode measured while switching on and off the EDM process using the single-layer electrodes; (a) fixed-fixed device with 15 jim gap spacing; (b) cantilever device with 40  3.10  jim gap spacing  48  (a) Measured machining voltage showing charging cycles with different time constants for the electrodes with different areas; (b) built-in capacitance in static (Cb)  and dynamic (Cbb) modes vs. device area  viii  50  Acknowledgements  I would like to take this opportunity to acknowledge my advisor and supervisor Dr. Kenichi Takahata for his guidance and insight into the research I have been involved in. This work would not have been possible without his dedication and commitment towards research and I appreciate all his work. It has been a good learning experience and I cherish the entire period I have worked. I would also take the opportunity to thank Dr Edmond Cretu for initiating and introducing me to this research field. I would also like to acknowledge Dr. John Madden for his suggestions on different aspects during my entire education. I would like to thank Ms. Vijayalakshmi Sridhar and Mr. Greg Wong for assisting me with the fabrication process. I would also like to thank Mr. Mrigank Sharma and Ms. Akila Kannan for their fruitful discussions on various topics during the research work. I would like to thank all others who have helped me during the entire period of my research work.  ix  Statement of Co-Authorship  The research work presented in this thesis has been done in conjunction with the members of the Microsystems and Nanotechnology Group during the period 2006  —  2008.  This statement is to confirm that the author of this thesis is solely responsible for the research performed.  The author would like to acknowledge the contribution of co  authors for the manuscript preparation, written during the course of the research work.  x  Chapter 1 Introduction  Micromachining is a key manufacturing technology to produce microstructures and micromechanical parts with sizes ranging from sub millimeters to microns for microelectromechanical systems (MEMS).  The fabrication techniques developed for  integrated circuit industry can be utilized for micromachining and manufacturing by adding mechanical components like beams, springs, gears to the devices.  Surface  micromachining has been one of the earliest developed machining techniques and to a certain extent the mainstay of the industry. They were mainly used to form twodimensional structures on the surface of a silicon wafer. The main advantage of surface micromachined structures is the easy integration with integrated circuit (IC) components on the same silicon wafer, resulting in mass production of microsensors and actuators. But the structures fabricated were planar and with limited aspect ratio. High-aspect-ratio microstructures are often demanded in MEMS in order to achieve higher performance in terms of larger force, higher sensitivity and mechanical robustness. These constraints were overcome by using bulk micromachining that were able to build complex threedimensional shapes.  However, the materials that can be manufactured using these  technologies were mainly applicable to silicon and other limited materials like titanium. Special lithography techniques such as LIGA (a German acronym for lithography, electroplating and injection molding based on deep X-ray lithography) and SU-8 based processes (SU-8 is a negative, epoxy-type photoresist that can be used to pattern high aspect ratio micro structures) can be used to fabricate mechanical parts but machining real 3-dimensional (3-D) with free form shapes are still very difficult. Another constraint with the LIGA process is the high cost involved in the fabrication process based on the use of synchrotron radiation. The requirement for small dimensions, high precision, diverse materials and complex three-dimensional structures has opened up avenues for manufacturing technologies that does  not  follow  traditional  lithography-based  approaches.  Non-traditional  micromachining processes is a group of processes that remove excess material by various methods including mechanical (physical removing by milling, cutting, stamping, etc), abrasive (removal of material by stream of abrasive particles like water jet machining, ultrasonic machining, etc), thermal (thermal energy usually applied to small portion of workpiece, causing the portion to be removed by fusion or vaporization), and electrochemical (electrochemical energy to remove material) without using a sharp cutting tool as is the case with the conventional machining. For example the mechanical machining uses a sharp tool to mechanically cut the material to obtain the desired geometry and are capable of producing free-form structures in wide range of materials including common engineering alloys. Complex 3-D structures can be generated with shape accuracies by means of various mechanical processes that include micro-milling, micro-drilling and micro-grinding. Micro-electro-discharge machining ( iEDM) is an electrothermal micromachining 1 process for electrically conductive materials by using controlled sparks that occur between a microscopic electrode and a workpiece material in the presence of a dielectric fluid. When two electrodes are in close proximity in a dielectric liquid, application of a voltage pulse between them can break down the dielectric and produce a spark which thermally erodes the material of the workpiece at the breakdown point. The p.EDM process is a high repetition of this single removal cycle that erodes the material from the workpiece until the shape of the electrode is formed on the workpiece. 1 iEDM can be used for machining complex shapes by the use of numerical control (NC) machines for the positioning of the electrode on the workpiece but the throughput is low as it is a serial process and involves high costs due to NC machines.  The thesis presented in the  following sections is based on the principle of jiEDM.  A new approach to  micromachining using MEMS-based actuators is presented in this report for high throughput, low-cost and high-precision machining. In the following sections, Section 1.1 gives a brief overview of some of the non traditional machining processes, Section 1.2 gives the principles and technologies of IiEDM technique, Section 1.3 summarizes the past efforts for advanced 1 tEDM techniques, and Section 1.4 gives the outline of this thesis.  2  11  Non-traditional Micromachining Processes  In general, the use of non-traditional techniques allows one to utilize a broad range of engineering materials for microfabrication.  The major drawback of the mechanical  micromachining processes is the high machining force that affects the machining accuracy and the elastic deformation of the electrode or the workpiece [1]. These are a lot of challenges in the industry for machining materials (ceramics and composites) to meet the design considerations (high precision, high surface quality and complex geometry).  Advanced materials are increasingly used in industries like aircraft,  automobile, tool, die and mold making industries. The mechanical machining processes are one of the common methods (e.g. thermal, abrasive and chemical) for machining of components in such fields. However for micromachining they are not efficient and are expensive to machine the materials. To overcome these constraints other non-traditional machining processes can be used. Some of the non-traditional machining processes are discussed below. Electro-chemical machining is a method of removing metal by an electrochemical process [1].  An electrolytic cell is formed using the electrode as cathode and the  workpiece as anode in an electrolytic solution. The workpiece is removed in the form of sludge formed by the chemical reactions that takes place in the electrolytic cell. It is used for working with hard materials and can cut contours or cavities in hard steel and metals such as titanium and carbide.  Electrochemical grinding (ECG) is a form of  electrochemical machining that combines both electrochemical attack and abrasion to remove material from electrically conductive workpiece, hard, tough materials [2]. ECG makes use of a rotating grinding wheel (electrode) where direct current is passed through an electrolyte between the wheel and the workpiece. The majority of the workpiece material is removed by the electrochemical attack whereas the abrasive action removes only a part of the workpiece material.  ECG can be used to grind any electrically  conductive material, steel or alloy steel part, hard and tough materials. ECG can be used for variety of applications in the aerospace, automotive instrumentation, textile and medical manufacturing industries. Some of the features of the non-traditional machining processes are given in Table 1.1 along with the applicable materials [3].  3  Laser beam machining is a thermal removal process where an intense beam of collimated, single wavelength, in-phase light is focused by an optical lens onto the workpiece material to be machined [4]. It can be used for variety of materials: metals, composites, plastics and ceramics. In ultrasonic machining, the cutting is performed by abrasive particles suspended in a fluid. The abrasives contained in a slurry are driven at high velocity against workpiece by a tool vibrating at low amplitude and high frequency [5, 6]. The tool oscillation is perpendicular to the workpiece and is slowly fed into the workpiece creating a shape of the tool in the workpiece [7]. Ultrasonic machining can be used for machining brittle materials such as single crystals, glasses and polycrystalline ceramics and for increasing complex operations to obtain intricate shapes and workpiece profiles.  Since the  ultrasonic machining is non-thermal and non-chemical, there is no change in the chemical or physical properties of the workpiece thereby offering stress-free machined surfaces [5]. It is used extensively in manufacturing hard and brittle materials that are otherwise difficult to machine using the conventional processes. Abrasive waterjet machining can be used to cut metals and hard materials. The cutting is performed by a high speed, small-scale erosion process that enables abrasive waterjet to cut difficult to machine materials with very limited thermal and mechanical effects. It is widely used in cutting metals such as aluminum, brass, steel, titanium and nickel based alloys, as well as glass, stone and composites [8].  Abrasive waterjet  machining can be used in a number of applications such as drilling, polishing, turning and milling. iEDM is a machining process based on material removal by melting and partly by 1 vaporization with miniaturized discharge spark pulses [9]. iiEDM has been leveraged for fabricating small features and micro-components such as nozzle holes, slots, and shafts in any conductive material including all kinds of metals and alloys offering exceptional micromachining opportunities. In the next section, details of jtEDM is reviewed.  1.2  Micro-Electro-Discharge Machining  The 1 iEDM process involves thermal erosion induced by miniaturized spark discharge pulses generated between a microscopic electrode tip and the workpiece in a dielectric 4  Table 1.1: Non-traditional Micromachining Processes. Size  Machining method  Aspect ratio  Mode  Materials  (depth/widt  (Serial/Batch)  Structures  Serial  Thin  &  h) 100  Electrochemical  Holes  Machining  as small as  metals  and  (ECM) [10, 1 1]  100 im  alloys,  turbine  blades,  pistons,  size  films  of  fuel-injection nozzles size of 20-30  Mm  EDM [9]  Serial/Batch  Any electrically conductive  5m  materials such as stainless  steel,  graphite  and  doped semiconductors Laser  Beam  Holes from  Machining  10  (LBM) [3, 12]  1.5mm  Plasma  Used  Beam  m  50  Serial  in hard materials  to  for  Complex profiles  Serial/Batch  -  Very  Machining  thick films:  temperature  (PBM)  25 im  materials  [31  Ultrasonic machining 13, 14]  [3,  Holes from  2.5 urn for a  50 urn to 75  250 jim hole  Serial/Batch  Hard and brittle materials such as ceramics,  mm  and carbides  5  high  glass  fluid. jiEDM is a non-contact process that requires little force between electrode and workpiece and is capable of machining ductile, brittle or super hardened materials. It is possible to achieve high precision and quality machining.  