Swimming microrobots are expected to play an important role in medical actuators most important elements of the micro fins device




PPy Au Bilayer Actuator Actuated Micro-fins Device for Underwater Microrobotics 
 
 
 

Designed by Kuangwen Hsieh 
 

Fall 2004 ENME 489F

Prof. E. Smela

 
 

1. Introduction

1.1 Motivation

      The obvious reason behind microfabricating the micro-fins device is so that the device can be eventually implemented on the ANT microrobot.  The advantage to add swimming mechanism to walking or crawling microrobots is intuitive.  This addition makes the microtobots amphibious, gives them advantages over those that cannot swim, and adds more functionality.  Furthermore, the development of underwater, swimming microrobots that only swim also receives growing attention.  Swimming microrobots are expected to play an important role in medical and industrial applications such as surgical operations, cell manipulations and pipeline maintenance [1, 2]. 

      Current actuators used in these swimming microrobots, however, usually have shortcomings including small displacement, large voltage requirement and power consumption and bioincompatibility [1].  For example, thermal bimorph actuator that consists of two layers with different thermal expansion coefficients actuates due to thermally-induced strain.  However, this type of actuator also requires larger amount of heat and power for actuation to occur.  A design of polyimide bimorph actuators reported by Ataka et al. show that the temperature rises to 260°C for their 500-µm-long cantilever to deflect 150 µm vertically.  In addition, the input current of 65 mA at 16 V is supplied to their design [3].  The heat required for actuating bimorph thermal actuators will kill most cells, rendering the microactuators useless in medical applications.  The voltage level, at 16 V, is also too high. 

      Conjugated polymer actuators, on the other hand, offer many advantages including the ability to be electrically controlled, having a large strain and thus large displacement, requiring low voltages for actuation, working at room and/or body temperature and ability to operate in liquid electrolytes, including body fluids [4].  Typical gold and polypyrrole bilayer actuators, developed by Smela et al. and will be the focus of this paper, operate between 0 to -1 V in room temperature liquid electrolyte [5].  These advantages motivate the merge between conjugated polymer based actuators and microrobotics.  

       

1.2 Device Description

      Before introducing the major components of the device, a few abbreviations and nomenclatures must be introduced first for clarity.  Table 1 contains all abbreviations and nomenclatures used in this report.  The standard abbreviations such as Si for silicon and SiO2 for silicon oxide will be used throughout the report.  Moreover, gold, chromium and polypyrrole will be abbreviated as Au, Cr and PPy respectively.  Finally, a negative photoresist called SU8 is used to fabricate the device.  SU8 can be used to fabricate high-aspect-ratio structures.  Moreover, after baking at 150 to 200°C for around 15 minutes, SU8 becomes a permanent structure, which is desirable for the micro-fins device.  Detailed description and data sheet for SU8 can be retrieved online at http://www.microchem.com/products/su_eight.htm.       

      A swimming microrobot using gold (Au) and polypyrrole (PPy) bilayer as the micro-actuator is introduced in this design project.  The purpose of this project is to demonstrate that PPy/Au bilayer actuators can be implemented in swimming microrobotics application.  PPy is chosen over other conjugated polymers because it is readily available in University of Maryland.    The microrobot features a pair of microfabricated, flapping “micro-fins” actuated by the PPy/Au bilayer actuators.  Figures 1 and 2 below show a top view diagram and a three-dimensional side view diagram of major parts of the micro-fins device.

      The components of the device include the electrode, the support beam and support frame, hinges, actuators and the micro-fins.  The electrode is made of Au deposited on SiO2.  The electrode serves as the connection to a voltage-controlling device called potentiostat so the device can be powered and tested.  The description for actuating and testing the device is found in the next section.  The support beam consists of (from bottom of the wafer to the top) SiO2, Si, SiO2 and SU8.  It makes the mechanical support for the hinges that are connected to the actuators.  The support beam is held in place by the surrounding support frame.  The support frame is the portion of the oxidized silicon wafer that is not etched during fabrication.  The frame also divides one device from another on the wafer.  The purpose of the hinges is to mechanically hold the actuators in place.  The layer structure of hinges, from bottom of the wafer to the top, consists of SiO2, Cr, Au and SU8.  The Au under SU8 in the hinges, which cannot be seen from the top view diagram, connects the Au electrode to the actuators and serves as wires between the electrode and the actuators.  The actuators are the most important elements of the micro-fins device.  They are made of Au and PPy bilayers.  From the top view, only PPy can be seen.  Au is underneath PPy.  The actuation mode of these actuators is bending up and down.  Details for device actuation are described in the next section.  The actuators connect to the micro-fins so the fins can flap up and down as they actuate.  The micro-fins consist of Au and SU8.  Upon actuation, the fins flap up and down analogous to how human pedal when they swim, and create propelling force for swimming.     
 
 
 

SiO2

SU8

PPy

Background Color / Empty Space

Au

Figure 1: Top view of the micro-fins device.  The device diagram is color coded by the Material seen in the top view.  The Au electrode serves as the connection to a potentiostat so the device can be powered and tested.  The support beam consists of (from bottom of the wafer to the top) SiO2, Si, SiO2 and SU8.  It makes the mechanical support for the hinges that are connected to the actuators and micro-fins.  The hinges consist of SiO2, Cr, Au and SU8.  The Au under SU8 in the hinges, which cannot be seen from the top view diagram, serves as electrical connection between the Au electrode and actuators.  The actuators are made of Au and PPy bilayers.  In this diagram, Au is hidden underneath PPy so only PPy can be seen.  The actuation mode is bending up and down.  They connect to the micro-fins so the fins can flap up and down as they actuate.  The micro-fins consist of Au and SU8.  Upon actuation, the fins flap up and down to create force for swimming.  The support frame is the portion of the oxidized silicon wafer that is not etched during fabrication.  The support frame holds everything together and separates a device from another.  Note the black color in the picture is actually the background or empty space.  The region surrounded by the gray support frame and the green support beam indicates the etched out cavities so the micro-fins have room to flap. 
 
 
 

Figure 2: Three dimensional view of the micro-fins device.  The diagram is color-coded the same way as in Figure 1 – by the material seen in the top view.  From the diagram, it is clear that every element of this device is surface micromachined and locates above the substrate.   
 

1.3 Device Actuation

      The micro-fins device actuates by flapping the fins up and down.  The fins flap analogously to human’s pedaling legs when they swim.  In both cases, forces opposite to the direction of pedaling are created as the result.  The difference between micro-fins’ flapping and human feet pedaling is that both fins flap up and down synchronously.  The four pictures of Figure 3 below show the actuation mode of the device.  Figure 3 (a) and 3 (b) show two different three-dimensional side view diagrams of the micro-fins flapping down.  Figure 3 (c) and 3 (d) show the fins flapping up.  Note that the electrode and the support frame are not shown in the pictures.  Also note that Figure 3 can be misleading because the actuators in the figures remain straight as the fins flap up and down.  In reality, however, the micro-fins flap up and down due to the bending of the actuators.  The pictures in Figure 3 do not show the bending of the actuators due to the difficulty in drawing the diagrams.       
 
 

Figure 3 (a) 
 

Figure 3 (b) 
 

Figure 3 (c)

        
 

Figure 3 (d) 
 

Figure 3: Actuation modes of the micro-fins device.  Note that the electrode and the support frame are not shown in these figures.  Figure 3 (a) and 3 (b) show the fins flapping down while Figures 3 (c) and 3 (d) show the fins flapping up.  The fins repeat the flapping motion and create forward propulsion for the device.  The resulting direction of motion is shown by the arrow in each picture.  Note that these pictures are misleading because the actuators should bend in reality.  It is the bending of these actuators that causes the fins to flap up and down.  Figures 3 (a) through 3 (d) serve as demonstrations for micro-fins flapping and the resulting motion of the device.      
 

      The actuators used in the micro-fins device consist of two layers – a layer of mechanical passive thin film and a layer of polymer thin film.  In this case, the mechanical passive thin film is Au and the polymer thin film is PPy.  Therefore, the actuators are called PPy/Au bilayer.  Figure 4 (a) below shows the bilayer structure of the actuator. 

      PPy/Au bilayer actuators actuate electrochemically.  The reason that the actuation is electrochemical is because the actuation, which is a chemical reaction, is aided by electrodes.  Several requirements must be met for this electrochemical actuation to occur.  First, the actuator must be physically connected to a potentiostat as the working electrode (WE).  Next, a silver/silver chloride (Ag/AgCl) reference electrode (RE) and a counter electrode (CE) must be connected to the potentiostat as well.  In this case, a gold plate will work as the CE.  Then the actuator, RE and CE must be immersed in liquid electrolyte.  Finally, when a voltage is applied to the actuator, PPy will change volume, as well as its color and conductivity.  Au, on the other hand, does not change volume.  The discrepancy in strain between the two layers causes the actuator to bend.  This process is diagramed in Figure 4 (b).

       Another picture that shows the bending of the bilayer actuator is Figure 4 (c).  In Figure 4 (c), the top green PPy thin film reduces in volume and contracts the bottom yellow Au mechanical passive layer.  As the result, the PPy/Au bilayer actuator bends up.    Note again that the actuation diagramed in Figure 4 (c) will only occur under the conditions described above (as diagramed in Figure 4 (b)).

      Smela reported that PPy/Au bilayer actuators bend when PPy thin film oxidizes and straightens when PPy thin film reduces [5].  This is because PPy thin film changes volume as it oxidizes and reduces. 
 
 

https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

Figure 4 (a) 
 
 

https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

Figure 4 (b) 
 
 

https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

Figure 4 (c) 
 

Figure 4: The principle behind the actuators of the micro-fins device.  The structure of the actuators is given in Figure 4 (a).  The actuators consist of a layer of mechanical passive film and a layer of polymer film.  In this case, the mechanical passive film is Au and the polymer film is PPy.  In order to actuate the PPy/Au bilayer actuator, the set-up diagramed in Figure 4 (b) must be followed.  The actuator is connected to a potentiostat as the working electrode.  Then it is immersed in liquid electrolyte along with a silver/silver chloride reference electrode and a counter electrode that is usually gold.  When a voltage is applied through the potentiostat, PPy changes volume, but Au does not.  The result is the bilayer bending as in Figure 4 (b) and 4 (c).  The chemistry behind this volume-change phenomenon of PPy is decribed in section 2.1.      
 

 
 

2. Background

2.1 Volume-change Property of Polypyrrole

      For PPy, the primary factor that contributes to the volume-change property is believed to be the physical insertion and expulsion of ions and solvent molecules: when positive ions (cations) and water enter, the volume expands [7]. 

