UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Feasibility of miniature polypyrrole actuated valves Cole, Matthew 2006

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

Item Metadata

Download

Media
831-ubc_2007-0063a.pdf [ 11.12MB ]
Metadata
JSON: 831-1.0101166.json
JSON-LD: 831-1.0101166-ld.json
RDF/XML (Pretty): 831-1.0101166-rdf.xml
RDF/JSON: 831-1.0101166-rdf.json
Turtle: 831-1.0101166-turtle.txt
N-Triples: 831-1.0101166-rdf-ntriples.txt
Original Record: 831-1.0101166-source.json
Full Text
831-1.0101166-fulltext.txt
Citation
831-1.0101166.ris

Full Text

FEASIBILITY OF MINIATURE POLYPYRROLE ACTUATED VALVES by MATTHEW COLE B.A.Sc, Queen's University, 2004 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n THE FACULTY OF GRADUATE STUDIES ( E l e c t r i c a l and Computer Engineering) THE UNIVERSITY OF BRITISH COLUMBIA November 2006 © Matthew Cole, 2006 11 Abstract Polypyrrole is a conducting polymer that can change in volume as its oxidation state is altered. This change in volume, operable at low voltages, can be used to create small actuating devices. The goal of this thesis is to evaluate the use of polypyrrole for the creation of a low voltage polymer valve and to demonstrate the mechanisms that could be used to create these valves. There are a number of challenges in using polypyrrole in a commercially viable valve; it must be able to withstand large temperature variations, have a high strain (to minimize mechanical amplification), have high work density (to minimize the amount of polypyrrole required), have a long lifetime and be assemblable into a compact valve. To evaluate and meet these requirements: 1) The effect of modifying synthesis and actuation conditions on the electrochemical actuation of polypyrrole is investigated to find the conditions that give the highest electrochemical strain and strongest polypyrrole films. Stable and fast strains of 6% at up to 2 MPa for films grown in propylene carbonate and actuated in NaPF6( a q) are achieved. These films were stored for up to 3 months before use with no losses in strain but showed a loss of 0.06% of their maximum strain per electrochemical cycle. 2) Polypyrrole is exposed to temperature variations, showing that high temperature (up to 80°C) exposure has a deleterious effect on polypyrrole actuation. To try and minimize losses, the effect of temperature in both aqueous and organic electrolytes and the mechanism for degradation is investigated. -PF6 grown films in aqueous electrolytes show the least degradation. 3) A linear valve mechanism is built and demonstrates that it should be possible to achieve the force and displacement required to open and close a sliding plate valve. Empirical models suggest that it should be possible to use polypyrrole sliding oil sealed valves. 4) An encapsulatable trilayer is built that seals holes in a perforated plate and could also be used to make a valve. iii Contents Abstract '. u Contents iii List of Tables ••• v List of Figures vi List of Abbreviations x Acknowledgments xi 1 Introduction 1 1.1 Conducting Polymers 1 1.2 Conducting Polymer Actuators , 3 1.3 Polypyrrole : 3 1.4 Observations and Model of Polypyrrole Actuation 4 1.5 Aims of the Work 10. 2 Polypyrrole Synthesis 12 2.1 Electrochemical Synthesis 12 2.2 Review of Current Synthesis Methods 15 2.2.1 Kaneto Group... 16 2.2.2 Yamaura Group 17 2.2.3 Smela Group 17 2.2.4 Other growth techniques 18 2.3 Summary 18 3 Actuation Results 20 3.1 Yamaura Grown Film (PPy-PF6) 21 3.1.1 Electromechanical actuation in 0.1 M TBAP in PC 21 3.1.2 Electromechanical actuation in NaCl(aq) 25 3.1.3 Electromechanical actuation Sodium Acetate(NaCH3COO)(aq) 26 3.1.4 Electromechanical actuation in NaPF6(aq) 28 3.1.5 Electromechanical Actuation in BMM-BF4 29 3.1.6 Electromechanical Actuation in Cyphos IL 106 31 3.2 Kaneto Grown Films 33 IV 3.3 Phosphonium Ionic Liquid Grown Films 35 3.3.1 Electrochemical actuation in Phosphonium IL 36 3.3.2 Electrochemical actuation in NaPF6(aq) 38 3.4 Summary of Growth Methods 39 4 Temperature effects on Polypyrrole 41 4.1 Initial Strain Effects 42 4.2 Effect of temperature Exposure on Actuation 47 4.2.1 Change in maximum actuation 48 4.2.2 Changes in Rate of Actuation 50 4.2.3 Temperature Induced Change in Conductivity and Length 50 4.2.4 Cause of drop in maximum actuator strain and strain rate 51 4.3 Electrochemical Actuation vs Temperature 52 4.4 Young's Modulus vs Temperature 58 4.5 Engineering Implications 60 4.6 Conclusions 61 5 Polypyrrole Devices'.... 62 5.1 Linear Actuation 62 5.2 Bilayers 65 5.2.1 Polypyrrole/Mylar Bilayer 67 5.3 Trilayers 68 5.4 Polypyrrole/Millipore Separator Trilayer 74 5.5 Polypyrrole Lifting Valve 76 5.5.1 Valve Mechanical Analysis 78 5.5.2 Valve Control 80 5.6 Polypyrrole Sliding Valve 84 5.7 Conclusions 89 6 Conclusion and Future Work 90 6.1 Summary of Results 90 6.2 Major contributions of this research '. 91 6.3 Future Work 91 Bibliography 93 V List of Tables Table 3-1 : Summary of different film growth methods and actuation results 39 Table 4-1: Growth and Actuation Conditions of Polypyrrole films 47 Table 4-2: Conductivity, Thermal Strain, Electrochemical Strain, Max Strain Rate 49 Table 4-3: Drop in Charge Transferred from exposure to heat 58 VI List of Figures Figure 1-1: The chemical structures of some common conducting polymers 1 Figure 1-2: Example of doping a conjugated polymer. In this case a positive potential is applied to the polymer, removing electrons, and negative ions move into the polymer in order to balance charge. This is not necessarily always the case as depending on the size of ions, either negative and/or positive ions can move in and out of the polymer to balance charge 2 Figure 1-3: Switching between the contracted (oxidized) and expanded (reduced) forms of polypyrrole. 4 Figure 1-4: A three electrode electrochemical cell. WE, CE, and RE are the working, counter and reference electrodes respectively. In the case of actuating polypyrrole, typically polypyrrole would be the W E while the C E would be a high surface area electrode. R s is the solution resistance, while R u is the uncompensated resistance 6 Figure 1-5: a) Finite element model for polypyrrole in solution (A more accurate model would be continuous). Rppy and Rppy' represent the electronic resistance of polypyrrole along segments of length and thickness respectively. R i represents segments of the parallel portions of the ionic resistance within the polypyrrole. Cv represents the volumetric capacitance of a volume of the film due to double layer charging of ions in the electrolyte. Ru represents parallel portions of the total ionic resistance outside of the film, due to the reference electrodes distance from the working electrode. Rc represents the contact resistance b) Simplified circuit model for a thin film. Rs is the combination of uncompensated and contact resistance (Ru and Re). Ri is the combined resistance of segments in series from part a) 7 Figure 1-6: Actuation in PPy-BF^ In the step down from 0.7 V to -0.7 V the current spikes to -37 mA and the initial strain rate is high. In the step up to 0.7 V, there is no current spike and the strain rate is much lower 10 Figure 2-1: Mechanism of Polypyrrole synthesis. As proposed by Baker and Reynolds [25] 14 Figure 3-1: Strain, Voltage and Current for a PPy-PF6 film actuated in 0.1 M T B A P ( P C ) . A was C V applied at 3.6 mV/s 22 Figure 3-2: Actuation of PPy-PF6 in 0.1M TBAP. 1.2 V to -0.1 V applied vs Ag/AgCl at 0.001 Hz 24 Figure 3-3: Actuation in 1.0M NaCl, cycled between -0.9 V and 0.7 V vs Ag/AgCl at 0.017 Hz 25 Figure 3-4: Actuation in NaCH3C00, -0.9 V to 0.7 V vs Ag/AgCl, 0.017 Hz 27 Figure 3-5: Actuation in 1.0M NaPF6, 0.7 V to -0.7 V vs Ag/AgCl, 0.005 Hz. 28 Figure 3-6: Actuation in B M I B F 4 , 0.8 V to -0.5 V vs Ag/AgCl, Scan rate of 5 mV/s followed by 2.5 mV/s 30 Figure 3-7: Actuation in [BMIM]BF4, 0.8 V to -0.5 V at 0.005 Hz 31 Figure 3-8: Actuation in Cyphos IL 106, 0.05 Hz, +8 V to -8 V vs Counter electrode. 32 Vll Figure 3-9: PPy-BF4 film actuation in 1.0 M NaPF6(aq), 0.7 V to -0.9 V vs Ag/AgCl, 0.002 Hz : 34 Figure 3-10: Triisobutyl(methyl)phosphonium tosylate 36 Figure 3-11: Actuation of Cyphos IL 106 grown film. +15 to -15 vs CE at 0.05Hz 37 Figure 3-12: Actuation of Cyphos IL 106 grown film. 0.7 V to -0.9 V vs Ag/AgCl at 0.05Hz in NaPF 6 ( a q ) 38 Figure 4-1: Initial strain of a PPY-PF6 film exposed to 0.1 M TBAP in PC. The film initially swells and then gradually contracts over the next 12 hours. The propylene carbonate solution is then heated to 80°C and the polymer undergoes a fast contraction of 3 %. When the solution is returned to 20°C the polymer contracts by 0.75 % 43 Figure 4-2: Strain and Temperature for a pre-soaked PPy-PF6 film in 0.1 M TBAP(PC) 44 Figure 4-3: Strain and Temperature for a pre-annealed PPy-PF6 film in 0.1M TBAP(pc) 46 Figure 4-4: Typical temperature and cycling test of a PPy-BF4 film actuated in NaPFe(aq). On the first heating cycle there was a small expansion and on the second heating cycle a small contraction. Both annealing or soaking can reduce these effects 48 Figure 4-5: Strain and Strain to Charge of films exposed to 80°C thermal cycles 50 Figure 4-6: Actuation and Charge transferred during actuation vs. temperature 53 Figure 4-7: Amplitude of actuation of a PPy-PF6 film in 1.0 M NaCl(aq). 0.4 V to -0.4 V vs Ag/AgCl at 0.05Hz 54 Figure 4-8: Actuation of a PPy-PF6 film in 1.0 M NaCl(aq), -0.2 V to 0.2 V vs Ag/AgCl at 0.05Hz 55 Figure 4-9: Actuation of a PPy-PF6 film in 1.0 M NaPF 6 ( a q ), -0.2 V to 0.2 V vs Ag/AgCl at 0.05Hz 56 Figure 4-10: Electrochemical cycling of a PPy-PF6 film from 0.2 V to -0.2 V vs Ag/AgCl in NaCl(aq). The amplitude of strain, charge transferred in and out of the film per cycle and temperature are shown 57 Figure 4-11: Young's Modulus of a PPy-PF6 films vs temperature 59 Figure 5-1: Zig-Zag metal wire actuator. The bent gold wire confines the direction of actuator to up and down. When the polypyrrole expands or contracts it expands or contracts the scaffold. The gold also improves electrical contact to the polypyrrole and reduces voltage drops through the polymer : 63 Figure 5-2: A linear polypyrrole film that was fabricated pulling up a 5 g weight. Left: A zoomed out view of the film and weigh. Top Right: A picture of the weight when the polypyrrole is in the contracted state. Bottom Right: A picture of the weight when the polypyrrole is in the expanded state 63 Figure 5-3: Young's modulus as a function of oxidation state. The 0.15 Young's modulus should be used for high frequency (>1 Hz) actuation calculations while the 5 seconds Young's modulus used for slower actuation Figure 5-4: Operation of a bilayer. a) In the neutral state there is no stress and the bilayer is straight, b) Redox removes ions from the polymer causing it to contract and the bilayer to curve towards the polymer, c) Redox inserts 64 Vlll ions into the polymer causing the bilayer to bend away from the polymer. Directions of arrows outside the bilayer show the direction of ion flow while arrows inside the bilayer show the contraction or expansion of layers 66 Figure 5-5: Actuation sequence of a polypyrrole/mylar tape bilayer in NaPF6(aq). The frames are spaced at 2 second intervals. A 0.05Hz 2.8 peak-peak V square wave was applied to the device. The black film behind the bilayer is a piece of polypyrrole acting as the counter electrode 67 Figure 5-6: Operation of a trilayer. a) No application of voltage, so no force and the trilayer is straight b) Application of a voltage, redox of the polymer layers cause ions to flow into one layer and out the other. The resulting swelling and contraction of the films causes a stress gradient which bends the trilayer. c) Reverse of step b causes the trilayer to bend in the other direction 69 Figure 5-7: Dimensionless curvature with respect to (^hp/hg) The curvature is made dimentionless by multiplying by hg and dividing by a and p 71 Figure 5-8: Dimensionless plot showing the maximum trilayer curvature and ratio of layer thicknesses for that curvature as a function of the relative elastic moduli of the layers 71 Figure 5-9: Dimensionless plot showing trilayer force against the ratio of layer's Young's moduli 73 Figure 5-10: Geometry of a trilayer: F is the force applied to the end of the polymer. The polymer layers have a thickness hp and the separator layer has a thickness hg. The total length of the trilayer is L and the width is W 74 Figure 5-11: Polypyrrole Trilayer: Composed of a 110 jum inner layer of Millipore PVDF separator. Next are two 70nm thin layers of platinum. The outside is 30 fxm thick layers of PPy-PF6 74 Figure 5-12: Actuation of a PPy/Millipore trilayer in NaPF6(aq).. The voltage was stepped between 1.4 V and -1.4 V. There is Is between each frame 75 Figure 5-13: Operation of a trilayer valve. Left: Closed position, no voltage is applied to the valve and it is straight. Right: Open position, voltage is applied to the valve and it curls to unseal the hole below.. 76 Figure 5-14: Lifting valve in operation. This shows one trilayer lifting and sealing over holes. For a complete device there would need to a number of trilayers cover the whole series of holes. The trilayer shown is much longer than necessary given the size of the holes 77 Figure 5-15: Pictures from wide and short trilayer lifting valve, with the top figure showing the entire bottom edge lifting and the bottom figure showing it closing. There is a thin layer of oil on the plate 77 Figure 5-16: Forces acting on the trilayer as it separates from the surface, r is the radius of curvature of the meniscus. L is the length of the trilayer. F l a y e r is the force exerted by the trilayer at the tip. Fm e ninsus is the force due to the liquid 78 Figure 5-17: Strain vs Voltage. The voltage was scanned at 0.67mV/s against an equal size polypyrrole counter electrode. The hysteresis shown in this picture is due to the scan rate, and disappears as the scan rate is slowed.... 80 ix Figure 5-18: BJT voltage Scaling circuit along with simulation results. This circuit was built and worked to operate a trilayer and sliding valve 81 Figure 5-19: MOSFET circuit and simulation results 82 Figure 5-20: Actuation of trilayer coated in silicone conformal coating 84 Figure 5-21: Sliding valve mechanism 85 Figure 5-22: Sliding valve: Left - Valve in open position. Right - Valve in closed position 85 Figure 5-23: Volume of polypyrrole required vs the cross sectional area of the polypyrrole. Friction is IN, and displacement is 1.5mm 88 List of Abbreviations aq aqueous BMIM l-Butyl-3-methyl-imidazolium CE Counter Electrode Cv Capacitance per unit volume CV Cyclic Voltammogram Cyphos IL 101 tetradecyl(trihexyl)phosphonium chloride Cyphos IL 106 triisobutyl (methyl)phosphonium tosylate DBS Dodecylbenzenesulfonic E Young's modulus IIL Imadazolium Based Ionic Liquid MB Methyl Benzoate PC Propylene Carbonate PIL Phophonium Based Ionic Liquid PPy Polypyrrole PVDF Polyvinylidene Fluoride RE Reference Electrode TBAP Tetrabutylammonium Hexafluorophosphate T-EAP Tetraethyleammonium Hexafluorophosphate TFSI Bis(trifluoromethanesulfonyl)imide WE Working Electrode XI Acknowledgments It is interesting to think back on when I started my Masters. The field of conducting polymer actuators seemed so daunting and yet I had no idea of the complexities or complications that it entailed. Well it was a journey, one that was helped along by a great group of people. First, John, best supervisor ever, enough said. And then, in no particular order. Thank you Ali, not only did you make the lab a livelier place and help me with numerous projects, but somehow you seemed to know where everything in the lab was, where to buy anything that I needed, who to ask about anything I needed know. Dave, your help with all sorts of polypyrrole growths along with your enthusiasm was amazing. Navid, for your help with ionic liquids and for the effort dealing with low polypyrrole generated voltages. Jagjit, for the work on ionic conductivity. Robin and Dan, for your help with chemistry and letting me use the fumehood. Mya, you introduced me to the joys of electrodeposition, and your amusing frustrations inspired me, if not to work harder, at least to help me accept that things don't always work perfectly. Tissa, thanks for keeping everything running. I'm sorry I asked you to go to lunch every day that you were fasting. Eddy, always relaxed, you could be counted on to help derive solutions to any mathematical problems. Tina, you were a great neighbour and your parallel work with actuators was inspiring. Arash, you were a source of boundless knowledge and help. Alex, thanks for the conversations, the coffee invitations, and the inspirations. Raha, thanks for the evaporations in the cleanroom. And lastly Victoria, whose love and support were as important as all the technical help combined. 1 1 Introduction This chapter gives an introduction to conducting polymers, focusing on polypyrrole, and their use as actuators. A model of both polypyrrole actuation and charging is presented. Finally, the aims of the work in this thesis are explained. 1.1 Conducting Polymers Conducting polymers are conjugated polymers, which due to a degree of charge derealization exhibit conductivity. Examples of some common conducting polymers include polypyrrole, polyacetylene, polyaniline and polythiophene (see Figure 1-2). In the neutral state, conjugated polymers generally exhibit very low conductivities and act as semiconductors. However, in the highly doped state their conductivity is increased dramatically and can be as high as 105S/cm. Figure 1-1: The chemical structures of some common conducting polymers. 2 Unlike in silicon, where doping is accomplished by substituting electron donor or acceptor atoms in a crystal lattice, in conducting polymers, doping is a reduction-oxidation (redox) process. The redox can be accomplished either chemically, by exposing a polymer to a reductant or oxidant, or often electrochemically by applying a potential to the polymer. Figure 1 -2 shows the process by which a polymer can be electrochemically doped. O© © © 0 © ( = ) @ © e © ^ ° © e Conjugated Polymer, Undoped Low Conductivity Conjugated"Conduct ive" Polymer, Doped High Conductivity Figure 1-2: Example of doping a conjugated polymer. In this case a positive potential is applied to the polymer, removing electrons, and negative ions move into the polymer in order to balance charge. This is not necessarily always the case as depending on the size of ions, either negative and/or positive ions can move in and out of the polymer to balance charge. Another key difference between doping in conducting polymers and silicon is the level of doping. Highly doped silicon might have one dopant atom per thousand silicon atoms while conducting polymers have doping levels up to one doping ion for every three polymer monomer. This means that conducting polymer can have enormous charge capacity (lOOF/g) [1]. 3 1.2 Conducting Polymer Actuators The high degree of doping of conducting polymers^  combined with their porous and flexible structure, allows them to work as actuators. As shown in Figure 1-2, when a conducting polymer is doped ions move in and out of its structure. As well, since ions often have hydration spheres (closely attracted solvent molecules), they can bring solvent molecules into the polymer with them. This flux of ions causes a net volume change which can be used in a variety of different configurations such as bilayer, linear and helix tubes to perform work [2-5] . Examples of such configurations can be seen in Chapter 5. Note that one might expect that as the polymer is reduced or oxidized the number of anions moving out of the film would balance with the number of cations moving into the film and there might be no net actuation. However, polymer actuators are generally grown or actuated so that one of the ions is far more mobile within the polymer. Often electrolytes are used where one ion is far larger that the other or where one ion has a much larger solvation sphere so that movement for that ion within the polymer is very slow or impossible. 