Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Design and characterization of polymeric strain gauges for biomedical applications Almarghalani, Maan 2015

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

Item Metadata


24-ubc_2015_may_almarghalani_maan.pdf [ 32.2MB ]
JSON: 24-1.0167155.json
JSON-LD: 24-1.0167155-ld.json
RDF/XML (Pretty): 24-1.0167155-rdf.xml
RDF/JSON: 24-1.0167155-rdf.json
Turtle: 24-1.0167155-turtle.txt
N-Triples: 24-1.0167155-rdf-ntriples.txt
Original Record: 24-1.0167155-source.json
Full Text

Full Text

Design and Characterization of Polymeric Strain Gaugesfor Biomedical ApplicationsbyMaan AlmarghalaniB.ASc, The University of British Columbia, 2012A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMaster of Applied ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Electrical and Computer Engineering)The University of British Columbia(Vancouver)March 2015c©Maan Almarghalani, 2015AbstractThe market need for organic materials to be used in sensor design has increasedwith the growing interest in organic printed electronics. Therefore, it is importantto find and investigate the piezoelectric and piezoresistive properties of organicmaterials through the use of alternative rapid fabrication techniques. Poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate), commonly known as PEDOT:PSS,a conductive polymer widely used in organic electronics, can be possibly used aspiezoresistive element to measure the strain on flexible substrate electronics. UsingPEDOT:PSS and other metallic inks such as silver, the goal of this work is use alter-native microfabrication technologies to deposit PEDOT:PSS on flexible substratesand then to use these methods to design strain gauges. The targeted biomedicalapplications of the designed strain gauges vary from rehabilitation devices to smartbiomedical monitoring systems. In this work, PEDOT:PSS strain gauges are ini-tially designed using aerosol jet deposition on a flexible polyamide substrate. Thetechnology has proved to be very powerful in depositing lines with thickness lessthan 1um. In order to reduce the initial resistance of the strain gauges, it is desir-able to increase the thickness of the structure. For this reason, laser micromachin-ing etching is used to fabricate PEDOT:PSS strain gauges. The designed structureshave been tested mechanically and electrically in order to measure their gauge fac-tors to longitudinal and transversal mechanical strains. The resultant longitudinalgauge factor varied in the range of -1 and 2, while little change in the resistancewas noticed for transversal characterization. Using the same fabrication method,silver paint strain gauges are designed and characterized to have a high longitudi-nal gauge factor approximated to be higher than 10. The silver paint gauge factorbarely responded to transversal actuation. While the variability of the PEDOT:PSSiistrain gauges results seemed to be an issue, the reproducibility of silver ink straingauges proved the viability of the technological fabrication process presented inthis work.iiiPrefaceThis dissertation is original, unpublished, independent work by the author, M. Al-marghalani in the Adaptive Microsystems Lab at The University of British Columbia(UBC).ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Strain Gauges in Biomechanics . . . . . . . . . . . . . . . . . . . 11.1.1 Piezoresistivity and Gauge Factor . . . . . . . . . . . . . 21.1.2 Electromyography (EMG) and Strain Measurements . . . 51.2 Conductive Polymers . . . . . . . . . . . . . . . . . . . . . . . . 81.2.1 PEDOT:PSS . . . . . . . . . . . . . . . . . . . . . . . . 101.2.2 PEDOT:PSS Based Stress Sensors . . . . . . . . . . . . . 111.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151.4 Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.5 Chapters Overview . . . . . . . . . . . . . . . . . . . . . . . . . 182 Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.1 Substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202.2 Fabrication Methods of the PEDOT:PSS Strain Gauges . . . . . . 20v2.2.1 Deposition Using Aerosol jet Printing . . . . . . . . . . . 202.2.2 Laser Micromachining and Peel-off . . . . . . . . . . . . 222.2.3 Laser Micromachining Etching . . . . . . . . . . . . . . . 253 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283.1 White Light Interferometry . . . . . . . . . . . . . . . . . . . . . 283.2 Tensile Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.3 Beam Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.3.1 Analytical Model . . . . . . . . . . . . . . . . . . . . . . 313.3.2 Finite Element Analysis of Beam Deflection . . . . . . . . 354 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 374.1 Substrate Material Comparison . . . . . . . . . . . . . . . . . . . 374.2 Comparison of the Fabrication Technologies . . . . . . . . . . . 414.3 Electromechanical Characterization of PEDOT:PSS Strain Gauges 464.4 Commercial Strain Gauges . . . . . . . . . . . . . . . . . . . . . 504.5 Electromechanical Characterization of Silver Strain Gauges . . . . 514.6 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.1 Thesis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 565.2 Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60A Supporting Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 66A.1 LabView Virtual Instrument (VI) Design . . . . . . . . . . . . . . 66A.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . 70A.3 Measurement Results . . . . . . . . . . . . . . . . . . . . . . . . 71B Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75viList of TablesTable 1.1 Gauge factors of some metals [2]. . . . . . . . . . . . . . . . . 4Table 1.2 Some conjugated conducting polymers, their bandgap energyand conductivities, adapted from [9]. . . . . . . . . . . . . . . 9Table 3.1 Aluminium beam dimensions. . . . . . . . . . . . . . . . . . . 35Table 4.1 Initial structure resistance . . . . . . . . . . . . . . . . . . . . 45Table 4.2 Summary of Experimental Longitudinal Gauge Factor of PE-DOT:PSS, Silver Paint and Commercial Strain Gauges . . . . . 53viiList of FiguresFigure 1.1 A cross section of a long wire showing the resistivity, crosssection and length, adapted from [6] . . . . . . . . . . . . . . 3Figure 1.2 A simplified representation of the muscle contraction sensor(MC) measuring principle for the determination of the me-chanical and physiological properties of skeletal muscles (1):Sensor tip; (2): Skin and intermediate layer; (3): Measuredmuscle, adapted from [47] . . . . . . . . . . . . . . . . . . . 7Figure 1.3 Simultaneously recorded force from the Muscle Contractionsensor FMC, force from a dynamometerFD and EMG duringisometric contraction of biceps brachii muscle of subject 3(n3), adapted from [47] . . . . . . . . . . . . . . . . . . . . . 7Figure 1.4 Chemical structure of PEDOT:PSS post polymerization of EDOTin polystyrene sulfonic acid, adapted from [38] . . . . . . . . 10Figure 1.5 The device structure of the PEDOT:PSS strain gauge, adaptedfrom [37] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Figure 1.6 Lift-off process of a PEDOT:PSS strain gauge , adapted from[21] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Figure 1.7 PEDOT:PSS thin film deposited on an array of electrodes on aPCB, adapted from [5] . . . . . . . . . . . . . . . . . . . . . 14Figure 1.8 PEDOT ink deposition images; evaporation cast PEDOT inkresulting in CNT film displacement on the left and inkjet printedPEDOT resulting in fine lines and absence of CNT film dis-placement on the right, adapted from [3] . . . . . . . . . . . . 17viiiFigure 2.1 The final structure of the strain gauge fabricated using: (a)Aerosol jet Printing (b) Peel off and Laser Micromachining(c) Etching by Laser Micromachining . . . . . . . . . . . . . 19Figure 2.2 Flowchart of the process of designing PEDOT:PSS strain gaugesusing aerosol jet printing technology . . . . . . . . . . . . . . 22Figure 2.3 Dimensions of the PEDOT:PSS strain gauge designed usingaerosol jet printing technology . . . . . . . . . . . . . . . . . 23Figure 2.4 Exposed VHB layer for PEDOT:PSS deposition after laser mi-cromachining. . . . . . . . . . . . . . . . . . . . . . . . . . . 23Figure 2.5 Deposition of PEDOT:PSS ink on the exposed area of VHBusing a regular lab syringe. . . . . . . . . . . . . . . . . . . . 24Figure 2.6 (a) A side view of layers of the sensor incorporating VHB as asubstrate. (b) A top view of the PEDOT:PSS dimensions. . . . 24Figure 2.7 Flowchart of the process of designing PEDOT:PSS strain gaugesusing laser machining and peeling-off. . . . . . . . . . . . . . 25Figure 2.8 Deposition of PEDOT:PSS on Kapton using a regular lab syringe. 26Figure 2.9 A side view of layers of the sensor incorporating Kapton as asubstrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Figure 2.10 Flowchart of the process of designing PEDOT:PSS strain gaugesusing laser machining etching. . . . . . . . . . . . . . . . . . 27Figure 3.1 The principle of operation of white light interferometry, adaptedfrom [33] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Figure 3.2 Tensile testing setup. 4-point measurement technique is usedby supplying the current and measuring the voltage producedby the strain gauge mounted on an aluminium beam. . . . . . 30Figure 3.3 Four point measurement method to avoid contact resistance. . 31Figure 3.4 cantilever beam and the same beam when deflected . . . . . . 32Figure 3.5 FEA of a deflected beam with dimensions shown in Table 3.1. 36Figure 3.6 Stress along the beam as a function of the length of the beam . 36Figure 4.1 (a) Tensile testing using the Bose ElectroForce tensilometer,and (b) a closer look showing the clamped substrate. . . . . . 38ixFigure 4.2 Controlled force applied to 2 layers of Kapton. . . . . . . . . 38Figure 4.3 The typical displacement measured simultaneously while theforce in applied. . . . . . . . . . . . . . . . . . . . . . . . . . 39Figure 4.4 Best fit stress vs strain curve of a double layer Kapton substrate. 40Figure 4.5 Hysteresis curve of stress vs strain of VHB . . . . . . . . . . 41Figure 4.6 A dry printed PEDOT:PSS line using aerosol jet printer withsome missing areas. . . . . . . . . . . . . . . . . . . . . . . . 42Figure 4.7 (a) A 2-dimension topography picture showing the thickness ofthe printed PEDOT:PSS line using the aerosol jet printer and(b) a 3-dimension look into the same line. . . . . . . . . . . . 43Figure 4.8 (a) A 2-dimension topography picture showing the thicknessof the PEDOT:PSS laser micromachined structure and (b) a3-dimension look into the same structure. . . . . . . . . . . . 44Figure 4.9 (a) A 2-dimension topography picture showing the thicknessof the silver paint structure and (b) a 3-dimension look into thesame structure. . . . . . . . . . . . . . . . . . . . . . . . . . 45Figure 4.10 Displacement actuation of the aluminium beam at a frequencyof 0.1 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Figure 4.11 The force and voltage responses to a deflection of 5mm at (a)0.1 Hz and (b) 0.2 Hz. . . . . . . . . . . . . . . . . . . . . . 48Figure 4.12 The force and voltage responses to a deflection of 5mm at 0.1Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Figure 4.13 The longitudinal force and voltage responses of the silver paintstrain gauge to a deflection of 5mm at (a) 0.05 Hz (b) 0.1 Hzand (c) 0.2 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . 52Figure A.1 (a) The graphical user interface and (b) the implementation ofthe LabVIew Host VI . . . . . . . . . . . . . . . . . . . . . 68Figure A.2 The implementation of the LabView Target VI . . . . . . . . 69Figure A.3 The strain gauge mounted on the skin of the bicep in (a) relaxedstate, and (b) flexed state . . . . . . . . . . . . . . . . . . . . 71xFigure A.4 Simultaneous recording of the force (Fg), MC and EMG. TheFg and MC variables are normalised to the maximal value,adapted from [47] . . . . . . . . . . . . . . . . . . . . . . . . 