Raman Spectroscopy of Metal/Organic/Inorganic

Heterostructures and Pentacene-Based OFETs

von der Fakultät für Naturwissenschaften der Technischen Universität Chemnitz

genehmigte Dissertation zur Erlangung des akademischen Grades

doctor rerum naturalium (Dr. rer. nat.)

vorgelegt von MSc. Phys. Beynor Antonio Paez-Sierra

geboren am 24. September 1971 in Florián-Santander

eingereicht am 06. März 2007

Gutachter:

Prof. Dr. Dr. h.c. Dietrich R. T. Zahn

Prof. Dr. Veit Wagner

Prof. Dr. Michael Hietschold

Tag der Verteidigung: 20. Dezember 2007

http://archiv.tu-chemnitz.de/pub/2008

B. A. Paez-Sierra, Dedication ii

Dedication

This thesis is dedicated to my wife Viktoriia

B. A. Paez-Sierra, Bibliografische Beschreibung iii

Bibliografische Beschreibung

MSc. Phys. Beynor Antonio Paez Sierra Raman Spectroscopy of Metal/Organic/Inorganic Heterostructures and Pentacene-Based OFETs Technische Universität Chemnitz Dissertation (in englische Sprache), 2007 Im Rahmen dieser Arbeit wurden die Wechselwirkung von Indium (In) und Magnesium (Mg) als Topelektroden auf zwei Perylen-Derivativen, 3,4,9,10-Perylentetracarbonsäure Dianhydrid (PTCDA) und Dimethyl-3,4,9,10-Perylentetracarbonsäure Diimid (DiMe-PTCDI) untersucht. Die Metal/organische Schichten wurden auf S-passivierten GaAs(100):2x1-Substraten hergestellt und unter Ultrahochvakuum (UHV)-Bedingungens aufgedampft. Als Hauptcharakterisierungsmethode wird die Raman-Spektroskopie eingesetzt, die eine nicht-destruktive Methode ist, und auch in situ Untersuchungen des Wachstumsprozesses ermöglicht. Die experimentell Ergebnisse haben gezeigt, dass alle aufgedampft Metallen auf die organische Schichten von PTCDA und DiMe-PTCDI eine Verstärkung des aktive Raman Signals von interne Schwingungsmoden fördern, begleitet durch die Aktivierung von normalerweise Infrarotaktivemoden. Diesem Phänomen als Oberflächenverstärkte Raman-Spektroskopie (SERS) genannt ist. Die Verstärkungsfaktoren sind in bereich von 101 zu 102 für die chemische Verstärkung und von 101 zu 104 für die Elektromagnetische Verstärkung. Interessant war, dass das In Wachstum auf PTCDA und DiMe-PTCDI nur eine Einfluss auf der Molekülgeometrie hat, und mit eine starker Diffusion durch die PTCDA- als in DiMePTCDI- Schichten. Das Mg Wachstum auf beiden Molekularstrukturen wurde durch die viel niedrigere Diffusion des Metalls in die organischen Molekülen im Vergleich zum Indium, es war durch die Bewahrung des von externe molekulare Schwingungsmoden nach das Metallswachstum, und in ersten Mal in einem Ramanexperiment beobachtet. Die PTCDA/Mg Strukturen formen sich durch zwei Stufen des Metallwachstum, die erste gehört zu einer neuen molekularen Struktur für eine Mg Schicht dünner als 2.8 nm, wo das PTCDA Molekühl des Sauerstoff-Atoms von die dianhydride Gruppe verliert. Die zweite gehört zu das SERS Spektrum von die vorherige Struktur. Im Fall von Mg/DiMe-PTCDI Heterostrukturen, den Molekühl wird gut bewahrt, wo die Raman Verschiebung an der diimide Gruppe wird nicht modifiziert. Auch von dieser Struktur eine interessante Eigenschaft wurde durch die Kopplung zwischen diskret Moleküleigenschwingungen am 221 cm-1, 1291 cm-1 und 1606 cm-1 des organischen Materials und den elektronischen Kontinuum-Zuständen des Mg-Metallkontakts. Ihre entsprechenden Energieliniengestalten werden gut durch die Breit-Wigner-Fano-Funktion beschrieben. Die Untersuchungen auf dem vorherigen Heterostrukturen half, die Kanalbildung von Pentacen-basierten organische Feldeffekt-Transistoren (OFETs) experimentell zu analysieren, und in ersten Mal in einem Ramanexperiment durchgeführt. Der organische Kanal war gebildet durch die organische Molekularstrahldeposition (OMBD) unter UHV-Bedingungens der Pentacen Moleküle, und es war mit eine Evaporationsrate von ca. 0.65 Å/min aufgedampft. Nach jede Aufdampfung von ca. 0.1 nm des organische Moleküle, den Strom und den Ramansignal in den Kanal wurden in situ gemessen. Die minimale nominelle Dicke des organischen Materials erforderlich für den effizienten Ladungstransport durch den OFET Kanal wurde um ungefähr 1.5 nm nomineller Einschluss oder 1.1 Monolagen (ML) zu sein. Eigenschaften der ersten Monolagen werden gut im Vergleich mit dickeren Schichten definiert, wo die 1.1 ML eine gestrecktes Natur wegen seines direkten Kontakts mit dem Gate-Isolator präsentieren. Es wurde gefunden, dass der leitende organische Kanal bzw. -organische erhöhende Schicht (OBL)- eine Druckdeformierung hat. Dieses Phänomen durch die rote Verschiebung der Ramanbanden beobachtet war. Das Ausgangskennlinienfeld des OFETs wurden nach die letzte aufgedampft organische Schicht gemessen. Es wurde gefunden, dass der Drain-Strom einem Relaxationsprozesse mit zwei Zeitkonstanten hat, wo eine in der Ordnung von 101 min ist und die zweite unter 102 min. Ein ähnliches Experiment mit der Beleuchtung des Kanals mit einer 676.4 nm Laserquelle, es erhöht der Drain-Strom und lässt ummodifiziert die Zeitkonstanten. In der Ergänzung, die OFET-Strukturen waren ex situ durch Landungstransientspektroskopie (QTS) unstersucht. Die QTS Spektren zeigten positive und negative Banden zum Gesamtsignal der relaxierte Ladung in Bezug auf die einzigartige Biaspulsepolarität. Wir haben dieses Phänomen als ,,anomales Verhalten des QTS-Signals“ genannt, und in ersten Mal in einem QTS-Experiment beobachtet. Bei Wiederholung der QTS-Messung innerhalb ca. 100 min, die QTS-Spektre eine langsame Relaxationsprozesse von Störstellen am s5 μ in bereich ca. 63 min < 102 min hat. Die Einfangsquerschnitten sind Zeitabhängig, es bedeutet, dass die Störstellendichte nicht Konstant im Lauf der Betriebs des OFET bleibt. Dafür des Drain-Strom verändert sich und die Beweglichkeit unabhängige des elektrisches Feld ist. Experimentell Untersuchungen auf dem OFETs mit der Kombination der Ramanspektroskopie und elektrischen Felder zeigten eine Erhöhung des Ramanseinfangsquerschnitt in endliche Bereich als die chemische SERS-Verstärkung von In bzw. Mg auf die Perylen-Derivativen PTCDA und DiMe-PTCDI. Nach den Ausschaltung des elektrisches Felds den Ramansignal des Pentacen-basierten OFET eine Relaxationsprozesse mit Zeitkonstant von ca. 94 min hat. Deshalb ist die Summe von Störstellensdichte wegen dieser am organische/anorganische Grenze plus dieser dass die elektrisches Felds am die organische Halbleiter induziert. Schlagwörter Organische Moleküle, Organische Molekularstrahldeposition (OMBD), In, Mg, GaAs(100), Grenzflächen, Ramanspektroskopie, Oberflächenverstärkte Raman-Spektroskopie (SERS), Fanoresonanz, organische Feld-Effekt Transistor (OFET), Landungstransientspektroskopie (QTS), Störstellen, organische erhöhende Schicht (OBL), Beweglichkeit, Modellbildung, Aktivierungsenergie, Ladungsdichte, Stromsrelaxation.

B. A. Paez-Sierra, Table of contents iv

Table of contents

Bibliografische Beschreibung……………………………………………………………….…..…... ii

Dedication……………………………………………………………………………….....………... iii

Table of contents……………………………………………………………………………….....…. iv

List of abbreviations…………………………………………………………………………...……. viii

Chapter 1. Introduction……………………………………………………………………...……. 1.1

1.1. A brief historic review of organic materials …………………………………………...….. 1.1

1.2. Investigation of organic/ anorganic heterostructures …………………………………....... 1.2

References……………………………………………………………………………..………... 1.4

Chapter 2. Fundamentals of molecular structures and Raman spectroscopy …………………. 2.1

2.1 Aromatic molecules……………………………………………………………………….… 2.1

2.2. The Benzene ring and π-electron delocalization……………………………………..…… 2.2

2.3. Organic semiconductor solid…………………………………………………………...….. 2.2

2.4. Transport in organic materials………………………………………………………..….… 2.3

2.4.1. Some input parameters for the coupled electron-phonon system………………..…... 2.4

2.4.2. Field effect mobility………………………………………………………..………... 2.5

2.5. Light-matter interaction……………………………………………………………..……... 2.7

2.5.1. Raman spectroscopy…………………………………………………………………. 2.9

2.5.2. Basic theoretical background on Raman spectroscopy…………………………...…. 2.9

2.5.2.1. Classical description of the Raman effect……………………………..…….. 2.9

2.5.2.2. Quantum mechanical description of the Raman effect…………………...….. 2.10

2.5.3. Surface-enhanced Raman scattering (SERS)…………………………………..….…. 2.14

2.5.4. Combined Raman spectroscopy with electric fields…………………………...…….. 2.16

References………………………………………………………………………………..……... 2.17

Chapter 3. Experimental techniques, materials, and algorithms…………………………...….. 3.1

3.1. Combined Raman Spectroscopy and electrical characterization setups……………..…….. 3.1

3.2. Sulphur passivation of GaAs(100)………………………………………………..………... 3.3

3.3. Materials and structures……………………………………………………………………. 3.3

3.3.1. Structures I. Perylene derivatives thin films capped by metallic overlayers of

Indium and Magnesium……………………………………..……………………………....

3.4

3.3.1.1. Molecular beam deposition (MBD) and metal evaporation…………..……... 3.4

3.3.2 Structures II. Pentacene and C60 organic molecules as active layers in field effect

B. A. Paez-Sierra, Table of contents v

devices………………………………………………………………………………..…….. 3.6

3.3.2.1. Molecular structures: Pentacene and C60…………………………………...... 3.6

3.3.2.2. Field effect structures……………………………………………...………… 3.8

3.3.2.3. Molecular beam deposition: pentacene and C60…..…………….....………… 3.8

3.4. Charge transient spectroscopy (QTS)…………………………………………...………..... 3.9

3.4.1. Shallow and deep level states in semiconductors……………………………...…….. 3.9

3.4.2. The charge transient spectroscopy (QTS) technique………………………...……….. 3.12

3.5. Algorithms for data evaluation and simulations………………………………………..….. 3.14

3.5.1. Gauss-Legendre quadrature……………………………………………………..….... 3.14

3.5.2. QTS spectra line-profile……………………………….………………………...…... 3.15

3.5.3. Correlated fitting algorithm…………………………………………………….…..... 3.15

References…………………………………………………………………………...………….. 3.16

Chapter 4. Metal / Organic interface formation investigated by in situ surface-enhanced

Raman spectroscopy (SERS)……………………………………………………………………...

4.1

4.1. Introduction………………………………………………………………………………… 4.2

4.2. Interaction of metals with perylene derivatives…………………………………………..... 4.3

4.3. Morphology of the metal film…………………………………………………………….... 4.6

4.4. Phonons and interface structural properties…………………………………………….…. 4.10

4.5. Mg/DiMe-PTCDI structures and discrete-molecular coupling with continuum electronic

metal states……………………………………………………………………………………....

4.13

4.6.1. Chemistry, metal film morphology and metal indiffusion……………….………….. 4.13

4.6.2. Coupling of vibrational modes and electronic excitations…………………….…….. 4.15

Conclusions……………………………………………………………………………….….… 4.18

References……………………………………………………………………………………..... 4.18

Chapter 5. Organic Field Effect Transistors (OFETs)………………………………………..…. 5.1

5.1. Introduction………………………………………………………………………………… 5.1

5.2. Statistical mechanics of charge carriers………………………………………………….... 5.3

5.2.1. Density of states……………………………………………………………………… 5.3

5.3.2 Charge carriers density…………………………………………………………….…. 5.3

5.3.3. Fermi integral argument……………………………………………………………... 5.4

5.4. Charge carrier density of organic materials………………………………………………... 5.5

5.5. The field effect transistor (FET)..................................................................................….…. 5.8

5.5.1. Energy band structure of an OFET..........................................................................…. 5.9

5.6. Output characteristics of the OFET………………………………………………………... 5.10

5.6.1. ,,Linear” regime……………………………………………………………………… 5.10

5.6.2. Saturation regime…………………………………………………………………….. 5.12

B. A. Paez-Sierra, Table of contents vi

5.6.3. Field effect mobility effμ .......................................................................................…. 5.13

5.7. Threshold voltage shift and field dependence……..………………………………………. 5.16

Conclusions………………………..………………………………………………………….… 5.18

References………………………….………………………………………………………..….. 5.18

Chapter 6. Combined Raman spectroscopy and electrical characterization of the conductive

channel in OFETs…….………………………………………………….…………………………

6.1

6.1. Introduction………………………………………………………………………………… 6.2

6.2. Simultaneous in situ I - V characterization and molecular vibration measurements of

OFETs………………………………..…………………………………………………….……

6.3

6.3. Organic boosting layer (OBL) in OFETs….……………………………………………..... 6.7

6.4. Characteristic regions of the organic layer in OFETs…….…………………………..…… 6.9

6.5. Organic – Insulator electrodynamics…………………….……………………………..….. 6.10

6.6. Vibrational bands profiling of the active layer………….……………………………..…... 6.12

6.7. Bias-stress effects and multi-exponential current relaxation………………………………. 6.15

Conclusions…………………………………………………………………………………...… 6.18

References………………………………………………………………………………………. 6.19

Chapter 7. Influence of electric fields and illumination in OFETs…………………………..…. 7.1

7.1. Introduction…………………………………………………………………………….…... 7.1

7.2. Raman bands and external electric fields………………..………………………… 7.2

7.2.1. Band gap modification by external electric fields………………………….………... 7.3

7.2.2. Raman bands alteration by external electric fields……………………………..……. 7.5

7.2.2.1. Pentacene…………………………………………………………………...... 7.5

7.2.3. The C60 fullerene………………………………………………………………..….… 7.9

7.3. Illumination and charge transport in OFETs……………………………………………..... 7.11

7.3.1. Persistent effects and multi-exponential kinetics……………………………………. 7.11

Conclusions…………………………………………………………………………………...… 7.13

References…………………………………………………………………………………….… 7.14

Chapter 8. Trap distribution in OFETs and anomalous QTS…………………………….…….. 8.1

8.1. Introduction……………………………………………………………………….………... 8.1

8.2. Traps and charge density distribution…………………………………………….………... 8.3

8.2.1. Effect of the electric field………………………………………………….………… 8.3

8.3. Anomalous behavior of the QTS signal…………………………………….……………… 8.7

8.3.1. Advantage of floating gate configuration in QTS measurements……….…………… 8.7

8.3.2. OFET devices with interdigitated source-drain electrodes…………….…………….. 8.8

8.3.3. Single channel OFET devices……………………………………………..…………. 8.10

8.4. Approach to modeling the ,,anomalous QTS signal”……………………………….….… 8.13

B. A. Paez-Sierra, Table of contents vii

8.5. Current collapse……………………………………………………………………………. 8.17

8.5.1. Negative conductance and photodetachment………………………….……………... 8.17

8.5.1.1. Charge conservation and photodetachment…………………….……………. 8.19

8.6. Photodetachment………………………………………………………..………………….. 8.19

Conclusions………………………………………………………………..……………………. 8.20

References………………………………………………………………………………………. 8.21

Chapter 9. Summary………………………………………………………………………………. 9.1

List of figures……………………………………………………………………………………….. 10.1

List of tables……………………………………………………………………………………….... 10.7

Erklärung............................................................................................................................................. 10.8

Curriculum Vitae................................................................................................................................. 10.9

List of publications………………………………………………………………………………….. 10.13

Acknowledgements………………………………………………………………………………….. 10.16

B. A. Paez-Sierra, List of abbreviations viii

List of abbreviations

AFM Atomic Force Microscopy

B3LYP Becke’s three parameter hybrid functional

BWF Breit-Wigner-Fano

cf. Confer

CCD Charge-Coupled Device

CT Charge Transfer

DFG Deutschen Forschungsgemeinschaft

DFT Density Functional Theory

DH4T Dihexylquaterthiophene

DLTS Deep Level Transient Spectroscopy

DiMe-PTCDI N-N’-dimethyl-3,4,9,10- perylene tetracarboxylic diimide

DIODE Designing Inorganic/Organic Devices

EFMPM Electrostatic Force Microscopy Phase Mode

FET(s) Field Effect Transistor(s)

gIRSE generalized Infrared Spectroscopy Ellipsometry

HOMO Highest Occupied Molecular Orbital

IPES Inverse Photoemission Spectroscopy

IRAV Infrared Active Vibrational

IR Infrared

I-V Current-Voltage

LED(s) Light Emitting Diode(s)

LO Longitudinal Optical

LUMO Lowest Unoccupied Molecular Orbital

MBD Molecular Beam Deposition

ML Mono Layer

MML Multi Mono Layers

MOSFET Metal-Oxide-Semiconductor Field Effect Transistor

NEXAFS Near Edge X-Ray Absorption Fine Structure

OBL Organic Boosting Layer

OFET(s) Organic Field Effect Transistor(s)

OLED(s) Organic Light Emitting Diode(s)

B. A. Paez-Sierra, List of abbreviations ix P3HT Poly(3-hexylthiophene)

PABA Para-Amino-Benzoic-Acid

PCBM [6,6]-Penyl-C61-butyric acid methyl ester

PEI Polyethylenimine

PTCDA 3,4,9,10-perylene tetracarboxylic dianhydride

QTS Charge Transient Spectroscopy

RS Raman Spectroscopy

RSS Raman Scanning Spectroscopy

(S)FWHM (Semi) Full Width at Half Maximum

SERS Surface-Enhanced Raman Scattering

SFG Sum Frequency Generation

SML Submonolayers

MESFET(s) Metal Semiconductor Field Effect Transistor(s)

MIS Metal Insulator Semiconductor

PPM Pentagonal Pinch Mode

RFID Radio Frequency Identification

SKPM Scanning Kelvin Probe Microscopy

SWCNT(s) Single-Wall Carbon Nanotube(s)

TFT(s) Thin Film Transistor(s)

TPD Thermal-Programmed Desorption

UPS Ultraviolet Photoemission Spectroscopy

UHV Ultra High Vacuum

UV-vis Ultraviolet-visible Spectrophotometry

VHR Variable Hopping Range

XRD X-Ray Diffraction

α -T6 α -sexithiophene

Chapter 1. B. A. Paez-Sierra, Introduction 1.1

Chapter 1

Introduction

1.1. A brief historic review of organic materials Carbon-based compounds are considered the fundamental core of organic materials. These molecules

provide a rich variety of properties, attractive for fundamental investigations in physics, chemistry,

biology, engineering, and material sciences [Reic2005]. Therefore, special interest has been focused

on the structural, reactive, and transport properties of such materials.

Through careful design and manipulation of a wide range of carbon-based structures, organic materials

have been implemented as hybrid structures, capable of acting as electronic conductors,

semiconductors, and insulators. In the electronic field, pioneering research based on organic materials

dates back to the early 1900s, when the aim was to identify changes in the electrical conductivity of an

alcoholic solution of eosin due to its fluorescence [Nich1904, Regn1903], and, in a similar manner,

with anthracene [Howe1910]. Later on, in the late 1940s, organic molecular films were produced

under vacuum conditions and revealed semiconducting properties [Pope1999].

Parallel to the experimental activity, numerous theoretical studies related to the ground state energy in

benzene, azulene, naphthalene, anthracene, tetracene (naphthacene), and pentacene were carried out

[Mann1949, Pari1956]. Also, electrical conductivity measurements on metal-free (mf) phthalocyanine

(Pc) [Fiel1957, Heil1962], and on several metal Pcs (mPcs) with central metal atom like Cu, Ni, Co,

Mn, and Fe, were performed, where a thermoelectric power of about C/V50 °μ+ for mf-Pc and

CuPc was measured, proving in this way the dominant p-type character of the molecules [Fiel1957].

An interesting complement related to electronic properties of organic solids and based on mf-Pc was

summarized in a publication issued in 1960 [Toll1960]. On the other hand, the n-type organic

counterpart dates back to a work reported in 1961 by Kommandeur and Hall [Komm1961]. The

research was performed on pyrene-iodine and perylene-iodine samples, where a high electronic

conductivity in these organic molecules was recorded. Another important issue is that of charge

trapping phenomena, with a pioneering investigation into Pc single crystals reported by Barbe et al.

[Barb1970].

Chapter 1. B. A. Paez-Sierra, Introduction 1.2 Particularly in the last 40 years, markedly growing research activity in the field of organic materials

has been registered, with greater emphasis on two well-linked aspects, namely fundamental research

and technology, during the past decade. Currently, research on large or small organic molecules for

organic-based electronics is huge [Klau2006]. Applications range from single organic-based devices,

i.e., diodes, transistors, light emitting devices (LEDs) [Kali2005], and gas sensors, towards more

sophisticated structures, such RF identity tags, large area solar cells, passive and active-matrix light-

emitting displays [Hadz2000, Kafa2005, Klau2006, Some2005].

Consequently, many investigations are targeted at understanding the electronic and chemical structure

of organic/inorganic interfaces and bulk-organic materials in order to tune the device performance

[Kahn2003, Scot2003, Zahn2005-2006]. In general, the physical and chemical properties of small

conjugated molecules and semiconducting or metallic polymers [Brüt2004, Heeg2001] make them

challenging candidates to solve drawbacks of the silicon-based devices, i.e., temperature processing,

the ability to be structured on plastic substrates and low-cost production.

The investigated aromatic molecules, namely the perylene derivative 3,4,9,10-perylenetetracarboxylic

dianhydride (PTCDA), N,N’-dimethylperylene 3,4,9,10-dicarboximide (DiMe-PTCDI), C60, and

pentacene, have served as prototype molecules to assemble different organic-based structures.

Experiments have probed n-channel transistors based on PTCDA [Ostr1997, Xue2004], n-alkyl

perylene diimides [Ches2004], and C60 [Koba2003]; while p-channel OFETs have been realized with

pentacene and many other organic molecules [Klau2006]. Still, the efficiency of p-channel OFETs is

well above the most efficient n-channel organic-based transistor. Other devices, named diodes, have

been demonstrated with In/PTCDA/Ag [Hude2002-2003] and Ag/DiMe-PTCDI/GaAs

heterostructures [Thur2005]. 1.2. Investigation of organic/ anorganic heterostructures Several devices based on organic molecules have demonstrated the reliability of a new electronic era.

Particularly, the organic field effect transistor (OFET) has been shown to be one of the most promising

structures to be coupled as a key building block in complex technological applications [Bao2006,

Hadz2000, Klau2006, Some2005]. Although the field effect transistor (FET) device has been known

since the previous century and was proposed by Lilienfeld between 1926 and 1933 [Klei1998], there

are still several challenges to be examined in order to understand the OFET version.

Several factors can influence the OFETs' performance, i.e., substrate handling, electrodes patterning,

organic material purification, environment, and formation of interfaces, among others. Within the

framework of the present research, special interest was addressed to the strategic features presented in

Figure 1.1; they consist of an inorganic semiconductor/organic semiconductor/metal heterostructure.

Chapter 1. B. A. Paez-Sierra, Introduction 1.3 The project in which the present research activity was included was called, "Organische Feldeffekt-

Transistoren: strukturelle und dynamische Eigenschaften“, or in its English version, “Organic Field

Effect Transistors: Structural and Dynamical Properties”. This was a six-year project, started in 2001

and supported by a grant from the ",Deutsche Forschungsgemeinschaft,DFG” (German Research

Society). The project was coordinated by Prof. Dr. Christoff Wöll of Ruhr Universität Bochum. More

than 20 academic institutions, mainly from Germany, participated in this research [ofet].

Figure 1.1. Organic materials forming different interfaces, i.e.metal / organic semiconductor/ inorganic

semiconductor heterostructure.

The contribution of the present work is the investigation, using combined Raman spectroscopy (RS)

and electrical characterization of OFETs, as well as in situ RS of organic/metal interface formation of

perylene derivatives PTCDA and DiMe-PTCDI with top metal contacts indium and magnesium (cf.

Figure 1.1). Concerning the latter, previous studies were devoted to the same perylene derivatives,

with Ag as top electrode [Paez2003, Salv2003, Zahn2004].

This thesis is organized as follows: Chapter 2 outlines an introduction to the fundamental aspects of

aromatic molecules and their role in building up the molecular solid. Additionally, the basic theoretical

background of Raman spectroscopy (RS), surface-enhanced RS (SERS) and the combined external-

applied electric fields with RS, important to characterize the heterostructures and OFET devices, is

presented.

In chapter 3, the experimental setups, i.e., RS, charge transient spectroscopy (QTS) and current-

voltage, are discussed. Afterwards, the procedures related to substrate preparation, molecular beam

deposition and metal evaporation (In, Mg), and the algorithms developed to analyze the measured data

are described. In chapter 4, the in situ measurements of RS and SERS of the interface formation

Chapter 1. B. A. Paez-Sierra, Introduction 1.4 between the metals indium and magnesium onto the perylene derivatives PTCDA and DiMe-PTCDI

deposited on S-GaAs substrates are analyzed.