1.3 Polypyrrole The most studied conducting polymers for actuation is polypyrrole due to its chemical stability, high conductivity and ability to undergo large volume changes at high stresses. Figure 1 -3 shows a chain of polypyrrole switched between the contracted and expanded states. Polypyrrole actuators have been reported to work at stresses of up to 34 MPa, strains of up to 39% and strain rates up to 13.8%/s [6-8]. 4 +2 e- -2 contracted state A- 1^ -2G- + 2A-expanded state Figure 1-3: Switching between the contracted (oxidized) and expanded (reduced) forms of polypyrrole. However, these results were not obtained with the same actuators. There seems to be a trade off between the stress and strain of polymer actuators, with polypyrrole typically actuated at <5% strain and <2.5 MPa. Lifetime is also an issue although recent work has shown that polypyrrole can be actuated in ionic liquids for over 105 cycles and polyaniline can be cycled in ionic liquids for over 106 cycles without degradation [8-10]. As well, as stress is increased polypyrrole tends to undergo viscoelastic creep. As a comparison natural muscle can undergo strains of 20-30% at strain rates up 500%/s for over 109 cycles (with regeneration) [11]. However, typical stresses that muscle can operate under are only 0.1 to 0.35 MPa, which is far lower than that of polypyrrole actuators. 1.4 Observations and Model of Polypyrrole Actuation As stated, actuation in polypyrrole is primarily a function of volume change due to ion flux caused by reduction or oxidation. Actuation can also be affected by changes due to redox such as changes in Young's modulus and viscoelastic creep. However these effects 5 can be minimized by keeping stress on the polymer low [12]. For this reason most tests in this work were performed at 0.1 MPa. Redox is usually performed electrochemically rather than chemically since it is generally easier to apply a potential to the film than to change its chemical environment. A relationship between stress, a, strain, e, and charge per unit volume, p describes the behaviour of polypyrrole to the first order (it is accurate at low charge densities and stresses) [13-15]. s ( t ) , a . p ( « ) + Hia 0) E In this equation E represents the Young's modulus of polypyrrole in solution while a represents the strain to charge ratio for polypyrrole. These can vary depending on both growth and actuation conditions. Due to its porosity, polypyrrole acts as a large double-layer capacitor (a capacitor created when voltages are applied in solution and ions move very close to the charged surfaces) storing charge in its volume with a capacitance per unit volume of Cy- Since p = C v V the strain may be expressed as: E ( t ) = a . c v . v + ^ (2> E where V is voltage. The voltage range that this applies for depends both on the type of the polypyrrole and actuation conditions. A typical configuration for actuation of polypyrrole is shown in figure 1-3. Polypyrrole acts as the working electrode (WE) and the voltage is applied using a counter electrode (CE) but measured and controlled against a reference electrode (RE). Using a RE has two advantages. It minimizes the effect of ionic resistance which is reduced to Ru. It also eliminates the effect of any voltage drop across the double layer of the CE which can vary in magnitude depending on the composition and size of the CE. 6 j p l l l WE RE Electrolyte R I ' V Y W W V I \1 R —7 CE s Figure 1-4: A three electrode electrochemical cell. WE, CE, and RE are the working, counter and reference electrodes respectively. In the case of actuating polypyrrole, typically polypyrrole would be the WE while the CE would be a high surface area electrode. Rs is the solution resistance, while Ru is the uncompensated resistance. The configuration shown in Figure 1-4 can be represented using the circuit model in Figure 1-5. 7 a) Rppy <. Cv Rppy' Ri A A -Cv Ri A V -Cv .Rppy > Cv: R P W Rpw' Ri Ri r W V — ? - v W c I Ri v V A Cv Rppy' Ri Ru - A V -: CV Ru V A / -Rppy' o—VVV r/VV-•v - ' - Cv Cv X . . . . J v w -Rppy' Rppy' Rppy' Rppy < Rppy Ri pAAA-< Cv d= Cv = Rppy' Rppy Ri Ri V v V -YVA—X\< Ri AAA-Ru Rppy' Ri AAA-f-AAA-Cv =±= Cv: Rppy1 Rppy" Ri Ru V A ^ - t - A A A - J W A Rppy' : CV V A A —Ki v W — 1 Rppy' Rppy1 b) Rppy Rppy Rppy £ Rppy > Figure 1-5: a) Finite element model for polypyrrole in solution (A more accurate model would be continuous). Rppy and Rppy' represent the electronic resistance of polypyrrole along segments of length and thickness respectively. Ri represents segments of the parallel portions of the ionic resistance within the polypyrrole. Cv represents the volumetric capacitance of a volume of the film due to double layer charging of ions in the electrolyte. Ru represents parallel portions of the total ionic resistance outside of the film, due to the reference electrodes distance from the working electrode. Rc represents the contact resistance b) Simplified circuit model for a thin film. Rs is the combination of uncompensated and contact resistance (Ru and Re). Ri is the combined resistance of segments in series from part a). In this circuit model a strip of polypyrrole is connected at one end so there is an electronic resistance (Rppy) along each segment of the film. In parallel with Rppy along the length of the film is Rppy' the electronic resistance through the film. However, because films are generally extremely thin, on the order of 10-20 /xm, compared to a length on the order of millimetres, this resistance can be neglected in most actuation experiments. Cv represents the volumetric capacitance per unit volume throughout the 8 film. This volumetric capacitance is due to double layer charging, which occurs anytime potentials are applied to a conductive substrate in solution. In Faradaic reactions the double layer builds sufficiently for charge transfer to occur, however in polypyrrole actuation the potential is not generally large enough for this. Ri is the ionic resistance within the film itself and finally Ru represents the uncompensated resistance within the bulk solution. The model in Figure l-5a acts as three interconnected transmission lines (the number of circuit elements depends on the accuracy desired). While for small signal analysis it can be simplified, for large voltages changes it is complicated by the fact that as polypyrrole is oxidized or reduced both its electronic and ionic resistance change. Therefore, in order to simplify analysis the smaller resistance factors are usually ignored. Warren has performed an in-depth analysis of the different impedances depending on difference configurations [16]! In the case of experiments performed in later chapters, the polypyrrole strip is connected at one end and films are very thin so electronic resistance through the film is small compared to electronic resistance along the film. Therefore, the equivalent circuit can be represented by Figure l-5b. Further approximation that can be made include assuming that the electronic resistance of the films is small, and thus combining the electronic resistances in series. This allows us to combine that the parallel ionic resistances and volumetric capacitance to get a simple RC circuit. The change in oxidation state of the film along with the current response can then be determined by simulation. The parameters and implementation of this model can be viewed in Warren [17] but the implications that are most relevant for the work presented in this thesis are: 1) Due to electronic resistance through the length of the film, rate of actuation can be highly dependant on length. 2) Ionic resistance through the film is in series with electronic resistance (Figure 1 -5b) and thus changes in electrolyte, which change ionic resistance, can have a large effect on rate. Also, as films become thinner, ionic resistance is minimized and rates go up. 9 3) Due to changes in electronic and ionic resistance, rates of actuation can be different when oxidizing and reducing films. Changes in resistance can cause current profiles that do not look like those based on typical RC transmission line models, (see Figure 1-6). Figure 1 -6 shows a typical polypyrrole square wave actuation in aqueous sodium hexafluorophosphate (NaPF6). The voltage is stepped from -0.7 V to 0.7 V vs. a Ag/AgCl reference electrode. As the voltage is stepped down from 0.7 V to -0.7 V the current spikes up to -37 mA and decays as one would expect from a RC transmission line. However, when then voltage is stepped up from -0.7 V to 0.7 V the voltage gradually rises, levels off, and then finally begins to decay. The reason for this second response is that when the positive step is applied the film began in the reduced state where electronic resistance is very high. Since the resistance is high there is a large voltage drop across the length of the film and only the end of the film near the electrical connection experiences the full voltage change. Thus, most of the current comes from the polypyrrole near the electrical connection. Gradually as the film is oxidized, conductivity goes up and more of the film becomes active so current rises. Eventually most of the film is active and the current decays as expected in an RC circuit. Additionally, since the current is higher when reducing the film the initial rate of reduction is also much higher than that of oxidation. These effects are modelled fairly well by Warren, but depends on accurate measurements of electronic and ions resistance which depends on individual films conditions. 10 P P y - B F . film in 1.0M N a P F P 500 500 200 300 400 500 Tirne(s) Figure 1-6: Actuation in PPy-BF4: In the step down from 0.7 V to -0.7 V the current spikes to -37 mA and the initial strain rate is high. In the step up to 0.7 V, there is no current spike and the strain rate is much lower. 1.5 A i m s of the W o r k The goal of the work in this thesis is to develop a small, low power, low cost valve. The valve needs to run off battery level voltages and currents, have a lifetime on the order of years, have a bandwidth on the order of 0.1Hz, and be able to endure the environmental conditions, such as temperature and humidity of a typical consumer device. Polypyrrole was chosen for the valve primarily due its low cost and low voltage of operation. Other actuator technologies were considered, however they all failed to meet the requirements for the valve. Piezoelectrics have very high efficiencies and powers, however they require high voltages (-100 V) to operate and have low strains, typically -0.1%. Thus, they would need costly DC/DC converters and huge mechanical 11 magnifications of their strain. Electrically actuated SMAs have high strains and stresses, however they are difficult to control due to the high current they require and they have relatively short lifetimes. Dielectric elastomers, relaxor ferroelectric polymers and liquid crystal elastomers require very high voltages (>lkV). Molecular actuators could possibly meet the requirements but they are still too early in development as they have not yet been developed on a macroscopic scale. Carbon nanotubes can work under very high stresses, but strains are fairly low (<0.6%) and they are still very expensive. EPMCs can't be completely excluded but they are still somewhat expensive if using gold or platinum electrodes, consume quite a bit of power when holding a fixed position, and don't work in a linear mode, so conduction polymers seem more promising. Ferromagnetic SMA's require unfeasible bulky magnets [18]. While polypyrrole does show a lot of promise, it is still a fairly new actuator technology and needs to be optimized and characterized for the valve conditions. There are many different ways to both synthesize and operate polypyrrole that can affect its capabilities. Chapter 2 gives some background on polypyrrole prepared using different methods of synthesis. Chapter 3 describes actuation results using different types of polypyrrole. Little is known of polypyrrole's temperature response, which is addressed in chapter 4. In Chapter 5, two valve mechanisms are demonstrated and the mechanical concerns in actually building a valve are addressed. Polypyrrole needs to be encapsulated both to protect it from the environment as well as prevent it from drying out. There are many different ways for mechanical valve designs to magnify polypyrrole strain. As well, even though polypyrrole can work at very low voltage, the strain and stress it can provide scale by voltage so very low voltage changes, such as provided by a battery, may require scaling to improve performance. Taken together, the work reported in this thesis suggests that polypyrrole driven valves are feasible, with advantages over alternative technologies, however there are some limitations. 12 2 Polypyrrole Synthesis This chapter reviews current methods to synthesize polypyrrole. It does not provide an overview of all polypyrrole synthesis techniques as it focuses on those techniques which are used to produce actuating films. However, it gives a starting point for the investigation of the actuation of different types of polypyrrole, and provides as basis for the experiments on the different types of polypyrrole actuated in chapter 3. The synthesis of polypyrrole involves the oxidation of pyrrole monomers, which form oligomers in solution, nucleate and precipitate out of solution. Polypyrrole generally precipitates onto an electrode, forming a solid film. However, depending on the synthesis conditions it may form a particulate in solution. Figure 2-1 shows a more detailed explanation of the synthesis as proposed by Baker and Reynolds [19]. There are two common methods of polypyrrole synthesis; chemical and electrochemical. Spin casting or dip coating methods commonly used to form polymer films are not usually used to produce polypyrrole due to its poor solubility. Some work has been performed to produce soluble polypyrrole but no useful devices have been reported from it [20-22]. While chemical synthesis of polypyrrole does not require a conductive substrate and can produce large quantities of polypyrrole quickly, from our experience it results in films with relatively low conductivity as well as poor mechanical properties. For this reason, only electrochemical synthesis methods of polypyrrole will be discussed. 2.1 Electrochemical Synthesis Polypyrrole is electrochemically synthesized by researchers in many different ways according to desired properties or often simply based on previous experience. 13 Typically electrochemical synthesis of polypyrrole involves combining pyrrole monomers and a salt (the dopant) into an electrolyte. Alternatively ionic liquids which act as both electrolyte and dopant can be used. Two electrodes are placed in the electrolyte and a voltage is applied between them, with polypyrrole deposition occurring on the electrode at a positive potential. This should not imply that any combination of solvent, salt and electrode produce polypyrrole that can make a useful actuator. Many factors affect the properties of polypyrrole such as: a) Deposition Electrode: The electrode must be made of a relatively inert material so that it does not corrode at a lower potential than polypyrrole oxidation. However, depending on the solvent or dopant, passivation layers can form which allow growth even on active metals such as untreated iron [23]. Secondly, polypyrrole adhesion is affected by the substrate material and surface morphology. If adhesion is not strong enough, the polypyrrole tends to separate from the substrate as it grows forming a bubbly film. Adhesion can often be improved by roughening the surface, however it is a tradeoff as a smoother surface result in a more homogenous conductive film. Common electrode materials are glassy carbon, platinum, gold and titanium. b) Solvent: For synthesis to occur pyrrole must necessarily be soluble in the solvent. Some solvents create films that are more conductive and mechanically strong while others create more porous larger straining films. The reason some solvents are better than others is not known and the choice of solvent is generally a trial and error process. There are theories on solvents' effects on interactions between PPy chains, however this comes from observation rather than modelling [24]. 14 Synthesis of Polypyrrole The last step is in fact a repetition of the first steps beginning with oxidation, followed by coupling to either end of the polymer, and finally elimination of H + . The electrons are either removed via an electrode (electrochemical deposition) or chemically, e.g. F e + + + £ * • F e + + Note that the polymerization does not generally result in a neutral polymer shown above, but rather the backbone is charged, as below, such that the total number of electrons transferred per monomer is 2+a where a is generally between 0.2 and 0.5: where A" is an anion or dopant. Here a=l/3. During the initial phases of electrodeposition the oligomers remain in solution, eventually precipitating to form a solid with intercalated anions. Figure 2-1: Mechanism of Polypyrrole synthesis. As proposed by Baker and Reynolds [25] 15 c) Salt: Electrochemical synthesis of polypyrrole generally involves using a dissolved salt as a dopant. However, synthesis using liquid salts where the solvent itself is the dopant have been performed by Pringle and Fenelon [26;27] d) Temperature: Generally, lower temperature synthesis results in more conductive and mechanically tough films. The temperature is restricted by the freezing temperature of the electrolyte. e) Current Density: Polypyrrole growth rate is directly proportional to current density so that the higher the current density the faster the film grows. However, higher current densities usually produce less ordered, less conductive and weaker films. Also a high deposition voltage is required for high current densities which increases the likelihood of deleterious chemical side reactions on either the working or the counter electrode. Too low a rate of deposition results in short chains that do not adhere to the substrate. Other factors that can affect on polypyrrole synthesis are the concentrations of dopants and pyrrole, purities of chemicals and pH of electrolytes. 