Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Two-photon photoelectron emission studies of metal/conjugated polymer interfaces Sohn, Youngku 2004

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
ubc_2004-93179X.pdf [ 15.46MB ]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0061205.json
JSON-LD: 1.0061205+ld.json
RDF/XML (Pretty): 1.0061205.xml
RDF/JSON: 1.0061205+rdf.json
Turtle: 1.0061205+rdf-turtle.txt
N-Triples: 1.0061205+rdf-ntriples.txt
Original Record: 1.0061205 +original-record.json
Full Text
1.0061205.txt
Citation
1.0061205.ris

Full Text

Two-Photon Photoelectron Emission Studies of Metal/Conjugated Polymer Interfaces by Youngku Sohn B.Sc, Chungnam National University, Korea, 1991 M.Sc, Chungnam National University, Korea, 1997 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 2004 © Youngku Sohn, 2004 Abstract Two photon photoelectron spectroscopy (2PPE) has been used in the investigation of the electronic structure of interfaces between metals and conjugated polymers. This technique probes excited intermediate states in the polymers. The systems studied in this thesis are regioregular poly(3-hexlythiophene-2,5-diyl) (P3HT) and MEH-PPV polymers on a Au(lll) substrate. Electron kinetic energy distributions and quantum yields were recorded with varying polymer thicknesses to investigate the influence of the metal substrate. Spectra and yields were also measured as a function of various photon energies and incident laser power. The kinetic energy distribution increases in width by an amount equal to the increment in photon energy, characteristic of the 2PPE process with photoexcitation from a fixed intermediate state. Kinetic modeling of the emission yield as a function of incident laser power and photon energy suggests that the intermediate state accessed during the 20 ns laser pulse is the negatively charged polaron. The binding energy of the intermediate state decreases for films thinner than 4 nm. This behavior can be understood by polarization at the interface, or alternatively by charge transfer. Also, thickness has a very large effect on quantum yield. Intermediate state pumping efficiency is lowered more than -1000-fold for thin films compared to films more than 4 nm thick. Change in lifetime quenching with polymer film thickness maybe a consequence of diffusion and dipole coupling to the metal. ii Table of Contents Abstract ii Table of Contents iiList of Tables vi List of Figures viAcknowledgements xiv Chapter 1 Introduction 1 Chapter 2 Reviews 5 2.1 Review of Conjugated Polymers 5 2.1.1 Polymer Electronic Structure 6 2.1.2 Polymers in Devices 8 2.2 Photoelectron Spectroscopy 11 2.2.1 Theory 12.2.2 2PPE Studies on Metal Surfaces and Image States 14 2.2.3 2PPE Studies of Adsorbate induced Intermediate States 16 2.3 Reviews of Organic/Metal Interfaces 19 2.3.1 Energy Level Alignment at the Interface 19 2.3.2 Doping 24 2.3.3 Some Aspects of Photoexcitations 26 2.3.4 Doped Polymers 30 2.3.5 Some Aspects of P3HT 36 Chapter 3 Experimental 41 3.1 Ultrahigh Vacuum Systemiii 3.1.1 Vacuum Pumps 42 3.1.2 Sample Holder and Heating Dock 43 3.1.3 Electron Energy Analyzer 44 3.2 Sample Preparation 46 3.2.1 Cleaning of a Gold Substrate 43.2.2 Film Preparations 47 3.2.3 Metal Deposition on Polymer Film 43.3 Two-Photon Photoemission Spectroscopy 8 3.3.1 Major Components 49 3.3.2 Laser System 50 3.3.3 Data Acquisition 1 3.4 Auger and X-ray Photoelectron Spectroscopy 52 3.5 Scanning Probe Microscopy 53 3.6 Some Further CommentsChapter 4 Sample Characterization 55 4.1 Cleaning of Gold Surface 54.2 Polymer Surface Morphology 7 4.2.1 Spin-and Dip-Coated Films 54.2.2 Annealing Samples 60 4.3 Polymer Thickness 3 4.3.1 ATM Method 64.3.2 XPS Method and Inelastic Electron Mean Free Path 68 4.3.3 FT-IR Method 69 4.4 Calcium Coverage 70 Chapter 5 2PPE of Gold Substrate 75 5.1 Bare Gold 75.2 Quantum Yield 7 5.3 Surface State 80 IV 5.4 Gold Substrate Dipped in Chloroform 82 Chapter 6 2PPE of Semiconducting Polymer Films 84 6.1 Characterization of 2PPE of P3HT 86.2 2PPE Analysis of P3HT Polaronic Energy Levels 93 6.3 Interfacial Effect on P3HT Energy Levels 98 6.4 P3HT 2PPE Yield versus Thickness 103 6.5 2PPEofMEH-PPV/Au 107 6.6 Annealing Effect 112 6.7 2PPE Laser Intensity Dependence for Thick Films t 121 Chapter 7 Summary and Conclusion 124 Bibliography 127 Appendix One 141 Appendix Two 145 V List of Tables Table 2.1 Ionization potential values of P3HT in literature 37 Table 2.2 Hole mobility in various P3HT based devices 9 Table 2.3 Assignments of the photoinduced absorption bands in conjugated polymers 40 Table 3.1 Concentration and tuning range of laser dyes used 5Table 4.1 Auger Sensitivity factors at 4 KeV electron beam 72 Table 4.2 Calculated photoelectron mean free paths using equation 4.8, An and Bn values were taken from the reference 73 vi List of Figures Figure 1.1 Interfacial electronic structure in a polymeric LED . 2 Figure 2.1 Structures of some conjugated polymers 6 Figure 2.2 Delocalized n -electrons by overlap of />-orbitals, (a) hybrid orbitals, and (b) delocalized polyacetylene 7 Figure 2.3 Band gap and the effect of conformation of conjugated polymers .... 7 Figure 2.4 Bond alternations for (a) polythiophene (aromatic and quinoid structures), and (b) polyacetylene 8 Figure 2.5 Schematic of a simple light emitting diode 9 Figure 2.6 Charge transfer at the D/A interface in photovoltaic cell 10 Figure 2.7 Schematic of a polymeric field-effect transistor 1Figure 2.8 Energy levels in metal and semiconductor 1 Figure 2.9 Emission processes for (a) UPS, (b) IPES, and (c) 2PPE 12 Figure 2.10 Scenarios in 2PPE probing (a) occupied state, (b) intermediate state, and (c) final state 13 Figure 2.11 An electron bound by an induced polarization, and the potential energy diagram 5 Figure 2.12 Schematic of electron transfer between metal and adsorbates 16 Figure 2.13 Two image state peaks A and B with the benzene coverage on Cu(111) 17 Figure 2.14 2PPE for C6H5S/Cu(l 11) (<D=3.7 eV) with photon energy. The two peaks (A and B) are induced peaks 18 Figure 2.15 Principle of UPS for studying organic/metal interface. A: vacuum level shift, O: work function of a metal substrate, IP: ionization potential of the polymer, Ep: Fermi level, hv is a fixed photon energy, and AKEmaX: change in KEmax cutoff 19 Figure 2.16 Interfacial electronic structure diagram in 2PPE for bare gold and polymer/metal 21 Figure 2.17 The electronic structures at the semiconductor-metal interfaces before and after contact. The width of the space charge is represented as W .... 22 Vll Figure 2.18 Possible interactions forming an interfacial dipole layer 22 Figure 2.19 Structures of doped systems (a) negative polaron (b) two negative polarons (c) negative bipolaron 25 Figure 2.20 Schematic of the self-localized state energies 2Figure 2.21 Simple schematics of photoexcited state decay channels in conjugated polymer. S, T and P represent singlet, triplet and polaron states, respectively 26 Figure 2.22 Typical UPS spectra for a doped polymer with coverage. New gap-states are shown in the spectra 31 Figure 2.23 Three cases of calcium on polymer (a) ideal metal on polymer, (b) doped layer at the interface, and (c) calcium oxide formation 33 Figure 2.24 Interfacial energy diagrams for calcium on pentacence/Au 34 Figure 2.25 Energy diagram of Ca/PFO/ITO 36 Figure 2.26 Possible regioregularity in alkyl substituted polythiophenes chain. H and T mean head and tail, respectively. Head carbon is located between sulfur and carbon with alkyl group. Tail carbon is located between sulfur and carbon with no alkyl group 37 Figure 2.27 Schematics of the gap states and symmetry-allowed optical transitions for polarons and bipolarons 40 Figure 3.1 Schematic of the 2PPE system 1 Figure 3.2 (a) Sample heating dock, and (b) ESCA-style sample-mounting holder 43 Figure 3.3 Schematic of the VG100AX concentric hemispherical energy analyzer 45 Figure 3.4 Diagram of a calcium deposition system 48 Figure 3.5 Schematic of the custom magnetic shield: top and side view 50 Figure 3.6 Schematic for the 2PPE yield measurement 52 Figure 4.1 AFM image (1 /umxl fjm) of a flame-annealed gold surface 56 Figure 4.2 Height profiles of the AFM image shown in Figure 4.1 56 Figure 4.3 S2p XPS spectrum for the P3HT on gold 57 Figure 4.4 Two possible stacking structures of P3HT 8 Figure 4.5 AFM images (2 fjmx2 fan) for unannealed (a) dip, and (b) spin-coated P3HT on gold, FFT analysis and their height line profiles. The inset of Vlll the AFM images shows their Fourier transform power spectrum ... 59 Figure 4.6 APM images (2 jum x2 jum) for (a) 2.3 nm and (b) 7.7 nm thick P3HT films annealed at 100°C, and the height line profiles 60 Figure 4.7 Surface roughness measured in 2 /um x2 jum plotted as a function of film thickness (measured by XPS). The films were annealed at 100°C ... 61 Figure 4.8 a) AFM images of dip-coated P3HT after a 2PPE experiment and the height profile of a circular hole, (b) 3D-view of a crater, and (c) topview of the crater 62 Figure 4.9 Peeling of polymer film by AFM tip. The scan (from A to F) was continued until all of the polymer films was removed. The scan range is 5 pm x 5 pm. The scratch is ~4 nm deep 63 Figure 4.10 Force curves of (a) P3HT and (b) Au contact: (i) no interaction region, (ii) jump-in-contact, (iii) contact region, (iv) retraction of the sample, and (v) jump-off contact 64 Figure 4.11 Types of force-distance pull-off curves and their corresponding interactions 65 Figure 4.12 AFM image and its line profile for a dip-coated P3HT after 2PPE experiment 66 Figure 4.13 AFM image of the scratched area and the line profile for the same film shown in Figure 4.7 7 Figure 4.14 AFM image (20 um x 20 urn) showing a sharp edge and the line profile for a-35 rim thick P3HT film 67 Figure 4.15 FT-IR intensities versus polymer thickness. Inset is the C-H stretching IR band with a linear background the fitting lines 69 Figure 4.16 Equilibrium contact angle (0) of overlayer on substrate 70 Figure 4.17 Three important thin film growth modes at surfaces for different coverages. (a) Layer-by-Layer Growth (Frank-Van der Merwe, FM). (b) Layer Plus Island Growth (Stranski-Krastinov,SK). (c) Island Growth (Vollmer-Weber,VW) 71 Figure 4.18 Typical Auger spectra for bare gold and calcium (1.4 ML) on gold in the range of 50 ~ 550 eV. Auger spectrum for a carbon contaminated gold is IX also shown 73 Figure 4.19 Experimental peak-to-peak intensity ratios (ICa(294) 1 1Aum) with Ca deposition time at various oven temperatures including the best fitting curves using layer-by-layer growth mode 74 Figure 5.1 2PPE spectrum for Au(l 11) taken with a photon energy of 3.65 eV . Inset is an expanded figure for the high energy cutoff 75 Figure 5.2 High energy cutoff as a function of incident photon energy 77 Figure 5.3 Photoemitted electron flux as a function of laser intensity for two different bare gold substrates, with an incident photon energy of 3.65 eV 78 Figure 5.4 Calculated absorbance of gold for p- and s-polarized 3.65 eV photons with respect to incidence angle from the surface normal 80 Figure 5.5 (a) Surface states for Au(l 11) taken at high (-24.0 V) and low (-10.0 V) bias, (b) IPS spectra for Au(l 11) at a photon energy of 11.0 eV as a function of incidence angle, SS: surface state, IS: n=l image state 81 Figure 5.6 2PPE spectra for a bare clean gold and a gold substrate dipped in blank chloroform solvent, with a photon energy of 3.65 eV . . 82 Figure 6.1 The 2PPE spectra for a bare gold and a 2 nm thick P3HT on gold, with a photon energy of 3.65 eV 85 Figure 6.2 Power law plots with photon energy for a 30 nm thick P3HT film on gold. The solid lines are fit curves discussed in text 87 Figure 6.3 The 2PPE spectra for a 2 nm thick P3HT on gold with three different photon energies. The spectra are normalized to the same peak height. The arrow indicates the low energy cutoff independent of the photon energy ... 88 Figure 6.4 The population density of polaron and singlet excitons as a function of Pump fluence for PIFTO and PIFTEH polymers 88 Figure 6.5 The 2PPE quantum yield (Q/I2) obtained at an emission level of -3000 electrons/pulse, and Ba obtained from the fits for the 30 nm thick P3HT film. All values are normalized to the values at 3.0 eV. Inset is an absorption profile for P3HT film 92 Figure 6.6 Kinetic energy distributions for a 2 nm thick P3HT on gold with three different incident probe energies. The arrows indicate the shift in position of the high energy emission edge with different laser wavelengths .... 94 Figure 6.7 Electronic energy levels for a 2 nm thick P3HT film and possible photoemission processes with a photon energy of 3.65 eV 96 Figure 6.8 The 2PPE spectra for a 2 nm thick P3HT film with ~1 A overlayer of deposited Ca, with a photon energy of 3.65 eV 97 Figure 6.9 Photoemission spectra of calcium (1.1 A) on P3HT (2 nm) with laser power, with a probe photon energy of 3.65 eV, Inset is a plot of two peak heights with laser intensity 98 Figure 6.10 2PPE spectra for P3HT film with varying thickness, with a photon energy of 3.65 eV 9 Figure 6.11 The energy diagram for spin-coated P3HT/Au annealed at 100°C, with respect to thickness. All spectra were taken with s-polarized light (ha> = 3.65eV). Values from the top to bottom are vacuum level shift, work function, intermediate state edge, and thickness of the films 101 Figure 6.12 Changes in HOMO and LUMO levels depending on polymer thickness by Hill et al. (1) molecule in gas phase, (2) isolated molecule on a metal surface, (3) one complete molecular layer on a metal surface, (4) first molecular layer on metal" surface, beneath a molecular film, (5) bulk molecular film, away from both the metal interface and the film surface, and (6) surface of a molecular film 102 Figure 6.13 Log-log plots of photoemitted electron flux (electrons/cm2 sec) vs. incident laser flux (photons/cm sec), with 3.65 eV photons and different thick P3HT films. Lines are fit as described in text 103 Figure 6.14 Relative pumping efficiency per absorbed photon, Q/ {[I(l-R)]2 (l-e"dA)-X,} vs. the average film thickness with A = 0.01 nm (A), X = 1.5 nm (•), and X= oo (o), with a photon of 3.65 eV 105 Figure 6.15 Best-fit ki and no parameters for the 2PPE data with polymer thickness 107 Figure 6.16 Plots of photoemitted electron flux (electrons/cm2 sec) vs. incident laser flux (photons/cm2 sec) for 2 and 20 nm thick MEH-PPV films, with 3.65 eV photons. The solid lines are fit curves discussed in text. Inset is a power law plot for 100 nm thick MEH-PPV film by Hale et al 108 Figure 6.17 2PPE spectra for 2 and 20 nm thick MEH-PPV films with a photon energy of 3.65 eV. Inset shows the 2PPE (hco =3.0 eV) for a 100 nm thick MEH-PPV 110 Figure 6.18 The 2PPE spectra for a few rim thick MEH-PPV on gold with three different photon energies. The spectra are normalized to the same peak height Ill Figure 6.19 Energy diagram for MEH-PPV/Au, with respect to thickness. Vacuum level is low E cutoff relative to EF. Intermediate state edge is high E cutoff relative to EF minus the photon energyhco = 3.65eV Ill Figure 6.20 2PPE spectra for a 30 nm thick P3HT film with respect to annealing temperature, with a photon energy of 3.65 eV 112 Figure 6.21 2PPE spectra for a 20 nm thick MEH-PPV film with respect to annealing temperature, with a photon energy of 3.65 eV 112 Figure 6.22 Energy diagram for P3HT/Au with respect to annealing temperature. Vacuum level is low E cutoff relative to EF. Intermediate state edge is high E cutoff relative to Ep minus the photon energy hco - 3.65eV. The HOMO edge is derived assuming that the IP of P3HT is 4.7 eV 113 Figure 6.23 Energy diagram for MEH-PPV/Au with respect to annealing temperature. Vacuum level is low E cutoff relative to EF- Intermediate state edge is high E cutoff relative to Ep minus the photon energy hco = 3.65eV. The HOMO edge is derived assuming that the IP of MEH-PPV is 5.3 eV 113 Figure 6.24 Work function with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV 114 Figure 6.25 XPS survey spectrum for an unannealed P3HT film 115 Figure 6.26 Photoemission yield (Q/I2) with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV ... . 116 Figure 6.27 Power-law slopes with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV 117 Figure 6.28 Power law plots with annealing temperature for a 30 nm thick P3HT film on gold, with a photon energy of 3.65 eV. The solid lines are fit curves xii discussed in text, which can be derived changing either kT or Ba ... 118 Figure 6.29 2PPE spectra for unarmealed and 100 °C annealed P3HT and MEH-PPV films, with a photon energy of 3.65 eV 119 Figure 6.30 Crystallinity and PL intensity with annealing temperature for P3DT measured by Yoshino et al 120 Figure 6.31 Relative 2PPE yield with respect to the thinnest film for annealed (100 °C) and unannealed P3HT films, with a photon energy of 3.65 eV .... 121 Figure 6.32 2PPE spectra for a 30 nm thick P3HT film measured at different laser powers, with a photon energy of 3.65 eV. Photon flux is 1020 ~ 1021 ph/ cm2 s. Electron flux is 1012~1013 el/ cm2 s 122 Figure 6.33 2PPE spectra measured at different laser powers for a 3 nm thick P3HT film, with a photon energy of 3.1 eV. Photon flux is 1021 ~ 1022 ph/ cm2 s. Electron flux is 1012 ~ 10° el/ cm2 s 123 Figure 6.34 2PPE spectra measured at different laser powers for a 20 nm thick MEH-PPV film, with a photon energy of 3.1 eV. Photon flux is 1021 ~ 1022 ph/ cm2 s. Electron flux is 1012 ~ 1013 el/ cm2 s 123 xiii Acknowledgements Everything goes to awesome God. I would like to thank my supervisor Dr. J. Todd Stuckless for his guidance, support and brilliant suggestions for my thesis project. Also many thanks to: Dr. D. Bizzotto for a review of my thesis. UBC Chemistry people: Technicians in mechanical and electronic service shops; Dr. D. Bizzotto and Mr. R. Stoodley (providing the use of FT-IR), Drs K. Mitchell and K. Wong (XPS data collection), Drs. M. Wolf and A. G. Pattantyus-Abraham (providing the use of AFM and MEH-PPV polymer), Mr. John Kim (for reading some parts of my thesis), and Elmer A. Ogryzlo (my memory back to 1996 -1998). My lab people: Dr. Richard Murdey (for discussion of instrumental parts), Ms. Jessica Anient (for setting up the 2PPE instrument), Dr. Jan Richter (for the discussion of kinetic modeling of the 2PPE data), Ms. Sabine Hirth (for helping to set up the metal deposition system), Mr. Sherman Hon (for the discussion of my experiments), Mr. Sam Liang (for his time in my lab). Dr. Jun-Gill Kang and professors at CNU chemistry, and researchers at Nuclear Chemistry Research Division, KAERI. And thanks to my family (including grand mom and dad in heaven), my relatives, my friends, and Elder B. J. Kim. Financial support was provided by. the. National Science and Engineering Research Council (NSERC), Canada Foundation for Innovation (CFI) and UBC. xiv Chapter 1 Introduction The discoveries of metallic behavior in a doped polymer in 1977 [1] and electroluminescence in a conjugated polymer by Burroughes et al. in 1990 [2] has opened a new era in conjugated polymer research and applications. The developments of electrically conductive polymers by Heeger, MacDiarmid, and Shirakawa led to the Nobel Prize in 2000. Conjugated polymers consist of alternating single and double bonds resulting in extensive derealization of jc-electrons along the polymer chain. The delocalized 7i-electron HOMO and LUMO states are responsible for the unique properties of the polymers [3]. The electronic structure consists of an occupied conduction band and unoccupied valence band separated by an energy gap, typically in the range of 1.5 ~ 3 eV. A wide range of band gaps can be tailored by changing the polymer backbones and side groups. Due to their unique optoelectronic properties, the potential applications of conjugated polymers are very broad including light-emitting diodes (LEDs), field-effect transistors (FETs), solar cells and nonlinear optical devices. Organic polymers have many advantages over inorganic materials: they are lightweight, flexible, cheap, and easy to process. High performance as well as a wide range of colors is possible. The major disadvantage is short operating lifetime due to ohmic heating [3]. Together with industrial applications, the fundamental understanding of physics of a system is a key element in efforts to improve the performance of electronic devices. As an example, Figure 1.1 shows a schematic of the energy levels in a polymeric LED where a conjugated polymer is sandwiched between two electrodes. In this device, an electron and a hole from two different electrodes are injected through the barriers and attracted together to form an exciton. The charge injection barriers are determined by the energy level alignment at the interface. The question is what the electron binding energies are, and whether the vacuum level is determined by interface dipole or band bending effects. Better detailed understanding of the basic operating mechanism is crucial to making better devices. 1 Metal cathodes are intrinsic to device construction. Conjugated polymers are commonly coated on high work function transparent conducting substrate, and the low work function metal is deposited by thermal evaporation. We should also understand the formation of new sub gap states (e.g. polaron and bipolaron) upon metal deposition. These states are important for charge transfer processes, and also can act as exciton dissociation sites [3]. space charge region High work function^ Conduction band metal © |Low work function metal Jectron-hole recombination) o • h+ Valence band Figure 1.1: Interfacial electronic structure in a polymeric LED. The electronic structures of organic/metal interfaces have been extensively studied [3] using ultraviolet photoelectron spectroscopy (UPS), x-ray photoelectron spectroscopy (XPS) or scanning tunneling microscopy (STM). UPS has been used to directly measure ionization potentials, work functions and HOMO levels. The LUMO level is derived using the optical band gap of the bulk polymer. However, the band gap may vary during interface formation. The electron affinities (EA) and LUMO levels can be determined by inverse photoelectron spectroscopy (JPES). Near-edge x-ray absorption fine structure (NEXAFS) has also been used to directly probe the unoccupied sates. Two-photon photoelectron spectroscopy is a second-order process [4]. The first photon excites an electron from an initial state below the Fermi level to an unoccupied 2 intermediate state. The second photon ejects the electron from this state to the final state above the vacuum level. The 2PPE spectrum provides information on both the occupied initial and unoccupied intermediate states with good spectral and time resolutions. Recent developments of femtosecond 2PPE techniques have opened a new stage of studying excited state dynamics at the interfaces. The early studies by Steinmann [4] using 2PPE focused on bare metals and their image states. Recently, molecules adsorbed on a metal have been extensively investigated. Zhu (U of Minnesota) and Wolf (Free University Berlin) have studied benzene, naphthalene and self-assembled monolayers adsorbed on a metal [5]. Harris (UC Berkeley), Fauster (University Erlangen-Niirnberg) and some other groups have studied dynamics at the surfaces and interfaces [6, 7, 8]. Halas (Rice U) have used 2PPE to study triplet exciton dynamics in MEH-PPV and the photo-oxidized polymer [9]. White (U of Texas) have studied tris(8-hydroxy-quinlonine) aluminum quinolate (Alq3) on a metal using 2PPE [10]. Haight (IBM) also have studied Alq3 using a two-photon technique where the probe beam photon energy i s much higher than that of the pump beam, to photoemit electrons simultaneously from the HOMO and photoexcited intermediate states [11]. This work has been motivated by a desire to understand the nature of photoexcitations and the energy of excited state charge carriers in conjugated polymers using 2PPE. The occupied HOMO levels have been fairly well studied and well defined. However, there has been lack of understanding of the unoccupied states. Understanding the unoccupied states which the electrons from the metal electrode inject into is as important as understanding of the occupied state in organic materials. The primary goal of this thesis is to introduce two-photon photoemission spectroscopy (2PPE) to studies of interfacial electronic structures and photoexcitation dynamics at metal/conjugated polymer interfaces. This work has a significant meaning; we are the first to comprehensively study metal/conjugated polymer interfaces using 2PPE. We demonstrate that 2PPE using nanosecond scale laser pulses is especially successful because the lifetime of the intermediate state in conjugated polymers is fairly long compared to metals [12], so that the electron emission from the metal substrate is negligible compared to that of the conjugated polymer. In UPS, the emission signal from 3 polymers is obscured by the strong background signal from the metal substrate. Moreover, because the photoemitted electron kinetic energy distribution has information on intermediate states, we then estimate the energy of excited charge carriers (polarons), analogous to the conduction band edge of conventional semiconductors. 2PPE provides "absolute energies" i.e. referenced to Evac and Ep, whereas optical absorption gives transition energies, referenced to band edges. In conjugated polymers, unlike inorganic semiconductors, when electron and hole are created by photoabsorption or injected from two different electrodes the relaxation and polarization properties change the band gap significantly [13]. In this case, the real carrier transport gap is different from that estimated from the absorption optical gap and HOMO level. Whether the photoexcitations have polaronic or excitonic properties are studied by changing photon flux and energies. 2PPE can also reveal band bending at the interface and the effect of the metal on the lifetime of excited states. Gold and calcium were selected as a high and low work function metals, respectively. The most well studied polymers; P3HT and MEH-PPV were chosen. It is known that regioregular P3HT has high carrier mobility and MEH-PPV has high luminescence yield. Chapter 2 is a review of conjugated polymers and their applications, and of the 2PPE technique. Studies of interfacial electronic structures of organics adsorbed on metals and of metal-doped organics are also reviewed. Chapter 3 provides detail of our experimental procedures. This includes Auger spectroscopy, scanning probe microscopy (SPM) and sample preparation methods. The metal deposition method is also explained. Sample characterization is described in Chapter 4. The techniques include Auger, XPS, AFM, and FT-IR methods. Chapter 5 is a discussion of our 2PPE results for bare gold films. Chapter 6 discusses the origin of photoemitted electrons in the 2PPE of P3HT/Au. 2PPE yields and power law plots with varying polymer thickness and photon energy are analyzed to explain photoexcitation dynamics. We suggest a simple kinetic model to explain the data in this chapter. Annealing effect on the polymer is also disscussed, including the surface morphology. New polaronic sub-gap states induced by calcium depostion on P3HT are also disscused. Some preliminary 2PPE results of MEH-PPV films are included for comparison purposes. Chapter 7 is the summary and conclusion of the thesis. The final remarks suggest a guide for future research. 4 Chapter 2 Reviews This chapter deals with reviews of conjugated polymers and their applications, and the theory and application of two photon photoemission spectroscopy. Electronic structures at metal/polymer interfaces are also reviewed. 2.1 Review of Conjugated Polymers Until the discovery of a metallic polymer in 1977 [1], polymers were generally considered only as insulators. Burroughes et al. demonstrated the first electroluminescent device based on poly(p-phenylene vinylene), conjugated polymer in 1990 [2]. This has rekindled research on conjugated polymers, which in recent years have been extensively developed for many technological applications due to their unique electronic and optical properties. The polymers have some advantages over inorganic materials [14,15,16]. They are lightweight, flexible, durable, cheap, and easy to process. You can make a thin non-planar lightweight device. You can solution-cast, spin-coat, print, or paint them over a large area. A wide range of colors and high performances can be achieved by tailoring polymer structures. There have been many efforts to synthesize ideal conjugated polymers, which require long polymer lifetime, good injection barriers and high quantum efficiency. The stacking properties and longer conjugation length of polymers are also indispensable for improving carrier mobility. Potential applications of conjugated polymers are very diverse. They include LEDs [17], laser gain media [18], photodiodes, photodetector, solar cells [19], biosensors [20], field-effect transistors (FET) [21], actuators [22], all-polymer optoelectronic devices [23], all-polymer integrated electronic circuits [24], waveguides [25], and electromagnetic radiation shielding [26]. In recent years, some electronic devices based on organic materials have been released in market by Eastman Kodak (digital camera display), RiTdisplay (cellular phones) and Philips (electric razor) [27]. In the near future, many other companies (including Cambridge Display Technology, IBM, Universal Display Corp., Toshiba, and 5 Samsung SDI) will likely introduce a lot more organic devices, for instance flexible large displays. 