72Figure A.5 The results of contraction (a) strain data, and (b) integrated EMG. 73xiAcknowledgmentsI would like to express my sincere gratitude to all the faculty, staff and fellow stu-dents who provided gracious support and advice throughout the way. In particular,I would like to thank my supervisor Dr. Edmond Cretu for all the help, advice,emotional and financial support throughout my masters. I also would like to thankDr. Jean-Sebastien Blouin for hugely contributing mentally and financially to myprogress. I am so grateful for Dr. John Madden and Dr. Mu Chiao for allowingme to use their labs equipment. I am very thankful for our lab manager Dr. AlinaKulpa for her assistance and insightful guidance during my work in the lab.I honestly could not have made it through all of this without the help and sup-port of some particular colleagues and friends in the program : Miguel Torres,Dr. Elie Sarraf, Dr. Mrigank Sharma, Daniel Au, Ahmad Sharkia, Siamak Moori,Carlos Gerrardo, Harrison Brown, Shirely (Lingyi Liu), Saquib Sarwar, MeisamFarajollahi, and Soheyl Kianzad.I can not express enough my gratitude for my family in Saudi Arabia and mydear friends here in Vancouver and around the world for the emotional supportthroughout my education. Thank you for believing in me and supporting me nomatter what.xiiTo my graciously invaluable family and friendsxiiiChapter 1IntroductionI have not failed. I’ve just found 10,000 ways that won’t work.— Thomas A. Edison (1931)The importance of measuring the structural properties of bridges, planes andor even human bodies has paved the way for the design of devices to sense suchproperties. Strain is just one significant indication of the structural properties andis heavily used in the environment around us. Strain gauges are sensors used tocharacterize the strain applied on any object from a building to our own muscles.Every muscle in our body incorporates such sensors that transform the strain in amuscle into a signal the brain can understand. The motivation of this thesis to pavethe way for the design of strain gauge sensor to be used to associate body strainto electromyogram (EMG) signals produced in muscles. While there are manymethods to measure strain, the design and characterization of a novel, yet cheap,material strain gauge is outlined in the thesis.1.1 Strain Gauges in BiomechanicsStrain is the measure of the relative change in dimensions and shape of an objectdue to internal or external force. When a certain stretching force is applied to anobject, it causes an increase in length of the object due to tensile strain. Also, whenthe object is compressed, the length of the object will reduce due to a compressivestress along the compression axis. An example of strain in the human body is when1a person uses the index finger to point at an object. The skin of the index fingerexpands (lengthens) due to the tensile strain applied by index finger muscles. Al-most every movement in the human body causes either an eccentric (lengthening)or concentric (shortening) contractions. Those contractions will produce a smallelectrical signal that is proportional to the neural activity. The signals can then bemonitored using an EMG amplifier. While EMG can provide physicians with a lotof information about the muscle activity, the force a muscle produces is dependenton the length of the muscle and the velocity of the contraction, which can not bereflected accurately by EMG alone [47]. Also, EMG amplifiers are expensive, asthey have to be certified for human use due to the chance of electrical shock.Strain gauges are devices that are able to measure strain, or stress for that mat-ter. Typical strain gauges have a spiral-like shape of a metallic wire placed on aninsulating stretchable substrate. When the gauge is stretched axially, the dimen-sions (increase in length in most cases) of the gauge changes and as a result aresistance change is induced. When gauge is stretched transversally, the area (in-crease in width mostly) of the gauge changes and again, as a result, a resistancechange can be measured. While metal foil strain gauges are the most common,some other material strain gauges such as semiconductors or organic polymers areemerging.Strain measurements can potentially complement EMG measurements to pavethe way for many biomedical applications including haptic interfaces and smartprostheses.1.1.1 Piezoresistivity and Gauge FactorAs mentioned before, the change in the mechanical dimensions causes a change inthe electrical resistance in strain gauges. This effect is known as piezoresistivity.The first part of the word piezo comes from Greek, which means press or compress.If we consider a long wire with cross section A, length l and resistivity ρ as shownin Fig. 1.1, the electrical resistance R can be calculated asR = ρ lA(1.1)2R = ρ lpir2 (1.2)Figure 1.1: A cross section of a long wire showing the resistivity, cross sec-tion and length, adapted from [6]A change in resistance divided by the initial resistance results in the following∆RR=∆ll−2∆rr+∆ρρ (1.3)Now we can use the Poisson’s ratio v, defined as the negative ratio of the strainin the transversal εt direction to the strain in the longitudinal εl direction, asv =−εtεl(1.4)or∆rr=−v∆ll(1.5)As a result, we can write Eq. 1.3 as shown∆RR= (1+2v)∆ll+∆ρρ = (1+2v+piE)ε (1.6)where pi is the piezoresistive coefficient. Eq. 1.3 can be also written as∆RR= (∆RR)l +(∆RR)t = GFlεl +GFtεt (1.7)where GFl , GFt ,εl ,εt are the longitudinal gauge factor, transversal gauge factor,longitudinal strain, and transversal strain, respectively.In metal foil strain gauges, the term (1+ 2v) in Equation 1.6 represents thegeometrical effect and is dominant, as the change in intrinsic piezoresistivity term3Table 1.1: Gauge factors of some metals [2].Metal Gauge FactorCopper 2.2Constantan 1.9Nickel 2.7Platinum 2.440% gold/palladium 1.9∆ρρ is small [12]. When a metal is pressed, the change of the mobility in electronsdue to the volume change will cause a small change added to the geometrical effectof the gauge. Hence, typical values of the gauge factor of metals are positive and inthe range of 1−2.7. Examples of gauge factor values of some metals representedin Table 4.1.For an isotropic material, the piezo resistive effect is given by∆ρxρx= pixσx (1.8)where σx is the stress and is related to the strain by Young’s modulus E accordingto Hook’s lawσ = Eε (1.9)In order to describe the effects of stress of all dimensions, including normaland shear stress of an anisotropic material, the piezoresistivity coefficients must bea tensor such that∆ρiρ0=∑jpii jσ j (1.10)In Equation 1.10 the index j of the stress term σ encompasses both the three normaland three shear components of stress. Similarly, the relationship of voltage andcurrent as expressed in Ohm’s law depends on the direction of the electric field andhas 6 components reflected in the index i. In a matrix form, Equation 1.10 can be4written as∆ρ1ρ0∆ρ2ρ0∆ρ3ρ0∆ρ4ρ0∆ρ5ρ0∆ρ6ρ0=pi11 pi12 pi13 pi14 pi15 pi16pi21 pi22 pi23 pi24 pi25 pi26pi31 pi32 pi33 pi34 pi35 pi36pi41 pi42 pi43 pi44 pi45 pi46pi51 pi52 pi53 pi54 pi55 pi56pi61 pi62 pi63 pi64 pi65 pi66×σ1σ2σ3σ4σ5σ6The number of independent piezoresistive components represented in the matrixcan be reduced for anisotropic materials due to symmetry. Silicon crystals haveonly 12 non-zero elements in the matrix and only three independent components,which are: pi11, pi12 and pi44. Only the piezoresistive coefficients in the longi-tudinal, which is parallel to the direction of current flow, and transversal, whichis perpendicular to the current flow, are considered in most applications [25]. Insemiconductor based strain gauges, the term ∆ρρ is much more dominant than thegeometrical effect. Increasing the gauge factor of strain gauges is very desirable asthe gauge factor represents the sensitivity of the sensor. Due to the high intrinsicpiezoresistivity of semiconductor strain gauges, the gauge factor of such devicesis approximately two order of magnitude higher than metallic gauges [42] [19].While semiconductor strain gauges have a relatively high gauge factor in com-parison to metallic foil gauges, they have much greater sensitivity to temperature-making them less reliable in applications where temperature variation is expected.Add to that, semiconductor strain gauges have a nonlinear relationship of the stressto resistance, but this drawback can be easily dealt with in today’s advanced com-puter systems [12].1.1.2 Electromyography (EMG) and Strain MeasurementsElectromyography (EMG) is a valuable technique to analyse and understand theelectrical activity in a muscle or a group of muscles. The EMG signal is generatedin the muscle fibres during a contraction of the muscle. The actual source of theelectrical signal comes from the depolarization process that happens in the mus-cle during a contraction and is separated from the recording electrode by layers of5tissues [29]. As mentioned earlier, while EMG can provide practitioners with alot of information about the muscle activity, EMG alone does not reflect the mus-cle force precisely. Tension measurements, in addition, can possibly complementthe information given by EMG about the muscle force [47]. Some studies havedemonstrated the potential of using strain measurements and correlating the resultswith EMG data [47] [48]. In 1952, a study of the relationship between EMG andtension of the bicep brachii in humans indicated a parallelism of the tension to theintegrated EMG signal [18]. While the first study used a dynamometer to measurethe tension produced by the isometric contraction, it did not incorporate a straingauge on site of measurement. Interest in measuring selectively the force of mus-cles in humans or animals led to implantation of strain transducers in animals atfirst. A study that used implantation of an EMG amplifier and a mercury-basedstrain gauge in rats demonstrated clearly a direct relationship of strain measure-ments and integrated EMG signal during lordosis reflex in rats [36]. However,EMG alone can not provide precise information, as it is the sum of motor unitspotentials, while the generated force also depends on the muscle size, and the fir-ing rate [48]. Precise measurements of force generated in muscles and tendons ledto surgically planting a strain gauge in a human achilles tendon for accurate forcemeasurements. The implantation of such sensors was made possible by followingthe same procedure done in cats earlier. The purpose of the study was to investi-gate clearly the role of the achilles tendon in jumps and to study directly the elasticforce generated and that was possible through the data provided by the implantedmetallic foil strain gauge [14] [13].While implantation of devices in humans can provide direct and accurate mea-surements, this method is cumbersome due to its invasiveness. A group in Sloveniahas developed a strain gauge based sensor to selectively and non-invasively mea-sure the muscle force [47] [48]. The development of high sensitivity and inexpen-sive strain gauges promises a bright future in non-invasive biomechanics measure-ments. The designed sensor setup and the results of the strain sensing of the bicepbrachii from [47] is shown in Fig. 1.2 and Fig. 1.3, respectively.6Figure 1.2: A simplified representation of the muscle contraction sensor(MC) measuring principle for the determination of the mechanical andphysiological properties of skeletal muscles (1): Sensor tip; (2): Skinand intermediate layer; (3): Measured muscle, adapted from [47]Figure 1.3: Simultaneously recorded force from the Muscle Contraction sen-sor FMC, force from a dynamometerFD and EMG during isometric con-traction of biceps brachii muscle of subject 3 (n3), adapted from [47]71.2 Conductive PolymersAs the demand for flexible organic electronics has been increasing in the past fewyears, there is a need for investigating organic materials that have the potential tobe used in organic sensor designs. Most of these materials embrace conductiveproperties that qualified them to substitute conventional metals in some applica-tions. Organic light emitting diodes, organic solar cells and organic transistors aresome of the key developments using conductive polymers [35] [43] [45].The field of organic electronics has firstly seen the light in middle of the twen-tieth century. Development in the field has led 3 researchers, Alan J. Heeger, AlanG. MacDiarmid and Hideki Shirakawa, to win the Nobel prize in chemistry "forthe discovery and development of conductive polymers" [41].Conjugated polymers (conductive polymers) have long chains of carbon withsingle and double bonds alternating. The extended pi-orbital system in the polymerallows electrons (or holes) to move from one end of the chain to the other [9] [34].Table 1.2 shows some examples of conjugated polymers, their bandgap energyand highest reported conductivities. It should be noted that the bandgap energy ofsilicon is 1.1eV . It can realized that the availability of p-orbitals is an importantfactor for the polymer to be intrinsically conductive, as this will provide an orbitalsystem for the charge carries to move along the polymer [9].8Table 1.2: Some conjugated conducting polymers, their bandgap energy andconductivities, adapted from [9].Organic conductors posses a few advantages over conventional materials. Theadvancement of such materials has been built upon the affordability and the flexi-bility of the conductive polymers [37]. Also, since the chemical structure of suchmaterials can be easily modified, a range of properties of the material can be pre-cisely tuned. For example, the doping of the polymer determines the conductivityof the material. While organic polymers are not normally conductive by nature,the process of transforming a polymer incorporates either partial oxidation withelectron acceptors, such as AsF5, or partial reduction with electron donors, suchas Na. While the doping process to transform a polymer from an insulator to aconductor is not simple, it adds the tunability to the conductivity of the material[9] [15]. Altering the chemical structure of a conductive polymer can also addbio-compatibility to its properties, which make conjugated materials favorable inbiomedical applications [37].Another important property of conjugated polymers is the ability to dilute thepolymer solution for better transparency, a feature that is highly desirable in op-toelectronics applications. As a result, there has been a lot of interest in the useof conjugated polymers in light emitting diodes and touch screen applications [37][35] [8].9While there are so many advantages of conductive polymers, there are stillsome challenges in the development of such materials. The life time expectancy oforganic materials after being exposed to environmental aspects posses the biggestchallenge in commercializing organic electronics devices. Packaging and encapsu-lation of organic devices can provide a promise to the evolution of organic devices.Another drawback of conjugated materials lay in the performance of such devicesin comparison to the mature metal or semiconductor devices. While the perfor-mance is lagging behind the mature microelectronics technology, organic electron-ics can still complement the market with cheap and disposable devices [24] [37].1.2.1 PEDOT:PSSPoly(3,4-ethylenedioxythiophene) poly(styrenesulfonate), also commonly knownas PEDOT:PSS, is an intrinsically complex conductive polymer widely used inorganic electronics and organic photovoltaic (OPVs). PEDOT is one of the deriva-tives of thiophenes and is a conjugated polymer extracted from ethylenedioxythio-phene (EDOT) monomers in the reaction shown in Fig. 1.4 [38]. While PEDOTis the conductive part of the solution, it is normally doped with PSS in order toincrease its solubility. [10]. Other dopants used with PEDOT including Tosy-late and phosphomolybdate [37]. The conductivity of PEDOT:PSS lies in a netFigure 1.4: Chemical structure of PEDOT:PSS post polymerization of EDOTin polystyrene sulfonic acid, adapted from [38]positive charge on the PEDOT chain that attracts the negative charge left on the10acid. As a result, PEDOT and PSS are closely connected the attraction betweenthe left positive charges on the PEDOT chains attract the negative charges on thePSS chains. Once the PEDOT:PSS compound is bonded, an unpaired pi electron ishighly mobile on the chain leading to high conductivity. PEDOT:PSS is availablecommercially under the trade name CleviosTM from Heraeus or Orgacon TM fromAgfa. The polymer was initially used as a preventing layer for static charge inphotographic films. Agfa coats about 200 million films a year with a thin layer ofPEDOT:PSS to prevent electrostatic discharge during the production of films [37].PEDOT:PSS has many advantages in comparison to other types of syntheticpolymers. The low oxidation potential and the moderate bandgap of PEDOT:PSSmake the polymer more attractive for organic electronics applications. The bandgapof PEDOT:PSS can vary, depending on the preparation of the solution, in the rangeof 1.6eV to 2.5eV , with a work function in the range of 5.2eV [38] [11]. Whileenvironmental stability of doped PEDOT is a key feature when compared to otherconducting polymers, it can also be transparent with relatively high conductivity-making it an economically beneficial choice for optoelectronics applications.Due to the flexibility of PEDOT:PSS, many attempts have been made to designelectromechanical sensors using this material. Strain sensors based on conductivepolymers can resist a high strain range, higher than metals and semiconductors,while still being relatively conductive [39].Fabrication of PEDOT:PSS sensors are so far based on traditional clean-roomprocesses. Photolithography was used by Mateui et al. to produce a PEDOT straingauge sensor on silicon substrate [28]. The use of traditional fabrication methodscan hinder the development of such technology, as it requires long processing timeand higher costs. Other alternative fabrication methods such as slot die coating,spin coating, inkjet or screen printing were reported [1] [38].1.2.2 PEDOT:PSS Based Stress SensorsPiezoresistivity and various patterning techniques of PEDOT:PSS have been re-ported in the literature. PEDOT:PSS can be patterned using conventional clean-room photolithography and polymerization processes, but this could be problem-atic as the fabrication process is expensive for small quantities of sensors. To the11author’s knowledge, the first attempt to characterize the piezoresistivity of PE-DOT:PSS used spin coating techniques to deposit a thin layer of PEDOT:PSS onKapton polyamide film. The layer deposited thickness was approximately 50nmat 4000rpm. After the deposition of the PEDOT:PSS layer, gold or silver contactswere deposited using shadow masking. The final device structure is shown in Fig.1.5 .Figure 1.5: The device structure of the PEDOT:PSS strain gauge, adaptedfrom [37]After characterizing the longitudinal piezoresistive coefficient, they concludedthat the measurable resistance is partially cancelled out by the geometrical effect,resulting in a gauge factor in the range of −1.1 to 0.3 [37]. One of the main issuesof this work was the initial resistance of the device was in the range of Mohm,making it hard to detect accurately the change in resistance.A group from Technical University of Denmark fabricated PEDOT strain gaugesusing conventional UV-lithography and reactive ion etching (RIE) processes andreported a gauge factor of 3.41 ± 0.42. The group showed the gauge factor ofPEDOT is comparable to metal and thus the polymer can be used in all polymerMEMS-based devices [28]. In 2008, another group presented a lift-off processfor depositing thin film PEDOT:PSS and showed a proof-of-concept test results ofpiezoresistivity of PEDOT:PSS [20]. They showed a new fabrication technique ofPEDOT:PSS incorporating lift-off as illustrated in Fig 1.6.12Figure 1.6: Lift-off process of a PEDOT:PSS strain gauge , adapted from [21]The method consists of depositing polyamide on a silicon wafer followed bypatterning photoresist- exposing the areas of interest for the PEDOT:PSS deposi-tion using spin coating. After PEDOT:PSS is deposited, the wafer if flipped in(d) and another layer of photoresist is deposited for dry etching under the areas ofPEDOT:PSS. They then characterized the pressure gauges designed and reported agauge factor of 0.48 ± 0.07 at 36.6 pm 3 relative humidity [21].PEDOT:PSS was fabricated elsewhere using a peeling technique, which incor-porates clean room processes such as etching and spin coating, and then character-ized using micro-bending. The results of the bending showed a high gauge factorof 17.8 ± 4, which is well above aforementioned results of PEDOT:PSS and othermetallic strain gauges [23].The range of operation and the reliability are two important factors in straingauges. As shown in [39], a flexible foam structure was used as a substrate forthe PEDOT:PSS to study the reliability of the material. A thin layer of adhesivepolyurethane (PU) was deposited at first to make the porous foam hydrophobicfollowed by spin coating of PEDOT:PSS. Laser micromachining was then usedto pattern the structure. It was shown that PEDOT:PSS can withstand a displace-ment of 17.7% for 60 cycles, which is much higher than the range of operation ofmetallic and semiconductor strain gauges.PEDOT:PSS has been of interest for large area tactile sensors to be used inrobotics applications. The main reason for choosing this polymer is the cost ofproduction drops significantly when compared to metals such as silver. Also, the13flexibility and ease of processing of PEDOT:PSS contribute significantly to thispolymer to be the first choice for new generations of sensors. In [5], PEDOT:PSSsheet was deposited using spin coating on an array of electrodes on a printed circuitboard (PCB) as shown in Fig. 1.7. When a a pressure is applied on the array, thecontact between the electrodes and the sheet increases-causing a detectable changein resistance.Figure 1.7: PEDOT:PSS thin film deposited on an array of electrodes on aPCB, adapted from [5]The flexibility of PEDOT:PSS has led to integrating the polymer as a sensingpiezoresistive sensor in clothing fabrics. Inkjet printing of PEDOT:PSS into thefabrics were shown to cover individual fabrics to the yarn throughout the entirethickness of the cloth. The experiment concluded that PEDOT:PSS has a negativevalue gauge factor in the range of −5 to −20, which compares to conventionalstrain gauges. Preliminary data of PEDOT:PSS incorporated in fabrics sensorsshow that such sensors can be used to monitor knee and wrist motions, whichpromises to be the future for rehabilitation applications [4] [7].The use of PEDOT:PSS mixed with other materials have also been reportedin the literature. The aim of doing so is to improve either the mechanical prop-erties of the blend or to add conductivity to it. In a paper characterizing the me-chanical properties of PEDOT:PSS mixed with Polyvinyl alcohol (PVA) , it wasshown that adding PEDOT:PSS to the mixture improves the Young’s modulus ofPVA from 0.0412GPa to 1.65GPa (50 wt% of PEDOT:PSS). While mixing PE-14DOT:PSS with PVA improved the tensile strength of the nanofibres, the conductiv-ity was hindered. The study suggested addition of 30%-40% of PEDOT:PSS for areasonable trade off of the mechanical to electrical properties [16]. Another studyincorporated this mixture in a design of a strain gauge sensor by electrospinningof PEDOT:PSS/PVA on a flexible substrate. The device demonstrated in the pa-per exhibited high sensitivity to very small deformation, fast response and a gaugefactor up to 396 [26].The trend in sensor technology tends to move towards all polymer integratedsensors. The addition of PEDOT:PSS as either part of a mixture or as an electrodematerial has been reported in other studies. In [40], PEDOT:PSS electrodes sand-wiched porous Polyniline doped with camphorsulfonic acid (PANI:HCSA). Whenexternal pressure is applied on the device, the resistance reduces due to the increasein the contact between the PEDOT:PSS electrodes through the porous mixture.While many attempts in characterizing PEDOT:PSS were taken in the litera-ture, the development of PEDOT:PSS strain gauges using cheap and rapid alterna-tive fabrication techniques has not yet been explored and evaluated appropriately.1.3 MotivationThe development of a conjugated polymer strain gauge has been the interest in re-search for the past decade. Applications varying from skin-mountable strain mea-surements to structural monitoring require flexible, yet sensitive sensors. Interestin monitoring joint movements in human has led to designing highly stretchablestrain gauges [27]. Also incorporation of such devices into fabrics for applica-tions such as rehabilitation paved the way for the development of smart textiles [4].While the flexibility of the material can be a determining factor is some biomedicalapplications, less flexibility with higher sensitivity have other applications in thefield. As mentioned earlier, PEDOT:PSS has many advantages over other syntheticpolymers due its relatively high conductivity, flexibility and transparency.Metal foil strain gauges have been used mainly for structural monitoring. Twoof the important requirements for biomedical strain sensing applications are: flex-ibility and