Katholieke Universiteit Leuven Faculteit Wetenschappen Departement Chemie

Dimitri Janssen

SELF-ASSEMBLING MONOLAYERS FOR ORGANIC THIN-FILM TRANSISTORS

Promotor Prof. Dr. W. Dehaen

Copromotor Prof. Dr. Ir. P. Heremans JUNI 2006

Katholieke Universiteit Leuven Faculteit Wetenschappen Departement Chemie

Self-assembling monolayers for organic thin-film transistors

Dimitri Janssen Juni 2006 Proefschrift voorgedragen tot het behalen van de graad van Doctor in de Wetenschappen

Promotor:

Prof. Dr. Wim Dehaen

Copromotor:

Prof. Dr. Ir. Paul Heremans

© Katholieke Universiteit Leuven – Faculteit Wetenschappen – Departement Chemie, Celestijnenlaan 200F, B-3001 Heverlee, Belgium All rights reserved. No part of the publication may be reproduced in any form by print, photoprint, microfilm or any other means without written permission from the publisher. D/2006/10.705/31 ISBN 90-8649-034-4

"The surest way to be late is to have plenty of time" L. Kennedy

Acknowledgements Het zit er op…, de tijd is voorbijgevlogen…, maar dat is niet abnormaal want ‘times flies when you're having fun’. En die ‘fun’ heb ik aan een hele hoop mensen te danken. Allereerst wil ik mijn promotor, Prof. Wim Dehaen bedanken. Ondanks het feit dat mijn thesis onderwerp niet helemaal in zijn specialiteit lag, aarzelde hij geen seconde om het promotorschap op zich te nemen. En hoe druk de agenda ook was, op de raarste momenten van de dag (of nacht, zoals op congres buffetten), steeds maakte hij tijd voor het nalezen van teksten, hulp bij problemen of gewoon een babbel. Ook de fijne, open en vrije sfeer op zijn labo (met de talloze labodrinks) waar ik steeds terecht kon bv. voor het gebruik van de infrastrucuur of het lenen van chemicalën, heb ik ten zeerste geappreciëerd. Evenveel dank ben ik verschuldigd aan mijn copromotor, Prof. Paul Heremans, ten eerste natuurlijk voor de geboden kansen, de onvoorwaardelijke steun en het uitdagende, interessante onderwerp. Maar ook voor de fijne, open groepsgeest en werkomgeving met talloze mogelijkheden waar we onbezorgd en vrij konden experimenteren, een eigen labo installeren of gewoon al eens iets konden laten vallen…;) Of course, i would also like to thank the jury members, Prof. G. Maes, Prof. S. De Feyter, Prof. J. Radecki and Prof. D. Knipp, for their willingness to take seat in the jury, for reviewing this thesis, for their kind suggestions and the nice collaborations. Het IWT, de KUL en Imec wil ik bedanken voor de financiële steun zonder dewelke dit project onmogelijk was. Zowel mensen van de oude ‘optoelectronics’ groep als de jongere generatie in de ‘polymer & molecular electronics’ groep ben ik ook zeer veel verschuldigd. Een speciaal woord van dank gaat uit naar Stijn Verlaak. Hij stond niet alleen mee aan de wieg van dit doctoraatswerk maar bleef het ook later met raad en daad bijstaan. Onze fijne samenwerking, de vele fun-experimenten, de gezellige babbels en blikjes-momenten zal ik nooit vergeten. Voorts wil ik mijn ‘party cubicle’-genoten en alle andere groepsleden over de jaren heen bedanken voor de fijne sfeer. Wat de PME groep zo speciaal maakt, is de unieke, open teamgeest waardoor meer dingen sneller verwezenlijkt kunnen worden. Er zijn geen eilandjes met een eigen onderwerp of problemen. Iedereen, ondanks de uiteenlopende onderwerpen, werkt samen en helpt waar nodig.

i

Dit multidisciplinaire doctoraatswerk is daar een mooi voorbeeld van. ‘After work’ manifesteerde dit zich in de vele gezellige activiteiten zoals pizza- en filmavonden, cubicle drinks en sportevenementen. Reiner, Cathleen, Tom, Wim, Yadong, Johan, Xiaobin, Hongzhen, Lihuan, Xinpei, Jan, Jef, Stijn, Stijn, Soeren, David, Sarah, Carsten, Dietmar, Robert, Cedric, Kris, Nathalie, Peter, Hans, Vladimir, Andriy, Joris, Maarten, Johan, Erwin, Didier, Barun en de vele thesisstudenten, bedankt! Graag wil ik ook de mensen van labo Dehaen bedanken voor al hun hulp tijdens mijn eerste synthese- maanden en nadien bij allerhande analyses, experimenten en problemen. Ook al kwam ik niet vaak in den F langs, of skipte ik al eens een labo-vergadering, steevast werd ik uitgenodigd op de vele interessante en gezellige congressen, uitstappen en labo-drinks. Mario, Raf, Stefan, Stefaan, Wouter, Bert, Misha, Prasad, Nathalie, Tine, Maarten, Kris, Karel, Suzanne, Chantal, Wienand, Annelies, Gu, Ahmed en iedereen van de oude en jonge garde die ik ongetwijfeld nog vergeet, bedankt! Verder wil ik nog een aantal belangrijke mensen bedanken. Zonder hun hulp was mijn onderzoek nooit zo snel gestart (en misschien wel nooit afgeraakt ;) ). Hun hulp en raad, de gastvrijheid in hun labo’s en hun vriendschap heb ik enorm geappreciëerd. Wim, Filip, Hoon, Koen, Jean-Michel, Kristien, Roel, Randy, Gunter, Jan, Zhou, Andrew en alle overige bio-mensen, bedankt! De talloze, andere MCP collega’s, zoals onze buren van de ‘magneto’ en epitaxie, wil ik ook graag bedanken voor hun hulp en de gezellige atmosfeer op het verdiep, op de vrijdag-meetings,… bedankt! De sympatieke UCP groep van Imec verdient ook een woord van dank. Zij was niet alleen mijn vast adres voor een korte vlucht uit de ‘vuile’, organische wereld maar ook de toegang tot het propere silicium universum waar met TXRF mijn SAMs met ongekende precisie gemeten konden worden. David, Wim, Frank, Bart, Nausikaa, Jens, bedankt! Bedankt ook aan alle mensen die op het departement Chemie of op Imec instonden voor de technische en administratieve ondersteuning: Marion, Inge & Chantal, de aankoop-, personeels- en financiële diensten, de receptionistes, de mechanische en elektronische werkplaatsen, het magazijn en de CASP, de sympatieke kuismannen in de support, de helpdesk, de mensen van de veiligheid, de glasblazerij,… allemaal bedankt! Natuurlijk wil ik ook mijn vrienden van in Leuven en daarbuiten en mijn familie danken voor hun interesse en steun! Tot slot wil ik mijn ouders en broer bedanken voor alle kansen die ze mij gegeven hebben, voor hun interesse en steun, hun geduld en aanmoedigingen,… bedankt! Kristel, bedankt voor alles en nog veel meer! Dimi ☺

"Nothing is a waste of time if you use the experience wisely." A. Rodin

ii

Abstract Despite a lower performance compared to their inorganic relatives, organic electronic devices potentially offer significant advantages like processing on flexible, large area substrates at low cost. Organic light-emtting diodes (OLEDs) are currently finding their way to the market but for organic thin-film transistors (OTFTs), further advances in performance, materials and processing are required. In this multidisciplinary thesis we have investigated the use of self-assembled monolayers (SAMs) to improve and replace OTFTs. Self-assembled monolayer are ordered assemblies of functional molecules which are formed spontaneously on an appropriate surface and are typically used to modify surface properties of materials. Several known and new SAM deposition and characterization techniques first were explored and optimized to obtain the required, qualitative SAMs for use in organic thin-film transistor structures. Also, an extensive study of the solvent wettability of various SAM modified surfaces was performed which is highly useful for processing organic materials to make organic devices. Various SAMs were then applied on the different interfaces in OTFTs which allowed for better control of the morphology and properties of the organic semiconducting material. This resulted in a significant improvement of the electrical performance of the organic transistors and even enabled a new way to pattern the organic semiconductor. Also, a more gentle method for contacting materials namely by pressing a rubber stamp with metal contacts against the material, was briefly studied. The possibilities of replacing the various functional parts of an OTFT with self-assembling structures proved much tougher. Different families of self-assembled monolayers were evaluated as a potential iii

replacement for the organic semiconductor but did unfortunately not yield the desired results. Also, a process was developed to selectively ‘grow’ a metal layer from solution by self-assembly on patterned substrates although some problems impeded the practical application of this process. Nevertheless, the key issues that need to be resolved for future application of these self-assembling OTFT replacements were identified. To finally conclude, this work has made important contributions in the effort to enhance the performance and processing of organic thin-film transistors and to replace them with self-assembling structures.