2.1.1 Polymer Electronic Structure Poly(p-phenylene vinylene) (PPV) and polythiophene and their derivatives have been widely studied due to their high luminescence efficiency and carrier mobility, respectively. The solubility in common organic solvents can be increased by grafting long alkyl side chains, which is of great advantage for fabricating thin films. In PPV, alkoxy substitutions such as MEH-PPV of Figure 2.1 are generally tailored to increase solubility, and to modulate the electronic and optical properties of the polymers. polyacetylene polyparaphenylene poly(phenylene vinylene),PPV polypyrrole polyfuran polythiophene L O Jn ^0 poly(3-hexlythiophene-2,5-diyl) poly[2-methoxy,5-(2'-ethyl-hexyloxy)-p-phenylene vinylene] P3HT MEH-PPV Figure 2.1: Structures of some conjugated polymers. A conjugated polymer backbone consists of alternating single and double bonds between carbon atoms, which have sp2 hybridization, and these bonds have resonance type delocalized structures along the polymer backbone. The delocalized rc-electrons act 6 as charge carriers along the chain resulting in semi-conducting properties of the polymer. Polyacetylene shown in Figure 2.2 (b) is one of the simplest conjugated polymers and consists of only single and double bonds between carbon atoms. (a) (b) Figure 2.2: Delocalized ^-electrons by overlap of p-orbitals, (a) hybrid orbitals, and (b) delocalized polyacetylene. The bonding and antibonding orbitals correspond to a valence band (HOMO) and conduction band (LUMO). Charge carriers can be introduced, i.e. by chemical or electrochemical doping, into the electronic structure to increase conductivity [3]. The polymer chains generally only weakly interact with each other by van der Waals or electrostatic dipole interactions. Conductivity typically decreases with increasing side group size in the polymer backbone. This can be explained as shown in Figure 2.3 [3] by a decrease in conjugation length accompanied by an increase in the HOMO-LUMO gap, caused by twisting of main chain. a Eg=band gap mm WgM Ordered Disordered Long conjugation length Short conjugation length High electrical conductivity Low electrical conductivity Figure 2.3: Band gap and the effect of conformation of conjugated polymers. 7 Conjugated polymers can be divided into two groups: degenerate and nondegenerate polymers [28]. Most conjugated polymers including P3HT and MEH-PPV are nondegenerate polymers. The degenerate polymers such as trans-polyacetylene have two identical ground state minima where the alternating single and double bonds are reversed. In polythiophene, the bond alternation, shown in Figure 2.4 (a) exhibit two different structures: aromatic and quinoid, resulting in two different ground states. aromatic structure (a) quinoid structure (b) Figure 2.4: Bond alternations for (a) polythiophene (aromatic and quinoid strucutures), and (b) polyacetylene 2.1.2 Polymers in Device i) Organic Light Emitting Diodes (LEDs) A simple LED shown in Figure 2.5 consists of a conjugated polymer sandwiched between anode and cathode materials with different work functions. The cathode is an electron injection electrode with a low work function. The anode is an electron injection electrode with a high work function. The transparent indium-tin oxide (ITO) has commonly been used as an anode so that light can pass through. Balanced electron-hole injection with good recombination of charge carriers is crucial to the design of efficient polymer light-emitting diodes. Single-layer devices tend to have a low efficiency because electron injection efficiency is lower than the hole injection efficiency; consequently e lectron-hole recombination occurs near the cathode surface, and the injection current is dominated by holes. The performance of an EL device is mainly limited by the electron injection efficiency [29]. The low work function 8 metal strongly determines the lifetime of a device. The device with a low barrier height requires a low driving voltage, which minimize stress and temperature rise during operation. Two layers, one a good electron transporter and the other a good hole transporter can be used to improve the EL efficiency [30]. In a quantum mechanical point of view, it has been believed that an EL quantum efficiency of 25 % is the limit, where only one singlet exciton forms for every three triplet excitons [31]. However, an EL efficiency of up to 63 % has been observed [32]. Electron Injection Electrode (Ca, Al; low work function metal) Light Emitting Layer R A (conducting polymer) + + + mm glass Hole Injection Electrode (ITO) Light(radiative singlet excitons) Figure 2.5: Schematic of a simple light emitting diode ii) Photovoltaic devices Photovoltaic devices such as solar cells are typically composed of silicon p-n junctions [33] similar to photo diodes but with light energy converted into electric energy. A solar cell with an efficiency of 10% can generate 100 W/m2. Photocells are made with hole and electron acceptor organic heterojunctions to improve efficiency. Hole acceptors include MEH-PPV and P3HT polymers [34]. Electron acceptors include C6o and CN-PPV. Photo generated excitons diffuse to and separate efficiently at the interface between the two different molecular semiconductors, then electrons and holes in the two materials transport to collection electrodes. In a heteroj unction photovoltaic cell using MEH-PPV/C60 shown in Figure 2.6 [19,35], the efficiency is more than two orders of magnitude higher than that of a single MEH-PPV layer device. The carrier collection efficiency (r)c) and energy conversion 9 efficiency (r)e) of the cell are 29 % and 2.9 %, respectively. In a laminated MEH-CN-PPV/POPT blend device, a power conversion efficiency of 1.9 % and external quantum efficiency of 29 % were achieved [36]. Donor Acceptor exciton >-# v i I MEH-PPV Ceo 1 1 ITO Ca or Al Figure 2.6: Charge transfer at the D/A interface in photovoltaic cell. iii) Field Effect Transistors (FETs) Field-effect transistors (FETs) based on a conjugated polymer, as shown in Figure 2.7, were demonstrated in 1980s [37,38]. The charge carrier mobility in FETs have improved since then from values of ~10"5 cm -V" -s* to approach that of amorphous silicon which 2 11 is 0.1 ~ 1.0 cm -V" -s" . The electron and hole carrier mobilities for a Si single crystal are 1500 and 400 cm^V'-s"1, respectively [37]. }-VG Figure 2.7: Schematic of a polymeric field-effect transistor (FET). The performance of an organic FET is determined primarily by the charge carrier 10 mobility and injection efficiency at the source and drain interfaces. Recently, Sirringhaus et al. [21] demonstrated a solution processed FET based on regioregular P3HT and MEH-PPV, with a mobility of 0.05 ~ 0.1 cm^V'V1 and an ON-OFF ratio of >106. The mobility strongly depends on the regioregularity and the 71-71 interchain stacking orientation of the P3HT polymer [21], being two orders of magnitude higher with the interchain stacking vector parallel to the substrate than with it perpendicular to the substrate. Minority carrier effects can be neglected for a unipolar FET. Two recent articles [37,38] review organic field-effect transistors in detail. 2.2 Photoelectron Spectroscopy Photoelectron spectroscopy is a widely used surface sensitive technique due to a short electron escape depth. This section focuses on two-photon photoemission (2PPE) spectroscopy in comparison with one photon ultraviolet photoemission spectroscopy (UPS). 2.2.1 Theory IP t t EA Conduction band LUMO HOMO Valence band Metal Semiconductor Figure 2.8: Energy levels in metal and semiconductor. Figure 2.8 shows the energy level of a metal and a semiconductor, respectively. The vacuum level (Evac) is the energy of an electron at rest outside the material far enough away to be in a field free space. The electron affinity (EA) of a semiconductor is the 11 energy needed to remove an electron from the bottom of the conduction band (LUMO) to the vacuum level. The work function (O) is the energy needed to remove an electron from the Fermi level to the vacuum level. The ionization potential is the energy required to remove an electron from the top of the valence band (HOMO) to the vacuum level. The band gap (Eg) is the energy difference between the HOMO and LUMO edges. Following common usage, the thermodynamic chemical potential of the electrons is labeled the Fermi level. For a metal at 0 K this is equivalent to the energy of the highest occupied state. More generally, for a system of delocalized orbital states obeying Fermi statistics it corresponds to the energy at which the Fermi-Dirac distribution function /(e,) = 1 )/tr +1) is equal to 50 %. In UPS, shown in Figure 2.9(a) [4], the incident photon energy is high enough to directly remove electrons from the valence levels. The electron binding energy is the difference between the energy of the incident light and the kinetic energy of the emitted electron in the vacuum. The photoemitted kinetic energy distribution gives information on the occupied states. efinal n EF > e- Einitial Ekin <- Ekii Efinal > e-tifinal H nv (a) (b) (c) Figure 2.9: Emission processes for (a) UPS, (b) IPES, and (c) 2PPE [4,39,40]. In the case of metal, where there are occupied sates at the Fermi level, the work function can be estimated using the incident photon energy and the maximum kinetic energy £™ • <$> = hv-E™x 2.1 where the kinetic energy is measured with the electron at the vacuum potential. 12 The initial binding energy of electrons with respect to the Fermi level can be expressed as: EF-Einilia,=hv-0-Ekin 2.2 This technique is a powerful tool for directly measuring the ionization potential, the work function and the HOMO level. However, it only probes the occupied states. The unoccupied states can be studied by IPES, shown in Figure 2.9(b) [4]. The incident low energy electrons on a metal decay into an unoccupied state with accompanying photon emission. The emitted photon energy gives information on the unoccupied state. One of the drawbacks in IPES is damaging of the sample induced by high electron fluxes. The energy resolution is also limited to about 0.5eV, caused by the energy distribution of the incident electron beam. When the incident photon energy is less than the work function of a material but greater than one half of the work function, electrons can still be emitted by a two-photon process [4]. In 2PPE, the first photon excites an electron from an occupied state below the Fermi level into an unoccupied intermediate state and a second photon induces the electron emission from that state to a final state above the vacuum level, hence the electron energy distribution has the information of the unoccupied electronic state as well as the occupied state. hv hv ''initial AEki„ = 2-AM; hv' hv' hv hv AEkin = l-Mv AE„„=0-Ahv hv' hv' kin hv hv ''final hv'. hv' (a) (b) (c) Figure 2.10: Scenarios in 2PPE probing (a) occupied state, (b) intermediate state, and (c) final state. 13 By changing the wavelength of the probe photon, additional information can be obtained. Figure 2.10 shows three scenarios of photoelectron emission mechanism depending on the probed states. If the photoelectrons are emitted from a given occupied states by direct two photon process, the photoemitted electron kinetic energy with incident photon energy varies as AEkin = 2-h(v'-v) If the variation of the kinetic energy is AEkin =l-/z(v'-v), the electrons are emitted from a fixed intermediate state populated by different initial states. In cases where the emission is determined by a state above the vacuum level, E&„ is independent of the incident photon energy. In conjugated polymers, long-lived intermediate states might also be populated by relaxation of the initial photoexcitation [12] as shown in figure 2.10(b). In this case, Efmai ~ ^' na> is n°t necessarily equal to Einilial. 2.2.2 2PPE Studies on Metal Surfaces and Image States Early studies using 2PPE focused on the relaxation dynamics of photoexcited charge carriers and image states [8]. Intermediate dynamics can be studied using time-resolved two-photon photoemission (TR-2PPE) spectroscopy with two different laser pulses and variable time delays [41]. The photoelectron spectrum is recorded as a function of time delay between the pump and the probe pulses. An electron near a metallic surface can be trapped by a positive induced polarization charge, called an image charge, inside the metal. The attractive force is expressed asF(z) = -e1 I (2z)2, where an electron is at a distance of z. A bound energy state due to this long-range columbic interactions is located below the vacuum level. The quantized states form a hydrogenic Rydberg series with binding energies EB («) written in equation 2.3 [7]. The states are localized in the direction normal to the surface. EB(n)*Emc-^^,n=\,2,3--- 2.3 n Figure 2.11 shows the induced polarization and potential energy diagram with n=l and n=2 image states [7]. The transitions are excited by optical electric fields 14 perpendicular to the surface. Because electron-electron and electron-phonon interactions in the image states are weak compared to states of the metal lattice, the electrons have a relatively longer lifetime [39]. Image state lifetimes are limited by coupling to electronic states of the metal, and are generally only observable when located in a partial band gap of the metal. The line width and binding energy of the states have information on the surface. Image states in metals have also been studied by the IPES [42]. Image states are strongly perturbed by adsorbed layers on a metal surface. The lifetime of an image potential state generally decreases due to strongly bound adsorbates. On the other hand, weakly bound adsorbates are known to reduce the image-state binding energy but increase its lifetime with increasing number of adsorbate layers. This implies that the electron transfer is efficient in strongly adsorbed systems. The binding energy is dependent on the adsorbates and the thickness [6,43]. For weakly adsorbed systems for instance Xe/Ag(lll) [44], Xe/Cu(lll) [45] and N2/Cu(lll) [46], the binding energy of the first image state is reduced. The dielectric continuum model (DCM) [6,47] has explained the image state binding energy and lifetime for physisorbed adsorbates on metal. Harris and coworkers have used the model to explain various adsorbate/Ag(lll) systems [6]. 15 2.2.3 2PPE Studies of Adsorbate Induced Intermediate States In recent years, 2PPE has been used to study adsorbate/metal systems in a time-resolved manner [4,48]. The adsorbates range from atoms to large molecules. They are strongly (chemisorption) or weakly (physisorption) interacted with the substrate surface. The 2PPE studies on adsorbate/metal interfaces better enable us to understand electron transfer (ET) from a metal to adsorbates, photochemistry at metal surfaces and charge injection in semiconductor devices. If excited electrons in the substrate, which quickly scatter and lose energy, are in resonance with unoccupied states of the adsorbates, the electrons can indirectly transfer into this state. Figure 2.12 shows both direct and indirect electron transfer at an adsorbate/metal interface [49]. The electrons are then probed by a second photon using different photon energies and different delay times. Figure 2.12: Schematic of electron transfer between metal and adsorbates. Steinmann et al. first introduced 2PPE to study adsorbate/metal systems [48]. They observed occupied and unoccupied states induced by atomic oxygen on Cu(lll). Wolf and coworkers examined CO/Cu(lll), and found that interactions between CO In* orbitals and the occupied states of the metal are predominant in CO-Cu bonding [50]. Petek et al. investigated an antibonding state in Cs/Cu(l 11) and Cs desorption dynamics using a time-resolved 2PPE [51]. excited electron distribution intermediate state 16 Recently, Zhu [52,53,54] and Wolf [43,54,55] have studied the electronic structures of large molecules such as benzene, naphthalene, and Ceo [53] at a metal surface. Figure 2.13 shows two different image peaks with benzene coverage on Cu (111) [43]. They argue that the first layer is parallel to the surface and the second layer is perpendicular to the first 1 ayer. T he nelu LUMO for b ilayer b enzene o n C u(l 11) i s 1 ocated at 0.6 e V above the vacuum level. For hexafluorobenzene (CeF6) on Cu (111) [53,55], the orientation of molecule with coverage is similar to benzene. The smaller change of the work function compared to benzene at the same coverage is due to lower electron-donating character and 7t-electron polarizability of Q^^. The image state lifetime decreases, but the molecular resonance lifetime increases with coverage. This implies an efficient decay from the image state into the CT* LUMO adsorbate level. The decrease of lifetime is not expected for weakly bound adsorbates. t I a i—< a o 'SS I Jm Pt, e I I Image States A and B in Benzene Bilayer Regime on Cn(lll) 2.0 ML /A \ / B\ 15MLj |AI Patch B 1.2MlJ B\ LOME/ A\ Patck A -15 -1.0 -0.5 0.0 0.5 Binding Energy below Evacuum(eV) Figure 2.13: Two image state peaks A and B with the benzene coverage on Cu(111) [43]. 17 Zhu have also studied self-assembled systems [54,56,57]. For a self-assembled thiophenolate (C6H5S-) on Cu(lll), two induced interfacial states were observed. In Figure 2.14, the A peak stays the same position with photon energy while the B peak varies as AZsfa.„ = 1 • Ah v. They are located -0.4 eV below and -2.7 eV above the vacuum level. The states are localized in C-S-Cu and are independent of the hydrocarbon groups bonded to sulfur. They suggest that the thiolate is insulating for electron transport, but can be conducting for hole transport. For a self-assembled C6F5S- on Cu(lll), the HOMO and LUMO are located at -6.4 eV and -3.2 eV, respectively. The HOMO-LUMO gap is reduced from 8.9 eV (C6F5SH) to 3.2 eV due to strong electronic coupling between the metal and absorbate orbitals [57]. 5.5 6.0 6.5 7.0 7.5 8.0 8.5 Final state energy above EF (eV) Figure 2.14: 2PPE for C6H5S/Cu(111) (0=3.7 eV) with photon energy. The two peaks (A and B) are induced peaks [54]. Naaman et al. [58] used wavelength- and time-dependent 2PPE with ns laser pulses for rhodamine B (RB) and sulforhodamine (SR) adsorbed on a doped silicon. The two molecules have a similar structure and electronic properties. The work function for the strongly bound SR decreases more than for the weakly bound RB. The lifetime of photoexcitiations in the SR decreases, but for the RB the lifetime increases. 18 Ultra-fast time resolution 2PPE has also been employed to study electron wave packets created by coherent electron excitation [7], optical dephasing and phase control of electrons in metals [7], real-time motion of adsorbate atom [59], and spin-resolved electron relaxation in ferromagnetic systems [60]. 2.3 Review of Organic/ Metal Interfaces Metal/semiconductor interfaces are of great importance for determining the performance of polymer based electronic devices. The geometry and electronic structure of a conjugated polymer are commonly different at the interface than in the bulk. 2.3.1 Energy Level Alignment at the Interface Figure 2.15 shows principles of one photon photoemission to study energy level alignment at organic/metal interfaces [61]. KEn™ cut-off Organic/Metal Metal > Ek Metal Metal Organic Figure 2.15: Principle of UPS for studying organic/metal interface. A: vacuum level shift, O: work function of a metal substrate, IP: ionization potential of the polymer, EF: Fermi level, hv is a fixed photon energy, and AKEmax: change in KEmax [61]. 19 In the UPS spectra from metal, the high energy onset corresponds to the emission from the Fermi level, while the low energy cutoff corresponds to the electrons which have a kinetic energy of zero. The kinetic energy measured at the spectrometer depends on the contact potential difference (CPD) between the spectrometer and the sample, which is the sum of the applied bias and the difference in work function of the sample and spectrometer. In Figure 2.15, A is the difference in contact potential corresponding to the change in work function of the sample. For the adsorbate/metal interface, the high energy onset corresponds to the emission from the HOMO. The shift of the low-energy cutoff is a vacuum level shift (A), and the change in KEmax is the hole barrier height. For a Schottky-Mott type interface, the vacuum levels of the metal and semiconductor before contact are assumed to be the same, then the hole and electron barriers are simply given by: A,B=IP-®M 2.4 t,B=^M-EA = Eg-<f>hB 2.5 If a vacuum level shift (A) exists, the electron (0eB) and hole (0hB) barriers should be modified by adding the shift (A): *k.B=IP-Ou-* 2.6 te.B=a>u-EA.+ A. = Eg-faiB 2.7 In the 2PPE process as shown in Figure 2.16, when the photon probes a fixed intermediate state, the energy of the highest intermediate state below the vacuum level is the photon energy minus the kinetic energy maximum (KE^). If that intermediate state also corresponds to the lowest allowed energy for electrons above the occupied states then the electron injection barrier (Be) from the Fermi level equals the difference between the probe photon energy (hv) and change in the high energy cutoff (&KEmaii). 20 EF X AKEmax=hv-Be HOMO Au Polymer Figure 2.16: Interfacial electronic structure diagram in 2PPE for bare gold and polymer/metal: O: the metal work function, IP: the ionization potential of the polymer, A : vacuum level shift, KEmax: the width of kinetic energy distribution, AKEmax: change in the high energy cutoff, Be: electron injection barrier, Bh: hole injection barrier. Figure 2.17 shows the energy band diagram at the interface between a metal and a conventional semiconductor before and after equilibrium [62]. In the simplest case of semiconductor-metal interface, the vacuum levels are already at the same level before contact. However, in the general case at metal/organic interfaces, electron redistribution realigns the electronic energy levels. As shown in Figure 2.17, charge transfer occurs until the Fermi level aligns at the same level, and a space charge region occurs at the interface with electrostatic gradient. The charge distribution in the region can be expressed as Poisson's equation [63]: d2V(x) I dx1 = -p(x) I ee0, where p(x) is the space charge density and s is the dielectric constant of the polymer. In a conventional junction, unless the Fermi level is close to band edges, charge concentration in the space charge region equals the dopant concentration. If this is constant through the region, the band bending described by the Poisson equation is quadratic. However, if the Fermi level of the metal is lower than the HOMO level, a big charge transfer occurs, and the bending is not quadratic. 21 EA LUMO HOMO E, A ... L w Figure 2.17: The electronic structures at the semiconductor-metal interfaces before and after contact. The width of the space charge is represented as W. Other than the charge transfer band bending, the observed potential changes have also been attributed to an interface dipole formation. The interface dipole is localized within a molecular layer. Seki and coworkers [61] proposed six possible mechanisms, shown in Figure 2.18 for the formation of an interface dipole; (a) charge transfer localized to the adsorbed 1st layer, (b) induced molecular polarization stabilized by image potential, (c) pushing back the electron cloud tail that extends out from the metal surface by the organic material, (d) chemical interactions, (e) formation of interface state, and (f) the permanent dipole of the organics. Charge Transfer Cation Formation Anion Formation Mirror Force Surface Rearrangement ii <a1) Chemical Interaction (a2) (b) Interface State (c) Permanent Dipole D (d) (e) Figure 2.18: Possible interactions forming an interfacial dipole layer [61]. 22 The magnitude of an interface dipole may generally depend on the adsorbate and on the metal work function. For 4,4'-AyV'-dicarbazolyl-biphenyl (CBP) [64] and zinc tetraphenylporphyrin (ZnTPP) [61], the surface dipoles are insensitive to the work function of metals. For CBP on Ag, Mg, and Au metals, the electron barriers are 0.5 eV, 0.7 eV, and 1.2 eV, respectively, and the interface dipole shifts are -0.5 eV [61]. For 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) on various metals, the interface dipole strongly varies with the work function of a metal substrate. However, for some organic materials, Gao and coworkers observed that the interface dipole has a linear relationship with the work function of metals [65,66]. The interface dipole and hole injection barrier at organic/organic interfaces are much smaller than those at the polymer/metal interfaces. Koch et al. used PEDOP/PSS and Au with nearly identical work functions, as substrates [67]. For pentacene and A/^V-diphenyl-A/,Ar'-bis(l-naphthyl)-l,rbiphenyl-4,4'diamine (a-NPD) on the Au substrate, interface dipole shifts are measured to be 1.05 eV and 1.15 eV, respectively, whereas on the PEDOP/PSS substrate, the interface dipole shifts are 0.3 eV and 0.1 eV, respectively [67]. They have explained that gold has a strong surface electronic component, with the electron density tailing from the gold greatly influenced by the adsorbates, resulting in a substantial change of the work function. Hill and coworkers have also determined dipole shifts and HOMO levels for various organic molecular semiconductors on metal and polymer substrates [64]. The relative importance of electrostatic band bending and interface dipole models has been a topic of debate. However, both probably occur simultaneously. Schlaf et al. proposed a method to discriminate between the two using XPS and UPS [68]. Generally, for undoped organic materials (with low charge carrier concentrations) the abrupt change in potential within the first molecular layer has been explained by an interface dipole effect (due to strong polarization), and the change in wider region is likely due to a band bending effect [69,70]. However, the potential change within the first layer is also likely contributed from the band bending effect [70]. Schlaf et al. observed a large band bending of 1.47 eV and a dipole shift of 0.71 eV at tris(8-hydroxyquinolato) gallium(Gaq3)/Pt interfaces, due to a large work function difference [68]. 23 2.3.2 Doping Top electrode metals with low work function in devices are commonly deposited by thermal evaporation. The very first stage of metal deposition is sometimes called the "doping stage". For n-type doping, an electron donor such as alkali metals donates an electron to the polymer chain to form an ionic complex. For p-type doping, an electron acceptor (I2, NOPF6, H2SO4, etc.) withdraws an electron from the polymer chain to form an ionic complex. Doping is employed to introduce charge carriers into the polymer, increasing the electrical conductivity. The charge transfer process induces local relaxations of the polymer chain. The equilibrium geometry of the ionized state is different from that of the ground state. The geometrical relaxation and charge redistribution induce new localized electronics states in the band gap [28]. In conjugated polymers with a non-degenerate ground state and highly anisotropic quasi-one-dimensional structure, the charged species are called polarons and bipolarons. Solitons exist only in systems with a degenerate ground state. Upon generation of an exciton by photoexcitation, if one of the charges hops to neighbor chains without recombination, a polaron is formed [71]. A bipolaron can be formed by the interaction of two polarons. It is believed that polarons and bipolarons exist competitively. Their coexistence or dominance may depend on the polymer systems, doped ions and the concentrations. In some cases, the assignment of new gap states is ambiguous. Figure 2.19 shows the self-localized orbital structure suggested for charged polaron and bipolaron states in doped polymeric systems [28]. The dots represent unpaired electrons in pz orbitals. Figure 2.20 shows qualitatively the energy states for these self-localized states suggested by Su-Schrieffer-Heeger (SSH) theory. In case of a negative polaron, the lower polaron level is filled and destabilized with respect to the HOMO of a neutral state; in addition, the higher level is half filled and stabilized with respect to the LUMO. For a negative bipolaron, both levels are filled, and the gap is smaller than polarons. The self-localized excitations have been extensively studied using optical absorption, IR, Raman [72], ESR, UPS and NEXAFS [73]. Photoinduced absorption spectroscopy has been widely used for studying new subgap states. By measuring the transmission change (AT IT) of probe beams due to the transient states induced by 24 pump beam, new subgap states as well as photoexcitation dynamics can be investigated. In t he t echnique, - AT I T = Aad = Nad , w here N i s t he p hotoexcitation d ensity, cr is the absorption cross-section, Aa is the absorption change and d is the sample thickness [74]. (a) (b) (c) Figure 2.19: Structures of doped systems (a) negative polaron (b) two negative polarons (c) negative bipolaron. Neutral iillll •Iii HI Positive Negative Positive Negative Singlet Triplet Polaron Polaron Bipolaron Bipolaron Exciton Exciton Spin 1/2 Spin 1/2 Spin 0 Spin 0 Spin 0 Spin 1 Charge+e Charge-e Charge+2e Charge-2e Charge 0 Charge 0 Figure 2.20: Schematic of the self-localized state energies. 25 2.3.3 Some Aspects of Photoexcitations The mechanism for photogeneration of charge carriers in conjugated polymers are still in controversy [12]. Under what conditions are the primary photoexcitations tightly bound electron-hole pairs (excitons), charged polarons or both? Are free carriers generated from secondary processes such as exciton-exciton annihilation? Do the carriers self-localize subsequently to form excitons due to electron phonon interactions? Figure 2.21 depicts the possible relaxation pathways of an initial photoexcitation labeled Sn [9]. An exciton dissociates to form polarons with an efficiency (l-r|) which depends on energy and derealization of Sn. Otherwise it may quickly relax to the manifold of bound delocalized exciton states. The exciton returns to a ground state via radiative (kr) or non-radiative {knr) recombination. The spin state can change by an intersystem crossing (kisc); the larger the spin-orbit coupling in a system, the higher chance of intersystem crossing. The spin-orbit coupling is stronger for polymers containing heavy atoms, for instance sulfur containing polythiophenes. Triplet-triplet annihilation can also occur with second order rate constant y . The whole excitation can also transfer to the neighbor via Forster energy transfer process. At high excitation concentrations, the excitons can also be dissociated by exciton-exciton annihilation. Figure 2.21: Simple schematics of photoexcited state decay channels in conjugated polymer. S, T and P represent singlet, triplet and polaron states, respectively [9]. 