It can micromachine any  electrical conductor (e.g. hard steel graphite, permanent magnet) including doped semiconductors.  The technique is capable of producing real three- dimensional  microstructures while achieving the smallest size of 5 jim with submicron tolerance with aspect ratio of 20-30 [15]. These attractive features have been leveraged for producing micro components as well as prototyping various bulk-metal-based MEMS such as antenna stent, micro Kelvin probe and capacitive pressure sensor [16, 17].  1.2.1  Principle of j.tEDM Operation  The basic components of the jiEDM process consists of an electrode tool and the workpiece material with the electrode immersed in dielectric fluid. The principle and operation of jiEDM is illustrated in Figure 1.1. The electrode and the workpiece are separated from each other by a dielectric fluid. A voltage difference is applied between the electrode (that serves as a cathode) and the workpiece (which acts as an anode). If the electrode and the workpiece are moved close enough together and the voltage is high EDM fluid  EDM fluid  Electron  !eO! 0  —T  Ionization channel  Posthve  9  (b) Ionization channel and atomic charge  (a) Initiation of electric discharge  EDM fluid  EDM fluid  I  tf  Electrode wear  Metal  particl  I ..  Workece  .  (d) Crater formation  (c) Thermal erosion of workpiece material  Figure 1.1: Mechanism of jiEDM [18].  6  enough, the dielectric fluid will break down and conduct an electric current, causing an electrical discharge (a spark) between them. The sparks will produce an extremely high temperature (of the order of 10,000K) at localized spots on the workpiece and brings particles of the workpiece to a vaporized state [7]. These particles immediately resolidify into small particles and are flushed away by the dielectric fluid, leaving a small crater eroded in the workpiece for a single spark. jiEDM is an EDM technique that uses low discharge energies (10w  —  i0 Joules) and low volume rate of removal (0.05 —500 jim ) 3  of workpiece material for producing small features [9].  1.2.2  Setup for j.tEDM  Figure 1.2 shows a basic setup of jiEDM apparatus. A sample is held on an X-Y stage and an electrode is on a Z stage. There are two types of pulse generation circuits that can be used with EDM: transistor type pulse generator and resistance capacitance (RC) pulse generator [19]. During jiEDM, the removal of material in each discharge should be controlled to ensure a stable processing. The material removal is related to the discharge energy of each pulse. The discharge energy depends on the discharge voltage, discharge current peak value and the pulse-on time. The short pulse-on time enables the small removal of material per discharge which is required for jiEDM. The transistor-type pulse generator has been used in conventional EDM (Figure 1.3).  The pulse discharge is  achieved using the switching component. The discharge process can be controlled by  V (60’-’-’l  (Ø-300pm)  (C’ : Stray capacitance)  Figure 1.2: Setup for jiEDM tests [18].  7  I  .  V  Electrode Workpiece  FET  Pulse circuit Figure 1.3: A transistor-type pulse generation circuit. monitoring the discharge state at the gap using the transistor-type pulse generator. But there is a delay of the transmission due to the switching component and the pulse control circuit components that makes it difficult to use for iEDM due to the difficulty in the pulse formation on the order of nanoseconds [19]. Many of tEDM techniques uses a RC pulse generation circuit as it provides a short pulse on-time and ceramics for selected mechanical components, which is very important for minimizing parasitic capacitance (C’) to lower the discharge energy E=(C  +  C)V / 2 for better quality and surface 2  roughness of the machined structures (Figure 1.2).  The pulse energy is about one  hundred times smaller than that in macro-scale EDM [18]. The typical electrode material used for tEDM is tungsten as its high melting point reduces the electrode wear. The j.tEDM technique can be used for three-dimensional machining using the NC machines (Figure 1.4 and Figure 1.5) [20]. The sample shown in Figure 1.5 was produced with a 5axis iiEDM system with two rotational axes. iEDM are typically fabricated by different types of The microelectrodes used for 1 tEDM and other methods. One of the most common methods is wire electro discharge grinding (WEDG). WEDG is a fabrication process that uses electrical discharges in a dielectric fluid to erode material from conductive wires.  The electro discharge is  conducted between the wire that is guided by a wireguide and the workpiece. The wire that is melted or diminished is replaced by new wire as it is continuously run around the wireguide.  WEDG can be used for high precision micromachining of cylindrical 8  electrodes as the desired shape can be machined by having the wire electrode rotate during the machining reducing the machining error (Figure 1.6).  1.2.3  Application Areas of jiEDM  EDM is used to machine a wide variety of miniature and microparts from electrically conductive materials such as metals, alloys, sintered metals, cemented iEDM may also be used to produce molds and dies that carbides, ceramics and silicon. 1 can themselves be utilized to manufacture other microparts from both conductive and  Workpiece  —  Electrode  / ;ross-section  Figure 1.4: 3D geometry by EDM using single electrode [18].  Figure 1.5: An SEM image of a machined sample [20].  9  Wire  Electrode  —  Wireguide  Figure 1.6: Wire electro discharge grinding. non-conductive materials such as plastics.  There are number of applications using  IIEDM process. tEDM can be used to obtain microshafts and pins that are important in the assembly of miniature devices [21, 22]. Micropunches with a diameter of 70 jim, used for the mass production of inkjet printer heads, have been fabricated by jiEDM [23]. iEDM allows precision holes to be made on conductive materials with high aspect ratios 1 [24]. The fabrication of ink-jet nozzles are one of the major applications for jiEDM [25]. Some of the other commercial applications include magnetic heads for digital VCRs [18]. Fuel injector valves, parts and components for medical devices, fiber optic connectors, micromachining, micro-mold making, stamping tools, and microelectronic parts are some of the examples of miniaturized and smaller size parts produced by the jiEDM technology. Some other applications include machining gear trains [26], photomasks [27], and forming tools [28].  1.3  Advanced tEDM  jiEDM offers exceptional capabilities to micromachine a variety of bulk metals and iEDM has some drawbacks associated with it, alloys with high precision. However 1 iEDM is a serial process that uses a single mainly low throughput as pointed out earlier. 1 electrode tip for the machining resulting in low throughput. This approach also affects the precisiOn for long processing as the corner wear of the electrode increases during 10  serial machining and the electrode tip degrades the machining accuracy of the workpiece (Figure 1.4). Microelectrodes are manufactured using pEDM process such as WEDG (described in previous section) which are then used for making holes by drilling, forming complex molds by pEDM milling process [23].  During this entire process, the  microelectrode is consumed continuously and the electrode has to be replaced when it cannot maintain high dimensional accuracy. It also takes time to shape and produce the electrode since they are individually produced. There are some techniques to compensate the electrode wear to minimize the degradation of machining precision [29, 30]. A very long electrode is generally required to remove large amount of material, to compensate for the electrode wear that results in instability of the electrode. However generally such electrodes tend to have bending issues. To overcome these issues associated with serial processing EDM, new types of methods have been developed towards achieving batch processing for iEDM. The following sections summarizes about the batch-mode jtEDM and outlines the new approach for iiEDM using MEMS actuators investigated in this effort.  1.3.1  iEDM Batch-Mode 1  To overcome the constraints of low throughput and to avoid the use of numerical control iEDM that uses arrays of highof the electrode tip and the workpiece, a batch-mode 1 aspect-ratio microelectrodes was demonstrated to achieve high parallelism/throughput of the process [16]. In this approach, the arrays were fabricated using the LIGA process [31, iEDM 32] and were advanced into the workpiece using the vertical NC stage in an 1 apparatus (Figure 1.7). It was observed that the throughput was substantially improved over the traditional process due to the use of the arrays. All the electrodes in the initial study were connected to a single pulse generation circuit.  It was reported that the  machining rate was reduced as the number of electrodes increased since individual sparks were generated thereby increasing the time of discharge spark cycle. It was found that there was an improvement in the machining rate when the electrodes were connected to individual pulse generation circuits thus allowing several sparks at the same time (Figure 1.8). The process still requires the NC capability for vertical positioning of the arrays. In another effort, jiEDM was used to fabricate the electrode arrays instead of LIGA process. 11  It provided the advantage of wide range of materials and also the cost was reduced further [33]. However, the electrode arrays were fabricated using wire electro-discharge  LIGA  Li  I  /  /  /  —  .  . —  EDMed )\structure / \arraY  Figure 1.7: Batch-mode tEDM using single pulse generating circuit [17].  • All electrodes connected to a single pulse timing circuit Dielectric substrate Plaling base Eleode  • Electrodes electrically divided • Each section has a timing circuit  Pulse timing  f  V2  ..I!I Discharge pulses in parallel  Discharge pulses in series Frequency at each electrode (i.e. machining rate) reduced as the number of electrodes increased  Machining rate is independent of the number of electrodes  Figure 1.8: Batch-mode tEDM pulse-timing circuits (a) electrodes connected to a single pulse generation circuit; (b) electrodes connected to individual pulse generation circuits [18]. 12  machining (WEDM) that results in reduced dimensional accuracy and loss of compatibility with batch production for lithography.  1.3.2  New Approach to jiEDM  The use of MEMS to control the 1 iEDM is proposed in this study to overcome the constraints associated with the batch-mode pEDM. The new approach involves a 1 iEDM method where planar electrodes are microfabricated directly on the surfaces of the work material and are electrostatically actuated for controlled generation of discharge pulses. This approach potentially results in high-throughput by extending it to large-area micromachining by using array of electrodes. The precision accuracy is also increased as only a single electrode is used to machine a single structure on the workpiece material. The new approach also intends to eliminate the need for NC machines thereby drastically reducing the equipment costs for the machining process.  