The field effect transistor theory and its drawbacks, together with further improvements probed by

experimental measurements, are addressed in chapter 5. In chapter 6, the in situ combined Raman

spectroscopy and electrical characterization of the conductive channel in OFETs are presented. It is

demonstrated that the charge transport is developed in a two-dimensional regime, proving that the

minimum organic material needed to form the conductive channel scales well below a 10 nm organic

layer thickness. Additionally, the Raman bands assignments are based on density functional theory

(DFT) calculations. The chapter is concluded with a proposed model to describe the drain current

relaxation during operation of the OFET device.

In chapter 7, the influence of electric fields and illumination on the structural and transport properties

of the OFET channel is discussed. The results are complemented with DFT calculations. In chapter 8,

the intrinsic and field induced traps at the organic channel of OFET devices are investigated by QTS

measurements. Additionally, the anomalous behavior of the QTS signal is experimentally

demonstrated and analyzed by a proposed theoretical model. Finally, concluding remarks and a

summary of this research are given in chapter 9.

References

[Bao2006] “Organic field-effect transistors V”, edited by Z. Bao and D. J. Gundlach, Proc. SPIE 6336, Washington 2006. [Barb1970] D. F. Barbe and C. R. Westgate, “Bulk trapping states in β -phthalocyanine single crystals”, J. Chem. Phys.

52, 4046-4054 (1970). [Barb1970] D. F. Barbe and C. R. Westgate, “Surface state parameters of metal-free phthalocyanine single crystals”, Phys.

Chem. Solids 31, 2679-2687 (1970). [Brüt2004] “Physics of organic semiconductors”, edited by W. Brütting, phys. stat. sol. (a) 201, 1037-1371 (2004). [Ches2004] R. J. Chesterfield, J. C. McKeen, Ch. R. Newman, P. C. Ewbank, D. A. da Silva Filho, J. L. Brédas, L. L.

Miller, K. R. Mann, and C. D. Frisbie, “Organic thin film transistors based on N-alkyl perylene diimides: charge transport kinetics as a function of gate voltage and temperature”, J. Phys. Chem. 108, 19281-19292 (2004).

[Fiel1957] P. E. Fielding and F. Gutman, “Electrical properties of phthalocyanines”, The J. Chem. Phys. 26, 411-419 (1957).

[Hadz2000] “Semiconducting polymers: Chemistry, physics and engineering”, edited by G. Hadzüoannou and P. F. van Hutten, Wiley-VCH, Weinheim 2000.

[Heeg2001] A. J. Heeger, “Nobel lecture: Semiconducting and metallic polymers: The fourth generation of polymeric materials”, Rev. Mod. Phys. 73, 681-700 (2001).

[Heil1962] G. H. Heilmeier and G. Warfield, S. E. Harrison, “Measurement of the Hall effect in metal-free phthalocyanine crystals”, Phys. Rev. Lett. 8, 309 - 311 (1962).

[Howe1910] H. E. Howe, “The electrical conductivity of fluorescent Anthracene vapor”, Phys. Rev. 30 (series I), 453-459 (1910).

[Hude2002] R. Hudej, and G. Bratina, “Electronic transport in perylenetetracarboxylic dianhydride: the role of In diffusion”, J. Vac. Sci. Technol.A 20, 797-801 (2002).

[Hude2003] R. Hudej, and G. Bratina, “Electrical conductivity in metal/3,4,9,10-perylenetetrcarboxylic dianhydride structures”, J. Appl. Phys. 93, 6090-6094 (2003).

[Kafa2005] “Organic light-emitting materials and devices IX”, edited by Z. H. Kafafi and P. A. Lane, Proc. SPIE 5937, Washington 2005.

[Kahn2003] A. Kahn, N. Koch, W. Gao, “Electronic structure and electrical properties of interfaces between metals and π -conjugated molecular films” , J. Pol. Sci. Part B: Pol. Phys. 41, 2529 – 2548 (2003).

[Kali2005] J. Kalinowsky, “Organic light-emitting diodes, principles, characteristics and processes”, Marcel Dekker, New York 2005.

Chapter 1. B. A. Paez-Sierra, Introduction 1.5

[Klau2006] “Organic electronics”, edited by H. Klauk, Wiley-VCH, Weinheim 2006. [Klei1998] Ch. Kleint, “Julius Edgar Lilienfeld: Life and profession”, Prog. Sur. Sci., 57, 253-327 (1998).

[Koba2003] S. Kobayashi, T. Takenobu, S. Mori, A. Fujiwara, Y. Iwasa, “C60 thin-film transistors with high field-effect mobility, fabricated by molecular beam deposition”, Science and Technology of Advanced Materials 4, 371–375 (2003).

[Komm1961] J. Kommandeur and Frances R. Hall, “Electronic conduction in complexes of aromatic hydrocarbons with Iodine”, J. Chem. Phys. 34, 129-133 (1961).

[Mann1949] D. E. Mann and J. R. Platt, “Spectral resemblances in azulene and naphthalene”, J. Chem. Phys., 17, 481-484 (1949).

[Nich1904] E. L. Nichols and E. Merrit, “Studies of luminescence. IV. The influence of light upon the absorption and electrical conductivity of fluorescent solutions”, Phys. Rev. 19 (series I), 396-421 (1904).

[ofet] Schwerpunktprogramm 1121: Organische Feldeffekt-Transistoren: strukturelle und dynamische Eigenschaften www.ofet.de.Coordinator Prof. Dr. Christof Wöll. Activity in Chemnitz coordinated by Prof. Dr. D.r. h.c. D. R. T. Zahn and Dr. R. Scholz under the reference Za 146/4-2 as part of SPP 1121: Organic field effect transistors: Structural and dynamical properties.

[Ostr1997] J. R. Ostrick, A. Dodabalapur, L. Torsi, A. J. lovinger, E. W. Kwock, T. M. Miller, M. Galvin, M. Beggren, and H. E. Katz, “Conductivity-type anisotropy in molecular solids”, J. Appl. Phys. 81, 6804-6808 (1997).

[Paez2003] B. A. Paez, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Interaction of metals with perylene derivatives as a model system for contact formation in OFET structures”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 210-217 (2003).

[Pari1956] R. Pariser, “Theory of the electronic spectra and structure of the polyacenes and of alternant hydrocarbons”, J. Chem. Phys., 24, 250-268 (1956).

[Pope1999] M. Pope and C. E. Swenberg, “Electronic processes in organic crystals and polymers”, University Press, Oxford, second edition, 1999.

[Regn1903] Regner, Physikalische Zeitschrift, 4, 862, (1903). [Reic2005] E. Reichmanis, H. Katz, C. Kloc, A. Maliaka, “Plastic electronic devices: From materials design to device

applications”, Bell Labs Technical Journal 10, 87-105 (2005). [Salv2003] G. Salvan, “Metal/Organic/Inorganic semiconductor heterostructures characterized by vibrational

spectroscopies”, PhD thesis, http://archiv.tu-chemnitz.de/pub/2003/0109/index.html TU Chemnitz 2003. [Scot2003] J. C. Scott, “Metal-organic interface and charge injection in organic devices”, J. Vac. Sci. Technol. A 21,

521–531 (2005). [Some2005] T. Someya, Y. Kato, S. Iba, Y. Noguchi, T. Sekitani, H. Kawaguchi, and T. Sakurai, “Integration of organic

FETs With organic photodiodes for a large area, flexible, and lightweight sheet image scanners”, IEEE TRANSACTIONS ON ELECTRON DEVICES 52, 2502-2511 (2005).

[Thur2005] I. Thurzo, H. Méndez and D. R. T. Zahn, “Dielectric relaxation in a hybrid Ag/DiMe-PTCDI/GaAs device”. J. non crystal. Sol. 351, 2003-2008 (2005).

[Toll1960] G. Tollin, D. R. Kearns, and M. Calvin, “Electronic properties of organic solids. I. Kinetics and mechanism of conductivity of metal-free Phthalocyianine”, J. Chem. Phys. 32, 1013-1019 (1960).

[Xue2004] J. Xue and S. R. Forrest, “Bipolar doping between a molecular organic donor-acceptor couple”, Phys. Rev. B 69, 245322 1-8 (2004).

[Zahn2004] D. R. T. Zahn, G. Salvan, B. A. Paez, R. Scholz, “Interaction between metals and organic semiconductors studied by Raman spectroscopy”, J. Vac. Sci. Technol. A 22, 1482-1487 (2004).

[Zahn2005] D. R. T. Zahn, G. Salvan, G. Gavrila, and B. A. Paez, “Chemistry and morphological properties of metal interfaces to organic semiconductors”, Adv. sol. sta. phys. 45, 313-324 (2005).

[Zahn2006] D. R. T. Zahn, G. N. Gavrila, and M. Gorgoi, “The transport gap of organic semiconductors studied using the combination of direct and inverse photoemission”, Chem. Phys. 325, 99-112 (2006).

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.1

Chapter 2

Fundamentals of molecular structures and Raman

spectroscopy

2.1 Aromatic molecules

The π conjugation is what makes many organic molecules attractive for electronic applications, i.e.,

organic light emitting diodes (OLEDs) [Kafa2005], solar cells [Kafa2005], and OFETs [Bao2006].

The present work is based on a particular branch of π -conjugated aromatic structures named arenes,

with the benzene ring as their structural building block. Most worked out molecules in the arene

family are acenes, consisting of planar arrangements of benzene rings and sharing two carbons

between them.

Pentacene is a prototype molecule of the acene group (cf. Figure 2.1(a)). The acene family can be

described by the formula 4n22n4 HC ++ , n being the number of benzene rings constituting the molecule.

Another important arene sub-group is the perylene derivatives, with perylene, PTCDA, and DiMe-

PTCDI some representative molecular structures of this sub-family (cf. Figure 2.1(b)).

(a)

(b)

Figure 2.1. Some representative molecules of the arene family. (a) APentacene and (b) Perylene

derivatives. (*) Main structures investigated in the present research.

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.2 2.2. The benzene ring and π-electron delocalization

As an example of π orbitals conjugation, the benzene ring is considered. Its carbon skeleton is

arranged in a regular hexagon with bond lengths of 1.4 Å being halfway between the typical 22 spsp − single-bond distance of 1.46 Å and the 22 spsp = double-bond distance of 1.34 Å. Each

carbon atom is involved in 2sp hybrid atomic orbitals bonding to the hydrogen atom and the two

adjacent carbon atoms, thus describing the σ bonds presented in Figure 2.2(a). The fourth electron

takes up a zp orbital, which is for the most part situated at right angles to the plane of the molecule

(cf. Figure 2.2(b)). These zp orbitals are fairly extended in space and overlap to form a pair of π -

electron clouds on each side of the plane of the molecule, as illustrated in Figure 2.2(c) [Bowe2002,

Care2004]

(a) (b) (c)

Figure 2.2. σ−hybrids and π−molecular orbitals in benzene. a) Localized σ−orbitals, b) pz atomic

orbital, and c) delocalised π−orbitals with highest densities above and below the plane ring. (Thanks to

G. Gavrila).

The delocalized π -electrons can move freely inside the molecule and are, therefore, sometimes called

the mobile electrons or the unsaturated electrons, because they result from unsaturated bonds. Through

such mobile π−electrons, electrical perturbations are easily transmitted from one part of the molecule

to another.

2.3. Organic semiconductor solid

Organic materials are distinguished from inorganic solids basically due to the weak electronic

interaction between nearer molecules.

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.3 Thus, in a conglomerate of molecules forming the organic solid, the weak electron-electron interaction

takes the form of a kind of band structure through the π orbitals [Ishii1999, Pire1974-1984], cf. Figure

2.3. The effective potential well of an electron is formed by the atomic nuclei and other electrons. The

wells of the nuclei are merged in the upper part to form a broad well. Deep atomic orbitals are still

localized in the atomic potential well (core levels), but the upper ones interact to form delocalised

molecular orbitals (MOs). The outermost horizontal part of the potential well is the vacuum level

(VL). The energy separations of the highest occupied MO (HOMO) and lowest unoccupied (LUMO)

from the vacuum level (VL) are the ionisation potential (IP) and the electron affinity (χ) of the

molecular solid, respectively. Since the molecules interact only by the weak Van der Waals

interaction, the top part of the occupied valence states and the lower unoccupied states are usually

localised in each molecule, with narrow intermolecular bandwidths (typically lower than 0.1eV)

[Gutm1967, Kao1981]. Thus, the electronic structure of an organic solid largely preserves that of a

molecule and the validity of usual band theory (which assumes itinerant electrons) is often limited

[Duke1980]. The electronic structure in an organic solid is usually simplified and presented as the one

depicted on the right side of Figure 2.2. The Fermi level is also indicated there, since electrons fill the

energy levels following the Fermi statistics [Ménd2006].

Figure 2.3. Molecular states of an

organic solid [Ishii1999, Ménd2006,

Pire1974-1984].

2.4. Transport in organic materials

A charge transport in the organic semiconductor is developed through overlapping of π neighboring

molecular orbitals. The transfer of charge from site to site in the organic solid is assisted by internal

and external or libronic phonons. The electronic conductivity of organic materials does not fully

follow the classical diffusion models, and the mobility, together with other kinetic coefficients, is

thermally activated due to the phonon-assisted process. Accordingly, there is a need to introduce a

model able to describe the hopping process.

The dynamics of charge carrier-phonon interaction in an organic solid is given by the Hamiltonian of a

coupled fermion-boson system [Cruz1992, Well1996]:

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.4

( ) ( )( )( )∑

∑∑∑

σ

+σ−

+

σσ

+σσ

σσ

+σσ

+σ

+−ωε+

++ω+μ−+−=

,iiiiphn

iii2

1

,i,i,i,i

)j,i(,i,j,j,io

bb1n

bbcccccctH

h

h

, (2.1)

where ot− denotes the transfer amplitude (attractive interaction) between nearest neighbor pairs ( ji, ).

This quantity is given by the product between the ionization potential and the overlap integral. The

operator )(,ic +σ annihilates (creates) an electron at the site i with spin projection σ , σμ ,i is the

chemical potential and takes into account the mean number of particles at a given temperature, ω

denotes the bare phonon frequency, )(ib + is the phonon annihilation (creation) operator,

↓↑σ += iii nnn is the number operator (electrons), and phn−ε is the energy of the electron-phonon

coupling.

The Hamiltonian given in eq. (2.1) is the modern idea of the pioneering version of the coupling

between the electron and the lattice vibrations, leading to the formation of a composite particle named

polaron and introduced by Fröhlich [Fröh1950-1954], and later on by Lee [Lee1952], and in a

variational form with path integrals by Feynmann [Feyn1955]. From tight binding theory, it is found

that the band gap is approximately four times the transfer amplitude. Based on this result, in chapter 5

the charge carrier density for a two dimensional charge carrier gas is determined. An interesting

feature of the Hamiltonian in eq.(2.1) is that the last term also accounts for trap occupation, and the

coupling between the excess charge carriers with the phonons.

It is known that

( ) i

2/1ph

ii qm2

bb ⎟⎟⎠

⎞⎜⎜⎝

⎛ ω=++

h, (2.2)

phm being the phonon effective mass and iq the phononic-spatial coordinate. From eq.(2.2) and the

last term of the previous Hamiltonian (eq.(2.1)), one finds the strength between the phonon and the

excess charge carriers given by

phnphphn2

ph A2m2 −− εε=εω=ξ , (2.3)

with A the maximum amplitude of the normal mode with energy phε .

2.4.1. Some input parameters for the coupled electron-phonon system

In the present work it has been theoretically and experimentally proved by combined Raman

spectroscopy and applied electric fields that the vibrations of pentacene layers forming the channel in

FETs (chapter 7) are affected, in particular those where the ring and C-H vibrations are involved. The

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.5 representative Raman region is between 1120 cm-1 - 1210 cm-1, which amounts to energies ranging

between 139 meV and 150 meV. The affected vibrations basically correspond to gA modes, having

in-plane vibrations of C-H bonds at the outer rings, and vibrations of C-H bonds parallel to the main

axis of the molecule.

Table 1 summarizes some input parameters for the Hamiltonian in eq.(2.1). The quoted values are the

force constants K determined from density functional theory (DFT) at the B3LYP/3-21G level in

Gaussian 98 [Gaus1998]. These values are extracted after geometry optimization and calculations of

the vibrational Raman and infrared spectra of the pentacene molecule.

The next quantity in the table is the average energy of the local phonons >ε< ph , which are affected

by the external electric field; these quantities were extracted from experimental measurements of the

Raman spectra in pentacene-based FETs, discussed in chapter 7. Afterwards, the third column of the

table lists the maximum amplitude (A) of the normal modes with average energy >ε< ph . The next

column is the energy of the electron-phonon coupling >ε< −phn ; here the >ε< −phn is assumed to be

approximately the experimental activation energies actε described in chapter 8. Finally, the last

column is the electron-phonon coupling (ξ ) computed by means of the relation given in eq.(2.3).

Table 2.1. Parameters of the electron-phonon

coupling in pentacene-based FETs.

K

/ eV- Å-2

>ε< ph

/ meV

A

/ Å actphn ε≈ε −

/ meV

ξ

/ eV- Å-1

172.45 143.0 0.04 25.0 2.9

167.1 7.60

127.25 146.3 0.05 25.0 2.51

167.1 6.49

2.4.2. Field effect mobility

The organic-semiconducting solid is characterized by a charge transport with low mobilities in

conjunction with the monocrystalline or polycrystalline inorganic semiconductor. Usually, in most of

the cases, the mobility is lower than that of the organic-crystal form. The charge transport is

accompanied by thermally activated and electric-field dependent processes. Therefore, the charge

transport model given by the coupled fermion-boson system in eq. (2.1) is well complemented by

descriptions taking into account the explicit thermally activated and field dependent processes, i.e.,

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.6 Poole-Frenkel [Gill1972], disorder formalism [Bäss1981-2000], and multiple trapping and release

processes [Nool1977].

A model that might summarize most of the features stated in the previous descriptions is the “variable

hopping range” (VHR) by Vissenberg and Matters [Viss1998]. The VHR model predicts a thermal-

activated and gate voltage dependent mobility of the OFET device. 1

orgoB

2s

11c

3

3o

eff Tk2)VC(

)1()1(B)~2(e

−ςς

−−⎥⎥⎦

⎤

⎢⎢⎣

⎡

ε⎥⎦

⎤⎢⎣

⎡ς+Γς−Γα

ςπσ=μ g , (2.4)

here oσ is a prefactor, e is the electron charge, α~ is the effective overlap between hopping sites, cB

gives the percolation criterion of the organic layer, T/To=ς , ∫∞ −−=Γ0

x1x dxet)x( is the gamma

function. The thermal energy TkB is below the width oBTk of the exponential distribution of the

DOS, and the terms sC , gV , orgε correspond to the gate-dielectric capacitance per unit area, gate

voltage, and the dielectric constant of the organic material, respectively.

Considering a drop potential through the insulating dielectric layer and leaving eq.(2.4) as a simpler

power in dependence with the gate voltage, the drop potential yields [Meij2002]

( )kTgoeff VV −μ=μ , (2.5)

with oμ the none-field dependent mobility, TV the threshold voltage, and k an exponent given by )1( −ς in eq.(2.4). A typical value of the mobility in a-Si is about 1.1 cm2 V-1 s-1 comparable with the

hole mobility in several thin film organic-based transistors.

The thermally-activated field effect mobility, based on the Bolttzmann statistics and with a single trap

state following an exponential distribution, is proportional to [Ches2004]

⎟⎟⎠

⎞⎜⎜⎝

⎛ ε−≈μ

TkexpS

B

acteff , (2.6)

the factor S a fitting parameter.

If the material is free of traps, then the field effect mobility is described as meff T −∝μ being 1≤ m≤4.

This power dependence is produced via scattering of charge carriers with phonons. When the

temperature increases, the exponent m has been found to increase for holes and decrease for electrons.

Experimental results have shown larger mobilities for holes than for electrons. On the other hand, if

the organic solid has a relative trap concentration c, the mobility is given by [Wolf1997] 1

B

teffeff Tk

expc1)T,0c()T,c(−

⎥⎦

⎤⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛ ε+=μ=μ , (2.7)

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.7 the first factor accounts for the mobility without traps, and the second one is the trapping factor.

Results presented in eqs.(2.4-2.7) are a fairly good estimate of the mobility. Those models are lacking

a description of the field dependence of threshold voltage ( TV ). This has been experimentally

evidenced by Gomes et al. [Gome2004], and in this work has been proved by the correlated fitting of

the output characteristics in OFETs (chapter 5), and by calculation of the time-dependent capture cross

section based on QTS measurements (chapter 8).

Additionally, field-dependent QTS measurements on OFETs with open gate configuration (chapter 8)

proved that the trapping mechanism has several components which depend on the way the organic

material is bias-pulsed. Although the gate-dielectric/organic interface is affected by traps, the applied

electric field induces additional trapping sites in the organic material (chapters 5-8). Therefore, the

conductive channel itself and likewise its boundaries, i.e., electrodes, bulk organic material, should not

be excluded from the trapping phenomena.

2.5. Light-matter interaction

2.5.1. Raman spectroscopy

The interaction between photons and matter has gained an important position, fueling pure research

and applications at different industrial levels. Among the optical techniques, Raman spectroscopy

plays a very important role in determining the vibrational properties of matter. The vibronic spectrum

is a fingerprint of each substance. Indeed, any physical or chemical alteration of the investigated

structure modifies its vibrational spectrum. The previous section addressed the importance of phonon-

assisted processes of charge transport in organic semiconductors.

Historically, the Raman effect was predicted by Smekal [Smek1923] in 1923 and the first experiment

was performed by Sir C. V. Raman in 1928 [Rama1928]. Since then, innumerable experiments in

diverse research fields, i.e., physics, chemistry, astronomy, biology, and many others, have been

realised. Two years after the experimental discovery by C. V. Raman, an article by D. H. Andrew

about the Raman spectra in organic molecules and the relation between the involved atoms and the

formed bond nature [Andr1930] was reported, which can be considered as pioneering research in the

vibrational molecular branch.

The usefulness of the method is mainly due to its sensitivity and non-destructiveness. It is capable of

providing information on chemical identity [Gloc1943, Schm2006, Kuzm1988], charge states

[Auss1986, Fleu1967, , Shan1972], processes at interfaces [Mile2006, Paez2003-2004, Zahn2001],

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.8 and structural order in organic materials [Colo2003], among many other interesting molecular

features.

Raman spectroscopy [Berm1988, Brig1980, Kuzm1988, Schü2006, Sing2005] is based on the

inelastic interaction between light and matter. When light is scattered by any form of matter, the

energies of the majority of the photons are unchanged by the process, corresponding to elastic or

Rayleigh scattering. A minor number of photons, approximately one per million, are involved in the

inelastic scattering process of creation (Stokes process) or annihilation (anti-Stokes process) of

excitations within the medium. Therefore, the scattered photons will have an energy lower (Stokes) or

higher (anti-Stokes) than that of the incident light. Most routine Raman experiments use the red-

shifted Stokes peaks, because they are more intense at room temperature.

Figure 2.4. Spectrum of scattered light showing the Raman Stokes, Rayleigh, and Raman anti-Stokes

bands.

Figure 2.4 shows the Stokes, Rayleigh, and anti-Stokes processes, where a photon with energy excωh

and momentum exck is scattered by the creation or annihilation of an elementary excitation with

energy iωh and momentum iq . The scattered photon has an energy ω′h and momentum exck ′ . This

means each scattered photon in the Stokes component is associated with a gain in energy iωh by the

sample. Similarly, the sample loses energy iωh for each scattered photon in the anti-Stokes

component.

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.9 The conservation energy reads:

iexc ω±ω=ω′ hhh , (2.8)

where the minus (plus) sign is for the Stokes (anti-Stokes) process. Conservation of momentum

requires

iexc qkk ±=′ , (2.9)

Here, the plus (minus) sign is for the Stokes (anti-Stokes) process. Schematically, a Raman spectrum

with energy iωh looks like that displayed in Figure 2.4. There is a very strong component due to

elastic scattered photons, i.e., with energy excωh .

In chapter 3, the experimental Raman setup used during this investigation to characterize the

vibrational bands of organic / inorganic interfaces (chapter 4) and OFETs (chapters 6-8) is discussed.

2.5.2. Basic theoretical background on Raman spectroscopy

2.5.2.1. Classical description of the Raman effect

Classically, Raman scattering can be explained by molecular polarizability (α ) modulation. When a

molecule is subjected to the electric field ∑ ω=l

l )tjexp(oEE of a multi-energetic ( lhω )

electromagnetic beam, its dipole momentμ is given by

Εαμμ += )0( , (2.10)

with )0(μ the permanent dipole and Εαμ =)ind( the induced dipole moment by the electric field Ε ,

and α the polarizability, which can be expressed by the superposition due to elastic (Rayleigh) and

inelastic (Raman) contributions )Raman()Rayleigh( ααα += , (2.11)

with the first term on the right hand side being the induced-Rayleigh polarizability and the second the

Raman induced polarizability.

The matrix form of eq.(2.10) in Cartesian coordinate is written as follows:

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

ααα

ααα

ααα

+⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

μ

μ

μ

=⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

μ

μμ

z

y

x

zzyzxz

zyyyxy

zxyxxx

)0(z

)0(y

)0(x

z

y

x

E

EE

. (2.12)

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.10 The character of the polarizability tensor depends on the symmetry of the molecule. On the other hand,

it must be noted that the dipole moment and the electric field do not necessarily share similar unit

vectors (sect. 8.8).

Since Ε is a frequency dependent electric field, the coordinates of the atoms might be affected.

Therefore, instead of specifically using the atom position, the normal coordinates )t(qi are

considered. The number of normal coordinates equals the total number (N) of vibrational modes. For

small vibrations, )t(q i can be approximated by

)tcos(q)t(q ioi ω= , (2.13)

where oq is the maximum amplitude and iω is the vibrational frequency of i-th normal mode.