26 Emission with a relatively narrow linewidth is indicative of phosphorescence from the localized triplet excitons [75]. Generally phosphorescence in these materials is very weak, especially with excitation in the visible region. The triplet state in P3HT is populated by means of upper excited-state transfer [76], from the higher energy S„ states since the energy difference of Sn - Tn is smaller than that of Si - Ti, allowing more efficient intersystem crossing. Kraabel et al. reported an intersystem crossing (Si -Ti) time of 1.2 ns for a regioregular P3HT in xylene solution [77]. They observed the same intersystem crossing time at 4 eV (higher excitation) and 2.4 eV (lower excitation) of excitation energies. Photoluminescence (PL) is a radiative recombination of optically created excitons. Electroluminescence (EL) is a recombination of excitons formed by an electrical excitation. The decay rates of the optically and electrically created excited states are similar [78]. An exciton extended over two adjacent polymer chains is called an interchain exciton, sometimes referred to as polaron-pair or charge transfer exciton. The interchain excitons are less efficient and non-radiative. Luminescence is mainly due to intra-chain singlet excitons. The luminescence efficiency is strongly influenced by the exciton lifetime and binding energy. The inter-chain interactions are believed to reduce luminescence efficiency in films with no remarkable change in the lifetime of emissive excitons [79]. The PL intensity in films is lower than in solutions and diluted blends where polymer chains are well separated from each other. The nonemissive inter-chain charge pairs [80] and excited-state dimers have been suggested to explain PL quenching resulting from interchain interactions. The luminescence quantum yield of 0.10 in MEH-PPV film is lower than that of 0.35 in solution. Excimer states with a long lifetime and broad red-shifted emission has been observed in CN-PPV [81,82], ladder poly(paraphenylene) (L-PPP) [83] and MEH-PPV [81,83]. The luminescence decay time in films is faster than in solution [84] due to a formation of polaron pairs. The luminescence decay time in MEH-PPV film is also dependent on the excitation intensity. A longer lifetime is observed at lower intensity [81]. The quantum efficiency of RR-P3HT is about an order of magnitude lower than that (8 %) of RRa-P3HT due to enhanced interchain interactions [85]. The interchain charge pairs (polaron pairs) are generated with an efficiency of -20 % in polythiophenes [80]. 27 The PL lifetime of substituted polythiophenes increase with increasing side chain length [86]. The difference between the emission and absorption edges is likely due to a lattice relaxation energy gained in the creation of an exciton [87]. Excitons and presumably polarons can migrate to a more conjugated (lower energy state) chain segment prior to recombination. This causes a red-shift and narrowing of the luminescence peak. For less conjugated polymers, a larger red-shift is obtained due to higher concentration of defects. MEH-PPV is less conjugated than PPV [88]. In most experiments the photon energy used is near the band gap energy, and the time-scale probed is sub-nanosecond, so the main excited species are excitons. In inorganic semiconductors, the optical excitations are well described in terms of a weakly bound electron-hole pair called a Wannier exciton, with a binding energy of the order of 0.01 eV. In organic molecules, the exciton is confined in a single molecule (Frenkel exciton), leading to a binding energy of the order of 1 eV. In photovoltaic devices, the exciton created by photoabsorption dissociates into a free electron and hole which move to different electrodes. For this reason, small binding energy resulting in fast separation of charges is desirable. In LEDs, large binding energy increases the chance of fast radiative recombination of electron and hole pairs created by carrier injection from two different electrodes. Increased luminescence quenching with excitation density has been explained by exciton-exciton annihilation. This bimolecular exciton-exciton annihilation is commonly reported in conjugated polymer systems [12,89]. The bimolecular and monomolecular exciton decay processes are simply expressed as — = G(t) yn1 2.8 dt r where G(t) is a generation rate of photoexcitations, x is an excitation lifetime without interactions between excitons, and y describes the bimolecular annihilation process. For a -sexithienyl films probed using pump-probe photoinduced absorption measurements, there are two main relaxation decay paths: one with a lifetime that depends on the excitation density, and the other which is constant. The fast decay is a exciton-exciton annihilation with a bimolecular rate constant of about 3 x 10"13 cm3 s"1 28 and the slow decay is a monomolecular process by exciton diffusion towards recombination centers within a disordered system [90]. Dicker et al. [91] have used a microwave impedance technique to study the photogeneration of charged carriers in polymers. The change in absorbed microwave power in the sample during simultaneous irradiation with a laser pulse is directly related to the transient photoconductivity. The square-root dependence on the laser intensity for P3HT and substituted PPV indicates that the relaxation mechanism for the charged species is a second order process. The decay kinetics are strongly dependent on temperature, with more long-lived charge carriers at low temperatures. Dicker et al. used an incident photon energy with a 2.5 ns pulse width and ~1014 photons/cm2 (0.04 - 1.75 mJ) per pulse, slightly higher than the intensities used collecting the 2PPE yields and spectra shown in this thesis. In those experiments they used 2.0, 2.5 and 3.5 eV photon energies and found the carrier yield increased with photon energy. At low laser intensities, Binh et al. observed a linear relationship of DC photoconductivity with laser intensity for poly(3-alkylthiophene) [92]. Up to 2.5 eV photon energy (absorption maximum), the photoconductivity follows the absorption profile, but then further increased up to 3.5eV although the absorption coefficient decreases [92]. It is known that the probability of a charge separation is increased using higher energy photons. Kohler et al. [34] obtained photocurrent quantum yields for pristine MEH-PPV and C6o-blended MEH-PPV films as a function of excitation energy. For the pristine MEH-PPV, the photocurrent increases with increasing photon energy. Although the absorption maximum is at around 2.5 eV, the photocurrent yield at the maximum is negligible. For the C6o-blended polymer film, the photocurrent spectrum follows the absorption profile of the film. This indicates that C6o efficiently facilitates charge separation. The enhanced charge separation is accompanied by a PL quenching at higher excitation photon energies and enhancement of bipolaron and triplet excitons [93]. Popovic and coworkers [94] observed a fluorescence lifetime quenching for MEH-PPV at high- energy excitation. Lim et al. have attributed fluorescence quenching to a wavelength dependant creation of charge-separated polarons [89]. It should be noted that higher excitation energies also increase the rate constant for relaxation into bound exciton states [95]. 29 Population of the long-lived triplet and polaron states leads to a depopulation of the ground state, giving rise to a bleaching of the absorption [96]. Polymers showing relatively easy saturation of excited states exhibit a large absorption cross section and long excited-state lifetimes. Consequently, the depletion of ground state could result in, for instance, fluorescence bleaching [89]. One of the great tasks, in conjugated polymers is the determination of the exciton binding energy. Experimental methods include photoconductivity measurements [97] and STM [98]. Equation 2.9 defines the exciton binding energy (Eb) for a conducting polymer in the long chain limit [99]. Eh=E -E , 2.9 b gsp g.opt where Egsp is the single particle energy gap, the energy difference between the electron and hole polaron energies. Eg!0pt is the energy required to create an exciton. The optical band gap (Eg;0pt) is commonly defined as the energy difference between the lowest unoccupied conduction band and the highest valence band, and can be determined from a well-defined optical absorption edge. An exciton requires an excess energy greater than the binding energy to ionize and produce charged free polarons. The exciton binding energy reported for the most studied system, PPV ranges widely from thermal energy to ~1 eV [98,99]. 2.3.4 Doped Polymers ESR is a powerful tool to study the nature of charged states generated upon doping. The upper negative polaron is a single-charged state and shows a positive response to ESR. For bipolarons, the state is fully occupied by electrons and is zero spin state showing no ESR signal. The decrease of ESR signal with increasing doping level indicates that bipolarons become dominant. UPS and XPS are also used to probe the states. Ettedgui et al. used NEXAFS technique and observed new states upon calcium deposition on Ca/DP-PPV. On the contrary, no new state was observed upon Al deposition [73]. 30 Figure 2.22 shows UPS spectra for a doped polymer showing new sub-gap states [100]. The bipolaron band splitting for Na-doped CN-PPV is 1.05 eV smaller than 2.0 eV for Na-doped PPV at saturated doping level. This has been explained by a strong confinement of bipolaron wavefunctions in CN-PPV. The depth of the bipolarons into the sample is about 20 ~ 40 A [101,102]. Binding Energy (eV) Figure 2.22: Typical UPS spectra for a doped polymer with coverage. New gap-states are shown in the spectra [100]. For Cs and Ca-doped p-sexiphenyl (6P), bipolarons are reported as the stable charged species [103,104]. The cesium does not diffuse very readily through the oligomer, and no shift of the bipolaron peak was observed with increasing doping levels. It was reported that the characteristic spectra of pure metallic calcium did not become visible after more than 5 nm of deposited Ca. Salaneck and coworkers also investigated Al, Na and Ca on diphenyltetradecaheptaene (DP7) using XPS, UPS and quantum chemical calculations [105]. If the Fermi level is located at the new peak in the density of states (DOS), the new state is assigned to a polaron. For bipolaron, the EF is located above the new peak 31 DOS position. For high doping levels, the intra-gap emission is due to bipolarons. The Na and Ca were said to diffuse into the polymer and form ionic bonds inducing bipolaron states in the forbidden gap. In contrast, Al initially forms covalent bonds with no gap state formation. In Rb-doped PPV, polaron and bipolaron states are formed at low and high coverages, respectively. However, when Cs is doped in PPV, only bipolaron states were observed over all the doping stages [106]. During metal deposition, diffusion and clustering of metal atoms is involved depending on the metal and surface condition. Reactive, metals are less likely to diffuse deeply into the polymer than noble metals [107]. It has been observed that in the initial stages of calcium deposition on oxygen-containing organic materials, the charge transfer complex Ca2+ O2", or more likely Ca2+0"e" where e" is a delocalized electron, is formed [108]. In addition, oxygen may migrate towards the surface to react with the calcium. The migration has been observed for other highly reactive metals such as K, Mg, and Al [109]. Angle resolved XPS results showed that the concentration of oxygen on the surface is increased substantially during the initial stages of Al or In deposition on MEH-PPV and PPV with 6 % bulk impurity oxygen. For Cu deposition, copper diffuses readily into the near surface of MEH-PPV and PPV polymers with no outstanding migration of oxygen towards the interface [110]. Potassium diffuses more readily than any other low work function metals, showing a distinct charge transfer doping effect and oxygen migration [111]. It is commonly believed that without oxygen and impurities in polymers, calcium diffuses into the clean polymer. For Ca on DP7 [105] and DHPPV [112], calcium diffuses and donates electrons to the 7t-system of the polymer to form Ca2+ ions. The Na, K or Rb diffuses uniformly into polymers. For Rb and K depositions, a longer time is required to reach equilibrium than for Na deposition [112]. As sketched in Figure 2.23 [108], in oxygen-containing surfaces, an insulating layer (interfacial oxide) is initially formed, followed by the metal deposition on top. If the oxide layer is thin enough, electrons can tunnel from the cathode into the polymer [108]. It is believed that the thin interfacial layer entirely determines the efficiency of a device, thus controlling the layer is very important. Cao et al. demonstrated that using an ultrathin Ba cathode with Al or Ag capping layer improves the device lifetime and 32 electron injection [113]. On the other hand, the thick layer (>50 A) ofBa metal enhances device degradation. The EL efficiency can be enhanced by inserting a thin CsF layer between MEH-PPV and Al. The CsF may be dissociated and doped in the polymer [114]. M PPV Ca doped layer with oxygen (a) (b) (c) Figure 2.23: Three cases of calcium on polymer (a) ideal metal on polymer, (b) doped layer at the interface, and (c) calcium oxide formation [108]. If a chemical reaction is involved, the interfacial structure is even more complex. For Ca on tris-(8-hydroxy quinoline) aluminum (Alq3) [115,116], calcium-Alq3 complexes are initially formed by charge transfer reaction, and at high coverage Ca reacts chemically with abstracted oxygen to form Ca-0 bonds resulting in decomposition of the Alq3 molecule. At low coverage, calcium first reacts with nitrogen in the pyridine group. Nitrogen is a more reactive site than oxygen. In the case of Al deposition, Al reacts preferentially with the quinolate oxygen. Exposure to molecular oxygen eliminates the interactions between the deposited Alq3 and Ca substrate. Even 1 A of Al deposition induces new peaks in XPS and gap state peaks in UPS [116]. Unlike the cases of Au and Ag, Ca deposition on pentacene [66] as shown in Figure 2.24, exhibit a dipole shift of 0.35 ~ 0.4 eV and a hole barrier height of 1.65 ~ 1.7 eV. The HOMO level observed in pentacene on Ca is achieved at 8 A of calcium deposition on pentacene. The metallic Ca2p XPS peaks appear a fter 32 A of Ca deposition. The work function of clean metallic calcium is achieved at 64 A of calcium deposition. At low coverage, the high binding energy (BE) Ca2p peaks imply Ca2+ species. The depth of the doped region estimated by XPS is larger than 50 A [66]. 33 Ca Y Pentacene Au \ Pentacene \ Ca Figure 2.24: Interfacial energy diagrams for calcium on pentacence/Au [66]. For metal deposition on PPV, a gradual band bending with metal coverage has been observed unlike the metal-inorganic semiconductor where the bending occurs with a few monolayers of overlayer metal film [117]. It has been observed that the surface oxidation of a polymer inhibits band bending. The metal-oxide layer screens the polymer and then retards the doping effect. If the oxide layer is thick, the chances of electron transfer through the layer are significantly reduced. Hsieh et al. observed that the bend bending at the Ca/PPV interface occurs very late in the deposition process and the interaction between calcium and PPV is not strong [118]. No significant shift of Cls XPS peak was observed until a sudden shift between a 8 A and 15 A of calcium deposited. In contrast, for Al on PPV a faster band bending was observed and the bending saturates at 5 A of Al deposition. The Cls XPS peak gradually shifts to lower binding energies. Hsieh et al. studied an effect of surface species (impurities) on metal interface formation by varying the contents of oxygen and sulfur impurities in PPV [118]. A sulfur-containing surface induces an upward bending. On the contrary, the sulfur free surface induces a downward bending resulting in a higher Cls binding energy. Aluminum deposition on the polymer exhibits a similar band bending, but more rapidly than calcium. For ~10 % of oxygen containing PPV, Hsieh et al. noticed no band bending during Al deposition. Konstadinidis et al. also observed no indication of bend bending for Al deposition on ~4 % of oxygen-containing PPV [119]. The Ca and Al react immediately with surface oxygen to form oxides, which may inhibit charge transfer and retard the band bending. Gao et al. introduced a band bending modified tunneling 34 (BBMT) model for Ca/MEH-PPV/ITO to explain the deviation of I- V measurement from Fowler-Nordheim tunneling model [120,121]. Park et al. observed band bending using XPS and UPS for a Ca doped impurity-free PPV oligomer, 5PV [117]. A broadening of the Cls XPS peak indicates a charge transfer from calcium to the polymer accompanied by a peak shift to lower binding energy and a formation of new gap states. They found a total band bending of 0.5 eV and a space-charge region of about 10 nm. For 5PV deposition on calcium, no gap state was observed. The presence of metal film on a luminescence polymer is commonly accompanied by a quenching of luminescence efficiency. Even 0.1 A of calcium deposition on phenylene vinylene oligomer quenches the PL intensity dramatically by 50 % with no change of PL shape [120,122]. Choong et al. found three stages of the luminescence quenching [122]. The first stage of PL quenching (1 A) has been explained by a nonradiative decay channel via polaron and bipolaron states created by calcium deposition at the interface, which act as exciton dissociation sites. In the second stage with no indication of metallic calcium, the decrease is much slower. This was explained as due to an overlap of quenching radii defined by the exciton diffusion length. For thick (>30 A) calcium deposition stages, optical attenuation of excitation and emission photons due to the overlayer metal is likely a main factor for the PL quenching. A complete PL quenching thickness of 200 A is related with an exciton migration distance. It is believed that PL quenching is also due to dipole interactions in the presence of metal. Park et al. observed that the PL quenched by metal is recovered upon oxidation of calcium. They attributed the PL recovery to the disappearance of mid-gap states formed in the fresh Ca/4PV [123]. Broms et al. demonstrated that Ca/CN-PPV/ITO LEDs prepared in the presence of background oxygen pressure have better performance than the devices fabricated with a clean polymer in UHV condition [124]. Liao et al. studied oxygen effects on a calcium/polyfluorene interface shown in Figure 2.25 [125]. An exposure of oxygen removes gap states induced by calcium doping. An excess exposure produces a very high electron barrier height of 3.1 eV and a very wide band gap. For calcium deposition on poly(9,9-dioctylfluorene) (PFO), no change of work function was observed after 16 A of calcium deposition. It indicates the formation of metallic calcium. An energy separation between the two bipolaron peaks is 1.8 eV. In 35 addition, a broadening of the LUMO was observed. Evac Figure 2.25: Energy diagram of Ca/PFO/ITO [125]. 2.3.5 Some aspects of P3HT A wide range of ionization potentials has been reported for P3HT, ranging from 4.2 eV [126] to 5.2 eV [127]. Bredas and coworkers obtained an ionization potential of 5.0 eV for un-substituted polythiopehene and P3HT using theoretical calculations [128,129]. For the un-substituted polythiopehene, Eckhardt et al. report an IP of 5.2 eV and a band-gap of 2.24 eV using electrochemical potential spectroscopy (ECPS) [130]. In ECPS, one measures the electrochemical potential as a function of charge put in or taken out of the polymer electrode. Onoda et al. obtain an IP of 4.7 eV for P3HT (< 20nm) using a low-energy photoelectron spectroscopy instrument in air [131]. Onoda et al. reported that the ionization potential of P3HT on ITO decreases with increasing thickness up to 10 nm. The variation is larger at higher temperatures due to an enhanced electron transfer. However, for P3HT on gold, the ionization potential of 4.7 eV was independent of the polymer thickness, taken to indicate an absence of charge transfer between gold and P3HT. The IPs of P3HT are summarized in table 2.1. 36 IP (eV) 4.9 5.0 5.2 Ref. 126 131 132 128, 133 134 127 Table 2.1: Ionization potential values of P3HT in literature. The optical band-gap (Eg) is estimated by the onset edge of the absorption spectrum. The band-gap decreases as the conjugation length of a polymer increases. If the polymer is thin, the conjugation length is likely influenced by the substrate. For highly regioregular P3HT, the band-gap in a bulk state has been reported to be 1.7 [135] and 1.8 eV [136]. Higher values have also been reported [137,138]. The band gap for an electrochemically deposited P3HT (560 nm thick) film was reported to be -2.0 eV [134]. The drop-casting films have a smaller band-gap than spin-coated films [137]. Chen et al. estimated the band gap of P3HT ranging from 1.7 eV (98.5 % RR) to 2.1 eV (50% RR) depending on regioregularity of the polymer [135]. Regioregularity as shown in Figure 2.26 is a measure of head-to-tail couplings (HT-HT) couplings in the polymer. (c) HT-HH (d) TT-HH Figure 2.26: Possible regioregularity in alkyl substituted polythiophenes chain. H and T mean head and tail, respectively. Head carbon is located between sulfur and carbon with alkyl group. Tail carbon is located between sulfur and carbon with no alkyl group. 37 The high ordered crystalline structure obtained by a self-organization of polymers on metal results in strong interchain interactions; consequently, charge carriers are no longer confined to a single chain and increased charge carrier mobility is observed. Mobility is a measure of the time interval between collisions for a carrier moving through a semiconductor lattice, and the effective mass of the charge carrier (// = qr I m'). The carrier mobility in organic materials has been measured using various techniques including time-of-flight (TOF) measurement. In this technique, one measures the transit time of carriers generated by low intensity laser pulses, and the carrier mobility is expressed as p = d2/xV, where x is the transit time of the carriers, dis the film thickness, and Vis the applied voltage [139]. The mobility is highly dependent on the solvent used in the film casting as well as the regioregularity of the P3HT [140]. Films prepared from chloroform solution have two orders higher mobility than from THF [140]. Spin-coated films generally have lower mobility than solution-casting films. The slower evaporation of solvent during solution-casting presumably allows more time to grow more ordered films [140]. The field-effect mobility in FETs prepared by a printing technique [141] was reported to be 0.01 ~ 0.03 cm2/V s. Dicker et al. [142] obtained a higher mobility (0.08 cm2/V s.) in drop-casting films than in spin-coated films (0.05 cm2/V s). However, for non-regioregular P3HT films, Kobashi et al. [143] reported a higher mobility for spin-coated than for the cast films. A high-mobility field-effect transistor (FET) with p, = 0.05 to 0.1 cm2 V"1 s"1 was demonstrated using spin-coated RR-P3HT on a SiO*2 substrate treated with a silylating agent to improve self-organization of the polymer chains [21]. The mobility of ultra-thin P3HT films has lower value than that of thicker films [144,145]. In a LB film device composed of 2 ~ 5 LB layers (4 to 10 layers), although the mobility is insensitive to the film thickness, the mobility is higher than that of a single LB layer [146]. The lower mobility in ultra thin films may be due to a charge carrier confinement restricting hopping of carriers to the next chain. The mobility is also dependent on the molecular weight (MW) of a polymer. Kline et al. showed that the mobility of RR-P3HT increases with increasing MW of the polymer [147]. Literature mobility values for P3HT are summarized in table 2.2. According to early reports by Heeger and coworkers [148-152], charges were 38 predominantly in the form of bipolarons in photoexcited and metal-doped polythiophenes. The bipolaron state should have no ESR signal and two induced optical absorption peaks. On the contrary, the polaron state should have a positive ESR signal and three allowed transitions in a photoinduced absorption spectrum. If the symmetry is considered, only one transition for bipolarons and two transitions for polarons should be observed as shown in Figure 2.27. In case when the symmetry is broken, disorder may assist bipolaron formation, and bipolarons can exhibit two transitions [148,149]. P3IIT Flole mobility (cm2 V'1 s"1) Thickness, Regioregularity (RR) Substrates ref. Spin FET lO^-lO"5 lfJ7 Regioirregular/10"8 S 50-100 nm, and thinner Si02/Si 145 Spin and cast FET 0.01 (spin) 0.045 (cast) regioregular Si02/Si 153 LB-FET 0.02 - 0.003 regioregular Si02/Si 146 Spin-FET 0.05-0.1 96% hydrophilic Si02/Si 21 Spin-FET 10"4~ 10"3 5-10 nm, >98.5 %,chloroform Treated Si02/Si 144 Electrochemical lO"4 NA ITO 134 Table 2.2: Hole mobility in various P3HT based devices. Using photoinduced absorption and ESR measurements, Osterbacka et al. have assigned two strong peaks at 0.35 eV (Pi) and 1.27 eV (P2) to ID localized polarons, and peaks at 0.06 eV (DPi) and 1.8 eV (DP2) to 2D delocalized polarons resulting from increased interchain coupling [13]. The polaron formation mechanism has been explained by introducing a bimolecular recombination process [144]. For a ladder-type poly(p-phenylene) with a rigid backbone structure, a low polaron peak at -0.1 eV implies small polaronic relaxation energy [154]. Furukawa reviewed photo-generated self-localized excitations and concluded that 39 polarons are the major species in nongenerate polymers. In his review article [12], he reassigned bipolarons to polarons summarized in table 2.3, based on the symmetry and the absorption intensity. conduction band Figure 2.27: Schematics of the gap states and symmetry-allowed optical transitions for polarons and bipolarons. Polymer film Band (eV) Early assignments New assignments [12] PT 0.45, 1.25 1.95 bipolaron exciton polaron PT 1.8 polaron -P3HT 0.35, 1.3 bipolaron polaron P3HT 0.5, 1.1 bipolaron polaron PPV 1.36 triplet exciton -PPV 0.55, 1.46 bipolaron polaron MEH-PPV 0.43, 1.36 bipolaron polaron DMO-PPV 0.45, 1.35 bipolaron polaron Table 2.3: Assignments of the photoinduced absorption bands in conjugated polymers. 40 Chapter 3 Experimental This chapter deals with the instrumentation and experimental techniques used during the thesis project. 3.1 Ultra-High Vacuum System The major design features required for these experiments were 1) Ultra-High Vacuum (UHV), in order to maintain a contamination-free surface during the experiment, and to prevent scattering of the photo-emitted electrons by residual gases, 2) a hemispherical deflection analyzer, to obtain electron kinetic energy distributions, 3) a sample emission region that was extremely well isolated from stray electrical and magnetic fields, 4) a tunable excimer pumped dye laser. Cylindrical Mirror Figure 3.1: Schematic of the 2PPE system. 41 Figure 3.1 shows a schematic of the 2PPE system. Our 2PPE system consists of two main parts; a preparation chamber and a photoelectron spectroscopy chamber, separated by a 6"gate valve. The preparation chamber is equipped with a cylindrical mirror electron analyzer (CMA) with an on-axis electron gun, a sputtering ion gun, a liquid nitrogen trapped diffusion pump which is isolated from the chamber by an 8" gate valve, various sample positioning tools including a XYZ sample manipulator, a sample heating dock, a sample signal feedthrough, and several view ports. The sample heating dock is attached to the XYZ manipulator. The modular manipulator allows xy (±25 mm), and z-axis (±50 mm) motion and full axial rotation (360°) of the sample. It allows the sample to move to desired positions for sputtering, AES and deposition. The transferring and positioning of samples was done using the manipulator (Fisons OMNIAX 100, Fisons RD2 and Fisons SH2), a wobble stick (VG ZWS225) and a magnetically coupled transporter (MDC MTM-24). The wobble stick is equipped with a handling fork. The fork transfers a sample from the heating dock to the magnetic transporter and vice versa. The magnetic transporter allows a linear motion of the sample, up to 60 cm. By using it, the sample can be transferred from the preparation chamber to the 2PPE chamber. It also allows 360° rotation of the sample around the axis of linear motion. The 2PPE chamber includes a concentric hemispherical analyzer (CHA) for measuring the kinetic energy of photo emitted electrons, a water-cooled titanium sublimation pump, a p-metal shield, and a fused silica (Suprasil-1) UV-grade laser entry window. The pressure was monitored using a Bayard-Alpert ionization gauge (KJL G8140-DI) connected to a KJL IG-4400 or a VG IGP3 controller. 