2.2 Review of Current Synthesis Methods Due to the the different properties that result from variations in synthesis conditions, a large number of methods of synthesizing polypyrrole have been developed and are continuing to be developed. The following section will cover some of the most used and/or promising of these methods. The methods are sorted by the group that has developed them or currently does the most work with them. 16 2.2.1 Kaneto Group Recent work by Kaneto has shown dramatic advances in rate and magnitude of electrochemical actuation of polypyrrole [7;28]. Kaneto synthesizes two different groups of films. The first are based on doping with relatively small ions such as tetrafluoroborate (BF4) or trifluoromethanesulfonate(CF3S03). Growth is performed at 0.2 mA/cm2 in aromatic esters solutions such as methyl benzoate and on substrates such as Ti. Doping salts such as tetraethyl ammonium tetrafluoroborate(TBABF4) are used where the TBA + cation is much larger and less mobile than the anion. These film exhibit large strains(12-15%) and stresses (20-22 MPa) [24]. These types of fdms will be referred to as PPy-BF4 or PPy-CF 3S0 3. The second class of fdms that Kaneto synthesizes are based on doping with large bis(trifluoromethanesulfonyl)imide based ions(CF2n+iS02)2N" (n = 1..4). These fdms are grown in aromatic esters with salts such at TBATFSI. They show very large initial strains of 23% to 40% corresponding to the size of the doping ion. These films have shown extremely fast strain rates (up to 13.8% per second for a free standing film). This is attributed to high porosity due to swelling in organic solvents. However, these films have low tensile strengths (<30 MPa) and show a decrease in strain over multiple cycles, so for practical use they need to be used with a supportive structure and at lower than maximum strains[7]. These films will be referred to as PPy-TFSI. PPy-BF4, PPy-CF3SC>3 and PPy-TFSI have been used to form a variety of devices such as coil composite actuators, zigzag metal wire actuators, trilayers and linear devices [29-31]. As of November 2006 videos of many of these devices are available at http://www.eamex.co.jp/index e.html. 17 2.2.2 Yamaura Group A very common method for synthesizing polypyrrole was developed by the Yamaura group [32]. It is based on doping films with hexafluorophosphate (PF6~) ions. Growth is performed in propylene carbonate (PC) with 0.05M tetraethyl ammonium hexafluorophosphate(TEAP) as a dopant, 0.06 M distilled pyrrole and 1% water by volume. Tetrabutylammonium hexaflurophosphate can be substituted for TEAP since it results in films with similar mechanical and electrical properties and is less expensive. Deposition is carried out galvanostatically at 0.125mA/cm on polished amorphous carbon electrodes. This method is currently used by researchers at Wollongong, MIT and UBC [10;33;34]. Deposition generally requires a voltage of between 2.5 V and 3 V. The voltage increases as the distance from the WE to CE is increased as well as when the size of the CE is reduced. On a large glassy carbon electrode, over long periods of time, films grow at approximately 1.6 jtim/h and have densities of approximately 1.5 g/cm3. This type of polypyrrole is highly conductive (as high as 450 S/cm), and very tough (has been actuated at up to 34 MPa [35]. Films grown using the Yamaura technique will be referred to as PPy-PF6 films. 2.2.3 Smela Group Smela's group synthesizes polypyrrole films doped with the dodecylbenzenesulfonic (DBS) ion (this type of polypyrrole will be referred to as PPy-DBS). Synthesis is performed in an aqueous solution of NaDBS on gold substrates at current densities of 2 mA/cm2 [36]. This is approximately 10 times higher than Yamaura or Kaneto depositions. Unlike films grown using Kaneto or Yamaura techniques, where growth is performed with doping cations that are much larger than doping anions, in PPy-DBS growth the anion, DBS", is much larger than the cation, Na+. For this reason the strain to charge direction in PPy-DBS actuation is reversed from typical Kaneto and Yamaura polypyrrole actuation. 18 Experiments on PPy-DBS films are generally carried out by Smela on bilayer type microelectromechanical (MEMS) devices, so peak stresses and strains are difficult to calculate. However based on models of curvature, films can show between 3-20% strain depending on thickness with is usually less than 3/an [37;38]. Conductivity of freestanding films is 70±10S/cm. A variety of micro actuators have been fabricated using PPy-DBS such as movable silicon hinges and micro valves [3;39]. 2.2.4 Other growth techniques Numerous other electrochemical growth techniques have been used by other researchers. Pringle synthesized polypyrrole in immidazolium based ionic liquids [40]. Pei and Inganas have grown polypyrrole in aqueous electrolytes and doped with a number of different anions such as nitrate and methysulfonate. [41]. Otero has synthesized polypyrrole in acetonitrile with LiC104 [42]. Delia Santa has synthesized films doped with benzensulfonate [43]. Kassim et al has synthesized films in an aqueous medium using camphor sulfonate as the dopant [44]. Hutchison synthesized polypyrrole in both 4-toluenesulfonic acid sodium salt (pTs) and sodium perchlorate [34]. The techniques mentioned for synthesizing polypyrrole are aimed at growing polypyrrole with good actuation properties (such a high mechanical strength, conductivity and strain). However, polypyrrole is also often electrochemically synthesized for other purposes. There has also been a huge amount of work growing polypyrrole films on metals aimed at preventing corrosion [45-52]. In these cases adhesion and stability are much more important properties. Also looked at is the use of polypyrrole in polymer transistors in which case high mobility and good electrical contact are critical [53]. 2.3 Summary As shown in this chapter there is not one specific type of polypyrrole. There is a huge variation in synthesis conditions that have an enormous effect on film properties. This is 19 advantageous in the fact that the properties of polypyrrole can be tailored for different uses but disadvantageous in that it makes comparisons between different films difficult and makes characterization of polypyrrole hard to generalize. Additionally, even small changes in synthesis conditions, such as the orientation of the electrodes or temperature gradients in the electrolyte can affect films. Therefore, there can even be differences between polypyrrole taken from different sections of an electrode. This chapter provides a good summary of currently popular synthesis techniques, however these are constantly changing as improvements are made. It gives a starting point for investigating the actuation of different types of polypyrrole. The following chapter addresses the actuation of polypyrrole films under many different conditions. 20 3 Actuation Results The following chapter covers actuation experiments on polypyrrole films. The goal of these experiments was to obtain a fast, high strain film. Polypyrrole films grown using the Yamaura and Kanato techniques were actuated under different conditions since they had previously demonstrated high strain. The same actuation measurements were not necessarily made under all the different conditions as these experiments were done over a long period of time and some conditions showed more promise than others. A more comprehensive study of all the different polypyrrole films would measure properties such as viscoelastic creep response, temperature response, long term electrochemical actuation at different rates, Young's modulus and strain to charge ratios. The primary property compared for all films was maximum electromechanical strain (actuation) under low load conditions (~0.1 MPa). This was performed over a set voltage range using low frequency square waves so that films were left at a specific voltage for a long time and that rate of actuation didn't affect the amount of actuation. Additionally, the Young's moduli of some films were measured, and a qualitative analysis of films strength is reported. The actuation of all films was measured using a Dual Mode Lever System (Aurora Scientific). Unless stated differently, voltages were controlled relative to a Ag/AgCl reference electrode with a large piece of carbon fiber paper as the counter electrode. Actuation was measured for either square wave or cyclic voltammograms (CVs). The voltages were applied using a Solartron 1260A impedance/gain phase analyzer interfaced with a Solartron 1287A potentiostat. The chapter has been divided into sections according to the synthesis methods used to create the film (see Chapter 2), and further divided within those sections by the actuation conditions. A summary of all the film growth and actuation results is provided in section 3.4. 21 3.1 Yamaura Grown Film (PPy-PF6) Films were synthesized using 0.05 M tetrabutylammonium hexafluorophosphate (TBAP) (Aldrich 98%), 0.06 M distilled pyrrole (Aldrich) and 1 % water in PC (Aldrich 99%) on a glassy carbon crucible. The glassy carbon was polished before use using a one micron diamond scrub followed by polishing with Jewellers rouge. Deposition was performed galvanostatically at 0.125 mA/cm 2 for 8 hours resulting in an 11 /xm thick, smooth, shiny and tough film. The resulting film was rinsed in PC, allowed to dry for at least 24 hours and peeled from the substrate. The film conductivity varied between 350 and 400 S/cm. Films were stored in air for up to 3 months with no measurable losses in conductivity. 3.1.1 Electromechanical actuation in 0.1 M T B A P in P C The operation of PPy-PF 6 has been characterized extensively in TEAP in PC and as such provides a good baseline for polypyrrole actuation [14;54;55]. For these experiments, T B A P was substituted for tetraethylammonium hexafluorophosphate (TEAP) to reduce costs. Figure 3-1 shows the position vs. voltage for a PPy-PF6 film actuated 0.1 M T B A P ( P C ) V S . an Ag/AgCl reference electrode. 22 Actuation in 0.1M TBAP, (PC) S , o CT5 7 1 ; — 1, 1 _ — — 7 ^ — — .1 ,1 1 \ 1 / 1 1 1 10 15 20 25 SO- BS ,40 10, 1.5 20 25 Time .(rain) 30. 35 40 Figure 3-1: Strain, Voltage and Current for a PPy-PF6 film actuated in 0.1 M TBAP (PC). A was CV applied at 3.6 mV/s The interesting results from Figure 3-1 are that while the strain of the polypyrrole reached an expected global maximum at IV, there was also a second local maximum at around -0.8 V . Assuming that only the PF6" anion was mobile then one would assume the strain should be at a maximum at 1 V since PFg" ions flow in, and at a minimum at -0.8 V since PF 6" ions flow out. (Note that because the CV has a finite rate, in this case 3.6 mV/s, the strain peaks would actually slightly follow the voltage peaks since the polypyrrole redox is relatively slow). There are two possibilities for the double peak in this system. The first is the low, but finite mobility of TBAP + within the polypyrrole film. At low voltages, most of the PF6~ has been forced out of the polymer and in order to balance the charge some T B A P + moves. However, contradicting this possibility is N M R data, which shows that PF 6" concentration is linear with voltage [56]. The second and likely cause of the double peak is the change in Young's modulus of polypyrrole. 23 Equation (2) is actually a simplified equation for the strain in polypyrrole that changes in polypyrrole. In reality, as Spinks [12] has shown, since Young's modulus can change for polypyrrole, strain should really be represented by: where As is the electrochemical strain, As0 is the strain with no load, a is the external stress on the the polymer, E' is the Young's modulus of the polymer in the expanded state and E is the Young's modulus of the polymer in the contracted state. The electrochemical strain at zero stress can be calculated as shown in equation (2). If E' is larger than E, then as the external stress goes up the strain goes down. This is the case for actuation with TBAP in propylene carbonate. We did not measure the change in Young's modulus for these conditions (some measurements of this effect on polypyrrole in NaPF6(aq) are shown in section 5-1), however Spinks saw a change in Young's modulus from 0.2 GPa in the reduced state to 0.02 GPa in the oxidized state [57]. Using equation 3, the applied stress of 0.1 MPa, and assuming the modulus change occurs near the fully oxidized state there would then be a maximum reversal in strain of 0.45%, since we had a stress of 0.1 MPa for this system. This is larger than we saw, however the change in Young's modulus occurs throughout the system so the effective reversal would be different. Using the results from the CV, the polypyrrole is cycled with a square wave from it maximum strain voltage of 1.2 V to its minimum strain voltage of -0.1 V to obtain the maximum strain (see Figure 3-2). (3) 24 < c 0 O Actuation in 0.1 M TBAP 10 ( P C ) 15 Time (min) 20 25 30 Figure 3-2: Actuation of PPy-PF6 in 0 . 1 M TBAP. 1.2 V to -0.1 V applied vs Ag/AgCl at 0.001 Hz PPy-PF6 in 0.1M TBAP( P C ) had a peak strain of 1.73 % at 0.001 Hz with a strain to charge ratio(tx) of 5.8* 10"11 m 3/C. Peak strain is defined as the difference between the strain of the film at its most expanded position and its most contracted position. Note that 1.73 % strain does not represent the peak strain that can be obtained. In the highly oxidized state the film does not stop expanding during the entire 500 seconds that a positive voltage is applied. The rate of expansion and magnitude of current drops significantly yet there is still a significant amount of additional strain that could have been obtained had the frequency been decreased below 0.001 Hz. 25 3.1.2 Electromechanical actuation in NaCI ( a q) PPY-PF6 films were actuated in 1.0 M N a C l ( a q ) at -0.9 V to 0.7 V vs. an Ag/AgCl reference electrode. Figure 3-3 shows an example of square wave actuation in NaCl( a q) at O.lMPa. The peak strain is 3.8% with a peak strain rate of 3.1 %/s and a of 3*10"u m 3/C. P P y - P F 6 film-in 1.0M NaCL Time(s) Figure 3-3: Actuation in 1.0M NaCl, cycled between -0.9 V and 0.7 V vs Ag/AgCl at 0.017 Hz The peak strain of 3.8 % and strain rate of 3.1 % are relatively large for conducting polymer actuators. This increases rate of actuation is a significant advantage for . actuation in NaCl ( a q ) rather than in 0.1 M T B A P ( P Q and in a 11 um thick, 10 mm long film connected at one end strain levels out after less than 20 seconds. However, it does have a similar double peak problem. Note that in the expansion (oxidative) phase the polymers expansion gradually levels off. In this direction polymer strain can be characterized using equations (1) and (2) with a constant a so that strain is proportional to 26 charge and position can be set by voltage. However, in the reduction cycle the polymer has as initial larger contraction, which drops off and is followed by a slight expansion. Since the direction of current has not changed the polymer must no longer have a constant a and thus strain is much more difficult to control. In this case, the presumed ion flux rate of Cf ions is much faster than the rate of Na+ flux so the initial large contraction is CI" moving out while the small reversal is likely Na+ ions moving in. It is surprising that this occurs in the reduction cycle. It could be that the Na+ ions move into to the polymer to balance charge since some of the initial PF6" ions might be trapped in the polypyrrole. In the expansion direction, where a is fairly constant, it is approximately 3*10"" m3/C. Actuation using NaCl is advantageous in that NaCl is cheap and relatively benign. As well, chloride is the primary anion in seawater and blood. If polypyrrole operates well in NaCl it offers the possibility that it could be used without encapsulation in internal human or marine applications. Smela has reported that polypyrrole can be operated successfully in blood although this is with PPy-DBS where the Na+ ion is exchanged rather than the CI". [38] 3.1.3 Electromechanical actuation Sodium Ace ta te (NaCH 3 COO) ( a q ) PPY-PF6 films were actuated in 1.0 M NaCH3COO at -0.9 V to 0.7 V vs. an Ag/AgCl reference electrode at 0.1 MPa. Figure 3-4 shows actuation at 0.1 MPa. 27 Actuat ion of P P Y - P F 6 • in• 1:0M : N a G H 3 G G 0 : Time(s) Figure 3-4: Actuation in NaCH3C00, -0.9 V to 0.7 V vs Ag/AgCl, 0.017 Hz Actuation in Sodium Acetate is small with a maximum initial actuation of approximately 2 %. However, this actuation gradually drops off and shows very large double peaks. This makes this solution undesirable for use. There are two likely possibilities for the decrease in actuation. The first is that initial larger actuation is due to exchange of PF6 ions which exists in the "as grown" film. Later cycles then involve the exchange of acetate ions that does not lead to large actuation. The second is that rather than only having one ion mobile throughout the experiment, the positive ions become more and more mobile in the film. Eventually as both ions move in the film there is no net ion flow in or out of the film and while current continues to flow actuation drops off. 28 3.1.4 Electromechanical actuation in NaPF 6(aq) PPy-PF6 6 films were actuated in 1.0 M NaPF 6 at -0.9 V to 0.7 V vs an Ag/AgCl reference electrode at 0.1 MPa. Figure 3-5 shows an example of actuation at 0.1 MPa. P P y - P F 6 film in 1..DM N a P F 6 I I I I l " I I I I / ^ _ _ y I H j i i i i i i i i 0 50 100 -150 200 250 300 350 400 450 500 0.2 I 1 1 77 2 I i i i i i j i i i l. 0 50 1.00 150 200 250 300 350 400 450 500 i 1 1 1 1 r j j I l I l l I i I ° "0 50 100 150 200 250 300 350 400 450 500 Time(s) Figure 3-5: Actuation in 1.0M NaPF6, 0.7 V to -0.7 V vs Ag/AgCl, 0.005 Hz The electromechanical strain in NaPF6(aq) of PPy-PF6 is approximately 6 % with a maximum strain rate of 1.26 %/s and a of 4.5*10" m /C. Note that actuation in NaPF6(aq) does not show a double peak so position can be held constant by setting a DC voltage. While this is not the highest actuating film measured it offers a good combination of a strong highly conductive film and relatively high strain. For that reason it was used in the devices in chapter 5. Since these films were actuated so often they helped give an idea of whether storage of films over time affected actuation. Films actuated 24 hours after growth showed the same magnitude of actuation as films left for three months in air suggesting that storage in air has little negative effects. o. 1 <_ < n CD IT) CO 29 3.1.5 Electromechanical Actuation in BMIM-BF4 PPy-PF6 films were actuated in l-Butyl-3-methyl-imidazolium tetrafluoroborate. Anquetil has shown that large strains of up to 14 % can be generated with PPy-PF6 films in this electrolyte [58]. Figure 3-5 shows initial actuation using CVs from 0.7 V to -0.5 V vs Ag/AgCl. There are two differences between actuation in BMIM-BF4 compared to most other electrolytes. The first is that in BM1M-BF4 actuation is due to cation rather than anion movement. At negative voltages, large immidazalium ions move into the polymer causing it to swell while at positive voltages they are forced out. This is surprising since the BF 4 ions is much smaller than the immidazolium ion and appears it should dominate ion motion. The second difference is that while initially actuation and current are very small, they rise on each cycle. This is likely due to decreasing ionic resistance within the polymer caused by polymer swelling and chain reconfiguration. The polymer is grown doped with PF6 so initially spacing between chains only needs to fit these ions. Unfortunately, as cycling continues the polymer begin creeping significantly and when cycling at 2.5 mV/s has a unrecoverable creep of 2 % per cycle. 30 P P y - P F 6 f i l m in BMIMESF4 Time(s) Figure 3-6: Actuation in BMIBF 4 , 0.8 V to -0.5 V vs Ag/AgCl, Scan rate of 5 mV/s followed by 2.5 mV/s On top of increased actuation and creep, polypyrrole actuated in BMLBF4 becomes a double peaked system over time. Figure 3-6 shows the same film in BMIBF4 cycled using square waves from 0.7 V to -0.5 V vs Ag/AgCl. While there is a peak strain of 3 % in each cycle, it decays over time as the slower moving BF 4 ion moves in and out of the polypyrrole. This effect, called salt draining, has been observed by Pei and De Rossi [41 ;43]. Using only the initial peak strain of the film the strain to charge was 1.1 * 10"10 m3/C. 31 P P y - P F R film in B M I M B F 4 E a CD O O <; < to CD TO 20' . 0 -20 900 1 0.5 0 _ -0.5 £ 900 1000 1100 1200. 