Assuming a monochromatic light source ( 1=l ) to excite the sample, and expanding the polarizability

in terms of the normal coordinates, the total dipole moment is then given by:

, (2.14)

where the different contributions to the induced dipole moment are commented on in the equation, the

term O(3) meaning third order derivatives. For the present research, only terms up to first order

derivatives are considered. If 0q 0i

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂α , then the corresponding -i-th- mode will not appear in the

Raman spectrum. A similar situation occurs for the infrared spectrum if 0q 0i

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂μ . It has been proved

that if the molecule has an inversion center of symmetry, the Raman active modes are not observed in

infrared, and those active in infrared will be absent from the Raman spectrum.

2.5.2.2. Quantum mechanical description of the Raman effect

The classical theory correctly predicts the frequency dependence for Rayleigh scattering and

vibrational Raman scattering. It also shows correctly the dependence of the Rayleigh scattering tensor

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.11 ( )Rayleigh(α ). On the other hand, the classical description has some limitations. It cannot be applied to

molecular rotations. The result for the induced polarizability ( )Raman(α ) of the Raman scattering tensor

is only partly correct.

The classical picture does not provide information about how the )Raman(α is related to the properties

of the molecule, in particular its characteristic transition frequencies, or to the frequency of the

incident radiation. Fortunately, the quantum mechanical model provides this information and bridges

the gap from the classical description to a complete treatment of all aspects of Raman scattering

[Long2002].

The electric dipole moment (μ ) given in eq.(2.10) can vary, due to the dynamics of the atomic motifs

comprising the molecule; consequently, vibronic transitions at any electronic state are expected. As an

example, in Figure 2.5(a) the vibronic energy levels of a diatomic molecule are depicted with the wave

functions on the intemolecular Morse potential, and are distributed for each electronic state with

energy Ee and the associated quantum numbers, i.e., principal (n), angular (j), orbital (l) and magnetic

( m~ ). Each vibrotional state is characterized by a frequency iω and normal coordinate modes iq . The

likely vibronic transitions at this electronic level are summarized by the infrared transitions shown in

Figure 2.5(b).

Another process is developed when the electronic state remains unmodified, thus the transitions are

developed between tne vibrational levels at the electronic ground state. If the scattered photon has the

same energy as the incoming radiation, the process is named Rayleigh scattering (cf. Figure 2.7(a)).

(a) (b)

Figure 2.5. Molecular Morse potential of the (a) ground electronic and bound vibrational states. (b)

Infrared activity described by transitions between vibrational states at a given electronic state

mljneE .

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.12 Otherwise, the normal Raman scattering is achieved and described by the Stokes (cf. Figure 2.6(b))

and anti-Stokes (cf. Figure 2.6(c)) scattering. There is an extension of the previous description when

the excitation energy is comparable to that of the first excited electronic state or above; then similar

scattering phenomena are developed, where the excitation attains the manifold electronic states and

consequently the resonant Raman effect is achieved as described in Figure 2.7.

It can be deduced from Figure 2.7 that the Raman bands (transitions) are most intense when the wave

function of the upper vibronic state resembles that of the ground state of the vibrational wave function.

The amplitudes of these vibronic transitions are proportional to the square of the probability of the

corresponding initial ( gor;nj − ) and final ( eor;mj −′ ) electronic excited vibronic states [Fran2000,

Joha2005].

(a)

(b)

(c)

Figure 2.6. Non-resonant Raman effect involving (a)

elastic light scattering or Rayleigh scattering, and the

inelastic processes: (b) Stokes scattering with the

scattered photon energy higher than the incoming one,

and (c) anti-Stokes where the scattered photon

possesses lower energy than that of the excitation

source.

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.13 In this scheme the vibronic wave function ( )r( on,jψ ) of the Morse potential is described by

)(L)2/exp()1nj2()nj(!n2njr)r( n2j2

nnj

oon,j ξξξ−+−Γ−β

==ψ −− , (2.15)

with )rexp()1j2( oβ−+=ξ , )(L n2j2n ξ− the associated Legendre polynomials, and the bonding

energy of the state nj expressed as 222n )nj)(m~2/(E −β−= h . The amplitude is proportional to

the transition 2

eogo mj;rnj;r ′−− , which can be readily integrated to determine the overlap

between vibronic states. Note that as long as the difference of the minimum position between the non-

excited and excited states is small, the vibrational transition probability increases. This is attributed to

allowed dipole transitions.

Figure 2.7. Resonant Raman

effect (a) vibronic transitions

pointing to the Raman-Stokes,

Raman-anti-Stokes, and Rayleigh

scattering.

For a molecular structure, those states are mainly between HOMO and LUMO. The intensity of the

Raman scattered light with polarization ρ of the excitation light is given by [Garo1976, Kürt1991]

24foII σρσρ αω∝ , (2.16)

where σoI is the intensity of the incoming light with polarizability σ , and inelastically scattered with

frequency fω . The polarizability in this frame is described in the form [Hass2004]:

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.14

∑ ⎥⎥

⎦

⎤

⎢⎢

⎣

⎡

Γ+ω+−+

Γ−ω−−=α

ρσσρ

σρf~ f~excff~f~excgf~ i)EE(

gˆf~f~ˆf

i)EE(

gˆf~f~ˆf

hhhh

μμμμ, (2.17)

where σρμ is the electric dipole operator and takes into account the electronic and nuclear

contributions. The states f , f~ , and g denote the final, intermediate, and initial states,

respectively. The corresponding energies are fE , f~E , and g~E . The factor f~Γ is a damping

coefficient. If the energy of the incident radiation ( excωh ) approaches the molecular transition energy

)EE( gf~ − , then the interaction will be in resonance.

The efficiency of the Raman signal is measured by the photon-scattering cross section [Joha2005] 2

n totgevibexc42

o

exc3

4oRS in

)n,0(f)n,1(fc6

p ∑ Γ+Ω−ω−ωπεωω′

=σhhhh

, (2.18)

with

vibexc ω−ω=ω′ , (2.19)

2/2/ 0,radvibphtot γ+γ+γ=Γ , (2.20)

0,radγ : interaction with the vacuum fluctuations expressed by Fermi’s golden rule

0,radγ : dephasing rate

geΩh : energy level separation

dipo lep = : dipole with elementary charge e and length dipl

vibΩ corresponds to the pure vibrational frequencies of the associated harmonic Hamiltonian

The typical order of RSσ is about 229 cm10− . The next section describes the procedure for enhancement

of the Raman signal. It involves the magnification of the cross section.

2.5.3. Surface-enhanced Raman scattering (SERS)

In the previous section, it was pointed out how small the Raman signal is in comparison to the

intensity of the incoming excitation laser energy. In order to enhance the detected Raman spectrum

intensity, several technological efforts have been targeted at circumventing this detection limit.

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.15 An interesting phenomenon which developed locally in the tested sample is observed when the

material is embedded in a metal-matrix environment. Although a perfect metal surface behaves as a

mirror for the electromagnetic radiation, this can be avoided by texturing the surface in a way that the

patterned layer will no longer be a quasi or perfect reflector for the light. Indeed, the modified surface

geometry substantially enhances the local electric field. The resulting local field produced in the

clusters is much higher, then the material in the vicinity is electronically and structurally altered by the

field, therefore, parallel signatures are induced, organic electric dipole ( )ind(μ ) coupling with the

metallic cluster plasmon, molecular polarizability ( )()Raman( ωα ), and charge transfer (CT) (cf. Figure

2.8) among others. The Raman signal is enhanced, usually by several orders of magnitude. The

phenomenon is referred to as surface-enhanced Raman spectroscopy SERS [Aroc2006, Mosk2002].

Figure 2.8. Examples of SERS from a (a) pentacene (30 nm)-based OFET, and a (b) In (15 nm)/

PTCDA (15 nm)/S-GaAs heterostructure.

The SERS signal is divided into two principal contributions, one from the interface formed between

the first molecular layer and the metal, called chemical enhancement or first layer effect. The other

one is a long-range contribution mediated by the huge local electric field, referred to as

electromagnetic enhancement.

The magnification factors (based on this research) are estimated to be 101-102 for the chemical

contribution or first monolayer effect [Pers2006], and 102 or higher for the electromagnetic

contribution. Therefore, the cross sections are magnified from ca. 10-29 cm2 to 10-17-10-14 cm2

[Joha2005, Knei1997, Otto2001]. Due to this pronounced enhancement of the photon cross section,

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.16 SERS is widely used in analytical science and technology [Aroc2006, Cao2005, Mosk2002, Vo2004].

Detailed theoretical descriptions about SERS are found elsewhere [Aroc2006, Joha05, Knei1997,

Otto2001, Pers2006]. As an example, figs. 2.9(a) and 2.9(b) illustrate the Raman spectra of a

pentacene (30 nm) layer forming the channel in an OFET structure. It should be noted that the Raman

signal of molecules on the gate dielectric is much lower that that produced by molecules deposited on

the Au electrodes; the enhancement factor for this experiment was about 31.

The panel given in Figure 2.8(b) consists of the Raman spectra from bare PTCDA (15 nm) deposited

on S-GaAs and an In layer of 15 nm deposited on the previous structure; the Raman bands are

enhanced by a factor of about 102. The metallic overlayer produces a breakdown of the molecular

symmetry, since the activated modes were originally infrared active modes of the bare molecular

layer. Additionally, a well-defined charge transfer (CT) is developed. It can be observed that both

structures, pentacene (30 nm)-based OFET and In/PTCDA, have a markedly different base line,

depending on whether the molecular layer is on the inorganic substrate or is interfacing with the metal

electrode.

2.5.4. Combined Raman spectroscopy with electric fields

The application of an electric field affects the optical absorption and the scattering of electromagnetic

(EM) radiation by optical phonons; both have been observed in inorganic semiconductors and

aromatic molecules [Auss1986, Fleu1967, Shan1972]. These changes result from several effects of the

modulating field: First it lifts the degeneracies of both the phonon and the electronic states, and second

it affects the symmetries of the corresponding wave functions. As a consequence, the Raman-active

phonon modes can shift energetically, and, due to changes of the selection rules, new scattering

transitions can contribute to the Raman signal. When an external electric field is applied to a molecule,

the electronic charges are redistributed between different atomic sites, resulting in induced dipole

moments; therefore, the energy potential described in section 2.5.2.1 is distorted. Thus the Frank-

Condon transitions are subjected to the new molecular organization, provoking changes in the Raman

cross section.

In chapter 7, the experimental results of the Raman signal with applied electric fields in organic-based

field effect transistors (OFETs) are described. Moreover, it is theoretically described how the applied

field (E) in a given direction produces modifications of the molecular dipole moment, resulting in

effective contributions that in general do not follow the direction of the field (eq. (2.12)).

The nature of the applied field can be a time dependent or a static perturbation. Both fields were

applied to the OFETs (chapter 6-8). In addition, in chapter 7, the experimental and theoretical

measurements of combined Raman spectroscopy with electric fields and the effect on the structural

properties of the pentacene layer in OFETs are described. Furthermore, after the removal of the

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.17 applied bias, the molecular layer follows a relaxation process, which was demonstrated by the Raman

bands relaxation.

The decay time was confirmed by charge transient spectroscopy (QTS), revealing novel phenomena

never seen in pentacene-based FETs [Thur2006], addressed by the authors as “anomalous QTS”. On

the other hand, the calculated cross sections from QTS are comparable with the photon cross section,

due to chemical enhancement in SERS. From both techniques, it has been inferred that the application

of the electric field induces electric dipole states at the organic material.

The theoretical simulations of the vibronic properties in the present investigation were done within the

framework of the density functional theory (DFT), using the package Gaussian 98 [Gaus1998] at the

B3LYP level [Beck1993] with 3-31G basis sets having Gaussian-like wave functions [Helg2000]. The

fundamentals of the DFT are reported in detail elsewhere [Koch2002, Parr1989].

References

[Andr1930] D. H. Andrews, “The relation between the Raman spectra and the structure of organic molecules”, Phys. Rev. 36, 544-554 (1930).

[Auss1986] F. R. Aussenegg and M. E. Lippitsch, “On the electric field dependence of Raman scattering”, J. Raman Spectrosc. 17, 45-49 (1986).

[Aroc2006] R. Aroca, “Surface-enhanced vibrational spectroscopy”, Wiley, Chichester, 2006. [Bao2006] “Organic Field-Effect transistors V”, edited by Z. Bao and D. J. Gundlach, Proc. SPIE 6336, Washington

2006. [Bäss1981] H. Bässler, “Localized states and electronic transport in single component organic solids with diagonal

disorder”, phys. stat. sol. (b) 107, 9-54 (1981) [Bäss2000] “Charge transport in random organic semiconductors” in “Semiconducting polymers: Chemistry, physics and

engineering”, chap. 12. Edited by G. Hadzüoannou and P. F. van Hutten, Wiley-VCH, Weinheim 2000. [Beck1993] A. D. Becke, “Density-functional thermochemistry. III. The role of exact exchange”, J. Chem. Phys. 98,

5648–5652 (1993). [Berm1988] D. Bermejo, “Itroduction to Raman spectroscopy”, OEA, Spain, 1988 (in spanish). [Bowe2002] D. I. Bower, “An introduction to polymer physics”, Cambridge University Press, UK 2002.

[Brig1980] E. B. Wilson, Jr., J. C. Decius and P. C. Cross, “Molecular vibrations: The theory of Infrared and Raman Vibrational Spectra”, Dover publications, Inc. New York, 1980.

[Cao2005] L. Cao, P. Diao, L. Tong, T. Zhu, and Z. Liu, “Surface-enhancement Raman scattering of p-Aminothiophenol on a Au(core) / Cu (shell)”, Chem. Phys. Chem. 6, 913-918 (2005).

[Care2004] F. Carey, “Organic chemistry”, 5th Edition, The McGraw-Hill Co., 2004. [Ches2004] R. J. Chesterfield, J. C. McKeen, Ch. R. Newman, P. C. Ewbank, D. A. da Silva Filho, J. L. Brédas, L. L.

Miller, K. R. Mann, and C. D. Frisbie, “Organic thin film transistors based on N-alkyl perylene diimides: charge transport kinetics as a function of gate voltage and temperature”, J. Phys. Chem. 108, 19281-19292 (2004).

[Colo2003] P. Colomban, “Raman analysis and smart imaging of nanophases and nanosized materials”, Spectroscopy Europe 15 (6), 8-16 (2003).

[Cruz1992] M. G. Cruz, “Path integral approach to the small polaron model in the atomic limit”, Am. J. Phys. 60, 1127-1131 (1992).

[Duke1980] C. B. Duke and L. B. and Schein, “Organic solids: is energy-band theory enough?”, Phys. Today. 33, 42-48 (1980).

[Feyn1955] R. P. Feynman, “Slow electrons in a polar crystal”, Phys. Rev. 97, 660-665 (1955). [Fleury1967] P. A. Fleury and J. M. Worlock, “Electric-field-induced Raman effect in paraelectric crystals”, Phys. Rev.

Lett. 18, 665 - 667 (1967). [Fran2000] A. Frank, A. L. Rivera, and K. B. Wolf, “Wigner function of Morse potential eigenstates” ,Phys. Rev. A 61,

054102 (2000). [Fröh1950] H. Fröhlich, H. Pelzer, and S. Zieuau, Philos. Mag. 41, 221 (1950). [Fröh1954] H. Fröhlich, “Electrons in lattice field”, Adv. Phys. 3, 325-361 (1954). [Garo1976] M. Garozzo and F. Galluzzi, “A comparison between different approaches to the vibronic theory of Raman

intensities”, J. Chem. Phys. 64, 1720-1723 (1976).

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.18

[Gaus1998] Gaussian 98 (Revision A.1), M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G., Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A., Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M. W. Gill, B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. Head-Gordon, E.S. Replogle and J.A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.

[Gill1972] W. D. Gill, “Drift mobilities in amorphous charge-transfer complexes of trinitrofluorenone and poly-n-vinylcarbazole”, J. Appl. Phys., 43, 5033-5040, (1972).

[Gloc1943] G. Glockler, “The Raman effect”, Rev. Mod. Phys. 15, 111-173 (1943). [Gome2004] H. L. Gomes, P. Stallinga, F. Dinelli, M. Murgia, F. Biscarini, D. M. de Leeuw, T. Muck, J. Geurts, L. W.

Molenkamp, and V. Wagner, “Bias-induced threshold voltages shifts in thin-film organic transistors”, Appl. Phys. Lett. 84, 3184-3186 (2004).

[Gutm2004] F. Gutmann and L. E. Lyons, “Organic semiconductors”. Wiley, N. Y. (1967). [Hass2004] S. Hassing and E. Nørby, “The correct form of the Raman scattering tensor”, J. Raman Spectrosc. 35, 87-90

(2004) [Ishii1999] H. Ishii, K. Sugiyama, E. Ito and K. Seki, “Energy level alignment and interfacial electronic structures at

organic/metal and organic/organic Interfaces” Adv. Mater. 11, 605-625 (1999). [Helg2000] T. Helgaker, O. Jørgensen, and J. Olsen, “Molecular electronic-structure theory”, (Hohn Wiley & Sons,

England, 2000). [Joha2005] P. Johansson, H. Xu, M. Käll, “Surface-Enhanced Raman Scattering and fluorescence near metal

nanoparticles” ,Phys. Rev. B 72, 035427 (2005). [Kafa2005] “Organic Light-Emitting materials and devices IX”, edited by Z. H. Kafafi and P. A. Lane, Proc. SPIE 5937,

Washington 2005. [Kao1981] K. C. Kao and W. Hwang, “Electrical transport in solids”. Pergamon Press, Oxford (1981).

[Knei1997] K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, and M. S. Feld, “Single molecule detection using Surface-Enhanced Raman Scattering (SERS)”, Phys. Rev. Lett. 78, 1667–1670 (1997).

[Koch2002] W. Koch, M. C. Holthausen. “A chemist’s guide to density functional theory”, Wiley-VCH, Germany, 2002. [Kuzm1988] H. Kuzmany, “Solid-State Spectroscopy, an introduction” (Springer-Verlag, Berlin Heidelberg, 1988).

[Kürt1991] J. Kürti and H. Kuzmany, “Resonance Raman scattering from finite and infinite polymer chains”, Phys. Rev. B 44, 597-613 (1991).

[Lee1952] T. D. Lee and D. Pines, “The motion of slow electrons in polar crystals”, Phys. Rev. 88, 960 - 961 (1952). [Long2002] D. A. Long, “The Raman effect: A unified treatment of the theory of Raman scattering by molecules”, JOHN

WILEY & SONS, LTD, Chichester, 2002. [Ménd2006] H. Méndez, , “Organic modification of Metal / Semiconductor contacts”, PhD Dissertation, http://archiv.tu-

chemnitz.de/pub/2006/0124/index.html TU Chemnitz (2006). [Meij2002] E. J. Meijer, C. Tanase, P. W. M. Blom, E. van Veenendaal, B.-H. Huisman, D. M. de Leeuw, and T. M.

Klapwijk, “Switch-on voltage in disordered organic field-effect transistors”, Appl. Phys. Lett. 80, 3838-3840(2002).

[Mile2006] A. G. Milekhin, R. Meijers,T. Richter, R. Calarco, S. Montanari, H. Lüth, B. A. Paez-Sierra, and D. R. T. Zahn, “Surface enhanced Raman scattering by GaN nanostructures obtained by bottom-up and top-down approach”, J. Phys.Cond.Matt. 18, 5825-5834 (2006).

[Mosk2002] M. Moskovits, Ll. Tay, J. Yang, T. Haslett, “SERS and the single molecule”, in “Optical properties of nanostructured random media”, Series: Topics in Applied Physics , Vol. 82, 215-226, Edited by Shalaev, Springer Berlin Vladimir M. 2002.

[Nool1977] J. Noolandi, “Multiple-trapping model of anomalous transit-time dispersion in a-Se”, Phys. Rev. B 16, 4466-4473 (1977).

[Otto2001] A. Otto, “Theory of first layer and single molecule surface enhanced Raman scattering (SRS)”, phys. stat. sol. (a) 188, 1445-1470 (2001).

[Paez2003] B. A. Paez, M. Bartzsch, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Combined electrical and Raman characterization of C60 based organic field effect transistors”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 63 (2003).

[Paez2003] B. A. Paez, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Interaction of metals with perylene derivatives as a model system for contact formation in OFET structures”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 210 (2003).

[Paez2004] B. A. Paez, G. Salvan, S. Silaghi, R. Scholz, T. U. Kampen and D. R. T. Zahn, “Raman monitoring of In and Ag growth on PTCDA and DiMe-PTCDI thin films”, Appl. Surf. Sci. 234, 168-172 (2004).

[Parr1989] R. G. Parr and W. Yang, “Density-functional theory of atoms and molecules”, (Oxford University Press, New York1989).

[Pers2006] B. N. J. Persson, K. Zhao, and Z. Zhang, “Chemical contribution to surface-enhanced Raman scattering”, Phys. Rev. Lett. 96, 207401 (2006).

[Pire1974] J. J. Pireaux, J. Riga, R. Caudano, J. J. Verbist, J. M. Andre, J. Delhalle and S. Delhalle, “Electronic structure of fluoropolymers : Theory and esca measurements”,J. Electron Spectrosc. Relat. Phenom. 5, 531-550 (1974).

[Pire1984] J. J. Pireaux and R. Caudano, “Experimental picture of the band structure formation in a solid”, Am. J. Phys. 52, 821-826 (1984).

[Rama1928] C. V. Raman and K. S. Krishnan, “A new type of secondary radiation”, Nature 121, 501–502 (1928).

Chapter 2 B. A. Paez-Sierra, Fundamentals of molecular structures… 2.19

[Schm2006] M. Schmitt and J. Popp, “Raman spectroscopy at the beginning of the twenty-first century“, J. of Raman Spectroscopy 37, 20-28 (2006).

[Schü2006] C. Schüller, “Inelastic light scattering of semiconductor nanostructures, fundamentals and recent advances”, chap. 4., Springer Tracts in Modern Physics Volume 219, Berlin Heidelberg, 2006.

[Shan1972] M. L. Shand, W. Richter, E. Burstein, and J. G. Gay, “Electric field induced Raman scattering in CdS”, J. Nonmmetals 1, 53-62 (1972).

[Sing2005] A. Singha, P. Dhar, and A. Roy, “A nondestructive tool for nanomaterials: Raman and photoluminescence spectroscopy”, Am. J. Phys. 73, 224-233 (2005).

[Smek1923] A. Semkal, “Zur Quantentheorie der Dispersion”, Naturwiss. 11, 873-875 (1923). [Thur2006] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, “Anomalous charge relaxation in channels of

pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 203, 2326-2340 (2006).

[Viss1998] M. C. J. M. Vissenberg, M. Matters, “Theory of the field-effect mobility in amorphous organic transistors”, Phys. Rev. B 57, 12964-12967 (1998).

[Well1996] G. Wellein, H. Röder, and H. Fehske, “Polarons and bipolarons in strongly interacting electron-phonon systems”, Phys. Rev. B 53, 9666–9675 (1996).

[Wolf1997] U. Wolf, H. Bässler, P. M. Borsenberger and W. T. Gruenbaum, “Hole trapping in molecularly doped polymers”, Chem. Phys. 222, 259-267 (1997)

[Zahn2001] D. R. T. Zahn, “Raman studies of molecular thin films”, phys. stat. sol. (a) 184, 41-50 (2001).

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.1

Chapter 3

Experimental techniques, materials, and algorithms

In this chapter, the investigated molecular structures and the main experimental characterization

techniques involved to scope the organic/inorganic structures are discussed. The most frequently used

experimental methods to characterize the samples during this research were Raman spectroscopy,

scanning Raman spectroscopy, current voltage (I-V) characteristics, and charge transient spectroscopy

(QTS).

The molecular structures consist of perylene derivatives named 3,4,9,10-perylene tetracarboxylic acid

dianhydride (PTCDA) and 3,4,9,10- N-N’-dimethyl-3,4,9,10-perylenetetracarboxylic diimide (DiMe-

PTCDI). These organic molecules were evaporated by molecular beam deposition (MBD) under UHV

conditions and deposited on S-GaAs substrates. Indium or magnesium were deposited on these organic

molecules and they were characterized in situ by surface-enhanced Raman spectroscopy (SERS)

experiments.

The organic molecules to form the active layer of the FET structures were pentacene and C60; minor

work was done on the C60–based devices. The data evaluation was carried out primarily in special

routines written in MatLab [MATL2003], with a particular code based on a consistent recursivity of

initial boundary conditions developed to achieve efficient data evaluation. Furthermore, in some

theoretical modeling involving numerical integration, a routine based on the Gauss quadrature was

constructed. As a result, high accuracy and convergence speed in comparison with standard

geometrical integration methods were achieved.

3.1. Combined Raman spectroscopy and electrical characterization setups

In chapter 2, the fundamentals of Raman spectroscopy were given. In this section, the experimental

setup used for this research is outlined.

The Raman spectrometer employed during this work was a triple monochromator Dilor XY 800 model

from Dilor. The system is equipped with a Peltier-cooled CCD (256x1024 pixels) camera for

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.2 multichannel detection in a back scattering configuration geometry. This allows the setup to be used in

macro- and micro-configuration.

The samples can be excited with different monochromatic laser lines, provided by Ar+, Kr+ or HeCd

gas-based lasers. During the macro-Raman experiments, the laser power density on the samples was

set to 25.4 Wcm-2 for C60 (to avoid photopolymerization), 70 Wcm-2 for pentacene active layers in

FETs, and 140 Wcm-2 for PTCDA, DiMe-PTCDI and related interlayers.

The spectral resolution ranged from 2 cm-1 to ~ 3.5 cm-1

as determined from the FWHM of the

Rayleigh line.