      Ions are transported in and out of PPy thin films when they go through oxidation and reduction (redox).  Oxidation involves the removal of electrons from polymers while reduction involves the addition of electrons to the polymers.  Oxidation and reduction level is controlled by an applied voltage.  When PPy thin films are oxidized, positive ions (cations) are expelled, and the volume of the thin films is reduced.  On the other hand, when PPy thin films are reduced, cations and solvents are inserted.  As the result, PPy thin films expand and volume increases [5].  In order for this reaction to be accomplished, a source and sink of ions outside of PPy must be supplied.  This source and sink of ions is usually a liquid electrolyte (salt water) [7].  Figure 5 below diagrams the volume-expansion property of PPy.  
 
 

https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

Figure 5: The principle behind the volume-change property of PPy.  When a positive voltage is applied, the PPy chain is reduced.  Positive ions (cations) and solvents are inserted in the PPy chain.  As the result, the volume expands.   
 

2.2 Applications of Polypyrrole/Gold Bilayer Actuators

      The use of PPy/Au bilayer actuator is pioneered and extensively studied by Smela et al.  The bilayer actuators are primarily used as hinges for different layers in their application.  They demonstrated many devices using PPy/Au bilayer actuators, including microvalves, self-assembly box, micro-origami and electrochemically driven PPy bilayer device for moving and positioning bulk micromachined silicon plates [4-7].  Figures 6 through 9 demonstrate these devices.

      The most relevant device to the micro-fins device, out of these applications, is the PPy bilayer actuator that moves and positions bulk micromachined silicon plates.  A picture of the device is given in Figure 9.  In the picture, the silicon plate is rotated at about 135°.  The difference between the work presented by them and the design presented in this report is that the plate being manipulated in their work is bulk micromachined while the micro-fins presented in this report will be surface micromachined.  An interesting finding that they reported is that the bilayer actuator rotates down to -90° when a 200-Å-thick layer of Cr was used with 800 Å of Au as the passive layer along with 1-µm-thick layer of PPy thin film as the active layer.      
 
 

https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

Figure 6: Application of PPy/Au bilayer actuaors: microvalve using an PPy/Au bilayer actuator as the hinge.  This microvalve design aims to prevent urinary incontinence, and it operates in urine [4].  When the bilayer is straight, the valve is closed.  When the bilayer is bent, the valve opens and allowed urine to flow.

   
 

http://www.wam.umd.edu/~smela/actuators.htm

Figure 7: Application of PPy/Au bilayer actuaors: self-folding box.     
 

http://www.wam.umd.edu/~smela/actuators.htm

Figure 8: Application of PPy/Au bilayer actuaors: micro-origami.

   
 

http://www.wam.umd.edu/~smela/actuators.htm

Figure 9: Application of PPy/Au bilayer actuaors: electrochemically driven PPy bilayer device for moving and positioning bulk micromachined silicon plates.  In this picture, the silicon plate is rotated at approximately 135°.  
 

2.3 Other Swimming Microrobots

      The field of swimming microrobotics is receiving growing attention.  Many different actuatiors that can be used in microrobotics are being studied.  They include using electrostatic actuators, thermal bimorph actuatiors, magnetic actuators, biomimetic actuators, piezoelectric actuators, and polymer based actuators [1, 2, 8-10, 12-16].  Different polymers that can be used as polymer microactuators are being explored as well [1, 16].

      An interesting swimming mechanism, actuated by external magnetic fields, is introduced by Honda et al.  Their device features a small cubic magnet attached to a spiral copper wire.  The schematic of their design is given below in Figure 10.  The swimming mechanism actuates in the presence of an alternating magnetic field or a rotational magnetic field.  The magnet rotates due to magnetic torque.  As the result, spiral waves propagate along the wire tail and then the mechanism can propel itself in the direction opposite to that of the wave propagation [8].  The four size parameters of the device are diameter of the copper wire (2a), diameter of the spiral (2b), linear wavelength of the spiral (λ), and total length of the wire when stretched straight (L).  Typical values are: 2a = 0.15 mm, 2b = 1 mm, λ = 3 mm, and L = 21.7 mm.  The shortcoming of this swimming micro-mechanism is that it is only tested in silicone oil, which has a much higher kinetic viscosity than water (tens of St to 0.01 St).       
 

T. Honda, K. I. Arai, K. Ishiyama, “Micro swimming mechanisms propelled by external magnetic fields,” IEEE Trans. Magnetics, vol. 32, issue 5, pp. 5085-5087, Sept. 1996.

Figure 10: Schematic of the swimming micro-mechanism introduced by Honda et al.  The magnet rotates due to magnetic torque in the presence of an alternating magnetic field or a rotational magnetic field.  As the result, spiral waves propagate along the wire tail and then the mechanism can propel itself in the direction opposite to that of the wave propagation.  Typical values are: 2a = 0.15 mm, 2b = 1 mm, λ = 3 mm, and L = 21.7 mm.  This device is only tested in silicone oil, which has a relatively high kinetic viscosity on the order of tens of St.

  

      Edd et al. presented a swimming microrobot featuring biomimetic propulsion that aims to remove kidney stones.  The propulsion mechanism for this design is inspired by the use of cilia and flagella in bacteria and spermatozoa.  Carbon nanotubes vertically grown from the substrate that will be actuated will serve as the micromachined flagella in this device.  Maneuver of the device will be accomplished by independent rotation of four plates, each having its own flagella.  Compare to PPy/Au bilayer actuator, this design has higher efficiency.  It is modeled to have 2% hydrodynamic efficiency.  Unfortunately, this design has yet been completed [9].   

      Other types of polymer based actuators, including ionic conducting polymer films (ICPF) actuator, parylene thermal actuators and polyaniline (PANI) actuators are currently being developed as micro-actuators for underwater microrobotics by Zhou et al. [1].  ICPF actuators actuate by stress gradient induced by ionic movement due to electric field.  Parylene thermal actuators actuate due to the induced stress gradient across a structure made of different layers of materials with different thermal expansion coefficients.  PANI actuators actuate due to its volumetric change caused by a reversible electrochemical oxidation-reduction reaction.  The actuation mechanism of PANI actuators is the same PPy/Au bilayer actuators.  These three polymer based actuators require less power input than conventional MEMS actuators.  Zhou et al. reported the capability to actuate all three actuators underwater using less than 7 V [1].  These actuators can be actuated underwater and thus can be advantageous over PPy/Au bilayer actuators.  Since these materials are relatively new to the MEMS community, research effort is focused on fabrication processes and procedures.  The use of these new polymer based actuators in microrobotics, however, has yet been extensively explored.  Figure 11 below shows a parylene actuator in action.  An applied voltage passes a current through the actuator, increases the actuator temperature, and triggers the actuation. 
 
 

J. W. L. Zhou, H. Y. Chan, T. K.H. To, K. W. C. Lai, W. J. Li, “Polymer MEMS actuators for underwater micromanipulation,” IEEE/ASME Trans. Mechatronics, vol. 9, issue 2, pp. 334-342, June 2004.

Figure 11: Parylene thermal actuator developed by Zhou et al. in action.  In this sequence of pictures, two probes make physical contact with conducting pads and pass a current through the device to trigger actuation.  The actuator bends up 90° at 2.5 V [1].   
3. Fabrication

3.1 Fabrication Sequence

In the description for fabrication sequence below, both diagrams and explanations are given.  Diagrams for both cross section view and top view are given.  There are also a few steps that have specified cross section view and top view denoted by dashed lines and letters.  In the explanation portion, justification for each fabrication step is given unless the step involves photoresist or appears in earlier steps.  Alternative fabrication methods are considered for some fabrication steps.  Comments are made to explain the functionality of certain materials, to further justify the fabrication step, or to clarify diagrams.  The following color code is used.

Si

SU8

PPy

SiO2

Au

Photoresist

Cr

 

 
 

 

Fabrication Sequence Diagrams

(Note: The diagrams are not drawn in scale)

 Steps and Explanations
SiO2

Si

1. Start with commercial oxidized Si wafer

Step 1 bottom view

SiO2

 Step 1: Start with Commercial Oxidized Si Wafer 

Comments

SiO2 serves the important role as the base layer for structures such as the micro-fins, hinges, and actuators to be built upon.  SiO2, an insulator, also separates Si from metal electrodes so the electrodes are not shorted out.

 
 

 
 

 

Photoresist 1813

2. Pattern photoresist 1813 with mask 1

Step 2 bottom view

SiO2

Photoresist 1813

Mask 1

 Step 2: Pattern Photoresist 1813 with Mask 1 

Explanation

The developer that will be used is developer 352.  Photoresist 1813 protects SiO2 that is underneath it from etchants.  The attached mask 1 is used during the UV light exposure step of the patterning process.  Since photoresist will almost always work, no alternative is considered. 

 

3. BHF etch away SiO2

Step 3 bottom view

Photoresist 1813

Si

 Step 3: Etch SiO2 with BHF 

Explanation

SiO2 is etched in order to reach the silicon surface underneath it and then etch the silicon to form cavities so the micro-fins can flap.  BHF is chosen to wet etch SiO2 for a couple of reasons.  1. It is readily available in University of Maryland.  2. It can be done in a few minutes, representing a faster process.  3. Since BHF does not attack silicon, the etch result is usually satisfactory.  The drawback of using BHF, however, is undercutting.  Undercutting occurs when SiO2 under photoresist is also etched. But since my goal is to etch silicon and form cavities, and since the dimension of the cavities is not that critical, undercutting will not be a very serious problem.  The dry etching alternative to BHF wet etching is using commercial gases such as CF4 + O2 (10%) or CHF3.  Unfortunately, the RIE system in University of Maryland is not equipped with such gases.  Purchasing these gases may be over the budget for MEMS 2 class.  So there is actually no alternative method to etch SiO2 in University of Maryland.

 
 

 

4. DRIE etch away Si

Photoresist 1813

SiO2

Step 4 bottom view

 Step 4: Etch Si using DRIE 

Explanation

In this step, silicon is etched away in order to make cavities so eventually the micro-fins will have room to flap during the testing stage.  DRIE is chosen because straight sidewalls are desired.  The straight sidewalls will reduce the overall size of devices and ultimately the cost, since more devices can be fabricated on one wafer.  Another reason is the availability.  University of Maryland has a DRIE system that is ready to go.  However, using DRIE has several drawbacks.  1. The whole 500 μm of silicon will be etched away, it will take a very long time (in the order of 10 hours) to dry etch silicon that thick with DRIE.  2. The expense of using the machine is great.  The alternative to DRIE is wet etching of silicon.  Out of the chemicals that can be used, KOH is the safest and readily available in University of Maryland.  Therefore, in the case that DRIE machine breaks down, wet etching silicon with KOH will be the alternative procedure.  The downside to this alternative procedure is that the mask must be re-designed as the result.  In the case of KOH etching, the sidewalls will be slanted at an angle of 54.7˚.  As the result, the dimension of the cavities shown on the mask must be much larger.

