Organic electronics on micro and nano fibers : from e-textiles to biomolecular nanoelectronics

(1)

Organic electronics on

micro and nano fibers

from e-textiles to biomolecular nanoelectronics

Mahiar Hamedi

(2)

Organic electronics on micro and nano fibers

from e-textiles to biomolecular nanoelectronics Mahiar Hamedi

Serie: Linköping studies in science and technology. Dissertations, No. 1224

©, 2008, Mahiar Hamedi, unless otherwise noted Printed by LiU-Tryck, Linköping, Sweden 2008 ISBN: 978-91-7393-763-4

(3)

“Systems of theories are tested by deducing from them statements of a lesser level of universality. These statements in their turn, since they are inter-subjectively testable, must be testable in a like manner – and so

ad-infinitum”

-Karl Popper (1959) “The logic of scientific discovery”

(4)

(5)

Abstract

Research in the field of conjugated polymers (CPs) has led to the emergence of a number of interesting research areas and commercial applications, including solar cells, flexible displays, printed electronics, biosensors, e-textiles and more.

Some of the advantages of organic electronics materials, as compared to their inorganic counterparts, include high elasticity, and mechanical flexibility, which allows for a natural integration of CPs into fabrics, making them ideal for e-texile. In this thesis, a novel approach for creating transistors is presented, through the construction of electrolyte gated transistors, directly embedded on functional textile fibers. Furthermore theoretical and experimental results of the integration of functional woven devices based on these transistors are shown. The realization of woven digital logic and design schemes for devices that can be placed inside living tissue, for applications such as neural communication, are demonstrated. Reducing feature sizes in organic electronics is necessity just as in conventional microelectronics, where Moore's law has been the most impressive demonstration of this over the past decades. Here the scheme of self-assembly (SA) of biomolecular/CP hybrid nano-structures is used for creating nano electronics. It is demonstrated that proteins in the form of amyloid fibrils can be coated with the highly conducting polythiophene derivative (PEDOT-S) through molecular self-assembly in water, to form conducting nanowire networks and nanodevices at molecular dimensions. In a second SA scheme, large area patterning of connected micro-nano lines and nano transistors from the conducting polymer PEDOT-S is demonstrated through assembly of these from fluids using soft lithography. Thereby the problems of large area nano patterning, and nano registration are solved for organic electronics.

The construction of functional nanoscopic materials and components through molecular self-assembly has the potential to deliver totally new concepts, and may eventually allow cheap mass production of complex three dimensional nano electronic materials and devices.

(6)

(7)

Populärvetenskaplig

sammanfattning

Ordet polymer kommer från grekiskans “poly” som betyder många, och “mer” som betyder del, och syftar till väldigt långa molekylkedjor som kan bestå av tiotusentals sammanlänkade atomer.

De polymerer som bokstavligen står oss närmast är livets polymerer. Proteiner och DNA är exempel på polymerer som ingår i allt levande och är känt för de flesta idag. En DNA molekyl kan exempelvis bestå av en kedja av miljoner sammanlänkade atomer.

En annan klass av allmänt kända polymerer är de konstgjorda polymerer som vi känner till som plaster. Det tillverkas idag genom petrokemiska processer hundratals miljoner ton plast i världen årligen. Dessa otroliga mängder av material har många användningsområden. Plaster kan bli formgjutna, exempelvis till läskflaskor, möbler, förpackningar och elektronikhöljen, eller formade till texilfibrer som exempevis nylon. De kan också appliceras från lösning på ytor i form av färger.

Livets och dagens petrokemiska framställda polymerer är inte elektroniska material. Vi är istället bekanta med metaller som strömförare och lyselement, som till exempel i elkablar och glödlampor. Vi känner också till att vanliga elektroniska kretsar består av kisel och att det finns andra hårda material som är elektriskt aktiva material.

Elektronik baserad på metaller och kiselbaserade kretsar har förändrat samhället radikalt under det senaste århundradet. Plast har också haft stor påverkan på samhället som ett viktigt material i många produkter. Elektroniken och plastens värld har dock varit separerade under många år. Men år 1977 gjorde en grupp forskare i USA en spännande upptäckt som bröt denna separation när de hittade en klass av plaster/polymerer som kunde leda ström. Dessa plaster har sedan dess utvecklats, diversifierats och tillförts förbättrade funktioner såsom elektroniska lyselement i alla färger, relativt hög ledningsförmåga, och bättre stabilitet- och processegenskaper. I och med denna utveckling börjar nu den tidigare döda plasten i vissa av våra vardagsprodukter att få liv. Vi börjar idag se helt nya produkter och prototyper baserade på ledande plaster, såsom nästa generations platta

(8)

skärmar från SONY, lysande plastfilmer från General Electrics, elektroniska papper från Plastic Logic, och plastsolceller från Konarka.

I första delen av denna avhandling har ledande plaster undersökts i samband med textila mikrofibrer. Det visas att man kan bygga transistorer och därmed digitala kretsar enbart med hjälp av vanliga textilfibrer, genom att kombinera dessa med ledande plaster. Detta är ett steg mot elektronisk textil, där man slutligen ska kunna väva textiler från nästa generations syntetiska klädfibrer, och forma avancerad elektronik som är helt inbäddad i själva tyget.

I och med att ledande plaster nu börjar att användas till att bygga alltmer avancerade kretsar ställs man också inför samma utmaning som den kiselbaserade elektroniken, nämligen att skapa väldigt många komponenter på väldigt små ytor. Ett chip i din dator har idag nästan 1 miljard transistorer på en area av någon kvadratcentimeter. För att göra detta krävs det att man utvecklar metoder för att mönstra plaster på nanoskalan (1 nanometer = 1 miljondels millimeter). I denna avhandling har en metod att skapa nanomönster i en ledande plast demonstrerats. Detta görs genom att en lösning av plasten formas i en slags gjutform med mikro- och nanometer stora kanaler. Metoden visar möjligheter hur man att på ett enkelt och effektivt sätt kan framställa delar av framtidens plastnanokretsar.

Ännu än metod som i denna avhandling utvärderar ledande plaster på nanometer skalan, bygger på att man förenar en av livets byggstenar, i form av en lång proteinkedja, med en ledande plast. Resultatet är att den ledande plasten fastnar på proteinkedjorna och formar ett elektrisk ledande skal runt dessa. Effekten är att man skapar ledande nanofibrer som är tiotusentals gånger mindre än exempelvis textilfibrer. Denna demonstration visar också på de enorma möjligheter som uppstår i föreningen av biologi och ledande plaster.

(9)

Included Papers

Paper 1

“Towards woven logic from organic electronic fibres” Hamedi M, Forchheimer R, Inganäs O.

Nature Materials 6 (2007) 357

Paper 2

“Electrochemical devices made from conducting nanowire networks self-assembled from amyloid fibrils and alkoxysulfonate PEDOT”

Hamedi M, Herland A, Karlsson RH, Inganäs O.