For the in situ measurements the angle between the incident and scattered light is fixed by the position

of the corresponding UHV windows. The sample surface was always oriented parallel to the collection

window, i.e. the scattered light was analysed in the direction parallel to the normal of the sample. The

incidence angle was 30°. The diameter of laser light focused on the sample was ~ 300 μm.

Additionally, in this research the polarization of the incoming and scattered beams is given according

to the Porto coding, which takes into account the propagation of the incoming and scattered beams

together with the polarization, i.e., z(xy)z’: means the incident light is propagating along the z axis and

then back scattered (z’) in the same axes; xy in the brackets labels the electric field polarizations with

respect to a fixed reference plane holding the sample.

Figure 3.1. Experimental setup for combined Raman spectroscopy, scanning Raman spectroscopy, and

current-voltage (I-V) characteristics measurements.

The incoming electric field is polarized in the “x” direction and the scattered electric field in the “y”

direction. The codes z(yy)z’, x(zx)y, x(zz)y, etc., are set up in a similar manner.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.3 The samples investigated in situ by Raman spectroscopy or in combination with electric fields were

placed in an analysis chamber with a base pressure of 2x10-10 mbar and with the capability of

performing organic beam deposition, metal evaporation from Knudsen cells, on-line spectroscopy and

electrical characterization of the structure formation, as sketched in Figure 3.1. The current-voltage

characteristics were obtained with a Keithley 238 high current source/meter unit controlled by a

computer with a code written in LabView [LabV2004] or MATLAB [MATL2003], as depicted in

Figure 3.1.

3.2. Sulphur passivation of GaAs(100)

The substrates were Si doped n-GaAs(100) wafers acquired from Freiberger Compound Materials

GmbH (FCM) with a doping concentration of ~ 318 cm108.1 −× estimated by infrared spectroscopy

measurements.

The passivation procedure of GaAs(100) surfaces was done in a S2Cl2 solution [Salv2003]. The

procedure is described in Figure 3.2.

Figure 3.2. Substrate passivation:

ex situ chemical treatment and in

situ annealing and material

deposition.

3.3. Materials and structures

The structures are divided into two types, one series intended for metal organic interface formation

investigations (structures I), the others consisting of organic molecules deposited on field effect

devices to form organic-based field effect transistors (OFETs) and referred to as structures II.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.4 3.3.1. Structures I: perylene derivatives thin films capped by metallic

overlayers of indium and magnesium

The molecular structures are described in fig 3.3. Additionally, Figure 3.3 indicates the group

symmetry, together with the character of the vibrational bands, i.e., Raman active, infrared active or

silent modes. In order to estimate the appropriate working energy for the resonant Raman experiments,

the absorption spectra of both molecules are shown in Figure3.3(c). Further description on crystal

structures and the molecular alignment on similar substrates are reported elsewhere [Ferg2006,

Frie2003, Gavr2006, Kobi2004, Ménd2006, Salv2003]. Details on the organic/inorganic interface

formation and the in situ investigation by Raman spectroscopy are presented in chapter 4.

3.3.1.1. Molecular beam deposition (MBD) and metal evaporation

The molecular and metallic materials were thoroughly degassed one by one in the vacuum chamber

for at least 24 h prior to deposition and were evaporated from two separate Knudsen cells at 280 °C

and 270 °C for PTCDA and DiMe-PTCDI, respectively. These parameters led to a growth rate of

about 0.3 nm/min for both molecules. The metals were evaporated at 830 °C for indium and 360 °C

for magnesium, with an evaporation rate of 2 nm/min.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.5

PPTTCCDDAA

3,4,9,10- PeryleneTetraCarboxylic DiAnhydride

E0-0 = 2.21 eV

CC2244HH88OO66

Symmetry D2h

Raman active: 19Ag+18B1g+10B2g+7B3g

IR active: +10B1u+18B2u+18B3u

Silent: + 8Au

108 internal vibrations

(a)

DDiiMMee--PPTTCCDDII

3,4,9,10- PeryleneTetraCarboxylic DiImide

E0-0 = 2.14 eV

CC2266HH1144OO44NN22

Symmetry C2h

Raman active:

44Ag+22Bg

IR active:

+23Au+43Bu

132 internal vibrations

(b)

(c)

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.6 Figure 3.3. Molecular structure of perylene derivatives with the associated symmetry group of (a)

PTCDA and (b) DiMe-PTCDI [Kobi2004, Salv2003]. (c) UV-vis absorption spectra of organic layers

deposited on quartz.

The film thickness was monitored in situ using a quartz crystal microbalance positioned near the

substrate. Ex situ thickness calibration was performed by atomic force microscopy (AFM).

Reevaporation of the molecules during metal deposition and laser exposure during the Raman

spectroscopy measurements was not achieved. The stepwise deposition of submonolayer metal

coverage was performed in short intervals and with a significant off-time of several minutes between

subsequent depositions. Furthermore, the substrate temperature was continuously monitored. In

complement, the double, or, in some cases, triple collection of the vibronic spectrum confirmed the

structure integrity.

This complies with recent published results related to experimental investigations of thermal

desorption (TPD) of PTCDA deposited on Ag(111) by Umbach and colleagues [Zou2006], who

revealed evaporation temperatures of the organic layer from metallic substrates at (538 ± 30) K for the

first monolayer in intimate contact with the metallic surface, while for the second monolayer it was

about (494 ± 30) K, as could be deduced from their publication. This also proves the physical

properties of the structures discussed in this work.

3.3.2 Structures II: pentacene and C60 organic molecules as active layers in

field effect devices

3.3.2.1. Molecular structures: pentacene and C60

The model molecules to construct organic-based field effect transistors were pentacene and C60. A

summary of the symmetry group, vibrational activity, UV-vis absorption spectra [Kolo2005], and

working energies for resonant Raman experiments is provided in Figure 3.4.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.7

Pentacene C22H14

E0-0 = 1.84 eV

Symmetry D2h

102 internal vibrations

Raman active: g3g2g1g1 B17B7B9A18 +++

IR active: u3u2u1 B9B17B17 ++

Silent: uA8

(a)

(c)

CC6600

E0-0 = 2.02 eV

Symmetry : Icosahedron

Raman active: )eneracy(degA2H8 gg +

IR active: )eneracy(degF4 u1

Silent:

uuu2ugg2g1 H73G63F5AG6F4F3 ++++++

176 Internal modes most of them with degeneracy

(b)

(d)

Figure 3.4. Molecular structure of (a) pentacene (C22H14) and its 102 internal vibrational modes divided

into Raman active, IR active, and silent bands, belonging to the D2h symmetry group [Ross2002], (b)

Fullerene C60, which belongs to the symmetry group of the truncated icosahedron [Kost1994]. (c), (d)

Absorption spectrum of a 30 nm pentacene and 30 nm C60 film deposited on quartz, respectively

[Kolo2005]. The excitation energies for resonant Raman spectroscopy measurements are indicated on the

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.8 spectra.

3.3.2.2. Field effect structures

The field effect devices were provided by Prof. H. von Seggern from TU Darmstadt. The field effect

substrates shown in Figure 3.5(a) were fabricated using heavily doped n-type silicon substrates (3–5

Ωcm resistivity) covered by a a layer of high quality thermally grown SiO2 about 285 nm thick, acting

as gate electrode and gate oxide, respectively. A thin chromium adhesive layer was deposited on the

entire oxide surface, before a 50 nm Au layer was deposited.

The Au source and drain electrodes were photolithographically structured. They were configured as

interdigitated fingers with a channel length of 5 μm and a channel width of 2 cm [Hepp2003]. Similar

substrates with single channels consisting of two parallel contacts for drain and source were provided

by C. Pannemann from the group of Prof. U. Hilleringmann in Paderborn University. A detailed

description of the substrates is presented elsewhere [Pann2004].

Width = 5 μm

Length = 2 cm

dSiO2 = 285 nm

(a)

(b)

Figure 3.5. Field effect structures used for the fabrication of organic-based field effect transistors

(OFETs). (a) Interdigitated structures and (b) single channel structures.

3.3.2.3. Molecular beam deposition: pentacene and C60

Prior to the molecular beam deposition, the material pentacene was purified, while C60 from Tokyo

Kasei Kogyo Co. Ltd. was acquired with a purity as high as 99.9%. The materials were degassed

during 24 h in Knudsen cells in the analysis chamber. Before the organic deposition, the cleanliness of

the FET substrates was electrically tested by measuring the current-voltage characteristics in order to

detect possible leakage currents. Additionally, the Raman bands of the prominent spectral region of

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.9 the involved organic material was measured. During independent experiments under UHV conditions,

pentacene and C60 were evaporated at 155 oC and at 400 oC, respectively; the respective evaporation

rates were 0.65 Å / min and 1.7 Å / min. The layers were deposited onto field effect structures as

depicted in Figure 3.6. After each molecular coverage, the Raman bands and the channel current were

measured (chapter 6).

Figure 3.6. Field effect structure and formation of

the channel by organic molecules of pentacene or

C60. The polarity of the Vg and Vd depends on the

charge carrier type i.e. n or p (chapter 5).

The resulting monolayers (below 5 nm) deposited on the gate dielectric built up the organic channel.

Though further molecular coverage up to about 30 nm did not significantly contribute to the drain

current (chapter 6) [Paez2005], it helped to improve the protection of the working device against

atmospheric conditions (chapters 6-8).

3.4. Charge transient spectroscopy (QTS)

The interface formation between organic/inorganic or organic1/organic2 plays a decisive role in charge

injection into and transport through the device. Particularly the metal/organic interfaces differ

significantly from their inorganic counterparts, since the depletion depth is larger than the organic

layer thickness. Therefore, the Fermi level in the organic material and the charge injection barriers

basically depends on the interface offset. These factors might make it difficult to investigate charge

states at the interface by means of capacitance-based spectroscopies.

3.4.1. Shallow and deep level states in semiconductors

Shallow states in a semiconductor are located in the band gap and are close to the band edges of the

valence (HOMO) or conduction (LUMO) band in an inorganic (organic) semiconductor. These states

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.10 introduce minor perturbations in the lattice and are usually ionized at room temperature, since their

energy is comparable to kT ( meV25≈ ) or much lower in reference to the band gap of the material.

A procedure to investigate shallow traps entails tuning the Fermi level towards the ionized shallow

level. Thus the trap center becomes neutral and can be ionized with a suitable bias voltage (sect.

3.4.3). Another type of trap level is referred to as a deep level. These states are deeper in the band gap

and away from the band edges, which makes them more localized.

In Figure 3.7, the trapping/detrapping process for a single deep level is summarized. For simplicity,

the recombination processes are excluded. The first process (a) is a capture of an electron and

described by the capture coefficient nc . Electrons in the conduction state might migrate to the ionized

trap. The second process (b) consists of emission of the electron from the trap center towards the

conduction band and is described by the coefficient ne .

It should be noted that in the situation described by (a) there is a density of carriers that can fill the

trap, while the emission involves only one electron with a spin up or down (double degeneracy of the

trap). A similar situation is developed for hole-related traps where the capture (c) and emission (d) are

described by the coefficients pc and pe , respectively.

Figure 3.7. Band diagram for a semiconductor

with a single deep level trap (recombination

processes between HOMO-LUMO or intermediate

states are excluded).

The deep level state can switch between a filled or an occupied state. When it is occupied by an

electron or a hole, the state is named Tn (donors) or Tp (acceptors), respectively. The total number

of deep levels occupied by electrons and holes is

TTT pnN += . (3.1.)

A trap is only occupied by an electron or a hole and will be in a neutral state. The opposite happens

when an external electric field induces dipole sites in the structure; then anion-dipole or cation-dipole

related traps are formed. This was experimentally demonstrated with pentacene-based field effect

transistors in the present work and is discussed in chapter 8.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.11

The free charge carriers, either for the electron density (n) or hole density (p), can be increased by the

released carriers from the trap centers or decreased due to the capture process. Therefore, the electron

or hole time rate process is described by

⎥⎥⎦

⎤

⎢⎢⎣

⎡

+−−+

+−−+=⎥

⎦

⎤⎢⎣

⎡−−

=⎥⎥

⎤⎢⎢

⎡

TpnTTnp

TnpTTpn

TpTp

TnTn

p)enc()pN)(epc(

n)epc()nN)(enc(pncpenpcne

pn

tdd

. (3.2)

In general, all terms in eq.(3.2) are time-dependent variables. If one type of deep level is dominant,

i.e., Tn or Tp , then the electron or rate balance equations are solved independently. In addition, if n

and p are time independent quantities, the solutions of the balance equations given in eq.(3.2) yield

⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢

⎣

⎡

⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛

τ−−

+++

++⎟

⎠⎞

⎜⎝⎛

τ−

⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛

τ−−

+++

++⎟

⎠⎞

⎜⎝⎛

τ−

=⎥⎦

⎤⎢⎣

⎡

texp1pcence

N)epc(texp)0(p

texp1pcence

N)enc(texp)0(n

)t(p)t(n

ppnn

TnpT

ppnn

TpnT

T

T , (3.3)

where )0(n T and )0(pT are the initial situation (t = 0) of trap densities occupied by electrons or

holes, respectively, and the time constant τ is given by

pcence1

ppnn +++=τ , (3.4)

here the capture coefficients p,nc depend on the capture cross section p,nσ of the deep level and the

charge carrier thermal velocity ( thv ). Therefore,

thp,np,c vc σ= . (3.5)

Assuming one type of deep level for electrons ( pn ee >> ) or holes ( np ee >> ), the balance time

process given in eq.(3.3) is reduced to

⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢

⎣

⎡

⎟⎟⎠

⎞⎜⎜⎝

⎛

τ−

⎟⎟⎠

⎞⎜⎜⎝

⎛τ

−

≈⎥⎦

⎤⎢⎣

⎡

pT

nT

T

T

texpN

texpN

)t(p)t(n

. (3.6)

nτ and pτ being the relaxation time constants of electron- and hole-related traps.

Considering the Fermi statistics for the trap density distribution and from the equilibrium conditions

(no external fields) in eq. (3.3), one finds the relation between the emission rate and the trap energy

level,

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.12

⎟⎠⎞

⎜⎝⎛ −−

σ=

τ=

⎟⎠⎞

⎜⎝⎛ −−

σ=

τ=

kTEEexp

gv N

1 e

kTEEexp

gv N 1 e

HTthpL

pp

TLthnL

nn

. (3.7)

Here LN is the effective density of states in the LUMO, g is the degeneracy factor of the deep level,

and kT is the thermal energy.

Assuming the constant effective mass approximation, eqs.(3.7) are simplified by

⎟⎟⎠

⎞⎜⎜⎝

⎛ Δ

σγ=⎟

⎠⎞

⎜⎝⎛ −

σγ=τ

⎟⎠⎞

⎜⎝⎛ Δ

σγ=⎟

⎠⎞

⎜⎝⎛ −

σγ=τ

−

−

kTE

exp1kT

EEexp1 T

kTEexp1

kTEEexp1 T

pT

pp

HT

pp

2p

nT

nn

TL

nn

2n

, (3.8)

To determine the trap energy TEΔ level, it is necessary to perform the experiment at different

temperatures, and the slope of the plot )Tln( 2p,eτ as a function of T/1 will give the )p,n(TE −Δ

value. The intercept on the )Tln( 2p,eτ axis gives the information related to the capture cross section

( p,nσ ) through ))/(1ln( p,np,n σγ .

3.4.2. The charge transient spectroscopy (QTS) technique

This section describes the spectroscopy of defects in the channel of pentacene based OFETs via bias

pulse excitation of the organic semiconductor between source and drain electrodes, followed by

processing of the transient charge Q(t) by means of a filter of time constants (rate window concept).

The rate window concept is based on the assumption of first order (exponential) kinetics of either

capacitance or charge decay after the excitation.

Here a transient charge decay is assumed as Q(t) = Q0 exp(-t/τ), where Q0 is the full released charge

and τ the time constant of the relaxation (eq.(3.7)). Application of the rate window concept means

designing a filter with a response peaking at a specified (programmed) delay ti that is related to τ

through a constant. To transform the exponential decay to a peak on the time scale ti three samples of

Q(t) are used at delays ti, 2ti, 4ti and combined in order to obtain a weighted sum [Thur1994, 2005-

2006],

)t4(Q2/1)t2(Q2/3)t(Q)t(Q 111 +−=Δ . (3.9)

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.13 For comparison, the double and triple boxcar integrators are compared in Figure 3.8 to measure deep

levels, figs. 3.8(a) and 3.8(b), respectively. The accuracy of the latter method is higher than that of the

double boxcar, since it samples the emptying signal at three different times.

Actually, to obtain a peak on the time scale ti (cf. Figure 3.8 (b)), it is sufficient to combine and weight

two sampled values only, as demonstrated by Kirov and Radev [Kirov1981], and Farmer et al.

[Farm1981]. After adding the third sample (eq. (3.9)), an improvement in selectivity as well as

immunity against any constant or linear component in Q(t) is achieved [Thur1994]. When illustrating

the performance of the filter defined by eq. (3.9), it is convenient to set t1 = ti and replace the time axis

by x = log10(t1) – Figure 3.9. Three QTS spectra belonging to three different time constants are shown

while scanning the delay t1.

(a) (b)

(c) (d)

Figure 3.8. Charge transient spectroscopy (QTS) based on the rate window concept. (a) Sample

wiring, and (b) applied bias pulse. (c) Triple boxcar integrator and; (d) output signal displayed as a

function of rate window time, the QTS maximum coincides with the relaxation time constant τ of the

trap (see sample spectra below for three different time constants) [Thur1994, 2005-2006].

It must be noted that both the FWHM and the peak height ΔQm are invariant against τ; the response

depicted by circles corresponds to simulating the QTS signal ΔQ(x) by an amplitude Gaussian with

variance w*. It is evident that the signal is peaking when the condition t1m ≈ τ is fulfilled with sufficient

accuracy; the peak height ΔQm corresponds to approximately 0.17Q0.

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.14 Throughout the present study, only the isothermal regime, i. e., scanning of the delay t1 at a constant

temperature T, will be considered. As to the pulse duration (pulse), two modes were alternatively used:

i) pulse = 1 ms; ii) pulse = 4t1 (both pulse and t1 scanned). In other words, the respective repetition

periods of excitation are (10-3 s + 4t1) and 8t1 for the two modes. For a correct understanding of what

follows, a sign convention is necessary, namely that signal ΔQ is of the same sign as ΔU.

Figure 3.9. Normalized QTS

responses to three exponential

decays with different time

constants τ are peaking when the

processing starts at t1 = τ. It should

be noted that both the height ΔQm

and the FWHM are invariant

against τ; approximation of the

fastest charge relaxation by a

Gaussian is shown by circles; w*

stands for the variance of the

Gaussian.

3.5. Algorithms for data evaluation and simulations

Most of the data evaluations were carried out in MatLab [MATL2003]. In some calculations that

involved numerical integration (Fermi-Dirac integrals of arbitrary order), a routine based on the

Gaussian quadrature and described below was used.

3.5.1. Gauss-Legendre quadrature

In order to evaluate some Fermi integrals, an algorithm based on the Gauss-Legendre quadrature

[Pres2002] was developed. The scheme is as follows; given a smooth function f(x) in an interval (a,b),

its integration can be approached by the summation of the weighted function in the interval (-1,1)

)(R)(g)(w2

d)(g2

xdf(x) kn

n

1kkk

1

1

ζ+ζζ≈ζζ= ∑∫∫=−

a-ba-bb

a

, (3.10)

with )(w kζ the weighting coefficients evaluated at the zeros ( kζ ) of the Legendre polynomial ( nP )

and defined by

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.15

( )[ ]2kn2k

k )(P12)(w

ζ′ζ−=ζ . (3.11)

nP′ is the derivative of the Legendre polynomial, and

( )[ ] )(g)!n2(1n2

)!n(2)(R )n2(3

41n2

kn ζ+

=ζ+

, (3.12)

is the rest of the sum.

The implemented algorithm allows the integrals to be evaluated with an accuracy of less than 1 %,

compared with the Simpson’s method; obviously, this depends on the order (n) of the Legendre

polynomial. The evaluation of the Fermi Dirac integral in this work is an example of the advantage of

the implemented routine.

3.5.2. QTS spectra line-profile

The QTS spectra were fitted by means of the filtering formula described by eq.(3.9), since it is

electronically implemented in the QTS setup to sense the sample signal. It is obvious that this method

of data evaluation gives greater confidence than the approximation given by the Gaussian profile

counterpart.

3.5.3. Correlated fitting algorithm

The fitting of the Raman spectra, current-voltage characteristics, QTS spectra and some other

measured quantities was carried out in a program written in MatLab [MATL2003]. The fitting

algorithm was based on the “Levenberg Marquardt” method. Although MatLab has a function to

perform the fitting at a given tolerance, it is not totally optimized and the convergence is quite

dependent on the initial boundary conditions.

To overcome this limitation, a routine with a correlated self-consistency algorithm was developed (cf.

Figure 3.10). The written program allows the boundary conditions to be optimized in such a way that,

after each iteration, the initial conditions are self-adjusted until the set tolerance is reached. Also, it is

possible to carry out back and forth fittings in order to determine a common set of boundary

conditions for two consecutive spectra. Figure 3.10 shows the flux diagram of the correlated fitting

algorithm.

The recursivity of the algorithm considerably reduces the computational time (by ca. 85 %). As an

example, the spectra shown in Figure 6.1 and the corresponding parameters presented in figs. 6.3(a)-

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.16 6.3(b) were fitted in an Intel Pentium 1.6 GHz (512 RAM) after 7.5 h, while with the non-optimized

routine and with lower correlation factors (used earlier) the fitting took approximately 48 h.

Figure 3.10. Flux diagram of the correlated fitting algorithm.

References

[Farm1981] J. W. Farmer, C. D. Lamp, and J. M. Meese, “Charge transient spectroscopy”, Appl. Phys. Lett. 41, 1063-1065 (1981).

[Ferg2006] A. J. Ferguson, T. S. Jones, “Photophysics of PTCDA and Me-PTCDI thin films: Effects of growth temperature”, J. Phys. Chem. B 110, 6891-6898 (2006).

[Frie2003] M. Friedrich, G. Gavrila, C. Himcinschi, T. U. Kampen, A. Yu Kobitski, H. Méndez, G. Salvan, I. Cerillo, J. Méndez, N. Nicoara, A. M. Baró, and D. R. T. Zahn, “Optical properties and molecular orientation in organic thin films”, Phys. Condens. Matters 15, S2699-S2718 (2003).

[Gavr2006] G. N. Gavrila, “Electronic properties and chemistry of metal / organic semiconductor / S- GaAs(100) heterostructures”, PhD thesis http://archiv.tu-chemnitz.de/pub/2006/0004/index.html TU Chemnitz (2006).

[Hepp2003] A. Hepp, H. Heil,W. Weise, M. Ahles, R. Schmechel, and H. von Seggern, “Light-emitting field-effect transistor based on a tetracene thin film”, Phys. Rev. Lett. 91,157406 1-4 (2003).

Chapter 3 B. A. Paez-Sierra, Experimental techniques… 3.17 [Kirov1981] K. I. Kirov and K. B. Radev, “A simple charge-based DLTS technique”, phys. stat. sol. a 63, 711-716 (1981). [Kolo2005] The samples consisting of pentacene and C60 deposited on quartz with a nominal thickness of 30 nm were

prepared and characterized with UV-vis by MSc. Phys. Viktoriia Kolotovska, TU Chemnitz (2005). [Kobi2004] A. Yu Kobitski, R. Scholz, D. R. T. Zahn, “Theoretical studies of the vibrational properties of the 3,4,9,10-

Perylene Tetracarboxylic Dianhydride (PTCDA) Molecule”, J. Molec. Struct. (Theochem.) 625, 39-46 (2004). [Kost1994] B. Kostant, “Structure of the truncated icosahedron (such as fullerene or viral coatings) and a 60-element

conjugacy class in PSl(2, 11)”, Proc. Nati. Acad. Sci. USA 91, 11714-11717(1994) Mathematics. [LabV2004] LabView® 5.0, National Instruments, 2004.

[MATL2003] MATLAB® 6.1, The MathWorks, Inc. 2003. [Ménd2006] H. Méndez, “Organic modification of Metal / Semiconductor contacts”, PhD Dissertation, http://archiv.tu-

chemnitz.de/pub/2006/0124/index.html TU Chemnitz (2006). [Paez2005] B.A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman

spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F 1-9 (2005).

[Pann2004] Ch. Pannemann, T. Diekmann, and U. Hilleringmann, “Degradation of organic field-effect transistors made of Pentacene”, J. Mater. Res. 19, 1999-2002 (2004).

[Pres2002] W. H. Press, S. A. Teukolsky, W. T. Vetterling, B. P. Flannery, “Numerical recipes in C, The Art of Scientific Computing”, (Cambridge University press second edition, New York 2002).

[Ross2002] D. Ross and R. Aroca, “Efective medium theories in surface enhanced infrared spectroscopy: the pentacene example”, J. Chem. Phys. 117, 8095-8103 (2002).

[Salv2003] G. Salvan, “Metal/Organic/Inorganic semiconductor heterostructures characterized by vibrational spectroscopies”, PhD thesis, http://archiv.tu-chemnitz.de/pub/2003/0109/index.html TU Chemnitz (2003).

[Thur1994] I. Thurzo and K. Gmucová, “Switching between deep-level transient spectroscopy and feedback charge capacitance modes in a versatile time-domain spectrometer”, Rev. Sci. Instrum. 65, 2244-2248 (1994).

[Thur2005] I. Thurzo, H. Méndez and D. R. T. Zahn, “Analysis of charge transient spectroscopy data originating from Gaussian densities of electron states in organics”, phys. stat. sol. (a) 202, 1994–2007 (2005).

[Thur2006] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, “Anomalous charge relaxation in channels of pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 203, 2326-2340 (2006).