1500 r n i i 1 L i " 1 1 1 1300 1400 1500 '• 1 1 1 1 1 - • 1 \ 1 \ -1000 1100 1200 Time(s) 1300 1400 1500 Figure 3-7: Actuation in [BMIM]BF4, 0.8 V to -0.5 V at 0.005 Hz 3.1.6 Electromechanical Actuation in Cyphos IL 106 PPy-PF6 films were actuated in phosphonium ionic liquid triisobutyl (methyl)phosphonium tosylate (Cyphos IL 106). While polypyrrole actuated in immidazolium based ionic liquids(IIL) shows long lifetimes and large strains, to date there is no literature reporting on polypyrrole actuation in phosphonium based ionic liquids. Voltages of 8 V to -8 V vs were applied vs a polypyrrole counter electrode. The reason for these high voltages is to compensate for the high solution resistance in the phosphonium ionic liquid. A n Ag/AgCl reference electrode could not be used due to the large size of the ionic liquid ions that were unable to penetrate the reference electrode vycor tip. Similar to actuation in the IILs, actuation initially started very small but gradually increased in magnitude. However even after many cycles actuation was still much smaller than in aqueous electrolytes or immidazolium based ionic liquid 32 electrolytes. A maximum actuation of 0.1 % was achieved when actuated at 8 V to -8 V with a 0.016 Hz square wave, with a relatively high strain to charge ratio of -3.81 x 10"10 m 3/C. Calculate from 0.9mC in a cycle, and 0.0631% Figure 3-8 shows the initial actuation and current. The direction of actuation (expansion at negative voltages) indicates that ion flux is generated primarily by the motion of the large phosphonium cations. There was also a gradual contraction of the PPy-PF6 in the IIL likely caused by the exit of the polypyrrole's original solvent (PC). P P y - P F 6 film in Cyphos IL 106 0 i 1 1 1 1 1 ~i ' Time(s)' Figure 3-8: Actuation in Cyphos IL 106, 0.05 Hz, +8 V to -8 V vs Counter electrode 33 3.2 Kaneto Grown Films Two films were grown and tested based on Kaneto's methods. The first was grown with 0.25 M pyrrole(Aldrich), 0.2 M TBABF4(Aldrich) in methyl benzoate(Aldrich) on a polished Ti electrode positioned upside down. Deposition was performed galvanostatically at 0.2 mA/cm2 for 4 hours in a 0 °C ice bath. The first film had a thickness of 16 fim with a conductivity of 40 S/cm. The resulting film appeared rough and had poor mechanical properties, making it very difficult to mount in the muscle analyzer. The second film was grown under identical conditions except with 0.1 M TBACF3SO3 substituted for T B A B F 4 . The film thickness was 25 |U.m with a conductivity of 45 S/cm. Qualitatively the films were brittle and broke easily with a rough looking surface. An interesting property of their growth was that their thickness and toughness depended significantly on the orientation of the Ti substrate. A growth on an upside down substrate produced a thin tough film, a growth on a right side up substrate produced a thick powdery film that could not even be placed in the muscle analyser. A vertical growth produced a film with properties between the upside down and right side up films. Since the upside down film was the easier to handle it was used for the measurements shown. These films will be referred to as PPY-BF4 and PPy-CFaSCh respectively. Films grown vertically or right side up showed the same actuation strain but with lower maximum stresses. Figure 3-9 shows actuation of a PPy-BF4 film in NaPF6. Actuation for PPy-CF3SO"3 films was also performed, with similar results to the PPy-BF4 films that are shown on table 3-1. 34 PPy-BF 4 film in 1 .OM NaPF 6 , 1000 1500 20 2 -20 3 o -40 1Q00 _ r r r 1 r 1500 1500 Time(s) Figure 3-9: PPy-BF4 film actuation in 1.0 M NaPF6(aq), 0.7 V to -0.9 V vs Ag/AgCl, 0.002 Hz PPy-BF 4 and PPy-CF 3S03 films had maximum actuations of 8.96 and 8.9 %. This is lower than the actuation 12-15 % strain reported by Kaneto for these films. However, the maximum actuation for these films was measured after several cycles so that strain had stabilized, unlike the measurements done by Kaneto?. If strain is measured based on the first cycles it could appear larger either due to irreversible contraction or expansion that often occur on the first cycle as synthesis ions and solvent are replaced by actuation ions and solvent or creep. As well, growth was performed on a solid Ti substrate while Kaneto films were grown on evaporated Ti. This could have resulted in different substrate roughness and therefore different film morphologies which may have affected strain. These films show promise in that they have almost 50% more strain than PPy-PF6 films in NaPF6(aq). However, the films are much weaker than PPy-PF6 films making them very difficult to use without breaking. As well their electrical conductivity is ten times lower, making them slow if used without electrical contact throughout their length. 35 3.3 Phosphonium Ionic Liquid Grown Films In recent years a great deal of attention has been paid to ionic liquid due to their chemical and thermal stability. Operation of polypyrrole in ILs shows great promise due to the possibility of extremely long cycle lives and large strains [59]. Pringle has synthesized polypyrrole using phosphonium (PIL) and immidazolium (IIL) based ionic liquids and shown that films synthesized in IIL's have greater electrochemical activity than film synthesized in regular organic solvents. She has also shown that films grown in PILs can completely expel their phosphonium cations when cycled in organic solvents indicating that they might have good actuation properties [26;40]. However there are no published reports of actuation of films grown in PILs nor of PIL grown films actuated in any solvents. Growth of polypyrrole was attempted with six phosphonium based ionic liquids: • Cyphos IL 101: tetradecyl(trihexyl)phosphonium chloride • Cyphos IL 103: tetradecyl(trihexyl)phosphonium decanoate • Cyphos IL 104: tetradecyl(trihexyl)phosphonium(bis 2,4,4-trimethylpentyl)phosphinate • Cyphos IL 105: tetradecyl(trihexyl)phosphonium dicyanamide • Cyphos IL 106: triisobutyl(methyl)phosphonium tosylate • Cyphos IL 110: tetradecyl(trihexyl)phosphonium hexafluorophosphate Growth in Cyphos IL 103. and 110 did not work at room temperature due to PIL's solid (wax like) form. When heated they became liquid and up to 10 % acetonitrile and up to 10 % water were mixed with them at high temperatures to try and make them liquid at lower temperatures. However, upon cooling to room temperature they reverted to a solid form. 36 Growth in the other PILs was performed galvanostatically using 0.1M pyrrole on a glassy carbon sheet at 0.125 mA/cm2 for 5 hours. Only growth in Cyphos IL 106 formed a solid film on the electrode. All of the.solution darkened noticeably over time, indicating that polymerization was occurring but either the polypyrrole didn't stick to the electrode or the pyrrole didn't form long enough chains to create a solid film. The film grown in Cyphos IL 106 had a thickness of 20 fim and a conductivity of 20 S/cm. Figure 3-10 shows the chemical structure of Cyphos IL 106. The other PIL's had very small anions compared to cations, but whether this had any effect on their lack of successful growth is unknown. Figure 3-10: Triisobutyl(methyl)phosphonium tosylate 3.3.1 Electrochemical actuation in Phosphonium IL Films were actuated in Cyphos 101 and 106. Actuation was performed at +3 V to -3 V, +8 V to -8 V and +15 V to -15 V vs. a much larger polypyrrole counter electrode. In Cyphos 101 there was no measurable actuation at any voltage. In Cyphos 106 at +/- 8 V and +/- 15 V, there was 0.05 % and 0.1 % actuation respectively at 0.05 Hz (Figure 3-11 shows actuation at ±15 V). 37 0.1 .E 0 ro GO E -0.1 0.5 V - 0 Z5 o -0.5 0 s o Cn ft. < CO > CD cn ro 20 £ T20 o > 0 10 PPy-Cyphos IL.106 film in Cyphos IL 106: i i L . J - J -0 w 20 30 40 5 0 ~ O T 7 0 80 90 '100 J I _ L L 1 0 20 30 50 60 70 80 90 100 20 _ i i ^ — 30 40 50 6.0 Time(s) ' 70 80 90 100 Figure 3-11: Actuation of Cyphos IL 106 grown film. +15 to -15 vs CE at 0.05Hz The current and actuation of PIL grown films is very small compared to traditional films. Much higher voltages are also required for actuation due to the large ionic resistance of the phosphonium ionic liquids. Looking at the direction of actuation, the film contracts on positive voltages, so ions must be leaving the film which means that the positive phosphonium cation is responsible for the majority of ion flux. One could expect that since the phosphonium ion is so large that there would be a large strain to charge ratio. However, in ionic liquids there are no uncharged solvent molecules so ions do not have hydration spheres which they bring into the film. Using the data from actuation at 15 V the strain to charge ratio is -4.7 10"" m3/C which is similar to the magnitude of the strain in NaCl ( a q ) or TBAP ( P C). 38 3.3.2 Electrochemical actuation in NaPF6 ( a q ) A phosphonium grown film was actuated at 0.7 V to -0.9 V vs Ag/AgCl in NaPF6(aq) in order to compare actuation of PIL grown films to Yamaura films under similar conditions. Actuation and currents were very small with a strain of less than 0.1 % at 0.05 Hz. While films grown in Cyphos 106 have reasonable conductivity, they form films with very low electrochemical activity. 5? 0.8 h 0.4 P P y - C y p h o s IL 106 film in N a P F 6 ( a q ) " n 1 1 r n r I I L I J I J L_ 20 40 60 80 100 120 1.40 160 180 200 ^ 0 20 40 60 80 100 120 140 160 180 200 co <: CO < o > CO 5 -1 ? 0 20 40 60 80 100 120 1.40 160 180 200 Time(s) — 1 1 1 1 i i i 1 L i i Figure 3-12: Actuation of Cyphos IL 106 grown film. 0.7 V to -0.9 V vs Ag/AgCl at 0.05Hz in NaPF 6(aq) 3.4 S u m m a r y of Growth Me thods Table 3-1: Summary of different film growth methods and actuation results. Growth Current C o n d u -Density T e m p ctivity So lvent So lute (mA/cm 2 ) Substrate (°C) S/cm P C MB 0.05M T B A P 0.2M T B A B F 4 0.2M T B A B F 4 0.125 0.2 •0.2 Glassy Carbon Ti Ti -35 400 40 40 Actuat ion So lvent Solute Max imum Voltage Actuat ion app l ied (V) Strain(%) Strain to Cha rge Volumetr ic Doub le p0' Capac i tance Peaks W / C ) P C Water Water Water 0.1 M T B A P 1.0M N a P F 6 NaCl N a Acetate 1-Butyl-3-methyl-imidazolium tetrafluoroborate ( B M I M B F 4 ) triisobutyl(methyl) phosphonium tosylate (Cyphos IL 106) Water Water N a P F 6 NaCl 1.2 to-0.1 (vs Ag/AgCl) 0.7 to -0.9 (vs Ag/AgCl) 0.7 to -0.9 (vs Ag/AgCl) 0.7 to-0.9 (vs Ag/AgCl) 0.7 to -.45 (vs platinum psuedorefe-rence electrode) 10 to -10 (vs C E ) 0.7 to -0.9 (vs Ag/AgCl) 0.7 to -0.9 (vs Ag/AgCl) >1.7* 3.8 <1 >0.1* 8.96 Over some ranges No Yes , small Yes , large Yes , after many cycles No No No 5.8 45 -11 -3.81 30 (F/mnQ 0.23 0.08 0.79 0.24 0.2 Not Not Measured Measured ; 0.2M : T B A -: CF3SO3 0.2 Ti 0 45 Water- I N a P F 6 0.7 to -0.9 (vs Ag/AgCl) 8.9 No Not Measured Not Measured Cyphos IL 106 0.125 Glassy Carbon 22 20 triisobutyl(methyl) phosphonium tosylate (Cyphos IL 106) 15 to -15 (vs C E ) >0.1% No -4.7 ** 15 to -15 0 0 tetradecyl(trihexyl)phosphonium (vs C E ) chloride (Cyphos IL 101) 0.7 to -0.9 Water N a P F 6 (vs Ag/AgCl) >0.1* 5.68 0.01 * Actuation was very slow and the film was not left for sufficient time to finish actuation so the maximum actuation was not determined **The volumetric capacitance could only be calculated for films operated against a R E since the voltage on the W E is needed There are a large number of types of films that could be used to make polypyrrole devices. However, PPy-PF6 films in NaPF6 showed the best combination of high conductivity, high and stable strain, and very importantly high strength. These properties made it the best combination for use in the devices shown in Chapter 5. 41 4 Temperature effects on Polypyrrole For polypyrrole to be used to make a useful valve it must function over a wide temperature range. For example, many consumer devices specify that they can be stored between -40°C and 85°C and operated between 0°C to 60°C. While there has been quite a lot of research performed looking at the conductivity of polypyrrole due to exposure to high temperature [60-62]; very little work has looked at actuation properties. Anquetil has characterized the effect of 90°C vacuum on polypyrrole mechanical properties as well as some strain effects and effects on conductivity [58]. However, he did not measure actuation at high temperatures nor did he look at the actuation of pre and post annealed films. Hara actuated polypyrrole at 60°C, showing increased strain and strain rate but did not look at any long terms effects of this temperature [63]. There are several different properties of polypyrrole that may be affected due to temperature exposure. In terms of using polypyrrole as an actuator, the most important property that is affected is electrochemical strain. Characterizing the effect of temperature on electrochemical strain is not as simple as measuring the electrochemical strain over a range of temperatures and finding a simple relationship. The effect electrochemical strain depends on both the environment in which the polypyrrole is exposed to heat, as well as the length of exposure at different temperatures. Additional important properties that can be affected by temperature exposure are; thermal coefficient of expansion, conductivity, permanent length changes, and mechanical properties. Section 4.1 looks at initial strain effects due to exposure to temperature in T B A P ( P Q . Section 4.2 looks at the effect of cycling the temperature from 20°C to 80°C on the actuation of polypyrrole in T B A P ( P Q , NaCl(aq) and NaPF(aq). Section 4.3 looks at the actuation of polypyrrole over a wide temperature range in NaCl(aq) and NaPF(aq). Section 4.4 looks at the effect of temperature on the Young's modulus of polypyrrole. 42 4.1 Initial Strain Effects We initially noted that polypyrrole undergoes an irreversible contraction upon exposure to high temperatures in PC. This contraction occurred only upon the first temperature exposure after which polypyrrole acts as a typical material in terms of thermal expansion and contraction. Anquetil also noted this contraction and attempted to eliminate it by pre-annealing polypyrrole in vacuum. This was successful in that the film didn't contract upon heating, however rather than stay a constant length it expanded once in was re-heated in PC [58]. Using these results we attempted to more fully characterize the initial strain effects in both in PC along with the effect of pre-annealing or pre-soaking the polymer. Temperature cycling was performed using a Fisher Scientific Isotemp Refrigerated Circulator Model 900. It had a fluid volume of 4 litres, which meant that while it could maintain a very constant temperature its heating rate was only 2.9°C/min and its cooling rate was between 2°C/min at 80°C and 0.1°C/min at 20°C. Films were initially heated from 20°C up to 80°C in 0.1 M TBAP (PC).at 0.1 MPa (see Figure 4-1). We cycled in 0.1 M TBAP ( P C ) since it is the same electrolyte that the films were grown in. This eliminates effects that might occur due to ions from the solution replacing ions in the film. 43 Film in 0.1MTBAP PC Strain vs Temperature over time Time(h) Figure 4-1: Initial strain of a PPY-PF6 film exposed to 0.1 M TBAP in PC. The film initially swells and then gradually contracts over the next 12 hours. The propylene carbonate solution is then heated to 80°C and the polymer undergoes a fast contraction of 3 %. When the solution is returned to 20°C the polymer contracts by 0.75 %. Initially the polypyrrole film expands by almost 5 %. This swelling is due to solvent intake as films are dried for at least 2 weeks prior to any experiments. The film then gradually contracts over the next 12 hours. Note that this contraction is very slow, less than 0.5 %/hour, and would typically not be noticed in actuation tests. It might also be neglected under high load conditions as it might be exceeded by creep. The temperature is then ramped up to 80°C (in this case there was an accidental overshoot to 87°C), where the polypyrrole contracts by 3 % in the 30 minutes. The large expansion followed by contraction could pose a challenge in conducting polymer driven devices since these strains are similar to the actuation strains, and could lead to misalignment of actuated structures over time or with temperature. 44 Contraction upon exposure to heat (the Guch-Joule effect) is typical of some polymer networks, however this is not the effect here as the contraction is irreversible [64]. After the period of contraction, the position stabilizes, but when the temperature is dropped back down to 20°C then polypyrrole contracts by another 1%. This begs the question of whether the contraction upon heating was truly a heat-inducted contraction or whether the contraction it is simply the same contraction shown over the first 12 hours, but speeded up. In order to answer this, another piece of the same film was left in 0,1 M T B A P ( P Q for two weeks before being placed in the muscle analyzer (see Figure 4-2). Strain vs Temperature over Time, in 0.1M T B A P P C Presoaked 100 r 1 1 : — i 1 1 • 1 1 r — CD E •80 t 60 CD 40 :20 Strain Temperature _L L _L I I I L LTD 0 2 4 6 8 10 12 14 16 18 Time(h) Figure 4-2: Strain and Temperature for a pre-soaked PPy-PF6 film in 0.1 M T B A P ( P Q In a pre-soaked film, there was initially an expansion of 1% when it was put in the muscle analyzer. Since its soaking solution was the same as its solution in the muscle analyzed, this was likely due to a slight creep from the 0.1 MPa the muscle analyzer applied. This film was left for 10 hours and the strain stabilized. Upon an increase in 45 temperature up to 80°C the film expanded by 1.5%. This expansion was a thermodynamic expansion due to temperature and can be used to determine a coefficient of thermal expansion for polypyrrole. Polypyrrole then contracted by less that 0.5% over the next six hours. Upon return to 20°C the film contracted by 1% to return to its preheated position. These results indicate that the contraction of polypyrrole upon exposure to heat is actually simply increasing the gradual contraction that polypyrrole undergoes when left in 0.1 M T B A P ( P Q . There was still a slight contraction (0.5%) from the heat exposure so it is possible that this contraction is purely caused by heat but it may be that the film wasn't left to soak long enough. We can use these results to determine the coefficient of thermal expansion of polypyrrole in propylene carbonate, 246.6 *10"6 /K. This is almost six times higher than that found in air of 44*10"6/K [58]. This is either due to reduced Young's modulus of polypyrrole when wetted rather than dry or perhaps the films can absorb more solvent when it is warmer and thus swell making the Young's modulus appear lower While pre-soaking eliminates the temperature-induced contraction in films, a second question is whether the contraction over time only happens in PC or can be induced by exposure to high temperature when not in solution. To determine this films were pre-annealed in as vacuum at 90°C for two hours and then exposed to a temperature cycle in 0.1 M T B A P ( P C ) (see Figure 4-3). 46 Strain vs Temperature over T ime, in 0.1 M T B A P P C , preAnnealed 100 i 1 — — i : 1 ••—i — i 15 o tS 50 cu E CD .1-Figure 4-3: Strain and Temperature for a pre-annealed PPy-PF6 film in 0.1M T B A P ( P Q The pre-annealed film showed a 1 0 % contraction upon exposure to the vacuum. Upon exposure to 8 0 ° C in P C it expanded rather than contracted like non pre-annealed films. This expansion can be explained by the fact that pre-annealing dries out films, removing most of the solvent causing them to contract by approximately 10%. When they are re-exposed to solvent, the solvent moves in between polymer chains and the films swell. As shown in Figure 4-3 after six hours of exposure to high temperature the expansion had still not completed, nor made up for the 1 0 % contraction in vacuum. Given that the expansion of pre-annealed films is as large as the contraction of non-annealed films it appears that pre-annealing films is not a good way to ensure a stable position in 0.1 M T B A P . Additionally, films re-immersed in 0.1 M T B A P ( P Q after annealing showed slower and smaller rates of actuation. However, it is difficult to say whether the reductions were due to damage or due to the fact the film did not fully re-absorb the 47 propylene carbonate. Since pre-annealed films in NaCl( a q) still showed the full 3.8 % actuation there is an indication that there was little or no damage to films from pre-annealing. The relatively small thermal expansion and contraction will need to be compensated for in mechanical designs, but this magnitude of strain is at least less than the maximum active strain. 