Nano Letters 8 (2008) 1736

Paper 3

“Fiber embedded electrolyte-gated organic TFTs for e-textile”

Hamedi M, Herlogsson L, Marcilla R, Crispin X, Berggren M, Inganäs O.

Advanced Materials, Accepted

Paper 4

“Bridging dimensions in organic electronics:

assembly of electroactive polymer nanodevices from fluids” Hamedi M, Tvingstedt K, Karlsson RH, Åsberg P, Inganäs O.

Nano letters, Submitted (2008)

Paper 5

“Construction of wire electrodes and 3D woven logic as a potential technology for neuroprosthetic implants”

Asplund M, Hamedi M, Forchheimer R, Inganäs O, Von Holst H,

(10)

Related work not included

Papers

“Limits to nanopatterning of fluids on surfaces in soft lithography” Wigenius, J.A., Hamedi, M., Inganäs, O.,

Advanced Functional Materials 18 (17), 2563-2571 (2008).

“Iron Catalysed Polymerization of Alkoxysulfonate-Functionalized EDOT gives Water-soluble PEDOT of High Conductivity”

Karlsson, R. H., Herland, A., Hamedi, M., Wigenius, J., Åslund, A., Inganäs, O., Konradsson, P.

Chemistry of Materials (Submitted) (2008).

Patents

“Micro and nano structures in elastomeric material“ Hamedi, M., Tvinstedt, K., Åsberg, P., & Inganäs, O.

WO/2006/096123 (2006).

“Electronic circuitry integrated in fabrics”

Hamedi, M. Forcheimer, R. Asplund, M. Inganäs, O.

WO/2008/066458 (2008)

Highlights in the news and media

“Designer Logic Comes to E-Textiles”

Service, R. F.

ScienceNOW 3 April 2007: 1

“Electronic textiles: A logical step” De Rossi, D.

Nature Materials 6 (5), 328-329 (2007).

“NextWorld: Intelligence"

http://dsc.discovery.com/tv/next-world/next-world.html

(11)

(12)

Tackord

Denna höst gulnar de vackra löven för femte gången sedan jag började som doktorand hos Olle Inganäs. Doktorandperioden har inneburit en otrolig personlig utveckling för mig. Jag har under dessa år blivit skickligare på att omsätta ideer till verklighet i både akademisk och kommersiell form, och jag har kommit att fördjupa min förståelse för vetenskap och för den oändliga process i vilken vetenskapen utvecklas. I denna utvecklingsprocess finns det ingenting som fortsätter att fascinera mig mer än den vackra dynamik som finns i de möten och band som uppstår mellan oss människor, och mellan människa och natur. Allting som jag har åstadkommit här är ett resultat av dessa möten och band, mellan mig och så många vackra personer som jag för evigt är bunden till och har allt att tacka för.

Olle Inganäs är en person med enorm dynamik och djup i sina tankar, och en person som verkligen tror på, och generöst stödjer individens frihet. Jag har verkligen haft tur att få jobba med en person med sådana sällsynta kvalitéer. Mina kollegor i vår forskningsgrupp har alla varit fantastiska människor som jag har haft många bra stunder med. Ni har alla bidragit direkt och indirekt till utvecklingen av detta arbete. Jag vill tacka Per Björk, Anna Herland, Mattias Andersson, Kristofer Tvingstedt, Maria Asplund, Wan-yu Lin, Fengling Zhang, Manoj M, Xiangjun Wang, Abay Gadisa, Peter Åsberg, Peter Nilsson, Roger Karlsson, Tomas Johansson, Nils-Krister Persson, Jens Wigenius, Sophie Barrau, Bekele, Shimelis, Wataru, Viktor Andersson. Mina kollegor och vänner i Norrköping har jobbat med närbesläktade forskningsområden och både inspirerat och bidragit till utvecklingen av mitt arbete. Jag vill tacka Payman Tehrani, David Nilsson, Peter Andersson, Fredrik Jakobsson, Joakim Isaksson, Elias Said, Lars Herlogsson, Nathaniel Robinson, Xavier Crispin, och Magnus Berggren

Robert Forchheimer, har varit min ”inofficiella” biträdande handledare. Ditt stöd har betytt mycket för mig.

Jag tackar Jens Birch, har varit min mentor och kommit med flera goda råd längs min resa.

(13)

Hans Von Holst för att du tror på organisk elektronik som ett medicinsk tverktyg.

Jones Alami, tack för att du trodde på mig som forskare innan jag själv gjorde det.

Personalen på IFM, som hjälpt till med lab och pappersarbeten, som möjliggjort min forskning. Jag vill främst tacka, Stefan Klintström, Agneta Askendal , Ann-Marie Holm, Mikael Amle´, och Bo Thunér.

Mina kollegor på vårt företag Donya Labs. Jag vill tacka Martin Ekdal, Gustaf Johansson, Ulrik Lindahl, Koshi och Matt Connors för att ha delat en vision med mig och hjälpt till att förverkliga den genom sitt fulla engemang med start och drift av bolaget.

Ett tack til mina nära vänner som alltid uppmuntrar mig i mina vägval, Amir, Toni, Murat, Ali, Said, Ulrik.

Min nära släkt, som alltid har uppmuntrat och inspirerat min nyfikenhet. Daryoush var den första som lärde mig vad en transistor var, redan när jag var ett barn, jag har dig att tacka mycket för i detta arbete. Azita och Azar har alltid lyssnat på mina tankar och teorier som ett litet barn och som vuxen. Göran, och Anders för att alltid visa sitt genuina intresse för mitt arbete. Mina kusiner som är mig så nära Farzad, Dornoosh, Danesh, och Hanna. Utan min familjs stöd skulle jag inte ha klarat min doktorandtjänst.

Koshiar du har hjälpt mig med att koppla min vetenskap till konstens värld, och jag skulle inte ha kunnat publicera några av mina artiklar eller denna bok utan dig. Maziar jag ha haft många djupa diskussioner med dig om mina arbeten, du inspirerar mig till att vara open-minded i ordets sanna bemärkelse.

Mamma och Pappa, allt jag är och kommer att vara är tack vare er. Älskar er. Min kärlek Anna Herland. Du har varit och är min stora förebild, både i forskningen och privat. Vi har många vackra stunder kvar att dela.

(14)

Chapter 1 General introduction

Polymers consist of large molecules i.e. macromolecules. The word has Greek origins, like many other western scientific words, where, “poly” means many and “mer” means part. The entanglement of these long polymer chains, give polymeric materials a wide range of useful mechanical properties. The most widely known man made organic polymers are petrochemical plastics, which are currently used in various everyday products, such as plastic bags, insulating bodies of electronic products and even in awkward products such as plastic flowers, which indeed have a common factor with real flowers, as polymers are also the building blocks of all life on earth. Proteins and DNA are for example polymers of life, and well known to almost anyone.