iv

Abbreviations 2D

two dimensional

3D

three dimensional

AFM

Atomic force microscopy

APTES

3-aminopropyltriethoxysilane

ATR

Attenuated total reflection

BA

Brewster angle

BPTS

bromophenyltrichlorosilane

BUDMS

11-bromoundecyldimethylchlorosilane

BUTS

11-bromoundecyltrichlorosilane

CA

Contact angle

CB

Conduction band

CUTS

11-cyanoundecyltrichlorosilane

CV

Cyclic voltammetry

CVD

chemical vapor phase deposition

DCC

N,N'-dicyclohexylcarbodiimide

DMAP

4-dimethylaminopyridine

DMF

N,N-dimethylformamide

DMSO

dimethylsulfoxide

DSH

dodecanethiol

v

DTS

decyltrichlorosilane

EOS

Equation-of-State

FDTS

1H,1H,2H,2H-perfluorodecyltrichlorosilane

FET

Field-effect transistor

FTIR

Fourier-transform infrared (spectroscopy)

GA-IRRAS

Grazing angle infrared reflection and absorption spectroscopy

HMDS

hexadimethylsilazane

HOMO

Highest-occupied molecular orbital

iMO

Intermolecular orbital

IR

Infrared

ITO

indium tin oxide

LED

Light-emitting diode

LUMO

Lowest-unoccupied molecular orbital

MO

Molecular orbital

MOSFET

Metal oxide semiconductor field-effect transistor

MPTES

3-mercaptopropyltriethoxysilane

NMP

N-methylpyrrolidone

OFET

Organic field-effect transistor

OLED

Organic light-emitting diode

OMBD

Organic molecular beam deposition

OTFT

Organic thin-film transistor

OTS

octadecyltrichlorosilane

OVPD

Organic vapor phase deposition

OWRK

Owens-Wendt-Rabel-Kaelble

oxUETS

Oxidized UETS

P3HT

poly-(3-hexylthiophene)

PDMS

polydimethylsiloxane

PEDOT

poly(3,4-ethylenedioxythiophene)

PEDOT:PSS

poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate)

PEG6/9

2-[methoxy(polyethyleneoxy)propyl]trimethoxysilane

PETS

phenethyltrichlorosilane

PPV

poly(p-phenylene vinylene)

PTES

propyltriethoxysilane

PTS

phenyltrichlorosilane

vi

PVD

Physical vapor deposition

RF

Radio frequency

RFID

Radio frequency identification

RMS

Root mean square

SAM

Self-assembled monolayer

SAM-FET

Self-assembled monolayer field effect transistor

SBAH

sodium bis(2-methoxyethoxy)aluminum hydride

SC1

Standard clean 1'

SEM

Scanning electron microscopy

STM

Scanning tunneling microscopy

TESU

11-(triethoxysilyl)undecanal

TFT

Thin-film transistor

TTF:TCNQ

tetrathiafulvalene:tetracyanoquinodimethane

TXRF

Total reflection x-ray fluorescence

UETS

10-undecenyltrichlorosilane

UTS

undecyltrichlorosilane

UV

Ultraviolet

UV/VIS

Ultraviolet/visible (light)

VB

Valence band

XPS

X-ray photon spectroscopy

vii

viii

Table of contents Acknowledgements

i

Abstract

iii

Abbreviations

v

Table of contents

ix

Publications & contributions

xi

Objectives & Outline

1

1.

Introduction

3

1.1 Self-assembled monolayers

3

1.1.1

The organosulfur family or ‘thiols’

6

1.1.2

The organosilicon family or ‘silanes’

10

1.1.3

Characterization of SAMs

15

1.1.3.1 Contact angle measurements

15

1.1.3.2 Ultraviolet/Visible absorption and fluorescence spectroscopy

16

1.1.3.3 Fourier-transform infrared spectroscopy

18

1.1.3.4 Cyclic voltammetry

19

1.1.3.5 Scanning probe microscopy : atomic force and scanning tunneling

21

1.1.3.6 Other techniques

22

1.2 Organic semiconductors and thin-film field-effect transistors

22

1.2.1

Introduction

23

1.2.2

Organic semiconductors

25

1.2.2.1 Molecular orbital theory

26

1.2.2.2 Charge transport

31

1.2.2.3 Materials

34

1.2.2.4 Processing

36

1.2.2.5 Applications

37

ix

1.2.3

2.

Organic thin-film field-effect transistors

1.3 References

46

Monolayer-related improvements

49

2.1 SAM Deposition techniques

49

2.1.1

From the vapor phase in a reflux setup

50

2.1.2

From the vapor phase in vacuum

52

2.2 "Buried metal layer" technique for FTIR

57

2.3 Total reflection X-ray fluorescence on SAMs

61

2.3.1

Introduction

61

2.3.2

Surface density

63

2.3.3

Applications

68

2.4 SAM surface enthalpy and wettability

3.

38

71

2.4.1

SAM deposition and contact angle measurement

71

2.4.2

Surface enthalpy determination

76

2.4.3

Wettability & wetting envelope

82

2.5 Conclusion

84

2.6 References

85

Application of self-assembly in organic field-effect transistors

87

3.1 Improving OTFT characteristics with SAMs

88

3.1.1

The organic semiconductor

89

3.1.1.1 Nucleation growth theory

89

3.1.1.2 Improved electrical characteristics of the organic semiconductor

96

3.1.1.3 Patterning

96

3.1.1.4 Improved ‘overgrowth’

100

3.1.1.5 Conclusion

103

3.1.2

The insulator

103

3.1.3

The contacts

107

3.1.3.1 The use of thiol SAMs on gold contacts

107

3.1.3.2 ‘Soft’ contacts

111

3.2 Replacing OTFT parts with SAMs 3.2.1

116

The organic semiconductor

116

3.2.1.1 Introduction

116

3.2.1.2 The thiol SAM approach

118

3.2.1.3 The two-step silane SAM approach

123

3.2.1.4 Conclusion

135

3.2.2

The insulator

136

3.2.3

The contacts

138

3.3 Conclusion

141

3.4 References

143

General conclusion & recommendations

147

Algemeen besluit

153

x

Publications&contributions Journal publications D. Janssen, R. De Palma, S. Verlaak, P. Heremans, W. Dehaen, Static solvent contact angle measurements, surface free energy and wettability determination of various self-assembled monolayers on silicon dioxide, Thin Solid Films, accepted (2006)

S. Steudel, D. Janssen, S. Verlaak, J. Genoe, P. Heremans, Patterned growth of Pentacene, Appl. Phys. Lett. 85 (23), 5550 (2004) S. Steudel, D. Janssen, S. Verlaak, P. Heremans, Patterned growth of organic small-molecule layers, Mat. Res. Soc. Symp. Proc. 814, i12.6 (2004)

J. Reyaert, D. Cheyns, D. Janssen, G. Borghs, J. Genoe, P. Heremans, Ambipolar injection in a submicron channel light-emitting tetracene transistor with distinct source drain contacts, Mater. Res. Soc. Symp. Proc. 846, DD7.13.1 (2004)

J. Reynaert, D. Cheyns, D. Janssen, R. Muller, V.I. Arkhipov, J. Genoe, G. Borghs, P. Heremans, Ambipolar injection in a submicron channel light-emitting tetracene transistor with distinct source and drain contacts, J. Appl. Phys. 97 (11), 114501 (2005)

xi

S. Steudel. S. DeVusser, S. De Jonge, D. Janssen, S. Verlaak, J. Genoe, P. Heremans, Influence of dielectric roughness on the performance of pentacene transistors, Appl. Phys. Lett. 85, 19 (2004)

S. Verlaak, S. Steudel, P. Heremans, D. Janssen and M. Deleuze, Nucleation of organic semiconductors on inert substrates, Phys. Rev. B 68(19),195409 (2003)

S. Verlaak, S. De Jonge, S. Noppe, D. Janssen, S. Steudel, S. De Vusser and P. Heremans, Intra-Grain and Oligo-Grain Top-Contact Organic Thin film Transistors, Mat. Res. Soc. Symp. Proc. 771, L6.9.1 (2003)

K. Myny, S. De Vusser, S. Stoedel, D. Janssen, R. Müller, S. De Jonge, S. Verlaak, J. Genoe, P. Heremans, Self-aligned surface treatment for thin-film organic transistors, Appl. Phys. Lett. 88, 1 (2006).