1.4  Thesis Outline  In the following chapters, two manuscripts are presented on 1 iEDM done over the course of research in the Microsystems and Nanotechnology Group. In Chapter 2, the planar electrodes are actuated towards the workpiece using the hydrodynamic force. In Chapter 3, a MEMS-based iEDM is presented where the planar electrodes are electrostatically actuated towards the workpiece for machining purposes. All the experimental results and the theoretical estimations for the electrode devices are presented in this report. Chapter 4 concludes the overall effort.  13  References [1]  H.  El-Hofy,  Fundamentals  of Machining  Processes:  Convectional  and  Nonconventional Processes, CRC Press, 2006. [2]  W. Cubberly and R. Bakerjian, Tool and Manufacturing Engineers Handbook, Society of Manufacturing Engineers, 1989.  [3]  B. Bhushan, Tribology Issues and Opportunities in MEMS, Springer, 1997.  [4]  R. Crowson, The Handbook ofManufacturing Engineering, CRC Press, 2006.  [5]  D. Kramer, “Ultrasonically assisted machining”, Mech. md. Mater., vol. 48, pp. 1521, 1995.  [6]  T. B. Thoe, D. K. Aspinwall, and M. L. H. Wise, “Review on ultrasonic machining”,  mt. i Mach.  Tools Manufact., vol. 38, no. 4, pp. 239-255, 1998.  [7]  J. McGeough, Micromachining ofEngineering Materials, Marcel Dekker, 2002.  [8]  M. Hashish, “Deep hole drilling in metals using abrasive materials”, Proc.  th 13  mt.  Conf on Jetting Technology, vol. 38, no.4, pp. 239-255, 1998. [9]  T. Masuzawa, “State of the art micromachining”, Annals of the CIRP, vol. 49, pp. 473-488, 2000.  [10] K. P. Rajurkar, G. Levy, A. Malshe, M. M. Sundaram, J. McGeough, X. Hu, R. Resnick, and A. DeSilva, “Micro and nano machining by electro-physical and chemical processes”, CIRP Annals, vol. 55, pp. 643-666, 2006. [11] W. Zhao, X. Li, and Z. Wang, “Study on micro electrochemical machining at micro to meso-scale”, Proc. IEEE Conf NEMS, pp. 325-329, 2006. [12] M. Henry, P. M. Harrison, 1. Henderson, and M. F. Brownell, “Laser milling: a practical industrial solution for machining a wide variety of materials”, Proc. SPIE, pp. 627-632, 2004. [13] R. Singh, and J. S. Khamba, “Ultrasonic machining of titanium and its alloys: A review”, J Mater. Process. Tech., vol. 173, pp. 125-135, 2006. [14] T. Li, and Y. B. Gianchandani, “A micromachining process for die-scale pattern transfer in ceramics and its application to bulk piezoelectric actuators”, I MEMS, vol. 15, pp. 605-6 12, 2006. [15] L. L. Chu, K. Takahata, P. Selvaganapathy, Y. B. Gianchandani, and J. L. Shohet,  14  “A micromachined Kelvin probe with integrated actuator for microfluidic and solid-state applications”, I MEMS, vol. 14, pp. 69 1-698, 2005. [16] K. Takahata, and Y. B. Gianchandani, “Bulk-metal-based MEMS fabricated by micro-elecctro-discharge machining”, Proc. IEEE Canadian Conf Electr. Comput. Eng. (CCECE), pp. 1-4, 2007. [17] K. Takahata and Y. B. Gianchandani, “Batch mode micro-electro-discharge machining”, I MEMS, vol. 11, pp. 102-110, 2002. [18] K. Takahata, Batch manufacturing technology based on micro-electro-discharge machining and its applications to cardiovascular stents. PhD Thesis The University of Michigan, 2005. [19] F. Han, L. Chen, D. Yu, and X Zhou, “Basic study on pulse generator for micro EDM”, mt. I Adv. Manuf Technol., vol. 33, pp. 474-479, 2007. [20] K. Takahata, S. Aoki, and T. Sato, “Fine surface finishing method for 3dimensional micro structures”, JEICE Trans. Electronics, E80-C, 2, pp. 29 1-296, 1997. [21] H. H. Langen, T. Masuzawa, and M. Fujino, “Modular method for microparts machining and assembly with self-alignment”, Annals of the CIRP, vol. 44, pp. 173-176, 1995. [22] H. H. Langen, T. Masuzawa, and M. Fujino, “Reverse micro-EDM and its applicability to microassembly”,  mt.  I Electrical Machining, vol. 1, pp. 53-57,  1996. [23] T. Masuzawa, M. Fujino, K. Kobayashi, and T. Suzuki, “Wire electro-discharge grinding for micro-machining”, Annals ofthe CIRP, vol. 34, pp. 43 1-434, 1985. [24] D. M. Allen, “Micro-electrodischarge machining”, PCMII, vol. 66, pp. 7-8, 1996. [25] D. M. Allen and A. Lecheheb, “Micro electro-discharge machining of ink jet nozzles: Optimum selection of material and machining parameters”, I Materials Processing Technology, vol. 58, pp. 53-66, 1996. [26] N. Shibaike, H. Takeuchi, K. Nakamura, and N. Shimizu, “Approach to higher reliability in 3d micro-mechanisms”, Proc. SPIE: Micromachining Technology for Micro-Optics, 2000. [27] S. H. Yeo, and G. G. Yap, “A feasibility study on the micro-electro-discharge  15  machining process for photomask fabrication”,  mt.  i Manufacturing Technology,  2001. [28] E. Uhlmann, G. Spur, N. A. Daus, and U. Doll, “Application of micro-edm in the machining of micro structures forming tools”, SME Technical Paper MF 99-285, 1999. [29] Y. Z. Yu, T. Masuzawam and M. Fujino, “Micro EDM for three-dimensional cavities: development of uniform wear method”, Annals of the CIRP, vol. 47, pp 169-182, 1998. [30] P. Bleys, J. P. Kruth, and B. Lauwers, “Sensing and compensation of tool wear in milling EDM”, J Mater. Process. Technol., vol. 149, pp. 139-146, 2004. [31] H. Guckel, K. J. Skrobis, T. R. Christenson, and J. Klein, “Micromechanics for actuators via deep X-ray lithography”, Proc. SPIE Symposium on Microlithography, pp. 39-47, 1994. [32] H. Guckel, “High-aspect ratio micromachining vs deep X-ray lithography”, Proc. IEEE, pp. 1586-1593, 1998. [33] Y. M. Sang, P. S. Mi  L. S. Young, and C. N. Chong, “Fabrication of stainless  steel shadow mask using batch mode micro-EDM”, I Microsyst. Technol., vol. 14, no. 3, pp. 411-417, 2007.  16  Chapter 2* EDM: MEMS-enabled micro-electro-discharge 3 M machining 2.1  Introduction  Micro-electro-discharge machining (jiEDM) is a non-contact micromachining technique that can be used to cut any type of electrically conductive materials. The technique is capable of producing real three-dimensional microstructures while achieving the smallest size of 5 jim with submicron tolerance [1]. These attractive features have been leveraged for producing micro mechanical components as well as prototyping various micro electro-mechanical systems (MEMS) and devices [2, 3]. However, the throughput is inherently low because the traditional technique is essentially a serial process that uses a single electrode tip together with numerical control (NC) of the tip and the workpiece, iEDM that uses microelectrode arrays producing structures individually. Batch-mode 1 fabricated by a deep x-ray lithography (LIGA) process [4] was demonstrated to achieve high parallelism/throughput of the process [5]. The use of LIGA, however, incurs high costs in the electrode fabrication. There have been some efforts that attempt to address the cost-effectiveness issue in the electrode fabrication, at the expense of compatibility with photolithography-based methods [6, 7]. In addition, the batch-mode method still requires an NC stage for advancing the arrays into the material. It has recently been shown that jiEDM can be implemented using electrodes that are microfabricated  directly  on  the  surfaces  of the  workpiece  using  standard  photolithography and etching processes [8]. This method exploits the machining voltage to electrostatically actuate movable microelectrodes, eliminating the need of NC machines from the machining process. This approach is suitable for selected electrode structures and applications for relatively shallow machining due to the limitation of the  Copyright © Institute of Physics and lOP Publishing Limited 2008. “A version of this chapter has been EDM: MEMS-enabled micro 3 published”. Alla Chaitanya Chakravarty Reddy and Kenichi Takahata, “M electro-discharge machining”, J Micromech. Microeng., vol. 18, 105009, 7pp, 2008. *  17  electrostatic actuation range.  This paper reports a new MEMS-based rnicro-EDM  EDM) method where planar electrodes microfabricated on the workpiece are actuated 3 (M with an external force for controlled generation of discharge pulses. The method aims to enable the machining process without the constraint attributed to the electrostatic actuation thereby achieving high applicability of the technique. The movable electrodes are microfabricated using dry-film photoresist processes, targeting at potential application of tEDM to very-large-area and/or non-planar samples with high precision and throughput at low cost. This paper is constituted as follows. Section 2.2 describes the machining principle and method. The design of the movable electrode devices is discussed in conjunction with the actuation mechanisms in Section 2.3. Section 2.4 presents details of the fabrication processes for the devices. The results of experimental characterization for the fabricated iEDM with the devices are reported in Section 2.5, devices and the demonstration of 1 followed by discussion including the analysis of the experimental results in Section 2.6. Section 2.7 concludes the overall effort.  22  Machining Principle  EDM method uses planar electrodes that are suspended by the anchors 3 The developed M above the surfaces of the conductive workpiece with a relatively large gap in a dielectric liquid (Figure 2.1). A resistance-capacitance (RC) pulse generation/timing circuit [1, 5] is coupled between the device and a DC voltage source (80-100 V) so that the electrode serves as a cathode whereas the workpiece is an anode. The parasitic/built-in capacitance that is present between the electrode-anchor structure and the workpiece is leveraged to form the RC circuit as shown in Figure 2.1 [8].  The voltage applied between the  workpiece and the suspended electrode causes a breakdown when the gap separation between them reaches the threshold distance (typically a few microns in jiEDM). The breakdown leads to a spark current and an instant voltage drop due to a discharge from the capacitor. The spark current causes a thermal impact that removes the material, leaving a crater-like shape at the breakdown point on the workpiece surface. The voltage is restored with time as the capacitor is recharged through the resistor of the circuit for the next breakdown. The machining is implemented by the repeated removal of the 18  material by pulses of the spark discharge while feeding the electrode into the material. The tolerance and surface quality of tEDM depends on the energy of the single electric discharge, EDSC, which is expressed as [1]: (2.1)  EDSC =-Cv2  where C is the capacitance of the RC circuit and V is the machining voltage. If the electrode is supported by a spring so as to be movable in the vertical direction, with a proper structural design and certain operating conditions, it can be displaced towards the workpiece with electrostatic force produced by the machining voltage to reach the breakdown gap and generate a discharge current. The “pull-in” mechanism can  Static state EDM fluid flow  1’ !i  4, Electrode lifted-off afterEDM  Spark discharge ;z:, -  l  iEDM and its process steps. Figure 2.1: Cross sectional view of MEMS-based 1 optionally be used to obtain a relatively large displacement [8, 9]. The repeated cycle of attraction of the electrode, breakdown and voltage drop, release of the electrode, and recovery of the voltage maintains the pulse generation and automatically advances the electrode in a self-regulated manner while removing the material.  