3.1.1 Vacuum pumps There are several vacuum pumps used in this system. A two-stage rotary vane pump (Alcatel 2010) and a sorption pump were used to pump the gas manifold. A diffusion pump (Varian VHS-4) was used to pump the chamber, backed by a two-stage rotary vane pump (Edwards RV102) and augmented by a liquid nitrogen trap to enhance pumping speed and to reduce oil backstreaming. One major disadvantage of diffusion pumps is oil backstreaming into the vacuum chamber. To minimize sample contamination, we used a high quality vacuum pump fluid, Santivac 5P polyphenyl ether (Scientific Instrument 42 Services, Inc.) which has lowest backstreaming characteristic, excellent thermal stability, and is suitable for untimate pressures below 10"10 torr. The advantages of diffusion pumps include a high pumping speed and efficient pumping of all gases. They have a lower initial cost than turbomolecular pumps and ion pumps, but a high operating cost in liquid-N2 consumption for the trap. A lower pressure is achieved after a bakeout procedure which degasses the chamber. This uniform heating of the system removes c ontaminants adsorbed on the chamber's internal wall especially adsorbed water. Heating was provided by the strip heaters (Omega OTF), silicon band tapes (Omega SRFG) and heating tapes (Thermolyne). The system was insulated by covering with a tent of aluminized fiberglass (Burnaby Insulation Supplies) during the bakeout for several days. Variacs controlled the heaters. The temperature was monitored using a Cr/Al thermocouple. The temperature was maintained in the range of 130 ~ 150°C. During and after the bakeout, the titanium sublimation pump (TSP) was used. The TSP is highly effective for pumping active gases such as H2, N2 and CO. 3.1.2 Sample holder and Heating dock The ESCA-style sample-mounting holder shown in Figure 3.2(b) was made of OFHC (oxygen free high conductivity) copper and coated with water-based colloidal graphite called Aquadag (Acheson Colloids Company) in order to ensure a uniform work function and to reduce opportunities for charging [155]. 43 The sample heating (annealing) dock shown in Figure 3.2(a) consists of a solid OFHC Cu cylinder, a UHV button heater (Heat Wave 1136) for resistive heating, and a solid Macor base for electrical insulation. A Cr/Al thermocouple attached to the face of the OFHC Cu dock was used to measure the temperature of the sample. The heating dock is also attached to vacuum feedthroughs for the thermocouple, and electrical supply to the button heater. A homemade controller monitored and controlled the temperature. The heating dock is mounted on a modular UHV manipulator to move the sample to a position for Auger spectroscopy or sputtering. 3.1.3 Electron Energy Analyzer The emitted electrons are collected and separated by an electron energy analyzer according to their kinetic energy. The analyzer determines the sensitivity and resolution of the spectra. There are several types of analyzers. According to the types of instrumentation, one uses a retarding field, time-of flight (TOF), or a deflection analyzer. We use a concentric mirror analyzer (CHA) for 2PPE [156], and a cylindrical mirror analyzer (CMA) for Auger spectroscopy. The CHA (VG Microtech VG100AX) shown in Figure 3.3 also known as a spherical deflection analyzer [156], includes three main components: lens, analyzer and channeltron detector. A potential AV is applied to the two concentric hemispheres of inner radius Ri and outer radius R2, the inner sphere is positive and the outer is negative. Higher energy electrons strike the outer sphere while the lower energy electrons deflect towards the inner sphere. Consequently, only electrons in a narrow energy region (called the pass energy) succeed in traveling all the way round the hemispheres to the detector. There are two lens modes: 1:1 and 1:3 modes. In 1:1 mode we use, both lenses have the same potential. This mode is generally used when the incident beam is relatively large. In 3:1 mode, the rear lens is at the earth potential. An area on the sample slightly larger than one third the size of the entrance slit is analyzed. This mode is used when the source beam is relatively small. 44 Outer hemispher (Negative) \ Figure 3.3: Schematic of the VG100AX concentric hemispherical energy analyzer [156]. Two operation modes in the CHA include constant analyzer energy (CAE) and constant retard ratio (CRR) modes. In the CAE mode, the pass energy (the manufacturer's setting is 7.5 eV) is constant irrespective of the kinetic energy. This mode ensures analyzer resolution is constant over all kinetic energies because the resolution is proportional to pass energy. In the CRR mode, by ramping both the retarding potential (R) and pass energy (HV), the retard ratio, (R + HV) IHV remains at a constant value. In this case, the pass energy increases with increasing the photoemitted kinetic energy, and the resolution is better at lower energies. However, we found that the throughput of electrons through the spectra was significantly less in the CAE mode than the CRR mode with R=0 and with a bias acceleration of 10 V on the sample. For the narrow energy range of our spectra (-10 to —11.5 eV) and broad spectral peaks, the change in resolution across the range is less important than optimizing the throughput. 45 A channel electron multiplier (Channeltron, Galileo 7010) detects the electrons passed through the CHA. There are two channeltron operation modes. In analog mode, the channeltron voltage is set low enough that the output current is proportional to the input electron flux. In pulse counting mode, an input electron produces a fixed amplitude saturated output pulse 10 to 20 ns long. During this time the channeltron does not respond to new input electrons. Therefore, if more than one electron is produced at each laser pulse of 20 ns pulse width, the output is saturated. At the peak maximum in our 2PPE spectra, because there are more than 10 electrons, the analogue mode is appropriate. This required we design a "gated" analogue detection interface, rather than a conventional pulse counting. For scanning the high energy cutoff region with weak electron emission (less than one electron), the pulse counting mode can be used. 3.2 Sample Preparation 3.2.1 Cleaning of Gold Substrate Gold substrates were purchased from Arrandee™ (http://www.rnhs-gmbh.com) and Molecular Imaging (MI) (http://www.molec.com). The size was about 11 mm x 11 mm. The substrate from MHS-GmbH was a 200 ~ 300 nm thick gold film o n borosilicate glass. The substrate from MI was -150 nm thick gold film on mica. The samples were cut to fit on the sample holder. Each gold film was washed in hot HN03:H2S04 (1:1) solution (or hot methanol only) and subsequently in deionized water. The substrate was flame-annealed with a butane micro torch. Atomically flat terraces with 100- 200 nm across can typically be achieved by this flame annealing method with preferred (111) orientation. Chapter 5 will discuss AFM images of the gold films. For 2PPE experiments, cleaning of the Au surfaces was achieved by Ar+ sputtering, using a Physical Electronic 04-161 ion gun with a beam voltage of 1 KV and emission current of 15 mA. The incident angle of Ar+ ions was about 6° with respect to the sample surface normal. The base Ar pressure during the ion irradiation was about 4 xlO"5 torr. The temperature of the substrate was maintained at 550 K to enhance sputtering yield and 46 especially herringbone reconstruction [157,158], followed by annealing at 760 K. The cleanliness of the surface was checked by Auger spectroscopy. 3.2.2 Film preparations The regioregular poly(3-hexlythiophene-2,5-diyl) (P3HT) was purchased from Aldrich Co. The regioregularity is greater than 98.5 % of head-to-tail conformation. The polymer is soluble in chloroform (CHCI3), but insoluble in most organic solvents. The polymer has a molecular weight (MW) of ~87,000 g/mol and a conductivity of 10"6 ~ 10"7 simens/cm. We dissolved the polymer in anhydrous chloroform (CHCI3) and prepared various concentrations of polymer solution. Polymer films on gold were prepared by dip- or spin-coating. Exposure of the cleaned gold substrates to air during the polymer coating was less than 100 sec, and then the sample was enclosed in an N2 envelope during transferring. We check the 2PPE of air exposed and pure-solvent dipped gold samples. They are not remarkably changed from that of the clean gold, showing only a moderate work function decrease, and not the characteristic giant yields and spectral structure of the polymer treated samples. In the dip-coating method, the gold substrate was dipped in a solution and pulled out slowly (-10 sec) from the solution to prevent an aggregation of the polymer because of fast evaporation of the solvent. The thicknesses were controlled by varying the concentrations of the solution. In the spin-coating technique, the thickness was controlled by changing the spin speed at a fixed concentration or changing the concentration at a fixed spin speed. It was difficult to make very thick uniform films (> 100 nm) by spin-coating. The sample was mounted on the ESCA-style sample holder and was inserted into the heating dock for annealing of the sample. Annealing procedures will be described later with results. 3.2.3 Metal Deposition on Polymer Film Calcium deposition was done by a resistive heating method. Figure 3.4 shows a schematic of the homemade deposition system. 47 Shutter with orifice -33-• Substrate Quartz crucible Power supply Figure 3.4: Diagram of a calcium deposition system. The evaporation of Ca metal (Alfa Aesar, -16mesh, 99.8 %) was performed in a quartz crucible (KJ Lesker, 1.5" long, 0.375" ID) wrapped with tungsten coils. A high current was applied to the boat, which undergoes resistive heating. The power supply for the evaporator is a step-down transformer controlled by a Variac. Temperature at the metal source was measured using a Cr-Al thermocouple. To control the flux rate, the temperature was varied between 620 and 690 K corresponding to 41 -46 Amps. The polymer target was kept at room temperature during the deposition. The base pressure during the deposition was about 10"8~ 10"7torr depending on the deposition rate. Before the deposition on the polymer, the flux rate (thickness) was measured by Auger spectroscopy of Ca deposited on an Au substrate taken before and after each metal deposition. Calcium was baked at -200 °C for 24 hours to remove residual gases prior to the deposition. 3.3 Two-Photon Photoemission Spectroscopy The 2PPE experiments were performed in the UHV chamber at a base pressure of 7 x 10"10 to 2 x 10"9 torr. For the experiments on bare gold substrates, the pressure was less than 8 x 10"10 Torr. The laser incident angle was 75° from the surface normal. The laser intensity was minimized to avoid space charging [159]. A negative bias voltage of 10 V 48 is applied to the sample to clearly observe the zero kinetic energy secondary electron emission cutoff. The observed kinetic energy of emitted electrons is the sum of the contact potential and the electron kinetic energy at the sample. The bias voltage distorts the initial angle of the emitted electrons, with the angular resolution reduced by the bias voltage, with more electrons collected. The nominal spectrometer resolution is 50 meV. But angular effects from the bias voltage make the resolution worse. We estimate from the sharpness of the low energy cutoff that a Gaussian representation of the instrumental resolution function has a FWHM of ~75 mV. 3.3.1 Major Components The laser beam was reduced to a diameter of 1mm by two iris diaphragm apertures (Newport ID-10), and passed through a calcite polarizer and fused silica focusing lens. The laser light then passes a MDC fused silica (Suprasil-1) viewport to strike the sample. We find the laser intensity is attenuated about 10% at each element. The attenuation of the viewport was considered in calculating the actual incident laser intensity. The optics are mounted on precision translation stages (Oriel 16121) and a 360° rotation stage (Newport RSP-2) to set a desired position. The spot size was estimated to be about 2.6 x 3 2 10" cm" derived using the distance from the iris diaphragm aperture to the sample, beam divergence and the incident angle of the laser beam. The low energy photoelectrons are easily deflected by magnetic fields. The sources include the Earth's magnetic field and laboratory instrumentation. A magnetic shield was fabricated from 0.050" CO-NETIC® AA alloy and coated with Aquadag. The custom u-metal shield (Magnetic Shield Corporation) shown in Figure 3.5 is a cylindrical tube with a diameter of 2.0" and length of 8.0" and it contains four ports [155]. The ports are equipped with cylindrical extensions longer than the hole diameter. Two ports are for a passage of the incident and reflected laser beams. The other two allow an entry of the sample, and photoemitted electrons towards the analyzer. The top and bottom of the shield are open to facilitate vacuum pumping. The sample was transferred into the shield using a linear transfer device, which is attached to the 2PPE analysis chamber. 49 photoemitted electrons ToCHA sample in 'sample' laser out / Figure 3.5: Schematic of the custom magnetic shield: top and side view [155]. 3.3.2 Laser System A XeCl excimer laser (Lambda Physik, COMPex 102) and dye laser (Lambda Physik, SCANmate 1) [160,161] was used to perform the experiments. The XeCl excimer laser produces 308 nm (4.0 eV) radiation with a nominal pulse duration of 20 ns at full width half maximum (FWHM), repetition rates of 1 - 20 Hz, and pulse energies of up to 200 mJ. The buffer gas (2.4 % HC1 / 1.8 % Xe / 0.14 % HC1 / 0.02 % H2/Ne) was purchased from Air Product Canada LTD. Manufactures specification for the beam divergence is 1.5 mrad and bandwidth is 0.15 cm"1. A laser dye was dissolved in an appropriate solvent shown in table 3.1 [162]. The laser was controlled by the Scanmate software. Dye Turning range (nm) Solvent Conc.(g/L) P-Terphenyl(PTP) 332 - 360 Dioxane 0.08 DMQ 346 - 377 Dioxane 0.07 Furan2 388 - 426 Methanol 0.17 Coumarinl02 460- 510 Methanol 0.8 Rhodamine 6G 569 - 608 Methanol 0.4 Table 3.1: Concentration and tuning range of laser dyes used. 50 The electric field vectors for p- and s-polarization are parallel and perpendicular to the incidence plane, respectively. The beam of the excimer laser is unpolarized. The laser beam of the SCANMATE dye laser is preferentially polarized vertical to the lab floor, which is s-polarization in our experimental geometry. Polarization was set using a Glan-Taylor polarizer (Halbo Optics PH10-M). The p-polarization was obtained using a Fresnel rhomb quarterwave retarder. The Fresnel rhomb utilizes the different phase changes on internal reflection of the p- and s-components of an incident plane polarized beam. The excitation laser intensity to be used was determined by considering the signal to noise (S/N) ratio, space charge and polymer damage effects. Typically, the laser intensities used are < 2 mJ/cm and -20 uJ/cm for bare gold and polymer films, respectively. 3.3.3 Data Acquisition Depending on experiment either CHA channeltron or the total photoemitted charge per pulse were measured using a charge-sensitive preamplifier (Oxford TC174). Figure 3.6 shows a schematic for the measurement of the total charge from the sample as the recharge pulse from the ground through the bias. The charge-sensitive preamplifier (Oxford TCI74) output voltage (1012 V/C) is fed to a DC amplifier (pre-amp stage of EG&G 5209) which functions as a line filter (60 and 120 Hz) and a secondary signal amplifier. In Auger spectroscopy, the CM A output passes through a current preamplifier to the lock-in amplifier that performs not only line filtering and secondary signal amplification but also a phase-sensitive detection. The electrical signal from the lock-in amplifier is acquired using a data acquisition card (Computerboards PCI-DAS 1602/16). The software program sends commands via this card to control the pass energies of the CMA and CHA. The program was programmed using LabView software (National Instruments). Details of the program are described in elsewhere [155]. The incident laser power was measured using a Joulemeter (Molectron J25). The pyroelectric detector generates a voltage as the absorbed optical energy degrades into heat. The calibrated response of the meter is 8.35 V/J. 51 Ground Figure 3.6: Schematic for the 2PPE yield measurement. 3.4 Auger and X-ray Photoelectron Spectroscopy The cleanliness of the gold substrate and calibration of the deposited calcium thickness were checked by Auger spectroscopy [163]. The Auger system includes a cylindrical mirror analyzer (Omicron CMA 100) with integral linear retraction, and a high intensity on-axis electron gun (Omicron EKI 50). The resolution is better than 0.5 %. The CMA channeltron is operated in analogue mode. An operating channeltron voltage of 1.2 kV was used in the analog mode. An electron beam current of 0.2 uA, filament current of 2.2 A and primary beam energy of 4 keV were used. XPS was used to measure the thickness and verify the surface composition without damaging the polymer film. The XPS measurements were carried out using a Leybold Max200 X-ray photoelectron spectrometer equipped with a hemispherical energy analyzer. The XPS data collection was done by Dr. K. Wong in the interfacial analysis and reactivity laboratory (IARL) led by Dr. K. Mitchell. The base pressure was about 5 x 10"10 torr. The non-monochromated Al Ka (1486.6 eV) X-ray source was operated with a power of 10 kV x 20 mA. A pass energy of 192 and 24 eV for survey scans and Cls, Ols and S2p high resolution spectra were used, respectively. The analyzed area was ~ 2 x 4 mm2 and the takeoff angle was 90°. All the core-level spectra were referenced to the Au4f peak at 74.0 eV, in case of surface charging. 52 3.5 Scanning Probe Microscopy Ex-situ AFM images for spin and dip-coated samples before and after the 2PPE experiments were obtained using an Autoprobe CP (Veeco Instruments) in contact or tapping mode [164]. There was no distinct topographical difference between the images recorded using the two modes. The piezoelectric ceramic tube scanner allows a maximum xy scan size of 100 um x 100 um and vertical (Z) range of greater than 7.5 um. For the contact AFM mode, we used a commercially available gold-coated silicon nitride microlever with a spring constant of 0.01 N/m, 0.03 N/m, or 0.1 N/m. The image was also taken using a tapping (non-contact) AFM mode to check tip-induced effects. A stiff cantilever with a spring constant of 17 N/m was used for the tapping mode because adhesion to the surfaces easily traps the AFM tip. For the scratch test, a force of 100 nN and a fast scan speed were applied to the cantilever. We used a conical silicon cantilever with a force constant of 0.26 ~ 2.1 N/m. The force-distance measurement using AFM was used to discriminate bare and polymer covered Au surfaces. The AFM was controlled using ProScan 1.51b software program. IP 2.0 software was used for analysis of the surface image including line depth profiles and 2D and 3D-surface images. 3.6 Some Further Comments Many of the groups doing 2PPE experiments have used time-of-flight (TOF) analysis, [5,6] in which one measures the time between the laser pulse and each electron arriving at a detector some fixed distance from the sample. The kinetic energy of the 1 , electron is given byKE = — me(llt) , where / is the flight distance, t is flight time, and me is the electron mass. The energy resolution of a TOF spectrometer depends mainly on time resolution, varying as AE cc At- E31211. The advantages are lower price, and most importantly, multi-channel detection of the energy of each electron at the detector, rather than a single channel analyzer scanning across the energy spectrum. However that is not critical in these experiments since the photoelectron emission from the 53 heterogeneous films will not exhibit fine structure and the spectral range is not wide, so there are typically only about ~ 20 channels needed per spectrum. The most important disadvantage of the TOF is that the long flight paths required for energy resolution make it more susceptible to stray fields. The long flight path also defines a small collection angle. For experiments on disordered films, where angular resolution is not required, this low collection efficiency just negates the multi-channel advantage. Finally, the time-of-flight technique is most applicable to experiments involving very short laser pulses, where the time resolution is best. As shown in this thesis, one can obtain exceptionally low background emission from metal substrates by instead using long laser pulses, and this is important for interfacial studies of ultra-thin films. Femtosecond laser systems are much more expensive than the excimer system which we use, but have the fundamental advantage of allowing the study of fast photo-excitation dynamics. However in this thesis we focus on the longer time scales over which the diffusion and recombination of charge carriers should occur, an important regime for device applications. 54 Chapter 4 Sample Characterization Several techniques for sample characterization are discussed in this chapter, namely Auger, AFM, XPS and FT-IR methods. Measuring the thickness of ultrathin adsorbed layers is a complex task and done by various techniques like XPS, Auger, ellipsometry, scanning tip techniques, or optical absorption methods. In tip probe techniques, e.g. AFM, a sharp film edge is required to determine the thickness, which in many experimental situations can be difficult to obtain. To determine the thickness using ellipsometry, the refractive index of an adsorbed layer is required, which can be very difficult to determine. In X-ray photoelectron and Auger spectroscopic methods, one first needs the electron mean free path value. The accuracy of this value determines the accuracy of the thickness. Surface morphology can also affect the analysis. 4.1 Cleaning of Gold Surface A bare gold film was flame-annealed in air and cleaned by Ar+ sputtering in UHV before the 2PPE experiments. A complicated surface reconstruction is common for the gold surface. Following the references [157], we try to use lower Ar+ ion flux and a substrate temperature of 550 K (refer to section 3.2) to obtain a smoother surface and herringbone reconstruction during the sputtering. Prior to polymer coating, the gold substrates were washed with hot methanol followed by flame annealing. This flame annealing produces large atomically flat terraces (-100 nm) with relatively few impurities. We also used gold.substrates cleaned by Ar+ sputtering before the polymer coating, and found that the 2PPE results for P3HT films on gold show no differences between the two substrates. 55 Figure 4.1: AFM image (1 pm\\ jum) of a flame-annealed gold surface. nm 1.8-1.4-1.0-0.6 -! 0 400 8001 nm Figure 4.2: Height profiles of the AFM image shown in Figure 4.1. Figure 4.1 and 4.2 show an AFM image and a height profile for a flame-annealed gold substrate. The atomically flat regions for the flame-annealed film are 200 ~ 300 nm wide. The flame-annealed Au films dipped in pure CHCI3 also show flat surfaces. Annealing treatment also enhances atomic features of gold in STM image [165]. Small grains merge to form large grains. However, it is known that the flame annealing is not successful for gold films exposed long in air [166]. The monatomic step height of -2.7 A in the AFM line profile is close to the literature value of -2.4 A [167]. We use this value to calibrate the AFM vertical height. DeRose et al. observed that gold films on mica show a smoother surface than those on glass after flame annealing [168]. However, our 2PPE data did not show significant 56 differences. As flame-annealed gold substrates age, the surface becomes rough, like unannealed gold films. 4.2 Polymer Surface Morphology X-ray photoelectron spectra were recorded to check interactions between the polymer and gold substrates, including polymer degradation. Figure 4.3 shows S2p3/2 and S2pi/2 XPS peaks with an intensity ratio of 4:2. It is commonly known that the sulfur in thiols [169] strongly interacts with gold surfaces, and new extra sulfur XPS peaks are observed due to strong Au-S interactions; however, the interaction of sulfur in P3HT with gold is negligible [170] since no new peaks were observed. Figure 4.3: S2p XPS spectrum for the P3HT on gold. 4.2.1 Spin- and Dip-Coated Films In the spin-coating method, the polymer solution is dispensed on the substrate spinning at high speed so that it is spread across the substrate. The polymer film thickness and morphology depend on polymer concentration and spin speed [171]. The n-n stacking structure of P3HT plays a crucial role in device performance. Two possible stacking structures are shown in Figure 4.4. For dip-casting films, thiophene rings are stacked with face-to-face distance of 0.37 ~ 0.38 nm between the 158 161 164 167 Binding Energy (eV) 170 57 adjacent polymer chains, and the main chain is parallel to the substrate [21]. The n-n stacking direction is parallel to the substrate as shown in figure 4.4(a). The P3HT film in this 2D-lamellar type structure has an edgewise intermolecular spacing of 1.60 ~ 1.67 nm [21], which increases with increasing length of the alkyl side chain [172]. The weak x-ray diffraction (XRD) peaks indicate that the spin-coated films show less ordered structure than drop-casting films [137]. The edgewise distance estimated from STM data is 1.33 nm, lower than the XRD values. It has been commonly observed that dip-coated P3HT films yield higher mobilities than spin-coated films (refer to chapter 2), implying a better ordered structure [153]. According to Sirringhaus et al. [21], lbw-regioregular (81%) spin-coated P3HT films have the structure shown in Figure 4.4(b). However, dip-coated low-regioregular, or spin-coated highly regioregular (96 %) P3HT films prefer the structure in Figure 4.4 (a). (a) (b) Figure 4.4: Two possible stacking structures of P3HT. Recently, self-assembly of P3HT drop-cast on HOPG from chloroform solution has been studied using STM [173]. For low partial coverage of polymer, the lamina tend to lie parallel to the substrate partly despite aggregation of polymer chains as in Figure 4.4 (b) [173]. For higher coverage, the structure is like Figure 4.4(a), and mobility is higher in the parallel stacking direction [21]. Figure 4.5 shows the surface topography of our dip and spin-coated P3HT films. The difference between dip and spin-coated films is obvious after performing fast Fourier Transform (FFT) of the topography. Interestingly, we found remarkably different 2D 58 FFT patterns. This is likely related to wetting and stacking pattern of polymer chains on gold substrate. jum Figure 4.5: AFM images (2 fjmxlp.m) for unannealed (a) dip, and (b) spin-coated P3HT on gold, and their height line profiles. The inset of the AFM images shows their Fourier transform power spectrum. The spin-coated films more strongly adhere to the Au. They withstand higher temperature before dewetting and need more force to scratch off with an AFM tip. Figure 4.5 shows height line profiles for dip- and spin-coated films. In the figure, especially for spin-coated sample, features of about 2 nm height are distributed on the line profile. The features may correspond to the height of a single polymer chain stacked edgewise [137, 174]. 59 4.2.2 Annealing Samples The annealing process is often desirable to remove impurities and to achieve flat and homogeneous surfaces. In many cases, device performances are also improved upon annealing treatment. We measured 2PPE for the annealed polymer films. Figure 4.6 shows AFM images and the line profiles for P3HT films annealed at 100°C. For thinner films, the roughness is likely limited by the width (~2 nm) of a polymer chain. For thicker films, the aggregation between polymer chains may increase the roughness. We see in the line profile that the height variations are much larger than the width of a polymer chain. jum Figure 4.6: AFM images (2/umx2jum) for (a) 2.3 nm and (b) 7.7 nm thick P3HT films annealed at 100°C, and the height line profiles. 60 Annealing changes morphological structures which may influence the measured 2PPE efficiency and spectra. The surface topography did not change significantly upon further annealing at 150°C. The roughness estimated in 2 fim x 2 fjm with respect to thickness is plotted in Figure 4.7. The roughness is about 4 nm for films thicker than 10 nm. For an untreated -8 nm thick film shown in Figure 4.5(b) the roughness is -1.5 nm which is much lower than for annealed ones shown in Figure 4.6(b) with a roughness of about 3.5 nm. The increase of roughness can be attributed to dewetting effects, and it is not as pronounced as in dip-coated films. 4.5 4.0 -E c 3.5 -CO 0 c 3.0 -D) Ro 2.5 -CO 2.0 -1.5 100 Film thickness(nm) Figure 4.7: Surface roughness measured in 2/j.m x 2/j.m plotted as a function of film thickness (measured by XPS). The films were annealed at 100°C. Since the interaction between polythiophene and gold is weak [170], dewetting may easily occur [175]. Figure 4.8 shows dewetting pattern of a dip-coated P3HT film on gold substrate after annealing to 100 °C. A dewetted hole is surrounded by a rim. A line profile shows the height of the rim is about 40 nm, which is much thicker than the initial thickness of -5 nm measured by AFM tip-scratch test. It was observed that the height and diameter of dewetted holes depend on the film thickness and heat treatment condition. Such remarkable dewetted holes were only observed in dip-coated films. This 61 dewetting pattern in dip-coated films was also observed in unannealed samples after XPS measurements. We believe the main cause of the dewetting is heating by the Bremsstrahlung radiation [176] from the non-monochromated Al X-ray source. It is known that the temperature of a thin Ta foil with similar heat sinking is found to increase to 42°C while operating with a non-monochromated Mg source at 300 watts for 90 min [177]. However, for dip-coated films on a natively oxidized silicon wafer no significant dewetting patterns w ere o bserved. This indicates a stronger interaction between P3HT and the silicon surface. The rms roughness for the annealed films on the silicon wafer was -1.0 nm which is not much different from unannealed ones. nm 60.0 (c) Figure 4.8: a) AFM images of dip-coated P3HT after a 2PPE experiment and the height profile of a circular hole, (b) 3D-view of a crater, and (c) topview of the crater. 