 
 

 

5. Remove photoresist 1813

Top SiO2 layer

Step 5 bottom view

Bottom SiO2 layer

Step 5: Remove Photoresist 1813 

Explanation

There are several resist strippers available to remove photoresist 1813.  Alternatively, one can expose the photoresist without any masks and then develop the photoresist.  In the case that photoresist is over-baked; piranha can be used to remove the photoresist as well.  In this fabrication sequence, assuming photoresist is not over-baked, UV light exposure with no masks and development of photoresist will be chosen as the method to remove photoresist.  The choice is rather arbitrary and is based on the familiarity with the material and the steps involved.  If KOH is used to etch silicon in the previous step, this step is not necessary.  It is shown repeatedly during the lab period that photoresist does not withstand KOH as it is removed as soon as wafers with photoresist are immersed in KOH.

Comments

In the bottom view picture, there are actually two different layers of SiO2 – the bottom layer that surrounds the cavities and the top layers that locate at the very bottom of the cavities.

 
 

 

Cr

6. Thermally evaporate Cr

Step 6 top view

Cr

 Step 6: Deposit Cr 

Explanation

There are three different methods to deposit Cr – thermal evaporation, electron beam evaporation, and sputtering.  University of Maryland is capable of all three methods.  The choice for one method over the other is arbitrary and may be determined primarily on equipment availability.  The film thickness should be around 3 nm because it serves only as an adhesion layer for Au portions that serve as hinges, wires and the electrode.

 
 

 

7. Pattern photoresist 1813 with mask 2

Photoresist 1813

B’

A’

B’

Step 7 top view

A’

Photoresist 1813

Cr

Mask 2

 Step 7: Pattern Photoresist 1813 on Cr with Mask 2 

Comments

Due to the complexity in shape of the photoresist pattern, dash cross section line is specified on the top view device diagram.  In the following steps, the cross section diagram is assumed to follow the same dash cross section line unless otherwise noted.

 
 

 

8. Cr etchant wet etch away Cr

Cr under photoresist 1813

Step 8 top view

 Step 8: Etch Cr with Cr Etchant 

Explanation

There may be more than one Cr etchant available in industry; however, there is only one type of Cr etchant available in University of Maryland.  Furthermore, dry etch of Cr is not a possibility.  So there is no choice in terms of choosing how to etch Cr in University of Maryland.  The drawback of this process is that it is time controlled and the etch duration for this process is rather short (in the order of 10 seconds).  If etched for too long, Cr underneath photoresist may be undercut.  It is possible, however, to dilute the Cr etchant to slow the etch rate down. 

 
 

 

9. Remove photoresist 1813

Step 9 top view

Cr

Step 9: Remove Photoresist 1813 off Cr 
 

 
 

 

10. Thermally evaporate Au

Au

Step 10 top view

Au

 Step 10: Deposit Au  

Explanation

Just like the case of Cr, Au deposition can also be done by thermal evaporation, electron beam deposition or sputtering.  In this case, since the step coverage is not an issue in the design, the choice for one method over the other is arbitrary and may be determined by equipment availability again.  The film thickness of the Au layer should be around 100 nm so Au will effectively cover the entire patterned Cr surface; additional explanation for Au film thickness is in section 4.  The technique to deposit Au on patterned Cr is part of a broader technique called differential adhesion demonstrated by Smela [4].  The reason this technique is used is because Au makes the bottom layer of both the actuator and micro-fins and needs to be released from SiO2 upon actuation.  This technique takes advantage of the fact that Au does not adhere to SiO2.  More details on differential adhesion will be given in the next step.

Comments

Au layer in this step serves both a mechanical role in the bilayer and as an electrode for the electrochemical deposition of PPy.

 
 

 

    11(a). Pattern photoresist 1813 with mask 3

A’

Photoresist 1813

B’

C’

Photoresist 1813

D’

    11(b). Pattern photoresist 1813 with mask 3

C’

B’

A’

Step 11 top view

Cr and Au under photoresist 1813

Au under photoresist 1813

D’

Mask 3

 Step 11: Pattern Photoresist 1813 on Au with Mask 3 

Explanation

Notice that mask 3 used to pattern Au is different from mask 2 that is used in Step 7 and 8 to pattern Cr.  Also notice that the Au in cross section from A’ to B’ is on top of a layer of Cr but the Au in cross section from C’ to D’ (the portion of actuator and micro-fins) is directly on top of SiO2.  Upon actuation, the actuator and micro-fins will lift off from SiO2 because Au does not adhere to SiO2 well. 
 

Comments

Notice there are two dash cross section lines in the top view device diagram to demonstrate the concept of differential adhesion.  In the following steps, if the diagrams are not specified, it is assumed that the cross section diagrams are based on the same cross section from A’ to B’.

 
 

 

A’

12. Au etchant wet etch away Au

B’

B’

A’

Step 12 top view

 Step 12: Etch Au with Au Etchant 

Explanation

As the case of Cr, there is more than one method to etch Au.  However, etching Au with wet iodine based etchant is effective and can be done in a relatively short period of time (in the order of 1 minute).  Again, the drawback is that the etch process is time controlled and undercut may occur.  But at University of Maryland, etching Au with wet Au etchant is a standard process and can be easily done.

 
 

 

13. Remove photoresist 1813

Only Au on SiO2

Cr and Au on SiO2

Step 13 top view

 Step 13: Remove Photoresist 1813 off Au 
 

 
 

 

14. Pattern SU8 photoresist with mask 4

D’

SU8

C’

Step 14 top view

D’

C’

SU8

Mask 4 (negative mask for SU8)

 Step 14: Pattern SU8 on Au 

Explanation

SU8 is chosen for this step because it is thick (on the order of a few microns to around a hundred microns), can be permanently deposited on Au and is readily available in University of Maryland.  SU8 serves two purposes for the micro-fins devices.  1. Increase the thickness and thus robustness of micro-fins.  Without SU8, the micro-fins consist of Au that is 100 nm (explained in section 4).  By adding a few microns of SU8, the fins will be thicker and more robust.  2. In the next step, PPy will be electrochemically deposited on Au that is not covered by SU8.  Since SU8 is not conductive, Au hidden underneath SU8 will not be electroplated with PPy.  So SU8 also acts as a patterning component for the next step.  The drawback, however, is that there is no alternative methods to SU8. 
 

Comments

Notice the mask for SU8 is the negative masks for those used with photoresist 1813.  This is because SU8 is a negative photoresist.  This means the portion of SU8 that is not exposed to UV light (black region on the mask) will actually dissolve away during development.

 
 

 

F’

15. Deposit PPy

PPy

E’

PPy/Au Bilayer

F’

PPy

Step 15 top view

E’

 Step 15: Electroplate PPy 

Explanation

PPy can be electrochemically deposited on a conducting surface such as Au [5].  This process is sometimes called electroplating as well.  When PPy is electroplated on Au, a bilayer forms.  This bilayer serves as the actuator of the device.  The reason PPy is chosen is because it is extensively studied in University of Maryland.  Therefore, PPy is well understood to some extend and is readily available in University of Maryland.  Another advantage to use PPy in PPy/Au bilayer structure is the large displacement in actuation.  There are other actuators that utilize similar bilayer ideas.  An example is thermal bimorph actuators.  Thermal bimorph actuators can potentially be an alternative to PPy/Au bilayer actuator if it fails.  However, the fabrication sequence may be significantly changed if thermal bimorph actuators are needed.  
 

Comments

Notice another different cross section between E’ and F’ is examined in the diagrams.  This cross section is used to demonstrate that PPy is electroplated on Au. 

F’

16. Release Micro-fins

E’

Step 16 top view

Cavity

F’

E’

 Step 16: Release Micro-fins 

Explanation

Smela reported that when PPy is electrochemically oxidized and reduced, PPy/Au bilayer actuator bends and lifts off from SiO2 surface [5].  Since the fins are also made by depositing Au directly on SiO2, the fins will (hopefully) be released from the SiO2 surface as they flap up.  Since the fins will be a few times thicker than the SiO2 layer, it is theorized that the fins will break the layer when they flap down. 
 

Comments

  1. Prior to actuating PPy/Au bilayer actuators and releasing the micro-fins, the wafer must be diced into individual devices because each device has separate connection between its own electrode and actuators
  2. The cavity will not look rectangular, as indicated in the diagram, because the SiO2 layer will be mechanically broken by the fins.    

 
 
 

3.2 Masks

      The magnified, but not-in-scale version of masks is diagramed in the previous section.  The real masks are attached below in Figure 13.  These masks are drawn in Adobe Illustrator CS.  For each mask, an individual mask and a collection of seven masks arranged on a quarter-wafer are shown.  Seven is the maximum number of devices that can fit on a quarter-wafer.    
 

Figure 13: The four masks that will be used to microfabricate the micro-fins device.  All of the masks are scaled correctly.  Refer to section 3.1 for larger but unscaled version of these masks.  In Figure 13, an individual mask and a collection of seven masks arranged on a quarter-wafer are shown for each mask.  Seven is the maximum number of devices that can fit on a quarter-wafer.  Mask 1 is used to pattern the SiO2 mask layer so Si can be etched to form cavities.  Mask 2 is used to pattern Cr.  Cr serves as the adhesion layer between SiO2 and Au.  Mask 3 is used to pattern Au.  The additional areas in mask 3 compared to mask 2 represents the actuators and the micro-fins.  The actuators and the fins will be released upon actuation.  This is why the actuators and the micro-fins are fabricated with Au base layer directly deposited on SiO2 without Cr.  Finally, mask 4 is used to pattern SU8.  Since SU8 is a negative photoresist, the dark field and the light field of mask 4 is revered.  The light field represents the area that will be covered with SU8 after SU8 photoresist development.

 
 

4. Device Design

      In order to fit the dimension of the ant-like microrobot, the support beam is set to be 4 mm in length.  The width is set to be 1 mm wide so it will be robust enough to hold the hinges mechanically.  As the result, the cavities will be 4 mm long as well.  The cavities are symmetric with respect to the support beam and are 2 mm in width.  They are 2 mm wide so the hinges, the actuators and the micro-fins can fit inside.  The support frame on each side for an individual device is 2.5 mm.  Therefore the two neighboring devices are separated by 5 mm on each side.  In addition, the support frame is 5 mm wide in the bottom.  So the separation between the top devices and the bottom devices is also 5 mm as well.  The overall size of each device, equipped with electrode and surrounded by the support frame is 20 mm in length and 10 mm in width.  Seven complete devices can fit on the quarter of a 4-inch wafer.  Figure 13 below shows 600%-magnified mask 1 and the dimension of the support beam, the cavities and the support frame. 
 