However, as the  material is removed, it reaches a state where the electrostatic force is no longer large enough to pull the electrode down towards the machining surface and sustain the breakdown gap against the restoring force through the spring, limiting further machining. The force can be increased by raising the voltage V, however, this results in larger 19  discharge energy EDSC leading to the degradation of tolerance and surface roughness in the machined structures. The application of force to the electrode externally can be a solution to this constraint. In addition, this approach permits one to make the original gap spacing larger beyond the range that the pull-in method is available, which, in fact, is required for the sacrificial removal of the dry-film photoresist described in the next section. The larger original gap also promotes easier flushing of the byproducts, which contains the particles removed from the workpiece and carbon residues produced from the EDM fluid during the process, when the electrodes are in resting state. There are various potential methods for applying external force to the electrode for the actuation towards the workpiece surfaces. Some micromachined actuators have been reported to use externally assisted actuation mechanisms [10-12].  For the iEDM  application, one of the simplest and most effective methods would be the use of downflow of the EDM fluid in which the electrodes and workpiece are immersed. The flow rate, i.e., hydrodynamic pressure can be controlled so that the electrodes are externally displaced to induce the repeated cycle of the pulse generation, implementing the EDM process. Furthermore, the down flow is expected to enhance the removal of the byproducts from the machining area during the process.  2.3  Electrode Design and Actuation Mechanisms  To experimentally demonstrate the machining method, various test electrodes were designed and fabricated for fixed-fixed and cantilever type configurations with sizes ranging from hundred’s of microns to millimeters. Sample designs are shown in Figure 2.2. For the feasibility study, this effort utilized simple construction and fabrication for the electrodes with single-layer structures without particular features underneath (as illustrated in Figure 2.1). The electrode structures are formed using stock copper foil with 18-jim thickness as the original material, providing a relatively thick yet uniform layer of copper with no residual stress. (Copper has been used as an electrode material for iEDM [5, 7, 8]). All the anchors used in the devices have the common dimensions of 2.5x2.5 mm . The maximum vertical deflection of a fixed-fixed or cantilever electrode, y, 2  with uniformly applied pressure,p, due to the fluid flow can be described by [13]:  20  pwl aEl  =  (2.2)  where E is the Young’s modulus of the electrode material, 1 is the length of the electrode, /12 where w and h are the width and thickness 3 I is the moment of inertia given by 1wh  of the electrode respectively, and a is a constant that depends on the electrode configuration (384 for fixed-fixed and 8 for cantilever). As discussed in the next section, the gap separation between the electrodes and the substrate used in this study is 70 jim that corresponds to the thickness of the dry-film sacrificial layer necessary for the electrode release. This separation requires high voltages to pull the electrodes down to the breakdown position. The pull-in voltage,  Vpj,  for the suspended electrodes can be  described by [9]: Vpi  =  /8Kg3 ‘d 27&A  (2.3)  where g is the initial gap separation, A is the area of the capacitive electrode, and permittivity of the EDM fluid. suspended KFF  =  structures,  which  is the  The parameter K is the effective stiffness of the is  defined  for  the  fixed-fixed  structure  as  2kP1[(kl/4)—tanh(kl/4)] where k=jPI(EI) and P is the axial force created by  the combination of the intrinsic stress in the suspended structure and the non-linear stress that arises due to the deflection of the structure. The effective stiffness for the cantilever can be represented as KCL constants (h= 18 jim,  =  3 /(3l) 2Ewh  .  Using equation (2.3) with the relevant  1 .59x 1 01 I F/rn for kerosene as the main component of typical  EDM fluids, E=120 GPa for copper, and the intrinsic stress assumed to be negligible), the pull-in voltages for the 70-jim gap are calculated to be approximately 730 V and 420 V for the fixed-fixed and cantilever electrodes in Figure 2.2, respectively. In either case, the pull-in voltage is far greater than a typical range of jiEDM voltage (60-1 1OV) [1]. The external actuation using hydrodynamic force can be an effective means to address the limitation in not only the feed depth but also initiating the breakdown even when the gap separations between the electrodes and the workpiece are large.  21  10000  (a)  1= 5000 Anchor  Electrode -  (h’s d)  0 0  jw=800  It)  (N  (Unit: pm) 2500 iEDM devices with (a) fixed-fixed and (b) cantilever Figure 2.2: Sample designs of the 1 configurations.  2.4  Fabrication  Figure 2.3 illustrates the cross-sectional view of the fabrication process for the suspended planar electrodes.  As described earlier, this effort explores the use of dry film  photoresists for all the lithography steps. The lamination of the dry films is commonly performed using a hot-roll laminator (XRL-120, Western Magnum Co., CA, USA) with a feed speed of 1.3 cm/s at 120 °C. The type-304 stainless steel in a form of 3” wafer was selected as the work material and served as the substrate for the fabrication in this effort. Two processes, shown as (a) and (b) in Figure 2.3, were developed for the device fabrication.  In the process (a), the devices are fabricated directly on the substrate,  whereas in the process (b), they are formed and supplied on a piece of the dry film that can be laminated and released on a selected surface of the workpiece.  The latter  approach can potentially be useful for processing the workpieces that have non-planar surfaces to be machined or very large dimensions that are not compatible with standard photolithography tools.  22  (b)  (a> 1) Laminate double resist  films  for  Hot-roll aminator—_.(  70-  pm-thick sacrificial? anchor layer Cu foil  2) Laminate 18-pmthick Cu foil  1) Laminate double resist films for 70pm-thick sacrificial! I. anchor layer on 18-pm-thick Cu foil  a’..  2) Laminate 15-pm-  thick resist on the [1iflh1iflhifl1i1)1ii11 other side of Cu foil  3) Laminate 15-pmthick resist UV light  3)  Patt:m the top  -Mylar mask  4) Pattern the top resist  Anchor  4) Wet etch Cu foil 5)  Laminate  on  the  workpiece  5) Wet etch Cu foil  / 6) Sacrificial etch of resist in developer and cleaning  i  Figure 2.3: Two dry-film processes developed for the fabrication of the movable  electrode devices on the workpiece. For the process (a), a negative photoresist (PM240, DuPont Co., DE, USA) with 35jim thickness is first laminated twice on a thoroughly cleaned wafer to form a sacrificial layer with the total thickness of 70 jim (step al). This thickness or greater was observed to be required for proper release of the designed electrodes performed at the last step of the process. Next, the 1 8-jim-thick copper foil is laminated on the sacrificial layer with the same laminator (step a2).  Then, a 15-jim-thick negative photoresist (SF306,  Macdermid Co., CO, USA) is laminated on the copper foil (step a3) and patterned using the mylar mask with the layout of the devices and a standard mask aligner (step a4). The SF306 photoresist is developed in an alkaline aqueous developer, which is then used as a mask for wet etching of copper in a ferric chloride solution (step a5). Finally, to release the electrodes, timed etching of the sacrificial resist is performed in the developer for 2.5 hours at room temperature without agitation (step a6).  2 perforations The 30x30-jim  defined in the electrodes (Figure 2.2) promote the undercutting during the sacrificial etching process, while leaving the resist to be the spacers at the anchors that have no holes. The stream of the developer is used for the last 5 minutes of the sacrificial etch to flush the resist residues, followed by cleaning in acetone. Figure 2.4 shows the electrode 23  devices fabricated by this process. The built-in capacitances of the fabricated fixed-fixed and cantilever electrodes in Figure 2.2 are measured in air to be 7.2 pF and 3.4 pF, respectively. For the process (b), the copper foil is laminated with the double-layer sacrificial film of PM240 photoresist on one side and with the SF306 photoresist on the other side of the foil (steps bi and b2). The SF306 resist is patterned (step b3) and used as a mask for the wet etching of copper while the protective film of the PM240 photoresist is kept intact to avoid any damage to the sacrificial layer during the etching process (step b4). Figure 2.5 shows a piece of the sacrificial dry film with arrays of the patterned copper electrodes prior to the lamination on a workpiece. After the completion of the electrode fabrication, the protective film of the sacrificial resist is removed and laminated on the sample to be machined (step b5). The sacrificial etch is performed in the similar manner as described in the process (a) above.  iless steel substrate chor  -..  (a)  1mm  (b) 500pm — •  ..  ..  —  ic)  AnchorEEIeCtrOd •‘•.  •  -—  •.l  1 mm—  Figure 2.4: An SEM image of (a) a fixed-fixed electrode and (b) a cantilever electrode both with the layouts shown in Figure 2.2, and (c) an optical image of the fabricated devices.  24  2 piece of sacrificial dry-film photoresist with patterned electrode Figure 2.5: A 6x6-cm devices (the 3” wafer underneath the resist film was placed for dimensional comparison with the film).  2.5  Experimental Results  Figure 2.6 shows a set-up used for pEDM tests as well as characterization of the electrode structures. The substrate with the fabricated devices was placed in an ultrasonic bath filled with low-viscosity dielectric EDM oil (EDM  TM, 185  Commonwealth Oil Co.,  ON, Canada). A 2O-K resistor was connected between the device and the DC voltage source to form the RC circuit with the built-in capacitance as shown in Figure 2.6. The electrical discharge pulses were monitored using an AC current probe, which has a minimal loading on the discharge circuit. A variable-speed motor pump was used to inject the EDM oil at a controlled rate to apply fluidic pressure to the electrodes for their actuation. The ultrasonic wave was applied to the fluid bath during the process to assist in the dispersion of the byproducts.  As described earlier with equation (2.1), the  discharge energy Esc, or machining quality depends on the built-in capacitance of the device, which is a dynamic parameter as it is partially determined by the movable electrode (in addition to the fixed anchors).  The behavior of the capacitance was  characterized using an HP 4275A LCR meter with varying flow rate of the EDM fluid as shown in Figure 2.6. 25  Figure 2.7 shows the built-in capacitance of the fixed-fixed electrode structure vs. flow rate, in both increasing and decreasing directions, measured while applying no voltage to the electrode. Compared to the 7.2 pF built-in capacitance measured in air, the increased static capacitance was expected due to the operation in liquid ambient. The plot in Figure 2.7 indicates a non-linear rise of the capacitance, i.e., decrease of the capacitive gap with the flow velocity as well as a highly elastic behavior of the electrode structure during the actuation. The capacitance was observed to become immeasurable at  Figure 2.6: A set-up used for the characterization of electrode actuation and iEDM tests.  