62 4.3 Polymer Thickness The thickness of ultrathin polymers has been measured by X-ray photoelectron spectroscopy. AFM tip-scratch method has also been used [178]. IR absorbance measurements have been used to show the correlation between the thicknesses measured by XPS and AFM methods. 4.3.1 AFM Method In AFM tip scratch method, the tip should not scratch the substrate. Depth is only accurate when the polymer is fully removed and the tip reaches the substrate. Figure 4.9: Peeling of polymer film by AFM tip. The scan (from A to F) was continued until all of the polymer films was removed. The scan range is 5 pm * 5 pm. The scratch is ~4 nm deep. Figure 4.9 shows the weakly a dsorbed dip-coated P3HT films on a gold surface. The tip for scratch tests was a silicon pyramidal tip with a radius of curvature of < 20 nm, 63 and the applied force was -100 nN. We found that the force causes no damage to the gold. Because the interaction of P3HT with silicon is stronger than that with gold, the polymer may adsorb on the silicon tip during the scratch scan. For spin-coated films, the same force was applied on the surface, and although there was a slight deformation, the spin-coated polymer films withstood the tip scratch, even when extended to long times. Because the scratch method was not efficient for thin spin-coated films, we used the XPS technique. Whether the polymer was fully removed was confirmed by using force-displacement measurement using the AFM tip. The force entirely depends on the interaction between the AFM tip and the surface. 100 c o > o 4— CD -a CD > _CD '43 C 03 o 50 0 -50 -100 -150 lit \ wJw~ „ V & (b); IB ,v ; , 1 .I.I. 0.0 0.2 0.4 0.6 0.8 Distance(piezo position,//™) Figure 4.10: Force curves of (a) P3HT and (b) Au contact: (i) no interaction region, (ii) jump-in-contact, (iii) contact region, (iv) retraction of the sample, and (v) jump-off contact. Figure 4.10 shows the different force-distance curves for contact with P3HT and Au. We used the same AFM tip for the scratch and the force test. Even with a blunt tip, the difference can easily be identified. In Figure 4.10, the sample on a piezoelectric transducer moves towards the AFM tip. When the tip is far from the sample, there is no interaction and no cantilever deflection. As the distance becomes smaller, the interactions between the tip and sample cause the cantilever to deflect downwards (then upwards). At 64 this stage, a sudden jump-in-contact occurs when the gradient of the interaction force exceeds the spring constant of the cantilever. When the sample moves still closer to the tip, short-range repulsive interaction occurs and the cantilever deflection will increase. The tip may also indent into the surface. After a maximum set deflection point (force), we then retract the tip from the sample. The sample and the tip are still in contact until the upward force of cantilever is greater than the attractive force between the tip and sample. The region is called jump-off-contact region [179]. The large pull-off force in curve (b) shown in Figure 4.10 is likely due to the H2O capillary force [179]. For P3HT surface, the force is likely due to adhesion or/and binding. Most force-distance tests on P3HT surfaces exhibited vertical jump-off force curves, not jagged force curve. When a polymer chain sticks on an AFM tip, the jagged force curve, shown in Figure 4.11(c), is obtained due to the binding between different polymer chains [179]. Adhesion Capillary Binding Force (a) (b) (c) Van der Waals Electrostatic (d) (f) Figure 4.11: Types of force-distance pull-off curves and their corresponding interactions [179]. 65 The AFM tip scratch and force-distance tests for dip-coated films were also performed after 2PPE experiment. Even without annealing, dewetting of the polymer film after the 2PPE measurement was observed. The dewetted hole depth, shown in Figure 4.12, is a good measurement for the film thickness confirmed by the scratch method. We scratched the hole through the polymer down to the metal substrate and measured the thickness. The depths of some holes before scratching shown in figure 4.12 and the depth after scratching shown in Figure 4.13 are fairly close. Therefore, some holes dewet completely and the gold surface is exposed. Figure 4.12: AFM image and its line profile for a dip-coated P3HT after 2PPE experiment. For thick samples, it is easy to measure the thickness accurately. A piece of scotch tape is used to peel the polymer film creating a sharp edge. An example for a -35 nm thick P3HT film is shown in Figure 4.14. 66 •0 0.5 1.0 1.5 2.0 fim Figure 4.13: AFM image of the scratched area and the line profile for the same film shown in Figure 4.7. 90-E 60 c 30 0: 0 10 /um 15 20 Figure 4.14: AFM image (20 urn x 20 um) showing a sharp edge and the line profile for a -35 nm thick P3HTfilm. 67 4.3.2 XPS Method and Inelastic Electron Mean Free Path In XPS, the attenuation of the substrate XPS signal with thickness can be used to measure the thickness of an overlayer, on condition that the thickness (d) is less than about 4 times the IMFP, where the Au intensity is about exp(-dVA-) ~2 % that of the bare substrate, and is difficult to accurately measure. The IMFP for organic compounds is given by Seah and Dench [180,181]. Xd(mglm2) = 49/E2 +0.1lVF 4.1 where E is the kinetic energy of the electrons, and Xd is inmg/m2. The IMFP expressed in nm using Xd in equation 4.2 is: kn (nm) = Xd I p 4.2 where p is the density in g/cm3 of the layer through which the electrons are traveling. We assume that the polymer films have uniform thickness. The relative XPS intensity with the polymer thickness can be derived using equation 4.3 [182]. 1ad _ ^ad \l g )_ 4 2 7 -dlA, cos# where Xs and A,a  are the IMFPs for the electrons from the substrate and overlayer, respectively , 6 is the emission angle with respect to surface normal, Iad and Is are the measured XPS intensities for the P3HT and Au electrons, respectively, and sad and ss are their XPS sensitivity factors. When the electrons from Au substrate travel through P3HT with kinetic energy of 1402.6 eV, the IMFP is calculated to be 3.65 nm assuming the polymer density of 1.1 g/cm3 [183]. The kinetic energy is the difference between 1486.6eV for Al K« x-ray source and the binding energy of 84.0 eV for Au4f. For P3HT, we use Cls XPS peak at a binding energy of 285 eV (kinetic energy of 1202.6 eV) and then X,ad = 3.21 nm. Given we observe relatively discrete height variations of 2 ~ 3 nm on the polymer surface, a more accurate formula might be similar to equation 4.7 in section 4.4. However, because the roughness of the films is smaller than the electron mean free path, the error due to the roughness should not be too significant. 68 4.3.3 FT-IR Method Thickness measurement using XPS is a time- and money-consuming method. Also, the XPS method doesn't work for d > 4X nm, where X is IMPF of Au electrons. Ideally the IR absorption peak height is proportional to the film thickness. If we obtain a good correlation curve between the IR signal and the thicknesses obtained by XPS and AFM, we can estimate the thickness more conveniently. The FT-IR spectra were obtained using a Bruker Equinox 55 spectrometer with a Hyperion IR microscope operated in reflection mode. We measure the intensity (Ro) for a bare gold substrate, and the intensities (R) for the polymer coated films. The IR absorbance (A) is then expressed as AR I R = \-RI R0. Figure 4.15 shows the FT-IR absorbance peak heights with a fitting curve, plotted against XPS and AFM thicknesses. Inset in the Figure is the AR/R showing the C-H stretching IR bands, and the peak height measurement at 2928 cm"1. 0.010 D) CD X 0.005 _^ CD CD CL LX 0.000 2928 2800 2900 3000 Wavenumber(cm') 3100 -O' # AFM thickness O XPS thickness J i I i i i i L • • ' 0 5 10 15 20 25 30 Film Thickness (nm) Figure 4.15: FT-IR intensities versus polymer thickness. Inset is the C-H stretching IR band with a linear background. 69 The thicknesses in this thesis are those measured by XPS and AFM. This is because for thinner films orientation of the polymer chains on the metal substrate may influence on the JR intensity via the surface selection rules [184]. In the early stage of P3HT adsorption, the lamella tends to parallel to the substrate partly due to aggregation of polymer chains [173]. As the coverage increases, it is likely that the lamellae become perpendicular to the substrate. 4.4 Calcium Coverage Auger electron spectroscopy (AES) was used to calibrate the calcium deposition flux for treated samples. For these measurements, calcium was deposited on a bare gold substrate at various crucible temperatures and deposition times, and Auger spectra were recorded afterwards. The degree of clustering of calcium on the Au is probably the main source of uncertainty in the Auger method. The growth mode (wetting behavior) of overlayer metals can be described by the spreading parameter S, which consists of surface and interface free energies [185]. The overlayer wets the substrate surface when S >0. S — Ysubstrate — (Yoverlayer Yinterface) 4.4 YAU=1-5 Jm"2[186] yCa = 0.5 J m"2 [186] Yca/Au = 0.8 J m"2 [187] When the overlayer partially wets the surface as shown in Figure 4.16, the relationship between the equilibrium contact angle (0) and the free energies (y) is expressed by Young's equation [185]. /I Ysubstrate ^interface ^ 1 A c cos 6' = -— = + 1 4.5 y overlayer ^ overlayer y overlayer ysubstrate interface Figure 4.16: Equilibrium contact angle (0) of overlayer on substrate. The Au/vacuum surface free energy is larger than the sum of Ca/vacuum and Ca/Au interface free energies. Therefore, the calcium overlayer likely wets the gold surface and 70 growth follows the layer-by-layer mode. In cases when the interface free energy changes after completion of the first overlayer, the growth mode may change. Layer-by-layer growth forms a smooth and continuous film, and Auger and XPS intensities can easily be analyzed in terms of thickness. < 1ML .continuous film 1-2ML clusters >2ML (a) (b) (c) Figure 4.17: Three important thin film growth modes at surfaces for different coverages, (a) Layer-by-Layer Growth (Frank-Van der Merwe, FM). (b) Layer plus Island Growth (Stranski-Krastinov.SK). (c) Island Growth (Vollmer-Weber.VW) [188]. Figure 4.17 shows three types of film growth modes: layer-by-layer, layer plus island and island growth modes [188]. If overlayer-overlayer interactions are strong, clusters are commonly observed. For continuous films, the ejected electrons from the substrate travel the same overlayer thickness. If clusters are formed, electrons travel different overlayer thicknesses; consequently the attenuation of the substrate electron intensity is dissimilar to that in the continuous films and should be modified. It is known that a sub-monolayer of Ca remains segregated at the surface and does not diffuse into the Au bulk [189]. If the coverage is less than 1 monolayer without cluster formation, the relative Auger signal intensity [190] can be expressed as equation 4.6. -i lad _ /Cad1 bulk V1 c ) A r 5 nuld(l-Zad) + Xad-e*<C°Se} 71 where Iad and Is is the Auger intensity for the adsorbate and the substrate, respectively. Xad is the fraction of substrate surface covered by adsorbate. The 9 (42.3°) is the acceptance angle of the CMA from the surface normal. Xad and A.s are the IMFP for the adsorbate and substrate Auger electron energies, respectively. The relative Auger signals, l£lk I Vbulk , for bulk substrate and adsorbate can be taken from sensitivity factors [191] shown in Table 4.1. It is assumed that the backscattering of electrons by the adsorbate is approximately equal to the backscattering by the substrate. For more than one monolayer, the expression for layer-by-layer growth model is: ex<™erbulk[(\-xad) + Xade^e} where xad IS not total amount of the adsorbate, just fraction covered by top layer, n is the # of layers including partial topmost layer. Au (74 eV) C (275 eV) Ca (294 eV) Sensitivity, Ibuik 2.1599 0.5453 0.8329 Table 4.1: Auger Sensitivity factors at 4 KeV electron beam. A general empirical form for calculating the Inelastic Mean Free Path (IMFP) is: A(nm) = ^T + Bny[KE 4.8 "v KE2 where Xn is IMFP in nm units, KE is the electron energy(eV), An and Bn are constants that depend on the material being considered. If the Au electrons with a KE of 74 eV travel through the Ca overlayer, the calculated Zn is 0.49 nm. For calcium with a KE of 295 eV, the IMFP is calculated to be 0.93 nm. We should keep in mind that the uncertainty of IMFP and the surface roughness could lead to deviations from the actual thickness. 72 Material Au , / Ca An 177 143 Bn 0.054 0.054 Table 4.2: Calculated photoelectron mean free paths using equation 4.8, An and Bn values were taken from the reference 180. 4-2-C/5 C CD 1 & 0 CD ^ O) -2 -4 carbon contaminated Au /—c KLL I ^ai_MM -Au MNN Bare gold Ca on gold * i'111111 • •1 1111 • • • i • • • • i i..., 100 200 300 400 500 Kinetic Energy(eV) Figure 4.18: Typical Auger spectra for bare gold and calcium (1.4 ML) on gold in the range of 50 ~ 550 eV. Auger spectrum for a carbon contaminated gold is also shown. Figure 4.18 shows typical Auger spectra for bare gold and calcium on gold. Auger spectrum for a carbon contaminated Au is also displayed for comparison. The Auger spectra were measured immediately after Ca deposition. We exclude the chances of calcium oxidation during the deposition because no oxygen peak at -510 eV was observed. The relative Auger signal was determined by the ratio of peak-to-peak heights 73 of the Ca signal at 294 eV and the Au signal at 74 eV. The deposition rate was controlled by changing the temperature. Figure 4.19 shows the relative Auger intensities (ICa{294) I IAu(1A)) for Ca/Au. The solid line represents our best fitting curves for layer-by-layer growth mode. Although this method is not a real time measurement, the repetition experiments showed that the flux rate was quite stable and reproducible during the metal deposition. An intensity ratio of 0.21 in Figure 4.19 corresponds to 1 ML thickness of calcium. < _< 0.1 fc-co of O 0.01 \r • r • i • 650 K / * O 670 K - • 685 K Layer-by-layer • 1 L , i . . . Growth • i 0 100 300 400 200 Time(sec) Figure 4.19: Experimental peak-to-peak intensity ratios ( 7Ca(294) / IAu{7A) ) with Ca deposition time at various oven temperatures including the best fitting curves using layer-by-layer growth mode. 74 Chapter 5 2PPE of Gold Substrate Two photon photoelectron spectra are recorded as a function of laser power and wavelength for gold substrates in the absence ofpolymer film. Work function and 2PPE quantum efficiency are measured including the effect of adsorption of chloroform, the solvent used in polymer film deposition. The surface state of Au(lll) is discussed. 5.1 Bare Gold Figure 5.1 shows the 2PPE spectrum for a clean bare gold sample, probed by 3.65 eV photons. 13 'to CD 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 E, -E, (eV) kin kin,maxv ' Figure 5.1: 2PPE spectrum for Au(111) taken with a photon energy of 3.65 eV . Inset is an expanded figure for the high energy cutoff. 75 The peak consists of two parts. The high energy region between the Fermi level and E = -2.0 eV is Au 6s electrons (s-p band). The lower energy region is likely dominated by Au 5d electrons [192,193]. According to UPS spectra of Au [192], the first d-band is placed between -2 eV and -4 eV, and the second d-band is located between -5 and -8 eV. However, the spectrum may be primarily due to the secondary electrons populated by scattering of photo-excited substrate electrons. Emission from the Au is weak at the Fermi level because there is a partial band gap [4] normal to the (111) crystal orientation which dominates the surface. The spectral shape is not affected much by the polarization of light. The low energy cutoff corresponds to the photoemission threshold at the vacuum level. The cutoff is shaper and easier to measure with low bias voltage. The inset shows the high energy cutoff region of emission from the bare Au substrate, obtained with more extended signal averaging, because of a weak signal in the region. The maximum kinetic energy electrons originate from the Fermi level at a fixed initial energy state. Therefore, the high energy cutoff shifts twice as much as each increment of photon energy (AKEmax=2- Ahco ). Figure 5.2 shows a plot of the kinetic energy width with the probe photon energy. The slope of two confirms that the photoemitted electrons originate from the initial occupied state by a two photon process. In Figure 5.1, the kinetic energy width is 2.2 eV probed by photon energy of 3.65 eV. Therefore, the work function O can be derived using equation 5.1. ® = 2htD-KE^ 5.1 5.1 eV = (2 x 3.65 eV) - 2.2 eV The work function was measured to be -5.1 eV for Au(lll). Our value is in good agreement with others. Cao et al. observed a work function of 5.1 eV for Au film (several hundred nm thick) on mica using 2PPE [193]. Sawatzky et al. also observed a work function of 5.1 ± 0.1 eV for Au film using UPS measurement [194]. Single crystal Au(lll) has a literature value of 5.31 eV [195]. The work function depends on surface morphology, and the lower value in our experiment than the single crystal can be attributed to surface roughness of films versus prepared single crystals. 76 The lowering of work function on rough metal surface can be understood in terms of the Smoluchowski effect [196], where smoothing of the charge density at steps and defects lowers the dipole associated with electron spill out. 3.0 3.2 3.4 3.6 3.8 4.0 Photon Energy (eV) Figure 5.2: High energy cutoff as a function of incident photon energy. 5.2 Quantum yield Figure 5.3 shows the total photoemitted charge per pulse as a function of laser power for two different gold substrates. In both cases, the slope of the log-log plot is 2. This quadratic power law behavior indicates a two photon process with the intermediate state population limited by only first-order kinetics. For substrate A which is -150 nm thick gold on mica, the photoemitted charge shows a deviation (slope > 2) from the quadratic law for laser intensities higher than ~ 0.1 MW/cm . However, substrate B which is -250 nm thick gold on glass does not show this deviation in this high intensity region. 77 10 13 substrate B slope =2 40 100 200 Laser intensity (KW/cm2) Figure 5.3: Photoemitted electron flux as a function of laser intensity for two different bare gold substrates, with an incident photon energy of 3.65 eV. Tomas et al. measured gold using p-polarized 2.33 eV photons with 25.5 ps pulse duration, and observed a deviation from the quadratic dependence above 300 MW/cm2 [197]. Steinmann observed space charging effects for Ag(l 11) above 50 MW/cm2 power density, without any indication of sample heating by 10 ns pulse laser [198]. Logothetis et al. observed an indication of heating of Au surface above 1 MW/cm2 during multi-photon experiments using 40 ns laser pulse [199]. There is no significant laser heating effect for bulk samples at our intensities [197,198]. But there can be a heating effect for thin films [199]. The thermal diffusion length (/ = ^2KT ) was estimated to be 2200 nm, where K is the thermal diffusivity of gold (1.18 cm2/s) [200], which is much larger than the Au film thickness. To check whether the power law slope larger than 2 is due to thermally assisted emission, we calculate a maximum temperature rise using AT = I-(l-R)-vl(Cv -d) [201], where I is the incident laser power, (1-R) (refer to Figure 5.4) is the absorbance of the pulse, r is the laser pulse duration, Cv (0.129 cm2 s" 78 2 K"1) is the heat capacity, and d is the thickness of the film. We estimate that the maximum temperature rise for ~200 nm thick gold film at 0.1 MW/cm2 input power is only about 26 K. This is in the limit that there is no heat transfer to the thick glass or mica substrates. This implies that the gold film showing a deviation from the quadratic is much thinner resulting from sputter cleaning. And, even though the temperature rise may not be big enough for thermoionic emission, the photoemission might be thermally assisted. The 2PPE yield in the quadratic regime at a photon energy of 3.65 eV was measured to be 6 x 10"35 cm2 s with s-polarized light. The yield for p-polarization was ~1 x 10"33 cm2 s. Our result is in reasonable agreement with the literature value of Bensoussan et al. They report a 2PPE yield of 4.9 x 10"33 cm2 s for an Au film (O = 4.8 eV) using 3.57 eV photons (p-polarized) at an incident angle of 30° from the surface normal [202]. Although, the absorbance at an incidence angle of 30° is slightly lower, a bit larger yield than our result maybe due to a contamination of the film they used, indicated by their low work function value. Also, the difference can be the result of our approximation of the laser pulse spatial and temporal profile. The absolute yield for p-polarization is about 18x larger than that for s-polarization. Cao et al. [193] observed that the 2PPE yield of Au(lll) film for p-polarized 3.2 eV photons at an incident angle of 45° with respect to the surface normal is about 6x higher than that for s-polarized light. For p-polarized light, the photoemission yield Yp, is the sum of bulk and the surface contributions [203]. Y =Y +Y 5 2 1p 1pjbulk T Ap,surf J-i-For s-polarized light, only emission from the bulk contributes to the yield. Y = Y 5 3 Neglecting any direct band transitions for these threshold photon energies, the bulk contribution to the 2PPE yield is proportional to the optical absorbtance, A=l-R. Y IY = A2 I A2 S4 Ip,bulk 1 1s,bulk ^p.bulk1 <Mk J' Figure 5.4 shows the absorbance of p- and s-polarized 3.65 eV photons calculated using Fresnel equations (refer to Appendix one) and the optical constants of Au. At an 79 incidence angle of 75° from the surface normal, the absorbance of p-polarized light is about 4x larger than that of s-polarized light. We estimate (Y I A2pbulk) l(Ys I A2bulk)«1.6 from the absorbance and the absolute yields of our data. So, the Yp surf « 0.6 • Yp bulk. According to Gao's results, (Yp I A2pbulk)l{Ys / A2bulk) * 2.7 , and the Ypsurf «\J-Ypbu!k . The magnitude of the surface effect may scale with the absorptivity of the sample, which is greater at 3.2 eV than at 3.65 eV of incident photons. Incidence Angle ($) Figure 5.4: Calculated absorbance of a bare gold for p- and s-polarized 3.65 eV photons with respect to incidence angle from the surface normal. 5.3 Surface State The crystal termination at the surface causes an intrinsic surface state localized in the projected sp-band gap of the Au(lll) surface. This wave function decays monotonically into the vacuum region. We assigned the peak located -0.35 eV below the Fermi level in our spectra to this state. The theoretical value reported is -0.5 eV below the Fermi level [204]. The value using normal photoemission at an incidence angle of 60° was reported 80 to be -0.4 eV below the Fermi level [205]. The energy position of the state is dependent on the take off angles. The surface state (SS) binding energy decreases with increasing take off angle from the surface normal [206]. At a high bias, our collection angle is wide. Figure 5.5: (a) Surface states for Au(111) taken at high (-24.0 V) and low (-10.0 V) bias, (b) IPS spectra for Au(111) at a photon energy of 11.0 eV as a function of incidence angle [206], SS: surface state, IS: n=1 image state. LaShell et al. reported that the sp-derived surface state is a doublet [207]. They have nearly the same intensities with a splitting of -0.1 eV. The energy and angular resolutions in our 2PPE system make it impossible to resolve the splitting. Temperature dependence of the surface states was studied by Paniago et al. [208]. The position and intensity of the state decrease with increasing surface temperature. The surface state was more clearly seen in the 2PPE spectra taken with higher photon energy. The surface state was not observed for unclean surfaces. The n=l image state binding energy for Au(l 11) is expected to be located at 0.8 eV below the vacuum level [4]. The image state binding energy obtained by inverse-81 photoemission studies is 0.6 eV [206]. The image state of gold in our 2PPE experiment using lower photon energies than 4.5 eV could not be observed, because the image state is located at ~ 4.5 eV above the Fermi level for <D = 5.1 eV. 5.4 Gold Substrate Dipped in Chloroform We also tested the gold substrate dipped in pure chloroform solvent, which was used for dissolving the polymer. Figure 5.6 shows the 2PPE spectra compared to the clean bare gold. The spectrum for the pure solvent dipped Au sample is a bit broader, but shows similar electron kinetic energy distributions. For polymer coated samples (in Chapter 6), the spectra exhibit much narrower features. | i I i i | f i i t i i | i t i i i i i i i I i i i i i i i i i I i i i i i i i i I 4 5 6 7 Final State Energy Above EF (eV) Figure 5.6: 2PPE spectra for a bare clean gold and a gold substrate dipped in blank chloroform solvent, with a photon energy of 3.65 eV. The work function was reduced by -0.45 eV. In many cases, adsorbates reduce the substrate's work function because the electron density tail out of the gold surface is modified by the adsorbate. We also observed as the surface contamination increased over 82 time in UHV, the work function of an initially clean gold was decreased. The contaminant coverage related to the surface work function can be determined by Auger spectroscopy. A small enhancement (about 50%) in 2PPE yield was observed for the dipped gold substrate. According to Fowler law [209], the photoelectron yield is Y = Fc(hco-<&)2, where Y is the quantum efficiency, Fc is the Fowler constant, hco is the photon energy, and O is the surface work function. The contribution from the equation is (2hco- 4.65)2/ (2 hco -5.1) ~ 1.45. The small enhancement is then attributed to the decrease in the surface work function. The surface state peak in this sample is not clearly seen. 83 Chapter 6 2PPE of Semiconducting Polymer Films 2PPE spectra and photoemission yields are measured for spin-coated regioregular poly(3-hexylthiophene)(P3HT) and calcium doped P3HT films on gold substrates. Kinetic modeling of the emission yield as a function of incident laser power and photon energy suggests that the intermediate state accessed during the 20 ns pulse is the negatively charged polaron. The kinetic energy distribution of emitted electrons is analyzed to determine the absolute binding energy of the polaronic levels in a Ca doped film, and the magnitude of the electron injection barrier at the gold/polymer interface as a function of film thickness. 6.1 Characterization of 2PPE of P3HT In this section we discuss the intermediate states accessed during the 2PPE process. Upon adsorption of organic molecules onto a metal substrate, the electron redistribution at the interface shifts the electronic energy levels [61]. Figure 6.1 shows the 2PPE spectra for a bare gold and a P3HT (2 nm) film on Au. In the spectra, we see the low and high energy cutoffs shift downwards. The maximum KE cutoff for Au corresponds to electrons excited to 2-hco = 7.3 eV above the Fermi level, and we report final state electron energies relative to this cutoff (assuming that the work function of the spectrometer does not change between experiments). The low energy cutoff of the kinetic energy distribution remains constant with changing photon energy, appropriate for the vacuum level threshold. The difference of the low energy cutoffs of the two samples corresponds to the work function change (AO). For the 2 nm thick film, the change in work function is 1.1 eV, which is a reasonable value for organic materials on gold substrates [61]. Since the work function of 4.0 eV for the 2 nm thick film is still larger than the photon energy of 3.65 eV in this 84 experiment the emission process still involves at least two photons. The decrease in work function will be discussed further later. l\Q bare gold < t "~ • ' ' • * 1 I * I i—i I,,.,i. —I-,,I,...I I 4 5 6 7 8 Final State Energy Above EF (eV) Figure 6.1: The 2PPE spectra for a bare gold and a 2 nm thick P3HT on gold, with a photon energy of 3.65 eV. Looking at the high energy cutoff for the 2 nm thick film, we see it extends only to about 5.3 eV above the Fermi level, compared to the total photon energy of 7.3 eV. This could indicate there are no initial-states available less than 2.0 eV below the Fermi level, but that would be equivalent to 6.05 eV below the vacuum level and the valence band edge of P3HT is certainly higher than that, with published values for the ionization potential ranging from 4.2 eV [126] to 5.2 eV [127]. What is not shown in Figure 6.1 is that the laser power is -17 times smaller for the P3HT sample to obtain the same amount of emitted charge as for the gold sample. The emitted charge from Au, which is quadratic function of laser power, would be -300 times smaller at this laser power. The increase is also due to a decrease of the surface work 85 function. According to the Fowler law, the enhancement contributed by the lowering of the work function w ould b e (2 hco-4.0)2/(2hco -5.1 )2~2.3 whenthe hco is 3.65 eV. This c ontribution i s s mall c ompared t o t he t otal e nhancement. T he m uch 1 arger 2 PPE yield compared to Au is expected, based on the relative lifetime of the intermediate states in the two materials. In conjugated polymers there are a variety of photoexcited states with lifetimes ranging up to the millisecond scale [12], whereas in a metal the excited electrons relax on a femtosecond time scale [193]. This is especially significant when using nanosecond scale laser pulses, and immediately has two beneficial consequences; 2PPE can be used to study the polymer/metal interface with negligible background signal from the bulk metal, and workable signal can be obtained with laser intensities low enough to avoid damage to the polymer. The absorbed photon density in these experiments is well below the (wavelength dependant) threshold of ~1020 cm'3 where irreversible modification of poly-alkylthiophene is observed [210]. Figure 6.2 shows the photoemitted electron flux as a function of incident photon flux at several different wavelengths, for a 30 nm thick P3HT film on Au. The 2PPE yield for the P3HT film increases with photon energy although the absorbance decreases from 2.5 eV to 3.65 eV. This is likely due to an increased concentration of long-lived polarons and/or triplet excitons as an intermediate state. Unlike the quadratic function of photon flux observed at each of these wavelengths in Au, for the P3HT samples, the dependence on laser flux covers a range which varies nearly from quadratic to linear. One possibility is that even though the work function is larger than any of these photon energies there may be a mix of one- and two-photon processes. However, Figure 6.3 shows that the kinetic energy distributions decrease only slightly with decreasing photon energy. If the emission using 3.1 eV photons was a one-photon process then the most of the electrons would be low energy threshold electrons. We argue that the behaviour of the power law of the photoemission yield is characteristic of 2PPE from conjugated polymers, where the intermediate state photoexcitations are prone to higher order relaxation kinetics and saturation bleaching effects at high densities. The emitted photoelectron flux can be written as a function of n, the excited intermediate state concentrations; Fel = Aa • Fph • n . The parameter A^, which is the quantum yield for emission per unit concentration of the intermediate state, includes an 86 photoexcitation cross-section and also the probability of the final-state electron propagating to and across the vacuum interface, as in the classic 3-step model of Spicer [211]. CD ID x c o o CD P 1.2 E o 1.0 0.8 0.6 0.4 8 0.2 o ^ 0.0 - A P3HT (30 nm) *l 3.35 eV 3.45 eV /A 3.56 eV , J 3.65 eVf id r J JA T 3.0 eV 3.1 eV a ~ it m AT ^ • 1 i i • t 1 i • 1 1 1 1 1 1 1 1 i i i i I i • i i 1 i i i i I i i i • 1 • ••• 1 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Incident Photon Flux (x1022ph/cm2s) Figure 6.2: Power law plots with photon energy for a 30 nm thick P3HT film on gold. The solid lines are fit curves discussed in text. The sub-quadratic yields observed in 2PPE would correspond to the intermediate state concentrations increasing sub-linearly with laser power over the 20 ns laser pulse. This could occur due to fast second order relaxation kinetics of the intermediate states, or due to depopulation of the ground state giving rise to a bleaching of the photoabsorption. In fact both of these phenomena are commonly observed in the photoexcitation of conjugated polymers [12]. For instance, Figure 6.4 shows the concentrations determined by Silva et al. from transient photoinduced absorption measurements for polarons and singlet excitons in polyfluorenes [212]. We see that the population density of photoexcited species become saturated at higher pump fluences. 