Figure 13: Magnified mask 1 and dimension of the support beam, the support frame and the cavities.  Mask 1 is used to pattern SiO2. 
 

      In order to establish physical connection between the electrode and the potentiostat, the electrode must be relatively big.  The dimension of the electrode will be arbitrarily chosen as 5 mm by 5 mm.  Alligator clips are usually used to clip on the electrode in order to make this physical connection from the potentiostat to the device.  When alligator clips are dipped in the electrolyte, electrochemistry will fail to work properly.  Therefore, there should be a relatively big separation between the electrode, which will not be immersed in the electrolyte, and the actuators.  The separation will be set at 6.8 mm to make sure the alligator clip will not be immersed in electrolyte by accident.  This is sufficient enough because alligator clips will not be clipped in such a way that it extends to the bottom end of the electrode.  The 6.8-mm gap includes 6 mm from the end of the electrode to the starting point of the support beam, a 0.2 mm separation between the sidewalls of the cavities to the, the first pair of hinges that are 0.2 mm wide as well, and 0.4 mm separation between the first pair of hinges and the second pair of the hinges. 

      The fins are 2.5 mm in length and 1.6 mm in width.  They are evenly symmetric with respect to the support beam in the middle.  The bottom end of the fins is separated from the sidewalls of the cavities by a 0.3-mm gap.  On the side, the gap between the sidewalls and the fins is 0.2 mm while the gap between the support beam and the fins is also 0.2 mm.  The fins are visible by eyes because the device must be vertically immersed in electrolyte when testing.  It will be hard to observe the flapping under a microscope this way of the device is not visible by eyes.   

      Smela reported that PPy/Au bilayer actuators 30 µm by 30 µm is capable of lifting rigid areas of 900 µm by 900 µm [5].  Since the rigid area of the each fin is 2500 µm by 1600 µm, the actuator for each fin must be several times bigger.  In order to be absolutely sure that the actuators will be able to lift the fins, two actuators, each 0.4 mm long and 0.2 mm wide are used.

      Finally, alignment marks are employed for masks 2, 3 and 4.  The alignment marks are important because there are many small features in the later steps of the fabrication sequence.  Figure 14 below shows the 600%-magnified mask 3.  Dimensions of the electrode, the micro-fins, the hinges, and the actuators are also given.

      The thickness of different layers should also be considered.  Cr should be really thin because it only serves as the adhesion layer.  Thickness under 100 Å will be sufficient enough for this application.  Au should be much thicker than Cr to cover the patterned Cr.  Again, Au serves both the passive layer in the bilayer actuator and as an electrode for electrochemical deposition of PPy.  Based on the report by Smela et al., 1000 Å is a good thickness for Au [7].  It is also reported by Smela et al. that actuation speed of PPy/Au bilayer actuators depend on the thickness of PPy thin film [7].  They find that the response time of the actuator increases as the thickness of PPy thin film increases.  Since the micro-fins cannot flap too slowly, the thickness of PPy should be relatively thin.  Smela et al. used PPy around 1 µm in thickness because the thickness is regulated by the thickness of photoresist in their fabrication process.  For the purpose of this project, 0.5 µm should be the appropriate thickness for PPy.  Finally, SU8 should be much thicker than all other materials.  It is hoped that the micro-fins should be able to break the SiO2 layer upon actuation and then flap up and down once they are free to move.  Since the thickness of the SiO2 layer is 0.5 µm, SU8 should be a couple of times thicker than SiO2.  Thickness of 4 to 5 µm for SU8 should be sufficient for this design.  Table 1 below records the thickness of each material. 
 
 

        Layer/Material Thickness (diff. Units)
        Cr < 100 Å
        Au 1000 Å
        PPy 0.5 µm
        SU8 4~5 µm

        

      Table 1: Different layers of material and their corresponding thicknesses. 
 
 

         

Figure 14: Magnified mask 3.  Dimensions of the electrode, the hinges, the micro-fins and the actuators are given as well.  Mask 3 is used to pattern Au.  In all three of mask 2, 3 and 4, alignment marks are added because the features are too small of naked-eye alignment.

 
 

5. Discussion

      Despite the advantages such as requiring low voltages for actuation and working at room and/or body temperature, PPy/Au bilayer actuators have some limitations, especially when applied in microrobotics.

      First, the efficiency of PPy/Au bilayer actuators is low.  Smela et al. reported 0.2% efficiency for their bilayer actuators used in moving and positioning bulk micromachined silicon plates [7].

      Next, the actuation of PPy/Au bilayer actuator is low.  Accodring to Smela et al., the fastest rate for bilayer actuators with PPy thickness of 1 µm to go from completely bent to completely straight and back was 2.5 Hz [5].  Usually, the actuation frequency for PPy/Au bilayer actuators to complete one cycle of actuation is around 1 Hz.  The actuation speed of the actuators is limited by ion transport in and out of PPy thin film.  As the result of slow actuation, the fins will flap slowly as well.  Slow flapping translates to slow swimming speed, which is undesirable.    

      Another limitation is that the device can only be actuated in liquid electrolyte.  This is because PPy thin film will only change volume in liquid electrolyte [7].  While this characteristic may be helpful in biomedical applications, it also limits the application of PPy/Au-bilayer-actuator-related devices.  For example, for investigation and maintenance of small pipes in a radiator, the swimming microrobot design presented in this report will not be applicable.

      Finally, the life-time of the actuator presents concerns as well.  Typical PPy/Au bilayer actuators fail after 1,000 to 10,000 cycles of actuation.  Failure is caused by delamination of PPy thin film from Au thin film [7].  This presents a problem for the long-term performance of the microrobot.

 

 
 

6. References

    [1]  J. W. L. Zhou, H. Y. Chan, T. K.H. To, K. W. C. Lai, W. J. Li, “Polymer MEMS actuators for underwater micromanipulation,” IEEE/ASME Trans. Mechatronics, vol. 9, issue 2, pp. 334-342, June 2004.

    [2]  Y. Okuda, S. Guo, Y. Hasegawa, K. Asaka, “A fin type of underwater microrobot with multi DOF,” in Proc. SICE, 2003, vol. 2, pp. 1656-1660.

    [3]  M. Ataka, A. Omodaka, N. Takeshima, H. Fujita, “Fabrication and operation of polyimide bimorph actuators for a ciliary motion system,” J. Microelectromechanical Systems, vol. 2, issue 4, pp.146-150, Dec. 1993.

    [4]  E. Smela, “Conjugated polymer actuators for biomedical applications,” Adv. Mater., vol. 15, No. 6, March 17 2003.

    [5]  E. Smela, “Microfabrication of PPy microactuators and other conjugated polymer devices,” J. Micromech. Microeng., vol. 9, pp. 1-18, 1999.

    [6]  E. Smale, O. Inganäs, I. Lundström, “Controlled folding of micrometer-size structures,” Science, vol. 268, pp. 1735-1738, 23 June 1995.

    [7]  E. Smela, M. Kallenbach, J. Holdenried, "Electrochemically Driven Polypyrrole Bilayers for Moving and Positioning Bulk Micromachined Silicon Plates," J. Microelectromechanical Systems, vol. 8, issue 4, pp. 373-383, 1999. 

    [8]  T. Honda, K. I. Arai, K. Ishiyama, “Micro swimming mechanisms propelled by external magnetic fields,” IEEE Trans. Magnetics, vol. 32, issue 5, pp. 5085-5087, Sept. 1996.

    [9]  J. Edd, S. Paven, B. Rubinsky, M. L. Stoller, M. Sitti, “Biomimetic propulsion for a swimming surgical micro-robot,” in Proc. IROS, 2003, vol. 3, pp.2583-2588.

    [10]  K. Ishiyama, K. I. Arai, M. Sendoh, A. Yamazaki, “Spiral-type micro-machine for medical applications,” in Proc. MHS, 2000, pp. 65-69.

    [11]  R. J. Wood, S. Avadhanula, M. Menon, R. S. Fearing, “Microrobotics using composite materials: the micromechanical flying insect thorax,” in Proc. ICRA, 2003, vol. 2, pp. 1842-1849.

    [12]  T. Fukuda, A. Kawamoto, F. Arai, H. Matsuura, “Mechanism and swimming experiment of micro mobile robot in water,” in Proc. MEMS, 1994, pp. 273-278.

    [13]  T. Fukuda, A. Kawamoto, F. Arai, H. Matsuura, “Steering mechanism and swimming experiment of micro mobile robot in water,” in Proc. MEMS, 1995, pp. 300.

    [14]  Y. Zhang, X. Wang, T. Mei, Q. Wang, P. Zhang, “Driving principle and dynamic analysis of a micro swimming robot,” in Proc. WCICA, 2004, vol. 5, pp. 4582-4586.

    [15]  M. G. Borgen, G. N. Washington, G. L. Kinzel, “Design and evolution of a piezoelectrically actuated miniature swimming vehicle,” IEEE/ASME Trans. Mechatronics, vol. 8, issue 1, pp. 66-76, March 2003.

    [16]  G. Laurent, D. Piat, “Efficiency of swimming microrobots using ionic polymer metal composite actuators,” in Proc. IEEE-ICRA, 2001, vol. 4, pp. 3914-3919. 
     






Download links for : << High confidence secure medical devices evidence based care with automated patient specific alerts million per year >>
"Swimming microrobots are expected to play an important role in medical actuators most important elements of the micro fins device"


How to Download
You may need eMule or Bittorrent to download ebook torrents or emule links.

Report Dead Link
Please leave a comment to report dead links, so that someone else may update new links.

Search More...