[Zou2006] Y. Zou, L. Kilian, A. Schöll, Th. Schmidt, R. Fink, and E. Umbach, “Chemical bonding of PTCDA on Ag surfaces and the formation of interface states”, Surf. Sci. 600, 1240-1251 (2006).

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.1

Chapter 4

Metal / organic interface formation investigated by in situ

surface-enhanced Raman spectroscopy (SERS)

In this chapter, the in situ investigation by resonant Raman spectroscopy of organic/metal

heterostructures prepared under UHV conditions is described. The heterostructures consist of indium

and magnesium deposited onto two perylene derivatives, 3,4,9,10-perylene tetra-carboxylic

dianhydride (PTCDA) and N, N' dimethyl-3,4,9,10-perylene tetracarboxylic diimide (DiMe-

PTCDI). The organic/metal structures were assembled on sulphur passivated Si-doped GasAs(100)

substrates (S-GaAs). The experimental results proved that all metals deposited onto the organic layers PTCDA or DiMe-

PTCDI promote enhancement of the Raman-active internal vibrational mode intensities, accompanied

by the activation of normally infrared-active modes. The developed phenomenon is called surface-

enhanced Raman spectroscopy (SERS). It is shown that metal coverages of several nanometer

thickness about 40 nm or above still allow the identification of vibrational bands, probing the

roughness of the metallic layers. In this research, the enhancement factors are estimated to be 101 for

the chemical contribution or first monolayer effect, and 103 for the electromagnetic contribution.

The comparison between organic monolayers capped by tiny amounts of metal, and thicker organic

films covered with similar metallic layers is used to extract information about chemical reactions at

the interface, diffusion of the metal into the organic material, and morphology of the metal film.

Indium deposition onto PTCDA and DiMePTCDI revealed molecular distortion, along with a

remarkable in-diffusion of In into In/ PTCDA structures, as demonstrated for the metal coverage on

organic monolayers and onto thicker PTCDA (15 nm) films. The phenomenon was much less apparent

in DiMe-PTCDI thin films.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.2

The deposition of Mg on both perylene derivatives revealed much less diffusion of the metal into the

organic layers, as demonstrated by the preservation of the external modes upon metal coverage. The

Mg/PTCDA structures undergo modification in two stages. The first one is the formation of a new

molecular structure at the interface and continued until there is a nominal metal coverage of about 2.8

nm, attributed to the removal of the oxygen atom from the anhydride groups. The second phenomenon

consists of the surface-enhancement of Raman signal of the former structure by further depositions of

Mg.

In the case of Mg/DiMe-PTCDI, it was found that the molecule exhibits a breakdown of selection

rules, and no formation of new molecular species compared to the Mg/PTCDA interfaces was

observed. This structure is characterized by the outstanding coupling between discrete molecular states

of the organic DiMe-PTCDI material and the electronic continuum of electronic states at the Mg metal

contact. The phenomenon is evidenced through the asymmetrical broadening of the Raman bands at

221 cm-1, 1291 cm-1 and 1606 cm-1 subsequent to the metal deposition. The line-shape of these bands

is well described by the Breit-Wigner-Fano function [Brow2001].

4.1. Introduction

The intimate contact of the organic/inorganic or organic1/organic2 junctions has been theoretically and

experimentally proven to be decisive in the further growth of the organic material, [Forr2003,

Hein2006, Otto2001-2005, Paez2005a, Pers2006, Schr2004, Wagn2003, Witt2004, Zahn2006,

Zou2006]. Therefore, the combined organic/inorganic or so called hybrid systems are of great interest

in different applications like organic-based field effect transistors (OFETs) [Paez2003a-b-2005b,

Scho2005, Thur2006, Xue2004] or organic modified diodes [Ménd2006, Park2002].

Several preparation methods to produce these types of structures are of current interest [Möll2003,

Nadi2005]. In particular, scientific reports concerning optical properties and molecular ordering

[Frie2003], electronic properties of PTCDA, PTCDI (3,4,9,10-perylenetetracarboxylic-diimide) and

DiMe-PTCDI investigated by surface sensitive PES (photoemission spectroscopy), IPES (inverse

photoemission spectroscopy) and NEXAFS (near edge x-ray absorption fine structure) are reported by

Gavrila [Gavr2006]. Extensions in conjunction with phthalocyanines “Pcs” (H2Pc, CuPc, F4Pc, F16Pc)

have been recently reviewed by Zahn and colleagues [Zahn2006]. The vibrational investigations based

on Raman spectroscopy and infrared spectroscopy have been a major focus of the group, which has

pursued investigations of metal/organic contact formation of PTCDA and DiMe-PTCDI with silver as

well [Salv2003].

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.3 The organic/metal contact heterostructures discussed in this thesis and OFET structures are

summarized in various publications [Paez2003a-c-2004a-b,-2005a-b, Salv02004a-b-2005, Zahn2004-

2005]. Also, an extensive effort has been directed at the organic-based devices and particularly

summarized in the projects “Designing Inorganic/Organic Devices” (DIODE) [diod2001], and the

“Organic Field Effect Transistors” (OFET). The latter project was carried out under the auspices of

the -,Deutsche Forschungsgemeinschaft, DFG, (German Research Promotion Society) [ofet], and

included this investigation (further details were given in chapter 1). Previous research, based on

electrical characterization of organic modification of metal/semiconductor Schottky contacts by

PTCDA [Park2002] and DiMe-PTCDI [Ménd2006, Thur2005a-b] has also been done by the group.

The perylene derivatives are known to be n-type materials, as demonstrated by the realization of n-

channel FETs [Ostr1997, Xue2004]. Complex structures based on coevaporated PTCDA and ZnPc (p-

type) [Dero2004] and bilayer structures formed by PTCDA/CuPc [Heut2005] have proved to be suited

for solar cell applications.

Likewise, p-n junctions between Pc and DiMe-PTCDI have provided power-conversion efficiencies of

approx. 1 % [McKe1998]. Recently, novel combinations of polyethyleneimine (PEI) with PTCDA

have been realized to produce functionalized structures where the PTCDA derivative molecules act as

the nanotube wall of the PEI-based backbone [Tian2006], hence leading to inexpensive and bendable

electronics.

The experimental results described in this chapter mainly relate to the resonant surface enhanced

Raman spectroscopy (SERS). Some authors have estimated the enhancement factors to be 102-106 for

the chemical contribution or first monolayer effect [Pers2006], and 1010 [Joha2005] for the

electromagnetic contribution. The magnified Raman cross section is about 10-17-10-14 cm2 [Joha2005,

Knei1997].

4.2. Interaction of metals with perylene derivatives

In Figure 4.1, the Raman spectra of 15 nm films of PTCDA are shown for metal coverages of 5 nm In,

4.5 nm Ag, and 5 nm Mg. The spectra in the low-frequency windows are normalized to the height of

the molecular breathing mode at 233 cm-1. The normalization in the high-frequency region is

performed with respect to the C=C stretch mode (1572 cm-1).

The deposition of Mg or In onto PTCDA leads to the appearance of the B1u band at 1243 cm-1 and the

increase in the relative intensity of the B3g mode at 1338 cm-1. These bands are strongly enhanced

compared to the Ag modes during the Mg deposition onto a monolayer of PTCDA on S-GaAs

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.4 [Zahn2004]. Therefore, they are considered to be a signature of the “first layer effect”, i.e. of the

molecules having direct contact with Ag [Zahn2004-2005].

The band at 1338 cm-1 was identified in reference [Kobi2002] as a B3g band on the basis of its

frequency and intensity in the crystal spectra, while a band 1292 cm-1 is likely to be a shifted variant of

the C-H deformation Ag mode at 1303 cm-1 in the single crystal [Salv2004b]. The other bands

correspond to modes that normally show infrared activity.

Figure 4.1. Raman spectra of In (5nm), Ag (4.5 nm) and Mg (5 nm) coverages on 15 nm thick PTCDA

films, compared with the spectrum of the bare PTCDA film in the spectral region of the internal

breathing mode (left) and in the spectral region of HC− deformation and C=C stretching modes

(right). (The Raman spectra involving Ag do not belong to this work, and are addressed elsewhere

[Salv2003]. They are presented here for comparison of metal contact formation on similar molecular

structures).

These modes are also activated in the spectra of PTCDA monolayers covered with indium, only with

higher intensities relative to those of the normally Raman active modes [Zahn2004]. In reference

[Zahn2004], the observed breakdown of the Raman-infrared selection rules was proposed to originate

from a weak charge transfer between the molecules and the metal surface mediated by molecular

vibrations. The Raman spectra of the (5 nm) Mg / (15 nm) PTCDA system also exhibit the break-

down of selection rules, with the occurrence of the modes observed in the other two metal/organic

heterostructures (cf. Figure4.1).

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.5 In addition, several modes with significant intensity appear at 307 cm-1, 502 cm-1, 598 cm-1, 696 cm-1,

1088 cm-1 and at 1225 cm-1, and 1433 cm-1. The assignment of these modes cannot yet be done

unambiguously. The frequency of the mode at 598 cm-1 is very close to the calculated value of 592 cm-

1 for a B3g mode of an isolated PTCDA molecule [Kobi2004].

Frequency calculations performed with the same basis set and density functional method in

Gaussian’98 as in the reference [Kobi2004] but for a modified PTCDA molecule, in which the central

O atom in the anhydride group is removed, deliver several frequencies that may be candidates for the

assignment of the experimentally observed modes 308 cm-1, 500 cm-1, 581 cm-1, 702 cm-1, 1090 cm-1.

Raman active modes in MgO microcrystals were observed at 595 cm-1, 719 cm-1 and 1096 cm-1

[Böck1974]. Thus the modes observed in the present work at 598 cm-1, 696 cm-1 and 1088 cm-1 may

also indicate the formation of MgO as a result of the interaction between Mg and PTCDA.

No modes of PTCDA or the modified molecule are found in the vicinity of 1225 cm-1. Whatever the

final assignment of the new modes is, they are not activated in the molecules in contact with either Ag

or In. Therefore, it can be concluded that the model of weak charge transfer is not sufficient to

describe the interaction involved at the Mg/PTCDA interface. Indeed, recent PES measurements

performed during the Mg deposition onto PTCDA show that the C1s and O1s core levels undergo

dramatic changes which can only be explained by the breaking of the OC− bonds in the PTCDA

molecule with formation of MgO [Gavr2006, Paez2003c-2004b].

In Figure 4.2, the spectra of 15 nm DiMe-PTCDI films for metal coverages of 5 nm In, 4,5 nm Ag and

5 nm Mg are shown. The spectra in the low-frequency windows are normalized to the height of the

breathing mode at 221 cm−1. The normalization in the high-frequency region is performed with respect

to the CC− stretch modes (1570 cm−1). In the case of DiMe-PTCDI, all the investigated metals, i.e.,

Ag, In and Mg, lead to the breakdown of selection rules with the occurrence of normally infrared

active modes at 1246 cm−1 and 1606 cm−1. The breathing mode at 221 cm−1 survives with increasing

metal coverage.

Consequently, a chemical reaction between these metals and the O atoms of DiMe-PTCDI molecules

can be ruled out [Zahn2004]. Interestingly, the features potentially assigned to MgO phonons do not

appear in the spectra, even for higher coverages of Mg. It can thus be concluded that the imide-methyl

group in the DiMe-PTCDI is less reactive compared to the O atoms in the anhydride group of PTCDA.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.6

Figure 4.2. Raman spectra of In (5nm), Ag (4.5 nm) and Mg (6 nm) coverages on 15 nm thick DiMe-

PTCDI films, compared with the spectrum of the bare DiMe-PTCDI film.

4.3. Morphology of the metal film Besides the occurrence of internal modes related to molecules in direct contact with the metal, the

totally symmetric modes are also enhanced (as shown by the normalization factors in Figure 4.1 and

Figure 4.2) in the spectra of Ag, In and Mg on 15 nm thick PTCDA and DiMe-PTCDI films. The latter

effect originates from the coupling of the incident and scattered radiation with localized and/or

collective plasmon resonances in the rough metal film. Accordingly, the intensity of the totally

symmetric Ag modes is very sensitive to the morphology of the metal film.

For a quantitative determination of the enhancement factors, curve fitting of each set of spectra

recorded during silver, indium and magnesium deposition onto PTCDA and DiMe-PTCDI was

performed using Lorentzian peaks. The dependence of relative area on metal coverage is plotted in

Figure 4.3 for a representative totally symmetric mode and for a normally infrared active mode of each

organic material. The relative intensities of a given metal coverage relate to the intensities of the

spectrum where the mode occurs for the first time. For example, the reference spectrum for the totally

symmetric Raman band is that of the pure organic film, while the reference spectrum for the normally

infrared active band is that taken after the first metal deposition.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.7

(a)

(b)

Figure 4.3. Enhancement factors of the Bu mode (1243 cm-1 in PTCDA and 1246 cm-1 in DiMe-

PTCDI) and of the C-C stretch Ag mode (1572 cm-1 in PTCDA and 1570 cm-1 in DiMe-PTCDI) for

PTCDA (a), and DiMe-PTCDI (right) as a function of the metal coverage (b).

The intensities of the Ag modes initially increase upon Ag and In deposition, reflecting an increase in

number and size of metal clusters as their plasmon energy approaches the energy of the laser field.

When Mg is deposited onto PTCDA, the intensities initially decrease, reflecting a reduction in the

number of Raman active PTCDA molecules. This corresponds to the conclusion drawn in the previous

section regarding the disruption of the PTCDA molecular structure upon reaction with Mg.

Above 2.8 nm Mg nominal coverage, however, the Ag Raman modes start to be enhanced, indicating

the formation of metallic clusters. Interestingly, the enhancement of the DiMe-PTCDI modes occurs

only above 20 nm nominal Mg thickness. The large difference in nominal thickness for which the

metallic character of Mg clusters is formed on PTCDA and DiMe-PTCDI is probably related to the

different morphology of the underlying organic layer. The DiMe-PTCDI films have very large empty

spaces between the organic islands, while the PTCDA films are much more compact [Frie2003].

The maximum enhancement of PTCDA modes for the Ag/PTCDA (15 nm) system is observed around

11 nm nominal Ag coverage (cf. Figure 4.3). This corresponds to the optimum cluster size for the

dipolar plasmon resonance.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.8 The In film thickness yielding the maximum enhancement for PTCDA and DiMe-PTCDI films is 26

nm and 5 nm, respectively. A further increase in the metal thickness leads to increasing size of the

metal clusters associated with screening of the inelastic-scattered radiation. Furthermore, the

absorption in the metal film also plays an important role in decreasing the Raman signal for higher

nominal coverages, when the clusters start to percolate. The signal from PTCDA and DiMe-PTCDI

internal modes remains visible even for a metal coverage of 43 nm, with higher intensity compared to

the pure organic film.

For Ag deposition onto DiMe-PTCDI, no saturation of the signal intensity was observed up to a

coverage of 263 nm. Considering that oI is the intensity of the light incident on the sample, d the

nominal thickness of the metal coverage and δ~

is the light penetration depth in the metal, then the

light intensity I scattered by the sample can be described by

)~d2exp(II o δ

−∝ . (4.1)

A summary of the values obtained from the fitting of the experimental decay of the enhancement

factors for the totally symmetric C=C stretching mode in all investigated heterostructures is given in

table 4.1. The obtained values are much larger compared to the penetration depth of 488 nm light into

smooth closed metal films. This is a clear indication that In and Ag films grown on PTCDA and

DiMe-PTCDI are not closed and have a high degree of roughness.

The apparent light penetration depth in Mg films grown on PTCDA and DiMe-PTCDI estimated from

the decrease in intensity of the C=C stretching mode has values comparable with the penetration depth

in a closed smooth Mg film. This indicates that the Mg film is smoother and that the efficiency of the

488 nm radiation in exciting dipolar resonances is lower for Mg.

Figure 4.4. AFM topographic

images of a 30 nm thick In film

on PTCDA. (a) (right part

showing PTCDA covered by In

clusters) and of a 113 nm thick

Mg film on PTCDA (b).

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.9 The AFM topographic images in Figure 4.4 confirm the higher roughness of In compared to that of

Mg films. It is recalled that the Ag modes are enhanced via the long range electromagnetic effect,

while the activation of Bu modes is characteristic for the molecules in intimate contact with or in the

very near vicinity of the metal surface. Therefore, the intensity of the Bu modes relative to that of the

Ag modes will be considered in the following to extract the metal diffusion depth into the organic

films.

Table 4.1. Skin depth of smooth metallic films, apparent penetration

depth of 488 nm light in In, Ag and Mg films grown on DiMe-PTCDI and

PTCDA layers

In Mg Ag

)filmsmoothdepthSkin(nm488=λΔ 8 nm 14 nm 2.5 nm

nm488=λδ (PTCDA) 49 nm 15 nm 50 nm

nm488=λδ (DiMe-PTCDI) 98 nm 15 nm --

Roughness (AFM) for a 15 nm

metal coverage on PTCDA

41 nm 28 nm --

In the case of Ag/PTCDA and Ag/DiMe-PTCDI, the intensity of Bu modes is low, indicating that only

a few molecules have intimate contact with Ag. This leads to the conclusion that the Ag atoms diffuse

very little into the PTCDA grains. On the other hand, the Bu bands are stronger compared to the Ag

modes in the spectra of In/PTCDA. This suggests that a large number of PTCDA molecules are in

direct contact with the metal, indicating a strong diffusion of In into the PTCDA islands. In/DiMe-

PTCDI represents an intermediate case between Ag/PTCDA and In/PTCDA..

The ratio between the area of the Bu mode at 1243 cm-1 (1246 cm-1) and that of the Ag mode at 1572

cm-1(1570 cm-1) in PTCDA (DiMe-PTCDI) is shown as a function of metal thickness in Figure 4.3. In

the case of Ag/DiMe-PTCDI, the maximum value of the ratio is observed for the first Ag deposition,

i.e., 0.4 nm Ag, whereas for PTCDA it increases up to the 1.4 nm nominal coverage of Ag. For In

deposition onto both organics, this ratio shows a saturation tendency only above 15 nm nominal In

coverage, but its value is lower for In/DiMe-PTCDI.

It is proposed that a maximum in the ratio defined above can be directly related to the metal diffusion

length in the organic film. Thus the Ag atoms arriving at the organic film surface diffuse into the

PTCDA or DiMe-PTCDI islands up to a nominal Ag coverage of 1.4 nm and 0.4 nm, respectively. An

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.10 exception is observed in the case of Mg/DiMe-PTCDI. Here the maximum in the defined ratio occurs

at the nominal metal coverage where the Ag modes start to be enhanced, i.e., around 20 nm.

The conclusions regarding the structural properties of the metal/organic interfaces drawn from the

enhancement factors of the internal modes are further confirmed by the spectral changes in the region

of external modes below 125 cm-1 that are discussed in detail in the reference [Salv2004a]. Whereas

the Ag and Mg deposition causes a broadening and a slight decrease in intensity of the phonon bands,

these smear out completely in the In/PTCDA system. This proves that indium diffuses much more

easily, as compared to Ag or Mg, into the organic islands, provoking the disruption of their crystalline

structure.

4.4. Phonons and interface structural properties

In Figure 4.5, the Raman spectra of 15 nm films of PTCDA in the region of external modes are shown

for 0.4 nm of Ag, In, and Mg. The spectra were recorded in crossed polarization configuration, i.e., the

incident electric field vector of the scattered light is perpendicular to that of the analyzed light or in the

Porto notation z(xy)z’ (see sect. 3.1).

Figure 4.5. Spectra of external Raman modes

from 15 nm thick PTCDA films capped with

0.4 nm metal layers, i.e., Ag, In, and Mg. The

spectral Raman shift between 25 cm-1 and 125

cm-1 corresponds to the libronic or collective

modes of the interacting molecules in the unit

cell [Salv2003].

In Figure 4.6, the spectra of the external modes are shown following stepwise metal deposition onto 15

nm PTCDA films.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.11

(a)

(b)

(c)

Figure 4.6. Raman monitoring in the external

mode region upon metal deposition: (a) Ag, (b)

Mg, (c) In. The experimental spectra are shown by

open symbols and the fitted spectra by red lines.

The Lorentzian functions used for the fitting of

the Raman spectrum of the pure PTCDA film are

shown by lines in the lower parts of the figures.

The spectra of Ag/PTCDA are normalized for a

better resolution of the phonons.

It can be seen in Figure 4.6(b) that for Mg overlayers, the external modes are still visible at 12 nm

coverage, whereas they are almost completely smeared out at 1.3 nm Ag. This is a clear indication that

the crystalline structure of the organic layers is less affected by the Mg deposition compared to Ag.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.12 However, it should be noted that the fitting of the spectra in the case of Ag/PTCDA is complicated by

the significant increase in the low-frequency background (see normalization factors in Figure4.6(a)).

The background evolution reflects an increasing degree of roughness, which is consistent with an

increasing number of metallic clusters that diffusely scatter the light. A strong increase in the low-

frequency background is also observed in the case of In deposition onto PTCDA, while it hardly

affects the spectra of Mg/PTCDA, supporting that the roughening due to Mg is lower compared to that

of the Ag and In films.

While the external molecular modes already disappear in the first deposition stages for In / PTCDA,

two new modes develop at 33 cm-1 and 112 cm-1 above an In coverage of 2.8 nm. They may

correspond to the transversal acoustic and longitudinal acoustic phonon peaks located at 33 cm-1 and

112 cm-1, respectively, in bulk indium [Flei2003].

This observation, corroborated by the concomitant increase in the low-frequency background,

indicates the formation of metallic In clusters. Moreover, the enhancement of the internal modes also

increases dramatically above 2.8 nm In, supporting the conclusion about the formation of metallic

clusters.

For a quantitative evaluation, the spectra of Ag/PTCDA and Mg/PTCDA were fitted using Lorentzian

functions. The evolution of the FWHM as a function of Ag and Mg thickness is plotted in Figure 4.7

for the external mode at 41 cm-1. This mode is fairly well separated from its neighbors and

consequently, the fitting parameters of the corresponding Lorentzian function are less correlated. As

the metal thickness increases, the FWHM of the external modes increases faster in Ag/PTCDA.

Figure 4.7. Evolution of the

FWHM of the external mode at

41 cm-1 as a function of the metal

coverage relative to the initial

values before the metal (Mg, Ag)

deposition. The dashed lines are

visual guides.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.13 4.5. Mg/DiMe-PTCDI structures and discrete molecular coupling with

continuum electronic metal states

In the previous sections, it has been found that In and Ag diffuses into PTCDA and DiMe-PTCDI to a

greater extent, as compared to Mg. On the other hand, it has been proven that at the Mg/PTCDA

interface a modified molecular structure is developed, possibly due to the withdrawal of the oxygen

atoms joining the anhydride groups of PTCDA.

In the case of DiMe-PTCDI, where the anhydride groups are replaced by imide groups, the Mg

overlayer on the organic layers undergoes interaction with the oxygen of the carboxylic groups and no

oxygen withdrawal has been evidenced by complementary PES measurements [Gavr2006]. An

interesting signature of the Mg/DiMe-PTCDI structure is the discrete molecular coupling with the

continuum electronic states of the metallic atoms, referred to as Breit-Wigner-Fano (BWF) resonance

[Brow2001, Cohe1992, Fano1961, Paez2005a, Zhou1993].

It has been found that when Mg is deposited onto a 15 nm DiMe-PTCDI layer the external molecular

modes are preserved up to large metal coverage. Since the external molecular modes are a fingerprint

for the molecular crystal, their preservation indicates a low diffusion of Mg into the DiMe-PTCDI

layer. Concerning the internal molecular modes, the Mg deposition induces a breakdown of selection

rules which is proposed to originate from a dynamic charge transfer between the DiMe-PTCDI

molecules and the metal. The line shape of the molecular breathing mode at 221 cm-1 becomes

asymmetrical at its high-frequency side above 0.3 nm nominal Mg thickness.

A similar effect is observed for the bands that occur at 1291 cm-1 and 1606 cm-1, but the asymmetry

appears on the low-frequency side. This line-shape asymmetry is likely to be related to a Fano

resonant coupling between the molecular vibration modes and the electronic continuum of states of

metallic clusters formed above 0.3 nm Mg coverage.

4.5.1. Chemistry, metal film morphology and metal indiffusion

Figure 4.8 shows the Raman spectrum of a bare 15 nm DiMe-PTCDI film and its evolution with step-

wise deposition of magnesium.

The first 0.3 nm Mg deposition leads to the occurrence of a band at 1252 cm-1. Upon further Mg

deposition, other modes develop at: 1291 cm-1, 1460cm-1 and 1606 cm-1. As shown above in Figure

4.8, all of these modes can be assigned to shifted normal modes of the DiMe-PTCDI molecule that are

only infrared active in the unperturbed molecule.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.14

50 100 150 200 250 300 350

GaAs

modesExternal

154/ 4

/ 4.4

/ 3.3

/ 0.8

/ 4.1

/ 0.8

/ 0.4

/ 0.4/ 0.4

/ 0.7

34

1x10-2

21

12258

12

6.48

In

tens

ity /

cts.

mW

-1s-1

Raman shift / cm-1

Mg thickness / nm

00.32.2

z(xy)-z

}*

(a)

1200 1280 1360 1440 1520 1600 1680

*154

/ 7.3/ 7.6

/ 2.7

/ 0.49.6

/ 4.7

/ 0.6

/ 6.4

/ 0.8/ 0.45

/ 0.4/ 0.4/ 0.6

34

1x10-1

21

12250

12

6.48

In

tens

ity /

cts.

mW

-1s-1

Raman shift / cm-1

Mg thickness / nm

00.32.2

28

*

z(xy)-z

(b)

Figure 4.8. Raman spectra of Mg/DiMePTCDI in the region of: (a) external modes and the breathing

molecular vibration mode. (b) C-C and C-H modes. The spectra in (a) and (b) are normalized with

respect to the intensity of the breathing mode and to that of the C-C stretching mode at 1570 cm-1,

respectively. An asymmetric broadening develops for the three modes marked with stars upon the Mg

deposition.