Patents P. Heremans, D. Janssen, S. Steudel, S. Verlaak, Method for producing pattened thin films, US Patent 10/885220 (2003)

Conferences & posters D. Janssen, W. Dehaen, P. Heremans, Synthesis and self-assembly of alkylsubstituted carbazole on gold surfaces, poster, 5th Sigma-Aldrich Organic Synthesis Meeting, Spa, Belgium, december 6-7, 2001

D. Janssen, W. Dehaen, P. Heremans, Coupling of conjugated moieties on amino-terminated selfassembled monolayers, poster, 6th Sigma-Aldrich Organic Synthesis Meeting, Spa, Belgium, december 2002

D. Janssen, W. Dehaen, P. Heremans, Amino-terminated self-assembled monolayers and coupling of conjugated moieties, poster, Nanoworkshop Egmond-aan-Zee, The Netherlands, juni 1-3, 2003

D. Janssen, W. Dehaen , P. Heremans, Vapour-phase deposition and TXRF characterization of silaneSAMs, poster, Chemistry and Self-assembly for Nanotechnology, Namen, Belgium, 2004

S. Verlaak, D. Janssen, B. Dutta, P. Heremans, Morphology and impurity effects in pentacene thin-film transistors, 44th Electronic Materials Conference, Santa Barbara, Ca., USA, June 26-28, 2002 xii

S. Verlaak, S. Steudel, D. Janssen, M. Deleuze and P. Heremans, The vapour phase growth of organic semiconductors on inert substrates, European-MRS, P-II.04, Strassbourg, France, June 10-13, 2003 S. Steudel, S.Verlaak, D. Janssen, S. De Vusser, J. Genoe and P. Heremans, Influence of traps on the dynamic properties of OTFTs, Organic Electronics Winterschool, , Planneralm, Austria, March 6-12, 2004

J. Genoe, S. Steudel, S. De Vusser, D. Janssen, and P. Heremans, Ongoing Work Towards Cheap RFID Tags In Organic Electronics, International workshop on RFID and wireless sensors, Kanpur, India, November 11-13, 2005

J. Genoe, S. Steudel, S. De Vusser, S. Verlaak, D. Janssen, and P. Heremans, Bias stress in pentacene transistors measured by four probe transistor structures, Proceedings of the 34th European Solid-State Device Research Conference - ESSDERC 2004, 413 -416, Leuven, Belgium, September 21-24, 2004

K. Bonroy, J. Friedt, F. Frederix, R. De Palma, D. Janssen, B. Goddeeris, G. Borghs, Realization and characterization of 3D gold substrates for biosensor applications, Qsense User Meeting, Leuven, Belgium, October 6-7, 2005

J. Genoe, S. De Vusser, S. Steudel, K. Myny, D. Janssen, S. Verlaak, and P. Heremans, Organic Circuits & Devices, Delft, the Netherlands, November 23, 2004

P. Heremans, J. Genoe, S. De Vusser, S. Steudel, R. Muller, K. Myny, D. Janssen, and S. Verlaak, Organic circuit components for pentacene RF-ID tags, VLSI-TSA Tech. Conference, Hsinchu, Taiwan, April 25-27, 2005.

P. Heremans, J. Genoe, S.De Vusser, S. Steudel, K. Myny, R. Muller, D. Janssen, and S. Verlaak, Is printed electronics compatible with low power?, Flexible Electronic Workshop, National Chiao Tung University, Hsinchu, Taiwan, April 27, 2005.

xiii

xiv

Objectives & outline Organic electronic devices can offer significant potential advantages despite a lower performance compared to their inorganic relatives. Thanks to their compatibility with and processibility on large area substrates, light weight, potential low price and simple packaging organic electronics have the prospect of a bright future. Although displays based on organic light-emtting diodes (OLEDs) are currently finding their way to the market, organic transistors have not developed that far yet and could find application in logic circuits (e.g. radio-frequency identification tags) and switching circuits for organic LEDs in displays. However, further advances in performance, materials and processing of organic thin-film transistors (OTFTs) are required. As interfaces are omnipresent in organic transistors, the ability to control the properties of these interfaces will prove essential for improving the final device in terms of processing and properties. The application of self-assembled monolayers (SAMs) on these interfaces allows such control and will be investigated in this work. In other words, the goal of this thesis is to improve and complement organic thin-film transistors with self-assembled monolayers. Surface modification is not the only interesting aspect of self-assembled monolayers. From processing point of view, it would be very interesting to exploit the self-assembling properties of these SAMs. Imagine a process in which organic transistors would assemble themselves! We will therefore also look at ways to ultimately replace the various parts of an organic thin-film transistor with self-assembling systems.

1

Logically, the outline of this thesis will follow the order described above. In chapter 1 we try to provide a general introduction to self-assembling monolayers and organic electronics focusing on the nature of the different types of SAMs and their characterization, and on organic thin-film transistors. While working towards our goal, we noticed that some existing SAM deposition and characterization techniques were unsuitable for our applications, so we had to adapt and improve them. This work is explained in chapter 2. On the other hand, our increased experience with SAMs led to a study of the solvent wettability of SAMs which proved very useful e.g. for processing organic materials to make organic devices and which is also discussed in chapter 2. In chapter 3 we will give an overview of our efforts, first, to improve and secondly to replace the various parts of an organic thin-film transistor (i.e. semiconductor, insulator and contacts) with selfassembled monolayers. The state-of-the-art and specific problems of SAMs and their application to improve or replace transistor parts, will be briefly discussed in each section. Finally we must note that the outline of this thesis does not really correspond to the chronological order in which the work was performed. After unsuccessful attempts to replace the organic semiconductor with a SAM in the beginning of this thesis, we shifted our attention towards the improvement of organic transistors using self-assembled monolayers which proved more successful.

2

1. Introduction This chapter intends to provide a general introduction to self-assembling monolayers and organic electronics. First, the concept of self-assembled monolayers, the different families and some characterization techniques will be introduced. In the second part we will focus on organic semiconductors and how these are used in devices like organic thin-film field-effect transistors.

1.1 Self-assembled monolayers A self-assembled monolayer (SAM) is a monomolecular, (ordered) layer or assembly which is spontaneously formed on an appropriate surface. The tendency of certain types of molecules to spontaneously form assemblies arises from their characteristic, amphiphilic chemical structure. Selfassembling molecules are typically composed of two groups: a head group which has high affinity for an appropriate surface and a tail group with a high affinity for similar tail groups of other, neighbouring molecules. When such molecules are brought in contact with the appropriate surface, the head groups of the molecules will physisorb or form a chemical bond to the surface (chemisorption). It is evident that the surface has to be clean because any contamination present will hamper or block interaction of the head group of the molecules with the surface. As the density increases, the tail group of the molecules on the surface will come closer and start to interact, typically giving rise to the order within the self-assembled monolayer. Often the molecules also contain a functional end group which allows for further physical or chemical functionalization. The typical structure of a self-assembling molecule is depicted in figure 1.1. Both the self-assembling and self-ordering properties of SAMs, combined with the presence of this functional end group and with

3

Chapter 1: Introduction

the SAMs stability (due to the surface bond) are the keys to a wide range of possibilities and applications such as surface engineering and surface modification for controlling adhesion, corrosion, lubrication, (bio)chemical sensing and mimicking etc. This is reflected in the strongly growing number of published SAM-related papers (e.g. in 1993 ~100 publications, in 2003 ~1100 publications).

functionalization

Self-assembling molecule “endgroup”

order “tail” assembly substrate

“headgroup”

Figure 1.1 Structure and interactions of self-assembling molecules.

Several ‘families’ of self-assembling molecules are known such as fatty acids on metals and metal oxides (e.g. AgO, Al2O3),1,2 alkanes on silicon3 and alkane phosphates on metals and metal oxides (e.g. Ti, TiO2, Ta2O5).4,5 The two most important families are the organosulfur and the organosilicon derivatives, both of which were used in this study and will be elucidated further in paragraph 1.1.1 and 1.1.2.