Polymers are currently not perceived as electronic materials by general public. Instead metals are generally known as conductors and seen in for example cables and light bulbs. In the microelectronics industry the mix of conducting metals and semi-conducting silicon based materials and other inorganic semi-conductor crystals are the main materials in all microchips. Silicon chips have changed the face of our modern world. These are embedded in almost every electrical product today, and the entire computer industry is currently very much connected with the word silicon, and still dominated by the hardware and software companies located in the infamous “silicon valley”.

This psychological, linguistic and commercial distinction between polymers and electronic materials are however on the verge of change, and they are becoming blended. This interesting change has its roots in the development of molecular electronics, and has taken off seriously since an important discovery was made in 1977. This year a group of researchers at Pennsylvania state university USA, discovered that a class of polymers could conduct high electrical currents [1].

The combination of the electro-optical and processability properties, being similar to plastics, has given rise to a number of fundamental discoveries and inventions since then, and only three decades later a number of commercial products are already emerging on the market, where color displays (e.g.

(18)

“XEL-1” OLED TV from Sony), light emitting plastic sheets (e.g. roll-to-roll manufactured plastic light-emitters from General Electrics) and plastic solar cells (e.g. from Konarka) are among the first big products.

If conducting plastics are regarded as the fourth generation of plastics we can think of using these for embedding electronics into the dead plastics that we have in our current products, and bring these to life. In this thesis this possibility is explored for textiles, where synthetic plastic microfibers such as nylon/polyester can be enhanced with conducting plastics to achieve electronic functions.

Since plastics are today mainly used for their mechanical properties at large scales, not much interest has been given to the nanopatterning of plastics, as compared to nanopatterning of many other materials. This will however be necessary as plastic electronics will undergo the same miniaturization process as conventional electronics has done, even if Moore’s law is not defined for plastics. In this thesis the possibility of nanopatterning of conducting polymers from fluids is demonstrated.

If conducting polymers are regarded as organic polymers structurally identical to the molecules of life, we can instead think of using them for embedding/interacting electronically with the building blocks of life. This intriguing possibility is also explored in this thesis by demonstrating protein based structures that can conduct electricity.

Although significant advance has been made during this 30 years, organic/plastic/polymer electronics is still at it infancy.

(19)

Chapter 2 Conducting polymers (CP)

2.1 Molecular electronics and CPs

The physics of conducting polymer electronics is based on the physics of the more general concept molecular electronics, where micro- or macromolecules (polymers) are the electron conductors. These molecules, and specially the macromolecules, are mainly carbon based, and therefore the word organic electronics is used .

In order to put conducting polymers into historical context, here a chronological overview on the history of molecular electronics is given, partly based on a summary by Hush. [2]. This section can be skipped or revisited after studying of the other chapters.

One of the initial ideas behind molecular electronics dates back to the work of Hund and Mulliken who together outlined the molecular orbital (MO) theory around 1930. The work on MO theory continued through Lennard-Jones who introduced the linear combination of atomic orbitals (LCAO) and was further completed by Mulliken who introduced self consistent field theory (SCF). The ideas take their modern form through the density functional theory of Kohn and the semi-empirical theories of Pople.

Mulliken also studied donor-acceptor charge transfer complexes, defining supra molecular ground states consisting of linear combinations of the orbitals of a donor- and an acceptor. [3] .

The electron donor acceptor (EDA) ideas initially led to the study of solid state EDA complexes such as semiconductors. In 1973 a metallic complex of (TCNQ) and (TTF) molecules was described [4], and later the first EDA superconductor was described [5].

An important contribution to the field of molecular electronics came from Albert Szent-Gyorgi, who in 1941 [6] suggested that if electrons in a crystal would be delocalized, then, theoretically, there is a possibility that common energy levels could exist in bimolecular systems where many molecules form one structural unit, for example in the chlorophyll system.

(20)

Later work by Evan and Gergely predicted bandgaps of 2-4 eV in an infinite hydrogen bonded protein, the work sparked off experimental work where electronic measurements through bio molecules were performed, with the conclusion that these were very poor semi-semiconductors (insulators). Another major advance in the field was the insight into the electronic structure of transition metal ions, from the development of the ligand field theory (1950- onwards), with the basic insight that the electronic degeneracy of ions could be lifted by the electric field of surrounding ligands.

Leading on from that work the fifth major advance in molecular electronics was the development of the theory of the kinetics of homogeneous electron transfer reactions. The basic theories underlying the kinetics of thermal and optical electron transfer are now well established, where R.A Marcus was awarded the Nobel prize 1992 for his work in this area. The main concept of this theory is to depict the complex electron-phonon (vibronic) coupling processes in terms of simpler and more tangible chemical properties.

The limits of the “Marcus theory” are coherent vibrationless through-molecule electronic transport and conduction via coherent geometric disturbances, such as solitons and polarons. A further advance that led to the Nobel prize in this area was awarded to Heeger, Shirakawa, and McDiarmid for the synthesis of conducting/conjugated polymers, and demonstration of high conduction in these materials. This led to the development of the theory of soliton transport in polyacetylene [7]. This theory facilitated the development of modern applications in molecular electronics based on conducting polymers (CPs). The modern theories of coherent and incoherent exciton conduction also fall under the general electron-transfer theory, (for example polaron-type hopping). The focus of this thesis is on systems utilizing conducting polymers, and a more detailed theoretical description is given in the following section mainly focusing on the modern theories of hopping transport in CPs.

The sixth major advance in molecular electronics could be the Aviram-Ratner proposal [8] that a single DBA molecule could act as a rectifier. It was shown that a molecule would be an insulator for low applied voltages and at a critical point suddenly switch on. The landmark proposal builds on the Mulliken donor-acceptor concept, the idea of conduction through a bridge and the theory of electron transfer.

Although the focus of this thesis is on micro and nano electronics based on conducting polymers, one might keep in mind that conducting small

(21)

molecules and CPs are already being used together in devices such as organic light emitting displays. Furthermore as nano electronics circuits based on CPs shrink, the combination of CPs with molecular components, such as those first suggested by Aviram-Ratner, can become interesting. The work by Akkerman, Blom et. al. demonstrates this beautifully by combining the conjugated polymer PEDOT and single molecule of alkanethiols [9]. The remaining part of this chapter is focused on conducting polymers.

2.2 Electronic structure

The structure of a CP is a chain of a carbon-based polymer, having conjugated units with alternating single and double bonds. The conjugation of this organic polymer is what distinguish them from regular plastics, which consists of only single bonds, and the conjugation leads to electronic properties. The words conjugated polymers and conducting polymers are therefore used interchangeably throughout the thesis and abbreviated by “CP”.