LL  516 c’ 14 a C”  C-)  0  1  2  3  4  5  6  Fluid flow velocity (mis) Figure 2.7 Built-in capacitance vs. fluid flow velocity measured with a fabricated device with the design shown in Figure 2.2a. 26  the flow velocity of 5.4 rn/s and greater, indicating physical contact of the electrode to the substrate. The resist spacers at the anchors were measured to show no detectable change in the thickness after immersing the devices in the EDM oil for 2-3 days, suggesting that the swelling effect due to the absorption of the oil from their sidewalls is negligible. With the application of machining voltage and the injection of the fluid to the electrodes, sequential pulses of micro spark discharge were successfully generated and sustained at the flow velocities of 3.9 m/s and 3.4 rn/s for the fixed-fixed and cantilever  1(A)  (a)  (b) Electrode +— edge  Figure 2.8: (a) Measured pulses of discharge current at the voltage of 90 V with an inset of single pulse close-up, and (b) an optical image of spark light captured through electrode’s holes. electrodes, respectively (Figure 2.8). The typical peak current and pulse duration were measured to be 2-3 A and 50 ns, respectively, in the set-up used. Figure 2.9a shows the stainless-steel workpiece machined using the cantilever device (Figure 2.2b) at 90 V for 27  about 15 minutes.  The pattern of the cylindrical structures in the machined area  corresponds to that of the holes of the electrodes. The machined structures were characterized using a WykoTM NT1IOO optical profiler (Figure 2.9b). The measurement indicates the removal depth of 20 jim and the average surface roughness of 520 nm in the machined areas. electrode.  Figure 2.9c shows an optical image at one of the holes in the  The image was taken after machining but before removing the electrode  structure from the workpiece, indicating a discharge gap of about 10 jim between the machined cylindrical structure and the perimeter of the electrode hole. Original sur  ..  (pm)  ..  -1.8  -  -6.8  11 8  --16.8  -21.8 -24.1  Workpiece surface  Top surface of copper electrode  (c)  Figure 2.9: Micromachined result obtained with a cantilever electrode: (a) An SEM image and (b) optically measured geometry of the machined structures (electrode removed after machining); (c) a top view at one of the holes of an electrode that was stuck to the workpiece during the machining, showing a circular surface of the workpiece through the hole with a discharge gap of-i 0 jim.  28  2.6  Analysis and Discussion  It is worth evaluating the measurement results and their consistency with theoretical estimations. The pressure applied onto the electrode immersed in a fluid by a flow of the fluid that is injected from a circular nozzle perpendicular to the electrode plane can be represented by [14]: 50 2 (HId)  (2.4)  2  where H is the normal distance between the exit of the nozzle and the electrode, d is the diameter of the nozzle, p is the density of the fluid, and v is the velocity of the fluid flow at the nozzle exit. Figure 2.10 plots the deflections of the sample electrodes with varying flow velocity of the EDM fluid obtained by equations (2.2) and (2.4) with the relevant constants associated with the experimental set-up used in this study (H=19 mm, d=1.3 3 for the EDM fluid used). mm, and u=796 kg/m  The hydrodynamic pressure p is  assumed to be uniform over the electrodes for this approximated estimation. The plot indicates the calculated displacement of 75 jim for the fixed-fixed electrode with the flow velocity of 5.4 rn/s at which the short circuit was observed, which matches well with the actual gap separation (70 jim) between the electrode and the workpiece under the assumption. In addition, with the same equations, the flow velocities that were required to initiate the breakdown are estimated to produce ‘—.40 jim displacement as seen in Figure 2.10, leaving 30-jim gap spacing. This is approximately consistent with the estimated pull-in gap of 25 jim at 90 V for the cantilever electrode obtained using equation (2.3). (For the fixed-fixed, this gap is estimated to be somewhat smaller as it is stiffened by the non-linear stress due to the forced deflection by flow already). The machining process/system as well as the device construction will need some improvement and optimization for increased performance and practicality of the process. It was observed that some of the electrodes were stuck to the substrates while implementing the process, preventing further machining (Figure 2.9c). This is likely caused by local welding of the electrodes to the substrates (as they were measured to be short-circuited to the substrates). The result can be partially due to open-loop control of fluid flow in the set-up used in this study, which could force the electrodes to physically  29  100  —  — —  Cantilever Fixed-fixed /  —80 60  Measured velocities (3.4 rn/s for cantilever, 3.9 rn/s  /  /  /  /  /  Fluid flow velocity, v (mis) Figure 2.10: Theoretical deflections vs. fluid flow velocity for the two electrodes in  Figure 2.2 calculated using the experimental conditions of the fluidic set-up used (Figure 2.6). touch the substrate. In addition, it was visually observed that the carbon residues tended to adhere to the electrodes and remained around the machining space. This can lead to two deleterious phenomena. One is the generation of secondary discharges through the carbon particles, which can cause excess material removal. The result of the relatively large discharge gap of 10 pm (Figure 2.9c) may be related to this effect. The other is the generation of irregular arcing between the electrode and the workpiece that produces a significant amount of heat [15]. This may have contributed to both the local welding, possibly along with the physical contact of the electrode to the workpiece, and the excess removal effect. These issues will be addressed by using feedback control for the fluid flow in synchronization with the discharge pulse generation as well as enhancing the dispersion of the carbon byproducts (by optimizing ultrasonic wave application, use of flushing steps, etc.) and/or minimize the carbon production using pure water as a dielectric fluid for the EDM process [16]. The electrodes need to be multi-layer constructions that incorporate custom features 30  onto the bottom of suspended plates (corresponding to the single-layer electrodes used in this effort). The addition of the microstructures at the bottom of the support can vary the electrostatic force depending on the geometry of the custom features. The use of external force for the electrode actuation would contribute to minimizing such dependences of the EDM method on application-associated factors. 3 M  The microfabrication of the  suspended multi-layer structures can potentially be approached using similar dry-film processes with additional patterning and electroplating steps  —  the sacrificial etching of  the dry-film resist will need to be optimized accordingly.  2.7  Conclusions  A MEMS-based iEDM method using hydrodynamic force for the actuation of microelectrodes fabricated on the workpiece has been studied. The dry-film photoresist processes with its sacrificial etching were developed for the fabrication of suspended electrode structures of copper with I 8-jim thickness.  The fabricated devices were  successfully utilized, as both movable electrodes and capacitive elements of the pulse generation circuitry, to produce pulsed micro sparks with the electrodes driven by controlled flow of the EDM fluid. Micromachining of stainless steel was experimentally demonstrated using the devices, achieving removal depth of 20 pm with the machining voltage of 90 V. The results obtained suggest that 1 iEDM can potentially serve as a nonNC, large-area batch processing technique with high applicability enabled by the external actuation approach. The theoretical analysis of the measured results obtained with the fabricated devices revealed that the behavior of the electrodes actuated by fluidic flow could be described well with the analytical models used.  Acknowledgments The authors would like to thank NSERC for their financial support to this research and Ms. Vijayalakshmi Sridhar at the University of British Columbia for assisting in the fabrication process.  31  References [1]  Masaki T, Kawata K and Masuzawa T 1990 Micro electro-discharge machining and its applications Proc. IEEE Conf on Micro Electro Mechanical Systems pp 21-6  [2]  Takahata K and Gianchandani Y B 2007 Bulk-metal-based MEMS fabricated by micro-electro-discharge machining Canadian Conf on Electrical and Computer Engineering (Vancouver, Canada, 22-26 April) pp 1-4  [3]  Reynaerts D, Meeusen W, Song X, Van Brussel H, Reyntjens S, De Bruyker D and Puers R 2000 Integrating electro-discharge machining and photolithography: work in progress I Micromech. Microeng. 10 189-95  [4]  Guckel H 1998 High-aspect-ratio micromachining via deep X-Ray lithography  [5]  Proc. IEEE 86 pp 1586-93 Takahata K and Gianchandani Y B 2002 Batch mode micro-electro-discharge machining IEEE I Microelectromech. Syst. 11 102-10  [6]  Liao Y, Chen S, Lin C and Chuang T 2005 Fabrication of high aspect ratio microstructure arrays by micro reverse wire-EDM I Micromech. Microeng. 15 1547-55  [7]  Yi S M, Park M S, Lee Y S and Chu C N 2008 Fabrication of a stainless steel shadow mask using batch mode micro-EDM Microsyst. Technol. 14 411-7  [8]  Alla Chaitanya C R and Takahata K 2008 Micro-electro-discharge machining by MEMS actuators with planar electrodes microfabricated on the work surfaces IEEE  [9]  Int. Conf on Micro Electro Mechanical Systems (MEMS 2008) pp 375-8 Pamidighantam 5, Puers R, Baert K and Tilmans H A C 2002 Pull-in voltage  analysis of electrostatically actuated beam structures with fixed-fixed and fixed-free end conditions I Micromech. Microeng. 12 45 8-64 [10] Judy J W, Muller R S and Zappe H H 1995 Magnetic microactuation of polysilicon flexure structures IEEE I Microelectromech. Syst. 4 162-9  [11] Nguyen NT, Troung T  Q, Wong K K, Ho S S and Low C L N 2004 Micro check  valves for integration into polymeric microfluidic devices I Micromech. Microeng. 14 69-75  32  [12] Lee S W, Kim D J, Ahn Y and Chai Y G 2006 Simple structured polydimethylsiloxane microvalve actuated by external air pressure I Mech. Sci. 220 1283-8 [13] Young W C 2002 Roark’c formulas for stress and strain (New York: McGrawHill) [14] Rajaratnam N 1976 Turbulentjets (New York: Elsevier Scientific Pub. Co.) [15] Jameson E C 2001 Electric discharge machining (Society of Manufacturing Engineers) [16] Lin C T, Chow H M, Yang L D and Chen Y F 2007 Feasibility study of micro-slit EDM machining using pure water mt. i Adv. Manuf Technol. 34 104-10  33  Chapter 3 MEMS-based micro-electro-discharge machining EDM) by electrostatic actuation of machining 3 (M electrodes on the workpiece  3.1  Introduction  Micro-electro-discharge machining (tEDM) is a micromachining technique applicable to any electrically conductive material and is used to produce micromechanical structures. EDM is a machining process where the sequential discharge of electrical pulses between the electrode and workpiece is used to remove the workpiece material in the presence of dielectric fluid.  The technique is capable of producing real three-dimensional  microstructures and offers smallest feature size of 5 jim with submicron tolerance [1]. IiEDM has been used for producing micro components as well as prototyping various bulk-metal-based MEMS [2, 3]. However, the throughput is inherently low because the technique is essentially a serial process that uses a single electrode tip and the machining also relies on numerical control (NC) of the electrode’s position. Batch-mode jiEDM that uses arrays of high-aspect-ratio microelectrodes has been demonstrated to achieve high parallelism/throughput of the process [4].  In this approach, the arrays were  fabricated using a LIGA process (a combination of X-ray lithography and electroplating) [5] and were advanced into the workpiece using the vertical NC stage in an jiEDM apparatus.  