biocompatibility. Metallic foil gauges have the ability to only measuremicro strain- not a suitable choice for joint measurements as the range of strain is15large [30]. However, customized printable metallic based strain gauges can possi-bly be used in some biomedical applications not requiring a high range of opera-tion. Semiconductor strain gauges have a large gauge factor, but the nonlinearityand the flexibility of the sensors present a problem when considered in biomedicalapplications.The PEDOT:PSS conduction mechanism in the sense that the alignment inthe chains might contributes to its piezoresistivity. That is, when the polymer isstretched, the alignment of chains can get distorted, leading to a change in the re-sistivity of the material. While many attempts where taken to develop PEDOT:PSSstrain gauges, commercialization of such devices require more research. Many ofthe developed PEDOT:PSS strain gauges reported seem to have different gauge fac-tors. The reported longitudinal gauge factors for PEDOT:PSS strain gauges rangefrom −20 to 17.8 depending on the fabrication and the characterization methods.Also, no data available in the literature regarding the transversal gauge factor of PE-DOT:PSS. Extracting the longitudinal and transversal gauge factors of PEDOT:PSSis a very important step towards commercializing PEDOT:PSS based strain gauges.The main goal of the thesis is to explore alternative microfabrication methodsfor the patterning of conductive layers (polymers or metallic) on flexible substrates.Strain gauge designs have been used as target, to explore the spread claims regard-ing the piezoresistivity of PEDOT:PSS, and to compare polymer and metallic inkbased strain gauges.1.4 Previous WorkThis thesis is a continuation of work started initially by developing PEDOT:PSSstrain gauges using inkjet printing. In [1], the initial goal was to develop cheapstrain gauges based on silver nanoparticles inks. Throughout the work, the rel-atively inexpensive PEDOT:PSS ink replaced the silver nanoparticles. The the-sis work showed the possibility to fabricate strain sensors based on cheap ma-terials using nonconventional cleanroom processes. While the thesis work hasclearly showed the potential of using inkjet printing technology for patterningstrain gauges, it identified problems with using this printing method. Limitedviscosity of the inks, compatibility of the ink with the printing methods, and the16roughness of the printing surfaces were some of the challenges and limitationsfound in the previous work. While it was reported that PEDOT:PSS does not holdpiezoresistive properties after inkjet printing, an experiment setup was designed toextensively study the piezoresistivity of PEDOT:PSS. The thesis work reported agauge factor of a PEDOT:PSS patterned sensor on Kapton to be 3.63 [1].The direction of the work switched from using PEDOT:PSS as the base mate-rial to synthesizing evaporation cast thin film carbon nanotubes (CNT) as a basematerial for strain gauges. In [3], the motivation of the thesis was to design CNTbased strain gauges for biological and structural health monitoring. CNT havebeen reported to have high gauge factors, in the range of 600 to 1000, due to theintrinsic piezoresistivity [31]. In this thesis work, CNT were incorporated into var-ious evaporation cast films in order to align the tubes during evaporation. Inkjetprinting with air flow evaporation casting were used in hope to achieve alignmentof the CNTs. PEDOT:PSS was used as a conductive medium to the CNT film tothe vacant space of the substrate. Initial deposition of PEDOT:PSS on the CNTfilm resulted in displacement of the CNT ink as shown in Fig. 1.8. As a result,it was essential to use an inkjet printer to deposit small and controlled amount ofPEDOT:PSS that would sit on top of the CNT film.Figure 1.8: PEDOT ink deposition images; evaporation cast PEDOT ink re-sulting in CNT film displacement on the left and inkjet printed PEDOTresulting in fine lines and absence of CNT film displacement on theright, adapted from [3]Tensile measurements of the CNT film was performed and resulted in a gauge17factor ranging from 0.1 to 4. The considerably low value of the gauge factor wasbelieved to be due to the lack of alignment of the CNTs embedded into the thinfilm.1.5 Chapters OverviewThe thesis is divided into a four chapters outlining the fabrication, characteriza-tion, results along with a discussion, and conclusions with future outlook. The nextchapter introduces the microfabrication techniques used to make the PEDOT:PSSstrain gauges. It also includes a brief comparison of 3 differently fabricated PE-DOT:PSS sensors. The characterization chapter focuses on the methods undertakento design the characterization setup including white light interferometry and ten-sile testing. After establishing the ground of characterization, the results chaptershows the obtained data with more analysis of the results. A brief comparison ofthe results of the different fabrication techniques used is outlined then. Finally,conclusions are presented of the work done in this thesis with some future outlookand potential ideas to improve the design of such sensors.18Chapter 2FabricationIn order to extract the piezoresistive coefficients of PEDOT:PSS, different fabrica-tion technologies were investigated to prepare PEDOT:PSS strain gauges. Flexiblepolymer based substrates were used to prepare the structure for the experiment.The design of the sensor was done by either aerosol jet printing or laser micro-machining. Silver paint and copper tape were used to form metal contacts to theresistive structures. The final design of the PEDOT:PSS-based devices is shownin Fig. 2.1. Laser micromachining etching is also used to design silver ink-basedstrain gauges.Figure 2.1: The final structure of the strain gauge fabricated using: (a)Aerosol jet Printing (b) Peel off and Laser Micromachining (c) Etch-ing by Laser Micromachining192.1 SubstrateTwo different substrate materials were considered in the design of the PEDOT:PSSstrain gauges. The first choice was polyamide, or Kapton R©, from DuPont. Polyamideis well studied in the literature and it has very robust mechanical properties- mak-ing it an excellent choice for a straining application [32]. Kapton R© is used inapplications requiring a large range of temperature stability, ranging from −269◦Cto 400◦C. The stress-strain relationship and the elastic modulus of the Kapton tapewere taken from the datasheet from Dupont website.PEDOT:PSS strain gauges were also designed on the high strength elastomerbonding tape, VHBTM from 3MTM. The reason of considering the VHB tape is thehigh flexibility it provides- making it extremely easy to stretch the PEDOT:PSSsensor. The elastomer has been reported to have a range of flexibility exceeding100% of its original length [17]. The elastomer also has layer of adhesion, but aftera few experiments it was noticed that PEDOT:PSS films do not adhere to the tape.The other issue with VHB is that it is viscoelastic- giving it hysteresis propertiesthat are not desired in strain gauges. Nonetheless, VHB was still used as a substratematerial for some samples as it will be mounted on a stiffer material, which forcesthe VHB to get back to its original length after stretching. Silver ink-based straingauges were only deposited in Kapton since the stretchability of VHB is too highfor a metallic material.2.2 Fabrication Methods of the PEDOT:PSS StrainGaugesThree alternative to cleanroom fabrication methods were adapted to design thePEDOT:PSS strain gauges, namely: Aerosol jet printing, laser micromachining andpeeling-off, and laser micromachining etching. Each of the fabrication techniquesis explained in detail in the following sections.2.2.1 Deposition Using Aerosol jet PrintingIn the fabrication of the PEDOT:PSS strain gauges, the Optomech Aerosol jetprinter (Optomech, USA) was initially used to pattern PEDOT:PSS, Agfa OrgacoTM20grade IJ-1005, purchased from Sigma Aldrich, lines. Silver paint (SPI, USA) isthen used to make electrical pads to connect the PEDOT:PSS lines to copper wiresfor further measurements. The principle of operation of the aerosol jet printer liesin atomizing the PEDOT:PSS ink using an ultrasonic atomizer, creating a denseaerosol composed of droplets. The aerosol is then transported to the depositionhead using a carrier gas. The sheath gas is then used to focus the aerosol on thesubstrate while the deposition head is patterning the structure. This process willresult in the desired structure printed on the substrate with a relatively high widthresolution. Kapton polyamide tape (DuPont, USA) is used as a substrate due itsflexibility and usability for printed electronics as mentioned earlier. The substrateis initially mounted on a glass slide to allow reasonable flatness during the print-ing process. The polyamide is cleaned using isopropanol and left to dry for 15minutes. In order to reduce the viscosity of the PEDOT:PSS before atomizing, thematerial was mixed with deionized water with a 1:1 ratio, making it easier to atom-ize using the ultrasonic atomizer. Once the PEDOT:PSS lines are patterned on thepolyamide, the structure is left to dry at room temperature for 30 minutes. Silverpaint is then applied and the structure was baked at 60◦C for 15 minutes. A flowchart of the process is shown in Fig. 2.2.21Figure 2.2: Flowchart of the process of designing PEDOT:PSS strain gaugesusing aerosol jet printing technologyAlthough the line width resolution of the aerosol jet printer is high, the width ofthe PEDOT:PSS lines patterned on the polyamide were increased to 400um in orderto decrease the initial resistance of the structure. The geometry of the structure isshown in Fig. Laser Micromachining and Peel-offThe design process of the PEDOT:PSS strain gauge incorporates laser microma-chining of a plastic mounted on a VHB elastomer layer. First, a small piece ofVHB is put on the adhesive side of Kapton. Then, laser micromachining cutting(Oxford Lasers, UK) is tuned to 355nm wavelength, 0.3mm/s translation platformspeed to output 400Hz pulses of 0.12mJ energy in order to cut the red plastic(mask) covering the VHB layer into the shape of the sensor as shown in Fig. 2.4.22Figure 2.3: Dimensions of the PEDOT:PSS strain gauge designed usingaerosol jet printing technologyFigure 2.4: Exposed VHB layer for PEDOT:PSS deposition after laser mi-cromachining.Once the mask is cut, the sensor area is removed, exposing the VHB substrate,and then 1mL of PEDOT:PSS is deposited on the mask, as shown in Fig. 2.5, usinga regular lab syringe.23Figure 2.5: Deposition of PEDOT:PSS ink on the exposed area of VHB usinga regular lab syringe.Afterwards, the structure is placed in a furnace for 3 hours at 60◦C to evaporatethe solvent in the PEDOT:PSS and left to anneal slowly to room temperature. Dueto the poor adhesion of the PEDOT:PSS to the VHB substrate, peeling off theplastic will cause some parts of the sensor to be come off. Therefore, another runof laser micromachining at a much less laser intensity (355nm wavelength, 1mm/stranslation speed, 400Hz pulses and 0.024mJ), is required to separate parts of thePEDOT:PSS from plastic mask. Now the mask is removed, silver paint and coppertape are used to probe the structure at 4 points, each of which is placed on a cornerof the structure. Finally, another layer of Kapton Tape is put on top of the structure,making the PEDOT:PSS layer sandwiched in between VHB and the Kapton. Fig.2.6 (a) shows a side view of the layers in the sensor and (b) the dimensions asviewed from the top.Figure 2.6: (a) A side view of layers of the sensor incorporating VHB as asubstrate. (b) A top view of the PEDOT:PSS dimensions.24The process of laser machining and peeling-off flowchart is shown in Fig. 2.7.Figure 2.7: Flowchart of the process of designing PEDOT:PSS strain gaugesusing laser machining and peeling-off.2.2.3 Laser Micromachining EtchingAnother approach to design ink-based strain gauges is to incorporate laser micro-machining etching. This method was considered due to some disadvantages of theVHB elastomer, discussed later in this thesis. For PEDOT:PSS-based gauges, thedesign process starts with depositing 3mL of PEDOT:PSS ink on the adhesive sideof Kapton tape in an area of 8mm height ×11mm width as shown in Fig. 2.8.25Figure 2.8: Deposition of PEDOT:PSS on Kapton using a regular lab syringe.The structure is then placed in a furnace oven set at 60◦C for 3 hours