4.2 Effect of temperature Exposure on Actuation The most important constraint on using polypyrrole at high temperatures is the effect of temperature on electrochemical strain. To determine this effect three different films were cycled in different solutions, (see Table 4-1) Table 4-1: Growth and Actuation Conditions of Polypyrrole films Film Growth environment Current Density (mA cm"2' Time (hrs) Thickness Orn) Conductivity (S/m"1) Electrochemical cycling environment 1 Methyl benzoate -0.25M Pyrrole -0.2M TBABF4 0.2 12 37 1.11E+03 Water -1.0M NaPF 6 2 Propylene carbonate -0.06M Pyrrole -0.05M TBAP 0.125 8 10 1.01E+04 Propylene Carbonate -0.05M TBAPF 6 3 Propylene carbonate -0.06M Pyrrole -0.05M TBAP 0.125 8 10 1.01E+04 Water -1.0M NaCl Testing of the films proceeded as: 1) Films were electrochemically cycled in solution at 20°C. Films 1 and 3 were cycled between 0.7 V to -0.9 V vs. Ag/AgCl while film 2 was cycled from 0.4 V to -0.8 V vs Ag/AgCl. 2) Films were left at 0 V vs. Ag/AgCl and the solution temperature was raised to 80°C, held for 30 min and then returned to room temperature. 3) Step 1 was repeated 4) Step 2 was repeated 5) Step 1 was repeated a second time. 48 For an example of this cycling, see Figure 4-4. Note that the frequency of actuation was modified during the experiment in order to ensure strain had stabilized and maximum actuation was achieved. These three different deposition and cycling environments provide a comparison of three commonly used electrolytes and solvents that have relatively high cycle life and show moderate to large strains [8;65]. A parallel set of experiments was performed on films whose conductivity was monitored after thermal cycling. No actuation or electrochemical cycling was performed on these films. Strain and Temperature of a P P y - B F 4 film 1.0M N a P F g ( a q ) 80 I 1 — — — i 1 1 1 1 .-••'••K., .......—i 1 112 0 0.5 1.5 2.5 3 3.5 Time(h) 4.5 Figure 4-4: Typical temperature and cycling test of a PPy-BF4 film actuated in NaPF6(aq). On the first heating cycle there was a small expansion and on the second heating cycle a small contraction. Both annealing or soaking can reduce these effects. 4.2.1 Change in maximum actuation Resulting maximum strains for each film are shown in figure 4-4 (solid bar). Filml (NaPF6) showed a gradual decrease in strain from 7.7% to 5.9% to 4.9%. Film 2 49 (TBAPF6) showed the most significant decrease in strain going from 1.17% to 0.64% to 0.36% strain. Film 3(NaCl) showed virtually no change in initial strain of 3.5 %. These results are plotted in Figure 4-5 below. Reduction in strain amplitude (the difference in strain between the maximum expanded and contracted state of a film), is a combination of reduction in strain to charge ratio and in the total charge transferred. Comparing the effects of these two factors can give an idea of the mechanism behind decrease in total strain of a film. Strain to charge was estimated by averaging the strain and current over several cycles, and subtracting the parasitic current. The parasitic current was assumed to be the stable current obtained near the end of electrochemical cycles where the strain rate had dropped to zero. In this case, changes in strain to charge can account for only a slight amount of the reduction in strain of Film 1 and 2 and strain to charge actually rose in film 3. Given this, the majority of reduction in strain is associated with a net reduction in total charge passed. Note that these experiments were performed with a minimal amount of electrochemical cycling between each thermal cycle to reduce the effect electrochemical cycling on strain. In this case the number of cycles was kept well below that causing significant decrease in strain. To assess whether it was likely that the film was simply breaking down as a result of the elevated temperature a TGA-FTER. (thermal gravimetric analysis - fourier transform infrared spectroscopy) experiment was run on a PF 6 grown film. The test indicated that Table 4-2: Conductivity, hernial Strain, Electrochemical Strain, Max Strain Rate Conductivity .Peak Strain Rate Thermal Strain Film Cycle S/m(x103) %/s % Strain 1 Pre Thermal 1.1 1.2 -BF 4film Post 1 Thermal 1.2 0.83. 2.1 in NaPF 6 Post 2 Thermal 1.2 0.38 -2.1 2. Pre Thermal 10 0.035 -PF 6film Post 1 Thermal 10 0.011 -4.8 in TBAPF 6,pc) Post 2 Thermal 5 0.005 -0.4 3 Pre Thermal 10 3.1 -PF 6film Post 1 Thermal 16 2.2 -3 in NaCloat Post 2 Thermal 10 1.72 -3 50 the first significant atmospheric mass loss begins at around 180°C, when the propylene carbonate begins to evaporate and the PF6 decomposes. There was no indication of thermal decomposition of the backbone until much higher temperature. This suggests that if there are any chemical changes to the polymer at 80°C they are the result of reactions with ions', solvent and polymer already present within the films. 4.2.2 Changes in Rate of Actuation More significant than change in strain is the drop in the rate of actuation, (see Table 4-2) This was calculated by measuring the change in strain in the first second after voltage was switched from its most oxidized to reduced state. Proportionally, the largest change was seen in Film 2 in which rate dropped 7-fold, from a peak strain rate of 0.035 %/s to 0.005 %/s. Film 1 had a 4-fold drop from 1.2 %/s and even film 3 which showed no net change in total actuation had a significant reduction in peak strain rate from 3.1 %/s to 1.72 %/s. Film 1 Film 2 Film 3 4 E? O ™ ^r- a Strain 9 O ~ Q a Strain.to 1 'S ^ Charge 0 3> Pre- Post 1 Post 2 Heating cycle cycles Pre- Post 1 Post 2 Heating cycle cycles Pre- Post 1 Post 2 Heating cycle cycles Figure 4-5: Strain and Strain to Charge of films exposed to 80°C thermal cycles 4.2.3 Temperature Induced Change in Conduct iv i ty and Length The films whose conductivity was monitored in a parallel set of experiments showed some variation in conductivity as shown in Table 4-2. However, this variation was very small and is not large enough to explain the change in rate of actuation. There were also 51 thermal contractions similar to those shown in section 4.1. The thermal contractions present an engineering challenge since they are similar in magnitude to active strain. There are two options to prevent this: 1) As in section 4.1 films can be held high temperatures for several hours before use or left in solution for several weeks. 2) Antagonistic configurations where films act against each other can be used. As long as the thermal conditions for both films are the same, then the length effects will be the same and there will be no net displacement. 4.2.4 Cause of drop in maximum actuator strain and strain rate There are three likely mechanisms for reductions in active strain of films 1 and 2 after thermal cycling. These include; (1) A change in the mechanical structure of the film which increases stiffness and thus reduces strain (2) Structural changes in the film, such as cross-linking, reducing accessibility of ions (3) Irreversible doping of sites on the film from preventing oxidation and reduction. The observed irreversible contraction induced by heating the films suggests a change in structure. However, i f the change in structure is causing a change in stiffness of the film then one would expect the strain to charge would decrease, which is not the case. On the other hand the change in structure could represent a reconfiguration that leads to regions of the film being less accessible to ions. Supporting this is work done by Otero on overoxidized films [66;67] He explains that over oxidation causes cross-linking of polypyrrole chains, forming lakes of electroactive oxidized material trapped within lakes of cross linked chains. This increases the diffusion time constant of the film, and thus increases the ionic resistance. In this case we are not over oxidizing the film but a similar effect could be taking place due to temperature. The last possibility is irreversible doping of sites on films, caused by ions in the solution. Research has shown that OH" ions can irreversibly oxidize active sites on polypyrrole [68]. 52 The changes in strain rates for the films was much larger than the changes in strain and can determine the mechanism behind both the changes in strain and strain rate. Since there was such a large loss in strain rate, and since strain rate is related to the current, there must be an increases in the RC charging time constant of films. This represents either a large increase in the electronic or ionic resistance or capacitance. But, as the total charge passed to the film goes down, capacitance is not going down. Measurements also show that films exposed to identical thermal conditions show almost no changes in electronic resistance. This leaves a major change in ionic resistance. This is well explained by increased cross-linking in the film although irreversible oxidations of some sites on film could make restrict ion movement within the film with the same result. 4.3 Electrochemical Actuation vs Temperature From Section 4.2 we can see that there is a loss in actuation when polypyrrole is exposed to high temperatures. However, this is a slow effect. Increases in temperature also have the effect of temporarily increasing actuation strain. In order to characterize this increase, films were electrochemically cycled quickly while exposed to a fast temperature rise. Figure 4-6 shows the results of this temperature rise on actuation in 1.0 M NaPF6( a q). The charge transferred per cycle has also been included in order to look at the cause of changes in strain. The strain amplitude was determined by electrochemical cycling the films constantly at 0.025Hz and measuring the difference between the maximum and minimum strain of the films. 53 Actuation vs Temperature, +-0.2Vvs Ag /AgC l in N a P F - ^ , 0.025Hz 0.8 I , , . , 1 Temperature (°C) Figure 4-6: Actuation and Charge transferred during actuation vs. temperature. The data in figure 4-6 shows that as the temperature increases, electrochemical strain amplitude increases. Strain to charge also remains constant but strain goes up, likely due to increased mass transport, as might be expected at higher temperatures. However, as the temperature is lowered the strain drops at a much faster rate than it rose and ends up much lower than it began. This is matched by the losses in charged transferred, and thus losses in strain from temperature can be attributed to reduced charging rates. This agrees with earlier results which only measured actuation at 20°C but showed that most of the loss in strain is due to losses in charge transferred rather than losses in strain to charge. 54 1.5 .E 1 05 CO 0.5 20 o O - a ai O l CD co cz CU i— J Z O 0.05 Actuat ion vs Temperature, in 1fvl N a C l L-30 40 50 60 Temperature ,(°C) Charge Transferred'vs Temperature, , in-1M N a C l 70 80 40 .50 60 Temperaturef°C) 80 Figure 4-7: Amplitude of actuation of a PPy-PF6 film in 1.0 M NaCl(aq). 0.4 V to -0.4 V vs Ag/AgCl at 0.05Hz. In the case of cycling in 1.0 M NaCl(aq),.the losses in charge transferred also match the losses in strain. Unfortunately, this test does not include the temperature being lowered back down to 22°C as the film broke during the experiment. Note that neither of these tests shows conclusively that the losses are due to the charges in temperature. Unlike previous tests where only a few electrochemical cycles were performed before the test and no cycling was done at high temperature, in this case the film is being electrochemically cycled constantly. This, electrochemical cycling could be causing significant losses that are not separated from temperature related losses. In fact if you look at the initial strain of the NaCl(aq) cycled film, over time these is a small loss in strain at 22°C (Going from 1.75 % to 1.4 %). In order to separate the effects of losses due to temperature, rather than losses due to cycling, other films were cycled for long period of time at 22°C (see Figure 4-8 and 4-9). 55 In the case of long term actuation of NaCl(aq) there is a reduction in strain which is matched almost exactly by a reduction in strain to charge. This is opposite of the effect we see due to heating without cycling where losses in strain are primarily due to losses in charge transferred. 56 Figure 4-9: Actuation of a PPy-PF6 film in 1.0 M NaPF6(aq), -0.2 V to 0.2 V vs Ag/AgCl at 0.05Hz. In NaPF6(aq) we see a similar effect when cycled at 22°C. In this case the change in strain to charge does not match the change in strain exactly so a little of the loss is due to changes in total charging, however the majority of losses are still due to a reduction in strain to charge. This film showed initial 0.06 % relative to the initial strain per cycle. In this case this represents an absolute loss of 0.00024% strain per cycle, with losses slowing over time. Optimally, using these results we could determine at exactly what temperature polypyrrole begins to thermally degrade. However, as temperature is increased, ionic resistance goes down so the amount of charge transferred per cycle goes up. Simultaneously, thermal degradation decreases the amount of charge transferred, decreasing actuation. As there is some variation cycle to cycle, it becomes very difficult to separate the two effects without many very long tests. Optimally, many separate films 57 of polypyrrole would be brought up to many different temperatures and held there while changes in charge transferred per cycle were measured over a long period of time. However, this was not feasible given our single muscle analyser and circulation bath. In an attempt to determine the temperature of degradation, a films was brought up to different temperatures and dropped down again while measuring strain and charge. Actuation at +-Q.2V vs Ag/AgCl in. N a C I ( a q ) , 0:025Hz '2 0.4 CO •H O 0.2 J 0 < 0 0 0-08 £ 0.06 o 1 0.04 5 0 0 2 o 8 0 £ 60 40 20 i. i I I , i. i • • • - r- 1 > i i i i i i i 5 10 15 20 25 30 35 40 45 i i i i i i i i 0 5 10 15 20 25 30 Tirne(rnin) 35 40 45 i i i i i i • >—NJ^ATAJ \ i i I I i. i I\: i i 10 15 20 25 30 35 40 45 Figure 4-10: Electrochemical cycling of a PPy-PF6 film from 0.2 V to -0.2 V vs Ag/AgCl in NaCl(aq). The amplitude of strain, charge transferred in and out of the film per cycle and temperature are shown. The result in Figure 4-10 show that even a small temperature increase up to 32°C results in a small decrease in charge transferred. However, while the drop in strain with cycling is significant even at 32°C it is not until 81°C that the drop becomes very large. Table 4-3 shows the drop due to exposure to each thermal cycle. These are sequential cycles on the same film so they would be different if imposed separately on a film. The drop is also in relation to the previous film charge transferred rather than its initial charge 58 transferred. The implications of this are that polypyrrole while polypyrrole should be stored and operated at the lowest possible temperature to improve lifetime, for short term operation it works better at higher temperatures. If polypyrrole must be exposed to high temperatures, a baseline requirement could be that it work with at least 50% of its original strain. The time till the polypyrrole has been reduced to 50% of its performance from heat is shown in Table 4-3. A worst case scenario is shown in table 4-3 for the length of time that films can be left before losses are more than 50%. Table 4-3: Drop in Charge r "ransferred from exposure to heat Temperature of Exposure Drop in Charge transferred Time before losses exceed (for approximately 100 min) (%) 50%. (hours) * 32 4.1 28 43 4.0 28 48 6.3 16 62 8.1 13 81 28.8 3 * This calculation makes the assumption that rate in the drop in charge transferred stays constant over time and is proportional to the current (not the initial) amount of charge transferred per cycle. This is likely still an overestimate as the drop for 43°C was less than for 32°C, so the rate of drop must actually decrease over time. 4.4 Y o u n g ' s M o d u l u s v s Tempera tu re As mentioned before, strain in polypyrrole is inversely proportional its Young's modulus. Thus changes in the Young's modulus of polypyrrole as a function of temperature could influence the strain of polypyrrole. To measure the effect of temperature on Young's modulus, a constant 2Hz 0.5 MPA sin wave was applied to a polypyrrole film. The film was then heated up to 80°C and cooled back down to 20°C. By measuring the amplitude of the displacement of the film as a function of oxidation state the modulus was determined. Figure 4-11 shows the resulting Young's modulus of the polypyrrole as it was heated and cooled. 59 Young's modulus vs Temperature in N'aPF. Figure 4-11: Young's Modulus of a PPy-PF6 films vs temperature These tests were actually performed simultaneously with earlier tests on electrochemical actuation vs. temperature. However, because the electrochemical actuation frequency was much lower than the force wave frequency, the displacement due to the force wave was separated out using a band stop filter. The direction of the arrows shows how the test proceeded over time. The film began with a Young's modulus of 880MPa which gradually dropped as the film was heated. On cooling the Young's modulus dropped to 660 MPa, however it eventually returned near the original modulus at 22°C. Considering the magnitude of the final change in Young's modulus (<10 %) these changes cannot explain the changes in electrochemical actuation. 60 4.5 Engineering Implications Reduction in active strain, strain rate and film contraction caused by temperature changes present obstacles to commercial use of polypyrrole actuators. It is quite likely that the effects observed by thermal cycling are reproduced in polymers operated only at room temperature, but simply over longer time periods. Considering that thermal contractions and expansions for these films are on the same order or larger than total strain, either efforts to minimize these contractions or use of designs that are unaffected by thermal contractions are necessary. Both pre-annealing or pre-soaking can have stabilizing effects on film lengths, although pre-soaking appears more promising [58]. If rate is important then since strain to charge remains relatively constant, films can be cycled galvanostatically as by Spinks [59]. This.approach should initially negate the rate effects of thermal aging, provided the overpotentials needed are not so large that deleterious reactions are induced. Alternatively, if the fastest possible rate is of interest, then a resistance compensation approach is desirable [54]. Since reduction in active strain is a function of electrolyte employed, in order to minimize reduction in active strain, other combinations of electrolytes should be tested. It is possible that in more stable electrolytes, such as ionic liquids the thermal effects on polypyrrole would be much smaller. There is also a temporary increase in actuation strain and strain rate of films at higher temperatures. Therefore, higher than room temperature actuation can provide better performance in the short term for polypyrrole films. The gains would be higher in low conductivity electrolytes where the electrolyte ionic resistance is significant and in uses where high rate is very important. 