87 13 c 3 O O c o o _Q) LU 3.10 eV 3.35 eV 3.65 eV •••••.•.•.•.4 4.0 4.5 5.0 5.5 Final State Energy Above Ep Figure 6.3: The 2PPE spectra for a 2 nm thick P3HT on gold with three different photon energies. The spectra are normalized to the same peak height. The arrow indicates the low energy cutoff independent of the photon energy. 10 19 6 o cr. 10 13 _ (a) singlets • „o • PIFTO O PIFTEH i i 100 1000 10 100 1000 Pump Flue nee (uJ/cm ) Pump Fluence (uJ/cm ) Figure 6.4: The population density of polaron and singlet excitons as a function of pump fluence for PIFTO and PIFTEH polymers [212]. We attempt to address then whether the intermediate state in our 2PPE experiments are the short lived singlet excitons created with high efficiency in polymeric 88 semiconductors, or more long-lived species such as charged polarons, or triplet excitons. Assuming that a linear 2PPE dependence represents saturation of the intermediate state, then the data of Figure 6.2 for 3.65 eV photons shows that this is obtained at incident photon fluences of less than 4X1013 /cm2. For a 30 nm thick film, the corresponding absorbed photon density calculated using the Fresnel equations is only 17 3 * 3X10 /cm . This places quite a low upper limit on the saturation density of the 2PPE intermediate state. The exciton species giving rise to sub-nanosecond transient absorption effects [89,213] in various conjugated polymers can be pumped up to densities on the order of ~1019 /cm3 (for reference, the monomer density in solid P3HT is 1X1021 /cm3 [214]). Silva et al. [212] report that polaron densities on the order of 1019 /cm can be obtained using sub-picosecond laser pulses, as shown in Figure 6.3. However charge separated electron and hole polarons would have a Langevin capture radius of q2/47isoskT = 22 nm [84] (The dielectric constant of polythiophenes is -2.5 [215]), corresponding in the limit of close packing to a density of only 1.6X1016 /cm3. We argue that field driven recombination within the Langevin capture radius is rapid on the time scale of our 20 nsec laser pulse (and of actual device operation), but slow enough to allow sub-nanosecond transients to higher concentration. The photoemitted charge vs. photon intensity can be described using a simple rate equation for the intermediate states. where / is the photon flux, Ba is the cross-section for excitation, n0 is the saturation concentration of excited states, and kx and k2 are the first and second order decay rate constants, respectively. This is a standard form of the rate equation describing excited state kinetics in these materials [12]. We have neglected a stimulated emission as it is not seen in RR-P3HT [85]. For 3.0 eV photons, since n is almost linear with laser flux, the bleaching and bimolecular terms can be neglected in Eq. 6.1, so we have dn = IBa (n0 - n) - kxn - k2n2 dt 6.1 6.2 89 For the 3.65 eV photons representing saturation of the intermediate states, the Q= Aa -I-n gives: ^3.65«o =y = 1-6x10~%ellph 6.3 Since the quantum yield Am will almost certainly not be smaller at 3.65 eV than at 3.0 eV, we have from Eqs. 6.2 and 6.3; B30 > A = k-f -A« > k . (3.4 x IO"24 / cm2 sec) 6.4 Fluorophore absorption cross-sections are typically on the order of 10"16 cm2 [89,213] and Ba would be even less than this if the excitation branches into other states besides that probed in the 2PPE experiment. Then must be smaller than 108 /sec. For singlet excitons in poly-alkyl thiophenes, kx is typically ~2xl09 /sec [13]. It seems likely then that the first-order decay in our 2PPE using 20 ns pulses must be smaller than that of the singlet exciton. Given these indications that the intermediate state is a free polaron, we attempt to fit the data of Fig. 6.2 using Eq. 6.1 with parameters appropriate to free charge carriers. This meant that kx is set to zero, and the saturation concentration is set to the Langevin radius packing density (n0= 1.6 x 1016 cm"3). The second order rate constant k2 is set equal to the bimolecular recombination rate constant k2 = AnDr, where r is the Langevin capture radius, D is the diffusion constant of the carriers. For D we use the Einstein relation D = fjkT I q , where p is the carrier mobility measured in conductivity experiments. A typical mobility value for spin-coated RR-P3HT is ~1 x lO"2 cm2/Vsec [153]. The fit is made assuming a constant Aa value, which by the 3.65 eV emission yield and the assumed saturation density must equal to 1.6 xiO"24 /cm3. We freely adjust the excitation cross-section Ba . The single adjustable parameter Ba nicely fits both the magnitude and curvature of the yields at each wavelength. We feel this is significant and it indicates this model is valid. 90 As a rough consistency check, we examine our extracted value for the emission quantum yield per unit concentration of intermediate-states, A3.65 = 1.7X10"24 cm3. The quantum yields measured in conventional one-photon photoemission from organic materials, using photon energies 1 eV above threshold, is typically on the order of ~10"3 -10"4 [216]. Given a ground-state density in the one-photon experiment of ~1021 /cm3, the two quantum yield values are in reasonable agreement. Figure 6.5 shows the fit values of Bm as a function of photon energy. Also shown is the effective 2PPE quantum yield calculated simply as the emitted charge divided by the square of the incident laser flux at an emission level of 5 x 1013 cm"2 s"1 (-3000 electrons/pulse). Both values strongly increase with photon energy even though the optical absorbance of P3HT measured by Dicker et al. [91] (inset in Figure 6.5) decreases monotonically from 3.0 to 3.65 eV. Some of the increase in yield with photon energy would be due to a larger electron escape probability as more energetic final-states are reached. This is especially important if there are intermediate-states far enough below the vacuum level that the final-states obtained at lower photon energies are actually sub threshold. However the onset of saturation in the polymer shows that the intermediate-state concentration has increased at shorter wavelengths even as the laser flux has decreased and higher-order relaxation and saturation processes have turned on, so certainly the cross-section for pumping into the intermediate-state has increased dramatically. Increase in photocurrent with increasing photon energy has been explained by an enhanced charge separation at higher excitations with only separated charge carriers contributing to the current [34]. However, Kraabel et al. observed that the transient photoconductivity in P3HT follows the absorption profile, and concluded that free carriers are directly formed by photogeneration, not by dissociation of excitons [77]. However, Dicker et al. found carrier density measured by microwave conductivity measurement increases from 2.5 eV to 3.5 eV photon energy [91]. Kraabel used P3HT thin films sandwiched between two electrodes, whereas Dicker used an "electrodeless" geometry. If excitons dissociate very quickly at the metal electrode, the photocurrent likely follows the absorption profile. Where photocurrents in organic materials are mediated by exciton diffusion to and dissociation at an electrode interface [84], our 91 action spectrum could be interpreted as due to increased photon penetration to the backside of the polymer near the metal decreasing its absorptivity. However the thin film is effectively quite transparent at all wavelengths used here, and also we find experimentally that the yield is very much quenched for thinner films where the photo-excitations are necessarily closer to the metal electrode. —i 1 i i i i i i i i ' • i • i 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 Photon Energy (eV) Figure 6.5: The 2PPE quantum yield (Q/l2) obtained at an emission level of -3000 electrons/pulse, and obtained from the fits for the 30 nm thick P3HT film. All values are normalized to the values at 3.0 eV. Inset is an absorption profile for P3HT film [91]. It is generally known that the probability of germinate exciton dissociation increases with photon energy [34]. This has been seen in photoconductivity experiments using UV-Vis light and in the decrease of exciton mediated luminescence. We argue then that the action spectrum in Fig. 6.2 supports the hypothesis that the intermediate-state is the charged polaron created by exciton dissociation. The cross-sections values, Ba, range from 2.9 x 10"16 cm2 at 3.0 eV to as large as 7.8 x 10"14 cm2 at 3.65 eV. It should be noted though that the saturation concentration 92 and the chromophore ground-state concentration are not necessarily the same [89]. Given a monomer density of lX1021/cm3 for P3HT [214], there are 4X104 monomer units within the 22 nm capture radius of each polaron. Therefore, the actual cross-section per ' ' ' '212 182 monomer for formation of the intermediate state varies from7X10"z' cmzto2XlO"10 cm . By comparison, given that the absorptivity of P3HT is a = 2X104 cm"1 at 3.0 eV and a = 1X104 cm"1 at 3.65 eV, the optical absorption cross-section per monomer varies from 17 2 17 2 2X10"" cm" to 1X10"" cm . Then the branching ratio for formation of the intermediate-state varies from 4X10"4 at 3.0 eV to 2X10"1 at 3.65 eV. 6.2 Analysis of P3HT Polaronic Energy Levels In order to investigate the energy levels of the intermediate photoexcitations, 2PPE spectra were measured at various photon energies. Figure 6.6 shows the electron kinetic energy distributions from a 2 nm thick P3HT film on Au at three different photon energies. The three spectra have about the same number of photoemitted electrons. A higher laser intensity was required for the lower incident photon energy to collect the same number of photoemitted electrons. A kinetic energy feature that increases in energy by an amount equal to the increment in photon energy (AEfa„ = l-Ahco) is characteristic of the 2PPE process with photoexcitation from a fixed intermediate state. We see that the high energy leading edge of the distribution corresponds to a fixed intermediate state energy. The maximum peak position in the kinetic energy distribution also does not change with wavelength, and is likely determined by the low energy cutoff superimposed over a sloping emission distribution. This is also possibly a characteristic of a final state located just above the vacuum level. Adsorbate-induced final state characteristics have been observed in CeHsS/Cutlll) [56], C6H6/Cu(lll) [43] andoligo(p-phenylene-ethynylene)thiol/Au [217]. The 2PPE emission peak is > 1 eV wide irrespective of the incident photon energy. The absorption spectrum of polarons measured in photo-induced absorption experiments suggest that there is a -0.55 eV wide band of polaronic states in RR-P3HT films [85], due to coupling between facing chains and also a distribution of effective conjugation 93 lengths. The 2PPE feature may be further broadened due to vibronic coupling to different vibrational excited states of the polymer after photo-detachment of the outgoing electrons. Additionally, the energy of a charged polaron depends on the polarization of surrounding medium. The relaxation of polarons at the vacuum interface could be approximately 0.2 eV less than that of polarons beneath the first layer [210]. And for the very thin film, different polarization at the metal interface is expected. Finally, in a photoelectron experiment there is inevitably some pile up of electrons towards lower energies due to inelastic losses during propagation of the final state outgoing electrons. 3 -4—' c o O c o -I—« o LU photon emission energy edge 3.10 eV 0.90 eV 3.35 eV 1.10eV 3.65 eV 1.30 eV t I t 0.0 0.5 1.0 1.5 Final State Energy Above E ^J vac Figure 6.6: Kinetic energy distributions for a 2 nm thick P3HT on gold with three different incident probe energies. The arrows indicate the shift in position of the high energy emission edge with different laser wavelengths. The intermediate state edge is located 2.35 eV below the vacuum level and also 1.65 eV above the Fermi level. We now want to compare this value to the electronic band structure of semiconducting polymer. The energy difference from the vacuum level to the 94 HOMO is the ionization potential which has been measured several times. Unfortunately, the published values for P3HT cover a wide range from 4.2 eV [126] to 5.2 eV [127]. We will use the value of 4.7 eV reported by Onoda et al. because it is near the median of reported values and because it is measured using a direct photoemission threshold technique [131]. The difference between the HOMO and LUMO edges is called a free particle band gap. This is very difficult to measure for these materials because the absorption edge is obscured by the high absorption cross section for exciton formation. The exciton peak is located at 2.0 eV. Zhokavets et al. [218] recently determined that the band gap is 2.67 eV for poly(3-octylthiophene), using the optical data fit to the density of states of a 1-D semiconductor. This is in good agreement with that inferred in the two-photon absorption measurements of Sakurai et al. [219]. Given the IP of 4.7 eV and band gap of 2.67 eV, the intermediate state is located 2.35 eV above the HOMO, and 0.35 eV below the LUMO. This implies the 2PPE electrons originate then from intermediate states which lie entirely in the band gap. This is opposite to our intuition based on fast 2PPE experiments with conventional semiconductors [220], where excited charge carriers are expected to relax and accumulate in the conduction band, but not in the forbidden band gap. In polymeric semiconductors, however, the charge carriers quickly become localized into polaronic species with a relaxed local bonding structure [12]. The difference in energy between the negative (positive) polarons and the edge of the LUMO (HOMO) band of the P3HT lattice is termed the polaronic relaxation energy [85]. According to Osterbacka et al. the localized polaron is 0.37 eV below the LUMO edge [13], in good agreement with our results. Sakurai and coworkers observed that the lowest energy singlet ('Bu) and triplet (3BU) exciton states of P30T are 2.0 eV and 1.55 eV above the HOMO, respectively [219]. For RR-P3HT, Jiang et al. [85] observed the' lowest energy singlet state at 2.0 eV. These energy levels are shown in Figure 6.7. We have already argued that singlet exciton states are too short-lived to contribute to the emission. Triplet exciton states, though, are long-lived. However they are too low in energy to be accessed by the sub 3.45 eV photons for which long lifetime intermediate states were inferred. And at higher photon energies we find a saturation density which is much lower than that expected for excitons. Finally, in 95 the magnetic resonance experiments on photoexcited RR-P3HT [85], there are no triplet excitons observed (although they are in RRa-P3HT). 1.3 eV High E cutoff 0 eV Ev hco = 3.65eV -2.05 eV -2.35 eV -3.0 eV -3.45 eV -4.7 eV • LUMO • singlet exciton triplet exciton . HOMO Figure 6.7: Electronic energy levels for a 2 nm thick P3HT film and possible photoemission processes with a photon energy of 3.65 eV. The previous discussion concerns only an upper edge. We do not see a low energy cut-off corresponding to the intermediate state since presumably the edge of the intermediate states lies more than 1 • hco below the vacuum level. Figure 6.8 is for a 2 nm thick P3HT film with an overlayer deposition of ~1 A of Ca. We can see the work function has been lowered to 3.2 eV. The low energy peak increases linearly with photon flux as shown in Figure 6.9. Also the high energy edge of the low energy peak is located I-hco above the Fermi level. Therefore, it represents a one-photon process, and is identical to the emission from calcium induced polaronic band-gap states located close to the Fermi level observed using ultra-violet photoelectron spectroscopy [221]. The 2PPE peakposition corresponds to an intermediate state energy at 2.85 eV below the vacuum level. The photo-induced absorption peak of k-doped RR-P3HT places the upper 96 polaron state at 1.75 eV above the HOMO [13], which is 2.95 below the vacuum level given a 4.7 eV ionization potential. This is in good agreement with our 2PPE result. Ca(1 A)/P3HT/Au 3.0 3.5 4.0 4.5 5.0 5.5 Final State Energy Above Ep (eV) Figure 6.8: The 2PPE spectra for a 2 nm thick P3HT film with ~1 A overlayer of deposited Ca, with a photon energy of 3.65 eV. A manifold of occupied band-gap states extending up to the Fermi level have been seen in UPS spectra of Rb doped poly(p-phenylenevinylene) [221] and Ca doped poly(9,9-dioctylfluorene).[125]. Our 2PPE feature is more than 1.5 eV broad. This broad range of unoccupied states down to the Fermi level agrees with that inferred from XPS shake-up features for Ca doped poly(9,9-dioctylfluorene) [125]. The kinetic energy distribution of the 2PPE peak was broadened, which may indicate a broadening of the band gap of the polymer resulting from the formation of new subgap states. If the increase of ~ 0.2 eV corresponds to a shift of LUMO, the total widening of the gap is estimated to be ~ 0.4 eV on condition that the HOMO level is 97 symmetrically shifted. It is in good agreement with a band gap widening of 0.4 ~ 0.5 eV for CIO4 -doped polythiophene [148]. In other words, the intermediate state is moved -0.2 eV towards the vacuum level and then the electron affinity is decreased. Osterbacka et al. also observed a broadening of -0.3 eV in the photoinduced polaron peak for la-doped RR-P3HT [13]. OlPPE#2PPE I lilt 1 —t... I 1 „ I 1 1 1 1 1 1 ' i i I • • i • I » • • i I 8.5 9.0 9.5 10.0 10.5 11.0 11.5 Kinetic Energy (eV) Figure 6.9: Photoemission spectra of calcium (1.1 A) on P3HT (2 nm) with laser power, with a probe photon energy of 3.65 eV, Inset is a plot of two peak heights with laser intensity. 6.3 Interfacial Effect on P3HT Energy Levels We have measured 2PPE spectra for ultra-thin films of P3HT on Au with varying polymer thicknesses to investigate the effect of the metal/polymer interface on the intermediate state energies. 98 Figure 6.10 shows the 2PPE spectra for spin-coated P3HT films of different average thickness (remembering that the films have roughness features of > 2 nm). The shift in low energy cutoff corresponds to a change in the local vacuum level, or work function, as the amount of P3HT on the Au substrate is increased. Interestingly the work function changes from 5.1 eV to 4.25 eV after just 0.7 nm average thickness, and has stabilized at 3.8 eV by 3.4 nm. In other words, the vacuum level shift abruptly and then stays constant after the 1st polymer layer is formed. This has commonly been observed in metal/organic interfaces. Seki have attributed this abrupt behavior to a dipole effect [61]. 3.5 4.0 4.5 5.0 5.5 6.0 Final State Energy Above EF (eV) Figure 6.10: 2PPE spectra for P3HT film with varying thickness, with a photon energy of 3.65 eV. 99 Alternatively, according to the band-bending model describing the interface between metal and semiconductors, a partial charge transfer occurs until the Fermi level aligns becoming constant throughout. There then exists a charged region at the interface. The change in the electrostatic potential is governed by the 1-D Poisson equation [63]; cfV I d2z = -p(z) I es0, where p(z) is the space charge density, E is the dielectric constant of the semiconductor, and e0 the dielectric constant of vacuum. In a conventional junction unless the Fermi level is close to the band edges, the charge concentration in this space charge region equals the dopant concentration. And if this is constant through the junction, the band bending described by the Poisson equation is quadratic. The curvature is larger for higher dopant concentrations. Regarding the behaviour observed for the P3HT/Au interface, we point out that while a doped semiconductor with a constant charge density depletion-region would exhibit smooth quadratic band-bending, that is not true if the Fermi level lies close to the polaron level at the interface. Then there is a high charge density and steep band-bending, but as this moves the Fermi level more than a few kT beyond the polaron edge into the gap then the charge density quickly decreases as shown in Figure 2.17, and the bending then becomes flat on the nanometer length scale probed in these experiments. Figure 6.11 shows the vacuum level corresponding to the low energy cutoffs, and also the intermediate state edges corresponding to the high energy cutoffs minus \-hco . For the 0.7 nm thick film, the high energy cutoff moves upward to 2.1 eV below the vacuum level compared to 2.65 eV for the 30 nm thick film. The binding energy of the intermediate state relative to the vacuum level is less for the thin films. These energy level shifts seen in the 2PPE experiment for the states above EF mirror those found by Hill et al. [69] for the states below EF in one-photon photoemission studies of metal/organic-semiconductor interface formation. 100 0.85 1-1 1.15 1.3 1.3 1.3 Vacuum Level (eV) 4.25 4.0 3.95 3.8 3.8 3.8 2.1 1.65 1.55 1.15 1.1 1.15 Intermediate state Edge (eV) Fermi level 0.7 2 2.5 3.4 7.7 30 Thickness(nm) Figure 6.11: The energy diagram for spin-coated P3HT/Au annealed at 100°C, with respect to thickness. All spectra were taken with s-polarized light (hco = 3.65eV). Values from the top to bottom are vacuum level shift, work function, intermediate state edge, and thickness of the films. In that study, the HOMO binding energy relative to the vacuum level was found to increase as the Alq3 films become thinner. Hill et al. argued there is less polarization stabilization of the positively charged hole when the molecules are surrounded by less material [69]. This is shown in Figure 6.12. The change in HOMO level from #2 to #6 is 0.6 eV. But polymer coils might look somewhat like #3 even at low coverage, so the total shift might be less for polymer films. We find a 0.5 eV change in binging energy. They could not observe the LUMO edge but predicted that occupied negatively charged states would have a decreased binding energy due to the same loss of polarization stabilization. Our 2PPE data confirms this behaviour, so there is an effective widening of the band-gap for sub-monolayer thick films. However, while the binding energy shifts of the 2PPE experiment are consistent with polarization effects we can still argue for the alternate mechanism of charge transfer to the polymer, since it has been argued that the HOMO-LUMO band gap will widen at a 101 charged interface due to rehybridization upon formation of a high density of polarons [63]. Figure 6.12: Changes in HOMO and LUMO levels depending on polymer thickness by Hill et al. (1) molecule in gas phase, (2) isolated molecule on a metal surface, (3) one complete molecular layer on a metal surface, (4) first molecular layer on metal surface, beneath a molecular film, (5) bulk molecular film, away from both the metal interface and the film surface, and (6) surface of a molecular film. Smith and coworkers have theoretically described band bending for metal-polymer interfaces [63]. They found that when a large charge density is transferred into the polymer layer, i.e. if the Fermi level of the metal contact is higher than the negative bipolaron formation energy per particle or lower than the positive bipolaron formation energy per particle, then the band gap of the polymer widens. It has been observed that the optical band gap of non-regioregular poly(3-methylthiophene) measured over a 6 nm ~ 1 pm thickness range is larger by ~0.2 eV for the thicker films [222]. This was attributed to a conjugation length increase due to better ordering in the thin films. Such an ordering effect may also contribute to our 2PPE results. 102 6.4 P3HT 2PPE Yield versus Thickness The 2PPE photoemission yields with film thickness are measured to further elucidate the influence of the metal substrate. Figure 6.13 shows the photoemitted electron flux as a function of incident photon flux at different polymer thicknesses. We use a log-log plot because it is useful to cover a wide range and to allow us to understand important features. We observe in the Figure that the 2PPE yield increases dramatically with increasing film thickness accompanied by a decreasing power law slope, and then the intermediate state finally saturates. 1E13 CM o x 13 c 1E12 o TJ CD LU 30nm 5nm 2.3nm 2nm 0.7nm bare gold 1E20 1E21 , i i—i—i i 11'' i— 1E22 1E23 Photon Flux (ph/cm2s) Figure 6.13: Log-log plots of photoemitted electron flux (electrons/cm2 sec) vs. incident laser flux (photons/cm2 sec), with 3.65 eV photons and different thick P3HT films. Lines are fit as described in text. 103 The kinetic modeling for this data is similar to that in section 6.1, but allows parameters to depend on distance of excitation from the metal surface. In particular, one expects that intermediate state species lifetimes can be shortened by proximity to the metal due to diffusion into the metal as a charge sink. The carrier mobility in spin-coated P3HT films can be higher than ~10~3 cm2 V"1 s"1 [153] which yields a diffusion constant D of-2.5 x 10"5 cm2/sec, so the carriers could diffuse as far as -40 nm during the 20 ns laser pulse. Also, it is known that coupling of the dipole of an electronic transition to a metal substrate can have a dramatic effect on the lifetimes of excited states near a metal [216]. Any modeling of 2PPE must take into that the amount of light absorbed in the polymer which depends on thickness. This can be calculated using the Fresnel equations (refer to Appendix one) [223]. Also, of course, there is more emitting material present in thicker film although emission from polymer buried beneath the vacuum/polymer interface will be attenuated by scattering as the excited electron propagates through the polymer. To account for this, photoemission has been described in terms of Spicer's classical three-step model: excitation, transport, and emission of electrons [211]. The emitted electron flux Q should depend on the instantaneous concentration of excited intermediate states as d Q= ^n-cr2-I2-e'zU dz 6.6 0 where d is the film thickness, X is the final state electron mean free path, a2 is the cross-section for coupling the excitation to vacuum emission via absorption of a second photon, and Iz is local laser intensity. . The intermediate concentration n will vary with depth if the excitation lifetime depends on the distance from the metal surface, and also if I varies through the depth of a film. We have started some finite element analyses taking into account the concentration profile that could develop through the film. However, in order to simplify analysis, for the sake of this thesis we will approximate a constant n through the film, and a constant I = I0-(l-R), where I0 is incident laser intensity, and R is thickness dependent reflectance. Then we integrate Eq. 6.6 to define an effective efficiency for pumping of intermediate 104 states Q/{[r(l-R)]2-(l-e~da)-A,]. The change in relative efficiency with thickness seen in Figure 6.14 shows the yield increase is not just due to more absorbed photons (l-R) or more emitting polymer(l-e'dU). The relative efficiencies with X = co and X~ 0 are plotted to put limits on the amount of increased emission due simply to the amount of emitting polymer. Note that these limiting values are independent of the surface roughness. The values based on a reasonable electron mean free path (1.5 nm for organic materials in this energy range [180]) are well bound by these limits. There is an increase in pumping efficiency or i ntermediate 1 ifetime with thickness in the first ~4 nm film. Above this thickness, the photoexcitations can be free from the metal influence, and consequently the efficiency does not increase remarkably with further increasing thickness. 