andannealed slowly to room temperature. After that, the laser was set to 355nm wave-length, 0.3mm/s translation platform speed to output 400Hz pulses of 0.12mJ en-ergy to etch out all the unwanted areas of PEDOT:PSS. Again, silver paint andcopper tape are used afterwards to electrically connect 4 electrodes to the sensor.Finally, another layer of Kapton is placed on top of the structure to ensure that theforce is transformed between the Kapton layers to the PEDOT:PSS structure. Aside view of the structure is shown in Fig. 2.9.Figure 2.9: A side view of layers of the sensor incorporating Kapton as asubstrateThe process of laser itching flowchart is shown in Fig. 2.10.26Figure 2.10: Flowchart of the process of designing PEDOT:PSS strain gaugesusing laser machining etching.For the silver ink gauges, the process is almost the same, excluding the anneal-ing process. The silver ink used dried at room temperature without the need offurnace. This is advantageous since the throughput of silver ink gauges is higher.27Chapter 3CharacterizationAfter the fabrication process of the PEDOT:PSS strain gauges sensors, it is impor-tant to understand the characterization methods required to identify the physicalproperties of the design. First, white light interferometry techniques are used toextract the topography of the structure. Then, tensile testing is used to study theeffect of the mechanical strain on the electrical resistance of the sensors. In orderto perform the mechanical testing of the strain gauges, a design of a beam for thesensors to be mounted on was required. This chapter covers the principles of whitelight interferometry, tensile testing, and beam design.3.1 White Light InterferometryIn order to characterize the topography of the PEDOT:PSS structure, Polytec MSA-500 Micro System Analyzer is used. The topography measurement the systemincorporates is based on white light interferometry with high speed electronics andsoftware for image processing. The principle of operation is basically splitting alight source into 2 parts, as shown in Fig. 3.1, the first of which goes to a referencemirror, and the other is incident to the test object [44].28Figure 3.1: The principle of operation of white light interferometry, adaptedfrom [33]When the distance changes between the sample and the interferometer, a changein phase of the reflected wave off the object changes at every depth of the sample incomparison with the reference beam. During the interference scan, a video camerais recording the interference patterns. Then, the software processes the video toacquire the topography of the structure.3.2 Tensile TestingIn order to accurately characterize the designed strain gauges, simultaneous elec-tromechanical tensile testing is required. The Bose ElectroForce R© tensilometer(Bose, USA) can either control accurately the displacement or the force applied tothe test object, while simultaneously reading electrical signals. The setup of theexperiment using the tensilometer available in Dr. Madden’s lab is shown in Fig.3.2.29Figure 3.2: Tensile testing setup. 4-point measurement technique is used bysupplying the current and measuring the voltage produced by the straingauge mounted on an aluminium beam.Initially, the strain gauges are clamped directly in the tensilometer. The testis done by controlling the load while reading the displacement. However, due tohysteresis issues, it was thought that mounting the strain gauges on a material witha high modulus of elasticity and low hysteresis profile can provide better measure-ments. Therefore, a design of an aluminium beam was required and is discussed indetail in the following section.In order to avoid contact resistance in the measurements, a 4-point measure-ment technique is incorporated. Electrical current was applied to the structureusing Keithley model 6221 current source (Keithley, USA) while measuring thevoltage drop in the strain gauge resistance. The method avoids the contact resis-tances, as the input impedance to the voltmeter (tensilometer) is very high, makingthe current flowing through the least resistive path (the strain gauge in this case) asshown in Fig 3.3.30Figure 3.3: Four point measurement method to avoid contact resistance.3.3 Beam DesignAs mentioned earlier, due to the hysteresis of the substrate materials and someslippage in the clamping of the sensor, mounting the strain gauges on a stiffer beamwith a low hysteresis is essential for precise measurements. While the dimensionsof the beam were determined by the physical dimensions of the tensilometer, it is ofgreat importance to understand the areas of stress exhibited on the beam. For thisreason, analytical calculations of the beam followed by finite element modellingwere carried out.3.3.1 Analytical ModelConsider a beam as shown in Figure 3.4. Since the value of interest of the stressis located on the top surface as the sensor will be mounted there, a computation ofthe strain on the surface is required. Assumptions made here are:31Figure 3.4: cantilever beam and the same beam when deflected(i) perfect adhesion between the substrate of the strain gauge and the cantileverbeam is made in this computation. That is, the strain of the cantilever beam is per-fectly transmitted to the strain gauge. If imperfect adhesion occured, the resultantstrain can be multiplied with 1−α , where α is a loss factor.(ii) the presence of the strain gauge on the surface of the beam does not influ-ence the elasticity of the cantilever beam, as the dimensions of the strain sensor aremuch smaller than the beam.From (i), when the cantilever bends, the strain sensor (resistor and substrate)will bend with the same radius of curvature r in mm, and for a small curvature:κ = 1r=dxdφ (3.1)The natural axis is at y= 0 and εx = 0. The height at the surface of the cantilevery= t2 of the strain gauge sensor is at y=t2 +ts, where t is the thickness of the beamand ts is the thickness of the strain gauge sensor in mm . The change in length Lcan be calculated as follows:32dL = (r+ y)dφ = rdφ + ydφ = dx+ ydφ (3.2)dx(y) = (r+ y)dφ + rdφ + ydφ = dx(0)+ ydx(0)r(3.3)where dx(0) is the natural section of the beam. The deformation can then becomputed as:dx(y)−dx(0) = ydx(0)r(3.4)Hence the strain as a function of the height is:εx(y) =dx(y)−dx(0)dx(0)=yr(3.5)The strain at the surface of the cantilever is:εx(t2) =t2r(3.6)The strain at the surface of the sensor is:εx(t2+ ts) =t2 + tsr(3.7)The radius of curvature is related directly to the bending momentum M(x) as:κ = 1r(x)=−M(x)EI(3.8)where E is Young’s modulus of elasticity in Pa and M(x) is:M(x) =−F(L− x) (3.9)I is the moment of inertia of the cross section area A, in mm4, as shown:I =∫y2dA =Wt312(3.10)The Von Mises equivalent stress σ ′ is equal to the normal stress σx. The bend-ing stress is given by33σx =M(x)t2I(3.11)Therefore, the stress σx can be defined asσx =6F(x−L)Wt2(3.12)The deflection of the beam δ in mm can be derived from Euler-Bernoulli rela-tionship and is given byδ = Fx2(x−3L)6EI(3.13)The maximum deflection occurs at the point load F and is given byδmax =FL33EI(3.14)So, the force can be written as:F =3EIδmaxL3(3.15)The strain εx at the surface of the cantilever is:εx(t2) =3tδmax(L− x)2L3(3.16)The strain εx at the surface of the strain gauge is:εx(t2+ ts) =3( t2 + ts)δmax(L− x)L3(3.17)As mentioned earlier, the dimensions of the beam cantilever were predeterminedby the dimensions of the tensilometer as shown in Table 3.1.34Table 3.1: Aluminium beam dimensions.Parameter Value [mm]Length L 80width W 40thickness t 1stress location x 10The maximum displacement (deflection) allowed by the tensilometer is 5mm.Hence, calculating the strain at the surface of the beam yields:εx(t2) =3(1mm)(5mm)(80mm−10mm)2(80mm)3= 0.00102 (3.18)Calculating the strain at the surface of the sensor yields:εx(t2) =3(1mm2 +100um)(5mm)(80mm−10mm)(80mm)3= 0.00123 (3.19)3.3.2 Finite Element Analysis of Beam DeflectionCOMSOL Multiphysics R©is a finite element analysis (FEA) software used to sim-ulate and visualize the beam with the dimensions as mentioned in Table 3.1. Thesimulation helps visualizing the stress on the surface of the beam in order to de-termine the optimal location for the strain gauges to be placed. Fig. 3.5 shows theplanar stress on the xy− plane of the beam when deflected by 5mm.35Figure 3.5: FEA of a deflected beam with dimensions shown in Table 3.1.Fig. 3.6 shows a graph of the stress along the beam when deflected by 5mm.This graph can be used to approximate the surface stress acting on the straingauges. Note that the beam in the simulation is clamped at 80mm, meaning thatthe ending at 0mm is free to move.Figure 3.6: Stress along the beam as a function of the length of the beam36Chapter 4Results and DiscussionThis chapter presents the experimental results of the designed structures. First, ten-sile testing of the substrates results are shown to determine the elastic modulus ofKapton and VHB. A brief comparison of the fabrication technologies are presentedwith topography measurements. Then, results of the piezoresistivity characteriza-tion of PEDOT:PSS and silver strain gauges are presented. Finally, a comparisonof the results with a commercially purchased strain gauge are presented with adiscussion of the results.4.1 Substrate Material ComparisonIn order to determine the mechanical properties of the substrates used for the straingauges, it is important to determine the elastic modulus. Tensile testing using BoseBiaxial ElectroForce is performed as shown in Fig. 4.1.37Figure 4.1: (a) Tensile testing using the Bose ElectroForce tensilometer, and(b) a closer look showing the clamped substrate.The purpose of this test is to determine the level of hysteresis in the chosensubstrates. Each of the substrates is clamped in the machine while the force iscontrolled and the displacement is measured simultaneously. The test was donethree times for each substrate in order to see any variations of the results in betweenthe tests. Fig. 4.2 shows the force applied on the 2 layers of Kapton and Fig. 4.3shows the change in displacement due to the force applied in one of the samples.Figure 4.2: Controlled force applied to 2 layers of Kapton.38Figure 4.3: The typical displacement measured simultaneously while theforce in applied.The thickness of the Kapton tape used as the substrate is 2.5um and the widthof the tape is 1.27cm . Knowing the thickness t and the width w of the substrate wecan calculate the stress applied as followsσ = FA=Fwt(4.1)where F is the force applied on the substrate A is the area. The stress σ can now beplotted as a function of strain ε to extract the tensile modulus of the material. Fig.4.4 shows the best fit stress vs strain curve of the double layer Kapton substrate.39Figure 4.4: Best fit stress vs strain curve of a double layer Kapton substrate.The slope of the graph is the elastic modulus of the material under test and iscalculated to be E = 1.3±0.08GPa.Similarly, tensile testing was done on the VHB substrate. Hysteresis stressvs strain curve of VHB is shown in Fig. 4.5. The curve clearly shows a lot ofhysteresis in the elastoviscous material. The time required for the substrate toreturn to its original length is not calculated in this work but reported elsewhere tobe up to 141s [22]40Figure 4.5: Hysteresis curve of stress vs strain of VHBWe can first notice that VHB requires much less force in order to produce thesame displacement as Kapton, this is due the flexibility of the elastomer. However,we see from the VHB figures that hysteresis is clearly presented.For this reason, Kapton was considered as the main substrate material used forthe design of the strain gauges. However, VHB was used later to mount the sensorson the skin for a biomechanics application of the designed strain gauges.4.2 Comparison of the Fabrication TechnologiesAs mentioned earlier, two fabrication techniques were used to fabricate the PE-DOT:PSS strain gauges- namely aerosol jet printing and laser micromachining.The two technologies are compared in terms of the thickness of the sensors pro-duced and resulting electrical resistances of the patterned structures. While theaerosol jet printed can achieve printing resolutions down to 10um in width, theprinting resolution depends on the material used, the atomization method, the sub-strate and the width of the nozzle. The smallest thickness of PEDOT:PSS linesachieved using the aerosol jet printed was approximately 70um, due to the viscos-ity of the