Every carbon atom in the conjugated chain is sp2-hybridised. Three of four sp2 orbitals lie in one plane with 120 degrees separation forming strongly localised σ-bonds, and the remaining fourth pz orbital forms a π-bond. The

formations of bonding and anti-bonding orbitals from the π-orbitals give rise to a bandgap with energy states in Lowest Occopied Molecular Orbitals (LUMO) band corresponding to the conduction band (π*_{-band), and Highest}

Occupied Molecular Orbitals (HOMO) corresponding to the valence band (π-band).

(22)

Figure 2.1 Schematic picture of the splitting of HOMO and LUMO levels with increasing

lengths of an acethylene molecule, towards the formation of bands for polyacethylene. Figure 2.1 shows how HOMO LUMO levels split up with increasing CP chain lengths, to form bands.

A simple model for understanding the bandgap in CPs comes from Kuhns [10] description, which is based on the electron in a box potential. In this model the π -electrons are considered to be in a box with boundaries corresponding to the length of the molecule L.

2 2 2 2 2mL n E_n = ! !

With N electrons the energy difference distance between the highest filled

n=(N/2) electron level defining the HOMO and the first unoccupied energy

level LUMO would then define the bandgap Eg of the molecule. i.e

n C E E Eg _n _n 1 1 " ! = +

The important result of this model is that the band gap is lowered when making the CP longer. A correction to this model which prevents the band gap from going to zero as n becomes large is given by the Peierls distortion

(23)

[11]. Since all carbon-carbon lengths are not equal in a CP, the bonds are dimerized (Peierls Distortion). The dimerization would increase the periodicity from L to 2L and add an additional term to the equation, i.e.

2

1 _C

n C

Eg = +

This formula is more successful in describing the band gap of oligomers and some polymers.

2.3 Charge carriers

The linear chain of the CP backbone confines the electrons in the π-orbitals to one dimension. Furthermore the described band gap of the polymer gives a low density of free electrons in the conduction band, hence the CP can be regarded as a 1-D semiconductor (with the exception of broadening). The CPs thus have a conductivity which span from close to insulators to semiconductors. Upon doping of the polymers however, the conductivity can be increased by several orders of magnitude (see figure 2.2). Doping is achieved either by withdrawing electrons (p-doping) or by adding electrons (n-doping), to the polymer backbone. p-doped polymers are usually more stable and mainly used in CP applications today.

Figure 2.2 Conductivity of doped and undoped CPs PEDOT and PA, versus other materials.

The considerable change in conductivity is due to the fact that new electronic species/quasi particles are introduced in the polymer upon doping, which act

(24)

as charge carriers. These charge carriers are transported through the π -bonded polymer chain, and can be either solitons for degenerate polymers and polarons or bipolarons for non-degenerate polymers. The solitons and polarons are quasi-particles with strong electron-phonon coupling, and these give rise to new energy states within the bandgap of the CP resulting in lowered bandgap and also changed color, a phenomenon described as electrochromism. Figure 2.3 shows schematically the formation of the more common bi-polaron band upon p-doping.

Upon doping the number of charge carriers are increased and the charge carrier mobility is also increased due to formation of new band states and so is the conductivity σ= neµ, where n is the concentration of charge carriers, e is the electron charge and µ is the charge carrier mobility. The mobility of the charge carriers in CPs is mainly described by hopping transport. This is described in more detail in section 2.6.

Figure 2.3 Polaron / bi-polaron energies in the band gap of a CP.

2.4 Electrochemistry in CPs

The most convenient way of dynamically altering the doping state of a polymer is through electrochemical redox reactions. Here the CP is brought into contact with an electrolyte and the doping/de doping of the polymer is made by applying a potential between the CP and a counter-electrode in

(25)

contact with the electrolyte. A reference electrode can also be used for more exact potential measurements such as in cyclic voltammograms. The electrochemical doping of a CP can occur not just at the surface but throughout the bulk of a CP film because the polymer acts as a mixed ion and electron conductor. This mixed conduction property is described in more detail in the next section. The more common process of oxidation of a neutral polymer P0 to oxidized state P+, through insertion of an anion A- and withdrawal of an electron e-can be described by

P0 → P+(A-) + e

-In other forms, the oxidation can occur from a polymer-ion complex form, described by

P0(M+A-) → P+(A-) + M+ + e-

The n-doping reaction can be described by a similar chemical reaction with cations instead of anions.

The doping and de-doping of CPs is not a symmetric process, which is reflected by the asymmetric voltammogram illustrated in figure 2.4. One explanation for this is that the system undergoes conformational changes during doping [12,13]

Figure 2.4 shows a schematic complete CV where both n-doping and p-doping can occur in an electrochemical setup with a bithiophene [14].

(26)

Figure 2.4 CV of both p-doping and n-doping.

This kind of CV measurements can be useful for characterization of polymer films and measurements of parameters such as electrochemical bandgap [15]. Usually in these measurements classical electrochemical setups are used comprising liquid electrolytes.

The dynamic and reversible redox response of polymers in electrochemical setups are however also interesting for the construction of electronic components. Two types of electrolyte based components in the form of transistors, are at focus in this thesis, and will be described in detail in chapter 3.

2.5 Ionic Transport in polymers and CPs

2.5.1 Ionic transport in polymers

As already described, the important process of CP doping requires a mixture of electronic and ionic charge transport, and they are therefore closely connected in these materials. The conduction of ions in polymers, and use of

(27)

polymer electrolytes (PE) was discovered about a decade earlier than the discovery of electronic conduction in polymers. The understanding of ionic conduction in CPs could therefore be based on the understanding of ionic conduction in polymers.

Figure 2.5 Schematics ions dissolved inside a polymer matrix.

One of the most common PEs is polyethylene oxide (PEO). PEO can dissociate low molecular weight salts into ions that carry the charge (seen schematically in figure 2.5)

The high ion transport properties in amorphous PEO are associated with transport by hopping between positions, which are dynamically created by the slower dynamics of polymer chains. In weak electrolytes, the ionic conductivity is the sum of partial conductivities of all ionic species. If an amorphous phase is assumed, then the thermal motion of polymer chains cause the movement of ions and the ionic conductivity is similar in behavior to the mechanical properties of the polymer. The ionic conductivity of PEs can be related to the glass temperature Tg through

! " T

_{( )}

= exp #a T # T

(

g

)

T₀+ T # T

(

_g

)

$ % & & ' ( ) )

The glass temperature is connected to the free volume of the polymer, which is a quantity that directly affects the thermal motion of polymer chains in inter-intra chain reactions. Furthermore the ionic species in polymer electrolytes are expected to reside in the free volume. The physic of ionic

(28)

charge transport is therefore intertwined with the packing, and the free volume between the polymer chains.

2.5.1 Ionic transport in CPs

In CPs just as in regular polymer electrolytes the free volume defines the ionic conductivity. In electrochemical reactions however the structure of the CP can change radically and in extreme cases such as in so called micro muscles [16], where the polymer can increase in volume as ions move in and out of the polymer film .