It was observed that the throughput was substantially improved over the  traditional process due to the use of the arrays, but the drawback is the high costs incurred in the LIGA process. In addition, the process still requires the NC capability for vertical positioning of the arrays.  “A version of this chapter is in preparation for journal submission”. Alla Chaitanya Chakravarty Reddy EDM) by electrostatic 3 and Kenichi Takahata, “MEMS-based micro-electro-discharge machining (M actuation of machining electrodes on the workpiece”.  34  In general, non-traditional micromachining techniques such as EDM, mechanical, and electrochemical machining that use single-tip microtools are performed with ultraprecision NC apparatus, which incur high costs of ownership, and the energy and space required to operate such machines are as large as those for macro-scale machining apparatus and do not scale well with the size of the objects to be machined. There have been some efforts for the realization of “micro factory” to address such issue. A micro lathe was developed as part of the concept [6]. The basic approach was an extension of precision engineering, i.e., the miniaturization of conventional NC systems. The use of MEMS may be one way to implement selected micromachining processes. This paper EDM) method where 3 describes a MEMS-enabled micro-electro-discharge machining (M planar electrodes are microfabricated directly on the surfaces of the work material and actuated for controlled generation of discharge pulses towards high-throughput, low-cost, and high-precision processing that does not require NC machines.  This study  investigates the M EDM method based on the electrostatic actuation of the electrodes 3 using machining voltage of the tEDM.  The design and fabrication of the movable  microelectrode arrays for custom machining and the experimental results are reported. This paper is constituted as follows. Section 3.2 describes the machining principle using the developed M EDM method. The design of the movable electrode devices with 3 different configurations for the electrostatic actuation is discussed in Section 3.3. Section 3.4 deals with the fabrication of the electrode devices. Section 3.5 presents the results of .iEDM experimental characterization for the fabricated devices and the demonstration of 1 with the devices. Section 3.6 discusses the analysis of the experimental results from Section 3.5. Section 3.7 concludes the overall effort.  3.2  Machining Principle  The developed M EDM method uses planar electrodes that are suspended by the anchors 3 through the tethers above the surfaces of the conductive workpiece with a relatively large gap as shown in Figure 3.la. A resistance-capacitance (RC) pulse generation/timing circuit, which is a typical configuration for tEDM to achieve reduced parasitic capacitance in the circuit [1], is coupled between the device and a DC voltage source (80Portion of this manuscript has appeared in the conference abstract in [7].  35  140 V) so that the electrode serves as a cathode whereas the workpiece is an anode. The In addition to the  workpiece with the electrode is immersed in dielectric EDM oil.  external capacitance Ce as part of the RC circuit, as indicated in Figure 3.la, there is a parasitic/built-in capacitance, Cb, between the electrode-anchor and the workpiece, which contributes to an increase in the total capacitance of the system. Figure 3.lb illustrates the mechanical and electrical behaviors of the device in the Dielectric anchor/spacer  (a)  / Electrode  (2) Pull in & breakdown (3) Release Electrode lifted-off after EDM  -1  r  r,  Voltage at electrode  v,I  (b)  t  0 Discharge current  Figure 3.1: (a) Cross sectional view of the MEMS-based iEDM and its steps; (b) dynamic behavior of discharge voltage and current corresponding to the steps.  36  machining process. Initially, the device is disconnected from the power supply and the RC circuit (Ce is fully charged) hence the electrode is stationary (step-i). Upon the connection to the circuit, Cb is also charged up to the supply voltage, electrostatically driving the electrode towards the workpiece.  With properly designed structures at a  selected voltage, the phenomenon known as “pull-in” takes place when the restoring spring force through the tethers can no longer balance the electrostatic force with the decrease of the gap spacing [9].  This results in a breakdown before the electrode  physically touches the surface of the workpiece, producing a spark current due to a discharge from the capacitors (step-2).  The thermal impact generated by the spark  removes the material and leaves a crater-like shape on the surface. The discharge lowers the voltage at the capacitors, i.e., between the electrode and the workpiece, releasing the electrode (step-3).  Simultaneously, the capacitors are charged through the resistor,  restoring the voltage at the gap and inducing the electrostatic actuation again.  This  sequence of pull-in and release of the electrode is used to achieve self-regulated generation of discharge pulses that etch the material by repeating the unit removal by a single discharge. The mechanism with the large separation at the static state is also intended to automatically prevent irregular continuous arcing as it lowers the gap voltage, releasing the electrode back to the original position and physically terminating the arc. The wide gap also promotes easier flushing of byproducts produced during the machining. Another M EDM method using external (hydrodynamic) force produced by controlled 3 flow of EDM fluid was reported in [8]. The method using electrostatic actuation with the machining voltage is useful for etching with relatively smaller depth compared to the externally actuated method. This approach enables a significantly simpler machining set up as it utilizes the standard tEDM generator for the actuation without using an external force generator such as the fluidic control module in the system.  3.3  Device Design  In this study, two types of electrode devices were designed. The first design uses a single-layer construction for the planar electrode-suspension structures without particular features underneath the planar structure. This design was intended to serve for feasibility testing of the method using the electrostatic actuation principle with the simple  37  construction and fabrication of the devices. The single-layer design is also useful for custom machining for certain patterns where electrodes can be continuous planar structures with some holes corresponding to the target patterns. (For example, machining of micro pillars can be implemented with this type of single-layer electrodes with through holes). Various planar electrodes were designed and fabricated with sizes ranging from hundred’s of microns to millimeters for vertical (fixed-fixed and cantilever) and torsional actuation. A sample layout of the electrode structure designed for the torsional actuation is shown in Figure 3 .2a. The electrode-tether structures were constructed by thin-layer of copper that has been used as an electrode material for tEDM [4]. A fixed-fixed and cantilever electrodes were also designed with length and width dimensions of 5.OxO.952 electrode and 2.05x0.80-mm mm 2 electrode respectively.  All the above mentioned  designs were using single layer structures and common 2.5x2.5-mm 2 anchors.  As  described later, a copper layer with 1 8-j.tm thickness was used for both types of designs to fabricate the suspended electrode-tether structures, and 30x30-11m 2 holes are provided in the structures to enhance the sacrificial etching for releasing during the fabrication process. The other design uses the electrodes with double-layer structures to incorporate micro features on the bottom surfaces of the planar electrodes with custom patterns to be machined that require individual electrode structures and a supporting base for the structures.  (For example, micro-hole machining can be implemented by this type of  double-layer electrodes with micro-pillar structures underneath). The supporting base structures of this design are structurally identical to the single-layer electrodes with the tethers, anchors, and through-hole arrays for the sacrificial-etching purpose. Figure 3.2b shows the sample layout design of the electrode device using the multiple-layer construction. To support larger planer electrodes, this particular design selected a crab leg-type tether/suspension. A pull-in voltage,  VpITS,  for torsional actuation in the similar configuration can be  described by [10, 11]: /O.83Kd KTL  =ab3[_.3.361___J]  38  (3.1)  9500 -  (a)  O Hoç O 3 2 :rn x  1  (b)  Custom design electroplated structures (50x50 Jim ) 2  8 8  1 Anchor  “V  Movable electrode plate  U ni.  .•  pm  Figure 3.2: A sample layout of the tEDM device for (a) torsional design with the singlelayer scheme; (b) custom design structures with the double-layer scheme.  where d is the original separation between the electrode and the workpiece,  is the  permittivity of the EDM oil, L is the length of the electrode, W is the width of the electrode,  KTL  is the torsional spring constant, G is the shear modulus of elasticity, 2a is  the width of the tether, 2b is the thickness of the tether and us the length of the tether. (The lateral dimensions are noted in Figure 3.2a.). Similarly, a pull-in voltage, Vp 1 for vertical actuation of the electrode can be described by [9]: /8K d 3 P1  —  27  2E h 3 ‘  [(kit / 4)  —  tanh( k1 1 /  4)1’  CL =  3l  where A is the area of the capacitive electrode. The parameter K is the effective stiffness  39  of the suspended structures, defined as  KFF  for the fixed-fixed and KCL for cantilever  structures respectively, 11 is the length of both the electrodes and  k =.JP/(EJ)  where P is  the axial force created by the combination of the intrinsic stress in the suspended structure and the non-linear stress that arises due to the deflection of the structure, E is the Young’s modulus of the electrode material and I is the moment of inertia given by Pwh / 3 12 where w and h are the width and thickness of the electrode respectively. These  expressions were used to design the device structures for torsional and vertical actuation of the electrode using the relevant constants (G=45 GPa  ,  E=120 GPa h18 im for  8=1.59x10 F/rn for kerosene-based EDM oil). For an EDM voltage of 80— copper and 11 140 V the structures in Figure 3.2a were designed to be pulled in when the gap, d, varies from 16 23 im for torsional design whereas the fixed-fixed and cantilever designs with -  the dimensions mentioned above were designed to be pulled in relatively large gaps of 20 —28 m and 23 -33 jtm, respectively, with the same voltage range.  3.4  Fabrication  Figure 3.3 illustrates the cross-sectional view of the fabrication process for the suspended planar electrodes, which uses stock copper foil as the starting material [8j. The type-304 stainless steel in a form of 3” wafer was selected as work material and served as the substrate for the fabrication in this effort. Figure 3.3 shows two processes (a) and (b) that are developed to fabricate the electrode devices with single-layer and double-layer structures described in the preceding section respectively with different design configurations. For the process (a), first, a stainless-steel wafer is thoroughly cleaned and degreased with acetone. A layer of hexamethyldisilazane adhesion promoter is spun on the wafer. A thick photoresist (SPR22O, Rohm and Haas Co., PA, USA) is then double coated on the wafer to form a sacrificial layer with the thickness of 3 0-40 tim, which is followed by a soft baking process on a hot plate  90°C  for 5 mm (step a-i). The sacrificial layer can  also be a single coating on the wafer with the thickness of 15  —  18 jim depending on the  electrode designs. Next, a 1-jim-thick photoresist (S1813, Rohm and Haas Co., PA, USA) is spun on the sacrificial layer (step a-2), and then the copper foil with 18-jim thickness is laminated on the S1813 that serves as an adhesive layer between the copper 40  1) Spin coat 401.tm thick sacrificial resist and soft bake on the workpiece  -  Resist  Workpiece (Stainless steel wafer) SPR22O  2) Spin coat ijim thick  iCu foil  3) Laminate 18 jim thick Cu foil and soft bake  (a)  I Mylar çmask Resist SPR22O  4) Spin coat 5jim thick resist and pattern  5) Wet etch Cu  6) Sacrificial etch in acetone and release electrodes  1) Spin coat 30-pmthick resist exose [ and pattern on Cu foil 2) Electroplate Cu 3)  ‘I,  ‘I’  ‘I —  Glass plate  Glass plate  I  Resist S181 3  Spin coat s:cnficiaI layers on the workpiece  (b)  Resist SPR22O Cu foil with alignment markers  / —  Workpiece (Stainless steel wafer)  ‘I,  ‘I  ‘I  SPR22O (soft-baked)  ‘I’  4) Spin coat 5pmthick resist on Cu and pattern I  5) Wet etch Cu foil to shape electrode and anchor  —Anchor  Electrode array  6)Sacnficial etch in electrodes  Figure 3.3: Two processes developed for electrode fabrication with (a) single-layer structures; (b) double-layer structures.  