61 4.6 C o n c l u s i o n s Work in this chapter shows that exposure to high temperatures cause a reduction in the electrochemical strain and strain rate in polypyrrole.All the films show a significant reduction in strain rate, while a PF6 grown film operated in NaCl( a q) showed the highest thermal actuation stability. The losses in electrochemical strain and rate correspond with reductions in total charge transferred and a drop in current respectively. The current and strain rate drop was greater the drop in total strain. The cause of the drops in strain is possibly due to cross-linking of polypyrrole chains. This reduces the number of active redox sites and thus reduces the strain amplitude. However, it also creates isolated lakes of ions, which significantly increase the ionic resistance of polypyrrole. Interestingly, thermal exposure losses can be separated from electrochemical cycling induced losses by looking at losses in charge transferred (likely accelerated by heat) vs. reductions in strain to charge (associated with cycling. Temperature also has other effects on the polypyrrole such as a first exposure irreversible contraction. This effect can be minimized by pre-soaking films, or its consequences eliminated by setting up polypyrrole in a trilayer or any other opposing configuration. Young's modulus changes in the process of heating polypyrrole but this effect does not appear to affect actuation significantly. 62 5 Polypyrrole Devices A number of different devices such as an oxygen controller [69], integrated braille screen [70], electrochromic pixel [3] and circulation pump [6] have been created using polypyrrole, although none have been commercially developed. Within devices such as these, polypyrrole is used in two different configurations; bending and linear. Bending devices generally act as bilayers (section 5-2) or trilayers (section 5-3), however they can be stacked as many layers to increase forces[71]. Linear devices are used either as unsupported films, performing work in tension (Section 5-1), or in mechanically supported designs such as helical coils or zig-zag metal wire actuators [2;29-31]. This chapter covers the use of polypyrrole for linear actuators in section 5-1, mechanics of bilayers in section 5-2, along with a simplified analysis of a polypyrrole/Mylar bilayer. Section 5-3 covers the mechanics of a trilayer and the fabrication and analysis of a polypyrrole trilayer used as a valve. Section 5-4 covers the use and design of a linear polypyrrole sliding valve. 5.1 L inear Ac tua t i on This section introduces linear devices, looking at the work they can perform, in order to help with the design of the linear valve shown in section 5-4. Linear based polypyrrole devices work either in; A supported configuration where polypyrrole is grown onto a supportive "scaffold" so that the polypyrrole position is constricted and expansion or contraction of the polypyrrole moves the "scaffold" (for one example of a supported configuration see Figure 5-1). An unsupported configuration where two ends of a polypyrrole film are clamped, tension is applied and when the polypyrrole contracts work is performed (for an image of polypyrrole used in an unsupported configuration see Figure 5-2); 63 or - Gold Wire Scaffold — Polypyrrole Figure 5-1: Zig-Zag metal wire actuator. The bent gold wire confines the direction of actuator to up and down. When the polypyrrole expands or contracts it expands or contracts the scaffold. The gold also improves electrical contact to the polypyrrole and reduces voltage drops through the polymer. •5 mm Figure 5-2: A linear polypyrrole film that was fabricated pulling up a 5 g weight. Left: A zoomed out view of the film and weigh. Top Right: A picture of the weight when the polypyrrole is in the contracted state. Bottom Right: A picture of the weight when the polypyrrole is in the expanded state. 64 The maximum work that a linear polypyrrole device can perform in one stroke is determined by the maximum combination of the device strain multiplied by the device load. We can calculated the maximum stress and strain using equation 3 (Chapter 3) For the PPy-PF 6 films in NaPF 6 which showed the best combination of strength and high strain the Young's modulus vs. potential is shown in Figure 5-3. The Young's modulus was measured by applying a 0.1 Hz, 0.5 MPa square wave to the polypyrrole film. The oxidation state of the film was then varied between -0.7 V and 0.7 V vs Ag/AgCl in 0.1 V increments and held at each voltage for 5minutes. The figure includes two sets of data, the almost first order Young's modulus taken from the 150 ms displacement response and the longer time period effective modulus taken 5 seconds after the force had been applied to the film (including any creep in that time period). (Note that the area and length of the film used to calculate the Young's modulus was the original area and length of the film. In reality the area and length change as a function of oxidation state, however the changes are small and we are interested in an engineering Young's modulus that can be used for equation 3.) y ai c o >-700 600 H 500 400 300 200 100 0 • Young's Modulus 0.15 seconds • Young's modulus 5 seconds -0.8 -0.6 -0.4 -0.2 0 0.2 Voltage vs Ag/AgCl (V) 0.4 0.6 0.8 Figure 5-3: Young's modulus as a function of oxidation state. The 0.15 Young's modulus should be used for high frequency (>1 Hz) actuation calculations while the 5 seconds Young's modulus used for slower actuation. For the short time scale response there is a change in Young's modulus from 630 MPa in the fully expanded (reduced) state compared to 380 MPa in the fully contracted (oxidized state). In the longer time scale, the change is from 540 in the reduced state to 353 in the 65 oxidized state. These changes are the opposite direction from the change in propylene carbonate shown by Spinks [12]. This change in modulus means that strain is higher at high stresses, a result that we see for low load increases in this system. For example, for a PPy-PF6 film going from a stress of 0.64 MPa to 2.32 MPa, an increase of 1.64 MPa, there was an increase in strain from 4.95 % to 5.16 %, an increase of 0.21%. Using equation (3) and the results from Figure 5-3 (the 5 second Youngs modulus) the strain goes up by 0.1 % I MPa, giving an expected increase of 0.16% for the previous example. Unfortunately, at even higher stresses there is larger creep and a chance that the film will break in the fully reduced state. A typical PPy-PF6 film in NaPF6(aq) shows a relatively small creep rate when the stress is less than 2MPa, and tends to break when fully reduced for a long time at more than lOMPa. Since it is generally desirable to restrict creep, it is suggested that the operating stress be kept at or below 2 MPa. As long as stress is less than 2 MPa the change in Young's modulus induces less than a 0.25% strain, a strain that can generally be neglected without significant error. 5.2 B i l aye rs The section covers the mechanics of bilayers along with an example of a fabricated bilayer. It shows the simplest type of polypyrrole bending device and provides a starting point for the analysis of the more complicated trilayers used for the valve in section 5.4. A polypyrrole bilayer consists of a layer of polypyrrole attached to a non-actuating material (see Figure 5-4). Upon redox the polypyrrole layer contracts or expands. Since the non-actuating material does not actively undergo a change in volume but is firmly attached to the polypyrrole this causes a stress gradient, and results in bending of the bilayer. Because the polypyrrole layer must exchange ions with another separate electrode bilayers are generally constrained to operating in solution. 66 b) Removal of Ions c) Insertion of Ions (Contracted) (Expanded) b) Neutral PPy Backing Figure 5-4: Operation of a bilayer. a) In the neutral state there is no stress and the bilayer is straight, b) Redox removes ions from the polymer causing it to contract and the bilayer to curve towards the polymer, c) Redox inserts ions into the polymer causing the bilayer to bend away from the polymer. Directions of arrows outside the bilayer show the direction of ion flow while arrows inside the bilayer show the contraction or expansion of layers. Researchers have created many different polypyrrole bilayers[37;42;72-75]. Bilayers magnify the relatively small strain of polypyrrole into large motions. Equations relating the bending of bilayers to the geometry, polymer expansion and mechanical properties were derived by Pei and Inganas in 1992 [41;75;76]. A simplified equation resulting from this derivation is: 6a-p R R; Er(h2)2 - E 2 N (4) E r E 2 - h r h 2 - (h , +h 2 ) + 4 . (h 1 + h 2 ) R represents the radius of curvature of the bilayer. Rj represents the initial radius of curvature. Ei and E 2 are the Young's modulus of the polypyrrole and backing respectively, hi and h 2 are the thickness of the polypyrrole and backing layer, a and p 67 are the strain to charge ratio and charge density of the polypyrrole respectively. This equation assumes that the Young's modulus of the polypyrrole and backing remains constant, which is not always the case. However, since the polymer does not generate very high forces on its own, the load on a film is low so the effect of changes in Young's modulus is low. 5.2.1 Polypyrrole/Mylar Bilayer Figure 5-5 shows the operation of a bilayer built from a thin film of polypyrrole and mylar tape. The bilayer was fabricated by pressing a 60 fim thick piece of mylar tape (Furon CHR Division) onto a l l [im thick PPy-PF 6 film. Electrical contact was made to one end of the polypyrrole using carbon fiber paper and stainless steel tweezers which were then wrapped in more mylar tape to isolate the them from the solution. Figure 5-5: Actuation sequence of a polypyrrole/mylar tape bilayer in N a P F 6 ( a q ) . The frames are spaced at 2 second intervals. A 0.05Hz 2.8 peak-peak V square wave was applied to the device. The black film behind the bilayer is a piece of polypyrrole acting as the counter electrode. In this case, the bilayer only curls in one direction. The reason for this is that during the first several cycles of actuation in NaPF6( a q) there is an overall contraction in the polypyrrole. This stops after several cycles however the median position remains that shown in the middle frame (Rj is not zero). 68 Note the enormous mechanical amplification in motion caused by the bilayer configuration. The polymer only had a strain of 3.5% and a total length of 30mm. Thus it only contracted by 1.05 mm while the tip of the bilayer moved to a position 6mm above the top of the bilayer, a displacement of over 36mm. Of course, this magnification of strain comes with a corresponding reduction in force. The bilayer fabricated above has the advantage over most actuators using polypyrrole in that it is extremely easy to fabricate and generates very large motions. However, like most bilayers it can only be operated in a liquid environment and is slow due to the fact that the working and counter electrodes are relatively far apart, and thus there is a relatively large solution resistance and RC charging time. 5.3 T r i layers The section covers the mechanics of trilayers along with an example of a fabricated trilayer. The trilayer is the type used for the valve in section 5.4 A polypyrrole trilayer works in a similar fashion to a bilayer except that it has two layers of polypyrrole sandwiching a separator material (Figure 5-6). 69 Figure 5-6: Operation of a trilayer. a) No application of voltage, so no force and the trilayer is straight b) Application of a voltage, redox of the polymer layers cause ions to flow into one layer and out the other. The resulting swelling and contraction of the films causes a stress gradient which bends the trilayer. c) Reverse of step b causes the trilayer to bend in the other direction. A trilayer has several advantages over a bilayer. The first is due to the fact that the separator material can be made of an ionically conductive material. This means that there is a very short ionic path length between the polypyrrole layers, and thus a very low ionic solution resistance. This results in a faster rate of actuation than in bilayers where a separate counter electrode is used. As well, using an ionically conductive separator allows a trilayer to operate out of solution. When removed from solution the separator can store ions that are exchanged between the polypyrrole layers. The separator will dry out in time, however using a very stable electrolyte such as propylene carbonate or ionic liquids can give a lifetime in air of over 50 hours [70]. The trilayer can also be encapsulated in a flexible material to seal in from the environment and eliminate evaporation. Another advantage of trilayers over bilayers is that separator stiffness does not have as much affect on bending curvature as the backing material stiffness has in bilayers. Compared with bilayers there is more polymer to bend the non actuating material and for soft non actuating materials - which actually result in very little bending in bilayers (using equation 10) - the polypyrrole acts against itself so bending is still high. Since the separator is in the middle, even i f it is relatively stiff the strain and hence the stress in it is small. (Strain is proportional to the distance from the neutral axis). 70 The bending equations for a trilayers have been derived by P. Madden [71] using the geometry in Figure 5-10 for two cases: The curvature at zero force is: R p-( h > i h g J 2 - 1 2 -h g . f h V (E \ P g \ + -JL + —* _ i i h g J l E P J. (5) where K is the curvature, R is the radius of curvature, a is the strain to charge ratio, and p is the charge density. E p and h p are the polymer modulus and thickness. E g and h g are the separator modulus and half the separator thickness. In this case the two important determinants of trilayer bending are the ratio of polypyrrole thickness to one half of the separator thickness, h p/h g which will be referred to as \p and the ratio of polypyrrole Young's modulus to separator Young's modulus, E p / E g which we wil l refer to as E ' . The curvature for any trilayer can be determined based on these ratios (see Figure 5-7). GO GO Figure 5-7: Dimensionless curvature with respect to \p(hjhs) The curvature is made dimentionless by multiplying by hg and dividing by a and p. 0 0.5 1 1.5 2 2.5 3 3.5 F = E p / E g Ratio of layer thicknesses (y = hp/hg) Curvature fi * h g / a * p ) Figure 5-8: Dimensionless plot showing the maximum trilayer curvature and ratio of layer thicknesses for that curvature as a function of the relative elastic moduli of the layers. 72 Values have been normalized so that these results can be used for any trilayer. From Figure 5-7 we can see that as E ' goes up the curvature goes up and peaks for a lower value of \p. Figure 5-8 shows the relationship between the E ' and the maximum curvature of a trilayer. For low values E ' the maximum curvature is very low and a very thick layer of polypyrrole is needed compared to the separator thickness. As E ' goes up the curvature goes up and the relative polypyrrole thickness goes down. Therefore, for thin layers of polypyrrole that operate quickly the separator should be as compliant as possible. The second case is the static force at zero curvature: F(w) AAA- T ' E p - a -p-W-h .0 + v|/) - L (6) where is the force with respect to \p, i.e (hp/hg). In this case the separator modulus has no effect so the force can be plotted against the ratio of Young's moduli (see Figure 5-9). 73 15 1 1 10 p-W-a-Ep-hg 5 0 1 1 0 1 2 3 Figure 5 -9: Dimensionless plot showing trilayer force against the ratio of layer's Young's moduli. The force is at zero strain is approximately linearly proportional \p at low values and quadratically proportional to \p at higher values. Therefore, for a higher force trilayer we want a thicker polymer layer, although this causes a reduction in the rate of force generation as well as the rate of bending. Another important factor to note is that force is inversely proportional to length. Therefore, to generate large forces it is better to have a large number of short trilayers rather than one long one. However, this reduces the magnification of strain. 74 F I 1 yf. conducting polymer 1 separator ' . conducting polymer < L ; > Figure 5-10: Geometry of a trilayer: F is the force applied to the end of the polymer. The polymer layers have a thickness hp and the separator layer has a thickness hg. The total length of the trilayer is L and the width is W. 5.4 Polypyrrole/Millipore Separator Trilayer 30pm PPy 110|im 30|im PPy Mllllpore Seperator Platinum Figure 5-11: Polypyrrole Trilayer: Composed of a 110 jtm inner layer of Millipore PVDF separator. Next are two 70nm thin layers of platinum. The outside is 30 jtm thick layers of PPy-PF6 Figure 5-11 shows the configuration of polypyrrole trilayer. The design is based on the trilayers built by Spinks [59]. To fabricate them 70 nm of platinum was evaporated onto a 110 iim PVDF Millipore separator. A 30 /xm layer of PPy-PF6 was then deposited onto the platinum using the method outlined in chapter 3. 75 This design has 5 separate layers so technically it is not a trilayer, however the platinum layer is very thin (430 times thinner than the polypyrrole layers) and not perfectly continuous so it has little effect on the actuation of the device and can be ignored. However, the platinum does help to make uniform contact along the length of the polypyrrole which speeds up charging. Trilayers were created on thinner separators, however they had problems with shorting through the separator. For our devices E p = 500 MPa, Eg = 625 MPa [77] and hg = 110 /rm. Using (6) the curvature is maximized when xj/ = 0.81. Therefore, for maximum actuation of our trilayer we would want a film that was 44.5 urn. However, our film was actually not that thick because having the maximum radius of curvature must be balanced with the increased charging time as films become thicker and the difficulties in fully charging very thick films [77]. Figure 5-12 shows the actuation of a trilayer in a NaPF6 solution. These trilayer can operate out of solution, however they operate at a faster rate in solution since the solution provides an additional ionic path around the separator and into the opposite side of the polypyrrole. Figure 5-12: Actuation of a PPy/Millipore trilayer in NaPF6 ( a q ).. The voltage was stepped between 1.4 V and -1.4 V. There is l s between each frame. 