When using larger ions and specially anions, the result can be difficulties in moving ions through remaining free volume after an oxidation or electro-polymerization reactions. An effect of structural changes in regard to CVs has already been shortly discussed.

What therefore differentiates CPs that have ionic conductivity as compared to pure PEs is the coupling between electric fields, ions and electrons, during electrochemistry. The analysis of ionic diffusion and field-driven ionic movement is therefore more difficult in CPs. In most cases it can be assumed that the electronic drift and diffusion is faster, (especially when the polymer is doped), than the ionic movement, and therefore ionic movement should be the main limiting factor in fast processes.

The intertwined ionic and electronic mechanism during electrochemical reactions is of great importance for understanding of the operation of electrolyte based devices such as electrochemical transistors, and electrochromic windows, where bulk doping is dominant, and in electrolyte gated OFETs where the formation of double layers and doping at the interface of CPs and polymer electrolytes is dominant. Electrolyte-gated transistors will be discussed in more detail in chapter 3.

2.6 Electron charge transport theory

2.6.1 Towards a unified theory for disordered CP films

The dominant mechanism of electron charge transport in classical in-organic electronics such as metals and in-organic crystalline semiconductors is based on band like motion of charges, where charges are delocalized through the crystal. The delocalization is due to the good electron overlap between adjacent atoms in the crystal. Here the electron mobility is limited only by

(29)

deviations in the lattice such as impurities and dislocations that cause electron scattering. The result is a high mobility, which decreases with increasing temperature, as higher temperature increases scattering events. The first description of such transport was presented more than a century ago by Drude [17], and therefore referred to as Drude transport. Sommerfeld further developed the Drude model to include quantum mechanical and statistical physics in the free electron gas model.

In CPs film, however, the high disorder of the material and the weak van der Waals interaction between the molecules lead to localized sites/states for the charges, where the charge is mainly transported via thermal hopping between the sites. Here the energy between the sites is disordered both in the energetic and the spatial domain. This, for CPs common type of transport, is only apparent in extreme cases for amorphous in-organics, with very high impurities.

Low temperature transport in such in-organic systems was first described in the 1930s, before the discovery of conducting polymers, by Sir Neville Mott using variable range hopping (VRH). The VRH conductivity is gives by

! "="0exp # T0 T $ % & ' ( ) 1/(d +1) * + , - . /

where d is the dimension of the system. This formula correctly describes the dependence of conductivity on temperature also for CPs. The correct understanding of charge transfer in molecular systems such as CPs however should be based on charge transfer, where the Marcus theory can be applied for the charge carriers (i.e polarons).

(30)

Figure 2.6 Gaussian density of energy states.

In pioneering work by Bässler [18], a model was proposed that disregards the exact chemical nature of the molecules, and instead assumes a Gaussian DOS for the transport sites (see figure 2.6)

!

g(") = 1

2#$ exp %" /2$

2

(

)

The effect of polarons is further neglected in the model as σ for the site distribution is assumed to be larger than the polaron binding energy. As a result the hopping probability is here described by using the more simple Miller-Abraham [19] hopping probability, as compared to Marcus Theory. The formula used is

"ij ="0 exp(#2$Rij# %j#%i KBT ) , %j &%i exp(#2$Rij) , %j <%i ' ( ) * )

(31)

One disadvantage of this model is that there exist no simple analytical solutions, however simulations with Monte-Carlo methods by Novikov et. al. using 2D- 3D sites have shown that this empirical formula [20] holds for disordered CP films: ! µ(E,T) =µ0exp " 3# 5kBT $ % & ' ( ) 2 " C0 # kBT $ % & ' ( ) 3 / 2 " * $ % & & ' ( ) ) eaE # $ % & & ' ( ) )

This formula is based on a slightly more sophisticated correlated Gaussian disorder model GDM, and explains the mobility over wide range of fields and temperatures.

The types of formulas based on the GDM and Miller-Abrahams hopping master equations are the most successful formulas to this date for describing mobility in disordered CPs. The Bässler model has been successful in unifying mobility description for many types of CP based devices, and shown good consistency to experimental data for mobility in disordered CPs, over a wide range of electric fields, temperatures and charge carrier densities. The work by Blom et. al. describes for example a unified mobility for both FETs and LEDs devices based on disordered CPs [21].

Hulea et.al have studied the DOS of the Poly(p-Phenylene Vinylene) over a wide energy range [22], by using electrochemical transistors (described in more detail in chapter 4). The interesting results here is that the DOS can be fitted quite well with Gaussian Distributions and that the only effect that the induced ions in the CP have for these devices, is a broadening of the DOS, as suggested by Arkhipov et. al. [23] for PHT based devices. This means that charge transport even in electrolyte gated devices, which is the device in focus in this thesis, can be explained within the framework of the Bässler hopping model, with a Gaussian or exponential energy distribution.

2.6.2 Metallic and truly metallic conduction in CP films

In more rare cases at higher conductivities in CPs, it is possible to find temperature ranges where this thermal activation of conductivity is changed into a metallic type of temperature dependence, i.e where the conductivity does not increase but decreases with temperature. The highly conducting polymers have been described as disordered metals. Here the conductivity is described as a tunneling between metallic islands separated by non-conducting regions (barriers). These regions could consist of other materials that exist in blends or they could be regions with bad doping or very high

(32)

disorder. The conduction can in this case be described with a Fluctuation Induced Tunneling FIT model [24]. For small metallic islands FIT predicts

!

ln"# $T$1/ 2

This is similar to the 1D case for the VRH model.

In 2006 Lee and Heeger, et. al. [25] showed that a form of the CP polyaniline possesses behavior that fully resembles a metal where both the conductivity vs. temperature is Drude type in the full temperature range. Furthermore the reflection of the polymer shows a plasma frequency in the infrared region, and correct conductivity values could be estimated from these reflectivity measurements using the connection between reflectivity and conduction in the Drude model. One proposed explanation for the deviation from localized hopping transport is that conduction electrons screen local potentials from polarons and structural disorders [26]. The conductivity of these metallic polyaniline films are around 1000 S/cm at room temperature. It is noteworthy that this is the same as the highest conductivities measured in the polymer PEDOT that is extensively used in this thesis.

One of the important results of this model in the context of this thesis, is that although high conductivity is observed in CP films at larger dimensions, the presence of non conducting islands could lead to a scenario where the paths for hopping are broken. It is believed that this scenario occurs in PEDOT:PSS nano sized structures, where the non-conducting parts between the islands comprise the large counter ion PSS (see figure 2.7). A solution to this problem is to use highly conducting polymers that do not have any big counter-ions. In papers 2,4, the introduction of PEDOT-S as such a nano compatible materials solves this problem, and enables creation of metallic nanowires with large aspect ratios.

The physical phenomenon that underlie the miniaturization of conducting nano fibers are discussed in more detail in the next section where deviations from the VRH for nanofibers are discussed.