[share-ebook]Swimming microrobots are expected to play an important role in medical actuators most important elements of the micro fins device

Google

Related Books


Books related to :

<< High confidence secure medical devices evidence based care with automated patient specific alerts million per year

Advanced informatics techniques including enhanced sensor systems few years creating tremendous pressure on hospitals >>




See all the reviews
The New York Times rss

    ----------Nucleic Acids Research----------

    ----------Industrial Pharmacy----------

    ----------Drug Development----------
    ----------Material----------
    ----------Nanometer----------
    ----------Chemistry----------
    ----------Living-beings----------
    ----------Health----------
    ----------Polytechnic----------
    ----------Depression----------
    ----------Caregiver----------
    ----------Cancer----------
    ----------Back Pain----------
    ----------Asthma----------
    ----------Allergy----------
    ----------Radiation----------
    Google

    Swimming microrobots are expected to play an important role in medical actuators most important elements of the micro fins device

    PPy Au Bilayer Actuator Actuated Micro-fins Device for Underwater Microrobotics 
     
     
     

    Designed by Kuangwen Hsieh 
     

    Fall 2004 ENME 489F

    Prof. E. Smela

     
     

    1. Introduction

    1.1 Motivation

          The obvious reason behind microfabricating the micro-fins device is so that the device can be eventually implemented on the ANT microrobot.  The advantage to add swimming mechanism to walking or crawling microrobots is intuitive.  This addition makes the microtobots amphibious, gives them advantages over those that cannot swim, and adds more functionality.  Furthermore, the development of underwater, swimming microrobots that only swim also receives growing attention.  Swimming microrobots are expected to play an important role in medical and industrial applications such as surgical operations, cell manipulations and pipeline maintenance [1, 2]. 

          Current actuators used in these swimming microrobots, however, usually have shortcomings including small displacement, large voltage requirement and power consumption and bioincompatibility [1].  For example, thermal bimorph actuator that consists of two layers with different thermal expansion coefficients actuates due to thermally-induced strain.  However, this type of actuator also requires larger amount of heat and power for actuation to occur.  A design of polyimide bimorph actuators reported by Ataka et al. show that the temperature rises to 260°C for their 500-µm-long cantilever to deflect 150 µm vertically.  In addition, the input current of 65 mA at 16 V is supplied to their design [3].  The heat required for actuating bimorph thermal actuators will kill most cells, rendering the microactuators useless in medical applications.  The voltage level, at 16 V, is also too high. 

          Conjugated polymer actuators, on the other hand, offer many advantages including the ability to be electrically controlled, having a large strain and thus large displacement, requiring low voltages for actuation, working at room and/or body temperature and ability to operate in liquid electrolytes, including body fluids [4].  Typical gold and polypyrrole bilayer actuators, developed by Smela et al. and will be the focus of this paper, operate between 0 to -1 V in room temperature liquid electrolyte [5].  These advantages motivate the merge between conjugated polymer based actuators and microrobotics.  

           

    1.2 Device Description

          Before introducing the major components of the device, a few abbreviations and nomenclatures must be introduced first for clarity.  Table 1 contains all abbreviations and nomenclatures used in this report.  The standard abbreviations such as Si for silicon and SiO2 for silicon oxide will be used throughout the report.  Moreover, gold, chromium and polypyrrole will be abbreviated as Au, Cr and PPy respectively.  Finally, a negative photoresist called SU8 is used to fabricate the device.  SU8 can be used to fabricate high-aspect-ratio structures.  Moreover, after baking at 150 to 200°C for around 15 minutes, SU8 becomes a permanent structure, which is desirable for the micro-fins device.  Detailed description and data sheet for SU8 can be retrieved online at http://www.microchem.com/products/su_eight.htm.       

          A swimming microrobot using gold (Au) and polypyrrole (PPy) bilayer as the micro-actuator is introduced in this design project.  The purpose of this project is to demonstrate that PPy/Au bilayer actuators can be implemented in swimming microrobotics application.  PPy is chosen over other conjugated polymers because it is readily available in University of Maryland.    The microrobot features a pair of microfabricated, flapping “micro-fins” actuated by the PPy/Au bilayer actuators.  Figures 1 and 2 below show a top view diagram and a three-dimensional side view diagram of major parts of the micro-fins device.

          The components of the device include the electrode, the support beam and support frame, hinges, actuators and the micro-fins.  The electrode is made of Au deposited on SiO2.  The electrode serves as the connection to a voltage-controlling device called potentiostat so the device can be powered and tested.  The description for actuating and testing the device is found in the next section.  The support beam consists of (from bottom of the wafer to the top) SiO2, Si, SiO2 and SU8.  It makes the mechanical support for the hinges that are connected to the actuators.  The support beam is held in place by the surrounding support frame.  The support frame is the portion of the oxidized silicon wafer that is not etched during fabrication.  The frame also divides one device from another on the wafer.  The purpose of the hinges is to mechanically hold the actuators in place.  The layer structure of hinges, from bottom of the wafer to the top, consists of SiO2, Cr, Au and SU8.  The Au under SU8 in the hinges, which cannot be seen from the top view diagram, connects the Au electrode to the actuators and serves as wires between the electrode and the actuators.  The actuators are the most important elements of the micro-fins device.  They are made of Au and PPy bilayers.  From the top view, only PPy can be seen.  Au is underneath PPy.  The actuation mode of these actuators is bending up and down.  Details for device actuation are described in the next section.  The actuators connect to the micro-fins so the fins can flap up and down as they actuate.  The micro-fins consist of Au and SU8.  Upon actuation, the fins flap up and down analogous to how human pedal when they swim, and create propelling force for swimming.     
     
     
     

    SiO2

    SU8

    PPy

    Background Color / Empty Space

    Au

    Figure 1: Top view of the micro-fins device.  The device diagram is color coded by the Material seen in the top view.  The Au electrode serves as the connection to a potentiostat so the device can be powered and tested.  The support beam consists of (from bottom of the wafer to the top) SiO2, Si, SiO2 and SU8.  It makes the mechanical support for the hinges that are connected to the actuators and micro-fins.  The hinges consist of SiO2, Cr, Au and SU8.  The Au under SU8 in the hinges, which cannot be seen from the top view diagram, serves as electrical connection between the Au electrode and actuators.  The actuators are made of Au and PPy bilayers.  In this diagram, Au is hidden underneath PPy so only PPy can be seen.  The actuation mode is bending up and down.  They connect to the micro-fins so the fins can flap up and down as they actuate.  The micro-fins consist of Au and SU8.  Upon actuation, the fins flap up and down to create force for swimming.  The support frame is the portion of the oxidized silicon wafer that is not etched during fabrication.  The support frame holds everything together and separates a device from another.  Note the black color in the picture is actually the background or empty space.  The region surrounded by the gray support frame and the green support beam indicates the etched out cavities so the micro-fins have room to flap. 
     
     
     

    Figure 2: Three dimensional view of the micro-fins device.  The diagram is color-coded the same way as in Figure 1 – by the material seen in the top view.  From the diagram, it is clear that every element of this device is surface micromachined and locates above the substrate.   
     

    1.3 Device Actuation

          The micro-fins device actuates by flapping the fins up and down.  The fins flap analogously to human’s pedaling legs when they swim.  In both cases, forces opposite to the direction of pedaling are created as the result.  The difference between micro-fins’ flapping and human feet pedaling is that both fins flap up and down synchronously.  The four pictures of Figure 3 below show the actuation mode of the device.  Figure 3 (a) and 3 (b) show two different three-dimensional side view diagrams of the micro-fins flapping down.  Figure 3 (c) and 3 (d) show the fins flapping up.  Note that the electrode and the support frame are not shown in the pictures.  Also note that Figure 3 can be misleading because the actuators in the figures remain straight as the fins flap up and down.  In reality, however, the micro-fins flap up and down due to the bending of the actuators.  The pictures in Figure 3 do not show the bending of the actuators due to the difficulty in drawing the diagrams.       
     
     

    Figure 3 (a) 
     

    Figure 3 (b) 
     

    Figure 3 (c)

            
     

    Figure 3 (d) 
     

    Figure 3: Actuation modes of the micro-fins device.  Note that the electrode and the support frame are not shown in these figures.  Figure 3 (a) and 3 (b) show the fins flapping down while Figures 3 (c) and 3 (d) show the fins flapping up.  The fins repeat the flapping motion and create forward propulsion for the device.  The resulting direction of motion is shown by the arrow in each picture.  Note that these pictures are misleading because the actuators should bend in reality.  It is the bending of these actuators that causes the fins to flap up and down.  Figures 3 (a) through 3 (d) serve as demonstrations for micro-fins flapping and the resulting motion of the device.      
     

          The actuators used in the micro-fins device consist of two layers – a layer of mechanical passive thin film and a layer of polymer thin film.  In this case, the mechanical passive thin film is Au and the polymer thin film is PPy.  Therefore, the actuators are called PPy/Au bilayer.  Figure 4 (a) below shows the bilayer structure of the actuator. 

          PPy/Au bilayer actuators actuate electrochemically.  The reason that the actuation is electrochemical is because the actuation, which is a chemical reaction, is aided by electrodes.  Several requirements must be met for this electrochemical actuation to occur.  First, the actuator must be physically connected to a potentiostat as the working electrode (WE).  Next, a silver/silver chloride (Ag/AgCl) reference electrode (RE) and a counter electrode (CE) must be connected to the potentiostat as well.  In this case, a gold plate will work as the CE.  Then the actuator, RE and CE must be immersed in liquid electrolyte.  Finally, when a voltage is applied to the actuator, PPy will change volume, as well as its color and conductivity.  Au, on the other hand, does not change volume.  The discrepancy in strain between the two layers causes the actuator to bend.  This process is diagramed in Figure 4 (b).

           Another picture that shows the bending of the bilayer actuator is Figure 4 (c).  In Figure 4 (c), the top green PPy thin film reduces in volume and contracts the bottom yellow Au mechanical passive layer.  As the result, the PPy/Au bilayer actuator bends up.    Note again that the actuation diagramed in Figure 4 (c) will only occur under the conditions described above (as diagramed in Figure 4 (b)).

          Smela reported that PPy/Au bilayer actuators bend when PPy thin film oxidizes and straightens when PPy thin film reduces [5].  This is because PPy thin film changes volume as it oxidizes and reduces. 
     
     

    https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

    Figure 4 (a) 
     
     

    https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

    Figure 4 (b) 
     
     

    https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

    Figure 4 (c) 
     

    Figure 4: The principle behind the actuators of the micro-fins device.  The structure of the actuators is given in Figure 4 (a).  The actuators consist of a layer of mechanical passive film and a layer of polymer film.  In this case, the mechanical passive film is Au and the polymer film is PPy.  In order to actuate the PPy/Au bilayer actuator, the set-up diagramed in Figure 4 (b) must be followed.  The actuator is connected to a potentiostat as the working electrode.  Then it is immersed in liquid electrolyte along with a silver/silver chloride reference electrode and a counter electrode that is usually gold.  When a voltage is applied through the potentiostat, PPy changes volume, but Au does not.  The result is the bilayer bending as in Figure 4 (b) and 4 (c).  The chemistry behind this volume-change phenomenon of PPy is decribed in section 2.1.      
     