10000 I o I 1000 o it LU D) C Q. E CL > 100 10 03 01 1 X = 0.01nm X - 1.5nm 0 5 10 15 20 25 30 Thickness(nm) Figure 6.14: Relative pumping efficiency per absorbed photon, Q/ {[l-(1-R)]2-(1-e'da)- 1} vs. the average film thickness with X = 0.01 nm (A), X = 1.5 nm (•), and X = oo (o), with a photon energy of 3.65 eV. The decrease of the power law slope with increasing thickness shown in Figure 6.13 can be understood by the kinetic models of section 6.1. When the excitation is decoupled 105 from the metal for thicker films, the intermediate concentration is increased dramatically, and the 2nd order and saturation effects occur resulting in a decrease in slope. We have attempted to fit the data with same n0 and k2 as the previous L angevin model, and introducing a new thickness dependent kx. However, we found it is necessary to also make n0 a function of thickness. We can see this in the data of Figure 6.13, where the power law slope for the 2.3 nm thick film is less than 1.5, indicating already some saturation, even though the yield and intermediate state concentration are much less than for thicker films. The Ba and Aa values for the 30 nm thick film are taken from the previous fit (refer to section 6.1), but depend on thickness as: B'm = Bm,0nm • (1 - R) I (1 - R)30nm, and Anw = A„,0nm • (1 - e"'") • (1 - R) /(l - R)30nm 6.7 The optical coefficients of P3HT are based on the data of Musa et al. [224]. The complex refractive index of P3HT is taken as h= 1.3 + 0.03-/ for an incident photon energy of 3.65 eV, and the absorption constant is ~10"3 nm"1. The laser intensity I is assumed to be constant through the film. The complex refractive index of gold at 3.65 eV photons is taken as n= 1.63 + 1.74-i. It is known that this value is unchanged for Au annealed in vacuum or aged in air [225]. Fits are shown in Figure 6.15. As seen in Figure 6.15, the kx obtained from the fit is larger closer to the metal surface, and also the saturation concentration decreases. The change in saturation concentration may indicate the intermediate state has a larger quenching radius near the metal. Alternatively, the intermediate species actually have a very non-uniform distribution in the films, with a "dead-zone" nearest the metal. Then the apparent saturation concentration averaged over the entire thin film is low because the emission is coming from just a small volume region furthest from the metal, with a high local concentration. Understanding the saturation and decay parameters is of importance for device operation. This is critical in determining the input and output efficiencies, as well as switching response time. This is also important in determining the operating conditions and the appropriateness of polymers for real devices. 106 1 10 Thickness (nm) Figure 6.15: Best-fit kx and n0 parameters for the 2PPE data with polymer thickness. 6.5 2PPE of MEH-PPV/Au We compare the 2PPE results of P3HT with the preliminary results of MEH-PPV. In addition to our studies of P3HT, we have taken a limited data set of 2PPE for MEH-PPV samples. As in the P3HT films, the 2PPE yield increases with increasing film thickness. While it is not shown here, the 2PPE yield also increases with decreasing laser wavelength, the same qualitative behavior as for P3HT. Similar wavelength dependence is seen in the published data for MEH-PPV photo-current devices, and was taken there as evidence [34] that the photoexcitations above 3 eV are more likely to form charged separated polarons. Previously, Hale et al. studied a 100 nm thick MEH-PPV film on Au using 2PPE [9]. The photon energy in Hale's experiments was 3.0 eV. The pulse width was -300 fs (repetition rate of 12 ns), compared to our 20 ns pulse width. Hale et al. obtains an emission of -1000 e- with an instantaneous photon flux of 3 xlO26 ph/cm2 s, much higher than ours. However, the time averaged intensity in those experiments was 6 xlO21 ph /cm2 s, which is very similar to the fluxes we use to obtain -1000 e- emission. 107 The results are consistent with ours if the intermediate lifetime is larger than the 12 ns repetition rate, so that the intermediate concentration n during the pulse is proportional to the averaged intensity rather than the instantaneous intensity. I • i i  i i 4E21 1E22 3E22 Photon Flux (ph/cm2s) Figure 6.16: Plots of photoemitted electron flux (electrons/cm2 sec) vs. incident laser flux (photons/cm2 sec) for 2 and 20 nm thick MEH-PPV films, with 3.65 eV photons. The solid lines are fit curves discussed in text. Inset is a power law plot for 100 nm thick MEH-PPV film by Hale ef al [9]. Figure 6.16 shows the power-law plots for 2 and 20 nm thick MEH-PPV films. We fit the data for the 20 nm thick film using the Langevin free carrier recombination model (refer to section 6.1 and 6.4). For MEH-PPV, we use p = 2.5 x 10"3 cm2/Vsec [226] and E = 3.0 [227], and then we have a saturation density no= 2.5 x 1016 cm"3 and a collisional recombination rate constant of 1.5 x 10"9 cmVsec. For the 20 nm thick film, the ki is still set to zero assuming long-lived intermediate states. The cross-sections value (Bm) at 3.65 eV photon energy obtained in this fit is 1.2 x 10"15 cm2 for the MEH-PPV film, which is -65 x smaller than for the 30 nm thick P3HT film. However, the absorptivity at 3.65 eV for MEH-PPV is 9 x 104 /cm [228], which is actually -9 x larger than that of P3HT. We 108 hypothesize that the lower cross-section for creation of intermediate state species is due to the much lower charge carrier mobility in MEH-PPV. This would make it harder for the high energy geminate excitons initially created to dissociate into separated polarons before losing enough energy that the charges remain trapped as a bound exciton. It is generally believed that in MEH-PPV, the excitation tends to remain localized as intra-chain excitons [12,229,230]. This gives a high luminescence cross-section "with radiative recombination competing against non-radiative recombination without effective dissociation. However, Rothberg et al. report that over 50 % of the photoexcitations (hco =2.4 eV) in MEH-PPV directly lead to long-lived interchain species [79]. Hale et al. argue that the intermediate state in 2PPE from MEH-PPV is actually the triplet exciton [9]. However w hile t riplet e xcitons h ave b een o bserved a 11 ow t emperatures i n M EH-PPV films, they are not seen at room temperature [231]. More details regarding the controversy over photogeneration species in MEH-PPV can be found in a recent review article [232]. The data is not fit well by our model, the curves shown in Figure 6.16 have a slope of 2, c ompared t o 1.5 f or o ur d ata. It m ay n ecessary t o h ave k2 much 1 arger t han t he Langevin recombination constant, and/or saturation at a much lower excited state/ground state ratio. The saturation or bimolecular annihilation in the 2PPE data of the MEH-PPV films may occur by other mechanism involving charge-separated interchain species (also referred to as "polaron pairs") [229,230]. Stimulated emission has not been considered because this is only observed in MEH-PPV solution or diluted in a solid matrix [232]. Even though the fit is not good, the variation in fitting parameter with thickness shows obvious metal proximity quenching. Figure 6.17 shows the 2PPE spectra for 2 and 20 nm thick MEH-PPV films. The change in work function with thickness is negligible, unlike for P3HT. Also, the change of -0.2 eV in the intermediate state edge with thickness is much less than for P3HT. It implies that there is less charge transfer at the interface. This is not surprising since the ionization potential of 5.3 eV [233,234] for MEH-PPV is larger than that of P3HT. Inset shown in Figure 6.17 is a 2PPE spectrum using 3.0 eV photons for a 100 nm thick MEH-PPV film by Hale et al. [9]. The high and low energy edges are difficult to determine in Hale's spectrum, which they attribute to detection of spurious low-energy electrons. We 109 note this is probably because they used a cylindrical mirror analyzer which requires electron entry at 42.3°, while the relatively large bias of 4 eV compared to threshold kinetic energies will accelerate electrons normal to the spectrometer. They do not determine the Fermi reference energy so we can not compare our data directly to theirs. 4.0 4.5 5.0 5.5 6.0 Final State Energy Above EF(eV) Figure 6.17: 2PPE spectra for 2 and 20 nm thick MEH-PPV films with a photon energy of 3.65 eV. Inset shows the 2PPE (hco = 3.0 eV) for a 100 nm thick MEH-PPV. Figure 6.18 shows the 2PPE spectra with three different photon energies. The high energy edge increases asl- Ahco, which corresponds to a fixed intermediate state energy as seen in the P3HT films (refer to section 6.2). Figure 6.19 depicts the energy diagram derived using the low and high energy cutoffs shown in Figure 6.17. According to this limited data set, the edge for thicker films is located at 2.5 eV below the vacuum level, or, given an IP of 5.3 eV [233,234], at 2.8 eV above the HOMO. Rikken et al. measured an electron barrier height of 1.2 eV above the Fermi level using internal photoemission for a polymer similar to MEH-PPV [235]. Our intermediate state edge for the 20 nm thick film is located 1.5 eV above the EF. Internal photoemission probes buried interface like case #4 of Hill (refer to Figure 6.12). There the gap should be narrower than the case #3 or #6, and a lower barrier found. 110 Ui c 0 3.65 eV 3.35 eV 3.10 eV o, t t t 4.0 4.5 5.0 Final State Energy Above Ep (eV) Figure 6.18: The 2PPE spectra for a few nm thick MEH-PPV on gold with three different photon energies. The spectra are normalized to the same peak height. 5.1 3.95 1.7 2 nm 4.0 1.5 Vacuum level Intermediate state edge Fermi level 20 nm Thickness (nm) Figure 6.19: Energy diagram for MEH-PPV/Au, with respect to thickness. Vacuum level is low E cutoff relative to EF. Intermediate state edge is high E cutoff relative to EF minus the photon energy hco = 3.65eV. ill 6.6 Annealing Effect We discuss the annealing effect on the polymer/metal interfaces. As a reminder, the films in sections from 6.1 to 6.5 were annealed at 100 °C. Figure 6.20 and 6.21 show the 2PPE spectra for a 30 nm thick P3HT and a 20 nm thick MEH-PPV film for a series of successive heat treatments, starting from the film without any annealing. Figure 6.22 and 6.23 show the corresponding energy diagram for the samples. 3^5 4^0 4^5 5^0 5^5 6^0 6^5 Final State Energy Above EF (eV) Figure 6.20: 2PPE spectra for a 30 nm thick P3HT film with respect to annealing temperature, with a photon energy of 3.65 eV. 170 UC 100 UC 3.5 4.0 4.5 5.0 . 5.5 6.0 6.5 Final State Energy Above EF (eV) Figure 6.21: 2PPE spectra for a 20 nm thick MEH-PPV film with respect to annealing temperature, with a photon energy of 3.65 eV. 112 5.1 2.6 2.15 2.65 0.7 2.65 1.45 1.15 0.9 2.6 1.05 Intermediate state edge EF 0.95 HOMO unannealed 50°C 100°C 150°C Figure 6.22: Energy diagram for P3HT/Au with respect to annealing temperature. Vacuum level is low E cutoff relative to EF. Intermediate state edge is high E cutoff relative to EF minus the photon energy hco = 3.65eV . The HOMO edge is derived assuming that the IP of P3HT is 4.7 eV. \ 5.1 2.0 2.45 2.25 2.55 2.5 Intermediate state edge 1.6 1.25 1.5 085 1.45 1.5 1.3 HOMO unannealed 50°C 100°C 150°C Figure 6.23: Energy diagram for MEH-PPV/Au with respect to annealing temperature. Vacuum level is low E cutoff relative to EF. Intermediate state edge is high E cutoff relative to EF minus the photon energy hco = 3.65eV . The HOMO edge is derived assuming that the IP of MEH-PPV is 5.3 eV. 113 Figure 6.24 shows the change in work function for the two polymers. Above 150 °C annealing, the slight increase of work function for the MEH-PPV film may be due to thermal degradation of the polymer [236]. Before annealing, the Fermi level is almost in line with the valence band or HOMO edge for P3HT. It is unlikely that the EF could be below the band edge throughout the thick film, as this would produce an extremely large voltage shift. The data indicates then that the IP of P3HT is probably a bit larger than 4.7 eV. For MEH-PPV, the EF is located at 0.85 eV above the HOMO. After annealing, the EF for both polymers is located almost in the middle of the band gap as expected in undoped semiconductors. It is well known that these polymers can be p-doped by absorbed oxygen [237,238]. The oxygen can be completely driven out above 100°C annealing [239]. Therefore, the change in work function is consistent with a transition from a p-doped state to an un-doped state upon oxygen degassing. Interestingly, the binding energy relative to the vacuum level of the high energy edge for both polymers is nearly constant with annealing temperature. 4.8 h 4.6 c o +=4 4 o c LL- 4.2 O § 4.0 3.8 0 50 P3HT (30 nm) MEH-PPV (20 nm) 100 150 Annealing Temperature(°C) 200 Figure 6.24: Work function with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV. 114 Yang et al. studied 10, 30, and 55 nm thick MEH-PPV films on Au using contact potential difference (CPD) measurements [227]. They spin coated from tetrahydrofuran (THF) or chlorobenzene (CB) solutions, and annealed the films at 5 5 °C in air. They observed a work function of 4.7 eV for all samples prepared from THF solution. Films prepared from CB solution showed about 0.15 eV higher work functions. These values are higher than our 2PPE results. It is not surprising since they treated samples in air. The XPS survey scan shown in Figure 6.25 for the untreated P3HT film showed that oxygen is present in the film. The Ols XPS peak is at 530 eV. For unannealed MEH-PPV films, impurity oxygen other than the oxygen in the MEH-PPV structure is not noticeable in XPS spectrum [236]. However, the amount of impurity required for the EF shift is tiny (ppm level in conventional semiconductors), so the impurity oxygen, compared to the intrinsic oxygen atom content in MEH-PPV may not be detectable in XPS. Au4f C1S Au5d A c ' ' 1 1 • I 1 I . . 1 I I 200 300 400 500 600 Binding Energy(eV) 100 Figure 6.25: XPS survey spectrum for an unannealed P3HT film. 115 The 2PPE yield of P3HT films drastically increases upon annealing over a temperature range up to 150 °C. Figure 6.26 shows the 2PPE efficiency Ql I2 for P3HT and MEH-PPV films using 3.65 eV photons. Although the lowering of the work function upon annealing contributes to an increase of the photoemission yield, this effect is not significant compared to the two orders of magnitude increase observed for P3HT. According to the Fowler law, the enhancement contributed by the lowering of the work function from 4.75 eV to 3.8 eV would be (2fico - 3.8)2/(2/ia> -4.75 )2 -1.9 when the hco is 3.65 eV [209]. 1E-29 1E-30 o CM JZ 3tlE-31 CN o 1E-32 50 100 150 Annealing Temperature(°C) Figure 6.26: Photoemission yield (Q/l2) with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV. The power law slopes for P3HT and MEH-PPV films with annealing temperature are shown in Figure 6.27. For MEH-PPV, both the yield and slope change only slightly with annealing. It does seem though that the higher order annihilation and/or saturation terms may be more pronounced upon annealing. The large decrease in slope for P3HT is consistent with the increased yield. Following our previous kinetic modeling, we write a 116 simple rate equation for the photoexcitations as dnldt = IBlo(no -n)-kTn-k2n2, where n is the concentration of photoexcitations, kT is the decay constant of photoexcitations corresponding to the annealing temperature T, and k2 is the second order decay constant [12]. 2.0 1.8 CD Q_ O cn 1.6 I 1.4 o CL 1.2 0 P3HT (30 nm) MEH-PPV (20 nm) 50 100 Annealing Temperature(°C) Figure 6.27: Power-law slopes with annealing temperature for P3HT (30 nm) and MEH-PPV (20 nm) films, with a photon energy of 3.65 eV. We can use the same«0, k2, and Ba from section 6.1, and fit the P3HT data shown in Figure 6.28 by varying kT. The kT varies from 7 x 1010 Is without annealing down to negligible values (< 107 Is) after 100 °C annealing. We could also fit the data by just varying Ba with kT = 0, in which case Bm varies from 8 x 10"17 cm2 without annealing to 1.5 x 10"14 cm2 after 100 °C annealing. So we can not distinguish if annealing is making it easier to create the intermediate state or is increasing their lifetime. 2PPE experiments with variable time delay between pump and probe pulse would be interesting here. 117 100°C 50°C o unannealed 1E13 \- J i i i i i L 1E21 1E22 Photon Flux (ph/cm2s) Figure 6.28: Power law plots with annealing temperature for a 30 nm thick P3HT film on gold, with a photon energy of 3.65 eV. The solid lines are fit curves, which can be derived changing either kT or Ba. It is known that oxygen quenches triplet excitons by the formation of singlet oxygen [240]. The pronounced oxygen effect in P3HT could indicate that the photoexcitations in P3HT are triplet excitons. However, we have argued based on the triplet energy level position given in the literature [219], and Osterbacka's ESR experiments [13] that the intermediate state species are not triplet excitons. Figure 6.29 shows the 2PPE spectra before and after annealing for P3HT and MEH-PPV films. Since the triplet state 2PPE emission peak is expected to be located at -0.4 eV above the vacuum level [219], if the dramatic yield increase is mainly due to a decrease in triplet quenching, we will see a much narrower 2PPE peak. However, in Figure 6.29 we see the same spectral shape for P3HT with annealing temperature. For the preliminary result of MEH-PPV, even though the yield does not much change with annealing temperature, we see the peak narrowing. We intend to investigate this effect in the near future by using lower photon energies which can not access the triplet state (if present). 118 Final State Energy Above E (eV) Final State Energy Above E (eV) Figure 6.29: 2PPE spectra for unannealed and 100 °C annealed P3HT and MEH-PPV films, with a photon energy of 3.65 eV. Beside oxygen triplet quenching, there are other mechanisms by which annealing could affect the yield. It is known that oxygen acts as a recombination center [241], which would limit polaron concentrations. Also, the polymer can be photo-degraded in presence of O2 [242]. Also, there is a conformational change of P3HT with annealing, which could affect the yield. We observed using AFM (refer to section 4.2) a change in the surface roughness of the P3HT films upon annealing, which indicates a change in conformation. Nakazono et al. observed that the crystallinity of poly(3-dodecylthiophene)(P3DT) increases with annealing up to -100 °C, and then decreases at higher temperature annealing [243], as shown in Figure 6.30. They found that the PL intensity varies inversely with degree of crystallinity. The change in crystallinity in Figure 6.30 looks similar to the change in 2PPE yield in Figure 6.26. We argue that 2PPE and PL have opposite dependence on mobility and crystallinity. Low mobility keeps the excitation localized, allowing more efficient recombination. High mobility allow for charge separation before recombination, making long lived polarons. 119 0 40 80 120 160 Heating Temperature (°C) 40 80 120 160 Heating Temperature (°C) Figure 6.30: Crystallinity and PL intensity with annealing temperature for P3DT measured by Yoshino et al. The 2PPE yield for P3HT films (also for MEH-PPV film) measured before the sample was cooled down to room temperature exhibited a lower quantum yield than when measured at room temperature. It is believed that the ring torsion angle and conformational defect increase with temperature because of the aliphatic side chains. The conformational defects shorten conjugation length resulting in a wider band gap (a blue-shift of absorption peak). This is often known as a thermochromism effect in polythiophenes [244]. The yield may decrease with higher temperature due to poorer crystallinity and mobility, lowering the dissociation probability, and thus Bm . Alternatively, at higher temperatures thermally assisted carrier de-trapping (increasing interchain hopping) from deep potential wells is thought to increase carrier mobility in P3HT in the range of 84 K < T < 416 K [21,245]. If the mobility increases then k2 also increases, and this would force a lower polaron concentration [74]. As mentioned previously regarding annealing, time-resolved pump-probe studies should be very useful to distinguish cross-section effects from lifetime effects on the intermediate state concentration. Figure 6.31 shows that the yield enhancement upon annealing is more significant for the thicker films. This is at least in part because the thin film suffers quenching by metal proximity even when annealed. Also, for the thinner films more oxygen could diffuse out without annealing, or the film might be more ordered before annealing. 120 104k CD C x: I— CO > a 1CV § 10' s_ 10 10L o o JO b • ooo o annealed • unannealed _j • ' • ' I i i i_| ' • • • I i i i i • • • • I i u 1 10 100 Thickness(nm) Figure 6.31: Relative 2PPE yield with respect to the thinnest film for annealed (100 °C) and unannealed P3HT films, with a photon energy of 3.65 eV. 6.7 2PPE Laser Intensity Dependence for Thick Films If we use higher laser intensities than were used to collect the 2PPE kinetic energy spectra discussed in the previous sections, we find that the high energy cutoff becomes dependent on the incident laser power for thick films with 3.65 eV photons. This is seen in Figure 6.32, which shows the 2PPE spectra probed by 3.65 eV photons for a 30 nm thick P3HT film with respect to laser power. There was no space charge effect since the low energy cutoff did not shift. The high energy cutoff increases with increasing laser intensity with a total shift of ~0.3 eV. Figure 6.33 shows the 2PPE spectra for a 2 nm thick P3HT film, where it can be seen there is no shift due to laser intensity. Also, for a photon energy of 3.1 eV, as shown in Figure 6.34, the high energy cutoff stays constant with respect to laser power even for the thick film. In common for both the thin films at all photon energies, and for the thick film using 3.1 eV probe photons, the concentration of the intermediate state is relatively low. We 121 attribute the shift then to a broadening of the intermediate state energies at high concentration due to the formation of bipolarons or a polaronic lattice. In a conference abstract (otherwise unpublished) Hale et al. report for 2PPE of poly(3-octylthiophene) that there is a peak which shifts to higher kinetic energy with increasing laser intensity which they also assign to polaron-polaron interactions (bipolaron formation) [246]. 0.0 0.5 1.0 1.5 Final State Energy Above Evac (eV) Figure 6.32: 2PPE spectra for a 30 nm thick P3HT film measured at different laser powers, with a photon energy of 3.65 eV. Photon flux is 1020 ~ 1021 ph/ cm2 s. Electron flux is 1012~ 1013el/ cm2 s. 122 __l I I I I I I • ' I I I I I I I I L 0.0 0.5 1.0 1.5 Final State Energy Above EF (eV) Figure 6.33: 2PPE spectra for a 2 nm thick P3HT film measured at different laser powers, with a photon energy of 3.65 eV. Photon flux is 1021 ~ 1022 ph/ cm2 s. Electron flux is 1012~ 1013el/cm2s. Figure 6.34: 2PPE spectra measured at different laser powers for a 20 nm thick MEH-PPV film, with a photon energy of 3.1 eV. Photon flux is 1021 ~ 1022 ph/ cm2 s. Electron flux is 1012~ 1013el/ cm2s.. 123 Chapter 7 Summary and Conclusion In summary, we have made the first ever study of two photon photoelectron emission from conjugated polymer that combines tunable photon energy and electron kinetic energy measurements. We report work function values, the absolute binding energy of intermediate state species, and saturation and lifetime effects. Ours is the first 2PPE study of interfacial effects, as we vary the thickness of ultra-thin polymer films on flat gold substrates. The energy and concentration of the excited states are critically affected by the metal substrates. We propose that 2PPE with 20 nsec pulses provides a direct probe of polaronic charge carrier states between the vacuum and Fermi levels. This is especially important for determining the charge injection barrier and carrier mobility at electrode/polymer interfaces. The morphology for gold substrates and polymer films was measured using AFM. Gold films are atomically smooth while polymer films have roughness features typically about > 2 nm, which may correspond to the edgewise stacking separation between individual chains. The surface topography depends on annealing temperature and on whether the films are spin- or dip-coated. The 2PPE spectra and photoemitted charge were measured with varying photon energies and incident laser powers. In the 2PPE spectra of bare Au, the KE distribution increases twice as much as each increment of photon energy (A¥LEmax=2- Ahco), and the photoemission yield is a quadratic function of laser power indicating second order two-photon processes. The work function for a clean gold was measured to be 5.1 eV. The 2PPE yield for a 3.65 eV s-polarized photon was ~5 x 10"35 cm2 s. For p-polarization, the yield was about an order of magnitude larger than that for s-polarization. A surface state for Au(l 11) is located at ~0.35 eV below the Fermi level. The photoelectron emission yields from the P3HT films are orders of magnitude larger than that of bare gold at the same laser intensity..The much larger 2PPE yield is 124 expected, based on the relative lifetime of the intermediate states in the two materials. This is especially significant when using nanosecond scale laser pulses, and immediately has two beneficial consequences; 2PPE can be used to study the polymer/metal interface with negligible background signal from the bulk metal, and workable signal. can be obtained with laser intensities low enough to avoid damage to the polymer. The P3HT 2PPE yield increases with photon energy although the absorbance decreases from 2.5 eV to 3.65 eV, much like photocurrent behaviour with photon energy. This is consistent with more efficient fission of excitons at higher energy to give long lived polarons. The dependence on laser flux covers a range which varies nearly from quadratic to linear. This is because the concentration of excited intermediate state in 2PPE increases sub-linearly with laser power at high concentration, and reaches a constant saturation giving a linear dependence. The kinetic modeling also shows that the lifetime and saturation concentration of the excited states of P3HT probed by nanosecond laser pulses are consistent with free polarons limited by Langevin capture and recombination. Upon adsorption of polymer, a decrease in work function was observed, changing abruptly from 5.1 eV to 4.25 eV after 0.7 nm average thickness, and then stabilized at 3.8 eV after 3.4 nm. A decrease of 0.55 eV in the minimum binding energy of the intermediate states, with decreasing the polymer thickness on a gold substrate, is consistent with the behavior of occupied states seen in one-photon experiments, confirming a widening of the band gap for sub-monolayer organic semiconductor on metal substrates. The change in polaron binding energy, and a sharp decrease of 1.3 eV in the vacuum level energy, is consistent either with equilibrated charge transfer across the metal/polymer interface or to changes in the interface dipole and in the polarization stabilization of charged species at the interface. In the 2PPE of P3HT films, the kinetic energy distribution of photoemitted electrons increases in energy by an amount equal to the increment in photon energy. This is characteristic of the 2PPE process with photoexcitation from a fixed intermediate state. The intermediate state edge for a 2 nm thick P3HT film is located 2.35 eV below the vacuum level. This is likely the first reported absolute binding energy of charged polaronic species in P3HT. If we assume that the IP of P3HT is 4.7 eV [131] and the 125 band gap is 2.67 eV [218], then the intermediate state edge is located at 2.35 eV above the HOMO level and 0.35 eV below the LUMO. Therefore the intermediate states are within the gap. The 20 ns laser pulse does not effectively allow us to probe the short lived transient states excited 3.65 eV above the HOMO. The 2PPE yield increases with increasing film thickness accompanied by a decreasing power law slope with respect to laser power. This can be understood by assuming the long-lived intermediate state excitation couples to the metal surface. The data at different polymer thicknesses is fit by varying the saturation concentration and the first-order rate constant, with a fixed second-order rate constant. The change in lifetime quenching within ~4 nm from the metal is attributed to diffusion or dipole coupling. The photoemission yield for thicker films is limited by saturation of photoexcitations. By doping with Ca, we could also see an emission from the occupied band gap states. For a 0.2 A of calcium doped P3HT film (<D =3.2 eV), the intermediate state 2PPE peak position corresponds to an energy at 2.85 eV below the vacuum level and 1.8 eV above the HOMO, in good agreement with literature value for (unoccupied) polaron level of 1.75 eV above the HOMO for I2-doped P3HT. The 2PPE yield for P3HT films was drastically increased upon annealing. This correlates with known crystallinity and mobility behaviors reported in literature. We attribute this to easier fission to make charged polarons. As the thickness is increased, the high energy cutoff in the 2PPE spectra recorded using higher photon energy became dependent on incident laser power. We have attributed this to a bipolaron broadening effect. For future research, it would be a great idea to employ this technique in investigating other well studied conjugated polymer systems. We also intend to install a 2nd laser to study the intermediate states using time-resolved experiments with a variable pump-probe delay. 