(33)

Figure 2.7 Schematic demonstration of conduction in disordered metals by hopping

between metallic islands. The break of conductivity for small / nano structures is demonstrated due to the distance between the islands.

2.6.3 Charge transport in CP nanofibers

As each CP is inherently a quasi 1D nanostructure, the use of CPs as components in future organic nano-electronics is very interesting. The advances in nanopattering methods such as self-assembly, nano imprinting, soft-lithography and more, have resulted in construction and electrical characterization of a number of highly conducting CP nanostructures in materials such as Polyaniline, PEDOT and polypyrrole.

The general conclusion from measurements on these nano fibers so far is that there is a very strong power law behaviour for both I(V) and G(T), with power exponents between 2-7 for G(T). An argument of Aleshin et. al [24] based on the exponents in these power laws is that none of the nano systems can be described by either VRH models nor other models for 1D systems. By assuming that parts of the CP nano fibers can be described as

(34)

one-dimensional Fermi liquids with high repulsive short range electron-electron interaction (Luttinger Liquids LL) [27], it is suggested that CP nanofibers can be seen as quasi 1D systems composed of several LLs connected in series. Some interesting results that further point in these directions are magnetotransport studies that have suggested that conductivity in 3D bulk materials in highly doped Polyaniline is controlled by interfibrillar point contacts between nanofibers with high conductance.

Conclusions that can be drawn from these results is that the inherent 1D nature of conjugated polymers could be manifested in nano fibers that are carefully designed at both molecular level and mesoscopic level. In this thesis the demonstration self-assembled PEDOT-S nanowires through molecular self-assembly has enabled yet another material class and tool for the analysis of the physics of transport in highly conducting polymers [28,29].

(35)

2.7 Common CPs: PEDOT and P3HT

The structure and composition of the polymer repeating unit gives rise to a variety of different physical properties, such as optical, electrical and chemical. For example the polyfluorenes and poly(para-phenylene-vinylene) (PPV) are highly luminescent and are therefore serving as active materials in LEDs.

A class of high conductive polymers that can be produced at low cost in large amounts are based on polyaniline, however certain disadvantages such as the formation of hazardous compounds upon degradation has limited the use of polyaniline.

The first synthesis of polythiophene was reported by Yamamoto and Lin in the 1980s. Since then a variety of different synthesis routes based on the polythiophene have been reported and many applications have been demonstrated with these derivatives such as OLEDs, FETs, and Solar Cells. The most successful forms of the derivatives of the polythiophenes are:

(i) The polythiophene polymer where a side-group is added to the 3-position to give solubility.

(ii) Poly(3,4-ethylenedioxythiophene) PEDOT and derivatives thereof.

These classes of polymers, especially the PEDOT polymer, are the ones used most extensively in the thesis, and will be described in more details below.

Poly(3-hexylthiophene (P3HT)

One of the successful polymers of Polythiophene is Poly(3-hexylthiophene) P3HT with an alkane chain containing six carbons on the side chain (Figure 2.8). The pure and regioregular form of P3HT is today commercially available. This CP has good solubility and the ability to form regions with good packing (crystalline domains). The level of regio-regularity can be altered in P3HT to give different levels of crystallinity.

P3HT is semi-conducting (low conductive) in its pristine state. With a bandgap of around 2 eV P3HT can be used in photovoltaics for charge separation and hole transport, in OLEDs as electroluminescent material, and as channel material in OFETs or electrochemical transistors.

(36)

Panzer et. al. [30] showed that field effect mobilities of P3HT could reach as high values as 0.7 cm2V-1s-1, with 1022 charges percm3 and a room temperature conductivity of 1000 Scm-1, by using an electrolyte gated doping. (see more details of electrolyte-gated devices in chapter 4). Although one should be careful with the interpretation of these results as pure FET, these are the highest reported values. In this thesis P3HT is used as channel material for the creation of electrolyte gated OFETs on fibers (see section 4.4). Electrochemically doped P3HT conductivity values of around 20Scm-2 _{are reported by Jiang et al. [31].}

The highest efficiencies of CP based solar cells to date are reported by Kim, Heeger et.al. for tandem solar cell that are partly based on P3HT [32].

PEDOT-analogues

The most commercially successful polythiophene derivative is poly(3.4-ethylenedioxythiophene) PEDOT [33,34], developed in the 1980 by Bayer AG Germany. Bayer developed a water soluble form of this polymer having a charge balancing counter-ion poly(styrene sulfonate) PSS. The result is a water soluble dispersion called PEDOT:PSS, where PSS polymers that are significantly longer than the PEDOT polymers and present in excess contribute both to solvation, and to the formation a polyelectrolyte system. This dispersion is highly p-doped in the pristine state, resulting in a material with high conductivity. The clever introduction of a dioxythiophene on the thiophene, results in a very air-stable polymer although the system is p-doped and electron rich.

The highest conductivity achieved in PEDOT is reported for a controlled form of vapor phase polymerized films with a conductivity of 1000 S cm-1 [13]. The more interesting form of high conductivity variants of PEDOT has however been achieved by further development of PEDOT:PSS. This is done by adding various additives such as di-ethylene glycol and DMSO to create a phase separation between PEDOT and PSS in films [35]. Conductivities as high as 500 Scm-1 are today achieved for commercial PEDOT:PSS in films made by solution processing.

The doped state of PEDOT is almost transparent as the polaron bands absorb in the IR region. The de-doped form of PEDOT is dark blue with an absorption maximum located slightly above 600 nm.

(37)

The combinations of commercial availability of PEDOT:PSS in large quantities, the high air stability and high conductivity of this CP in pristine state, have resulted in numerous studies and demonstrated application, too large to be mentioned within the scope of this text. The most widespread applications to date include hole injection layers in OLEDs, anti-static coatings, and electrochromic windows [33].

As PEDOT is highly conducting in pristine state it can however not be used as channel material in depletion mode OFETs. Instead the good ion mobility of PEDOT:PSS and other PEDOT analogues, makes this material ideal for use in electrochemical devices. These will be discussed in more detail in section 4.1.

One of the weaknesses of PEDOT:PSS is that the excess of the large PSS- ions create dispersions of micellar structures resulting in domains with non conducting islands [36]. As already discussed in section 2.6.3, these high conducting blends can cause problems in the nano domain due to the character of hopping.

A solution to this problem it to create a PEDOT based material with only one component. Attempts to do this were initiated through the modification of the PEDOT monomer EDOT. Reynold’s group reported an alkane substituted EDOT, that could be polymerized into a chloroform-solube PEDOT, and derivatives thereof called PPproDOT [37]. An important chemical step was the synthesis of EDOT with a more polar group, hydroxymethyl, resulting in EDOT-M [37]. EDOT-M is a good starting material for the formation of a number of truly water soluble derivatives. One of these is alkoxsysulfonate EDOT-S where the S group very much resembles the sulfonate in PEDOT:PSS. We reported on a fully water soluble form of PEDOT-S through oxidative polymerization of EDOT-S [29]. This CP has a conductivity of above 1Scm-1 to date. The interesting character of PEDOT-S is that it is believed to be partially self-doped by its ionic side groups, and that the formation of micellar islands cannot occur here. In papers 2,4 PEDOT-S is used to create functional nano electronic structures at dimensions that are not realizable with PEDOT:PSS.