     
     

    2. Background

    2.1 Volume-change Property of Polypyrrole

          For PPy, the primary factor that contributes to the volume-change property is believed to be the physical insertion and expulsion of ions and solvent molecules: when positive ions (cations) and water enter, the volume expands [7]. 

          Ions are transported in and out of PPy thin films when they go through oxidation and reduction (redox).  Oxidation involves the removal of electrons from polymers while reduction involves the addition of electrons to the polymers.  Oxidation and reduction level is controlled by an applied voltage.  When PPy thin films are oxidized, positive ions (cations) are expelled, and the volume of the thin films is reduced.  On the other hand, when PPy thin films are reduced, cations and solvents are inserted.  As the result, PPy thin films expand and volume increases [5].  In order for this reaction to be accomplished, a source and sink of ions outside of PPy must be supplied.  This source and sink of ions is usually a liquid electrolyte (salt water) [7].  Figure 5 below diagrams the volume-expansion property of PPy.  
     
     

    https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

    Figure 5: The principle behind the volume-change property of PPy.  When a positive voltage is applied, the PPy chain is reduced.  Positive ions (cations) and solvents are inserted in the PPy chain.  As the result, the volume expands.   
     

    2.2 Applications of Polypyrrole/Gold Bilayer Actuators

          The use of PPy/Au bilayer actuator is pioneered and extensively studied by Smela et al.  The bilayer actuators are primarily used as hinges for different layers in their application.  They demonstrated many devices using PPy/Au bilayer actuators, including microvalves, self-assembly box, micro-origami and electrochemically driven PPy bilayer device for moving and positioning bulk micromachined silicon plates [4-7].  Figures 6 through 9 demonstrate these devices.

          The most relevant device to the micro-fins device, out of these applications, is the PPy bilayer actuator that moves and positions bulk micromachined silicon plates.  A picture of the device is given in Figure 9.  In the picture, the silicon plate is rotated at about 135°.  The difference between the work presented by them and the design presented in this report is that the plate being manipulated in their work is bulk micromachined while the micro-fins presented in this report will be surface micromachined.  An interesting finding that they reported is that the bilayer actuator rotates down to -90° when a 200-Å-thick layer of Cr was used with 800 Å of Au as the passive layer along with 1-µm-thick layer of PPy thin film as the active layer.      
     
     

    https://umd.blackboard.com/courses/1/200408_ENME489F_Smela/content/_11821_1/Lecture_13__Smela_research.pdf

    Figure 6: Application of PPy/Au bilayer actuaors: microvalve using an PPy/Au bilayer actuator as the hinge.  This microvalve design aims to prevent urinary incontinence, and it operates in urine [4].  When the bilayer is straight, the valve is closed.  When the bilayer is bent, the valve opens and allowed urine to flow.

       
     

    http://www.wam.umd.edu/~smela/actuators.htm

    Figure 7: Application of PPy/Au bilayer actuaors: self-folding box.     
     

    http://www.wam.umd.edu/~smela/actuators.htm

    Figure 8: Application of PPy/Au bilayer actuaors: micro-origami.

       
     

    http://www.wam.umd.edu/~smela/actuators.htm

    Figure 9: Application of PPy/Au bilayer actuaors: electrochemically driven PPy bilayer device for moving and positioning bulk micromachined silicon plates.  In this picture, the silicon plate is rotated at approximately 135°.  
     

    2.3 Other Swimming Microrobots

          The field of swimming microrobotics is receiving growing attention.  Many different actuatiors that can be used in microrobotics are being studied.  They include using electrostatic actuators, thermal bimorph actuatiors, magnetic actuators, biomimetic actuators, piezoelectric actuators, and polymer based actuators [1, 2, 8-10, 12-16].  Different polymers that can be used as polymer microactuators are being explored as well [1, 16].

          An interesting swimming mechanism, actuated by external magnetic fields, is introduced by Honda et al.  Their device features a small cubic magnet attached to a spiral copper wire.  The schematic of their design is given below in Figure 10.  The swimming mechanism actuates in the presence of an alternating magnetic field or a rotational magnetic field.  The magnet rotates due to magnetic torque.  As the result, spiral waves propagate along the wire tail and then the mechanism can propel itself in the direction opposite to that of the wave propagation [8].  The four size parameters of the device are diameter of the copper wire (2a), diameter of the spiral (2b), linear wavelength of the spiral (λ), and total length of the wire when stretched straight (L).  Typical values are: 2a = 0.15 mm, 2b = 1 mm, λ = 3 mm, and L = 21.7 mm.  The shortcoming of this swimming micro-mechanism is that it is only tested in silicone oil, which has a much higher kinetic viscosity than water (tens of St to 0.01 St).       
     

    T. Honda, K. I. Arai, K. Ishiyama, “Micro swimming mechanisms propelled by external magnetic fields,” IEEE Trans. Magnetics, vol. 32, issue 5, pp. 5085-5087, Sept. 1996.

    Figure 10: Schematic of the swimming micro-mechanism introduced by Honda et al.  The magnet rotates due to magnetic torque in the presence of an alternating magnetic field or a rotational magnetic field.  As the result, spiral waves propagate along the wire tail and then the mechanism can propel itself in the direction opposite to that of the wave propagation.  Typical values are: 2a = 0.15 mm, 2b = 1 mm, λ = 3 mm, and L = 21.7 mm.  This device is only tested in silicone oil, which has a relatively high kinetic viscosity on the order of tens of St.

      

          Edd et al. presented a swimming microrobot featuring biomimetic propulsion that aims to remove kidney stones.  The propulsion mechanism for this design is inspired by the use of cilia and flagella in bacteria and spermatozoa.  Carbon nanotubes vertically grown from the substrate that will be actuated will serve as the micromachined flagella in this device.  Maneuver of the device will be accomplished by independent rotation of four plates, each having its own flagella.  Compare to PPy/Au bilayer actuator, this design has higher efficiency.  It is modeled to have 2% hydrodynamic efficiency.  Unfortunately, this design has yet been completed [9].   

          Other types of polymer based actuators, including ionic conducting polymer films (ICPF) actuator, parylene thermal actuators and polyaniline (PANI) actuators are currently being developed as micro-actuators for underwater microrobotics by Zhou et al. [1].  ICPF actuators actuate by stress gradient induced by ionic movement due to electric field.  Parylene thermal actuators actuate due to the induced stress gradient across a structure made of different layers of materials with different thermal expansion coefficients.  PANI actuators actuate due to its volumetric change caused by a reversible electrochemical oxidation-reduction reaction.  The actuation mechanism of PANI actuators is the same PPy/Au bilayer actuators.  These three polymer based actuators require less power input than conventional MEMS actuators.  Zhou et al. reported the capability to actuate all three actuators underwater using less than 7 V [1].  These actuators can be actuated underwater and thus can be advantageous over PPy/Au bilayer actuators.  Since these materials are relatively new to the MEMS community, research effort is focused on fabrication processes and procedures.  The use of these new polymer based actuators in microrobotics, however, has yet been extensively explored.  Figure 11 below shows a parylene actuator in action.  An applied voltage passes a current through the actuator, increases the actuator temperature, and triggers the actuation. 
     
     

    J. W. L. Zhou, H. Y. Chan, T. K.H. To, K. W. C. Lai, W. J. Li, “Polymer MEMS actuators for underwater micromanipulation,” IEEE/ASME Trans. Mechatronics, vol. 9, issue 2, pp. 334-342, June 2004.

    Figure 11: Parylene thermal actuator developed by Zhou et al. in action.  In this sequence of pictures, two probes make physical contact with conducting pads and pass a current through the device to trigger actuation.  The actuator bends up 90° at 2.5 V [1].   
    3. Fabrication

    3.1 Fabrication Sequence

    In the description for fabrication sequence below, both diagrams and explanations are given.  Diagrams for both cross section view and top view are given.  There are also a few steps that have specified cross section view and top view denoted by dashed lines and letters.  In the explanation portion, justification for each fabrication step is given unless the step involves photoresist or appears in earlier steps.  Alternative fabrication methods are considered for some fabrication steps.  Comments are made to explain the functionality of certain materials, to further justify the fabrication step, or to clarify diagrams.  The following color code is used.

    Si

    SU8

    PPy

    SiO2

    Au

    Photoresist

    Cr

     

     
     

     

    Fabrication Sequence Diagrams

    (Note: The diagrams are not drawn in scale)

    		 Steps and Explanations
    SiO2

    Si

    1. Start with commercial oxidized Si wafer

    Step 1 bottom view

    SiO2

    		 Step 1: Start with Commercial Oxidized Si Wafer 

    Comments

    SiO2 serves the important role as the base layer for structures such as the micro-fins, hinges, and actuators to be built upon.  SiO2, an insulator, also separates Si from metal electrodes so the electrodes are not shorted out.

     
     

     
     

     

    Photoresist 1813

    2. Pattern photoresist 1813 with mask 1

    Step 2 bottom view

    SiO2

    Photoresist 1813

    Mask 1

    		 Step 2: Pattern Photoresist 1813 with Mask 1 

    Explanation

    The developer that will be used is developer 352.  Photoresist 1813 protects SiO2 that is underneath it from etchants.  The attached mask 1 is used during the UV light exposure step of the patterning process.  Since photoresist will almost always work, no alternative is considered. 

     

    3. BHF etch away SiO2

    Step 3 bottom view

    Photoresist 1813

    Si

    		 Step 3: Etch SiO2 with BHF 

    Explanation

    SiO2 is etched in order to reach the silicon surface underneath it and then etch the silicon to form cavities so the micro-fins can flap.  BHF is chosen to wet etch SiO2 for a couple of reasons.  1. It is readily available in University of Maryland.  2. It can be done in a few minutes, representing a faster process.  3. Since BHF does not attack silicon, the etch result is usually satisfactory.  The drawback of using BHF, however, is undercutting.  Undercutting occurs when SiO2 under photoresist is also etched. But since my goal is to etch silicon and form cavities, and since the dimension of the cavities is not that critical, undercutting will not be a very serious problem.  The dry etching alternative to BHF wet etching is using commercial gases such as CF4 + O2 (10%) or CHF3.  Unfortunately, the RIE system in University of Maryland is not equipped with such gases.  Purchasing these gases may be over the budget for MEMS 2 class.  So there is actually no alternative method to etch SiO2 in University of Maryland.