41  foil and the sacrificial layer. The S1813 resist is soft baked on a hot plate 90°C for 10 mm to solidify the layer, making the copper foil rigidly fixed down to the substrate through the sacrificial layer (step a-3). Then, a 5-jim-thick layer of SPR22O is spun on the copper foil and patterned using a mylar mask with the layout of the devices (step a-4). The developed SPR22O is used as a mask for wet etching of copper in a commercially available ferric chloride solution (CE- 100, Transene Co., Inc., MA, USA) (step a-5). Finally, timed etching of the sacrificial layer is performed in acetone, providing suspended electrodes while leaving the anchors attached to the substrate (step a-6). As noted earlier, the holes created in the electrode-tether structures promote undercutting during the sacrificial etching process while leaving the sacrificial layer as the anchors without the holes. Figure 3.4 shows the fabricated devices after the sacrificial etching using single layer structures. In the process (b), microstructures to be formed on the bottom of the electrode are fabricated by copper electroplating on the same copper foil used in the process (a). First, a 30-jim-thick resist SPR22O is coated on the copper foil, followed by a soft baking process on a hot plate 90°C for 30 mm. Prior to the resist coating through-hole markers are created in the copper foil for the alignment of the patterns to be created on the top and bottom of the copper foil. The resist is then exposed and patterned using the mylar mask to create the molds (step b-I). Next, copper is electroplated in the mold of the patterned SPR22O forming the 25-jim-thick microstructures on the copper foil (step b-2). Figure 3.5a shows an example of electroplated structures obtained after this step. As a separate process, 25-jim-thick resist SPR22O is coated on the stainless-steel workpiece to form the sacrificial layer, followed by a soft baking process on a hot plate 90°C for 5 mm. Next, a 1-jim-thick resist S1813 is coated on the sacrificial layer to serve as an adhesive layer for the bonding of the copper foil with microstructures underneath. The bonding is followed by the soft baking process on a hot plate 90°C for 10 mm (step b-3). Then, a 5-jim-thick SPR22O is coated on the top of the copper foil and patterned (step b-4) to be used as a mask for wet etching of copper to define the overall shapes (step b-5). The final timed sacrificial etching in acetone to obtain suspended electrodes is performed in the similar manner as described in process (a) above (step b-6). Figure 3.5 shows the fabricated devices for the custom design structures with the crab-leg design and electroplated arrays  42  of microstructures on the backside of the planar electrodes.  —  Iede  (a) ji  I  fJ w  I  I  I  ilUUjJiLL  1mm  I  Stainless Steel  (b) /  Cu electrode  1 mm  Figure 3.4: (a) An optical image of the fabricated devices on a stainless-steel substrate; (b) an SEM image of a sample device fabricated for the single-layer design in Figure 3.2a; 43  (a)  (b)  (c)  iece Figure 3.5: Fabrication results for the double-layer design: (a) an SEM image of the electroplated copper on the copper supporting electrode; (b) an optical image of the fabricated device with the design in Figure 3.2b; (c) an SEM image of the close-up.  44  3.5  Experimental Results  A set-up shown in Figure 3.6 was used for electromechanical characterization of the electrode devices and jiEDM tests. The fabricated device was placed in an ultrasonic bath filled with the EDM oil (EDM  TM, 185  Commonwealth Oil Co., ON, Canada). An  RC circuit with a 20-K2 resistor and a 100-pF (corresponding to the C in Figure 3.la) capacitor was connected to the device and the DC voltage source as shown in Figure 3.6. The electrical discharge pulses were monitored using a current probe (CT-2, Tektronix, Inc., TX, USA) inserted between the capacitor and the electrode. A laser displacement gauge (LK-G82, Keyence Co., NJ, USA) that provided the spot size of 70 jim with the resolution and response speed of 0.2 jim and 20 jis, respectively, was used to measure the vertical actuation of the electrode in situ. An jiEDM process produces byproducts that  Current probe  EDM fluid Ultrasonic bath Figure 3.6: A set-up for device characterization and jiEDM tests. include the particles removed from the workpiece and carbon residues produced from the EDM oil. The removal of the byproducts is important to prevent irregular arcing and stabilizing the process. This was achieved by applying ultrasonic waves during jiEDM, (i.e., machining was performed while continuously “cleaning” the workpiece in the ultrasonic bath). The sequential pulses of micro spark discharge were successfully generated using the electrostatic actuation mechanism (Figure 3.7). The peak current, pulse duration, and charging time constant were approximately measured to be 2.5 A, 50 ns, and 1 jis, respectively, in the set-up used.  Figure 3.8a shows the stainless-steel workpiece  machined using the torsional electrode in Figure 3.4b with 100 V for 10 mm.  45  The  removed depth was measured to be up to 20 jim.  The machining voltage used is  theoretically not large enough to pull-in this particular torsional with an initial gap separation of 40 lim. The successful pull-in and machining may be attributed to the aid of vertical actuation in addition to the torsional one of the design. Figure 3.9 shows a result of the in-situ measurement of the displacement of a single-layer fixed-fixed and cantilever electrodes with the designs noted in Section 3.3, which had approximate initial gap separations of 15 jim and 40 jim, respectively. The thickness of 15 ,im corresponds to the single-coated sacrificial layer and that of 40 jim to the double-coated one. For a clear visibility of the actuation effect, the temporal measurement was performed while switching the machining voltage (of 50V for fixed-fixed electrode and 100 V for cantilever electrode), i.e., EDMing on and off manually.  The result indicates the  approximate vertical actuation of 10 jim and 30 jim respectively at their measurement points (the center and at the tip) for the fixed-fixed and cantilever electrodes. From the result, the discharge/breakdown gap is estimated to be 5  —  10 jim for the electrodes and  setup.  (a)  (b)  Figure 3.7: (a) Measured current pulses of micro spark discharge; (b) single pulse close up; (c) an optical image of the micro sparks captured through electrode’s holes. 46  The machining that only used the built-in capacitance Cb as the capacitive element of the RC circuit without using the external capacitor Ce was also tested. The machining tests for the double-layer electrodes were implemented with this set-up.  Figure 3.8b  shows the hole arrays obtained using the crab-leg electrode with the machining voltage of 90V.  It was observed that many of the double-layer electrodes did not provide  displacements large enough to initiate and generate discharge pulses. It is expected that  (a)  (pm)  N  0  -6.5 (b)  -11.5 -16.5  -23.9 Figure 3.8: (a) An SEM image of the iEDMed stainless-steel substrate (electrode removed after machining) with inset of close-up; (b) optically measured geometry of the machined structures using  WykoTM  NTI 100 profiler. Batch structures are measured to  have the depth of— 20 rim..  47  the used voltage was not high enough, i.e. the electrostatic force to pull-in due to the relatively large gap of 55 tm between the workpiece and the supporting plate that provides the major part of the capacitive area.  Static level  :1  E Co  (a) ci)  2 a)  4 WI  o  5  EDM:  Off  On  15  20  25  10  30  35  40  45  50  Time (s) Static level  (b)  :i  -I’  0 15  s30  U  ci)  5 4 -D  2  EDM:  60 0  20  40  60  On  100 80 Time (s)  Off 120  140  160  180  Figure 3.9: Vertical displacement of the electrode measured while switching on and off the EDM process using the single-layer electrodes; (a) fixed-fixed device with 15 jim gap spacing; (b) cantilever device with 40 im gap spacing.  48  3.6  Analysis and Discussion  For better understanding of the machining process, it is worthwhile analyzing the electromechanical characteristics of the developed electrode device, which is used as both an electrostatic actuator and a capacitive element for the discharge pulse formation. This section discusses theoretical analysis of the experimental results for the built-in capacitance as well as mechanical behavior of the devices.  A. Built-in capacitance The discharge energy of a single pulse is given by 2 CV 1 2 [11, where C is the total capacitance that is the sum of the external capacitance (Ce) and the built-in capacitance (Cb) in the set-up shown in Figure 3.6 and V is the applied voltage. The energy determines the degree of the thermal impact and the size of the resultant crater-like shape hence the roughness and quality of the machined surfaces. In order to reduce the discharge energy for finer machining along with a simple circuit configuration, EDM can be effectively implemented by using only Cb to form the RC circuit without Ce (as demonstrated in the result in Figure 3.8b). For this implementation, it is important to evaluate the value of Cb, however, the direct measurement is difficult because Cb is a dynamic parameter as the actuation of the electrode modulates it, and Cb at the timing of a breakdown,  Cbb,  is the value of interest. The time constant, v=RCbb, in a charging cycle  can be used to indirectly estimate  Cbb.  The constant r was determined by probing the  voltage at the electrode and measuring the charging time.  Figure 3.lOa shows the  measured time constant for single-layer electrode devices which were calculated from the time taken to drop to 90 % of the voltage. Figure 3.1 Ob plots r measured in the dynamic mode (i.e., while EDM is on) for various devices with different electrode-tether areas as well as Cbb calculated from the results. The static value, Cb, measured directly by probing the devices without applying machining voltage using HP 4275A LCR meter is also plotted in Figure 3.lOb for comparison. The result indicates approximate linear dependence of both Cb and  Cbb  (or ) on the area of the device.  Figure 3.lOb also  suggests, as predicted, that Cbb is substantially increased from Cb, modulating the discharge energy. 49  (a)  40  R=20 kC, No external capacitor, V=100 V  LL  Dynamic mode (calculated from measured time constant)  ° 3 U)  C) C  (b)  800 CD QQ)  600.