Before we proceed to these two families of self-assembling molecules, we would like to spend a few words on a very important, practical aspect of SAM formation for all families of self-assembling molecules namely cleaning of the substrate. Contaminants like adsorbed gases and impurities (e.g. fatty acids, salts, metals,…) are typically present on a surface and it is quite obvious that a surface covered with such contamination is not available for interaction with the self-assembling molecules. The importance and optimization of surface cleaning has been investigated in the PhD thesises of W. Laureyn6 and F. Frederix.7 In practice, several of the investigated cleaning methods are used in this thesis. The first and most used method is very effective and commonly used for wafer cleaning in the silicon industry. It is known as the ‘RCA clean’ and consists of a Piranha cleaning step followed by a “Standard Clean 1” (SC1).8 In the first step, the substrates are submerged in a freshly made 7:3 mixture of H2SO4 and H2O2 which will actively oxidize and remove all organic contamination from

4

Chapter 1: Introduction

the surface. The second cleaning step is performed in a 5:1:1 mixture of H2O, H2O2 and NH4OH at 70 °C which removes any remaining organic residues and particles. In another, often used and effective method, substrates are placed in a chamber with an ultraviolet (UV) lamp that produces two main wavelenghts (185 nm and 253 nm).9 The UV light forms molecular oxygen and ozone (O3) from the (oxygen-enriched) air in the chamber. As depicted in figure 1.2, this ozone and molecular oxygen will attack the organic contamination − which is also activated by the UV light − and remove it as small, volatile molecules (CO2, H2O,…). It is clear that this UV/ozone method is not able to remove inorganic particles and contaminants. It is however a dry method, which can be advantageous. The system used in this thesis is a homebuilt system with a mercury grid lamp (40-50 mW/cm²) of BHK Inc. CONTAMINANT hν1 (253.7nm) MOLECULES

O2 O3

hν2 (184.9nm) hν1 (253.7nm)

Excited Molecules Free radicals Ions Neutral Molecules O3

VOLATILE MOLECULES (CO2, H2O,…)

[O], O3

Figure 1.2 Schematic representation of the cleaning mechanism of a UV/ozone system.

One last method that has been used in this thesis is cleaning by oxygen plasma. A plasma is generated at reduced oxygen pressure typically by radio frequency (RF) or microwave irradiation. The substrate is then exposed to the plasma and reactive species in the plasma such as ions and atomic oxygen will attack organic contamination in a similar way as in a UV/ozone system. Depending on the power of the plasma, also some etching can occur due to the bombardment of the surface with electrons and ions. Other methods that are sometimes used in literature for cleaning substrates include many other types of reactive plasmas (e.g. Argon, CF4,…) and electrochemical cleaning in which a potential is applied to the substrate effectively oxidizing any surface contamination.

Cleaned metals and oxides typically have a very high surface energy and will tend to be recontaminated quickly.10,11 To prevent recontamination, it is best and most convenient to use cleaned substrates immediately or to store them in ultrapure water.12 This way, contaminants first have to dissolve and diffuse to the surface before they can recontaminate the substrates.

5

Chapter 1: Introduction

1.1.1 The organosulfur family or ‘thiols’ In 1983 Allara and Nuzzo reported the formation of a self-assembled monolayer from organic disulfides on gold13 which initiated a widespread interest in self-assembling organosulfur compounds. Nowadays the organosulfur family is by far the biggest and most widely used and studied family of self-assembling compounds. It includes organic alkanethiols, cyclic and aromatic thiols (also heterocyclic), sulfides, disulfides, thiophenes, thioureas, thiocarbamates and dithiolcarbonates (also called xanthates) among many others.14,15 The activity for this large family of sulfur-bearing, surfaceactive molecules can be pinpointed to the strong affinity of the sulfur atom for transition metals. As illustrated in figure 1.4, organosulfur compounds are known to coordinate to surfaces or colloids of gold, silver, copper, platinum, iron, nickel, GaAs, InP,etc.

14,15

They can be deposited from solution

and from gas phase but the former method is most commonly used. Y Y

Y

Y

S

S S

S

Y

Y

S

Au, Ag, Cu, Fe, Ni, GaAs,…

S H

Figure 1.3 Schematic representation of a typical dialkyldisulfide and alkanethiol and their preferential substrates.

The simplest, and thus archetypal example of the organosulfur SAMs are the (long-chain) alkanethiols on gold surfaces with a (111) crystal orientation. When an alkanethiol e.g. in solution, is brought in contact with a clean gold surface, the sulfur atom in the alkanethiol will form a bond to the gold surface. The strength of this bond was determined to be approximately 167 kJ/mol and is formed in an exothermic process (21 kJ/mol).16 The true nature of this bond and its formation is quite complex and controversial and has been the focus of many studies over the past two decades. It is generally accepted that the Au-S bond is thiolate(RS-)-like with extensive covalent character as the sulfur atom appears to have a net negative charge of –0.2e from XPS studies and simulations.17,18,19 The mechanism involved in the reaction of the alkanethiols with gold(0) may be considered as an oxidative addition of the S-H bond to the gold surface, followed by a reductive elimination of the hydrogen atom, probably via H2 release:

6

Chapter 1: Introduction

R-S-H + Aun(0) → R-S-Au(I) + Aun-1(0) + ½ H2 SAMs formed from dialkyldisulfides appear to be indistinguishable from the corresponding alkanethiol SAMs and their adsorption energy was calculated to be ~100,5 kJ/mol.16 A simple oxidative addition of the S-S bond to the gold(0), effectively cleaving the disulfide and yielding two thiolates, is believed to be the mechanism:20 R-S-S-R + Aun(0) → 2 R-S-Au(I) + Aun-1(0) Organosulfur monolayers show the highest order on gold with a (111) crystal orientation.21 This orientation is the lowest energy surface and is preferred in the growth of thin Au films e.g. by slow evaporation on smooth surfaces such as silicon or mica.22 If necessary, larger and smoother terraces with a (111) orientation can be obtained by flame annealing23 or template stripping24 of Au polycrystalline thin films. Morphologies of typical gold surfaces are shown in figure 1.4. Alternatively, but much more expensive, Au(111) single crystal substrates can be used.

25 nm

200 nm

a)

0

125 Å

b)

100 nm

0

10.0 Å

0

40 Å

Figure 1.4 Scanning tunneling microscope image of a) a polycrystalline gold film and b) gold terraces. Early structural studies of alkanethiol SAMs on Au(111) by diffraction techniques,25,26,27 modelling28 and scanning tunneling microscopy29 revealed a crystalline, hexagonal (√3x√3)R30° order in the SAM which is commensurate with the underlying Au(111) structure. A closer analysis and more refined techniques showed a much more complex, subtle structure: two (or three30) different molecular states in the unit cell were discovered which break the simple hexagonal symmetry and give rise to a c(4x2) superlattice.21,31,32 But the alkyl backbone alone cannot explain the superlattice

7

Chapter 1: Introduction

and some studies indicate that the sulfur atoms are not perfectly hexagonally commensurate with the Au(111) but are slightly shifted. Figure 1.5 illustrates this surface structure of thiol SAMs.

a)

b)

c)

d)

e)

(111) FCC Figure 1.5 a) Face cubic centered structure of gold with the (111) intersect; b) the resulting Au (111) surface with the (1x1) unit cell; c) the sulfur atoms (dark gray dots) of the adsorbed alkanethiol layer have a (√3x√3)R30° structure commensurate with the underlying Au(111); d) the p(3x2√3) superlattice which is generally written as c(4x2) with the plane of the all-trans alkylchain indicated by the black diagonal in the dark gray dots; and e) STM topograph showing an octanethiol monolayer with a c(4x2) reconstruction on Au(111).33

Another point of discussion is the exact place of the sulfur-gold bond: four symmetrical positions can be discerned (atop, bridge, fcc hollow and hcp hollow) and originally it was assumed that the sulfur atom was bound in one of the hollow positions.25,26,34 More recent experiments and simulations suggest that the bond ideally occurs in the atop position but depending on the chain length and type it will deviate from this32,35,36 and even forms some sort of ‘dimer’ with other thiolates.37 Other crystal orientations (e.g. (001)) or other materials (e.g. Ag, Cu) are less common and will yield a different SAM structure due to differences in crystal lattice structure of the substrate and surface affinity of the sulfur atom for this substrate.15 Kinetic studies of the organosulfur monolayer formation revealed again a complex process with many intermediate phases but it can be roughly simplified into two steps:15,30,38 In the first step, which is very fast (minutes), molecules chemisorb on the surface. This step is driven by the formation of the strong gold-sulfur bond and is practically rate-limited by the diffusion of molecules to the surface. It is therefore strongly dependent upon organosulfur concentration and the whole process can be well described by diffusion-controlled Langmuir adsorption. Initially the molecules are lying flat on the surface in a disordered way but as the monolayer density increases, the alkyl-chains start to interact and form more ordered domains. This crystalline, flat-lying phase is called the “striped phase”. Figure 1.6 shows ordered, striped domains (A) in a disordered phase (B) after 4 minutes in a thiol solution.