material. It is worth noting that the PEDOT:PSS was diluted in DI waterwith 1 : 1 ratio prior ultrasonic atomization. The nozzle used to deposit the PE-DOT:PSS lines has a diameter of 200um. One of the main problems of diluting41PEDOT:PSS in water and depositing the lines using the aerosol jet printer is oncethe lines dry, some areas in the line does not have PEDOT:PSS content in them-causing the line to be non-conductive as shown in Fig. 4.6.Figure 4.6: A dry printed PEDOT:PSS line using aerosol jet printer withsome missing areas.For this reason and also to increase the thickness of the lines, in order to re-duce the initial resistance, multiple printing runs were required to achieve thickerlines. Fig. 4.7 shows the topography of a typical printed line on Kapton as a sub-strate using aerosol jet printing taken using the white light interferometry moduleof Polytec MSA-500 measurement equipment.42Figure 4.7: (a) A 2-dimension topography picture showing the thickness ofthe printed PEDOT:PSS line using the aerosol jet printer and (b) a 3-dimension look into the same line.As shown in Fig. 4.7 (a), the thickness of the line, after 3 runs at a depositionspeed of 1mms , is approximately 1um. Another issue of printing PEDOT:PSS usingthe aerosol jet printer is the temperature effect on the ink. The PEDOT:PSS ink isa solution dried inside the carrier tube during idle time, which in effect increasedthe tube pressure. To avoid this problem, the printing of the ink had to be done fastbefore the PEDOT:PSS start drying inside the tube. To avoid this issue, purgingthe tube after every printed line is required.Laser micromachining is another way used to fabricate both PEDOT:PSS andsilver paint strain gauges. The minimum reported resolution of the laser used is20um. However, in order to reduce the resistance of the structure, the width andthe thickness of the ink were increased by printing more lines on top of the existinglines. Again, Polytec white light interferometry was used to acquire topographymeasurements of both PEDOT:PSS and silver paint strain gauges as shown in Fig.4.8 and Fig. 4.9, respectively.43Figure 4.8: (a) A 2-dimension topography picture showing the thickness ofthe PEDOT:PSS laser micromachined structure and (b) a 3-dimensionlook into the same structure.44Figure 4.9: (a) A 2-dimension topography picture showing the thickness ofthe silver paint structure and (b) a 3-dimension look into the same struc-ture.From the topography measurements in Fig. 4.8, the uniformity of the PE-DOT:PSS thickness is shown with an average thickness value of approximately40um. The measured thickness of the silver paint structure is approximately 35umas shown in Fig. 4.9.The initial resistance of the each of the sensors built with different fabricationtechniques is measured and shown in Table 4.1. Note that due to the variabilityof the PEDOT content in the PEDOT:PSS solution and the deposition method, theinitial resistance value varies significantly.Table 4.1: Initial structure resistanceFabrication method Value [Ω] Material Thickness [um]Aerosol jet printing 149000 PEDOT:PSS 1±0.2Laser etching 344±130 PEDOT:PSS 40±5Laser etching 102±10 Silver paint 35±5454.3 Electromechanical Characterization of PEDOT:PSSStrain GaugesA PEDOT:PSS strain gauge was mounted on an aluminium beam in order to char-acterize the electromechanical characteristics of the sensor. As seen from the FEAmodel of the beam in Section 3.3.2 Fig. 3.6, the highest stress on the surface occursnear the fixed end of the beam. Hence, the sensor is mounted on the area with high-est stress. The area of interest in the beam was initially rubbed using isopropanolalcohol to clean the surface in preparation for the sensor mounting. Then, a smallamount of super glue adhesion is applied to the back of the sensor and laid on thealuminium beam surface. Once laid on the surface, a small finger pressure is ap-plied on the sensor for 5 minutes to ensure good adhesion between the surface andthe sensor. Finally, light copper wires were soldered on the copper tape pads of thesensor.After the sensor was mounted on the beam, the tensile testing is performed bycontrolling the displacement, while measuring both the the force and the voltage.The maximum displacement allowed by the tensilometer is 5mm. Consequently, si-nusoidal waves of varying frequencies with an amplitude of 0−5mmpk− pk wereapplied to the aluminium beam, as shown in Figure 4.10. While many different ac-tuating frequencies were attempted, two actuating frequencies of 0.1Hz and 0.2Hzwere chosen, as they show the most predictable behaviours.46Figure 4.10: Displacement actuation of the aluminium beam at a frequencyof 0.1 Hz.As mentioned earlier, 4-point measurement technique was adapted to avoidcontact resistances in the measurements. Current of values ranging from 1−10mA,depending on the structure’s resistance, are applied to the structure. The voltageread in the tensilometer using a 16-bit analog to digital (ADC) converter is ampli-fied by the following equation, as given in the tensilometer manualGain =6kRchosen+1 (4.2)The value of the resistance was chosen to be Rchosen = 2kΩ for a voltage gainof gain = 4. The force applied on the aluminium beam along with the voltageresponse to the change in displacement for the two actuating frequencies of 0.1Hzand 0.2Hz are shown in Figure 4.11 (a), and (b) respectively.47Figure 4.11: The force and voltage responses to a deflection of 5mm at (a)0.1 Hz and (b) 0.2 Hz.Note that the response of the sensor was not seen in the compression state,when the load is negative. Therefore, only the positive part of the signal was takeninto account for curve fitting purposes. As can be seen from the figures, the voltageclearly has the same period as the applied force period. From the variation of thevoltage, we can extract the gauge longitudinal gauge factor of the sensor. From thedisplacement and load results, we calculate the Young modulus of the beam with48the strain gauges mounted on it, as follows:E =FL33wt312 δmax=(6.1N)(80mm)33 (40mm)(1mm)312 (2.5mm)= 135.1GPa (4.3)The stress absorbed by the surface on which the strain gauge is mounted|σx|=6F(x−L)Wt2=6× (13.2N)(10mm−80mm)(40mm)(1.1mm)2= 114.55MPa (4.4)Hence, the strain on the surface is the devision of the stress by the tensilemodulusεx =σxE=114.55MPa135.1GPa= 0.00085 (4.5)As mentioned earlier, the longitudinal gauge factor G|| can be calculated asfollowsG|| =∆VV1εx=4.299V −4.293V4.293V10.00085= 1.6 (4.6)Now that we extracted the longitudinal gauge factors G||, we can extract thepiezoresistive coefficient as followsG|| = (1+2v+pi||Esubstrate) (4.7)where v is the poisson ratio of the substrate (Kapton) and is given from the ma-terials information sheet to be 0.33, pi|| is the longitudinal piezoresistive coefficient,and Esubstrate is measured in Section 4.1 to be 1.3GPa. Hence,pi|| =G||−1−2vEsubstrate=1.6−1−2(0.33)1.3×109=−4.6×10−11Pa−1 (4.8)The change in the resistance when the sensor is mounted transversally is verysmall. The transversal gauge factor G⊥and piezoresistive coefficient pi⊥ respec-tively can be calculated asG⊥ =∆VV1εx=4.290V −4.293V4.293V10.00085=−0.822 (4.9)49The change in voltage is very small. In fact, it is within the noise level of thesignal so the calculated value will be neglected.4.4 Commercial Strain GaugesThe results of the PEDOT:PSS based strain gauges are compared to a commercialgauge. The force and voltage response of the commercial strain gauge is shown inFigure 4.13.Figure 4.12: The force and voltage responses to a deflection of 5mm at 0.1Hz.As can be seen from the figure, the response of the commercial strain gaugeseems to be cleaner. However, the longitudinal gauge factor of the PEDOT:PSSof 1.6 is comparable to the calculated gauge factor of the commercial strain gaugeof 1.84. For this reason, it was thought that a metallic ink can be incorporated ina design of a strain gauge using the same fabrication method as fabrication of thePEDOT:PSS strain gauges.504.5 Electromechanical Characterization of Silver StrainGaugesSame approach of mounting the silver paint based strain gauge was taken in orderto characterize its piezoresistive properties. Displacement actuations at frequenciesof 0.05Hz, 0.1Hz and 0.5Hz were applied while measuring the force and voltage.The longitudinal response force and voltage at the three frequencies is representedin Figure 4.14 a, b and c, respectively.51Figure 4.13: The longitudinal force and voltage responses of the silver paintstrain gauge to a deflection of 5mm at (a) 0.05 Hz (b) 0.1 Hz and (c)0.2 Hz.The longitudinal gauge factor Gsilver|| and the piezoresistive coefficient pisilver||of the sensor is calculated to beGsilver|| =∆VV1εx=2.82V −2.79V2.79V10.0085= 12.6 (4.10)pisilver|| =Gsilver||−1−2vEsubstrate=12.6−1−2(0.33)1.3×109= 8.4×10−9Pa−1 (4.11)The transversal gauge factor Gsilver⊥ and the piezoresistive coefficient pisilver⊥of the sensor is calculated to beGsilver⊥ =∆VV1εx=2.775V −2.772V2.772V10.0085= 0.7220 (4.12)pisilver⊥ =Gsilver⊥−1−2vEsubstrate=0.7220−1−2(0.33)1.3×109=−7.215×10−10Pa−1(4.13)Again, the change of voltage (for constant current) when the sensor is mounted52laterally is very small. Therefore, the calculated value will be neglected.4.6 ComparisonThe results of both the PEDOT:PSS and silver paint based strain gauges are com-pared to a commercial gauge. The longitudinal gauge factor of the PEDOT:PSS of1.6 is comparable to the calculated gauge factor of the commercial strain gauge of1.84. It should be noted that the gauge factor value provided by the manufacturerof the commercial strain gauge is 2.1. The resulted experimental gauge factor,on the other hand, could have possibly suffered from the stress not being fullytransformed from the aluminium beam surface to the strain gauge itself. Since themounting procedure of all the strain gauges is the same, the assumption of smallloss in the stress from transforming completely to all the strain gauges should bevalid for all. The silver paint strain gauge has experienced a much higher gaugefactor of 12.6 in comparison to both the PEDOT:PSS and the commercial straingauges, but the non-smooth variation in the electrical resistance seems to indicatereversible disconnections of the conductive paths under larger stress.Table 4.2: Summary of Experimental Longitudinal Gauge Factor of PE-DOT:PSS, Silver Paint and Commercial Strain GaugesStrain Gauge Longitudinal Gauge FactorPEDOT:PSS 1.6Silver Paint 12.6Commercial 1.84It should be noted that due to the design of the strain gauges fabricated in thelab, the length of the strain gauge is maximized to increase the sensitivity of thesensors in the longitudinal direction, while the transitional areas are small to elim-inate the transversal loading effect on the sensors. Hence, both of the transversalgauge factors of the the PEDOT:PSS and silver paint based strain gauges shownon the table are very small. One of the main issues of the silver paint strain gauge,however, is the very small range of operation. Looking back at the resulted wave-form of the silver paint strain gauges, we see high spikes at higher values of stress.53This is due to reversible cracks in the silver lines, causing a sudden increase inresistance for larger stress, thus limiting the operating range. In order to avoid thisproblem, the silver based strain gauge can only operate at a smaller range.While the flexibility of PEDOT:PSS is considered a plus, the variability in thegauge factor results from one sample to another hinders the viability of the use ofthis polymer in sensing applications. The deposition method used in designing thePEDOT:PSS strain gauges may have caused random alignment of the PEDOT:PSSchains in the solvent. Therefore, it could be the reason in the high variability in theresults of the PEDOT:PSS strain gauges. It should be noted that while the nominalinitial resistances of each PEDOT:PSS structure varied greatly (in the range of 200ohm to 10 kohm), the response varied between one sample to another. Some sam-ples did not response to the strain at all, but rather the resistance stayed constant.The data did not vary in time for the same PEDOT:PSS sample for repeated testson the same day. The structures were not retested at different days to see if thereis a variations in the results between day to day. The data shown in this thesis isof one structure that responded in this manner. The silver ink strain gauges results,however, did not have high variability between