The IR spectra depicted in Figure 4.9, do not cover the spectral region below 650 cm-1; therefore, any

assignment of this band is difficult. However, from density functional theory calculations, this band

might belong to a normal mode with Au symmetry. The activation of modes with lower symmetry is

again observed, leading to the well known effect for molecules in contact with metal surfaces

[Otto2001-2005]. It can be induced by either molecular deformation in the vicinity of an interface or

by charge transfer processes from the molecule into the metal or vice-versa.

In addition to the spectral changes discussed above, the deposition of Mg up to the coverage of 21 nm

leads to a decrease in the overall signal, as reflected by the normalization factors in Figure 4.8. The

signal decrease is caused by the light attenuation in the metal overlayer. When the Mg coverage is

increased above 21 nm, however, the intensity of all bands again increases. The maximum

enhancement factor relative to the intensity in the spectrum of the bare organic film is observed for the

C-C stretching mode at 1570 cm-1 and it amounts to 7.6 for the Mg coverage of 122 nm. An increase

of the Raman signal of the organic layer upon metal deposition has also been observed for In and Ag.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.15 This effect occurs due to the enhancement of the electric field of the incoming and scattered radiation

induced by plasmons that are resonantly excited by the electromagnetic radiation in the metal clusters.

As a consequence, the enhancement of the Raman signal provides clear proof of a high degree of

roughness of the metal film.

900 950 1000 1050 1100 1150

Inte

nsity

/ ct

s.m

W -1

s -1

Raman shift / cm-1

DiMePTCDI (15nm)/ Mg (2.8 nm)

(a)

1200 1300 1400 1500 1600 1700 1800 1900 2000

Raman shift / cm -1

Mg/ DiMePTCDI

IR DiMePTCDI

Inte

nsity

/ ct

s.m

W -1

s -1

DiMePTCDI

DiMePTCDI (15nm) / Mg (2.8 nm)

(b)

Figure 4.9. Comparison between the Raman spectra of bare DiMe-PTCDI, Mg (2.8 nm) / DiMe-

PTCDI and the IR for the organic.

Upon metal deposition up to 21 nm, the external modes (phonons) below 120 cm–1 remain unchanged.

Considering that the external modes are a signature of the organic layer crystallinity, their preservation

indicates that the Mg atoms do not diffuse into the organic crystalline islands as much as In and Ag do

on both perylene derivatives.

4.5.2. Coupling of vibrational modes and electronic excitations

Another modification induced in the Raman spectra by the metal deposition is an asymmetric

broadening of the breathing mode at 221 cm-1, and the modes at 1291 cm-1 and 1600 cm-1. Similar

asymmetric line-shapes were observed for metal-doped fullerene films [Brow2001, Knei2001] and in

the Raman spectra of metallic carbon nanotubes [Zhou1993]. There, it has been suggested that the

asymmetric line-shape can be fitted with a Breit-Wigner-Fano (BWF) function that accounts for a

coupling between discrete states and an electron states continuum, so-called Fano resonance. The

phenomenon has been observed in the Raman spectra of inorganic systems such as highly doped Si

[Card1983, Cerd1973, Chan1978] and semiconductor superlattices [Pan1996, Kanz2000].

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.16 The BWF function reads:

( )

( ),

/)(1)/()(1

)( 2

2

Γω−ω+

Γω−ω+=ω

BWF

BWFo

qII (4.2)

This expression can be rewritten as: ( ) ,1qI~)(I 2

2

o ε+ε+

=ω where Γω−ω

=ε BWF is the reduced energy,

with BWFω being the Breit-Wigner-Fano frequency and Γ the resonance width related to the phonon

self-energy when the interaction with the electronic continuum takes place. q is the asymmetry

parameter and (1/q) is proportional to the degree of coupling.

Figure 4.10. Fitted Raman spectra of Mg / DiMe-PTCDI: from bottom to top: bare 15 nm DiMe-PTCDI

covered with 34 nm Mg and 122 nm Mg. The peaks fitted with the BWF function are represented by

thick black lines.

Here the BWF function is employed for fitting the modes at 221 cm-1, 1291 cm-1 and 1606 cm-1. All

other modes are fitted using Lorentzian functions. Figure 4.10 shows, from bottom to top, the

experimental Raman spectra along with the fitting curves for bare DiMePTCDI (15nm), coverage of

34 nm Mg, and 122 nm Mg thickness. The absolute values of q-1 range from 0.07 to 0.16.

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.17 For comparison, q-1 ranges from 0.18 to 0.27 for a band at 1543 cm-1 in the case of metallic carbon

nanotubes. The value of q-1 is influenced by the choice of the background. Therefore, it is difficult to

draw conclusions regarding changes in the degree of coupling with the increase in metal coverage. For

an excitation energy of 2.54 eV, the energy of the photons scattered by the molecular vibrations at 221

cm-1, 1291cm-1, and 1606 cm-1 amounts to 2.51 eV, 2.38 eV and 2.34 eV, respectively. It must be

noted that the asymmetry parameter of the band at 1606 cm-1 band is negative, while the other two

BWF line-profiles are characterized by positive asymmetry parameters.

According to reference [Zhou1993] a negative/positive value of the coupling parameter indicates that

the central frequency of the continuum lies below/above the discrete mode frequency. This would

imply that in this case the central frequency with the continuum is located between 160 meV and 200

meV above the molecular ground state. On the other hand, the band asymmetry occurs at metal

coverages for which the enhancement of the Raman signal due to dipolar plasmon resonances is also

observed.

Consequently, it might be proposed that the origin of the observed BFW line-shapes is a coupling

between the molecular electronic levels and the plasmons in the Mg clusters modulated by the

molecular vibrations. Interestingly, the band at 221 cm-1 stems from a breathing vibration of the whole

molecule and the band at 1606 cm-1 stems from a stretching vibration. Thus both involve the breathing

of the carbon rings.

In order to summarize the evaluated Fano resonance energies, Figure 4.11 illustrates the energy level

alignment of the DiMe-PTCDI/Mg interface determined by NEXAFS spectroscopy [Gavr2006]. The

Fano resonances distributions were determined from the enhancement factors of the SERS spectra

documented in Figure 4.8.

Figure 4.11. Energy level

alignment of the DiMe-PTCDI /

Mg heterostructure. The Fano

resonances indicated in the band

diagram were obtained from the

resonant Raman measurements,

while the other energy levels

were quoted from NEXAFS

spectroscopy measurements on a

similar sample [Gavr2006].

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.18 Conclusions

Ag, In and Mg deposition onto 15 nm thick PTCDA and DiMe-PTCDI films on S-GaAs(100):2x1

have been characterized in situ by Raman spectroscopy. The breakdown of selection rules in the

spectra of Ag/organic and In/organic heterostructures originates from a dynamic fractional charge

transfer process modulated by molecular vibrations. In the spectra recorded during the Mg deposition

onto PTCDA, bands assigned to MgO develop, indicating a reaction of Mg with PTCDA molecules

that causes the loss of the central O atom of the anhydride groups.

Upon Ag and In deposition, the totally symmetric modes are initially strongly enhanced.

Subsequently, the signal is attenuated exponentially with an exponent that is much smaller than the

penetration depth of the incident radiation in a smooth closed metal film, reflecting a high level of

roughness of the metal overlayer. The intensity of the normally infrared active modes relative to the

Raman active modes provides information on the metal diffusion depth in the organic films.

Complementary information on the metal diffusion depth into the organic layers is provided by the

attenuation rate of the external mode intensities as a function of the metal coverage. While Mg and Ag

form abrupt interfaces, In strongly diffuses into the organic layers. For Mg, however, additional

features are observed in the Raman spectra compared to those induced by Ag and In, clearly indicating

the reactive nature of the interface between Mg and PTCDA.

Additionally, in the DiMe-PTCDI/Mg structure, the observed Raman spectral bands at 221 cm-1, 1291

cm-1 and 1606 cm-1 amount to energy gap states above the HOMO of 30 meV, 160 meV and 200 meV,

respectively. These modes are broadened asymmetrically upon the metal deposition and their line-

shape is well described by the Breit-Wigner-Fano function. This effect is accompanied by an

enhancement of the Raman signal intensity, due to plasmon excitations in the metallic clusters. It is

proposed that the BWF line-shapes originate in a coupling between the molecular electronic levels and

the plasmons in the Mg clusters modulated by the molecular vibrations.

References

[Böck1974] H. K. Böckelmann and R. G. Schlecht, “Raman scattering from microcrystals of MgO”, Phys. Rev. B 10, 5225-5231 (1974).

[Brow2001] S. D. Brown, A. Jorio, P. Corio, M.S. Dresselhaus, G. Dresselhaus, R. Saito, K. Kneipp, “Origin of the Breit-Wigner-Fano lineshape of the tangential G-band feature of metallic carbon nanotubes”, Phys. Rev. B 63, 155414 1-8 (2001).

[Card1983] M. Cardona, “Light Scattering in Solids”, I, V. 8, chap. 4 (Springer-Verlag 1983). [Cerd1973] F. Cerdeira, T. A. Fjeldly, and M. Cardona, “Effect of Free Carriers on zone-center vibrational modes in

heavily Doped p-type Si. II. Optical Modes”, Phys. Rev. B 8, 4734 - 4745 (1973). [ Chan1978] M. Chandrasekhar, J. B. Renucci, and M. Cardona, “Effects of interband excitations on Raman phonons in

heavily doped n-Si”, Phys. Rev. B 17, 1623–1633 (1978). [Cohe1992] C. Cohen-Tanoudji, J. Dupont-Roc and G. Grynberg, “Atom-photon interactions”, pp. 49-66, 239-255. (John

Wiley & Sons, New York, 1992).

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.19

[Dero2004] H. Derouiche, J. C. Bernède, J. L’Hyver, “Optimization of the properties of bulk heterojunctions obtained by coevaporation of Zn-phthalocyanine/Perylene”, Dyes and Pigments 63, 277-289 (2004).

[diod2001] EU funded Human Potential Research Training Network (Contract No.: HPRN-CT-1999-00164) / Coordinator: Prof. DRT Zahn, http://www.tu-chemnitz.de/diode/ (2001).

[Fano1961] U. Fano, “Effects of configuration interaction on intensities and phase shifts” Phys. Rev. 124, 1866-1878 (1961).

[Frie2003] M. Friedrich, G. Gavrila, C. Himcinschi, T. U. Kampen, A. Yu Kobitski, H. Méndez, G. Salvan, I. Cerillo, J. Méndez, N. Nicoara, A. M. Baró, and D. R. T. Zahn, “Optical properties and molecular orientation in organic thin films”, Phys. Condens. Matters 15, S2699-S2718 (2003).

[Flei2003] K. Fleischer, S. Chandola, N. Esser, W. Richter, J. F. McGilp, “Phonon and polarized reflectance spectra from Si(111)-(4×1)In: Evidence for a charge-density-wave driven phase transition”, Phys. Rev. B 67, 235318 1-6 (2003).

[Forr2003] S. R. Forrest, “Organic–inorganic semiconductor devices and 3, 4, 9, 10 perylenetetracarboxylic dianhydride: an early history of organic electronics”, J. Phys. Condens. Matter 15, S2599–S2610 (2003).

[Gavr2006] G. N. Gavrila, “Electronic Properties and Chemistry of Metal/Organic Semiconductor/S-GaAs(100) Heterostructures”, PhD thesis http://archiv.tu-chemnitz.de/pub/2006/0004/index.html TU Chemnitz (2006).

[Hein2006] M. Hein, P. Dumas, M. Sinther, A. Priebe, P. Lilie, A. Bruckbauer, A. Pucci, and A. Otto, “Relation between surface resistance, infrared-, surface enhanced infrared- and surface enhanced Raman-spectroscopy’s of C2H4 and CO on copper”, Surf. Sci. 600, 1017–1025 (2006).

[Heut2005] S. Heutz, A. F. Nogueira, J. R. Durrant, T. S. Jones, “Charge Recombination in CuPc/PTCDA thin films”, J. Phys. Chem. B. 109, 11693-11696 (2005).

[Joha2005] P. Johansson, H. Xu, M. Käll, “Surface-enhanced Raman scattering and fluorescence near metal nanoparticles” ,Phys. Rev. B 72, 035427 (2005).

[Kanz2000] T. W. Kanzler, C. P. Holfeld, V. G. Lysenko, D. M. Whitaker, K. Köhler, and K. Leo, “Influence of Fano coupling on Wannier-Stark transitions in semiconductor superlattices”, phys. stat. sol. (a) 178, 39-44 (2000).

[Knei1997] K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, and M. S. Feld, “Single Molecule Detection Using Surface-Enhanced Raman Scattering (SERS)”, Phys. Rev. Lett. 78, 1667–1670 (1997).

[Knei2001] K. Kneipp, L. T. Perelman, H. Kneipp, V. Backman, A. Jorio, G. Dresselhaus, and M. S. Dresselhaus, “Coupling and scattering power exchange between phonon modes observed in surface-enhanced Raman spectra of single-wall carbon nanotubes on silver colloidal clusters”, Phys. Rev. B 63, 193411 1-4 (2001).

[Kobi2002] A. Yu. Kobitski, G. Salvan, R. Scholz, D. Tenne, T. U. Kampen, H. P. Wagner, D. R. T. Zahn, “Raman spectroscopy of the PTCDA–inorganic semiconductor interface: evidence for charge transfer”, Appl. Surf. Sci. 190, 386-389 (2002).

[Kobi2004] A. Yu Kobitski, R. Scholz, D. R. T. Zahn, “Theoretical Sudies of the Vibrational Properties of the 3,4,9,10-Perylene Tetracarboxylic Dianhydride (PTCDA) Molecule”, J. Molec. Struct. (Theochem.) 625, 39-46 (2004).

[McKe1998] N. B. McKeown, “Phthalocyanine materials: Synthesis, structure and function”, chap. 6, Cambridge University Press (monograph), (1998).

[Ménd2006] H. Méndez, “Organic modification of Metal / Semiconductor contacts”, PhD Dissertation, http://archiv.tu-chemnitz.de/pub/2006/0124/index.html, Technische Universität Chemnitz (2006).

[Möll2003] S. Möller, S. R. Forrest, C. Perlov, W. Jackson, and C. Taussig, “Electrochromic conductive polymer fuses for hybrid organic/inorganic semiconductor memories”, J. Appl. Phys., 94, 7811-7819 (2003).

[Nadi2005] N. Mamede, J. L. A. Sanchez de Almeida Prado, “Materiais híbridos orgânico-inorgânicos: preparação e algumas aplicações”, Quim. Nova 28, 281-288 (2005).

[ofet] Schwerpunktprogramm 1121: Organische Feldeffekt-Transistoren: strukturelle und dynamische Eigenschaften http://www.ofet.de, Coordinator Prof. Dr. Christof Wöll. Activity in Chemnitz coordinated by Prof. Dr. Dr. h.c. D. R. T. Zahn and Dr. R. Scholz under the reference Za 146/4-2 as part of SPP 1121: Organic field effect transistors: Structural and dynamical properties.

[Ostr1997] J. R. Ostrick, A. Dodabalapur, L. Torsi, A. J. lovinger, E. W. Kwock, T. M. Miller, M. Galvin, M. Beggren, and H. E. Katz, “Conductivity-type anisotropy in molecular solids”, J. Appl. Phys. 81, 6804-6808 (1997).

[Otto2001] A. Otto, “Theory of First Layer and Single Molecule Surface Enhanced Raman Scattering (SERS)”, phys. stat. sol. (a) 188, 1455–1470 (2001).

[Otto2005] A. Otto, “The chemical (electronic) contribution to surface-enhanced Raman scattering”, J. Raman Spectroscopy 36, 497-509 (2005).

[Paez2003a] B. A. Paez, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Interaction of metals with perylene derivatives as a model system for contact formation in OFET structures”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 210 (2003).

[Paez2003b] B. A. Paez, M. Bartzsch, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Combined Electrical and Raman characterization of C60 based organic field effect transistors”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 63-70 (2003).

[Paez2003c] B. A. Paez S, “Structural property studies based on Raman spectroscopy”, winter talk TU Chemnitz Halbleiterphysik 18.Nov.2003.

[Paez2004a] B. A. Paez, G. Salvan, S. Silaghi, R. Scholz, T. U. Kampen and D. R. T. Zahn, “Raman Monitoring of In and Ag Growth on PTCDA and DiMe-PTCDI Thin Films”, Appl. Surf. Sci. 234, 168-172 (2004).

[Paez2005a] B. A. Paez, G. Salvan, R. Scholz and D. R. T. Zahn, “Interface formation of Mg with DiMePTCDI studied by Raman spectroscopy”, phys. stat. sol. (c) 2, 4048-4052 (2005).

Chapter 4 B. A. Paez-Sierra, Metal / organic interface formation… 4.20

[Paez2005b] B. A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F 1-9 (2005).

[Pan1996] S. H. Pan, Z. H. Chen, K. J. Jin, G. Zhen Yang, Y. Huang and T. N. Zaho, “Fano interference in intrasubband Raman scattering of semiconductor superlattices” Z. Phys. B 101, 587-591 (1996).

[Park2002] S. G. Park, “Organic Modified GaAs Schottky contacts”, PhD Thesis, http://archiv.tu-chemnitz.de/pub/2002/0004/data/ TU Chemnitz (2002).

[Pers2006] B. N. J. Persson,1,2,3 Ke Zhao,4 and Zhenyu Zhang, “Chemical Contribution to surface-enhanced Raman scattering”, Phys. Rev. Lett. 96, 207401 1-4 (2006).

[Salv2003] G. Salvan, “Metal/Organic/Inorganic semiconductor heterostructures characterized by vibrational spectroscopies”, PhD thesis, http://archiv.tu-chemnitz.de/pub/2003/0109/index.html TU Chemnitz (2003).

[Salv2004a] G. Salvan, D.R.T. Zahn, and B. Paez, “Surface enhanced Raman scattering in organic thin films covered with silver, indium and magnesium”, J. Luminescence 110, 296-302 (2004).

[Salv2004b] G. Salvan, D. R. T. Zahn, “Evidence for strong interaction of PTCDA molecules with defects on sulphur-passivated”, Europhys. Lett. 67, 827-833 (2004).

[Salv2005] G. Salvan, B. A. Paez¸ S. Silaghi, and D. R. T. Zahn, “Deposition of silver, indium, and magnesium onto organic semiconductor layers: Reactivity, indiffusion, and metal morphology”, Microelectronic Engineering, 82, 228-235 (2005).

[Scho2005] R. Scholz, A.-D. Müller, F. Müller, I. Thurzo, B. A. Paez, L. Mancera, D.R.T. Zahn, “Comparison between the charge carrier mobilities in pentacene OFET structures as obtained from electrical characterization and potentiometry”, Proc. of SPIE 5940, 59400I 1-7 (2005).

[Schr2004] F. Schreiber, “Organic molecular beam deposition: Growth studies beyond the first monolayer”, phys. stat. sol. (a) 201, 1037–1054 (2004).

[Thur2005a] I. Thurzo, H. Méndez and D. R. T. Zahn, “Analysis of charge transient spectroscopy data originating from Gaussian densities of electron states in organics”, phys. stat. sol. (a) 202, 1994–2007 (2005).

[Thur2005b] I. Thurzo, H. Méndez and D. R. T. Zahn, “Dielectric relaxation in a hybrid Ag/DiMe-PTCDI/GaAs device”. J. non crystal. Sol. 351, 2003-2008 (2005).

[Thur2006] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, “Anomalous charge relaxation in channels of pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 203, 2326-2340 (2006).

[Tian2006] Y. Tian, Q. He, C. Tao, J. Li, “Fabrication of fluorescent nanotubes based on layer-by-layer assembly via covalent bond”, Langmuir 22, 360-362 (2006).

[Wagn2003] V. Wagner, T. Muck, J. Geurts, M. Schneider and E. Umbach, “Raman analysis of first monolayers of PTCDA on Ag(111)”, App. Surf. Sci., 212-213, 520-524 (2003).

[Witt2004] G. Witte and Ch. Wöll, “Growth of aromatic molecules on solid substrates for applications in organic electronics”, J. Mater. Res., 19, 1889-1916, (2004).

[Xue2004] J. Xue and S. R. Forrest, “Bipolar doping between a molecular organic donor-acceptor couple”, Phys. Rev. B 69, 245322 1-8 (2004).

[Zahn2004] D.R.T. Zahn, G. Salvan, B. A. Paez, R. Scholz, “Interaction between metals and organic semiconductors studied by Raman spectroscopy”, J. Vac. Sci. Technol. A 22, 1482-1487 (2004).

[Zahn2005] D. R. T. Zahn, G. Salvan, G. Gavrila, and B. A. Paez, “Chemistry and morphological properties of metal interfaces to organic semiconductors”, Adv. sol. sta. phys. 45, 313-324 (2005).

[Zahn2006] D. R. T. Zahn, G. N. Gavrila, and M. Gorgoi, “The transport gap of organic semiconductors studied using the combination of direct and inverse photoemission”, Chem. Phys. 325, 99-112 (2006).

[Zhou1993] P. Zhou, K. A. Wang, P. C. Eklung, G. Dresselhaus, M. S. Dresselhaus, “Raman-scattering study of the electron-phonon interaction in M3C60 (M=K,Rb)”, Phys. Rev. B, 48, 8412-8417 (1993).

[Zou2006] Y. Zou, L. Kilian, A. Schöll, Th. Schmidt, R. Fink, and E. Umbach, “Chemical bonding of PTCDA on Ag surfaces and the formation of interface states”, Surf. Sci. 600, 1240-1251 (2006).

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.1

Chapter 5

Organic field effect transistors

(OFETs)

In this chapter, the basic concepts of the organic field effect transistor (OFET) are presented. The main

topics discussed include the statistics of charge carriers, estimation of charge carrier density and

molecular density per cm2. Before this, a previous discussion concerning the Fermi integrals is

addressed.

After the statistics and transport description, the chapter focuses on the comparison between

experimental and theoretical OFET output characteristics. The analysis is done in order to identify

some of the drawbacks of the transistor equations. The electric field dependence of the mobility

( effμ )the threshold voltage ( TV ), the dynamics of boundary conditions, i.e., time dependence of traps

developing at interfaces and promoted in the organic material, structural relaxation, among others,

probe the necessity of reformulating the OFET modeling.

Special attention is dedicated to an analysis of existing formulations to describe the drain current in

different regimes, and some proposals are made in order to pursue a more reliable analysis of the

experimental data. As an example, a single channel device is considered. An accurate analysis of the

output characteristics data has shown that the field effect mobility ( effμ ) does not remain constant and

its relationship with the applied voltage is not limited to a simple electric gate-field relation. Although

models based on field-effect mobilities as a function of electric fields, i.e., the Frenkel-Poole

description, are available, these are limited to the sub-linear regime, meaning that an analysis of the

transistor as a switching device is not covered.

5.1. Introduction

The field effect transistor is a three-terminal device, where two leads serve to drive charge carriers

through a medium behaving as a channel and joining both contacts. A third terminal, isolated from the

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.2 other two, modulates the amount of charge density by an electric field. The first two terminals are

referred to as source and drain and the last one is called the gate terminal (cf. Figure 5.1.).

Further and novel details on the OFET device are

given in the present and next chapters. Concerning the

pioneering ideas about FETs, proposals date from

1926-1933 with Lilienfeld´s patented works about

"method and apparatus for controlling electric

currents” (1930 US Patent N. 1,745,175), "amplifier

for electric currents”, (1932 US Patent 1,877,140),

and "device for controlling electric current” (1933 US

Patent 1,900,018) [Klei1998]. The modern description

of an inorganic-based FET is found elsewhere

[Kaga2003, Kwok1995, Oda2006, Paul1994,

Raza2001, Shoc1952].

In the previous chapter, the metal contact formation on

Figure 5.1. Scheme of a field effect

transistor. The drain and source terminals

serve to drive the modulated current

through the channel shown in blue color.

The charge in the channel is modulated by

the capacitive effect of the third (isolated)

terminal named gate.

organic films was discussed. The metal/organic junctions on perylene derivatives have served as

model systems to investigate the physics developed at the organic/inorganic interface [Forr2003,

Paez2004, Xue2004, Zahn2004]. Vibrational spectroscopy has demonstrated great potential when

investigations of homo- and hetero-structure formation are required.

In the context of interface formation, the transistor is an effective means by which to probe the

formation of different boundaries, being one of the most appropriate configurations for the

organic/inorganic or organic1/organic2 scenarios to study several phenomena developed in the active

layer and its surroundings. The transistor geometry, electrodes configuration, active layer and many

other features make it a well-suited and almost closed system to identify not only charge transport

properties but many other combined properties that improve the knowledgement of basic signatures of

the organic channel.

After the experimental realization of the first organic field effect transistor based on polythiophene

[Tsum1986] and later on with small molecules, i.e., LuPc2 /ZnPc [Madr1987], numerous innovations

and intense research have been initiated. Applications in active matrices, such as flexible displays

driven by transistors with a polymer semiconductor [Bao2006, Bric2006, Huit2001, Kalu2006,

Some2005], printed pentacene and oligothiophene for RFID tags [Subr2005], and gas sensors

[Guil1998, Tane2005] are some of myriad examples involving the OFET device.

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.3 5.2. Statistical mechanics of charge carriers

5.2.1. Density of states

In this section, some preliminary ideas about the density of states ( εεΩ d)( ) and charge carriers

distribution are presented.

The density of states εεΩ d)( accounts for the number of cells per volume in the phase space,

h)()(with,dkdkdkV)(~zyx ε=ε==εΩ kpp , (5.1)

where V is the volume given by the configuration space, p the momentum, k the wave number and

h Planck´s constant divided by π2 .