     
     

     

    4. DRIE etch away Si

    Photoresist 1813

    SiO2

    Step 4 bottom view

    		 Step 4: Etch Si using DRIE 

    Explanation

    In this step, silicon is etched away in order to make cavities so eventually the micro-fins will have room to flap during the testing stage.  DRIE is chosen because straight sidewalls are desired.  The straight sidewalls will reduce the overall size of devices and ultimately the cost, since more devices can be fabricated on one wafer.  Another reason is the availability.  University of Maryland has a DRIE system that is ready to go.  However, using DRIE has several drawbacks.  1. The whole 500 μm of silicon will be etched away, it will take a very long time (in the order of 10 hours) to dry etch silicon that thick with DRIE.  2. The expense of using the machine is great.  The alternative to DRIE is wet etching of silicon.  Out of the chemicals that can be used, KOH is the safest and readily available in University of Maryland.  Therefore, in the case that DRIE machine breaks down, wet etching silicon with KOH will be the alternative procedure.  The downside to this alternative procedure is that the mask must be re-designed as the result.  In the case of KOH etching, the sidewalls will be slanted at an angle of 54.7˚.  As the result, the dimension of the cavities shown on the mask must be much larger.

     
     

     

    5. Remove photoresist 1813

    Top SiO2 layer

    Step 5 bottom view

    Bottom SiO2 layer

    		Step 5: Remove Photoresist 1813 

    Explanation

    There are several resist strippers available to remove photoresist 1813.  Alternatively, one can expose the photoresist without any masks and then develop the photoresist.  In the case that photoresist is over-baked; piranha can be used to remove the photoresist as well.  In this fabrication sequence, assuming photoresist is not over-baked, UV light exposure with no masks and development of photoresist will be chosen as the method to remove photoresist.  The choice is rather arbitrary and is based on the familiarity with the material and the steps involved.  If KOH is used to etch silicon in the previous step, this step is not necessary.  It is shown repeatedly during the lab period that photoresist does not withstand KOH as it is removed as soon as wafers with photoresist are immersed in KOH.

    Comments

    In the bottom view picture, there are actually two different layers of SiO2 – the bottom layer that surrounds the cavities and the top layers that locate at the very bottom of the cavities.

     
     

     

    Cr

    6. Thermally evaporate Cr

    Step 6 top view

    Cr

    		 Step 6: Deposit Cr 

    Explanation

    There are three different methods to deposit Cr – thermal evaporation, electron beam evaporation, and sputtering.  University of Maryland is capable of all three methods.  The choice for one method over the other is arbitrary and may be determined primarily on equipment availability.  The film thickness should be around 3 nm because it serves only as an adhesion layer for Au portions that serve as hinges, wires and the electrode.

     
     

     

    7. Pattern photoresist 1813 with mask 2

    Photoresist 1813

    B’

    A’

    B’

    Step 7 top view

    A’

    Photoresist 1813

    Cr

    Mask 2

    		 Step 7: Pattern Photoresist 1813 on Cr with Mask 2 

    Comments

    Due to the complexity in shape of the photoresist pattern, dash cross section line is specified on the top view device diagram.  In the following steps, the cross section diagram is assumed to follow the same dash cross section line unless otherwise noted.

     
     

     

    8. Cr etchant wet etch away Cr

    Cr under photoresist 1813

    Step 8 top view

    		 Step 8: Etch Cr with Cr Etchant 

    Explanation

    There may be more than one Cr etchant available in industry; however, there is only one type of Cr etchant available in University of Maryland.  Furthermore, dry etch of Cr is not a possibility.  So there is no choice in terms of choosing how to etch Cr in University of Maryland.  The drawback of this process is that it is time controlled and the etch duration for this process is rather short (in the order of 10 seconds).  If etched for too long, Cr underneath photoresist may be undercut.  It is possible, however, to dilute the Cr etchant to slow the etch rate down. 

     
     

     

    9. Remove photoresist 1813

    Step 9 top view

    Cr

    		Step 9: Remove Photoresist 1813 off Cr 
     

     
     

     

    10. Thermally evaporate Au

    Au

    Step 10 top view

    Au

    		 Step 10: Deposit Au  

    Explanation

    Just like the case of Cr, Au deposition can also be done by thermal evaporation, electron beam deposition or sputtering.  In this case, since the step coverage is not an issue in the design, the choice for one method over the other is arbitrary and may be determined by equipment availability again.  The film thickness of the Au layer should be around 100 nm so Au will effectively cover the entire patterned Cr surface; additional explanation for Au film thickness is in section 4.  The technique to deposit Au on patterned Cr is part of a broader technique called differential adhesion demonstrated by Smela [4].  The reason this technique is used is because Au makes the bottom layer of both the actuator and micro-fins and needs to be released from SiO2 upon actuation.  This technique takes advantage of the fact that Au does not adhere to SiO2.  More details on differential adhesion will be given in the next step.

    Comments

    Au layer in this step serves both a mechanical role in the bilayer and as an electrode for the electrochemical deposition of PPy.

     
     

     

      11(a). Pattern photoresist 1813 with mask 3

    A’

    Photoresist 1813

    B’

    C’

    Photoresist 1813

    D’

      11(b). Pattern photoresist 1813 with mask 3

    C’

    B’

    A’

    Step 11 top view

    Cr and Au under photoresist 1813

    Au under photoresist 1813

    D’

    Mask 3

    		 Step 11: Pattern Photoresist 1813 on Au with Mask 3 

    Explanation

    Notice that mask 3 used to pattern Au is different from mask 2 that is used in Step 7 and 8 to pattern Cr.  Also notice that the Au in cross section from A’ to B’ is on top of a layer of Cr but the Au in cross section from C’ to D’ (the portion of actuator and micro-fins) is directly on top of SiO2.  Upon actuation, the actuator and micro-fins will lift off from SiO2 because Au does not adhere to SiO2 well. 
     

    Comments

    Notice there are two dash cross section lines in the top view device diagram to demonstrate the concept of differential adhesion.  In the following steps, if the diagrams are not specified, it is assumed that the cross section diagrams are based on the same cross section from A’ to B’.

     
     

     

    A’

    12. Au etchant wet etch away Au

    B’

    B’

    A’

    Step 12 top view

    		 Step 12: Etch Au with Au Etchant 

    Explanation

    As the case of Cr, there is more than one method to etch Au.  However, etching Au with wet iodine based etchant is effective and can be done in a relatively short period of time (in the order of 1 minute).  Again, the drawback is that the etch process is time controlled and undercut may occur.  But at University of Maryland, etching Au with wet Au etchant is a standard process and can be easily done.

     
     

     

    13. Remove photoresist 1813

    Only Au on SiO2

    Cr and Au on SiO2

    Step 13 top view

    		 Step 13: Remove Photoresist 1813 off Au 
     

     
     

     

    14. Pattern SU8 photoresist with mask 4

    D’

    SU8

    C’

    Step 14 top view

    D’

    C’

    SU8

    Mask 4 (negative mask for SU8)

    		 Step 14: Pattern SU8 on Au 

    Explanation

    SU8 is chosen for this step because it is thick (on the order of a few microns to around a hundred microns), can be permanently deposited on Au and is readily available in University of Maryland.  SU8 serves two purposes for the micro-fins devices.  1. Increase the thickness and thus robustness of micro-fins.  Without SU8, the micro-fins consist of Au that is 100 nm (explained in section 4).  By adding a few microns of SU8, the fins will be thicker and more robust.  2. In the next step, PPy will be electrochemically deposited on Au that is not covered by SU8.  Since SU8 is not conductive, Au hidden underneath SU8 will not be electroplated with PPy.  So SU8 also acts as a patterning component for the next step.  The drawback, however, is that there is no alternative methods to SU8. 
     

    Comments

    Notice the mask for SU8 is the negative masks for those used with photoresist 1813.  This is because SU8 is a negative photoresist.  This means the portion of SU8 that is not exposed to UV light (black region on the mask) will actually dissolve away during development.

     
     

     

    F’

    15. Deposit PPy

    PPy

    E’

    PPy/Au Bilayer

    F’

    PPy

    Step 15 top view

    E’

    		 Step 15: Electroplate PPy 

    Explanation

    PPy can be electrochemically deposited on a conducting surface such as Au [5].  This process is sometimes called electroplating as well.  When PPy is electroplated on Au, a bilayer forms.  This bilayer serves as the actuator of the device.  The reason PPy is chosen is because it is extensively studied in University of Maryland.  Therefore, PPy is well understood to some extend and is readily available in University of Maryland.  Another advantage to use PPy in PPy/Au bilayer structure is the large displacement in actuation.  There are other actuators that utilize similar bilayer ideas.  An example is thermal bimorph actuators.  Thermal bimorph actuators can potentially be an alternative to PPy/Au bilayer actuator if it fails.  However, the fabrication sequence may be significantly changed if thermal bimorph actuators are needed.  
     

    Comments

    Notice another different cross section between E’ and F’ is examined in the diagrams.  This cross section is used to demonstrate that PPy is electroplated on Au. 

    F’

    16. Release Micro-fins

    E’

    Step 16 top view

    Cavity

    F’

    E’

    		 Step 16: Release Micro-fins 

    Explanation

    Smela reported that when PPy is electrochemically oxidized and reduced, PPy/Au bilayer actuator bends and lifts off from SiO2 surface [5].  Since the fins are also made by depositing Au directly on SiO2, the fins will (hopefully) be released from the SiO2 surface as they flap up.  Since the fins will be a few times thicker than the SiO2 layer, it is theorized that the fins will break the layer when they flap down. 
     

    Comments

    1. Prior to actuating PPy/Au bilayer actuators and releasing the micro-fins, the wafer must be diced into individual devices because each device has separate connection between its own electrode and actuators
    2. The cavity will not look rectangular, as indicated in the diagram, because the SiO2 layer will be mechanically broken by the fins.    

     
     
     

    3.2 Masks

          The magnified, but not-in-scale version of masks is diagramed in the previous section.  The real masks are attached below in Figure 13.  These masks are drawn in Adobe Illustrator CS.  For each mask, an individual mask and a collection of seven masks arranged on a quarter-wafer are shown.  Seven is the maximum number of devices that can fit on a quarter-wafer.    
     

    Figure 13: The four masks that will be used to microfabricate the micro-fins device.  All of the masks are scaled correctly.  Refer to section 3.1 for larger but unscaled version of these masks.  In Figure 13, an individual mask and a collection of seven masks arranged on a quarter-wafer are shown for each mask.  Seven is the maximum number of devices that can fit on a quarter-wafer.  Mask 1 is used to pattern the SiO2 mask layer so Si can be etched to form cavities.  Mask 2 is used to pattern Cr.  Cr serves as the adhesion layer between SiO2 and Au.  Mask 3 is used to pattern Au.  The additional areas in mask 3 compared to mask 2 represents the actuators and the micro-fins.  The actuators and the fins will be released upon actuation.  This is why the actuators and the micro-fins are fabricated with Au base layer directly deposited on SiO2 without Cr.  Finally, mask 4 is used to pattern SU8.  Since SU8 is a negative photoresist, the dark field and the light field of mask 4 is revered.  The light field represents the area that will be covered with SU8 after SU8 photoresist development.