8

Chapter 1: Introduction

(a)

(b)

Figure 1.6 a) Different steps in the formation of an alkanethiol SAM with the formation of the striped phase as a first step and the final closely packed phase as the second step; b) STM image (107 nm x 107 nm) of the striped phase (A) in a disordered phase (B).30

As the density continues to increase, the flat-lying phase will reach its saturation coverage. In what is considered as the second step, some molecules will stand up and slowly ‘crystallize’ into upstanding, ordered domains or ‘islands’ via some intermediate phases. These upstanding, crystalline domains gradually grow until the complete monolayer is transformed into the “standing-up” phase. This process is slow, in the timescale of hours or days, and is mainly driven by favourable Vanderwaals-interactions between the alkyl chains (4-8 kJ/mol per methylene unit).39 In the final phase, the distance between the sulfur atoms is about 4,97 Å based on the (√3x√3)R30° structure.25 The optimal alkyl chain spacing, however, is about 4,3 Å which is only possible if the chains are roughly tilted over 30° with respect to the surface normal. And indeed, this tilt was experimentally confirmed by Fourier-transform infrared (FTIR) spectroscopy.40 The kinetic process of SAM formation points to another peculiar property of the thiolate-gold bond: given enough thermal activation energy, it can easily move on the surface41,42 which allows the SAM to adapt its structure and minimize its energy for a given surface density. Similarly, structural defects in the SAM order will be repaired to lower overall energy and this interesting property is often called ‘self-healing’. Also, organosulfur SAMs should not be seen as static systems. Studies have shown that alkanethiols can desorb and exchange with other alkanethiols in solution or gas-phase, usually via dimerized disulfide.43 A organosulfur SAM is thus a dynamic system that is in equilibrium with molecules in the liquid or gas above.

Given the wide variety and the remarkable and versatile properties of organosulfur compounds, it is not surprising that these systems have found wide application.14,15 They are used as protective coatings e.g. against corrosion, as coatings that improve adhesion, wetting or friction, as patterning layers (by microcontact printing), as optical coatings e.g. for non-linear applications and − last but not least − as anchoring layer for attaching other layers, particles, molecules… in e.g. seed layers,

9

Chapter 1: Introduction

biosensors etc. Finally, organosulfur SAMs are also used in organic and molecular electronics as we will show in chapter 3.

For a more detailed treatment of organothiol SAMs, we gladly refer to some excellent papers in literature.44,45,46

1.1.2 The organosilicon family or ‘silanes’ Alkylchlorosilanes and alkylalkoxysilanes are the most prevalent members of the family of selfassembling organosilicon-derivatives. Although these molecules and their properties were already known for several decades, their self-assembling and monolayer-forming properties were firstly noticed by Sagiv in 1980.47 Similarly to organosulfur SAMs, alkylsilane-based SAMs are typically formed from solution48 but gas-phase deposition has also been shown.49,50 The typical structure of these alkylsilane compounds is shown in figure 1.7a. In the case of alkylchlorosilanes, the silicon atom bears at least one chloro group. For alkylalkoxysilanes, there is at least one alkoxy group (typically methoxy or ethoxy) connected to the silicon atom. These functions are responsible for the surface affinity of the head groups as they can form a bond with hydroxyl-terminated surfaces through a condensation reaction. For example a siloxane is formed through a strong, covalent Si-OSi bond upon reaction of these silane molecules with the surface silanol groups (-SiOH) present on a silicon dioxide surface. The tail part of these organosilicon compounds usually consists of an alkyl chain but can be aromatic too. Similar to organosulfur SAMs, interchain interactions will drive the ordering of the molecules. The terminal group, indicated with Y in figure 1.7, can be a methyl group but it can also be a functional group that allows further chemical modification. It is evident that terminal groups which are reactive towards the alkoxy or chloro group on the silicon atom cannot be used as this would result in inter- and possibly intramolecular polymerization. This unfortunately excludes the practically useful polar groups like carboxylic acids, alcohols and amines, and limits the range of terminal functions to more apolar functions such as a bromo, a vinyl or a nitrile group. As mentioned these organosilicon molecules can self-assemble on any surface bearing hydroxyl groups which includes many practically important surfaces such as glass, quartz, silicon dioxide and metal oxides (e.g. Al2O3, Ta2O5,…) but also mica, zinc selenide and even oxidized polymers.14,51 Again, a thorough and preferably oxidative cleaning is essential as it will uncover and create additional surface hydroxyl groups improving SAM formation.

10

Chapter 1: Introduction

a)

b)

Y

Y

Y

Y

c) Y

Y

Chlorosilanes: H20

with R= -Cl, R’= -Cl or -CH3

H20

Alkoxysilanes: with R= -OCH3, -OCH2CH3 R’= -OCH3, -OCH2CH3 or -CH3

R'

Si R' R

HCl OH

R'

Si Cl

R'

OH OH

H20 OH

R'

Si

R' OH

OH OH

Hydroxyl surfaces (SiO2, Ta2O5,…)

R'

OH

Si

R' OCH3 Si R' R' OH O OH

Si R' R' OH O OH

Hydroxyl surfaces (SiO2, Ta2O5,…)

Figure 1.7 a) Structure of self-assembling organosilicon derivatives, b) hydrolysis and condensation reaction c) resulting in the immobilization of the molecule on preferred substrates.

The surface bond arises between the organosilicon head group and the hydroxyl groups on the surface. Alkylchlorosilanes are very reactive and will hydrolyse rapidly with any trace of water present to form silanol groups. These silanol groups can then react with surface hydroxyl groups by a condensation reaction eliminating water. A direct condensation reaction of the chlorosilane with surface hydroxyl groups is also possible with HCl as byproduct. In both cases, the alkylchlorosilane is covalently anchored to the surface. Alkoxysilanes will not directly react with surface hydroxyl groups due to their relatively low reactivity. They first require hydrolysis of the alkoxy group, often acid assisted, to form a silanol group which can then similarly condense with surface hydroxyl groups. The main problem with alkoxy- and chlorosilanes is that the molecules in the hydrolyzed silanol state can react both with the hydroxyl groups on the surface and with each other. Especially during deposition from solution, the molecules have a high probability encountering each other rather than the surface. This will result in the formation of polymerized clusters of molecules which grow on the surface or grow in solution and subsequently bind to the surface. This is not the case for monofunctional silanes where an intermolecular reaction results in a dimer which is unable to bind other molecules or the surface. However, monofunctional silanes typically have, in lieu of chloro groups, two bulky methyl groups on the silicon atom which sterically hinder a close packing and lead to low-density monolayers. The two-step SAM formation mechanism and kinetics of organosilane compounds largely corresponds to that of thiol SAMs described earlier. However, due to the different head group surface chemistry for organosilane molecules there are some complications: first, the head group has up to three binding silanol groups and is larger compared to a sulfur atom. Secondly, the

11

Chapter 1: Introduction

formation of these silanol groups requires the presence of water. Third, when a bond is formed, it is irreversible i.e. the molecules become immobilized and cannot move or desorb unlike organosulfur compounds on gold. And finally, compared to e.g. Au(111) most hydroxyl-group bearing surfaces like oxides are amorphous and have a relatively low density of binding sites (~1015 surface hydroxylgroups per cm² i.e. ~1 per 20 Ų).52 It is therefore not surprising that the exact SAM formation process is still controversial and not fully understood yet. There is however some consensus: for chlorosilanes, the molecules hydrolyze in solution or in the ultrathin, adsorbed moisture layer which is typically present on clean oxide surfaces and depicted in figure 1.8a. This adsorbed water layer has been studied and seems to be important in the SAM formation process.53,54 Complete hydrolysis of the chlorosilane groups was confirmed by XPS and infrared spectroscopy.10,55 The hydrolyzed molecules are loosely held in place on the surface by intermolecular and surface hydrogen bonds as depicted in figure 1.8b.

a)

b)

Figure 1.8 a) Organosilane adsorption and hydrolysis in the adsorbed moisture layer, b) before condensation, the molecules form hydrogen bonds.56

Next, the condensation reactions occur. Based on the observation by Wasserman et al. that no chlorine could be detected in the complete SAM, one could assume that all silanol groups are bound to surface hydroxyl groups as depicted in figure 9a but the density of surface hydroxyl groups is too low to support this.57 Wasserman et al. and others then suggested that for trichlorosilanes one silanol group simultaneously binds to the surface while the two remaining silanol functions bind to neighbouring molecules creating cross-links.14,56,57 Others suggested that first the molecules assemble and then condense with each other forming a cross-linked two-dimensional polysiloxane network that ‘floats’ in the adsorbed water layer.58 Upon thermal treatment, this two-dimensional network then undergoes a condensation reaction with surface hydroxyl groups and thus becomes bonded to the surface. These models were supported by the observations that low quality monolayers were formed under strict anhydrous conditions with dried substrates,57,58,59 that intermolecular siloxane bonds were observed60 and that structurally similar SAMs were formed on

12

Chapter 1: Introduction

very different substrates.58 Figure 1.9b shows the typical representation of a silane SAM according to these two models.

a)

b)

Si O

O

Si O

O

O

Si O

O

O

O

Figure 1.9 Models of an organosilane SAM a) in which all silanol groups are surface-bound and b) in which each silane molecule is bound to the surface and to neighbouring molecules by two crosslinks.