,V

)(~2d)( εΩ=εεΩ (5.2)

Here the factor 2 accounts for the spin contribution. The accessibility of particles or quasiparticles is

determined according to the Fermi-Dirac or Bose-Einstein distributions ( )(f ε ) for fermions or bosons,

respectively. If traps are present in the material, then these are added to this expression.

Accordingly, the density of states given by eq.(5.2) is written as follows

∑=

ε−εδΛ+εΩ=εΩN

1ii )()()( , (5.3)

with )( iε−εδ being the Dirac delta function describing the discrete trap with energy iε and Λ a

scaling factor to keep the corresponding units.The total number of particles on per volume is given by

the probability of finding them in a given energy interval

εεεΩ= ∫∞

∞−d)(f)(n o , (5.4)

5.3.2 Charge carrier density

The charge carrier density is determined by means of eq.(5.4). Therefore, it is necessary to define or

measure the density of states ( εεΩ d)( ) in order to determine the charge density of the organic

material. In polyacenes, a quasi-constant density of states has been theoretically predicted [Rosa1991].

Then, the electron and hole carrier densities n and p are given through the equations

ε+−εΩ

= ∫∞

−ε d)exp(1)E(n

LUMO fETkE

LUMO , (5.5)

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.4

ε+

ε−Ω= ∫ ∞− ε− d

)exp(1)E(p HOMO

f

E

TkE

HOMO , (5.6)

with Ef Fermi level.

The expressions given by eqs. (5.5, 5.6) can be written in terms of a reduced energy using the

following substitutions

TkE~ LUMO−ε

=ε , (5.7)

TkEE LUMOF −=η , (5.8)

TkE~ HOMO ε−

=ε , (5.9)

TkEE FHOMO −

=η , (5.10)

where ε~ is the reduced energy and η the reduced chemical potential.

Another consideration is to replace the density of states by a power law of energy; a special case is the

electron gas in a metal, where the relation between energy and density of states obeys a square root

law ( 2/1d)( ε∝εεΩ ). In general, if there are constrictions of the electron gas, then the density of

states will hold those signatures and will involve modifications of the particle energy. Therefore, the

charge carrier density defined by eqs.(5.5, 5.6) is proportional to

εη−ε+

ε+Γ

=η ∫∞ ~d

)~exp(1

~

)j1(1)(F

0

j

j , (5.11)

this integral form is called the Fermi-Dirac integral and its order is defined by the power j .

The next section summarizes some values of the arguments and helps to identify whether the

molecular structure is highly-, middle- or non-degenerate. In this work, a MatLab [MATL2003] code

to evaluate the Fermi integral of any order was developed.

5.3.3. Fermi integral argument

In this section, some values of the quoted Fermi integral are shown; its complete evaluation is

performed in a developed code based on the Gauss-Legendre quadrature already described in chapter

3. Table 5.1 presents some values of the reduced (η ) and increment ( ηΔ ) of the chemical potential

along with the order (j) of the integral [Blak1987].

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.5

Table 5.1. Parameters reported for the Fermi-Dirac integral

[Blak1987]

η ηΔ j

]4,4[− 1.0 4;;3;;2;;1;;;1; 27

25

23

21

21

23 −−−

]10,4[ 2.0 4;;3;;2;;1;;;1; 27

25

23

21

21

23 −−−

As an example, Figure 5.2 shows the argument of some Fermi-Dirac integrals as a function of the

reduced energy ( ε~ ) and reduced chemical potential (η ).

(a) (b) (c)

Figure 5.2. Argument of the Fermi-Dirac integral as a function of the reduced energy ε , and the

reduced chemical potential η (evaluated in MatLab [MATL2003]). Orders (j) of the arguments

(a) 21 , (b) 2

1− , and (c) 0.

5.4. Charge carrier density of organic materials

In section 5.3.2, the integral expressions for the charge carrier density of organic materials were given.

A particular feature of aromatic molecules is that the band gap is approximately four times the transfer

integral ( ζ~

), as determined from tight binding calculations. Therefore, the interval for computing the

Fermi-Dirac integral indicated in eq. (5.5) is reduced to the interval ]~4,~4[ ζ−ζ− , and the charge

density can be determined as a consequence.

It has been demonstrated experimentally that the charge transport in OFETs is restricted to a two-

dimensional charge carrier gas [Muck2004, Paez2005]. Accordingly, the charge carrier density reads

⎥⎦

⎤⎢⎣

⎡

ζ+ηζ−ηζπ

=)~22/cosh()~22/cosh(ln

~m4pr

r2

r*

D2h

(5.12)

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.6 with Tk/~~

Br ζ=ζ the reduced transfer integral. The reduced chemical potential for pentacene is

about 34≈η at room temperature.

The charge density given in eq.(5.12) differs from that one proposed by Daraktchiev and colleagues

[Dara2005]. Presumably they considered further assumptions before obtaining their charge density

calculations.

As an illustration of the two-dimensional charge density given in eq.(5.12), the charge density as a

function of the reduced chemical potential and the reduced transfer integral is shown in Figure 5.3.

The estimated density values are scaled with respect to 2* /m4 hπ , *m being the effective mass of the

charge carriers; for pentacene o* m7.1m = and o

* m5.5m = [Wijs2003].

Figure 5.3. Charge density distribution as

a function of the reduced chemical

potential and reduced transfer integral.

The effective masses for pentacene are

o* m7.1m = and o

* m5.5m =

[Wijs2003]

Assuming a two-dimensional molecular packing and the experimental lattice constants of pentacene

determined by XRD measurements [Ruiz2004], then the number of molecules per 2cm is estimated

and given in table 5.2. The lattice constants a and b are parallel to the substrate while the molecule is

perpendicular and aligned on the diagonal of the rectangle formed by a and b (cf. Figure 5.4).

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.7

Figure 5.4. Triclinic crystal

structure to estimate the number

of molecules per 2cm .

Table 5.2. Lattice constants of the triclinic cell for pentacene

with cell parameters °≈α 978.76 , °≈β 136.88 , °=γ 415.84 ,

and density of molecules per 2cm .

The molecular densities summarized in table 5.2 are useful to

determine the number of charge carriers in the organic material.

More recent results have revealed a modified triclinic structure of pentacene submonolayers (SML)

and multimonolayers (MML) grown on native SiO2 [Ruiz2004]. Consequently, a scheme similar to the

one used in the former triclinic structure is employed for the modified crystal cell, and the molecular

densities are summarized in table 5.3.

Table 5.3. Lattice constants of the triclinic cell for SML and

MML Pentacene, with cell parameters °≈α 978.76 ,

°≈β 136.88 , °=γ 415.84 . The last column indicates the

density of pentacene molecules per 2cm .

a /Å b / /Å 2cm

P~

−

SML 7.62 * 5.90 * 141044.4 ×

MML 7.58 * 5.91 * 141046.4 ×

Pentacene

(distance H-H

central ring)

4.99

7.62

5.90

4.99

141078.6 × 141026.5 ×

*Lattice parameters reported in the literature [Ruiz2004]

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.8 5.5. The field effect transistor (FET)

A modified version of the FET architecture given in chapter 3 is presented here and illustrated in

Figure 5.5. The geometrical arrangement of the constituent elements is as follows: there is a substrate

on which an isolating layer of thickness id is deposited, on top of this layer there are patterned

contacts and in between a semiconducting material is deposited. Experimentally, the organic material

is deposited on the whole gate dielectric region, including the patterned contacts (cf. Figure 3.6). Here,

the addressed scheme shown in Figure 5.5 helps to strategically distinguish betweeen the different

boundaries and domains assembling the organic field effect structure.

During the experiments under UHV conditions, and prior to the deposition of the semiconducting

material, the current through any configuration of two contacts was tested to assure negligible leakage

currents, but this did not guarantee cleanliness of the gate dielectric between the top contacts on the

isolating layer. Therefore, organic-free paths on the dielectric were probed by Raman spectroscopy

measurements.

Before the deposition of the active layer, it had to be ensured that there was no current flowing

between the gate terminal and any of the top contacts.

The organic layer bridging the contacts is called channel and has length L, width W, thickness d, and

an effective thickness δ where the two dimensional charge transport mainly is developed (cf. Figure

5.5). A further discussion of the channel formation is presented in chapter 6.

Figure 5.5. Organic field effect transistor

To assure the device functionality, it is aplied

a voltage to the gate therminal, thereby an

excess charge density is induced in the

channel by the capacitive effect developed at

the isolating layer. The effect of the induced

charge is extended mainly to an effective

thickness (δ ) of the organic film (discussed in

chapter 6). Until this stage, the charge density

has only been modulated and no charge

transport between the electrodes joint by the

organic material is observed

The next step is to ground one of these contacts and set the other to a suitable potential driving the

gate-induced charge through the channel. The applied voltages depend on the majority charge carrier

of the organic layer and the work function at the metal organic interface. Finally, the grounded and

biased contacts forming part of the channel are named drain and source, respectively.

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.9 5.5.1. Energy band structure of an OFET

The energy band diagram of the organic material forming the different interfaces in the channel is

outlined in Figure 5.6. In the case of OFETs, the conduction and valence bands of the organic material

are replaced by the lowest unoccupied and highest occupied molecular orbitals LUMO and HOMO,

respectively.

(a) (b) (c)

(d)

Figure 5.6. Energy band diagram between the gate and the organic film with an isolating interlayer, (a)

under equilibrium conditions. (b) Accumulation mode of the organic field energy bands for negative

and (c) positive (c) gate voltages ( gV ) respectively. (d) Band diagram between the organic and the

contacts drain and source (UDS = 0 V).

The diagrams indicated in Figure 5.6 (a, b, c) correspond to the organic/insulator/gate (heavily doped

Si) structure; the first scheme (cf. Figure 5.6 (a)) indicates equilibrium conditions, while the second

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.10 and third diagrams depict the energy accumulation developed in the organic material when a negative

or positive gate voltage is applied. The different gate fields induce charge on the organic material. The

accumulated charge density per monolayer plane and parallel to the gate dielectrics is inversely

dependent on the square distance between them, i.e., monolayers away from the gate have lower

charge density in comparison with those close to the dielectric substrate (chapter 6).

The next energy diagram is the one between the molecular layer and the contacts source and drain (cf.

Figure 5.6 (d)). In this work, the metal contacts of the OFETs were gold. Therefore, a band diagram of

the interface Au/pentacene based on ultraviolet photoemission spectroscopy (UPS) and inverse

photoemission spectroscopy (IPES) measurements [Kahn2003] is documented in Figure 5.6(d). It

should be observed that the Fermi level (EF) alignment between Au/pentacene shows a more favorable

p type transport in the organic layer, but there is still the probability of having an n-type transport as

probed by charge transient spectroscopy (QTS) measurements.

5.6. Output characteristics of the OFET

5.6.1. "Linear” regime

As an example of the OFET output characteristics, Figure 5.7 shows an ideal and experimental drain

current -(Id) vs. drain-voltage (Vd) characteristics. The rising part of the dd VI − corresponds to the so

called "linear” regime, where the applied drain voltage does not exceed the potential at the organic

insulator interface induced by the applied gate bias. This means that the condition or the drain

potential )VV(V Td −< g must be fulfilled, with TV a threshold voltage or drop potential across the

organic/isolator/ interface.

The physical stationary (time-independent) description of a prototypic field effect transistor (FET) can

be found elsewhere [Kaga2003, Kwok1995, Oda2006, Paul1994, Raza2001, Shoc1952]. In the

isolated gate geometry, such as that shown in Figure 5.1, the field effect is produced by a capacitive

effect leading to an induced charge Q in the organic material, given by

)VV(CQ T−= g , (5.13)

with gV the gate potential, and TV the threshold voltage which takes into account the field free-

carrier concentration as well as details of the internal charge distribution and trapping at the interfaces

[Libs1993].

The induced charge density forms a region called an accumulation layer, with length L , width W ,

and thickness δ . As long as no other external field different from the gate bias is applied, the charge

density is confined to the volume δ××WL .

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.11 Otherwise, when a drain voltage is applied, there is a superposed electric field parallel to the channel.

Then the effective electric field is the contribution due to gate and drain, and the volume where the

charge density is distributed is modified. For drain voltages below the potential drop )VV( T−g , the

drain current density transported through the channel is proportional to the drain-source field and

given by

EJ σ=d , (5.14)

where σ is the effective conductivity of the channel, expressed as (in the one dimensional model)

eneffμ=σ , (5.15)

with effμ the effective field effect mobility, n the charge density and e the magnitude of the electron

charge.

Considering only the charge transport from source to drain and combining eqs.(5.13-5.15), the drain

current intensity ( dI ) is as follows

( )xdVd

V)VV(WLW

CenI dTeffeffd −−⎟⎟⎠

⎞⎜⎜⎝

⎛μ=μ= g , (5.16)

with io d/AC εε= the capacitance of the insulating gate material of thickness id , area LWA = and

dielectric constant ε .

Next step is to integrate eq. (5.16) from source to drain

( )∫∫ −−μ=dV

0Tseff

L

0d VdV)VV(WCxdI g , (5.17)

Td

2d

dTseff

d VVVwith2

VV)VV(

LWC

I −<⎟⎟⎠

⎞⎜⎜⎝

⎛−−

μ= gg , (5.18)

ios d/C εε= is the capacitance per unit area.

As mentioned above, eq.(5.18) counts only for the drain current intensity. This means that the charges

driven by the drain field are those induced solely by the gate field in the channel.

An additional contribution is the bulk current, which is due to the drain voltage without any applied

gate bias. The drain current given in eq.(5.18) has a parabolic dependence on the drain voltage dV ,

and it is linearly proportional to the drain field only if the condition Td VVV −<< g is satisfied.

The restriction on the drain voltage has some consequences. The first one is that the dynamics in the

channel do not fall in the event that the channel is pinched off. Second, the dd VI − is well described

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.12 by linear superposition of two contributions of the drain voltage; when )VV(V Td −<< g , then the

first-order contribution of dV to the drain current is dominant. In this region, the behavior of the drain

current with the drain voltage is linear.

The opposite happens when the quadratic component of dV is significant (eq. (5.18)). Then it is clear

that the latter situation no longer belongs to the linear regime. For this reason, the quotation marks ""

are used. The linearity of the drain current with the drain voltage holds if and only if

)VV(V Td −<< g .

5.6.2. Saturation regime

This regime occurs when the drain voltage cannot drive more charge to increase the intensity of the

drain current. The saturated drain current starts at the maximum drain value reached in the "linear”

(parabolic) regime. Thus from eq. (5.18)

0VdId

d

d = , (5.19)

this yields the maximum drain potential in the "linear” regime

Tmaxd VVV −=− g , (5.20)

2T

seffd )VV(

L2WCI −

μ= g . (5.21)

The dependence of the drain current on the drain voltage is depicted in Figure 5.7(a); the drain onset

potential joining both linear and saturation regions is clearly identified. In Figure 5.7(b), experimental

data of an organic field effect transistor (OFET) are shown, together with the simulated equations. The

first message of this inset is that the developed ideas are only approaches to the real data. Further in

this chapter, additional physical phenomena that can improve the theoretical model are considered.

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.13

(a)

(b)

Figure 5.7. Output characteristics of a field-effect transistor and maximum drain current where the

saturation starts. (a) Simulation of an ideal FET and (b) experimental drain current - drain voltage of a

pentacene (30 nm)-based FET; the solid lines correspond to the fitted output characteristics.

5.6.3. Field effect mobility effμ

An important quantity, not only for OFETs but also in other systems involving charge transport, is the

mobility, which gives information about the capability to drive charges in a well-defined medium

under internal and external fields.

The relations obtained in the previous section for the drain current involve the mobility; by means of

them, the field effect mobility in OFETs can be roughly estimated [Roic2004]

⎪⎪⎪

⎩

⎪⎪⎪

⎨

⎧

−≥−

−<

⎟⎟⎠

⎞⎜⎜⎝

⎛−−

=μ

Td2Ts

d

Td2d

dTs

d

eff

VVVwithregimesaturation,

2)VV(

LWC

I

VVVwithregime,,,

2VV)VV(

LWC

I

gg

g

g

linear"

(5.22)

A reference quantitity that shows how small changes in the gate field can modify the drain current in

the channel is

gm

g

gVI

lim d

0V ΔΔ

=→Δ

, (5.23)

the previous expression is usually presented in the literature as

Tddseff

.constV

dm VVVif,V

LWC

VI

−<<μ

=⎟⎟⎠

⎞⎜⎜⎝

⎛

∂∂

==

gg

d

g , (5.24)

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.14 from this expression, a value for the field effect mobility ( effμ ) in the so called "linear regime” is

obtained, when the quadratic contribution to the drain current is negligible, i.e, Td VVV −<< g . This

approximation is referred to as "gradual channel approximation”. Therefore, the field effect mobility

is found to be a constant.

In accordance with the parameters extracted from the fitted output characteristics of the experimental

drain current - drain voltage characteristics shown in Figure 5.7(b), the field effect mobility is not

constant and presents a field dependence as can be inferred from Figure 5.8. There is a dependence on

the applied gate voltage with a saturation voltage at about –15 V. Similar results have been reported in

n-type OFETs with n-alkyl perylene diimides as the active layer [Ches2004].

In chapter 7, the influence of electric fields on the molecular structure of the OFET channel is

discussed. It is experimentally demonstrated that the Raman signal increases with the applied gate

field and saturates at about –22 V. Therefore, charge carrier – vibration interaction mediated by

electrical polarization of the organic material is expected.

Figure 5.8. Field effect mobility in a pentacene

(30 nm)-based FET. The values (half filled

circles) were extracted from the output

characteristics, while the solid curve corresponds

to a power law with the applied gate field.

Withinn the framework of the previous approximation, some authors have also considered the

relationship between the transconductance ( mg ) and conductance ( dg ) [Horo1999, Kaga2003]

TdTseff

.constV

dd VVVif),VV(

LWC

VI

−<<−μ

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

==

ggd

g

g . (5.25)

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.15 Combining eqs.(5.24) and (5.25) yields [Horo1999]

Td2Ts

deff VVVif,

)VV(CV

WL

−<<−

==μ ggm

2d

gg

. (5.26)

A first conclusion from this result is that no effect of the bias gate voltage on the effμ is considered (a

constant value is assumed), contrary to what is found from the experimental results displayed in Figure

5.8. The quoted formulae derived under ideal conditions can lead to values of effμ with an uncertainty

of at least 100% [Roic2004]. These observations push to look for different procedures to determine

more realistic field-effect mobility values. One way is by fitting the output characteristics by a

correlated algorithm, i.e. all drain-current – drain voltages are fitted in a common set of simultaneous

equations.

For the previous fitting (cf. Figure 5.8), it can be determined that the mobility as a function of the

applied gate field follows a power law:

2k

1eff c)VV(csat

++−=μ g , (5.27)

with 62sat

61 105.6cand,81.1k,V03.16V,1003.4c −− ×==−=×= . The corresponding fitting is

presented in Figure 5.8 by the solid red line.

Actually, the mobility of pentacene-based transistors is about 112 sVcm5 −− [Kell2003]. The field

effect mobility can be improved in some extent if an appropriate surface treatment of the gate

dielectric is realized [Bao2006]. This subject was not investigated during the present work.

One should be aware of the measurement conditions, i.e., exposure to atmosphere, voltage-sweep

either the drain or the gate voltage. It has been observed that consecutive measurements of the output

characteristics makes the drain current lower; likewise, its monitoring over a period of several hours

has resulted in a decrease composed of multiple Debye-like decays (chapter 7-8) [Paez2005,

Thur2006, Gu2006]. It is a question whether the mobility is determined for the sub-linear regime, i.e.,

very small voltages, or derived from a flash measurement taking no longer than a few minutes.

It must be noted that eq.(5.27) differs from the expression for the mobility deduced from the "variable

hopping range” (VHR) described in chapter 2 eq.(2.5). Particularly here, the voltage satV does not

correspond to the threshold voltage. Additionally, the mobility dependence on the charge density

should be taken into account, since it involves natural and artificial trapps (chapter 8). The latter

trapping phenomenon is promoted by the applied fields leading to formation of dipoles which behave

as trapping centers for the charge carriers (chapter 8).

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.16 5.7. Threshold voltage shift and field dependence

The shift of the threshold voltage makes an additional contribution to the line-shape of the drain

current. For very low drain voltages, where the mobility has been shown to be dependent on the

applied field, the threshold shift is considered to stem from trapping of charge carriers at the

channel/isolator interface, accompanied by an extra trap-density developed in the conductive channel.

This is referred to as the defect-pool model originally applied to a-Si:H TFTs [Foma2005, Powe1987-

1992] and recently to organic based devices after having observed non-exponential drain current

relaxation [Stal2004].

Consequently, in the case of very low drain voltages, the current decay within the linear region can be

well described by a threshold voltage shift function, in which the effect of the gate voltage is involved

[Chia1998, Foma2005, Libs1993] and described by

⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

⎥⎥⎦

⎤

⎢⎢⎣

⎡⎟⎠⎞

⎜⎝⎛τ

−−Δ=Δβtexp1VV oT , (5.28)

with )kT/Eexp(o ττ=τ being the characteristic trapping time of carriers, and oVΔ approximately

the initial voltage drop value across the insulator towards the organic channel. In this model, the

thermal activation energy is given by β= τEEa , with β being the stretched-exponential exponent;

β=τ /EE a is interpreted as the average effective energy barrier that carriers in the a-Si:H channel

need to overcome before they can enter the insulator, and oτ is the thermal prefactor for emission

over the barrier.

The trap dynamics and threshold voltage shift in OFETs have been recently demonstrated for several

organic and polymer-channel materials, such as α -sexithiophene (α -T6), dihexylquaterthiophene

(DH4T), poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61-butyric acid methyl ester (PCBM)

[Gome2004-2005, Sall2003, Stal2004, Stre2003]. In the inorganic counterpart, such as hydrogenated

amorphous silicon [Powe1987-1992], the authors have found similar results.

During the present research, an interesting dynamic and anomalous behavior of the QTS spectra on

pentacene-based devices, with interdigitated source and drain contacts, and on single channel

structures were found. Part of the results are published elsewhere [Thur2006] and a further discussion

is addressed in chapter 8. The QTS results indicate the involvement of dipolar features, and deep

levels of minor charge carriers.

For some time the defect-pool model has been partially available in some commercial packages for

device simulation, but these tools are still lacking with regard to giving reasonable values in

comparison with those obtained from experimental measurements. The latest forecast for improving

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.17 organic-based device simulation involves the implementation of routines based on an extended

Frenkel-Poole [Gill1972] description and the Holstein [Well1996] model. These will obviously

improve the simulations of organic materials, but a dynamics of both boundary and initial conditions is

desired, phenomena proved in the present research by vibronic characterization of metal/organic

interface formation, and the combined Raman spectroscopy with electrical characterization of OFETs.

In the present investigation it has been experimentally demonstrated that the current

relaxation is related to the dynamics of the capture cross section (chap. 8).

As was stated in eq.(5.28), the threshold voltage probes the developed dynamics of the device

[Gomez2004]. Following the discussion of the experimental output characteristics described in Figure

5.8, the threshold voltage was determined and given in Figure 5.9. Assuming only the induced

capacitive charge, the threshold voltage can be considered as a quantity proportional to the trapped

charge at the organic/isolator interface. Therefore, with an organic channel of about 1.5 nm thick and a

surface capacitance of 2cm/nF23 , the trap density is about 218 cm10 − . The quoted value is so huge

that pure interfacial effects developed at the OBL/insulator interface can hardly reach or give a

comparable number of traps. The divergence between the computed trap density and the one resulting

from interfacial states suggests that additional trap sources in the device should be taken into account.

In chapter 8, the formation of additional traps -called "artificial traps” and generated by the applied

fields - is discussed.

Figure 5.9. Threshold voltage of a single

channel pentacene (30 nm)-based FET. The

values were extracted after fitting the output

characteristics depicted in Figure 5.8(b)

The threshold shift is also responsible for the

observed negative conductance in OFETs

(chapter 8) and the drain current collapse as

probed in inorganic-based FETs [Klei2003].

Although several scientific reports about OFETs

present this phenomenon in the output

characteristics, they are lacking a discussion or

explanation of its origins [Hepp2003,

Rost2004].

In chapter 6 the experimental and theoretical

descriptions of the Id-Vd characteristics are

given. Multi-exponential kinetics [Paez2005]

are assumed, where three drain current terms are

added together to reproduce the measured data.

One is independent of the trapping process,

while the other two are dependent on the

relaxation time constants of the OFET device.

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.18 Conclusions

A basic formula to model the OFET output characteristics was described. Signposts determined after

the confrontation with experimental data and the simulated results were established.

It was found that the mobility is strongly affected by the applied fields and its determination in the

sub-linear regime produces errors of about 100% in comparison with values obtained in the saturation

regime. In order to eliminate this substantial uncertainty, the set of equations describing the output

characteristics was considered in a self-consistent way, and all drain current–voltage measurements

were correlated by means of a back and forth procedure until the desired convergence with the

experimental data was achieved. The computation was performed by further extensions of the

correlated fitting algorithm discussed in chapter 3.

The detailed data analysis revealed that the extracted field-effect mobility follows a power law

proportional to the applied gate voltage. The mobility law presents a saturation at a given gate field;

afterwards it decreases, probably due to traps induced by the applied field.

Similar to the mobility, a dependence between the threshold voltage (VT) and the applied electric

fields was also determined. The VT is proportional to the trapped charge at the organic/isolator

interface; therefore, the extracted threshold-voltage values are strong indications that trapping effects

in the device should be considered. On the other hand, assuming only a charge induced by the

capacitive-gate structure, it was found that the amount of charge trapped at the interface is only a small

fraction of the huge relaxed charge observed, a finding prompting new concepts that will be discussed

in chapter 8.

Although there are models based on field effect mobilities and threshold voltage shifts as a function of

electric fields, they are limited to the sub-linear output regime of the OFET, meaning that the analysis

of the transistor as a switching device is not covered. There is a need to include temporal and field

dependences between the total drain current and the initial conditions. This would produce much

better correspondence between the experimental and simulated output characteristics.