Another positive side effect of PEDOT-S is that this polymer is a polyelectrolyte, (i.e a CP carrying permanent ionic charge), as compared to PEDOT. The polyelectrolyte character of PEDOT-S contributes to the possibility of interaction with biomolecules which often carry net charges [38]. In paper 2 self-assembly of PEDOT-S onto protein molecules is demonstrated for the creation molecular conducting nano wires.

(38)

These results indicate that PEDOT-S like classes of polymers may eventually fully replace PEDOT:PSS and become the next generation of commercial highly conducting stable CPs.

(39)

Chapter 3 Self-assembly of CP

micro and nanofibers

Micro and nano scaling of patterns in CPs is a necessity in the continued developments of all application areas based on CPs.

In organic electronics for example creation of patterns in micro and nano dimensions is important for increasing performance of individual devices and for increasing device density in circuits, just as for inorganic micro-electronics [39].

However the conventional methods that are used for pattering of in-organic materials are usually not compatible with organic electronics. Furthermore the character of CPs gives rise to completely new ways for creating micro and nano patterns. One of the most promising and successful methods for creating patterns in CPs are those that take advantage of solvated CPs for the creation of micro and nano patterns through different forms of self-assembly. In this thesis the two concept of self-assembly from fluids are used. These are self-assembly of CP onto biomolecular nano fibers, and fluidic patterning with soft lithography for making arrays of micro and nano wires. These methods are presented here in detail, and the results of patterning of PEDOT from paper 2,4 are presented.

In the coming chapters the use of these self-assembled micro and nano fibers/wires for making transistors and devices is discussed.

3.1 CP decorated biomolecular nanowires

All living systems constitute building blocks that are hierarchically self-assembled. Whiteside categorizes [40] two main types of self-assembly: - Static self-assembly involving systems that are at equilibrium and do not dissipate energy (e.g. crystals, and folded proteins). The formation of such structures can require energy to form, but once formed they are stable. Current research is mainly based on this type of self-assembly.

- Dynamic self-assembly occurs the interactions responsible for the formation of structures or patterns requires dissipation of energy from the system, such as in all life-forms. The understanding of the processes that

(40)

involve dynamic self-assembly in life is still very rudimentary. However biological systems provide us with many biomolecular building blocks that can be mimicked, modified or directly used for the purpose of bottom up construction of mesoscale structures.

One of the interesting SA geometries involves semi-one-dimensional structures such as nanowires and nanotubes. Functional SA nanowires can be used as building blocks in next generation nano-electronic circuits, were they can be used as transistor channels, emitters, nanowires in crossbars and more (see chapters 4,6 for more detail). In the field of inorganic nano electronics the functional self-assembled nanowires: carbon nanotubes, and semi-conducting crystalline nanowires [41] are the most extensively studied materials.

Biomolecules can in the same manner be self-assembled into biomolecular nanowires. The most well-known biomolecular nanowire is the DNA molecule. DNA can have very high aspect ratios of several thousands, and its unique functionality has enabled DNA molecules to be used as nanowires as well as in SA of quite complex 2D and 3D nanostructures [42]. Peptides and proteins are another class of biomolecules that can generate 1D nanowires. These are referred to as fibrils, filaments or protein fibers, and mainly self-assemble from other smaller protein polymers or from peptide monomers. In this thesis the focus has been on amyloid fibrils. These are biomolecules with a diameter of approximately 10nm [43], and lengths extending up to 10 µm. The structure of amyloid fibrils is based on extended β-sheets [44]. Furthermore these fibrils are very stable and can easily be functionalized.

As biomolecules are not easily electronically utilized, CPs offer a great opportunity for making biomolecules functional, where the biomolecule acts as a nanotemplate that is functionalized/coated with an electroactive CP. In previous works, DNA has been used as template together with conducting CPs. Polypyrrole has been chemically polymerized on DNA to form self-assembled nanowires [45], polypyrrole and poly(3,4-ethylenedioxythiophene) have been electrochemically polymerized on DNA to form chiral CP structures [46], and polyaniline nanowires have been formed on immobilized DNA on surfaces using chemical polymerization routes [47]. There are however problems with such polymerization methods such as degradation of the biomolecules, and poor CP quality. This is probably the reason for why none of the previously demonstrated polymerizations of highly conducting polymer have resulted in highly conducting nanowires.

(41)

A much more simple and natural route involves direct SA of CPs onto biomolecular nano templates in the fluid phase. This powerful method of combining the functionality of both biomolecules and CPs have been demonstrated in schemes based on combining the molecular recognition capabilities of DNA together with the optical properties of CPs to self-assemble supramolecular systems. Systems with optical properties that mimic logic operations [48,49] have been demonstrated.

The research group affiliated with this thesis work, have previously demonstrated that luminescent conjugated poly-electrolytes can be self-assembled onto amyloid fibrils [38,50], and also onto DNA [51] to form optically functional nanowires. None of these CP:biomolecular nanowires however exhibit measurable electrical characteristics. In paper 2 in this thesis we show highly conducting biomolecular nanowires, further described in the next chapter.

3.1.2 PEDOT:amyloid nanowires

The successful self-assembly of luminescent conjugated-poly-electrolytes [50] onto DNA and protein fibrils can not be demonstrated for the highly conducting polymer PEDOT:PSS. The reason is probably that PEDOT in PEDOT:PSS lacks ionic side chains and that the presence of the large and non-conducting macromolecule PSS prevents biomolecules to be used as nano templates. In paper 2 the newly developed form of PEDOT-S [29] is used. It is believed that the CP backbone in PEDOT-S is partly self-doped by the alkoxysulfonate ionic group, and that the rest of the ionic groups make the polymer water soluble (dispersible).

The ionic interactions of the ionic side groups can also together with other weak interaction forces, (such as hydrogen binding/hydrophobic interactions), be the mechanisms of binding between this CP and positively charged biomolecules.

The beauty of the SA process is the ease under which PEDOT-S binds onto amyloid fibrils directly in water without the need of any heat, and in a matter of a few minutes. One of the difficulties in making conducting nanowires, is however that the CP has to coat the entire fibrils so that no lack of PEDOT-S along the fibers would allow for conduction breaks. This is solved by using the CP in excess during self-assembly. It is believed that PEDOT-S chain assemble around amyloid fibrils until all assembly sites are filled. The remaining PEDOT-S is then filtered away since the size of these is smaller than the size of the coated fibrils. This is demonstrated schematically in figure 3.1.