76 5.5 Polypyrrole Lifting Valve A lifting valve was constructed using the polypyrrole/Millipore separator trilayer presented in section 5.4. Figure 5-13: Operation of a trilayer valve. Left: Closed position, no voltage is applied to the valve and it is straight. Right: Open position, voltage is applied to the valve and it curls to unseal the hole below. In the closed position, the trilayer remains in its natural straight position over a perforated steel plate. On application of a positive voltage, it curls upwards, allowing air to flow through the holes in the plate. While on its own this valve design allows easy air flow in the open position it does not seal particularly well in the closed position since the trilayer cannot apply a great deal of force against the plate. The trilayer is also not particularly rigid so edges can lift slightly allowing air to flow even in the closed position. To improve sealing a thin layer of oil is placed between the trilayer and the plate. In the closed position the trilayer lies against the plate with the surface tension from the oil sealing the edges. This does make it slightly more difficult for the valve to open as it must overcome the surface tension of the oil in order to lift away from the plate. Figure 5.14 shows some still pictures from a video taken of a trilayer lifting off an oil covered perforated plate. One problem with this valve design is that it requires headspace in order for the valve to lift far enough off the plate to allow any air to flow. Then the more headspace that is available the more air can flow. Assuming that hole is 1mm in diameter, for the hole to most constricted channel for air to flow through we need at least 1mm of headspace. Figure 5-14: Lifting valve in operation. This shows one trilayer lifting and sealing over holes. For a complete device there would need to a number of trilayers cover the whole series of holes. The trilayer shown is much longer than necessary given the size of the holes. In this case, the trilayer is much longer than necessary in order to make the motion and sealing more visible. A more appropriately scaled trilayer is shown in Figure 5-15. It also lifted up to unseal the holes however due to its short length and smaller motion this is more difficult to see. Figure 5-15: Pictures from wide and short trilayer lifting valve, with the top figure showing the entire bottom edge lifting and the bottom figure showing it closing. There is a thin layer of oil on the plate. 78 5.5.1 Valve Mechanical Analys is In the operation of the valve the oil layer seals between the trilayer and the perforated surface when the trilayer lies flat on the surface, however this means that for the trilayer to separate from the surface it must overcome the oil surface tension. Since the tip of the trilayer releases from the surface first and since the force is lowest at the tip only the maximum force required to separate the tip from the surface will be analyzed. The oil has a contact angle with the perforated surface of approximately 0° and in analyzing the highest force case we assume the oil sealant exhibits a very low contact angle the trilayer. The separation of the trilayer then looks as shown in Figure 5-16. Figure 5-16: Forces acting on the trilayer as it separates from the surface, r is the radius of curvature of the meniscus. L is the length of the trilayer. F t r i | a y e r is the force exerted by the trilayer at the tip. Fmeninsus is the force due to the liquid. The pressure generated by the meniscus can be calculated by: Ap = 2- (7) Where p is the pressure generated by the oil, y is the surface tension of the oil and r is the radius of curvature of the meniscus, r at the time of separation is equal to the half the height of the oil layer. In the case of the experiments this was approximately lOOjum. Then the force at the tip of the trilayer can be calculated using: F(x) = Ap(x)-A A F(x) = 2y — r (8) (9) 79 where A is the area of the trilayer (L * Width). Silicone oil has a surface tension of approximately 20 mN/m so for a 5 mm wide and 30 mm long trilayer this gives a maximum force of 60 mN. Using equation (6) the tip of the trilayer can generate a maximum force of 3.75 mN if the polypyrrole has its maximum strain of 6%, or 1.7mN if the strain is 2.75% (see section 5.5.2 for the reason for 2.75% strain). However, as seen from Figure 6-14 even the long trilayer can actually lift itself from the surface. The analysis of the maximum force must significantly overestimate the required force to overcome the oil surface tension, and is a substantial overestimate of the force the trilayers requires to separate from the plate. The reasons for the overestimate are that the trilayer does not necessary bend perfectly evenly, the contact angle between the oil and trilayer is not truly zero and the assumption that A represents the whole surface of the trilayer assumes that the whole trilayer curls off the surface, which may not be true of sections of the trilayer close to its contact point. We can however use the results to look at how the maximum forces on the trilayer scales as the trilayers size is changes. In this case, the force from the surface tension is proportional to the area of the trilayer and thus scales linearly with the length of the trilayer. On the other hand, the force generated at the tip of trilayer scales inversely with length. Combining these two relationships, the force required divided by the force generated scales quadratically with length. Therefore, if trilayer is made 6 times shorter it is 36 times easier to lift the trilayer. This means that if a trilayer is scaled down to the length shown in Figure 5-15, or even shorter, it will not have any trouble lifting from the plate even if its performance degrades over time. A second consideration is if there is a pressure difference below the perforated plate and over the polypyrrole. This would increase the force required for the polypyrrole to lift off the plate. Assuming there was a max pressure difference of 0.1 atm, the force exerted through one of the perforated hole would be lOkPa * 1 mm2 = lOmN total. This is significant but within the range of force the trilayer can generate. However, if the pressure difference got significantly higher it could become a concern as well if the forces on the section of trilayer in contact with the hole caused damage. On the other 80 hand, because the trilayer curls open it only needs to lift its tip slightly before some gas can flow and pressure is equalized. If pressure is exerted tries to force the valve open, the oil meniscus force could help keep the trilayer closed. 5.5.2 Valve Control The valve was initially designed to operate by opening and closing using a voltage change of 1 V to 1.275 V. However, displacement over this voltage range resulted in very low stresses and strains for the valve. Initially a 0.275 V change appeared sufficient. However as seen in Figure 5-17, the strain vs. voltage is approximately linear from -1 V to 1 V. but drops off above 1 V and below -IV. i 9 n 8 -7 - ^ — 6 -4 " " " " " " " " ' ^ 3 -- —• r — — — — 2 ^ 1 -, '• , G— , -1.5 -1 -0.5 0 0.5 .1 1.5 Voltage vsAg/AgCI (V) Figure 5-17: Strain vs Voltage. The voltage was scanned at 0.67mV7s against an equal size polypyrrole counter electrode. The hysteresis shown in this picture is due to the scan rate, and disappears as the scan rate is slowed. In order to get a large strain the voltage needs to be scaled somewhere between 1 and -1 V. Once voltage is scaled it is easy to increase the range as well, which also increases the force and strain the polypyrrole can generate thus reducing the total amount of polypyrrole required. A BJT based transistor circuit was made that scaled the voltage range of 1 -» 1.275 to l-» 0 respectively. Figure 5-18 shows a circuit diagram and simulation for the electric 81 circuit that was constructed and hooked up to the valve. Scaling the voltage down to 0 to 1 volt results in a strain of 2.75 %. This is the strain that will be used for calculations of force and movement for the valve designs. 5 5 V1 6 Power Supply o.ooo Source Output(V) 0.000 Valve Input (V) 0.000 . Valve current(mA) T 455k |200k$. j - ^ r-rook| • 1 -J 390k -ik< •10k* < ,0.02k — ••1 ooook. Transistor Circuit Polypyrroleitrilayer Trans Circuit 1.CIR 80.000 ' T.(Secs) • 80.000 T (Sees) Figure 5-18: BJT voltage Scaling circuit along with simulation results. This circuit was built and worked to operate a trilayer and sliding valve. The circuit diagram is actually complicated for the simple scaling that it accomplished. Originally, a simpler BJT circuit using only three BJT's was designed that scaled the voltage appropriately but it didn't supply enough current to quickly charge the polypyrrole and therefore opened and closed the valve very slowly. Unfortunately, due to biasing requirements, the BJT circuit draws a fair bit of current even when not switching (78 fiA when the valve is a 0 V and 130 itA when the valve is at IV.) A second circuit was designed (see Figure 5-19), but not fabricated, using MOSFETs to reduce this current. In this case, the circuit draws less than 22 /xA when the valve is at 0 V open and less than 33 /xA when the valve is at 1 V. The current draw is due to the finite 82 resistance from the source to the drain of the MOSFETs even when they are off. Both of these circuit designs assume very little leakage through the polypyrrole valve. In the case of the current at OV this is true as no voltage is applied to the polypyrrole so no current will flow. However, at IV there would be some leakage through the polypyrrole. This leakage is highly dependant on the purity and type of electrolyte used. For NaPF6( a q) which generates the fastest and largest actuation a series of trilayers covering a 600 cm plate would have a leakage current of less than 200 uA. However, if used with other solvents such as PC or Ionic liquids the leakage is much lower(<10uA).' Source Output ( ) Power Supply r r >< 70k < 4 M e g | 70k ^  |-)| R7 \ 70k *> H E 3 Transistor Circuit Valve Input ]02k V v V -> I 0000k Polypyrrole trilayer 0.900 200.000 ' 0.000 Valve Input 20.000 0.000 0.000 Valve current (mA) 40.000 SO .000 120.000 160.000 200.000 Figure 5-19: MOSFET circuit and simulation results. 83 5.5.3 Valve Encapsulation The lifetime of the valve above is limited by the fact that the trilayer dries out over time. With propylene carbonate as solvent the trilayers can work for several days, however using aqueous electrolytes (which show larger and faster strain) performance begins to degrade after several hours. In order to prevent the drying out several encapsulation methods were attempted. Trilayers were wrapped in different materials such as tape and saran wrap, however it wasn't possible to create a good enough seal to prevent solution from evaporation. The tape also gradually unstuck over time since it was exposed to a wet solvent. Trilayers are also very thin (< 200um) so sealing them mechanically with a thick material can affect their mechanical properties and reduces bending. A more successful encapsulation method was based on applying a thin conformal coating. First, the trilayers were dried in air for two weeks. The top section of the trilayers was compressed between two glass slides and the rest of the trilayer was sprayed with on both sides Silicone Conformal Coating 422A (MG chemicals) or Acryl Conformal Coating 419B (MG chemicals). This created a 2-8 /im thick coating around the outside of the trilayers. Aqueous NaPFg was then applied to the uncoated section of the trilayer, which wicked down the length of the trilayer. The trilayers were subsequently cycled from 1.4 to -1.4V. Figure 5-20 shows the successful actuation of a silicone coated trilayer. The acrylic coated trilayer did not work since the acrylic was too rigid and did not adhere well to polypyrrole. Figure 5-20: Actuation of trilayer coated in silicone conformal coating While the trilayer continued to work with the silicone coating, it did not prevent the trilayer from drying out after 3 hours. By eye and under a low magnification optical microscope the silicone coating appeared continuous, however since the trilayer dried out, either the sides of the trilayer were not coated properly or the silicone layer had pinholes. Pinholes may have formed due to the fact the polypyrrole is porous. It is possible that increasing the number of silicone layers could have eliminated leaks, however i f the silicone coating became too thick this would inhibit bending. A look at the actuation of the silicone coated trilayer shows that there was a curling in addition to the desired bending. This was caused by the silicone layer being thicker on one side of the trilayer than the other was. A dip coating method might increase the uniformity of the coating and the pinhole effect. A second option to improve sealing is to switch types of sealing altogether and use a parylene vacuum deposited conformal coating. This type of coating might result in a better seal and a parylene coating can be extremely thin, down to 0.4 jLtm (SC coatings) and still be pinhole free. 5.6 Polypyrrole Sliding Valve A linear sliding valve mechanism solves the problem of vertical space considerations for the lifting valve as well as allowing more airflow. On the other hand it is much more difficult to encapsulate a linear valve than sliding one, mechanical amplification is required and the polymer only performs work during its contraction cycle. A sliding 85 valve was created based on two perforated plates that slide against each other. When the two plates are lined up gas can flow through the holes, however when one plate moves relative to the other the holes no longer line up and the plates seal. Oi l can also be used between the plates in order to increase the level of seal. Figure 5-21 shows a sliding valve created using a polypyrrole films as linear actuators. 3.5mm 35mm l__ o— o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 0 0 0 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - :2.5mrH Figure 5-21: Sliding valve mechanism This polypyrrole sits in a NaPF 6( a q) with electrical contact made at the bottom end. Each strip is looped around the pivoting lever and clamped down at the bottom. Figure 5-22 shows this valve in action. Each film in the valve above is 26 mm long, 4 mm wide and (11*2) pm thick (each film is looped around the lever point so it is doubled). The valve shown above has a mechanical amplification of 10 times. Under a voltage of 1.275 V to -1.275 V , it had a 86 tip movement of 2.6mm. This represents strain of only 1%, much lower than the 4.8% expected based on tests of film in isolation. The reason for this has to do with the friction in the system. The major source of friction in the system is between the top and bottom plates. This friction varies depending on the liquid layer between the two plates. There is also some friction from the pivot of the lever in the acrylic base. Friction reduces the motion near the end of actuation where forces are lowest. Generating enough force to overcome this friction is one of the challenges in making a device using a minimum amount of polypyrrole. The loss in strain due to friction seems to contradict section 5-3, which showed that that the strain of films is relatively constant with constant load. However, the strain that has previously been mentioned was for a constant load . In the present case, the polypyrrole must work against friction. The maximum stress that a film can generate is: G m a x = S * E (10) where E is the contracted Young's modulus of the film. Given that the films had a cross sectional area of 88 * 10 ~9 m2 and a contracted Young's modulus of 530 MPa in the contracted state then the maximum stress the film can generate is 25.44 MPa, which is 0.22 N at the tip. However, as the film contracts, the maximum force it can generate scales down linearly with its level of contraction. Eventually, the force it can generate equals the dynamic force of the system and it can no longer move the tip. The maximum strain a film can generate for given friction can be calculated using: Ff"ct.on / \ F A ~ l^ o Emax/ cross c r . "(Ffriction" Eo' E Across) (12) &max * .p cross where Ffriti0n is the maximum friction (static friction is used as the system moves so slowly), s0 is the free strain of the polymer, A c r 0 Ss is the effective cross sectional area of 87 the film, E is the Young's modulus and smax is the maximum strain the polypyrrole can generate for that system. Assuming we need to move the top plate a distance D, then D = Emax* L (the effective length of the film) and: D ~(Ffriction ~ s o E A cross) (13) L Across E Note that we use effective lengths of the film, and effective cross section because these are changed by mechanical amplification in the system. For example, a three times mechanical amplification makes the effective length three times larger but the effective cross sectional area three times smaller. Equation 13 can then be used to determine the cross sectional area of polypyrrole required to move the valve a distance D (taking into account effective changes in D and FfHction due to magnification) if the film length is L. In the case of the result shown in Figure 5-22, it can be used to calculate the friction in the system if the reduction in strain in the system was occurring entirely due to friction. Using the conditions for the experiment above, with D as 2.5mm, L as 250mm (scaled up by 10 due to the magnification), A c r 0ss as 8.8 * 10 ~9 m2 (scaled down by 10 due to the magnification), E as 530 and s0 as 4.8 pm, the friction of the system was approximately 0.177N. In reality, the friction in the system shown in Figure 5-22 is lower than what would be required to form a sealing valve. Realistically, to effectively seal between the two plates, a layer of oil is required, which can increase the friction between the plates up to 3 N along with a minimum motion of 1.5mm. The following analysis using equation 13 shows how these specifications can be met using the minimum amount of polypyrrole. s0 of 2.75% is used which assumes that IV from the BJT circuit is used to operate the valve. A graph of the volume of polypyrrole required for the realistic case is shown vs cross sectional area of polypyrrole in Figure 5-23. 88 4 - 1 0 ' 6 -10 ' 8 10 ' Cross Sectional Area(m 2) Figure 5-23: Volume of polypyrrole required vs the cross sectional area of the polypyrrole. Friction is IN, and displacement is 1.5mm. In this case the minimum volume of polypyrrole occurs when the cross sectional area is 4*10"7m2 so that the volume is 4.75*10"8 m3, and thus the length is 119mm. This length is too long to fit on the valve mechanism, however by increasing the area by a factor of 4.75 and using 4.75 mechanical amplification we retain the same stress and strain conditions, and can use 25 mm long strips of film. Strips are currently 4mm wide and 11 um thick, for a cross sectional area of 8.8* 10"8 m2 per strip. A total cross sectional area of 1.6* 10"7 m2 is required so this means 18 strips need to be used on each side (strips can be made thicker and wider to later reduce the number of strips but this doesn't change the total volume of polypyrrole required). Then a total volume of polypyrrole, 8 mm wide, 200 urn thick and 25 mm long needs to be used to move the plates. 89 There are other possible causes for losses in strain. Firstly, the actuation frequency (0.05Hz) was too high for maximum actuation. At 0.05 Hz, we would not expect the full strain (the exact amount depends on the geometrical configuration), however the majority of the actuation would still occur. It was also very difficult to keep the entire film taught during operation. When one film was completely expanded, it became slightly loose. When it began contracting again, motion only began once the film became taught. This meant there was wasted actuation of the films. This was minimized by preloading the films, however, there was a limit to the static force that could be applied to the film without breaking becoming an issue. The films also had a finite width leading to a wide attachment point, so that the pivot point was not truly 3.5cm away from the pivot. This meant that motion of the side of the film far from the pivot had less effect than the side close to the pivot. 