These models can however not be correct: as Sagiv already suggested in 1980, the bond length of the intermolecular Si-O-Si cross-link is too small (4.1 Å ).47,61 Detailed structural calculations by Stevens indeed show that high-density SAMs can not be formed when extensive cross-linking occurs and when the head group is too large (for example in dimethylchlorosilanes).62 He calculated that silanol groups are small enough to allow a high-density SAM packing and suggested that two of the three hydrolyzed silanol functions do not react and remain pendant. Experimental data from Parikh et al. who showed an increase in SiOH infrared absorption with increasing SAM coverage, 60 and from Wange et al.61 seem to support this. Possibly, these unreacted silanols are hydrogen-bonded to water at the surface, to surface hydroxylgroups or to neighbouring silanol groups. To conclude, figure 1.10 shows the most probable mechanism for organosilane SAM formation in which preferably only one chloro or silanol group reacts with a surface hydroxyl group. Cross-linking can occur but will give rise to defects and a lower density of the SAM.

13

Chapter 1: Introduction

Figure 1.10 Organosilane SAM formation process.61

As one can easily understand, the trivalent head group, its sensitivity to water and its possibility to cross-link pose some practical difficulties for silane SAM deposition compared to organosulfur SAMs. Any trace of water present e.g. in solvents, in the atmosphere or on the surface of recipients will give rise to polymerization. For chlorosilanes, polymerization can mostly be avoided by working under strictly anhydrous conditions. For alkoxysilanes, it is simply impossible to avoid polymerization as their low reactivity requires adding water (and acid) which aids hydrolysis but also promotes polymerization. The irreversible, strong bond of organosilane SAMs also has advantages: they are mechanically61 and thermally very robust. For example, organosilane SAMs begin to oxidize in air around 200°C but oxidation under vacuum or nitrogen only begins around 250°C and 300°C respectively.63 Organosulfur SAMs are more sensitive to oxidation as the binding sulfur atom is easily oxidized to a non-binding sulfonate. Thiol SAM desorption is observed to start around 100°C16 and to be complete around ~210°C.16,20,64

The combination of self-assembling properties and a strong surface bond on practically highly relevant surfaces like glass or metal oxides makes the organosilane family well suited for a wide range of applications. Similar to organosulfur SAMs, they are used as coatings that improve or prevent adhesion, wetting or friction e.g. in micro-electromechanical systems,65 as patterning layers (by microcontact printing), as optical coatings e.g. for non-linear applications and as anchoring layer

14

Chapter 1: Introduction

for attaching other layers, particles, molecules… in for example (bio)sensors.66,67 Finally, organosilane SAMs are also used in organic and molecular electronics

68,69,70

as we will show in

chapter 3.

1.1.3 Characterization of SAMs The characterization of self-assembled monolayers is non-trivial: because of its molecular dimensions, a SAM is 'invisible' for any form of conventional microscopy. Therefore other techniques are required which typically probe only one property of the SAM. To get a more complete picture of the structure, order and properties of a SAM, application of several techniques is often necessary. In the following paragraphs we will briefly elucidate the techniques that have been used in this thesis.

1.1.3.1

Contact angle measurements

A simple and rather fast method for evaluating surface properties of a (SAM-modified) surface is the measurement of the contact angle (CA) of a solvent droplet, typically water, with this surface.71 Such wettability measurement gives a direct indication of the surface hydrophobicity or hydrophilicity. The contact angle of the liquid with the solid surface is determined by the Young equation :72

γsv = γsl + γlv cosθ

(1.1)

and is schematically depicted in figure 1.11.

a)

syringe

b)

γlv

droplet γsv

Figure 1.11: a) Picture of a water droplet on a surface; b) schematic representation of the liquid contact angle θ and the respective interfacial energies.

15

Chapter 1: Introduction

where θ is the measured contact angle and γ is the surface energy of the solid-vapor (sv), solidliquid (sl) and liquid-vapor (lv) interface. The unit for γ is J/m² but this can also be written as N/m. In the latter case we speak of a surface tension which is typically used for liquids while in the former we speak of a surface energy. Young’s equation however is only valid under thermodynamic equilibrium. This practically means that the system has to be static: the drop is not moving and the drop volume is not changing. Dynamic contact angle measurements in which the drop volume is varied, yield different contact angle values (namely the advancing and receding contact angle) and from these values and mostly their difference (the so-called hysteresis) surface properties can be deduced. Dynamic CA measurements are experimentally more difficult to perform accurately, especially on hydrophilic surfaces and therefore we will mainly focus on static contact angle measurements in this thesis. The experimental setup used in this thesis is an OCA-20 instrument from Dataphysics with SCA 20 software. This setup provides 4 motorized syringes for volume-controlled droplet deposition. A videocamera and accompanying software allows droplet imaging and contact angle extraction with various algorithms.

1.1.3.2

Ultraviolet/Visible absorption and fluorescence spectroscopy

In ultraviolet and visible (UV/VIS) spectroscopy, light of various wavelenghts (energies) is passed through a sample and the amount of absorbed light is recorded as a function of its wavelength. Practically, the wavelength range denominated by UV/VIS and measureable in air (due to atmospheric absorption) starts around 185 nm and extends up to 800 nm. Absorption of ultraviolet and visible radiation by an atomic or molecular species can be attributed to the excitation of electrons, generally from bonding orbitals and thereby creating a so-called ‘excited state’, schematically depicted in figure 1.12. The energy of the absorbed radiation can therefore be correlated with the types of bonds present in the studied species and help identify functional groups in a molecule.

16

Chapter 1: Introduction

ABSORPTION excited state S

EMISSION

4 3 2 1 1 0

ground state S0

4 3 2 1 0

wavelength Figure 1.12 Typical absorption and fluorescence spectrum and in overlay, electronic transitions between different vibrational levels of the ground state and the excited state.

Even more important is the application of UV/VIS absorption spectroscopy for quantitative determination of compounds containing absorbing groups as Lambert-Beer’s law describes a linear relationship between absorbance and concentration:

log(I0/I) = -logT = A = bcε

(1.2)

where I0 and I the intensity of the incident and respectively the emergent light beam, T the transmittance, A the absorbance, b the path length (in cm), c the concentration (in M or mol/l) and ε the molar extinction coefficient (in M-1cm-1). This is also schematically ilustrated in figure 1.13.

cε I0

I b

Figure 1.13 Schematical depiction of an absorption measurement of a solution with concentration c and molar extinction coefficient ε in a cuvet with pathlength b.

For a thin film, this can be rewritten as A = tDε where t is the film thickness (in cm), D (in mol/l) the molar density (i.e. the mass density ρ divided by the molecular mass) and ε the molar extinction coefficient (in M-1cm-1). For a monolayer, this can be rewritten as A = σε where σ is the molar

17

Chapter 1: Introduction

surface density (in mol/cm²) and ε again the molar extinction coefficient (in M-1cm-1). The instrument used in this thesis is a double-beam spectrophotometer UV-1601PC from Shimadzu.

Fluorescence spectroscopy studies the intensity and spectral properties of light emitted by a sample upon optical excitation. As mentioned above, absorption of light excites electrons creating an excited state. This excited state can decay to its ground state by emission of the excess energy as a photon. It is evident that this emitted photon cannot have a higher energy (or lower wavelength) than that of the excitation light. The energy of the emitted photon can again be related to the types of bonds and functional groups present in the studied species.

Similarly, the amount of emitted photons can be quantitatively related to the concentration of a species in a sample. Compared to absorption spectroscopy, fluorescence spectroscopy typically offers a higher sensitivity (up to 3 orders of magnitude) and wider linear concentration ranges but it is less widely applicable as not all absorbing compounds fluoresce. The instrument used in this thesis is a RF-5301PC spectrofluorometer from Shimadzu.