(42)

It should be noted that this types of self-assembly with CP being in excess has not been done in the previous studies where purely luminescent CPs are used. One reason is that luminescence is seen even if the fibers are not fully coated. Instead a smaller amount of CPs relative to biomolecules is used in order to avoid any filtering of the supernatant solution to remove excess of unbound CP.

AFM and TEM images of PEDOT-S:amyloid reveals nanowire geometries with very high aspect ratio and relatively high stiffness. Networks of these fibers are used as channel material in electrochemical transistors, for analysis of the electrical character of the nanowires. This is described in more detail in section 4.3.1.

Figure 3.1 Schematic figure of self-assembly of the CP, PEDOT-S, onto biomolecular

fibrills, and the filtering of residual PEDOT-S

The SA method of PEDOT-S:amyloid can most probably be applied for a number of other biomolecules as well. Figure 3.2 shows AFM and TEM pictures of SA PEDOT-S:biomolecule fibers where also a peptide NW coated with PEDOT-S is shown.

(43)

Figure 3.2 a) TEM picture of peptide NW network coated with PEDOT-S b) TEM picture

of amyloid fibrillar networks coated with PEDOT-S c) AFM figure of free standing amyloid fibrlls coated with PEDOT-S, amyloid network, and two PEDOT-S coated d) TEM picture of pure amyloid fibrillar network

(44)

3.2 Soft-lithography patterning

Soft lithography is a non-photolithographic set of micro/nano fabrication methods, where all of its members use a patterned elastomer as a stamp, mold or mask, instead of a rigid photo mask as in conventional photo lithography. The method was popularized in the beginning of 90s by George Whitesides et. al. [52]. The elastomers can be used to generate two- and three-dimensional patterns through a different set of techniques, which have been developed during the years. The smallest features that are demonstrated with SL are as small as around 1 nm [53], which means that SL can remain a nano patterning tool for the future. The powerful capabilities of different techniques and the experimental simplicity of SL allows for the use of these in a very wide range of applications.

The possibility of templated self-assembly [40] of a diversity of molecules and materials with elastomer templates, has been exploited in this thesis for the patterning of CPs. The techniques of molding, especially molding in capillaries, and soft embossing / soft nanoimprint lithography are the main techniques that has been used for micro and nano patterning of CPs.

Soft nano imprint lithography

Nanoimprint lithography NIL, refers to patterning methods where printing is done by pressing a pre-generated nano relief structure into a meltable material. If an elastomer is used as the relief structure and the fusible structure is UV curable resist (which is usually the case), the method can be called soft UV NIL [54].

Soft UV NIL has been used in order to pattern the passive structures in a number of CP based devices such as LED and TFTs 39. The use of NIL for direct patterning of CPs is more rarely occurring, but has for example been done for polyaniline [55].

In this thesis soft NIL is for patterning of passive resists, in order to make three-dimensional masks (see coming section 3.2.1).

Capillary molding patterning

The method of capillary molding comprises laminating an elastomer against a substrate to form micro or nano fluidic channels, into which liquid materials such as CPs in solutions can be filled. The filling process can be driven by pure capillary force (often referred to as micromolding in capillaries, MIMIC) or external pressure. When the liquid has filled channels

(45)

it is solidified through solvent evaporation or curing, and the solid micro/nano objects or molecules that were carried by the liquid are delivered to the substrate surface forming a negative of the channel. The mold is finally removed to complete the fabrication leaving only the patterns of active material.

Molding is an interesting SL technique since it allows direct patterning of organic electronic materials directly from fluids 56, without any additional step that could degrade the material. Molding allows patterning in both micro and nano dimension, with good feature size control especially in the planar directions, allowing for construction of dense nano structures such as dense arrays of nano lines.

Direct patterning of organic electronic materials by capillary molding in previous works include patterning of conducting carbon micro electrodes [57], and patterning of micro and nano structures in polyaniline [58].

3.2.1 Patterning of PEDOT in bridged micro- and

nanowire arrays

In the continued development of organic electronics, scaling of patterns sizes is an important issue. Two main challenges can be pointed out here:

1) Methods for scaling patterns (to sub micro, nano and molecular scales), that match the requirement of cheap and large area patterning, since this is required for organic electronics.

Previous examples that try to deal with this issues include self-aligned printing [59,60] where nano gaps are easily created between inkjet printed lines by dewetting from hydrophobic surfaces. However these schemes do not provide sub micron features throughout the pattern as the thickness of the lines that create the nano gaps still is in the ten micrometer regime.

2) The second challenge is alignment and registration of printed nano electronic devices over large areas. The reason is that large area flexible substrates such as plastics are structurally instable, meaning that alignment between several nano patterning steps is very difficult [39]. Furthermore alignment at submicron scales is not compatible with printing techniques such as ink-jetting.

(46)

One of the purposes of Paper 4 in this thesis is to show a method that deals with both of these issues by demonstrating large area patterning of seamlessly connected micro-nano patterns the CP PEDOT.

The idea here is to take advantage of the water solubility of PEDOT polymers, (which makes the polymer solutions compatible with elastomer materials such as PDMS and derivatives of PDMS), and create PEDOT patterns through molding in capillaries. This technique is used for patterning connected PEDOT wires that range from nano sizes and up to tens of micrometers thereby fulfilling the requirement of large area cheap nano patterning. The connection between the nano-micro features would further mean that the larger micro parts could be used as pads and connected to other structures, which in turn are aligned and patterned on-top using macro scale patterning techniques such as roll to roll printing (100 µm). In this way the issue of nano alignment for the PEDOT layer would also be automatically eliminated.

The reason for choosing PEDOT is that the material is multifunctional, and can act as many parts in organic electronics devices such as OLED backplane layer, conduction lines, crossbar lines, source-drain contacts, transistor channel material, and more.

The presented idea of capillary molding of connected features ranging from nanometers and up to micrometers is however not straightforward.

One reason is that the softness of the elastomer in SL limits the possible aspect ratios a (a=height/width) of the relief that can be used, when the relief in the stamp has one constant height, as is usually the case. If a is too low, the elastomeric material will deform/sag defining a lower limit al. If a is too

high the stamp will instead collapse under its own weight defining an upper limit ah The lower and upper limits of the aspect ratio define a working

range w=ah / al for the elastomer (see Figure 3.3). The most commonly used

material in soft lithography, poly(dimethylsiloxane) (PDMS), has a working range of w~10 based on calculations and experimental findings [61]. The idea of one step patterning of features in a range of some 100 nm, to some 100 µm requires working ranges in the order of at least 1000. In this thesis this is achieved by creating a three-dimensional hybrid templates with two height levels, where level one contains the nano structures with one height and level two the higher micro structures. These templates were produced using relatively cheap production methods (for more information, see methods paper 4, and patent application [62]).