5.7 Conclusions The results and models in the chapter show how polypyrrole could be used as a low voltage polymer valve. Models of bilayer and trilayer actuation shows how layer thickness and modulus need to be taken into account and should be modified in order to make devices with the right combination of force and displacement. Two different types of polypyrrole valve mechanisms, trilayer and sliding, have been made and the required kinematics demonstrated. A trilayer valve, which relies on an oil layer to seal a trilayer against a perforated plate has the advantage of simple encapsulation but requires vertical head space to allow high gas flow. It can generate sufficient forces to lift off the oil covered perforated plate and gains increased mechanical advantage as it is made shorter. A linear design can move a sliding plate and could be fabricated with 200 um thickness and 8mm width if frictional forces were 3N. However, it is a more complicated design and much more difficult to encapsulate. The two designs provide a basis for fabricating a complete valve, which could be evaluated in terms of total lifetime, reliability and air flow. 90 6 Conclusion and Future Work This chapter summarizes the results of this thesis, looks at the major contributions of the research and addresses future work that needs to be performed. 6.1 S u m m a r y of R e s u l t s The goal of this work was to evaluate the feasibility of a low voltage polypyrrole valve. For a polypyrrole valve to be feasible it must be able to withstand large temperature variations, have a high strain (to minimize mechanical amplification), have high work density (to minimize the amount of polypyrrole required), have a long lifetime and be possible to assemble into a compact valve. To find the best polypyrrole synthesis and actuation condition twelve different synthesis methods and actuation conditions were investigated. The best results were found with polypyrrole doped with PF<s" synthesized in propylene carbonate and actuated in NaPF6(aq). This film showed up to 6% strain, had low creep at stresses below 2 MPa, and decreases in strain by 0.06 % relative to its initial strain per cycle. Since, for the valve to work in a device, it has to work over a large temperature range, temperature tests on films were conducted. Results showed that films have an initial temperature induced contraction (which can be eliminated by pre-soaking) and gradually degrade at high temperatures. This degradation occurs faster at higher temperatures and causes a reduction in the charge that can be transferred into polypyrrole. However, cycling at high temperatures does show some temporary improvements in performance. Degradation caused by temperature can be separated from cycling losses as they cause loss in charge transferred rather than strain to charge. Finally, to evaluate the kinetics as well as evaluate the ease of fabrication demonstrations of valve mechanism were built. Two versions, one with a linear and one with a trilayer configuration were constructed and tested. The linear mechanism generated sufficient 91 force to lift off an oil covered perforated plate and gains increased mechanical advantage as it is made shorter. A linear design moved two perforated plates, opening and closing air flow holes, and could be fabricated with 200 um thickness, 25 mm length and 8 mm width if maximum frictional forces were 3 N for a movement of 1.5 mm. Both designs showed promise, with the trilayer design simpler to build and encapsulate but taking up more space and not as certain to seal properly. 6.2 Major contributions of this research • Evaluation of twelve different synthesis and actuation conditions for polypyrrole • First ever electrochemical actuation of polypyrrole in phosphonium ionic liquids and of polypyrrole grown in phosphonium ionic liquids • First demonstration of temperature effects on polypyrrole actuation • Only demonstration of polypyrrole encapsulated with a conformal thin film • Built two different polypyrrole based valve mechanisms and evaluated the requirements for a full valve. 6.3 Future Work Since the goal of this work was to evaluate a polypyrrole valve, any future work must address issues that were exposed from the results of this thesis. The critical issue is to determine whether polypyrrole definitely can or definitely cannot be used to create a low voltage valve. Firstly, all the films showed temperature degradation over time, which must be reduced if the valve is to be used. Using current methods films can be exposed to 43°C for up to 28 hours and 62°C for up to 13 hours before losing half their strain. However, no research has looked at the effect of using very purified electrolytes with 92 polypyrrole actuators. Work with polypyrrole supercapacitors has shown that pure systems increase lifetime and voltage ranges, suggesting that impurities are a problem, so it is likely this would reduce and might hopefully eliminate thermal degradation. Secondly, while demonstration version of valves were created they were not complete enough to use to evaluate gas flow. Complete versions of the linear valve require a smaller lever system as well as encapsulation. A version of the trilayer valve would require a series of trilayer to be coordinated to cover a large number of holes on the perforated plate. These valves should be built, attached to a sample power supply and the performance measured over time. Bibliography 93 1 Snook, G.A., Chen, G.Z., Fray, D.J., Hughes, M. and Shaffer, M., Studies of Deposition of and Charge Storage in Polypyrrole-Chloride and Polypyrrole-Carbon Nanotube Composites With an Electrochemical Quartz Crystal Microbalance, Journal of Electroanalytical Chemistry, 568 (2004) 135-142. 2 Spinks, G. M . , Wallace, G. G., Ding, J., Zhou, D., and Xi, B. Gillespie J. Ionic liquids and polypyrrole helix tubes: bringing the electronic Braille screen closer to reality. Smart Structures and Materials 2003: Electroactive Polymer Actuators and Devices. 5051, 372-380. 2003. 3 Smela, E., A micro fabricated moveable electrochromic "pixel" based on polypyrrole, Advanced Materials, 11 (1999) 1343-1345. 4 Kaneto, K., Kaneko, M., Min, Y. and MacDiarmid, A.G., "Artificial muscle" : Electromechanical actuators using polyaniline films, Synthetic Metals, 71 (1995) 2211-2212. 5 Madden, John D., Schmid, Bryan, Lafontaine, Serge R., Madden, Peter G. A., Hover, Franz S., McLetchie, Karl, and Hunter, Ian W. Conducting-polymer-driven actively shaped propellers and screws. Smart Structures and Materials 2003: Electroactive Polymer Actuators and Devices (EAPAD). 5051, 437-441. 2003-. San Diego, CA, USA. 6 Zama, T., Tanaka, N., Takashima, W. and Kaneto, K., Fast and Large Stretching Bis(Trifluoromethylsulfonyl)Imide (Tfsi)-Doped Polypyrrole Actuators and Their Applications to Small Devices, Polymer Journal, 38 (2006) 669-677. 7 Hara, S., Zama, T., Takashima, W. and Kaneto, K., Gel-Like Polypyrrole Based Artificial Muscles With Extremely Large Strain, Polymer Journal, 36 (2004) 933-936. 8 Madden, J.D., Rinderknecht, D., Anquetil, P.A. and Hunter, I.W., Cycle life and load in polypyrrole actuators, Sensors and Actuators A, (2005) 9 Lu, W., Fadeev, A., Qi, B., Smela, E., Mattes, B., Ding, J., Spinks, G., Mazurkiewicz, J., Zhou, D., Wallace, G., MacFarlane, D., Forsyth, S. and Forsyth, M., Use of ionic lquids for pi-conjugagted polymer electrochemical devices, Science, 297 (2002) 983-987. 10 Anquetil, Patrick A., Rinderknecht, Derek, Vandesteeg, Nathan A., Madden, John D., and Hunter, Ian W. Large strain actuation in polypyrrole actuators. Smart Structures and Materials 2004: Electroactive Polymer Actuators and Devices (EAPAD). 5385, 380-387. 2004. San Diego, CA, USA. 11 Hunter, LW. and Lafontaine, S.R., A Comparison of Muscle With Artificial Actuators, Solid-State Sensor and Actuator Workshop, 1992. 5th Technical Digest., IEEE ,(1992) 178-185 12 Spinks, G., Liu, L., Wallace, G. and Zhou, D., Strain response from polypyrrole actuators under load, Advanced Functional Materials, 12 (2002)437-440. 13 Mazzoldi, A., Delia Santa, A. and De Rossi, D., Conducting polymer actuators: Properties and modeling. In Y. Osada and D.E. De Rossi (Eds.), Polymer Sensors and Actuators, Heidelberg, 1999, 94 14 Madden, J.D., Madden, P.G. and Hunter, I.W., Characterization of polypyrrole actuators: modeling and performance. In Yoseph Bar-Cohen (Ed.), Proceedings of SPIE 8th Annual Symposium on Smart Structures and Materials: Electroactive Polymer Actuators and Devices, Bellingham WA, 2001, pp. 72-83. 15 Baughman, R.H., Conducting polymer artificial muscles, Synthetic Metals, 78 (1996) 339-353. 16 Warren, M., M.A.Sc. Thesis: Electronic and Structural Effects on the Electrochemistry of Polypyrrole, Vancouver, BC, 2005, 17 Warren, M., M.A.Sc. Thesis: Electronic and Structural Effects on the Electrochemistry of Polypyrrole, Vancouver, BC, 2005, 18 Madden, J.D., Vandesteeg, N., Anquetil, P.A., Madden, P.G., Takshi, A., Pytel, R.Z., Lafontaine, S.R., Wieringa, P.A. and Hunter, I.W., Artificial muscle technology: Physical Principles and Naval Prospects, IEEE Journal of Oceanic Engineering, (2004) 24 pages 19 Baker, C K . and Reynolds, J.R., A quartz microbalance study of the electrosynthesis of polypyrrole, Journal of Electroanalytical Chemistry, 251 (1988) 307-322. 20 Lim, H.K., Lee, S.O., Song, K.J., Kim, S.G. and Kim, K.H., Synthesis and Properties of Soluble Polypyrrole Doped With Dodecvlbenzenesulfonate and Combined With Polymeric Additive Poly(Ethylene Glycol), Journal of Applied Polymer Science, 97 (2005) 1170-1175. 21 Song, K.T., Lee, J.Y., Kim, H.D., Kim, D.Y., Kim, S.Y. and Kim, C.Y., Solvent Effects on the Characteristics of Soluble Polypyrrole, Synthetic Metals, 110 (2000) 57-63. 22 Song, K.T., Lee, J.Y., Kim, D.Y., Rhee, H.W., Kim, S.Y. and Kim, C.Y., Effect of Molecular Interactions on Solubility of Polypyrrole, Molecular Crystals and Liquid Crystals, 349 (2000) 363-366. 23 Bazzaoui, M., Martins, J.I., Costa, S.C., Bazzaoui, E.A., Reis, T.C. and Martins, L., Sweet Aqueous Solution for Electrochemical Synthesis of Polypyrrole Part 2. On Ferrous Metals, Electrochimica Acta, 51 (2005) 4516-4527. 24 Zama, T., Hara, S., Takashima, W. and Kaneto, K., The Correlation Between Electrically Induced Stress and Mechanical Tensile Strength of Polypyrrole Actuators, Bulletin of the Chemical Society of Japan, 77 (2004) 1425-1426. 25 Baker, C K . and Reynolds, J.R., A quartz microbalance study of the electrosynthesis of polypyrrole, Journal of Electroanalytical Chemistry, 251 (1988) 307-322. 26 Pringle, J.M., Macfarlane, D.R. and Forsyth, M., Solid State Nmr Analysis of Polypyrrole Grown in a Phosphonium Ionic Liquid, Synthetic Metals, 155 (2005) 684-689. 27 Fenelon, A.M. and Breslin, C.B., The Formation of Polypyrrole at Iron From l-Butyl-3-Methylimidazolium Hexafluorophosphate, Journal of the Electrochemical Society, 152 (2005)D6-D11 28 Hara, S., Zama, T., Takashima, W. and Kaneto, K., Free-Standing Polypyrrole Actuators With Response Rate of 10.8% S(-l), Synthetic Metals, 149 (2005) 199-201. 29 Hara, S., Zama, T., Takashima, W. and Kaneto, K., Polypyrrole-Metal Coil Composite Actuators as Artificial Muscle Fibres, Synthetic Metals, 146 (2004)47-55. 95 30 Hara, S., Zama, T., Tanaka, N., Takashima, W. and Kaneto, K., Artificial Fibular Muscles With 20% Strain Based on Polypyrrole-Metal Coil Composites, Chemistry Letters, 34 (2005) 784-785. 31 Hara, S., Zama, T., Ametani, A., Takashima, W. and Kaneto, K., Enhancement in Electrochemical Strain of a Polypyrrote-Metal Composite Film Actuator, Journal of Materials Chemistry, 14 (2004) 2724-2725. 32 Yamaura, M., Hagiwara, T. and Iwata, K., Enhancement of electrical conductivity of polypyrrole .. film by stretching: counter ion effect, Synthetic Metals, 26 (1988) 209-224. 33 Warren, M.R. and Madden, J.D., A Structural, Electronic and Electrochemical Study of Polypyrrole as a Function of Oxidation State, Synthetic Metals, 156 (2006) 724-730. 34 Hutchison, A.S., Lewis, T.W., Moulton, S.E., Spinks, G.M. and Wallace, G.G., Development of polypyrrole-based electromechanical actuators, Synthetic Metals, 113 (2000) 121-127. 35 Madden, J.D., Madden, P.G., Anquetil, P.A. and Hunter, I.W., Load and time dependence of displacement in a conducting polymer actuator, Materials Research Society Proceedings, 698 (2002) 137-144. 36 Shimoda, S. and Smela, E., The effect of pH on polymerization and volume change in PPy(DBS), Electrochimica Acta, 44 (1998) 219-238. 37 Christophersen, M., Shapiro, B. and Smela, E., Characterization and Modeling of Ppy Bilayer Microactuators - Part 1. Curvature, Sensors and Actuators B-Chemical, 115 (2006) 596-609. 38 Jager, E.W.H., Smela, E. and Inganas, O., Microfabricating Conjugated Polymer Actuators, Science, 290 (2000) 1540-1545. 39 Smela, E., Conjugated Polymer Actuators for Biomedical Applications, Advanced Materials, 15 (2003)481-494. 40 Pringle, J.M., Efthimiadis, J., Howlett, P.C, Efthimiadis, J., Macfarlane, D.R., Chaplin, A.B., Hall, S.B., Officer, D.L., Wallace, G.G. and Forsyth, M., Electrochemical Synthesis of Polypyrrole in Ionic Liquids, Polymer, 45 (2004) 1447-1453. 41 Pei, Q. and Inganas, O., Electrochemical Application of the bending beam method. 1. Mass transport and volume changes in polypyrrole during redox., Journal of Physical Chemistry, 96 (1992) 10507-10514. 42 Otero, T.F, and Broschart, M., Polypyrrole Artificial Muscles: a New Rhombic Element. Construction and Electrochemomechanical Characterization, Journal of Applied Electrochemistry, 36 (2006) 205-214. 43 Delia Santa, A., De Rossi, D. and Mazzoldi, A., Performance and work capacity of polypyrrole conducting polymer linear actuator, Synthetic Metals, 90 (1997) 93-100. 44 Kassim, A., Basar, Z.B. and Mahmud Hnme, Effects of Preparation Temperature on the Conductivity of Polypyrrole Conducting Polymer, Proceedings of the Indian Academy of Sciences-Chemical Sciences, 114 (2002) 155-162. 45 Grgur, B.N., Krstajic, N.V., Vojnovic, M.V., Lacnjevac, C. and Gajic-Krstajic, L., The Influence of Polypyrrole Films on the Corrosion Behavior of Iron in Acid Sulfate Solutions, Progress in 96 Organic Coatings, 33 (1998) 1-6. 46 Grgur, B.N., Zivkovic, P. and Gvozdenovic, M.M., Kinetics of the Mild Steel Corrosion Protection by Polypyrrole-Oxalate Coating in Sulfuric Acid Solution, Progress in Organic Coatings, 56(2006)240-247. 47 Herrasti, P., Recio, F.J., Ocon, P. and Fatas, E., Effect of the Polymer Layers and Bilayers on the Corrosion Behaviour of Mild Steel: Comparison With Polymers Containing Zn Microparticles, Progress in Organic Coatings, 54 (2005) 285-291. 48 Wallace, G. G , Dominis, A., Spinks, G. M., and Tallman, D. E. Factors Influencing the Performance of Inherently Conducting Polymers as Corrosion Inhibitors: the Dopant. 843, 103-123. 49 Paliwoda-Porebska, G., Rohwerder, M., Stratmann, M., Rammelt, U., Due, L.M. and Plieth, W., Release Mechanism of Electrodeposited Polypyrrole Doped With Corrosion Inhibitor Anions, Journal of Solid State Electrochemistry, 10 (2006) 730-736. 50 Branzoi, V., Branzoi, F., Pruna, A., Pilan, L. and Anghel, C , Corrosion Kinetics and Protection of Pure Zinc in Aqueous Nh4cl Solutions in the Presence of Some Organic Inhibitors, Revue Roumaine De Chimie, 50 (2005) 1019-1025. 51 Koene, L., Hamer, W.J. and De Wit, J.H.W., Electrochemical Behaviour of Poly(Pyrrole) Coatings on Steel, Journal of Applied Electrochemistry, 36 (2006) 545-556. 52 Tuken, T., Yazici, B. and Erbil, M., The Corrosion Behaviour of Polypyrrole Coating Synthesized in Phenylphosphonic Acid Solution, Applied Surface Science, 252 (2006) 2311-2318. 53 Liu, Y., Varahiamyan, K. and Cui, T.H., Low-Voltage All-Polymer Field-Effect Transistor Fabricated Using an Inkjet Printing Technique, Macromolecular Rapid Communications, 26 (2005) 1955-1959. 54 Madden, J.D., Cush, R.A., Kanigan, T.S. and Hunter, I.W., Fast contracting polypyrrole actuators, Synthetic Metals, 113 (2000) 185-193. 55 Madden, J.D. , Madden, P.G, Anquetil, P.A. and Hunter, I.W., Load and time dependence of displacement in a conducting polymer actuator, Materials Research Society Proceedings, 698 (2002) 137-144. 56 Jenny Tso, NMR study of polypyrrole oxidation, (To Be Published) 57 Spinks, G., Liu, L., Wallace, G. and Zhou, D., Strain response from polypyrrole actuators under load, Advanced Functional Materials, 12 (2002) 437-440. 58 Anquetil, P. A. Large Contraction Conducting Polymer Molecular Actuators. Ph.D thesis. 2005. Cambridge, MA. 59 Spinks, G.M., Xi, B.B., Zhou, D.Z., Truong, V.T. and Wallace, G.G., Enhanced Control and Stability of Polypyrrole Electromechanical Actuators, Synthetic Metals, 140 (2004) 273-280. 60 Sakkopoulos, S., Vitoratos, E., Dalas, E., Kyriakopoulos, N., Malkaj, P. and Argyreas, T., Shrinking Rate of Conducting Grains in Hci-Protonated Polyaniline, Polypyrrole, and Polypyrrole/Polyaniline Blends With Their Thermal Aging, Journal of Applied Polymer Science, 97 (2005) 117-122. 97 61 Satoh, M., iSHIKAWA, h., sMANO, k., Hasegawa, E. and Yoshino, K., Highly conducting polypyrrole prepared from homogenous mixtures of pyrrole/oxidizing agent and its applications to solid tantalum capacitors, Syntehtic Metals, 65 (1994) 39-44. 62 Samuelson, L.A. and Druy, M.A., Kinetics of the Degradation of Electrical-Conductivity in Polypyrrole, Macromolecules, 19 (1986) 824-828. 63 Hara, S., Zama, T., Takashima, W. and Kaneto, K., Artificial Muscles Based on Polypyrrole Actuators With Large Strain and Stress Induced Electrically, Polymer Journal, 36 (2004) 151-161. 64 GrosbergA.U and Khokhlove A.R, Giant Molecules, edn., 1998, 65 Zama, T., Hara, S., Takashima, W. and Kaneto, K., Comparison of Conducting Polymer Actuators Based on Polypyrrole Doped With Bf4(-), Pf6(-), Cf3so3-, and Clo4-, Bulletin of the Chemical Society of Japan, 78 (2005) 506-511. 66 Otero, T.F., Marquez, M. and Suarez, I.J., Polypyrrole: Diffusion Coefficients and Degradation by Overoxidation, Journal of Physical Chemistry B, 108 (2004) 15429-15433. 67 Otero, T.F. and Boyano, I., Characterization of Polypyrrole Degradation by the Conformational •Relaxation Model, Electrochimica Acta, 51 (2006) 6238-6242. 68 Schlenoff JB and Xu H, Evolution of physical and electrochemical properties of polypyrrole during extended oxidation, Journal of the Electrochemical Society, 139 (1992) 2397-2401 . 69 Spinks, G.M., Wallace, G , Zhou, D., Ding, J., Xi, B. and Gillespie, J., Ionic liquids and helical interconnects: bringing the electronic braille screen closer to reality, Proceedings of SPIE Smart Structures and Materials, 5051 (2003) 372-380. 70 Andrews, M.K., Jansen, M.L., Spinks, G.M., Zhou, D.Z. and Wallace, G .G, An Integrated Electrochemical Sensor-Actuator System, Sensors and Actuators a-Physical, 114 (2004) 65-72. 71 Madden, P.G.A., Ph. D. Thesis: Development and modeling of conducting polymer actuators and demonstration of a conducting polymer-based feedback loop, Cambridge, MA, 2003, 72 Delia Santa, A., Mazzoldi, A., Tonci, C. and De Rossi, D., Passive mechanical properties of polypyrrole films: a continuum poroelastic model, Materials Science and Engineering C, 5 (1997) 101-109. . 73 Otero, T.F., Rodriguez, J., Angulo, E. and Santamaria, C , Artificial muscles from bilayer structures, Synthetic Metals, 55-57 (1993) 3713-3717. 74 Kaneto, Keiichi, Min, Yonggang, MacDiarmid, and Alan G. Conductive polyaniline laminates . 96. 75 Pei, Q. and Inganas, O., Conjugated polymers and the bending cantilever method: electrical muscles and smart devices,' Advanced Materials, 4 (1992) 277-278. 76 Pei, Q. and Inganas, O., Electrochemical Muscles: bending strips built from conjugated polymers, Synthetic Metals, 55-57 (1993) 3718-3723. 77 Metz, P., Alici, G. and Spinks, G.M., A Finite Element Model for Bending Behaviour of Conducting Polymer Electromechanical Actuators, Sensors and Actuators a-Physical, 130 (2006) 1-11. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101166/manifest

Comment

Related Items