samples, proving the capability ofthe laser itching use in the patterning of conductive inks.Another factor of comparison is the price of production. While the commer-cially purchased strain gauge costed $12 per gauge, the in-house fabrication ofeither the PEDOT:PSS or the silver point strain gauges costs approximately $3 pergauge. It should be noted that the price estimation of the in-house strain gauges isbased on the cost of the materials purchased in small quantities. If, however, massproduction is required, in bulk purchasing of the materials can potentially reducethe price down to $1 per gauge. Besides, the in-house fabrication time of the straingauges can take up to a day per ∼ 10 structures. It should be noted that the fabri-cation time depends on the curing time of the material. As a result, fabrication thesilver paint strain gauges is actually faster than the PEDOT:PSS strain gauges.Lastly, one of the main advantages of exploiting the alternative microfabrica-tion techniques in the design of the strain gauges as discussed in the this thesis isthe customizability of the design for specific applications. In order to utilize thestrain gauges to their maximum sensitivity, variable strain gauges shapes mayberequired. For example, strain sensing of human skins due to movements can bene-54fited from specified gauge shape designs to promote highest sensitivity.55Chapter 5Conclusions5.1 Thesis OverviewThe main motivation of this work was to develop custom designed structures (straingauges taken as target example) using alternative microfabrication techniques, basedon either conductive polymers (PEDOT:PSS) or metallic inks such as silver. Thegoal for developing such sensors is to incorporate them in applications rangingfrom joint monitoring in humans to integration of such sensors in fabrics or sport-ing attire.Initially, the idea of developing printable sensors was initiated by the acquisi-tion of new pieces of equipment in Dr. Edmond Cretu’s Adaptive MEMS labora-tory. It is reported in the literature in [37] [28] [20] [21] [23] [39] for PEDOT:PSSto have intrinsic piezoresistive properties. However, the reported gauge factors inthe literature vastly varied from one paper to another. In addition, to the author’sknowledge, there is no attempts to characterize the transversal gauge factor of PE-DOT:PSS. In order to claim intrinsic piezoresistivity of a material, it is thought thatit is important to characterize both the longitudinal and transversal force response.Also, the fabrication techniques used in the literature are mostly conventional cleanroom processes, meaning a non-rapid slow manufacturing of sensors to be studied.The first attempt in this work to design PEDOT:PSS strain gauges was doneby using the Sonoplot Microplotter. The principle of operation relies on dipping asmall glass tip in the solution, and due to capillary forces, the fluid fills the tube.56Then, through ultrasonic vibration in a piezoelectric element mounted on the glasstube, controlled amount of the solution is deposited on the material. Less successwas seen depositing PEDOT:PSS using the microplotter due to several reasons,including: the viscosity of the material, the non-uniformity of the surface, the PE-DOT:PSS content of the solution deposited. The deposition of PEDOT:PSS lineswas then successful using the aerosol jet printer. The use of the aerosol jet printerproved to be a very good alternative to deposit PEDOT:PSS lines, but due to thevery small thickness of the deposited ink, multiple deposition if required to achievea thicker structure. The process of multiple depositions required close looks intothe tube pressure of the equipment since the PEDOT:PSS dried in the tube, causingunexpected clogs. Finally, it was though that laser micromachining etching canprovide the best results of patterning when it comes to strain sensors. Thick layersof ink is thought to be required for flexible printed strain sensors as it is importantfor the lines not to break when bent. Silver paint is then fabricated using the samemethod of laser micromachining etching as an alternative material.Using laser micromachining, both PEDOT:PSS and silver paint strain gaugeswere patterned and characterized longitudinally and transversally using control-lable tensile testing. Four-point measurement methodology was adapted in orderto roll out contact resistances from the gauge factor measurements. Tensile test-ing was initially done on different substrates including Kapton and VHB. Due tothe great hysteresis found in VHB, Kapton was chosen as the main substrate ma-terial for the strain gauges. The in-house designed sensors were mounted on analuminium beam along with a commercially available strain gauge. While thedisplacement of the tensilometer was controlled, both the force applied on the alu-minium beam and the voltages read from the strain gauge resistors were read si-multaneously. The average resultant longitudinal and transversal gauge factors ofthe PEDOT:PSS strain gauges were 1.6 and -0.82, respectively, with a large vari-ation from sample to sample. On the other hand, the longitudinal and transversalgauge factors of the silver paint sensors were 12.6 and 0.722, respectively. Al-though the PEDOT:PSS strain gauges have a small longitudinal gauge factor, theyare very stretchable, making them suitable for applications requiring a large rangeof operation. In fact, it was reported in the literature a stretchability of PEDOT:PSSof 17%. One of the main issues, however, of the PEDOT:PSS strain gauges is the57inconsistency in the measurements. It is believed that is caused due to the align-ment of the chains in the synthetic polymer. As a result, different fabrications anddepositions techniques result in inconsistent response of the sensor. Nonetheless,the work in this thesis has shown the potential for using alternative deposition andpatterning methods to fabricate very cheap disposable sensors.The main contribution of this work was to establish methods of patterning con-ductive inks on flexible substrates. The deposition of PEDOT:PSS on the VHBelastomer is one of the few examples of the capabilities of the established alterna-tive fabrication technologies investigated in this thesis.The use of sensors designedusing the studied alternative fabrication technologies can be used in biomedicalapplications. An example application will be relating EMG to force measurementsby mounting a custom patterned force sensor on the skin of a person and seeing theeffects of flexing the muscle while monitoring EMG.5.2 Future OutlookIn this work, a great deal of attention on the investigation of the piezoresistivityof PEDOT:PSS was given while exploring different fabrication methods to depositdifferent kinds of inks on flexible substrates. While the thorough investigation ofPEDOT:PSS as a synthetic material to be used in strain gauges proved so far thatit might not be a suitable material in terms of both sensitivity and reproducibilityof the structures„ the fabrication methods investigated showed the potential to beused with other types of inks. Mixing the synthetic polymer with other organicmaterials and using the same fabrication methods can seems the next logical step.While PEDOT:PSS in fact seems to be affected only by geometrical changes, theconductivity, flexibility, and the ease of processability of the material are just a fewadvantages. It is reported elsewhere that a mixture of PEDOT:PSS with Polyvinylalcohol (PVA) increases the stiffness of the material and results in a gauge factoras high as 396 [16] [26].Another alternative to intrinsically piezoresistive sensors, the geometrical ef-fect can be incorporated in designing a capacitive sensing strain sensor. The ideais to pattern small capacitive combs on flexible material. The change of the over-lapping area of the capacitor combs is a result strain. A study published in 201458showed the potential for developing capacitive metallic strain gauges for wirelessmonitoring [46]. Power consumption is the main advantage of developing suchsensors, as operating principle relies on a sensing capacitance dependence on theapplied strain. While the published study uses metallic capacitive combs, the useof a synthetic polymer to design a capacitive strain gauge can prove to be a cheaperoption.Further research is also required for musculoskeletal characterization. WhileEMG monitoring has always proved to be a method to monitor a muscle groupactivity, EMG acts only as an input to the system. Hence, EMG alone does notgive the full picture about the force produced in the muscle. Strain measurementsalong with EMG can help in identifying the system. Further research is requiredto first identify most sensitive strained areas, for a strain sensor to be mounted, inresponse to a specific muscle group activation. Also, since strain measurementsare conducted on the skin of the subject, it is essential to monitor all the musclegroups around the area of interest.59Bibliography[1] H. Al-Chami. Inkjet printing of transducers. Master’s thesis, University ofBritish Columbia, September 2010. → pages 11, 16, 17[2] P. Andreucci, L. Duraffourg, C. Marcoux, P. Brianceau, S. Hentz,S. MINORET, E. Myers, and M. Roukes. Nems comprising alsi alloy basedtransduction means, May 23 2012. URL EP Patent App.EP20,100,731,526. → pages vii, 4[3] D. T. Y. Au. Evaporation cast thin film carbon nanotube strain gauges.Master’s thesis, University of British Columbia, August 2013. → pages viii,17[4] Y. Bar-Cohen, editor. Piezoresistive sensors for smart textiles, volume 6524,April 2007. → pages 14, 15[5] J. Castellanos-Ramos, R. Navas-GonzaÂt’lez, H. Macicior, T. Sikora,E. Ochoteco, and F. Vidal-Verdu. Tactile sensors based on conductivepolymers. Microsystem Technology, 16(5):765–776, May 2010. → pagesviii, 14[6] O. College. Resistance and resistivity, September 2013 (Oct. 14, 2014).URL → pages viii, 3[7] S. Coyle, Y. Wu, K.-T. Lau, D. D. Rossi, G. Wallace, and D. Diamond.Smart nanotextiles: a review of materials and applications. MRS Bulletin, 32(5):434–442, May 2007. → pages 14[8] G. P. Crawford, editor. Flexible Flat Panel Displays. John Wiley and Sons,Ltd, Chichester, UK., 1st edition, June 2005. → pages 960[9] L. Dai. Intelligent Macromolecules for Smart Devices. Springer, xvi edition,2004. → pages vii, 8, 9[10] A. Elschner, S. Kirchmeyer, W. Lovenich, U. Merker, and K. Reuter.PEDOT: Principles and Applications of an Intrinsically ConductivePolymer. Taylor and Francis Group, 1st edition, November 2010. → pages10[11] A. Elschner, W. Loevenich, A. Eiling, and J. Bayley. Ito alternative: solutiondeposited cleviosTMpedot:pss for transparent conductive applications, June2012 (Oct. 28, 2014). URL → pages 11[12] O. Engineering. The strain gage, 1996 (Nov. 3,2014). URL → pages4, 5[13] S. Fukashiro, P. V. Komi, M. Järvinen, and M. Miyashita. In vivo achillestendon loading’ during jumping in humans. European Journal of AppliedPhysiology and Occupational Physiology, 71(5):453–458, May 1995. →pages 6[14] R. Gregor, R. Roy, W. Whiting, R. Lovely, J. Hodgson, and V. Edgerton.Mechanical output of the cat soleus during treadmill locomotion: In vivo vsin situ characteristics. Journal of Biomechanics, 21(9):721–732, August1988. → pages 6[15] A. J. Heeger. Semiconducting and metallic polymers: The fourth generationof polymeric materials. The Journal of Physical Chemistry, 105(36),September 2001. → pages 9[16] C. hsiu Chen, A. Torrents, L. Kulinsky, R. D. Nelson, M. J. Madou,L. Valdevit, and J. C. LaRue. Mechanical characterizations of castpoly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)/polyvinyl alcoholthin films. Synthetic Matels, 161:2259–2267, January 2011. → pages 15, 58[17] L. Hu, W. Yuan, P. Brochu, G. Gruner, and Q. Pei. Highly stretchable,conductive, and transparent nanotube thin films. Applied Physics Letters, 94(16):161108, 2009. doi: URL →pages 2061[18] V. T. Inman, H. Ralston, J. D. C. Saunders, M. B. Feinstein, and E. W. W. Jr.Relation of human electromyogram to muscular tension.Electroencephalography and Clinical Neurophysiology, 4(2):187 – 194,1952. ISSN 0013-4694.doi: URL →pages 6[19] Y. Kanda. Piezoresistance effect of silicon. Sensors and Actuators A:Physical, 28(2):83–91, July 1991. → pages 5[20] U. Lang, P. Rust, and J. Dual. Towards fully polymeric mems: Fabricationand testing of pedot/pss strain gauges. Microelectronic Engineering, 85(5–6):1050 – 1053, 2008. ISSN 0167-9317.doi: URL of the Micro- and Nano-Engineering 2007 