1.1.3.3

Fourier-transform infrared spectroscopy

Infrared (IR) spectroscopy is a widely applied technique in chemical and fysico-chemical research. It probes the vibrational behaviour of chemical bonds and interactions within and between molecules by looking at the wavelength-dependent attenuation of an incident infrared lightbeam. When the energy of the incident beam corresponds to the vibrational energy level of a bond, resonance occurs and energy from the incident beam is absorbed which is then detected. In principle, Beer’s law is followed and the amount of absorption can then be related to e.g. a concentration allowing for quantitative measurements. In practice however, deviations from Beer’s law often occur, both fundamentally and instrumentally, and meticulous care and calibration is required to obtain a similar accuracy and precision compared to those obtained with UV/VIS methods. Qualitatively, the spectral position of the absorption band can be related to the type of chemical bond while small deviations from these typical positions give information about the chemical nature surrounding this bond such as molecular conformation or intermolecular order. Currently, Fourier-Transform IR (FTIR) spectrometers use interferometry and Fourier transformation to increase speed, signal-tonoise ratio and resolution resulting in fast and highly sensitive measurements compared to conventional dispersive instruments. For solutions and powders, simple transmission spectroscopy can be used but thin films on surfaces like self-assembled monolayers require special setups such as attenuated total reflection (ATR),

18

Chapter 1: Introduction

Brewster-angle (BA) transmission or grazing-angle IR reflection and absorption (GA-IRRAS). Only the latter technique is available in Imec and was used in this thesis. In GA-IRRAS, an incident IR beam impinges under a low, grazing angle and is reflected off the sample where part of the light is absorbed during the reflection by any material e.g. a SAM present on the sample. Compared to ATR or BA, GA-IRRAS is less complex and allows the use of simple one-sided, flat substrates. This method however has several disadvantages. First, the sensitivity is rather low compared to ATR or BA as the lightbeam passes only once through the adsorbate. Secondly, refractive substrates like oxides have a very low reflectivity of the incoming IR radiation resulting in a very low signal. Therefore we are restricted to substrates with specular reflection like metals. Finally, for conductive substrates like metals there is the so-called ‘surface dipole selection rule’ which states that only vibrations giving rise to an oscillating dipole perpendicular to the surface can be detected. This is due to the fact that any electric field (e.g. from an incident lighbeam or a vibrating dipolar bond) which is parallel to a conductive surface, will be compensated by an induced opposite field in the metal. Nevertheless this technique is well-suited to characterize for example self-assembled thiolmonolayers on gold surfaces via the identification of typical vibrations of functional groups. Moreover detailed analysis of the band position of functional groups such as methylene (CH2) hints about the conformational order and packing density of the monolayer. Well-ordered thiol-SAMs on gold are known to have an antisymmetric CH2 vibration band around 2918 cm-1 while a band at 2926 cm-1 and higher is observed for disordered, ‘spaghetti-like’ monolayers.48 The FTIR instrument used in this thesis is a Mattson Galaxy Series 7000 spectrometer with a Spectra-Tech FT 85 accessory for grazing-angle measurements. Recently this instrument was replaced by a Bruker IFS66v spectrometer with a grazing angle accessory from Specac.

1.1.3.4

Cyclic voltammetry

Cyclic voltammetry (CV) is an electrochemical technique which is widely used for surface characterization.73 A linear potential change is cycled between two potential values (figure 1.14b) and applied to a working electrode and a counter electrode, both referenced to a reference electrode and all submerged in an unstirred electrolyte solution (figure 1.14a). Any current resulting from the potential sweep is detected and recorded (figure 1.14c).

19

Chapter 1: Introduction

Working electrode

Reference electrode

Counter electrode Reference electrode

a)

Counter electrode Working electrode

b)

c)

d)

Figure 1.14 a) Schematic cyclic voltammetry setup and b) image of the electrochemical cell used at Imec with c) the applied voltage sweep and d) the resulting voltammogram.

If an inert electrolyte is present, only a capacitive current will flow which can be used to determine for example the thickness of any insulating thin film (e.g. a SAM) present on the working electrode. When a redox couple (such as hexacyanoferrate/hexacyanoferrite (Fe(CN)63-/Fe(CN)64-), ferrocene/ferrocenium (Fe(C5H5)2)/ Fe(C5H5)2+) is present in the electrolyte, also a faradaic current due to electron transfer (oxidation and reduction of the couple) is detected. When a SAM is deposited on the working electrode, the electron transfer with the redox couple will be hampered and the measured current will be lower. The measured peak current (ip in A) can be related to the electrode area (A in cm²) via the Randles-Sevcik equation :

ip = (2.69x105) n3/2 A C D1/2 v1/2

(1.3)

where n is the number of moles of electrons transferred in the redox-reaction, C is the analyte concentration (in mol/cm³), D is the diffusion coefficient (in cm²/s), and v is the scan rate of the applied potential (V/s). This allows qualitative analysis of the density, stability and presence of defects in the self-assembled monolayer with respect to the type of assembled molecules and deposition parameters.

20

Chapter 1: Introduction

When a redox-active group is incorporated in a SAM on the electrode, the amount of transferred electrons (which is the area under the curve) can be directly related to the SAM density. The main disadvantage of cyclic voltammetry is the need for a conductive substrate but nevertheless it is an excellent technique for the study of reactions, mechanisms and kinetics of redox processes, and of self-assembled monolayers such as thiols on gold. The setup used in this thesis consists of a home-made electrochemical cell, depicted in figure 1.14b, with a Ag/AgCl reference electrode placed in a Luggin capillary and a Gamry PC3 potentiostat with CMS 100 and 130 software. Experiments were performed in a solution of 10-2 M K3Fe(CN)6 and 1M KCl with a scan speed of 100mV/s unless specified otherwise.

1.1.3.5

Scanning probe microscopy: atomic force and scanning tunneling

Scanning probe microscopy is relatively new and quite different from conventional microscopies: a sharp probe is brought into close proximity with a sample under study and by laterally scanning this probe over the sample, a physical interaction between the probe and the sample is detected and recorded to obtain a 2D ‘image’ of the sample. The use of piezo-actuators allows sub-nanometer resolution in the XY-plane and in the height Z. In the case of scanning tunneling microscopy (STM) the tunneling current between a conductive tip and sample is measured for a certain potential difference. The tunneling current is exponentially dependent on the tunneling gap (i.e. the tip-sample separation) enabling sub-angstrom height resolution. Single molecules and even atoms can be resolved by STM. Figure 1.15 schematically shows the operating principle of STM with, for example, an obtained image of the typical hexagonal packing of carbon atoms in graphite (figure 1.15b).

Detector

A Tip

Laser

V

Tip

Tunnel current 1 nm

a)

Sample surface

b)

0

3.5 Å

c)

Sample surface

Figure 1.15 a) Schematic operating principle of STM and b) an STM image of graphite showing the hexagonal packing of carbon atoms. The scale of the image is 7 by 7 nm. c) Schematic operating principle of AFM.

21

Chapter 1: Introduction

In atomic force microscopy (AFM), repulsive and attractive forces between tip and sample are measured by detecting the deflection of a laser beam focused on the backside of a tip as shown in figure 1.15c. AFM can be operated in different modes (e.g. contact, tapping and non-contact), each having its own characteristics. In this thesis we used tapping mode AFM where the tip is oscillated perpendicular to the sample and the tip-sample forces are detected via changes in oscillation amplitude and frequency. This method combines relatively high resolution with low applied force for measurement of soft materials. In contrast to STM, AFM does not require conductive tips and samples but the different nature of the interaction and the ‘large’ curvature radius of AFM tips (~10 nm) limits the AFMs resolution. Nevertheless both microscopies are excellent for imaging where conventional, optical microscopies fail. In this thesis, we used a PicoSPM system from Molecular Imaging with a PicoScan 2100 controller and Picoscan 4.19 and 5.2 software which is capable of performing both STM and AFM.

1.1.3.6

Other techniques

There are many other techniques for characterizing SAMs that have not been used in this study. For completeness, we briefly mention X-ray photoelectron spectroscopy (XPS) which is often used in literature and allows the detection of elements present on the surface (like Si, C, O, N,…) and their oxidation state. This semi-quantitative technique can be useful to study surface cleanliness, monolayer deposition and surface reactions. Another interesting but rather complicated technique for surface characterization is ellipsometry. Analysis of the polarization of the wavelength- and angle-dependent reflection of a lightbeam on a substrate allows the determination of the refractive index and thickness of layers like SAMs on this substrate.

1.2 Organic

semiconductors

and

thin-film

field-effect

transistors In this chapter, we will try to provide a general introduction to (organic) semiconductors and transistors in a rather intuitive, conceptual way. Paragraph 1.2.1 will briefly introduce the concept of semiconductors for novice readers. Paragraph 1.2.2 then elaborates on organic semiconductors, their properties and applications. For a more thorough, physical introduction we gladly refer to the PhD thesis of S. Verlaak.74 Finally in paragraph 1.2.3 we will introduce the physical basics of organic

22

Chapter 1: Introduction

thin-film transistors, one of the main applications in the field of organic semiconductors and therefore the focus in this thesis.