     
     

    4. Device Design

          In order to fit the dimension of the ant-like microrobot, the support beam is set to be 4 mm in length.  The width is set to be 1 mm wide so it will be robust enough to hold the hinges mechanically.  As the result, the cavities will be 4 mm long as well.  The cavities are symmetric with respect to the support beam and are 2 mm in width.  They are 2 mm wide so the hinges, the actuators and the micro-fins can fit inside.  The support frame on each side for an individual device is 2.5 mm.  Therefore the two neighboring devices are separated by 5 mm on each side.  In addition, the support frame is 5 mm wide in the bottom.  So the separation between the top devices and the bottom devices is also 5 mm as well.  The overall size of each device, equipped with electrode and surrounded by the support frame is 20 mm in length and 10 mm in width.  Seven complete devices can fit on the quarter of a 4-inch wafer.  Figure 13 below shows 600%-magnified mask 1 and the dimension of the support beam, the cavities and the support frame. 
     

    Figure 13: Magnified mask 1 and dimension of the support beam, the support frame and the cavities.  Mask 1 is used to pattern SiO2. 
     

          In order to establish physical connection between the electrode and the potentiostat, the electrode must be relatively big.  The dimension of the electrode will be arbitrarily chosen as 5 mm by 5 mm.  Alligator clips are usually used to clip on the electrode in order to make this physical connection from the potentiostat to the device.  When alligator clips are dipped in the electrolyte, electrochemistry will fail to work properly.  Therefore, there should be a relatively big separation between the electrode, which will not be immersed in the electrolyte, and the actuators.  The separation will be set at 6.8 mm to make sure the alligator clip will not be immersed in electrolyte by accident.  This is sufficient enough because alligator clips will not be clipped in such a way that it extends to the bottom end of the electrode.  The 6.8-mm gap includes 6 mm from the end of the electrode to the starting point of the support beam, a 0.2 mm separation between the sidewalls of the cavities to the, the first pair of hinges that are 0.2 mm wide as well, and 0.4 mm separation between the first pair of hinges and the second pair of the hinges. 

          The fins are 2.5 mm in length and 1.6 mm in width.  They are evenly symmetric with respect to the support beam in the middle.  The bottom end of the fins is separated from the sidewalls of the cavities by a 0.3-mm gap.  On the side, the gap between the sidewalls and the fins is 0.2 mm while the gap between the support beam and the fins is also 0.2 mm.  The fins are visible by eyes because the device must be vertically immersed in electrolyte when testing.  It will be hard to observe the flapping under a microscope this way of the device is not visible by eyes.   

          Smela reported that PPy/Au bilayer actuators 30 µm by 30 µm is capable of lifting rigid areas of 900 µm by 900 µm [5].  Since the rigid area of the each fin is 2500 µm by 1600 µm, the actuator for each fin must be several times bigger.  In order to be absolutely sure that the actuators will be able to lift the fins, two actuators, each 0.4 mm long and 0.2 mm wide are used.

          Finally, alignment marks are employed for masks 2, 3 and 4.  The alignment marks are important because there are many small features in the later steps of the fabrication sequence.  Figure 14 below shows the 600%-magnified mask 3.  Dimensions of the electrode, the micro-fins, the hinges, and the actuators are also given.

          The thickness of different layers should also be considered.  Cr should be really thin because it only serves as the adhesion layer.  Thickness under 100 Å will be sufficient enough for this application.  Au should be much thicker than Cr to cover the patterned Cr.  Again, Au serves both the passive layer in the bilayer actuator and as an electrode for electrochemical deposition of PPy.  Based on the report by Smela et al., 1000 Å is a good thickness for Au [7].  It is also reported by Smela et al. that actuation speed of PPy/Au bilayer actuators depend on the thickness of PPy thin film [7].  They find that the response time of the actuator increases as the thickness of PPy thin film increases.  Since the micro-fins cannot flap too slowly, the thickness of PPy should be relatively thin.  Smela et al. used PPy around 1 µm in thickness because the thickness is regulated by the thickness of photoresist in their fabrication process.  For the purpose of this project, 0.5 µm should be the appropriate thickness for PPy.  Finally, SU8 should be much thicker than all other materials.  It is hoped that the micro-fins should be able to break the SiO2 layer upon actuation and then flap up and down once they are free to move.  Since the thickness of the SiO2 layer is 0.5 µm, SU8 should be a couple of times thicker than SiO2.  Thickness of 4 to 5 µm for SU8 should be sufficient for this design.  Table 1 below records the thickness of each material. 
     
     

          Layer/Material Thickness (diff. Units)
          Cr < 100 Å
          Au 1000 Å
          PPy 0.5 µm
          SU8 4~5 µm

            

          Table 1: Different layers of material and their corresponding thicknesses. 
     
     

             

    Figure 14: Magnified mask 3.  Dimensions of the electrode, the hinges, the micro-fins and the actuators are given as well.  Mask 3 is used to pattern Au.  In all three of mask 2, 3 and 4, alignment marks are added because the features are too small of naked-eye alignment.

     
     

    5. Discussion

          Despite the advantages such as requiring low voltages for actuation and working at room and/or body temperature, PPy/Au bilayer actuators have some limitations, especially when applied in microrobotics.

          First, the efficiency of PPy/Au bilayer actuators is low.  Smela et al. reported 0.2% efficiency for their bilayer actuators used in moving and positioning bulk micromachined silicon plates [7].

          Next, the actuation of PPy/Au bilayer actuator is low.  Accodring to Smela et al., the fastest rate for bilayer actuators with PPy thickness of 1 µm to go from completely bent to completely straight and back was 2.5 Hz [5].  Usually, the actuation frequency for PPy/Au bilayer actuators to complete one cycle of actuation is around 1 Hz.  The actuation speed of the actuators is limited by ion transport in and out of PPy thin film.  As the result of slow actuation, the fins will flap slowly as well.  Slow flapping translates to slow swimming speed, which is undesirable.    

          Another limitation is that the device can only be actuated in liquid electrolyte.  This is because PPy thin film will only change volume in liquid electrolyte [7].  While this characteristic may be helpful in biomedical applications, it also limits the application of PPy/Au-bilayer-actuator-related devices.  For example, for investigation and maintenance of small pipes in a radiator, the swimming microrobot design presented in this report will not be applicable.

          Finally, the life-time of the actuator presents concerns as well.  Typical PPy/Au bilayer actuators fail after 1,000 to 10,000 cycles of actuation.  Failure is caused by delamination of PPy thin film from Au thin film [7].  This presents a problem for the long-term performance of the microrobot.

     

     
     

    6. References

      [1]  J. W. L. Zhou, H. Y. Chan, T. K.H. To, K. W. C. Lai, W. J. Li, “Polymer MEMS actuators for underwater micromanipulation,” IEEE/ASME Trans. Mechatronics, vol. 9, issue 2, pp. 334-342, June 2004.

      [2]  Y. Okuda, S. Guo, Y. Hasegawa, K. Asaka, “A fin type of underwater microrobot with multi DOF,” in Proc. SICE, 2003, vol. 2, pp. 1656-1660.

      [3]  M. Ataka, A. Omodaka, N. Takeshima, H. Fujita, “Fabrication and operation of polyimide bimorph actuators for a ciliary motion system,” J. Microelectromechanical Systems, vol. 2, issue 4, pp.146-150, Dec. 1993.

      [4]  E. Smela, “Conjugated polymer actuators for biomedical applications,” Adv. Mater., vol. 15, No. 6, March 17 2003.

      [5]  E. Smela, “Microfabrication of PPy microactuators and other conjugated polymer devices,” J. Micromech. Microeng., vol. 9, pp. 1-18, 1999.

      [6]  E. Smale, O. Inganäs, I. Lundström, “Controlled folding of micrometer-size structures,” Science, vol. 268, pp. 1735-1738, 23 June 1995.

      [7]  E. Smela, M. Kallenbach, J. Holdenried, "Electrochemically Driven Polypyrrole Bilayers for Moving and Positioning Bulk Micromachined Silicon Plates," J. Microelectromechanical Systems, vol. 8, issue 4, pp. 373-383, 1999. 

      [8]  T. Honda, K. I. Arai, K. Ishiyama, “Micro swimming mechanisms propelled by external magnetic fields,” IEEE Trans. Magnetics, vol. 32, issue 5, pp. 5085-5087, Sept. 1996.

      [9]  J. Edd, S. Paven, B. Rubinsky, M. L. Stoller, M. Sitti, “Biomimetic propulsion for a swimming surgical micro-robot,” in Proc. IROS, 2003, vol. 3, pp.2583-2588.

      [10]  K. Ishiyama, K. I. Arai, M. Sendoh, A. Yamazaki, “Spiral-type micro-machine for medical applications,” in Proc. MHS, 2000, pp. 65-69.

      [11]  R. J. Wood, S. Avadhanula, M. Menon, R. S. Fearing, “Microrobotics using composite materials: the micromechanical flying insect thorax,” in Proc. ICRA, 2003, vol. 2, pp. 1842-1849.

      [12]  T. Fukuda, A. Kawamoto, F. Arai, H. Matsuura, “Mechanism and swimming experiment of micro mobile robot in water,” in Proc. MEMS, 1994, pp. 273-278.

      [13]  T. Fukuda, A. Kawamoto, F. Arai, H. Matsuura, “Steering mechanism and swimming experiment of micro mobile robot in water,” in Proc. MEMS, 1995, pp. 300.

      [14]  Y. Zhang, X. Wang, T. Mei, Q. Wang, P. Zhang, “Driving principle and dynamic analysis of a micro swimming robot,” in Proc. WCICA, 2004, vol. 5, pp. 4582-4586.

      [15]  M. G. Borgen, G. N. Washington, G. L. Kinzel, “Design and evolution of a piezoelectrically actuated miniature swimming vehicle,” IEEE/ASME Trans. Mechatronics, vol. 8, issue 1, pp. 66-76, March 2003.

      [16]  G. Laurent, D. Piat, “Efficiency of swimming microrobots using ionic polymer metal composite actuators,” in Proc. IEEE-ICRA, 2001, vol. 4, pp. 3914-3919. 
       
    << High confidence secure medical devices evidence based care with automated patient specific alerts million per yearAdvanced informatics techniques including enhanced sensor systems few years creating tremendous pressure on hospitals >>