Conference {MNE}2007. → pages 12, 56[21] U. Lang, P. Rust, B. Schoberle, and J. Dual. Piezoresistive properties ofpedot:pss. Microelectronic Engineering, 86(3):330 – 334, 2009. ISSN0167-9317. doi: URL TheFourth {IEEE} International Symposium on Advanced Gate StackTechnoiogy (ISAGST 2007). → pages viii, 13, 56[22] B. Lassen, M. J. andC. Melvad, G. R. Kristjánsdóttir, and R. Jones.Hysteresis in dielectric electroactive polymers. In Proc. SPIE 7493, SecondInternational Conference on Smart Materials and Nanotechnology inEngineering, volume 7493, page 6. SPIE, October 2009. → pages 40[23] G. Latessa, F. Brunetti, A. Reale, G. Saggio, and A. D. Carlo. Piezoresistivebehaviour of flexible pedot:pss based sensors. Sensors and Actuators B:Chemical, 139(2):304 – 309, 2009. ISSN 0925-4005.doi: URL →pages 13, 56[24] Y. Li and C. Wong. Recent advances of conductive adhesives as a lead-freealternative in electronic packaging: Materials, processing, reliability andapplications. Materials Science and Engineering: R: Reports, 51(1–3):1 –35, 2006. ISSN 0927-796X.62doi: URL →pages 10[25] C. Liu. Foundations of MEMS. Prentice Hall Press, Upper Saddle River, NJ,USA, 2nd edition, November 2011. → pages 5[26] N. Liu, G. Fang, J. Wan, H. Zhou, H. Long, and X. Zhao. Electrospunpedot:pss–pva nanofiber based ultrahigh-strain sensors with controllableelectrical conductivity. Journal of Materials Chemistry, 21:18962–18966,November 2011. → pages 15, 58[27] N. Lu, C. Lu, S. Yang, and J. Rogers. Highly sensitive skin-mountable straingauges based entirely on elastomers. Advanced Functional Materials, 22(19):4044–4050, June 2012. → pages 15[28] R. Mateiu, M. Lillemose, T. S. Hansen, A. Boisen, and O. Geschke.Reliability of poly 3,4-ethylenedioxythiophene strain gauge.Microelectronic Engineering, 84(5–8):1270 – 1273, 2007. ISSN 0167-9317.doi: URL of the 32nd International Conference on Micro- andNano-Engineering. → pages 11, 12, 56[29] R. Merletti and P. J. Parker, editors. Electromyography: Physiology,Engineering, and Non-Invasive Applications. John Wiley and Sons, 1stedition, July 2004. → pages 6[30] M. Muraki, S. Takamatsu, K. Matsumoto, and I. Shimoyama. Organicsemiconductor based strain sensors for input system on flexible oleds. InMicro Electro Mechanical Systems, 2008. MEMS 2008. IEEE 21stInternational Conference on, pages 904–907, January 2008. → pages 16[31] W. Obitayo and T. Liu. A review: Carbon nanotube-based piezoresistivestrain sensors. Journal of Sensors, 2012, February 2012. → pages 17[32] M. Pecht and X. Wu. Characterization of polyimides used in high densityinterconnects. Components, Packaging, and Manufacturing Technology,Part B: Advanced Packaging, IEEE Transactions on, 17(4):632–639,November 1994. → pages 20[33] Polytec. Basics of white light interferometry, 2005 (Nov. 10, 2014). URL → pages ix, 2963[34] M. A. Rahman, P. Kumar, D.-S. Park, and Y.-B. Shim. Electrochemicalsensors based on organic conjugated polymers. Sensors, 8(1):118–141,January 2008. → pages 8[35] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, andK. Leo. White organic light-emitting diodes with fluorescent tube efficiency.Nature, 459:234–238, May 2009. → pages 8, 9[36] S. Schwartz-Giblin and D. W. Pfaff. Implanted strain gauge and emgamplifier to record motor behavior in unrestrained rats. Physiology andBehavior, 25(3):475–479, September 1980. → pages 6[37] T. M. Schweizer. Electrical characterization and investigation of thepiezoresistive effect of pedot:pss thin films. Master’s thesis, GeorgiaInstitute of Technology, April 2005. → pages viii, 9, 10, 11, 12, 56[38] L. Vanmaele. Conducting polymer materials for flexible opv applications:OrgaconTM pedot : Pss, 2011 (Oct. 28, 2014). URL →pages viii, 10, 11[39] V. K. Varadan, editor. Reliability of PEDOT-PSS Strain Gauge on foamstructure, March 2010. → pages 11, 13, 56[40] P.-C. Wang, W.-K. Lin, S.-Y. Hung, and H.-J. Lu. Fabrication andcharacterization of all-polymer pressure sensors integrated with transductionmodules based on conducting polymers. In Microsystems PackagingAssembly and Circuits Technology Conference (IMPACT), 2010 5thInternational, pages 1–4, October 2010. → pages 15[41] N. N. M. A. . Web. The nobel prize in chemistry 2000, 2000 (Oct. 27, 2014).URL →pages 8[42] A. Window, editor. Strain Gauge Technology, volume XIV. Springer, 2ndedition, 1993. → pages 5[43] D. Wöhrle and D. Meissner. Organic solar cells. Advanced Materials, 3(3):128–138, March 1991. → pages 8[44] J. C. Wyant. White light interferometry, 2002. URL → pages 2864[45] Y. Yamashita. Organic semiconductors for organic field-effect transistors.Science and Technology of Advanced Materials, 10(2), July 2009. → pages 8[46] R. Zeiser, T. Fellner, and J. Wilde. Capacitive strain gauges on flexiblepolymer substrates for wireless, intelligent systems. J. Sens. Sens. Syst, 3:77–86, April 2014. → pages 59[47] S. ÄRˇorÄS´evic´, S. Stancˇin, A. Meglicˇ, V. Milutinovic´, and S. Tomažicˇ. Mcsensor–a novel method for measurement of muscle tension. Sensors, 11(10):9411–9425, September 2011. → pages viii, xi, 2, 6, 7, 71, 72[48] S. ÄRˇorÄS´evic´, S. Tomažicˇ, M. Narici, R. Pišot, and A. Meglicˇ. In-vivomeasurement of muscle tension: Dynamic properties of the mc sensorduring isometric muscle contraction. Sensors, 14(9):17848–17863,September 2014. → pages 665Appendix ASupporting MaterialsIn this chapter, a potential application of designed strain gauges in presented. ThePEDOT:PSS strain gauge, due its flexibility, is chosen to monitor the deflectionin the skin around the bicep brachii, while integrated EMG is read off the musclegroup. The application presents the potential of dealing with the muscles responseas a system. That is, the input to the muscle is the EMG and the output in this case isdeflection. Also, this implementation of a strain gauge sensor on the body presentsthe possibility to integrate custom designed sensors for many human monitoringapplications including but not limited to: sensor integration in fabrics, and bodyarea networks. In this chapter, a brief explanation of the implementation of Na-tional Instruments (NI) LabView Virtual Instrument (VI) on a Field programmablegate array (FPGA) is described. Then, the experimental setup of measuring theintegrated EMG signal along with the strain measurements is presented. Finally,the measurement results are presented along with a discussion.A.1 LabView Virtual Instrument (VI) DesignThe NI LabView FPGA allows for a simple and rapid system level implementationfor verification of a system. The LabView FPGA module implements a graphicalprogramming technique that can be transformed into a low-level of abstractionthrough the LabView Embedded technology. The program design is implementedusing an FPGA module along with NI reconfigurable I/O (RIO) hardware. Due to66the high parallelism and flexibility LabView provides for FPGA programming, itis thought that using such system is ideal for this experiment.The system designed in LabView encompasses two Virtual instruments (VIs),namely: a host and a target. While the host encompasses all the signal processingand the graphical user interface (GUI), the target acts as an adaptive data acquisi-tion unit implemented on an FPGA. Fig. A.1 (a) shows the GUI and (b) the VIimplementation of the host and Fig. A.2 shows the implementation of the targetVI.67Figure A.1: (a) The graphical user interface and (b) the implementation ofthe LabVIew Host VI68Figure A.2: The implementation of the LabView Target VIThe communication between the host and the target VIs is done through a directmemory access (DMA) first-in-first-out (FIFO). This allows the target VI to use thehost RAM directly, allowing for a major improvement in the speed of acquisition.The FIFO used has two parts, the first of which is in the target and is used to savethe data read though the IO channel, and the second is in the host, which is usedto read out the data when the FIFO is full. The sampling rate of data acquisitionis specified by the loop rate of the target VI. For this experiment, it is specified at100µsec. The LabView target VI is also configured to output two analog signalsspecified by the user in the range of ±10V to power up and external devices.69A.2 Experimental SetupThe experimental setup consists of an integrated EMG amplifier (Advancer Tech-nologies Muscle Sensor v3) from Pololu Robotics and Electronics c© , general-purpose, disposable Ag-AgCl EMG surface electrodes, a PEDOT:PSS strain gauge,current source, and VHB elastomer for mounting purposes. First, the skin was pre-pared for electrode placements using alcohol pads and abrasion using Nuprep gel.This step is very important in order to provide better skin adhesion and to reducethe skin electrical resistance. Two differential electrodes are then placed on thebicep brachii muscle and the third electrode is placed on the elbow bone as a ref-erence. Once the electrodes are placed on the skin, a small piece of VHB tape isused to mount the PEDOT:PSS strain gauge on the skin. VHB was chosen due toits high flexibility and adhesion to the skin. The configuration of the electrodes andthe strain gauge are shown in Fig. A.3. In the same figure we see the flexing of therelaxation state and the flexing state of the bicep muscle.70Figure A.3: The strain gauge mounted on the skin of the bicep in (a) relaxedstate, and (b) flexed stateThe EMG amplifier requires powering signals of ±5V , provided by the Lab-View system and current is applied through the strain gauge resistor while moni-toring the voltage by the LabView VI.A.3 Measurement ResultsWith a similar setup, another research team have used a silicon based strain gaugeto develop a sensor, called muscle contraction (MC) sensor, to measure musclescontractions, as mentioned in [47]; their results are shown in Figure A.4.71Figure A.4: Simultaneous recording of the force (Fg), MC and EMG. The Fgand MC variables are normalised to the maximal value, adapted from[47]As shown in Fig. A.5, our measurement results of the induced strain (using PE-DOT:PSS strain gauges) indicate a certain level of correlation with the concurrentEMG measurements.72Figure A.5: The results of contraction (a) strain data, and (b) integratedEMG.One of the main challenges in performing such correlation is to find the perfectarea of the skin that experiences the most stress due to flexing the bicep brachii.Another main challenge is establishing the right functional dependence betweenthe skin stretching and the activation of the muscles. In this experiment, it wasnoticed that movement of the skin due to the tricep muscle activation caused somespikes in the strain measurements. However, focusing on the results between 30secand 50sec in Fig. A.4, we notice a clear correlation between the EMG of the bicepbrachii and the strain. Having other spikes in the results should not be surprising,73as the deflection of the skin (i.e. the stress applied on the strain gauge mounted onthe skin) is a function of not only the input EMG to the bicep muscle, but rather afunction of multiple muscles inputs.In order to find the cross correlation between the input signal EMG and the out-put signal (strain), we use the Mathworks c© MATLAB command xcorr. The cor-relation coefficients represent the matching between the input signal while shiftedacross the output with highest correlation found at zero delay of approximately60%.74Appendix BPublicationsAlthough not directly related to my thesis work, some other activities during mymasters resulted in 2 conference publications:1- Guillén-Torres, M. Á., Almarghalani, M., Sarraf, E. H., Caverley, M., Jaeger,N. A., Cretu, E., & Chrostowski, L. (2014, September). Silicon photonics charac-terization platform for gyroscopic devices. In Photonics North 2014 (pp. 92880U-92880U). International Society for Optics and Photonics. I created the hardware,firmware and software interfaces of the rotation table, as well as coding the rota-tion patterns. I wrote part of the manuscript, commented and edited the full finalmanuscript.2- Farajollahi, M., Takallo, S. E., Fannir, A., Almarghalani, M., Cretu, E.,Nguyen, G. T. M., Cédric, P., Sassani, F., & Madden, J. D. W. (2015 March).Stacking trilayers to increase force generation. In SPIE Smart Structures NDE2015. (pp. 9430-44). I applied some of the methods learned in the work of thisthesis (laser micromachining) into the fabrication of the stacked trilayer actuator.75


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items