126 Bibliography 1. C. K. Chiang, C. R. Fincher, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis, S. C. Gau, and A. G. MacDiarmid, Phys. Rev. Lett. 39, 1098 (1977). 2. J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H. Friend, P. L. Burns, and A. B. Holmes, Nature (London) 347, 539 (1990). 3. Handbook of Conducting Polymers, 2nd edition, T. A. Skotheim, R. L. Elsenbaumer, and J. R. Reynolds, eds., Mercel Dekker, New York, 1998. 4. Th. Fauster and W. Steinmann, Chap 8 In Photonic Probes of Surfaces, ed. P Halevi. Amsterdam: Elsevier (1995). 5. X. Y. Zhu, Annu. Rev. Phys. Chem. 53, 221 (2002). 6. C. B. Harris, N.-H. Ge, R. L. Lingle, Jr., J. D. McNeill, and C. M. Wong, Annu. Rev. Phys. Chem. 48, 711 (1997). 7. U. Hdfer, LL. Shumay, C. ReuB, U. Thomann, W.Wallauer, and T. Fauster, Science 277, 1480 (1997). 8. R. Haight, Surf Sci. Rep. 21, 275 (1995). 9. G. D. Hale, S. J. Oldenburg, and N. J. Halas, Phys. Rev. B 55, R16069 (1997); G. D. Hale, S. J. Oldenburg, andN. J. Halas, Appl. Phys. Lett. 71 , 1483 (1997). 10. W. Zhao, W. Wei, J. Lozano, and J. M. White, Chem. Mater. 16, 750 (2004). 11. M. Probst and R. Hight, Appl. Phys. Lett. 71, 202 (1997); K. Read, H. S. Karlsson, M. M. Murnane, H. C. Kapteyn, and R. Haight, J. Appl. Phys. 90, 294 (2001); H. S. Karlsson, K. Read, and R. Haight, J. Vac. Sci. Technol. A 20, 762 (2002). 12. Primary Photoexcitations in Conjugated Polymers: Molecular Exciton versus Semiconductor Band Model, edited by N. S. Sariciftci, World Scientific Publ., Singapore (1997). 13. R. Osterbacka, C. P. An, X. M. Jiang, and Z. V. Vardeny, Science 287, 839 (2000); R. Osterbacka, C. P. An, X. M. Jiang, and Z.V. Vardeny, Synth. Met. 116, 317 (2001). 14. A. R. Brown, A. Pomp, C. M. Hart, and D. M. de Leeuw, Science 270, 972 (1995). 15. F. Gamier, R. Hajlaoui, A. Yassar, and P. Srivastava, Science 265, 1684 (1994). 16. R. F. Service, Science 273, 879 (1996). 17. A. J. Heeger, Solid State Commun. 107, 673 (1998). 127 18. F. Hide, M. A. Diaz-Garcia, B. J. Schwartz, M. R. Andersson, Q. Pei, A. J. Heeger, Science 27, 1833 (1996); M. Berggren, A. Dodabalapur, R.E. Slusher and Z. Bao, Nature 389, 466(1997). 19. G. Yu, J. Gao, J. Hummelen, F. Wudl, and A. J. Heeger, Science 270, 1789 (1995); W. U. Huynh, J. J. Dittmer, and A. P. Alivisatos, Science 295, 2425 (2002). 20. L. Chen, D. W. McBranch, H. L. Wang, R. Helgeson, F. Wudl, and D. Whitten, Proc. Nat. Acad. Sci. USA 96, 12287 (1999); P. S. Heeger, and A. J. Heeger, Proc. Natl. Acad. Sci. U. S. A. 96, 12219 (1999). 21. H. Sirringhaus, N. Tessler, R. H. Friend, Science 280, 1741 (1998). H. Sirringhaus, T. Kawase, R. H. Friend, T. Shimoda, M. Inbasekaran, W. Wu, and E. P. Woo, Science 290, 2123 (2000); H. Sirringhaus, et al., Nature 401, 685 (1999). 22. E. W. H. Jager, E. Smela, O. Inganas, Science 290, 1540 (2000). 23. P. K. Ho, D. S. Thomas, R. H. Friend, and N. Tessler, Science 285, 233 (1999). 24. C. Drury, C. Mutsaers, C. Hart, M. Matters, and D. de Leeus, Appl. Phys. Lett. 73, 108 (1998). 25. H. Ma, A. K. Y. Jen, and L. R. Dalton, Adv. Mater. 14, 1339 (2002). 26. J. Joo and A. J. Epstein, Appl. Phys. Lett. 65, 2278 (1994). 27. W. E. Howard, Scientific American, February (2004). 28. Y. Furukawa, J. Phys. Chem. 100, 15644 (1996). 29. P. K. H. Ho, J. S. Kim, J. H. Burroughes, H. Becker, S. F. Y. Li, T. M. Brown, F. Cacialli, and R. H. Friend, Nature 404, 481 (2000). 30. Y. Cao, I. D. Parker, G. Yu, C. Zhang and A. J. Heeger, Nature 397, 414 (1999). 31. R. H. Friend et al, "Electroluminescence in conjugated polymers" Nature 397, 121 (1999). 32. M. Wohlgenannt et al, Phys. Rev. Lett. 88, 197401(2002). 33. C. J. Brabec, N. S. Sariciftci, and J. C. Hummelen, Adv. Funct. Mater. 11,15 (2001). 34. A. Kohler, D. A. dos Santos, D. Beljonne, Z. Shuai, J. L. Bradas, A. B. Holmes, A. Kraus, K. Mullen, and R. H. Friend, Nature 392, 903 (1998). 35. L. Smilowitz, N. S. Sariciftci, R. Wu, C. Gettinger, A. J. Heeger, and F. Wudl, Phys. Rev. B 47, 13835 (1993). 36. M. Granstrdm, K. Petritsch, A. Arias, A. Lux, M. Andersson and R. H. Friend, 128 Nature 395, 257(1998). 37. C. D. Dimitrakopoulos and D. J. Mascaro, IBM J. RES. & DEV. 45, 12 (2001); Christos D. Dimitrakopoulos, and Patrick R. L. Malenfant, Adv. Mater. 14, 99 (2002). 38. H. E. Katz, and Z. Bao, J. Phys. Chem. B 104, 671 (2000). 39. W. Steinmann, Appl. Phys. A 49, 365 (1989). 40. W. Steinmann, Phys. Stat. Sol. 192, 339 (1995). 41. E. Knoesel, A. Hotzel, T. Hertel, M. Wolf, and G. Ertl, Surf. Sci. 368, 76 (1996). 42. D. Straub and F. J. Himpsel, Phys. Rev B 33, 2256 (1986). 43. D. Velic, A. Hotzel, M. Wolf, and G. Ertl, J. Chem. Phys. 105, 9155 (1998). 44. J. D. McNeil, R. L. Lingle, Jr., N.-H. Ge, C. M. Wong, R. E. Jordan and C. B. Harris, Phys. Rev. Lett. 79, 4645 (1997). 45. A. Hotzel, K. Ishioka, M. Wolf and G. Ertl. Appl. Phys. B 68, 615 (1999). 46. D. McNeill, R.L. Lingle, Jr., R.E. Jordan, D.F. Padowitz and C.B. Harris. J. Chem. Phys 105, 3883 (1996). 47. N. H. Ge, C. M. Wong, R. L. Lingle Jr., J. D. McNeill, K.J. Gaffney, and C.B. Harris, Science 279, 202 (1998). 48. D. Reiger, T. Wegehaupt, and W. Steinmann, Phys. Rev. Lett. 58, 1135 (1987). 49. M. Wolf, A. Hotzel, E. Knoesel, and D. Velic, Phys. Rev. B 59, 5926 (1999). 50. T Hertel, K. Knoesel, E. Hasselbrink, M. Wolf, and G. Ertl, Surf. Sci. 317, LI 147 (1994). 51. H. Petek and S. Ogawa, Annu. Rev. Phys. Chem. 53, 507 (2002). 52. H. Wang, G. Dutton, and X. Y. Zhu J. Phys. Chem. B 104, 10332 (2000). 53. T. Vondrak and X. Y. Zhu. J. Phys. Chem. 103, 3449 (1999); G. Dutton and X. Y. Zhu. J. Phys. Chem. 105, 10912 (2001). 54. X. Y. Zhu, T. Vondrak, H. Wang, C. Gahl, K. Ishioka, and M. Wolf, Surf. Sci. 451, 244 (2000). 55. K. Ishioka, C. Gahl and M. Wolf. Surf. Sci. 73-77, 454 (2000). 56. T. Vondrak, H. Wang, P. Winget, C. J. Cramer, and X. Y. Zhu, J. Am. Chem. Soc. 122, 4700 (2000). 57. T. Vondrak, C. J. Cramer, and X. Y. Zhu, J. Phys. Chem. B 103, 8915 (1999). 129 58. Samokhvalov, and R. Naaman, J. Phys. Chem. B 104, 11248 (2000). 59. H. Petek, M. J. Weida, H. Nagano and S. Ogawa, Science 26, 1402 (2000); H. Petek, A.P. Heberle, W. Nessler, H. Nagano, S. Kubota, S. Matsunami, N. Moriya, and S. Ogawa, Phys. Rev. Lett. 79, 4649 (1997). 60. M. Aeschlimann, M. Bauer, S. Pawlik, W. Weber, R. Burgermeister, D. Oberli, and H.C. Siegmann, Phys. Rev. Lett. 19, 5158 (1997). 61. H. Ishii, K. Kiyoshi, E. Ito, and K. Seki, Adv. Mater. 11, 605 (1999). 62. X. Crispin, V. Geskin, A. Crispin, J. Cornil, R. Lazzaroni, W. R. Salaneck, J. -L Bredas, J. Am. Chem. Soc. 124, 8131 (2002). 63. P. S. Davids, A. Saxena, and D. L. Smith, Phys. Rev. B 53, 4823 (1996); G. Paach, P. H. Nguyen, and S. L. Drechsler, Synth. Met. 111-112, 321 (2000); N. Kirova, and S. Brazovskii, Synth. Met. 76, 229 (1996). 64. I. G. Hill, A. Rajagopal, and A. Kahn, J. Appl. Phys. 73, 662 (1998); I. G. Hill, A. Rajagopal, and A. Kahn, J. Appl. Phys. 84, 3236 (1998). 65. L. Yan, N. J. Watkins, S. Zorba, and Y. Gao and C. W. Tang, Appl. Phys. Lett. 81, 2752 (2002). 66. N. J. Watkins, L. Yan, and Y. Gao, Appl. Phys. Lett. 80, 4384 (2002). 67. N. Koch, J. Ghijsen, A. Elschner, R. L. Johnson, J.-J. Pireaux, J. Schwartz, and A. Kahn, Appl. Phys. Lett. 82, 70 (2003); N. Koch, A. Elschner, J. Schwartz, and A. Kahn, Appl. Phys. Lett. 82, 2281 (2003). 68. R. Schlaf, C. D. Merritt, L. A. Crisarulli, and Z. H. Kafafi, J. Appl. Phys. 86, 5678 (1999). 69. I. G. Hill, A. J. Makinen, and Z. H. Kafafi, J. Appl. Phys. 88, 889 (2000); I. G. Hill, A. J. Makinen, and Z. H. Kafafi,, Appl. Phys. Lett. 11, 1825 (2000); I. G. Hill, A. Kahn, Z. G. Soos, and R. A. Pascal, Chem. Phys. Lett. 327, 181 (2000). 70. G. Paasch, H. Peisert, M. Knupfer, J. Fink, and S. Scheinert, J. Appl. Phys. 93, 6084 (2003). 71. C. Botta, S. Luzzati, R. Tubino and A. Borghesi, Phys. Rev. B. 46, 13008 (1992). 72. K. Honda, Furukawa, K. Furuya, H. Torii, and M. Tasumi, J. Phys. Chem. A 106, 3587 (2002). 130 73. E. Ettedgui, H. Razafitrimo, K. T. Park, Y. Gao, B. R. Hsieh, W. A. Feld and M. W. Ruckman, Phys. Rev. Lett. 76, 299 (1996). 74. M. Westerling, R. Osterbacka, and H. Stubb, Phys. Rev. B 66, 165220 (2002). 75. B. Xu arid S. Holdcroft, J. Am. Chem. Soc. 115, 8441 (1993). 76. J. P. Yang, W. Paa, and S. Rentsch, Chem. Phys. Lett. 320, 665 (2000); D. Beljonne, Z. Shuai, G. Pourtois, and J. L. Bredas, J. Phys. Chem. A 105, 3899 (2001). 77. B. Kraabel, D. Moses, and A. J. Heeger, J. Chem. Phys. 103, 5102 (1995). 78. J. S. Wilson, A. S. Dhoot, A. J. A. B. Seeley, M. S. Khan, A. Kohler and R. H. Friend, Nature 413, 828 (2001). 79. M. Yan, L. J. Rothberg, E. W. Kwock and T. M. Miller, Phys. Rev. Lett. 75, 1992 (1995); L. J. Rothberg, M. Yan, F. Papadimitrakopoulos, M. E. Galvin, E.W. Kwock and T.M. Miller, Synth. Met. 80, 41 (1996). 80. A. Ruseckas, M. Theander, M. R. Andersson, M. Svensson, M. Prato, O. Inganas and V. Sundstrdm, Chem. Phys. Lett. 322, 136 (2000). 81. I. D. Samuel, G. Rumbles, and C. J. Collison, Phys. Rev. B 52, R11573 (1995); I. D. W. Samuel, G. Rumbles, C. J. Collison, R. H. Friend, S. C. Moratti and A. B. Holmes, Synth. Met. 84, 497 (1997). 82. E. M. Conwell, J. Perlstein, and S. Shaik, Phys. Rev. B 54, R2308 (1996). 83. T. Pauck et al, Chem. Phys. Lett. 244, 171 (1995). 84. S. Barth, M. Deussen and H. Bassler, Phil. Trans. R. SocLond. A 355, 749 (1997); S. Tasch, G. Kranzelbinder, G. Leising, and U. Scherf, Phys. Rev. B 55, 5079 (1996). 85. X. M. Jiang, et al.,Adv. Funct. Mater. 12, 587 (2002). 86. R. Hidayat, A. Fujii, M. Ozaki, and K. Yoshino, Jpn. J. Appl. Phys., Part 1 40, 7103 (2001). 87. K. Pichler, D. A. Halliday, D. D. C. Bradley, P. L. Burn, R. H. Friend, A. B. Holmes, J. Phys: Condens. Matter 5, 7155 (1993). 88. G. R. Hayes, I. D. W. Samuel, and R. T. Phillips, Phys. Rev. B 52, RI 1569 (1995). 89. S. H. Lim, T. G. Bjorklund, and C. J. Bardeen, J. Chem. Phys. 118, 4297 (2003). 90. G. Lanzani, M. Nisoli, S. De Silvestri, and F. Abbate, Chem. Phys. Lett. 264, 667 (1997). 131 91. G. Dicker, B. Wegewijs, J. Piris, T. J. Savenije, B. H. Huisman, D. M. de Leeuw, M. P. de Haas and J. M. Warman, Synth. Met. 121, 1451 (2001); B. Wegewijs, J. Piris, G. Dicker, T. J. Savenije, M. P. de Haas and J. M. Warman, Synth. Met. 121, 1357 (2001). 92. N. T. Binh, M. Gailberger, H. Bassler, Synth. Met. 47, 77 (1992). 93. N. C. Greenham et ai, Phys. Rev. B 53, 13528 (1996). 94. M. I. Khan, G. C. Bazan, and Z. D. Popovic, Chem. Phys. Lett. 298 (4-6), 309 (1998); M. Esteghamatian, Z. D. Popovic and G. Xu, J. Chem. Phys. 100 ,13716 (1996). 95. R. Kersting, B. Mollay, M. Rusch, J. Wenisch, G. Leising, and H. Kauffmann, J. Chem. Phys. 106, 2850 (1997). 96. S. C. Yang et ai, SPIE Proceedings 3797, 26 (1999). 97. D. Moses, J. Wang, A. J. Heeger, N. Kirova and S. Brazovski, PNAS 93, 13496 (2001); B. Schweitzer, H. Bassler, Synth. Met. 109, 1 (2000); see also more references therein. 98. S. F. Alvarado, P. F. Seidler, D. G. Lidzey, and D. D. C. Bradley, Phys. Rev. Lett. 81, 1082 (1998). 99. E. M. Conwell, Synth. Met. 83, 101 (1996). 100. G. Greczynski, M. Fahlman, and W. R. Salaneck, J. Chem. Phys. 113, 2407 (2000); G. Greczynski, M. Fahlman, W. R. Salaneck, N. Johansson, D. A. d. Santos, A. Dkhissi, and J. L. Bredas, J. Chem. Phys. 116, 1700 (2002). 101. M. Fahlman, P. Broms, D. A. dos Santos, S. C. Moratti, N. Johansson, K. Xing, R. H. Friend, A. B. Holmes, J. L. Bredas, and W. R. Salaneck, J. Chem. Phys. 102, 8167 (1995). 102. M. Fahlman, D. Beljonne, M. Logdlund, R. H. Friend, A. B. Holmes, J. L. Bredas, and W. R. Salaneck, Chem. Phys. Lett. 214, 327 (1993). 103. N. Koch, G. Leising, L. M. Yu, A. Rajagopal, J. J. Pireaux, and R. L. Johnson, J. Vac. Sci. Techno!. A 18, 295 (2000). 104. N. Koch, A. Rajagopal, J. Ghijsen, R. L. Johnson, G. Leising, and J. J. Pireaux, J. Phys. Chem. B 104, 1434 (2000). 105. P. Dannetun, M. Logdlund, C. Fredriksson, R. Lazzaroni, C. Fauquet, S. Stafstrdm, C. W. Spangler, J. L. Bredas, and W. R. Salaneck, J. Chem. Phys. 100, 6765 (1994). 106. G. Iucci, K. Xing, M. Logdlund, M. Fahlman, W. R. Salaneck, Chem. Phys. Lett. 132 244, 139(1995). 107. F. Faupel, R. Willecke, and A. Thran,Mater. Sci. Eng. Res. 22, 1 (1998). 108. J. Birgerson, M. Fahlman, P. Broms, and W. R. Salaneck, Synth. Met. 80, 125 (1996). 109. M. Atreya, S. Li, E. T. Kang, K. G. Neoh, Z. H. Ma and K. L. Tan, J. Vac. Sci. Technol. A 17, 853 (1999). 110. S. Li, E. T. Kang, and K. G. Neoh, Z. H. Ma and K. L. Tan, Surf. Sci. 454-456, 990 (2000). 111. S. Li, E. T. Kang, and K. G. Neoh, Z. H. Ma, K. L. Tan, and W. Huang, Appl Surf. Sci. 181, 201 (2001). 112. W. Salaneck, and J. L. Bredas, Adv. Mater. 8, 48 (1996). 113. Y. Cao, G. Yu, I. D. Parker, and A. J. Heeger, J. Appl. Phys. 88, 3618 (2000). 114. P. Piromreun et al, Appl. Phys. Lett. 77, 2403 (2000). 115. V. E. Choong, M. G. Mason, C. W. Tang, and Y. Gao, Appl. Phys. Lett. 72, 2689 (1998) . 116. Q. T. Lee, L. Yan, V. E. Choong, E. W. Forsythe, M. G. Mason, C. W. Tang, and Y. Gao, Synth. Met. 102, 1014 (1999). 117. Y. Park, V. Choong, E. Ettedgui, Y. Gao, B. R. Hsieh, T. Wehrmeister, and K. Mullen, Appl. Phys. Lett. 69, 1080 (1996); Y. Park, V. E. Choong, B. R. Hsieh, C. W. Tang, T. Wehrmeister, K. Mullen, and Y. Gao, J. Vac. Sci. Technol. A 15, 2574 (1997). 118. B. R. Hsieh, E. Ettedgui, and Y Gao, Synth. Met. 78, 269 (1996); B. R. Hsieh, E. Ettedgui, and Y Gao, Polym. Adv. Technol. 8, 408 (1997). 119. K. Konstadinidis, F. Papadimitrakopoulos, M. Galvin, and R. L. Opila, J. Appl Phys. 77, 5642 (1995). 120. Y. Gao, K. T. Park and B. R. Hsieh, /. Chem. Phys. 97, 6991 (1992); Y. Gao, K. T. Park and B. R. Hsieh, J. Appl. Phys. 73,7894 (1993); Y. Gao, Acc. Chem. Res. 32, 247 (1999) . 121. E. Ettedgui, H. Razafitrimo, B. R. Hsieh, and Y Gao, Appl. Phys. Lett. 67, 2705 (1995). 122. V. Choong, Y. Park, and Y. Gao, T. Wehrmeister and K. Mullen, B. R. Hsieh and C. W. Tang, Appl. Phys. Lett. 69 , 1492 (1996); V. Choong et al, J. Vac. Sci. Technol A, 15, 133 1745 (1997). 123. Y. Park, V. E. Choong, B. R. Hsieh, C. W. Tang, and Y. Gao, Phys. Rev. Lett. 78, 3955 (1997). 124. P. Broms, J. Birgersson, N. Johansson, M. Logdlund, and W. R. Salaneck, Synth. Met. 74, 179 (1995). 125. L. S. Liao, L. F. Cheng, M. K. Fung, C. S. Lee, S. T. Lee, M. Inbasekaran, E. P. Woo, and W. W. Lee, Chem. Phys. Lett. 325, 405 (2000), and Phys. Rev. B 62, 10004 (2000). 126. M. Jorgensen, and F. C. Krebs, Poly. Bull. 51, 23 (2003). 127. M. Onoda, K. Tada, A. A. Zakhidov and K. Yoshino, Thin Solid Films 331, 76 (1998); L. Micaroni, F. C. Nart, and I. A. Hummelgen, J. Solid State Electrochem. 7, 55 (2002) . 128. J. L. Bredas, R. L. Elsenbaumer, R. R. Chance, and R. Silbey, J. Chem. Phys. 78, 5656(1983). 129. B. Themans, W. R. Salaneck, and J. L. Bredas, Synth. Met. 28, C359 (1989). 130. H. Eckhardt, L. W. Shacklette, K. Y. Jen, and R. L. Elsenbaumer, J. Chem. Phys. 91, 1303 (1989); for more references see G. Yoder, B. K. Dickerson, and A. B. Chen., J. Chem. Phys. Ill, 10347 (1999). 131. M. Onoda, K. Tada, and H. Nakayama, J. Appl. Phys. 86, 2110 (1999); N. Nakanishi, K. Tada, M. Onoda, and H. Nakayama, Appl. Phys. Lett. 75, 226 (1999). 132. N. Stutzmann, R. H. Friend, and H. Sirringhaus, Science 299, 1881 (2003). 133. L. Burgi, T. J. Richards, R. H. Friend, and H. Sirringhaus, J. Appl. Phys. 94, 6129 (2003) . 134. R. Valaski, L. M. Moreira, L. Micaroni and I. A. Hummelgen, J. Appl. Phys. 92, 2035 (2002). 135. T. A. Chen, X. Wu, and R. Rieke, J. Am. Chem. Soc. Ill, 233 (1995). 136. R. D. McCullough, R. D. Lowe, M. Jayaraman, and D. Anderson, J. Org. Chem., 58, 904 (1993); B. D. Malhotra, W. Takashima, S. S. Pandey, R. Singhal, K. Endo, M. Rikukawa, and K. Kaneto, Jpn. J. Appl. Phys., Part 1 38, 6768 (1999); R. D. McCullough, S. Tristram-Nagle, S. P. Williams, R. D. Lowevt and M. Jayaramant, J. Am. Chem. Soc. 115, 4910(1993). 134 137. W.Takashima, S. S. Pandey, T. Endo, M. Rikukawa, N. Tanigaki, Y. Yoshida, K. Yase, K. Kaneto, Thin Solid Films 393, 334 (2001). 138. Z. Chiguvare, J. Parisi, and V. Dyakonov, J. Appl. Phys. 94, 2440 (2003). 139.1. H. Campbell, D. L. Smith, C. J. Neef, and J. P. Ferraris, Appl. Phys. Lett. 74, 2809 (1999). 140. W. Takashima , S. S. Pandey, T. Endo, M. Rikukawa, K. Kaneto, Curr. Appl. Phys. 1,90(2001). 141. Z. Bao, Y. Feng, A. Dodabalapur, V. R. Raju, and A. J. Lovinger, Chem. Mater. 9, 1299 (1997). 142. G. Dicker, T. J. Saveniji, B. H. Huisman, D. M. de Leeuw, M. P. de Haas, and J. M. Warman, Synth. Met. 137, 863 (2003). 143. M. Kobashi and H. Takeuchi, Macromol. 31, 7273 (1998). 144. M. Westerling, R. Osterbacka , H. Stubb, Synth. Met. 119, 623 (2001). 145. A. Assadi, C. Svensson, M. Willander, and O. Inganas, Appl. Phys. Lett. 53, 195 (1988). 146. G. Xu, Z. Bao, and J. T. Groves, Langmuir 16, 1834 (2000). 147. R. J. Kline, M. D. McGehee, E. N. Kadnikova, J. Liu, and J. M. J. Frechet, Adv. Mater. 15, 1519 (2003). 148. T. C. Chung, J. H. Kaufman, A. J. Heeger and F. Wudl, Phys. Rev. B 30, 702 (1984). 149. Z. Vardeny, E. Ehrenfreund, O. Brafman, M. Nowak, H. Schaffer, A. J. Heeger and F. Wudl, Phys. Rev. Lett. 56, 671 (1986). 150. Y. H. Kim, S. Hotta, and A. J. Heeger, Phys. Rev. B 36, 7486 (1987). 151. Y. H. Kim, D. Spiegel, S. Hotta, and A. J. Heeger, Phys. Rev. B 38, 5490 (1988). 152. M. J. Nowak, S. D. D.V Rughooputh, S. Hotta, and A. J. Heeger, Macromolecules, 20, 965 (1987). 153. Z. Bao, A. Dodabalapur, A. J. Lovinger, Appl. Phys. Lett. 69, 4108 (1996). 154. M. Wohlgenannt, C. P. An, and Z. V. Vardeny, J. Phys. Chem. B 104, 3846 (2000). 155. J. Ament, M.Sc. thesis, University of British Columbia 2001. 156. VG100AX Operating Manual, Document Number HA010006, VG Microtech, 1995. 157. M. V. R. Murty, T. Curcic, A. Judy, B. H. Cooper, A. R. Woll, J. D. Brock, S. Kycia, and R. L. Headrick, Phys. Rev B, 60, 16956 (1999). 135 158. M. V. R. Murty, A. J. Couture, B. H. Cooper, A. R. Woll, J. D. Brock, and R. L. Headrick, J. Appl. Phys. 88, 597 (2000). 159. T. L. Gilton, J. P. Cowin, G. D. Kubiak, and A.V. Hamza, J. Appl. Phys. 68, 4802 (1990). 160. Excimer Laser COMPex Instruction Manual, Lamda Physik, 1994. 161. Dye Laser SCANmate Instruction Manual, Lamda Physik, 1993. 162. Lamdachrome Laser Dyes Data Sheets, 2nd Ed., 1994. 163. Omicron CMA User's Guide and Instruction Manual, 1995. 164. Park Scientific Instrument (Thermo Microscopes Inc.), A Practical Guide to scanning probe microscopy and Autoprobe CP manual. 165. W. Haiss, D. Lackey, and J. K. Sass, J. Chem. Phys. 95, 2193 (1991). 166. M. H. Dishner, M. M. Ivey, S. Gorer, J. C. Hemminger, and F. J. Feher, J. Vac. Sci. Technol. A 16, 3295 (1998). 167. J. V. Barth, H. Brune, G. Ertl, and R. J. Behm, Phys. Rev. B 42, 9307 (1990). 168. J. A. DeRose, D. B. Lampner, S. M. Lindsay, and N. J. Tao, J. Vac. Sci. Technol. A 11, 776(1993). 169. A. Ulman, Chem. Rev. 96, 1533 (1996). 170. F. Elfeninat, C. Fredriksson, E. Sacher, and A. Selmani, J. Chem. Phys. 102, 6153 (1995). 171. M. Cecchi, H. Smith, and D. Braun, Synth. Met. 121, 1715 (2001). 172. H. G. O. Sandberg, G. L. Frey, M. N. Shkunov, H. Sirringhaus, R. H. Friend, M. M. Nielsen, and C. Kumpf, Langmuir 18, 10176 (2002). 173. E. Mena-Osteritz, A. Meyer, B. M. W. Langeveld-Voss, R. A. J. Janssen, E. W. Meijer, and P. Bauerle, Angew, Chem. Int. Ed. 39, 2680 (2000); B. Grevin, P. Rannou, R. Payerne, A. Pron, and J. P. Travers, Adv. Mater. 15, 881 (2003). 174. K. Kaneto, K. Harada, W. Takashima, K. Endo, and M. Rikukawa, Jpn. J. Appl. Phys. 38, L1062 (1999). 175. T. G. Stange, D. F. Evans, W. A. Hendrickson, Langmuir 13, 4459 (1997). 176. D. Zerulla, and T. Chasse, Langmuir 15, 5285 (1999). 177. A. Adnot, Surface Science Western E-mail list, http://www.uwo.ca/ssw/list.html, 1999. 136 178. C. Ton-That, A. G. Shard, and R. H. Bradley, Langmuir 16, 2281 (2000). 179. W. F. Heinz and J. H. Hoh, Tibtech 17, 143 (1999). 180. M. P. Seah, and W. A. Dench, Surf. Interface Anal. 1, 2 (1979). 181. C. J. Powell, and A. Jablonski, NISTElectron Inelastic-Mean-Free-Path Database -Version 1.1, National Institute of Standards and Technology, Gaithersburg, MD (2000). 182. P. J. Cumpson, Surf. Interface Anal. 29, 403 (2000). 183. S. Marchant, and P. J. S. Foot, Polymer 38, 1749 (1997). 184. Vibrational spectroscopy of molecules on surfaces, J. T. Yates and T. Madey, eds., Plenum Press, New Yourk 1987 185. R. Khanna, A. T. Jameel, and A. Sharma, bid. Eng. Chem. Res. 35, 3081 (1996); S. Qu, C. J. Clarke, Y. Liu, M. H. Rafailovich, J. Sokolov, K. V. Phelan, and G. Krausch, Macromolecules 30, 3640 (1997). 186. H. L. Skriver, and N. M. Rosengaard, Phys. Rev. B. 46, 7157 (1992): K. F. Wojciechowski, Surf Sci. 437, 285 (1999); L. Vitos, A.V. Ruban, H. L. Skriver, and J. Kollar, Surf. Sci. 411, 186 (1998). 187. P. A. Savintsev, A. A. Shebzukhov, Kh. T. Shidov Fiz. Khim. Poverkh. Yavlenii Vys. Temp.\91\, 196-200, V. N. Eremenko (Ed.) NaukovaDumka, Kiev. 188. D. J. O'Connor, B. A. Sexton, and R. St. C. Smart (Eds), Surface Analysis Methods in Material Science, Springer-Verlag, 1992. 189. Q. T. Jiang, T. Gustafsson, P. Haberle, and D. M. Zehner, Phys. Rev B 45, 14256 (1992). 190. O. Millo, A. Many, and Y Goldstein, J. Vac. Sci. Technol. A, 7, 2688 (1989); P. H. Holloway, J. Vac. Sci. Technol. 12, 1418 (1975). 191. Handbook of Auger Electron Spectroscopy, 3rd edition, Physical Electronics, 1995. 192. S. T. Lee, G. Apai, M. G. Mason, R. Benbow, and Z. Hurych, Phys. Rev. B 23, 505 (1981). 193. J. Cao, Y. Gao, H. E. Elsayed-Ali, R. J. D. Miller, and D. A. Mantell, Phys. Rev. B 58, 10948 (1998). 194. S. C. Veenstra, U. Stalmach, V. V. Krasnikov, G. Hadziioannou, H. T. Jonkman, A. Heeres, and G. A. Sawatzky, Appl. Phys. Lett. 76, 2253 (2000). 137 195. H. B. Michaelson, J. Appl. Phys. 48, 4729(1977); D. E. Eastman, Phys. Rev. B 2, 1, (1970). 196. M. Weinelt, J. Phys.: Condens. Matter 14, R1099 (2002); more references therein. 197. C. Tomas, E. Vinet, and J. P. Girardeau-Montaut, Appl. Phys. A 68, 315 (1999). 198. K. Giesen, F. Hage, F. J. Himpsel, H. J. Riess, and W. Steinmann, Phys. Rev. Lett. 55,300(1985). 199. E. M. Logothetis and P. L. Hartman, Phys. Rev. Lett. 18, 581 (1967). 200. Handbook of Chemistry and Physics, 84th Ed. 2003-2004. 201. D. Burgess, Jr., P. C. Stair, and E. Weitz, J. Vac. Sci. Technol. A 4, 1362 (1986). 202. M. Bensoussan, J. M. Moison, B. Stoesz, and C. Sebenne, Phys. Rev. B. 23, 992 (1981). 203. T. Lopez-Rios, and G. Hincelin, Phys. Rev. B 38, 3561 (1988). 204. E.V. Chulkov, M. Machado and V.M. Silkin, Vacuum 61, 95 (2001). 205. R. Courths, H. G. Zimmer, A. Goldmann, and H. Saalfeld, Phys. Rev B 34, 3577(1986); S. D. Kevan and R. H. Gaylord, Phys. Rev. B 36, 5809 (1987). 206. D. P. Woodruff, W. A. Royer and N. V. Smith, Phys. Rev. B 34, 764 (1986). 207. S. LaShell, B. A. McDougall, and E. Jensen, Phys. Rev. Lett. 11, 3419 (1996). 208. R. Paniago, R. Matzdorf, G. Meister, and A. Goldmann, Surf. Sci. 336, 113 (1995). 209. J. T. Stuckless, and M. Moskovitis, Phys. Rev. B 40, 9997 (1989). 210. N. Hosaka, H. Tachibana, N. Shiga, M. Matsumoto, and Y. Tokura, Phys. Rev. Lett. 82, 1672 (1999). 211. W. E. Spicer, Phys. Rev. B 111, 114 (1958). 212. C. Silva, D. M. Russell, A. S. Dhoot, L. M. Herz, C. Daniel, N. C. Greenham, A.C. Arias, S. Setayesh, K. Mullen and R. H. Friend, J. Phys.: Condens. Matter 14, 9803 (2002). 213. C. Silva, A. S. Dhoot, D. M. Russell, M. A. Stevens, A. C. Arias, J. D. MacKenzie, N. C. Greenham, S. Setayesh, K. Mullen, andR. H. Friend, Phys. Rev. B 64, 125211 (2001). 214. K. Tashiro, M. Kobayashi, T. Kawai, and K. Yoshino, Polymer 38, 2867 (1997). 215. A. Tivanski, J. E. Bemis, B. B. Akhremitchev, H. Liu, and G. C. Walker, Langmuir 19, 1929 (2003). 138 216. M. Pope and C. Swenberg, Electronic Process in Organic Crystals and Polymers, 2nd Ed. Oxford Univ. Press (1999). 217. C. D. Zangmeister, S. W. Robey, R. D. van Zee, Y. Yao, and J. M. Tour, J. Am. Chem. Soc. 126, 3420 (2004). 218. U. Zhokhavets, R. Goldhahn, G. Gobsch, and W. Schliefke, Synth. Met. 138, 491 (2003). 219. K. S akurai, H. T achibana, N. S higa, C. T erakura, M. M atsumoto a nd Y T okura, Phys. Rev. 5 56,9552 (1997). 220. K. Shudo, S. Takeda, and T.Munakata, Phys. Rev. B 65, 075302 (2002). 221. M. Logdlund, R. Lazzaroni, S. Stafstrom, W. R. Salaneck, and J. L. Bredas, Phys. Rev. Lett. 63, 1841 (1989); R. Lazzaroni, M. Logdlund, S. Stafstrom, W. R. Salaneck, and J. L. Bredas, J. Chem. Phys. 93, 4433 (1990). 222. R. D. McCullough, Adv. Mat. 10, 93 (1998). 223. P. Yeh, Optical waves in layered media, John Wiley & Sons, 1988. 224.1. Musa, and W. Eccleston, Synth. Met. 97, 69 (1998). 225. P. B. Johnson and R. W. Christy, Phys. Rev. B 6, 4370 (1972). 226. J. M. Warman, G. Ft. Gelinck and M. P. deHaas, J. Phys.: Condens. Matter 14, 9935 (2002). 227. J. Yang, I. Shalish, and Y. Shapira, Phys. Rev. B 64, 035325 (2001). 228. M. Tammer, and A. P. Monkman, Adv. Mater. 14, 210 (2002). 229.1. B. Martini, A. D. Smith, and B. J. Schwartz, Phys. Rev. B 69, 035204 (2004). 230. B. J. Schwartz, Annu. Rev. Phys. Chem. 54, 141 (2003). 231. Osterbacka, Phys. Rev. Lett. 91, 219701 (2003); A. S. Dhoot, and N. C. Greenham, ibid 91, 219702 (2003). 232. T. Q. Nguyen, I. B. Martini, J. Liu, and B. J. Schwartz, J. Phys. Chem. B 104, 237 (2000). 233. I. H. Campbell, P. S. Davids, D. L. Smith, N. N. Barashkov, and J. P. Ferraris, J. Appl. Phys. 72, 1863 (1998). 234. I. H. Campbell, T. W. Hagler, D. L. Smith, and J. P. Ferraris, Phys. Rev. Lett. 76, 1900 (1996). 235. G. L. J. A. Rikken et al, Syn. Met. 67, 115 (1994). 139 236. M. Atreya, S. Li, E. T. Kang, K. G. Neoh, Z. H. Ma, K. L. Tan, and W. Huang, Polym. Degrad. and Stab. 65, 287 (1999). 237. M. G. Harrison, J. Gruner, and G. C. W. Spencer, Phys. Rev. B 55, 7831 (1997). 238. J. D. McNeill and P. F. Barbara, J. Phys. Chem. B 106, 4632 (2002). 239. B. A. Mattis, P. C. Chang, and V. Subramanian, Mat. Res. Soc. Symp. Proc. 771, L10.35.1 (2003). 240. C.Yang, F. P. Orfino, and S. Holdcroft, Macromolecules 29, 6510 (1996); C.Yang, and S. Holdcroft, Synth. Met. 84, 563 (1997). 241. N. E. Fisher, and D. J. Willock, Phys.: Condens. Matter 4, 2533 (1992). 242. M. S. A. Abdou, F. P. Orfino, Y. Son, and S. Holdcroft, J. Am. Chem. Soc. 119, 4518 (1997). 243. M. Nakazono, T. Kawai, and K. Yoshino, Chem. Mater. 6, 864 (1994). 244. J. Corish, D. E. Feeley, D. A. Morton-Blake, F. Beniere, and M. Marchetti, J. Phys. Chem. B 101, 10075 (1997). 245. B. Wegewijs,F. C. Grozema, L. D. A. Siebbeles, M. P. de Haas and D. M. de Leeuw, Synth. Met. 119, 431 (2001). 246. G. D. Hale, Abstract for Meeting of the American Physical Society, Los Angeles, CA (March 1998). 140 Appendix ONE We consider the reflection, transmission and absorption of incident electromagnetic radiation of a thin polymer film bounded by vacuum and a semi-infinite metal substrate. m lo S R e \ e / xH^x Vacuum m ^\ Polymer m NT Mclal Figure 1: Radiation incident at angle 6 on polymer of thickness d. R, A and T are the fraction of the incident flux which are reflected into vacuum, absorbed in the polymer and transmitted into the metal. The optical constants of a thin film are represented by a complex refractive index m . The complex refractive index n is expressed in terms of its real (n) and imaginary (k) components: n = n+ ik where n is the refraction index and k is the extinction coefficient. The k is related with the absorption constant; a = Ank IA , where A is a wavelength of incidence light. In an unbounded medium, Iz = /0e_az. The relationship between a complex refractive index n and a complex dielectric constant is e = ea + i £b = n Thus: ea = n2 - k2 £b = 2 n k We assume that all the media are homogenous. The complex refractive indices are written as m , m, and m, and dielectric constants as ey, E2, and s3, respectively. 141 Based on conservation of tangential momentum, the wave vectors at each interface are; k\z kiz ' 2 -n\ sin20 ~ 2 -m sin2c9 /- 2 - 2 A:3Z = V«3 -«i sin # where 6 is the incidence angle from the surface normal. Based on continuity of tangential electric field vectors, according to the Fresnel's equations, the reflection and transmission coefficients for s-polarization are, respectively, k\z—kiz *• kiz—kiz ru=- ~— f23=-kiz+kiz k2z+kiz and 2ku ~ 2kiz t\2=— ~— tli= — k\z+k2z kiz+k^z The transmission and reflection coefficients are written as 2b " " b rn +r23e ~ tntne (\ + mr2ie2b) (\ + mr23e2b) where b = 2nk\2 —i. X The transmittance and reflectance are; R =( cos(6>) For p-polarization with continuity of tangential magnetic field vectors obtains, ^ 2 ~ ~ 2 ~ ~2~ ~ 2 -«2 k\z-n\ k2z ~ m k2z-n2 &3z )2 T=Re(k3) ,2 rn ~ 2~ ~ 2~ rri " 2~ ~ 2~ «2 &lz+«l &2z «3 ^2z+«2 &3z >v y\ ->v >*v -*v 2«1«2&1Z ~ 2ft2«3&2z ^2=^rr^ — ^23=^7 — «2 Ariz -t- «i &2z m &2z+«2 ky2 ru+riie * tntne " (\ + rnr2ie-2ih>d) " (X + rxir-ae-21*^) 142 The total absorbance of the polymer can be written as A = 1 - T - R. Example: Sample calculations for the P3HT film (« = 1.3 + 0.03-i) [224] on gold (n = 1.63 + 1.74-0 [225] at 3.65 eV photons. ir 1 1 1 1 1 o.s r 0 10 20 30 40 50 Thickness (nm) Figure 2: Reflectance, transmittance, and absorbance versus P3HT film thickness at an incident angle of 75° from the surface normal, with a photon energy (s-polarized) of 3.65 eV. In Beer-Lambert's Law, the T and A are T = Id 110 and A = - log T = -a-d, respectively, where Id = I0e'ad . Note A = 1 - R - T can be different from absorbance in Beer's law. The absorbance derived from the Fresnel equation is not linear with film thickness. In particular, E fields tangential to metal surface decay quickly on approach to metal, giving a steeper gradient of field and absorbance rate than predicted by Beer's law, as shown in Figure 3. 143 144 Appendix TWO 145 

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 30 2
Canada 11 0
China 5 30
Germany 5 1
Japan 4 0
France 3 0
Algeria 2 0
United Kingdom 1 0
Italy 1 0
Ukraine 1 0
Denmark 1 0
Russia 1 0
City Views Downloads
Mountain View 9 0
Vancouver 6 0
East Lansing 6 0
Unknown 4 1
Houston 4 0
Tokyo 4 0
Ottawa 3 0
Beijing 3 0
Guangzhou 2 0
Sunnyvale 2 0
Ashburn 2 0
Riverside County 2 0
Montreal 2 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061205/manifest

Comment

Related Items