<W ØCD  4,”  20 CU  o  30  400  Static mode (measured directly)  •  0 C)  0 CD D  10  200 D  0 5  10  15  20  25  30  0 35  Total area of electrode and anchor (mm ) 2 Figure 3.10: (a) Measured machining voltage showing charging cycles with different time constants for the electrodes with different areas; (b) built-in capacitance in static (Ci,) and dynamic (Cbb) modes vs. device area.  50  B. Mechanical Response  The planar electrodes used in this process are actuated toward the workpiece against thin film of the EDM oil that is present in the gap spacing between the electrode and the workpiece. The actuation mechanism of the electrodes under the influence of the applied voltage can be modeled using d’Alembert’s principle and is given by [13]:  x 2 d d 2 l2pG(A)S m—+c-—+k m x=F;c= e 3 N,ird di’ 2 dt  (3.3)  where m is the mass of the copper electrode, c is the squeeze film damping coefficient attributed to the narrow gap spacing between the suspended electrode and the workpiece, x is the displacement of the electrode, displacement,  ,  t  is the time taken by the electrode to make the  km is spring constant corresponding to the spring constant discussed in  Section 3.3), Fe is the electrostatic force, p is the fluid viscosity of EDM oil, S is the area of the electrode, N is the number of holes in the electrode and  G(A) =  —  —  —  where A is the fraction of the open area in the electrode. The mechanical response of the fixed-fixed electrode and cantilever electrode is analyzed based on the following assumptions. The electrode is at the breakdown gap (assuming 5 jim from the experimental result) after a spark discharge, the velocity at this position for the electrode is 0 and the electrostatic force (Fe) is negligible when compared to the spring force to pull-back the electrode to its initial gap separation. The mechanical response of the fixed-fixed and cantilever electrode using the equation (3.3) with the constants used in the experimental set-up for this study (p =2. lxi 0 Pa-s) and the mathematical modeling software MapleTM was measured for the fixed-fixed and cantilever electrodes with single-layer structures. The displacements that are made with these electrodes are calculated to be on the order of nanometers to sub-nanometers within 1 jis that is a typical measured pulse interval. They will be even smaller as the machining voltage, i.e. electrostatic force is restored during the cycle. This suggests the position of the electrode around the discharge gap during the jiEDM process.  The in-situ  measurement of the displacement of the cantilever electrode in Figure 3.9b suggests the  51  electrode is stabilized at the discharge gap for a relatively long-on time  (- 80 sec) which  agrees well with the mechanical response analysis.  3.7  Conclusions  A new technique for tEDM using the electrostatic actuation of the planar electrodes microfabricated on the workpiece has been studied. Laminated 18-jim-thick copper foil was patterned to construct the planar electrodes suspended with the anchors, which were actuated with the machining voltage for well-controlled pulse formation of micro spark discharges.  The planar electrodes microfabricated directly on the stainless steel  workpiece material using standard UV lithography and wet etching of copper foil are suspended above the anchors using the sacrificial etching of the resist Layers. Micromachining of stainless steel with the voltage of 90  —  100 V was demonstrated using  both the single-layer electrodes and the double-layer electrodes with custom patterns providing approximate removal depth of 20 jim for the machining time of 10  -  15 mins.  The built-in capacitance of the electrode structure was exploited to form the pulse generation circuitry and experimentally characterized both in dynamic and static modes. The mechanical response of the electrode was theoretically analyzed to estimate the vertical position of the electrode after each discharge spark cycle. It was observed that the mechanical response of the electrode was far slower than the electrical response resulting in stable iEDM process.  The developed method offers an opportunity to  eliminate the need for high-precision NC machines for jiEDM thereby potentially achieving dramatic improvements of the process in not only the cost reduction but also the machinable area by introducing large-area patterning techniques for the device fabrication.  Acknowledgments The authors would like to thank NSERC for their financial support to this research. They also thank Ms. Vijayalakshmi Sridhar and Mr. Greg Wong at the University of British Columbia for their assistance in the fabrication process.  52  References  [1]  T. Masaki, K. Kawata, and T. Masuzawa, “Micro Electro-Discharge Machining and its Applications,” Proc. IEEE mt. Workshop on Micro Electro Mechanical Systems  [2]  (MEMS ‘90), Feb. 1990, pp. 21-26. L.L. Chu, K. Takahata, P. Selvaganapathy, Y.B. Gianchandani, and J.L. Shohet, “A Micromachined Kelvin Probe with Integrated Actuator for Microfluidic and SolidState Applications,” I Microelectromech. Syst., vol. 14, no. 4, Aug. 2005, pp. 691698.  [3]  K. Takahata, and Y.B. Gianchandani, “Bulk-Metal-Based MEMS Fabricated by Micro-Electro-Discharge Machining,” Proc. IEEE Canadian Conf Electr. Comput.  [4]  Eng. (CCECE ‘07), Vancouver, Canada, Apr. 2007, pp. 1-4. K. Takahata, and Y.B. Gianchandani, “Batch Mode Micro-Electro- Discharge  [5]  Machining,”I Microelectromech. Syst., vol. 11, no.2, Apr. 2002, pp. . 1 102 10 H. Guckel, “High-Aspect-Ratio Micromachining via Deep X-Ray Lithography,” Proc. IEEE, vol. 86, no. 8, 1998, pp. 1586-1593.  [6]  Y. Okazaki, N. Mishima, and K. Ashida, “Microfactory  -  Concept, History, and  Developments,” I Manuf Sci. Eng., vol. 126, no. 4, Nov. 2004, pp. 837-844. [7]  C. R. Alla Chaitanya, and K. Takahata, “Micro-electro-discharge machining by MEMS actuators with planar electrodes microfabricated on the work surfaces,” in Tech. Dig. IEEE  mt.  Conf on Micro Electro Mechanical Systems (MEMS ‘08),  Tucson, AZ, Jan. 2008, pp. 375-3 78. [8]  EDM: MEMS-enabled micro-electro 3 C. R. Alla Chaitanya, and K. Takahata, “M discharge machining”, I Microelectromech. Syst., vol. 18, 105009, Sep. 2008, 7pp.  [9]  S. Pamidighantam, R. Puers, K. Baert, and H.A.C. Tilmans, “Pull-in Voltage Analysis of Electrostatically Actuated Beam Structures with Fixed-Fixed and Fixed-Free End Conditions,” I Micromech. Microeng., vol. 12, July 2002, pp. 458464.  [10] J. Cheng, J. Zhe, and X. Wu, “Analytical and finite element model pull-in study of rigid and deformable electrostatic microactuators,” I Micromech. Microeng., vol. 14, Jan. 2004, pp. 57-68. 53  [11] W.C. Young, Roark Formulasfor Stress and Strain,  7 t h  ed., McGraw-Hill, 2002.  [12] F. Zhang, A.K. Prasad, and S.G. Advani, “Investigation of a Copper Etching Technique to Fabricate Metallic Gas Diffusion Media,” I Micromech. Microeng., vol. 16, Sep. 2006, pp. N23-N27. [13] W. Weaver Jr., S. P. Timoshenko, and D. H. Young, Vibration Problems in Engineering,  5 t h  ed., John Wiley & Sons, New York, 1990.  [14] T. B. Gabrielson, “Mechanical-thermal noise in micromachined acoustic and vibration sensors,” IEEE Trans. On Electron Devices, vol. 40, May 1993, pp. 903909.  54  Chapter 4 Conclusions The goal of the tEDM technique developed in this research effort is to increase the throughput, cost-effectiveness and to achieve high-precision machining. In Chapter 2, a MEMS-based iiEDM method that uses the hydrodynamic force for the actuation of the electrode towards the workpiece was developed. The use of dry film photoresist was also implemented to fabricate the planer electrodes.  A resistance-capacitance circuit  connected between the electrode and the workpiece was used to achieve controlled generation of discharge pulses with the applied DC voltage of 80  —  140V. The planar  electrodes were then actuated using the downflow of the EDM fluid. Micromachining of stainless steel was demonstrated with a removal depth of 20 im at 90V. The built capacitance between the electrode and the workpiece was utilized to control the discharge energy which defines the roughness and quality of the machined surfaces.  The  experimental results obtained were also verified with the theoretical models of electrode deflection using the hydrodynamic force.  The fabrication approach using the  hydrodynamic force offers the scalability towards large-area, high-throughput batch processing with tEDM at low cost.  Some of the applications where MEMS-based  .tEDM can be targeted include machining precision holes, ink-jet nozzles and fabrication of micro-stencils. In Chapter 2, the 1 iEDM method enables the large deflection of the electrode using the external fluid control module. The control module requires additional setup for jj,EDM process thereby increasing the equipment costs. In Chapter 3, the planar electrodes are electrostatically actuated towards the workpiece thereby eliminating the need for the external fluid module. First, the planar electrodes were constructed to be single-layer structures using I 8-tm-thick Cu foil as electrode material that were suspended with the anchors, to demonstrate the machining principle with the electrostatic actuation.  Micromachining of stainless steel for the  single-layer electrode structure was demonstrated with a removal depth of 20 im at 100V. Second, the planar electrodes were constructed to be double-layer structures for custom machining. The microelectrodes with the patterns to be machined on the workpiece were 55  formed on the bottom surface of the planar structure suspended by anchors. Micromachining of stainless steel was demonstrated with custom design structures patterned on the workpiece.  In Chapter 3, the mechanical response of the planar  electrodes during the actuation mechanism was also analyzed. It was observed that the mechanical response of the electrode to pull-back to its original gap separation was slow when compared to the electrical response of 1 ps after each discharge cycle. It suggested that the electrode was around the breakdown gap during the pJEDM process. In the electrostatic actuation method developed in Chapter 3, the machinable depth is limited as the machining enlarges the gap spacing between the electrode and the workpiece, lowering the electrostatic force to pull in the electrode. Higher voltages can be used to increase the force while compromising the surface roughness and quality in the machined structures as they increase the discharge energy as well. To extend the machinable depth, applying external force to the electrodes towards the workpiece instead of using the electrostatic actuation to sustain the machining process can be used with the need of an extra module for the actuation.  The method in Chapter 2  implemented this approach using the hydrodynamic force with the use of the fluid control module. The jiEDM techniques developed in concluding chapters eliminates the need of NC machines resulting in high throughput, low costs and potentially offers an opportunity for extending the technology to large-area patterning for the device fabrication. It was observed that the copper foil which was used as electrode material was deformed partially during tEDM. This could be attributed to the high temperature zones created during the spark discharge cycle [1]. The future work would involve the thermal analysis of the copper foil during the j.tEDM process [2]. The use of alternative electrode materials (like tungsten) that provides higher melting point can be used to minimize electrode wear and damage due to the thermal effect.  56  References [1]  J. McGeough, Micromachining ofEngineering Materials, Marcel Dekker, 2002.  [2]  S. H. Yeo, W. Kurnia, and P. C. Tan, “Electro-thermal modeling of anode and cathode in micro-EDM”, I Phys. D: Appi. Phys., vol. 40, pp. 2513-2521, 2007.  57  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0066936/manifest

Comment

Related Items