1.2.1 Introduction Materials can be categorised into conductors, semiconductors or insulators depending on their ability to conduct electricity. Macroscopically, the conductivity (σ) relates the field (E) applied over the material with the resulting current density (j) which itself can be related to the density (n) and charge (q) of the charge carriers and their mobility (µ):

j = σ E and σ = n µ q

(1.4)

Microscopically, a material will show conductivity if mobile, free charge carriers are present in the material. These free charge carriers will move through the material when an external field is applied. Charge carriers can become mobile when sufficient thermal energy and incompletely filled energy levels slightly above the occupied energy levels are available in the material. Evaluation of this last condition requires knowledge of the electronic structure of the material. In classical, crystalline solid-state physics, a material is considered to be a three-dimensional, periodic lattice of atoms and ‘band theory’ is used to deduce and describe the electronic structure of solids. Band theory assumes that electrons in the material are subject to a periodic potential due to the atoms in the solid and are confined by potential barriers at the materials edge. Calculation of the energy levels in the material based on this model shows that these energy levels are grouped in bands that are separated by energy band gaps. In view of the condition for electrical conduction mentioned earlier, a fully occupied band without energetically nearby, empty bands cannot conduct electricity. Also, an empty band, not containing mobile charge carriers, does not conduct electricity. To further simplify the analysis, only the highest occupied and lowest unoccupied band are of interest for electrical conduction as the lower energy bands are completely filled and therefore cannot contain mobile charge carriers. The highest occupied band is called the valence band (VB; after the valence electrons that populate this band) and the lowest unoccupied band is called the conduction band (CB). Let us now compare the electronic band structure of conductors, semiconductors and insulators which is depicted in figure 1.16.

23

Chapter 1: Introduction

empty levels

filled levels donor acceptor

CB

doping

20eV >3eV

0,2-3eV

∆T VB Conductor

Insulator

Semiconductor

Figure 1.16 Schematic electronic band structure of a conductor, an insulator and an intrinsic (∆T) and extrinsic (doping) semiconductor. For conductors, there are two possible scenarios: a half-filled valence band, which occurs in materials consisting of atoms with only one valence electron per atom such as copper and silver, or a filled band overlapping with an empty band. In both cases, electrons can easily be promoted to a nearby, empty energy level and thus can move giving rise to electrical conduction. For insulators, the valence band is completely filled and separated from the next higher empty band by a large energy gap (>3 eV). There can be no mobile charge carriers present and therefore insulators do not conduct electrical current. A similar electronic structure can be seen for semiconductors but the band gap between valence and conduction band is much smaller (typically 0) and the energy released in this process (∆G) can be written as the number of transferred molecules (n) multiplied by the energy released per transferred molecule (∆µ) and corrected for the energy required to create additional crystal surface area:

∆G = −n∆µ +

crystal faces

∑ (A Γ ) i i

i

(3.1)

with Γi the specific free surface energy per area and Ai the additional surface area of each crystal face i (which is function of the number of added molecules). The specific free surface energy per area Γi is function of Ψi which is the interaction energy between a molecule and a neighbouring molecule or surface along the crystal direction indicated by the subscript. One of the more important factors in this last term is the interaction energy with the substrate, indicated by Ψmol-sub. In figure 3.3, the formation energy ∆G is plotted for a nucleus growing molecule per molecule for a certain supersaturation µ and this shows that a critical nucleus size needs to be overcome before a stable nucleus is created.

91

Chapter 3 : Application of self-assembly in organic field-effect transistors

Critical nucleus

Stable nucleus

Figure 3.3 Formation energy of a growing nucleus by adding molecule per molecule (at a constant supersaturation (illustration by courtesy of S. Verlaak).

The minimal dimensions of this critical nucleus and thus its thermodynamically most stable shape can be derived from equation 3.1. For the ‘height’ of the nucleus, two types of nuclei can be discerned depending on the supersaturation and interaction energies in direction normal to the substrate: nuclei in which the height scales with the lateral dimensions and which thus grow threedimensionally, and nuclei which grow as a monolayer (i.e. 2D nuclei). Analysis of the formation energy ∆G in function of the supersaturation ∆µ, shown in figure 3.4, indicates that 2D growth becomes possible above a critical supersaturation ∆µcrit and will be favoured above a certain transition supersaturation ∆µtr while 3D growth is possible for all supersaturations.

∆µ=0 ∆µcrit ∆µtr

* ∆G [e V ]

1000 100 10

*

∆G 3D

*

∆G 2D

1

0.2

0.4

0.6

∆µ [eV]

0.8

1.0

Figure 3.4 Energies of formation as a function of supersaturation ∆µ for forming 2D nuclei above critical supersaturation ∆µcrit (dash-dotted line) and 3D nuclei (solid line below and dashed line above transition supersaturation ∆µtr) (illustration by courtesy of S. Verlaak).

92

Chapter 3 : Application of self-assembly in organic field-effect transistors

For practical evaluation and application of this model, values for the interaction energies are needed and the supersaturation needs to be linked to experimental deposition parameters. S. Verlaak calculated the intermolecular interactions for the different crystal directions for various materials (pentacene, perylene and tetracene) by molecular mechanics with a MM3 force field. The supersaturation ∆µ can be linked to the flux Φ and substrate temperature Tsub with the following equation:

∆µ ≈ RTsub ln(

2 πMRTsub Φ P∞ ( Tsub )

)

(3.2)

where R is the universal gas constant, M the molar mass and P∞ the equilibrium vapor pressure at Tsub which can be found in literature.13,14 Based on this model, it is then possible to theoretically interpret and predict the deposition conditions for 2D and 3D growth e.g. of pentacene. Assuming there is no interaction with the substrate, the different growth regimes for varying substrate temperature Tsub and flux Φ can be calculated and are depicted in figure 3.5a. Above the solid line (where ∆µ=0), no growth will occur while between ∆µ=0 and the critical supersaturation ∆µcrit (dashed line) the growth mode will be 3D. Between ∆µcrit and ∆µtr (dotted line) both 3D and 2D growth are possible but 3D growth will be favoured. Finally, below ∆µtr preferred 2D growth occurs. When these conditions are calculated for a ‘real’ substrate like UV-ozone cleaned silicon dioxide (with a pentacene-substrate interaction energy of 0.105 eV) and compared to experimental results, they correspond well as is shown in figure 3.5b: the filled circles show experimentally determined growth conditions which yielded continuous, 2D films while the crosses yielded noncontinuous 3D films as predicted by the model.

93

Chapter 3 : Application of self-assembly in organic field-effect transistors

∆µ=0

∆µcrit

∆µ=0 ∆µtr

∆µtr

a)

b)

Figure 3.5 Parameter space for substrate temperature Tsub and flux Φ yielding different growth modes for pentacene a) for an ideally inert substrate (Ψmolecule-substrate=0): no growth above the solid line (∆µ=0); 3D growth between ∆µ=0 and the critical supersaturation ∆µcrit (dashed line); 3D and 2D growth possible between ∆µcrit and ∆µtr (dotted line) but 3D will be favoured; and preferred 2D growth below ∆µtr. For a UV-ozone treated silicon dioxide substrate (b), the experimentally observed film morphologies (full circles: 2D growth, crosses: 3D growth) correspond well with the model.

The difference between the plots in figure 3.5a and 3.5b also indicates the importance of the substrate interaction Ψmol-sub on the growth mode and thus the film morphology according to the nucleation growth model. Around the time of these experiments, we finished our work on implementing the vapour phase method for SAM deposition. This allowed us to modify substrates with a wide range of different, smooth surface treatments (and thus different molecule-substrate interactions) which we then used to study the influence of these surface treatments on the growth morphology of pentacene thin films. All SAMs were deposited from the vapor phase at standard deposition conditions mentioned in paragraph 2.1.2. For a constant flux, S. Verlaak determined the substrate temperature (~supersaturation) at which the film morphology changes from 2D to 3D for substrates modified with the different SAMs. These substrate temperatures are shown in table 3.1.

94

Chapter 3 : Application of self-assembly in organic field-effect transistors

Table 3.1 Water contact angle, substrate temperature at which 2D-to-3D transition occurs and disperse and polar surface enthalpy extracted earlier by the method after Wu in paragraph 2.4.

Surface chemistry

Water

Tsub (K) at

contact angle Φ=0.25± ±0.03Å/

Disperse surface

Polar surface

enthalpy (mN/m)

enthalpy (mN/m)

(°)

s

(Wu)

(Wu)

OTS

109 ± 1

343 ± 3

22.57 ± 0.12

2.81 ± 0.02

BUTS

90 ± 3

343 ± 3

26.48 ± 0.04

4.29 ± 0.01

UETS

95 ± 2

340 ± 3

25.20 ± 0.12

3.50 ± 0.01

CUTS

80 ± 1

336 ± 3

24.08 ± 0.04

15.59 ± 0.01

FDTS

111 ± 1

320 ± 3

12.64 ± 0.07

3.49 ± 0.10

clean SiO2 (UV/O3)