References

[Bao2006] “Organic field-effect transistors V”, edited by Z. Bao and D. J. Gundlach, Proc. SPIE 6336, Washington 2006. [Blak1987] J. S. Blakemore, “Semiconductor Statistics”, Appendix B. Dover, New York, 1987. [Bric2006] A. L. Briceno, M. Roberts, M. -M. Ling, H. Moon, E. J. Nemanick, and Z. Bao, “Patterning organic

semiconductors using dry poly(dimethylsiloxane) elastometric stamps for thin film transistors”, JACS 128, 3880-3881 (2006).

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.19

[Ches2004] R. J. Chesterfield, J. C. McKeen, C. R. Newman, P. C. Ewbank, D. A. da Silva Filho, J.-L. Brédas, L. L. Miller, K. R. Mann, and C. D. Frisbie, “Organic thin film transistors based on N-Alkyl Perylene Diimides: Charge transport kinetics as a Function of gate voltage and temperature“, J. Phys. Chem. B 108, 19281-19292 (2004).

[Chia1998] C. S. Chiang, J. Kanichi, and K. Takaechi, “Electrical instability of hydogenated amorphous silicon thin-film transistors for AM-LCDs“, Jpn. J. of Appl. Phys. 37, 4704-4710 (1998).

[Dara2005] M. Daraktchiev, A. von Mühlenen, F. Nüesch, M. Schaer, M. Brinkmann, M.-N. Bussac, and L. Zuppiroli, “Ultrathin organic transistors on oxide surfaces”, 2005 New J. Phys. 7 133, 1-8 (2005).

[Foma2005] A. A. Fomani, “Threshold voltage instability and relaxation in hydrogenated amorphous silicon thin film transistors”,Master thesis, Waterloo, Ontario, Canada, (2005).

[Forr2003] S. R. Forrest, “Organic–inorganic semiconductor devices and 3, 4, 9, 10 perylenetetracarboxylic dianhydride: an early history of organic electronics”, J. Phys. Condens. Matter 15, S2599–S2610 (2003).

[Gill1972] 1 W. D. Gill, “Drift mobilities in amorphous charge-transfer complexes of trinitrofluorenone and poly-n-vinylcarbazole”, J. Appl. Phys., 43, 5033-5040, (1972).

[Gome2004] H. L. Gomes, P. Stallinga, F. Dinelli, M. Murgia, F. Biscarini, D. M. de Leeuw, T. Muck, J. Geurts, L. W. Molenkamp, and V. Wagner, “Bias-induced threshold voltages shifts in thin-film organic transistors”, Appl. Phys. Lett. 84, 3184-3186 (2004).

[Gome2005] H. L. Gomes, P. Stallinga, M. Murgia, F. Biscarini, T. Muck, V. Wagner, E. Smits, and D. M. de Leeuw, “Meta-stability effects in organic based transistors”, Proc. of SPIE Vol. 5940 59400K-1-8 (2005).

[Gu2006] G. Gu, and M. G. Kane, "Hytheresis in Pentacene-based organic thin-film transistors”, Organic Field Effect Transistors V, edited by Z. Bao and D. Gundlach, Proc. SPIE 6336-63360H, H1-H13 (2006).

[Guil1998] G. Guillaud, J. Simon, J. P. Germain, “Metallophthalocyanines: gas sensors, resistors, and transistors”, Coord. Chem. Rev., 178-180, 1433-1484 (1998).

[Hepp2003] A. Hepp, H. Heil,W. Weise, M. Ahles, R. Schmechel, and H. von Seggern, “Light-emitting field-effect transistor based on a tetracene thin film“, Phys. Rev. Lett. 91,157406 1-4 (2003).

[Horo1999] G. Horowitz, R. Hajlaoui, D. Fichou, and A. El Kassmi, “Gate voltage dependent mobility of oligothiophene field-effect transistors”, J. Appl. Phys. 85, 3202-3206 (1999).

[Huit2001] H. E. A. Huitema, G. H. Gelinck, J. B. P. H. van der Putten, K. E. Kuijk, C. M. Hart, E. Cantatore, P. T. Herwig, A. J. J. M. van Breemen, D. M. de Leeuw, “Plastic transistors in active-matrix displays. The handling of grey levels by these large displays paves the way for electronic paper”, Nature 414, 599 (2001).

[Kaga2003] Ch. R. Kagan and P. Andry, "Thin-film transistors”, Marcel Dekker, 2003, New York. [Kahn2003] A. Kahn, N. Koch, W. Gao, "Electronic structure and electrical properties of interfaces between metals and π -

conjugated molecular films” , J. Pol. Sci. Part B: Pol. Phys. 41, 2529 – 2548 (2003). [Kell2003] T. W. Kelley, D. V. Muyres, P. F. Baude, T.P. Smith, and T.D. Jones, “High prformance oganic tin film

transistors” Mat. Res. Soc. Symp. Proc. 771, L6.5.1-L6.5.11 (2003). [Klei1998] Ch. Kleint, "Julius Edgar Lilienfeld: Life and Profession”, Prog. Sur. Sci., 57, 253-327 (1998). [Klei2003] P. B. Klein and S. C. Binari, “Photoionization spectroscopy of deep defects responsible for current collapse in

nitride-based field effect transistors”, J. Phys.: Condens. Matter 15, R1641–R1667 (2003). [Kwok1995] Kwok K. Ng., “Complete guide to semiconductor devices”, McGraw-Hill, New York, 1995.

[Libs1993] F. R. Libsch and J. Kanicki, “Bias-stress-induced stretched-exponential time dependence of charge injection and trapping in amorphous thin-film transistors“. Appl. Phys. Lett. 62, 1286-1288 (1993).

[Madr1987] M. Madru, G. Guillaud, M. Al Sadoun, M. Mairot, C. Clarisse, M. Le Contellec, J. –J. André and J. Simon, “The first field effect transistor based on an intrinsic molecular semiconductor”, Chem. Phys. Lett. 142, 103-105 (1987).

[MATL2003] MATLAB® 6.1, The MathWorks, Inc. 2003. [Muck2004] T. Muck, V. Wagner, U. Bass, M. Leufgen, J. Geurts, L.W. Molenkamp, “In situ electrical characterization of

DH4T field-effect transistors”, Synth. Met. 146, 317–320 (2004). [Oda2006] “Silicon nanoelectronics”, edited by Sh. Oda and D. Ferry, CRC, Boca Raton 2006. [Paez2004] B. A. Paez, G. Salvan, S. Silaghi, R. Scholz, T. U. Kampen and D. R. T. Zahn, “Raman Monitoring of In and

Ag Growth on PTCDA and DiMe-PTCDI Thin Films”, Appl. Surf. Sci. 234, 168-172 (2004). [Paez2005] B. A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman

spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F 1-9 (2005).

[Paul1994] R. Paul, “MOS-Feldeffekttransistoren”, Halbleiter-Elektronik 21, (Springer-Verlag, Berlin, 1994). [Powe1987] M. J. Powell, C. van Berkel, I. D. French, and Nicholls, “Bias dependence of instability mechanisms in

amorphous silicon thin film transistors”, J. Appl. Phys. 51, 1242-1243 (1987). [Powe1992] M. J. Powell, C. van Berkel, A. R. Franklin, S. C. Deane, and W. I. Milne, “Defect Pool in amorphous-silicon

thin-film transistors”, Phys. Rev. B 45, 4160-4170, (1992). [Raza2001] B. Razavi, “Design of analog CMOS Integrated circuits”, McGrawHill, Boston, 2001. [Roic2004] Y. Roichman, Y. Preezant, and N. Tessler, “Analysis and modelling of organic devices”, phys. Stat. Sol. (a)

201, 1246-1262 (2004). [Rosa1991] A. L. S. da Rosa and C. P. de Melo, “Electronic properties of polyacene”, Phys. Rev. B 43, 2183-2186 (1991). [Rost2004] C. Rost, D. J. Gundlach, S. Karg, and W. Rieß, “Ambipolar organic field-effect transistor based on an organic

heterostructure”, J. Appl. Phys. 95, 5782-5787 (2004). [Ruiz2004] R. Ruiz, A. C. Mayer, G. G. Malliaras, B. Nickel, G. Scoles, A. KazimirovH. Kim, R. L. Headrick and Z.

Islam, “Structure of Pentacene thin films”, Appl. Phys. Lett., 85, 4926-4928 (2004). [Sall2003] A. Salleo and R. A. Street, "Light-induced bias stress reversal in polyfluorene thin-film transistors”, J. Appl.

Phys. 94, 471-479 (2003).

Chapter 5 B. A. Paez-Sierra, Organic field effect transistors 5.20

[Shoc1952] W. Schockley, “A unipolar field effect transistor”, Proc. Of the I.R.E. 1365-1376 (1952). [Some2005] T. Someya, Y. Kato, S. Iba, Y. Noguchi, T. Sekitani, H. Kawaguchi, and T. Sakurai, “Integration of organic

fets with organic photodiodes for a large area, flexible, and lightweight sheet image scanners”, IEEE TRANSACTIONS ON ELECTRON DEVICES 52, 2502-2511 (2005).

[Stal2004] P. Stallinga, H. L. Gomes, F. Biscarini, M. Murgia, and D. M. de Leeuw, “Electronic transport in field-effect transistors of sexithiophene”, J. Appl. Phys. 96, 5277 (2004).

[Stre2003] R. A. Street, A. Salleo, and M. Chabinyc, “Bipolaron mechanism for bias-stress effects in polymer transistors”, Phys. Rev. B 68, 085316 1-7 (2003).

[Subr2005]

V. Subramanian, J. M. J. Frechet, P. C. Chang, D. C. Huang, J. B. Lee, S. E. Molesa, A. R. Murphy, D. R. Redinger, and S. K. Volkman, “Progress toward development of all-printed RFID tags: Materials, processes, and devices”, Proc. IEEE 93, 1330-1338 (2005).

[Tane2005] M. C. Tanese, D. Fine, A. Dodabalapur b, L. Torsi, “Interface and gate bias dependence responses of sensing organic thin-film transistors”, Biosensors and Bioelectronics 21, 782–788 (2005).

[Thur2006] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, “Anomalous charge relaxation in channels of Pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 203, 2326-2340 (2006).

[Tsum1986] A. Tsumura, H. Koezuka, and T. Ando, “Macromolecular electronic device: Field-effect transistor with a polythiophene thin film”, App. Phys. Lett. 49, 1210-1212 (1986).

[Well1996] G. Wellein, H. Röder, and H. Fehske, “Polarons and bipolarons in strongly interacting electron-phonon systems”, Phys. Rev. B 53, 9666–9675 (1996).

[Wijs2003] G. A. de Wijsa, C. C. Mattheusa, R. A. de Groota, T. T.M. Palstraa, “Anisotropy of the mobility of Pentacene from frustration”, Synthetic Metals 139, 109-114(2003).

[Xue2004] J. Xue and S. R. Forrest, “Bipolar doping between a molecular organic donor-acceptor couple”, Phys. Rev. B 69, 245322 1-8 (2004).

[Zahn2004] D. R. T. Zahn, G. Salvan, B.A. Paez, R. Scholz, “Interaction between metals and organic semiconductors studied by Raman spectroscopy”, J. Vac. Sci. Technol. A 22, 1482-1487 (2004).

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.1

Chapter 6

Combined Raman spectroscopy and electrical characterization

of the conductive channel in OFETs

In this chapter, the combined vibrational and electrical measurements of the organic layer in OFETs

are discussed. During the deposition of pentacene on a Si-SiO2 gate structure with Au bottom contacts

for source and drain, the film growth was monitored with simultaneous in situ macro Raman

spectroscopy and drain current (Id) measurements of the OFET. The deposition of the active layer was

carried out under UHV conditions at a growth rate of 0.6 Å/min.

The aim of the in situ characterization was to determine the minimum nominal thickness of the

pentacene layer required for efficient charge transport through the OFET circuit. It was found that at a

thickness around 1.5 nm nominal coverage, the first percolation paths through the first organic

monolayer develop, resulting in a sharp rise of the drain current. Up to a nominal film thickness of

about 5 nm, a subsequent slower increase of the drain current can be observed, revealing that the

percolation of pristine monolayers continues at a slower pace up to rather thick organic layers

contributing to a very minor extent to the overall device drain current.

Throughout the chapter, the conductive channel is referred to as an "organic boosting layer (OBL)”.

The interpretation of the OBL is based on some ideas used for inorganic transistors, in which the

active layer for particular configurations is divided into two arrangements.

The first set of monolayers closest to the gated dielectric is composed of a strained material

[Abst1985, Cell2005, Coll2003, Drak2003, Gámi2002, Lamm2003, Wu2004]. The second one, on top

of the first, is the normal layer with the corresponding bulk signatures. It is observed that the charge is

flowing mainly through the strained material and a much lower amount is transported through the

bulk material. Electrical and structural properties of both films might reflect differences between their

physical properties, i.e., the effective mass of the charge carriers, mobility, lattice constants, and/or

vibronic features.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.2

Within the framework of this research, the OBL is outlined by two properties: (i) the minimal number

of monolayers forming the conducting channel where the maximum percentage of charge carriers are

transported; and (ii) by its strained signature evidenced from vibronic features when the OBL forms

different interfaces in the OFET. It is shown that the OBL behaves as a two- dimensional confinement

for the charge carriers [Paez2005a]. Obviously, the OBL concept might be modified by the

configuration of the electrodes.

6.1. Introduction

As was already discussed in chapters 1 and 2, organic-based electronics has become very attractive for

technological applications. The delocalized carriers in the π molecular orbitals make organic

molecules suitable for charge transport. Devices based on organic materials are cheaper and easier to

manufacture than their inorganic semiconductor-based counterparts.

Some prototype organic materials can be printed out or sublimated on bendable substrates [Ridl1999,

Sun2005], thus providing a new trend in the electronic market. Indeed, several applications of organic-

based electronic devices are found as price tags, plastic electronics, transparent displays, and non-

volatile plastic memories with gate insulators based on polymer ferroelectric materials [Brab2001,

Kim2005, Liab2005, Rees2004, Yoo2001].

These interesting features, among many others, have been the focus of several experimental and

theoretical investigations in order to understand the organic/inorganic or organic1/organic2 interface

formation. In particular, the optimization of device performance is of great interest [Liab2005,

Sche2005, Schr2004, Stre2005]. In this work, the investigation of charge transport and the conductive

channel formation in organic field effect transistors is carried out by means of vibronic and electrical

characterization techniques. The experimental research was performed by in situ drain current-voltage

(Id -V) and Raman spectroscopy measurements.

The vibrational spectrum is a characteristic fingerprint of the investigated system. Therefore, Raman

spectroscopy facilitates the identification of the molecular vibrations and thereby provides information

about normal mode frequencies of the molecule under different conditions. Accordingly, information

concerning charge state processes at interfaces, or structural order in organic materials, can be easily

distinguished by using this non-destructive spectroscopic technique [Chen2003, Colo2003,

Paez2005a-b, Salv2005, Zahn2001-2005].

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.3

In addition, the drain current monitoring provides evidence of the minimal amount of the organic

material required to reach the charge transport in the OFET channel. Both techniques were employed

during the molecular beam deposition of the pentacene molecules. Therefore, correlations between the

pentacene film thickness with vibrational properties and the drain current were outlined. The vibronic

and drain current monitoring was achieved on sub-monolayer coverage of the organic material with

rather thicker films of approx. 30 nm.

6.2. Simultaneous in situ I - V characterization and molecular vibration

measurements of OFETs

Prior to the molecular evaporation, the basic structures were electrically tested for possible leakage

currents. Additionally, the Raman spectrum was taken to assure the absence of organic contaminants

on the substrate before starting the organic deposition. Along with the step-wise molecular beam

deposition, the substrate was biased with a V10V −=d and V6V −=g . This allowed the continuous

monitoring of the drain current as a function of the organic film thickness.

The 3D plot shown in Figure 6.1 presents the evolution of the pentacene Raman bands in accordance

with the film thickness. On the right hand side, the vertical plane depicts the drain current as a function

of the organic layer thickness. Some intermediate Raman spectra were intentionally skipped for better

presentation clarity.

From this figure, it can be clearly seen that with pristine organic molecule depositions, i.e., below 1.5

nm nominal thickness, the Raman bands start to appear while the drain current is approximately zero.

This indicates the formation of molecular clusters which do not percolate to form the conducting

channel. At a thickness of about 1.5 nm, the drain current increases abruptly and with further organic

molecule deposition there is an asymptotic increase of the Id. From this result, it is deduced that the

transport takes place in layers with intimate contact or very close to the gated dielectric.

The first monolayers form the OFET channel and behave as an organic boosting layer (OBL), since a

high percentage of the nominal drain current is transported. A similar effect has been observed in

strained-Ge buried-channel MOSFET structures [Shan2006] with a thickness of about 4 nm.

In situ experiments on mobility as a function of the molecular channel thickness in

dihexylquaterthiophene (DH4T)-based FETs were conducted by the group of Prof. Wagner and

colleagues [Muck2004] It has been experimentally demonstrated that the main charge transport for

this structure is developed in the first two monolayers. Similarly, experiments on sexithienylbased

FETs revealed that the charge carrier mobility rapidly increases with increasing coverage and saturates

at a coverage of about two monolayers [Dine2004].

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.4

Figure 6.1. Simultaneous monitoring of

Raman bands (black spectra) and drain

current (red curve on the right hand side of

the vertical plane) during the pentacene

deposition. The pointed Raman bands

correspond to the in-plane C-C ring and in-

plane C-H vibrations.

Further analysis of the present results (sec. 6.3) shows that the sub-monolayer pentacene deposition of

about 1.5 nm is significantly affected by the substrate. Molecules on the electrodes present a splitting

of the band at 1157 cm-1, while those on the gate dielectric are much less affected. This indicates that

the organic/inorganic interface is decisive for the OBL formation.

Additionally, detailed analysis of the molecular growth and Raman bands were extracted by fitting

each spectrum. The data evaluation was performed by using the correlated fitting algorithm described

in chapter 3. In Figure 6.2(a), some extracted Raman spectra at organic thicknesses of about 0.2, 1.5,

5.4, 10.2 and 33.3 nm, respectively, are illustrated from bottom to top. The spectra are normalized

with respect to the band at 1179 cm-1.

In order to have a better comparison, each experimental spectrum is plotted together with its fitting

and the corresponding line-shapes used for the data evaluation. Results indicate a red shift of the Ag

band at 1158 cm-1 for organic coverages above 10 nm. This is mediated by the bulky pentacene

formation. Additionally, in the same figure, the vibration symmetries of the considered line-shapes are

highlighted.

The peak position and symmetry assignments are based on density functional calculations carried out

with the three parameter hybrid functional B3LYP and the 3-21G basis set in the Gaussian 98 package

[Gaus1998]. The theoretical evaluation was done for a single pentacene molecule.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.5

1140 1160 1180 1200

Pentacene based OFETsRaman spectra at different thickness

Raman shift / cm-1

Inte

nsity

/ a.

u

A g 11

79.2

cm-1

? 118

0.4 cm

-1

? 11

77.8

cm-1

B 1g 11

63.8

cm-1

A g 11

55 cm

-1

5x10-2 mW. cts-1. s-1 A g

1158

cm-1

λ = 647.1 nm130 W cm-2

0.2 nm/ 1

1.5 nm/ 7

5.4 nm/ 7.2

10.2 nm/ 11

33.3 nm/ 108.4

(a)

(b)

Figure 6.2. Extracted intermediate Raman spectra of pentacene based OFETs at different organic layer

thicknesses, together with the corresponding fitting line-shapes (a); and comparison between the area

of the Raman band at 1179 cm-1 and the drain current as a function of the organic film thickness (b).

The drain current as a function of the organic film thickness is shown in Figure 6.2(b); the dependence

of the Raman intensity on the band at 1179 cm-1 is depicted here as well. The combined Figure 6.2(b)

shows a strong correlation between the electric current intensity and the molecular coverage below 1.5

nm, indicating the transition between separate molecular clusters and their percolation onset to form

molecular layers, and hence the channel for the corresponding charge transport.

Analyzing the total organic deposition of approximately 33 nm and the abrupt increase of the drain

current, it is concluded that the charge transport is mainly developed below 5 nm, for which thickness

approximately 70 % of the total charge is driven. Additionally, the development of the Raman

intensity of the band at 1179 cm-1 provides an indication that basically six molecular arrangements

should be considered, as follows:

• Overlayers below 1.5 nm: These are governed by the formation of molecular clusters which

seem to be isolated, since no drain current is registered. A pentacene layer of about 1.5 nm can

be distinguished as the threshold thickness between the isolated molecular clusters and the

layer percolation. This is suggested by the dramatic increase in the drain current.

• Coverages between 1.5 nm and 10 nm: This interlayer is likely to be characterized by the

formation of more compact films. Therefore, the higher percentage of the drain current

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.6

confirms the nature of the OBL in comparison with that obtained at 33 nm pentacene film

thickness.

Considering molecules in upright position and from the geometry optimization performed

with density functional theory, it is found that the height of a single pentacene molecule is

about 1.41 nm. Therefore, the conductive channel ranges between 1.1 ML and 7 ML of

pentacene.

Measurements with AFM of pentacene growth on several substrates have shown a terrace-like

growth [Kell2006, Ruiz2004, Tiba2003] where a sub-monolayer coverage of 0.5 nm

pentacene on SiO2 exhibited a lateral size of approximately 0.5 μm [Ruiz2004]. It is

interesting to note that when 5.7 nm of pentacene is deposited on Co substrates at 360 K, six

different height levels of terracing are revealed [Tiba2003].

• Between 10 nm and 20 nm: This regime is also manifested by a kink in the Raman bands

intensity as a function of the film thickness. Formation of substantial molecular islands and a

higher number of molecules in contact with the drain and source electrodes are likely. This

intermediate thickness contributes much less to the total drain current. Roughly this can be

explained by the square inverse dependence of the distance between the molecular plane

position and the gate dielectric interface (sec. 6.5).

• Between 20 nm and 25 nm thickness, the indicated Raman band intensities in Figure 6.2 (b)

saturate, forming a plateau, while the broadening of the pristine clusters is increased.

• Above 25 nm: the current does not increase significantly, showing an asymptotic dependence

on the molecular film thickness.

Figure 6.3(a) shows the plotting of the remaining Raman band intensities as a function of the

molecular layer thickness. The trend of these bands is similar to that given by the profile of the band at

1179 cm-1 as a function of the layer thickness and addressed in Figure 6.2(b). This indicates that each

of the Raman bands develops in a similar way as the organic molecules are deposited.

Additionally, the FWHM of the fitted bands in relation to pentacene deposition are illustrated in

Figure 6.3(b). It is found that there is a nearly linear dependence of the FWHM tendency on the

organic film thickness, with a slope of about 0.045 cm-1 /nm. The broadening of the features is

attributed to the modified molecular scenario. Therefore, a larger scattering of the FWHM is expected.

The behavior of the band at 1163 cm-1 is surprising; the FWHM decreases with molecular deposition.

This band - as can be seen from the inset of Figure 6.3(b) - has the highest broadening and makes a

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.7

tiny contribution to the overall spectral region in comparison to other bands. A detailed observation of

the 1163 cm-1 band reveals that the linear dependence is modulated by oscillations whose maxima and

minima correspond fairly well to the characteristic domains already discussed in the previous

paragraph.

0 5 10 15 20 25 30

Pentacene based OFETsband-area in dependenceof organic thickness

1155 cm-1

1157 cm-1

1163 cm-1

1178 cm-1

1179 cm-1

Pentacene thickness / nm

Pea

k ar

ea /

a.u

(a)

0 5 10 15 20 25 30

2

4

6

8

10

12FW

HM

/ cm

-1

Pentacene thickness / nm

Pentacene based OFETs FWHMin dependence of organic thickness

1155 cm-1

1157 cm-1

1163 cm-1

1178 cm-1

1179 cm-1

1180 cm-1

(b)

Figure 6.3. Fitting parameters in dependence on the molecular layer thickness in pentacene-based

OFETs (a) Raman band intensities and (b) FWHM broadening.

Although a low nominal coverage of about 10 nm is enough to have charge transport in the OFET, it is

not sufficient to protect the channel from deterioration when it is exposed to atmospheric conditions.

Therefore, a thicker layer of the organic material or a capping material is required in order to obtain

working devices under environmental conditions.

6.3. Organic boosting layer (OBL) in OFETs

In this section, the vibrational analysis of the sub-monolayer coverage is addressed. The Raman bands

of 1.5 nm pentacene growth on SiO2 and on Au/SiO2 substrates are illustrated in Figure 6.4. The

difference between the spectra indicates that the substrate significantly modifies the way the pentacene

grows. This can be inferred from the different line-shape and the splitting of the band at 1157 cm-1 for

molecule growth on Au. Molecules deposited on SiO2 substrates do not present selection rules in the

Raman spectrum.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.8

Results reported by Ruiz et al. on grazing incidence, x-ray diffraction, x-ray reflectivity and atomic

force microscopy have proved that the crystal cell parameters of sub-monolayer (0.5 nm) pentacene

differ from those for bulky layers (19 nm). A common signature of both regimes, sub-monolayer and

multiplayer, is that the volume of the unit cell is nearly unmodified [Ruiz2004-2005].

1140 1160 1180 1200

Raman shift / cm-1

Inte

nsity

/ a.

u

Splitting

A g 11

79.5

cm-1

? 118

0.9 cm

-1

? 11

77.1

cm-1

B 1g 11

65.4

cm-1

A g 11

56.6

cm-1

Pentacene(1.5 nm) forming different interfaces in FETs

2x10-2 mW. cts-1. s-1

A g 11

59.4

cm-1

λ = 647.1 nm130 W cm-2

on SiO2/ 1

z(yy)z' = z(xy)z'

z(yy)z'

z(xy)z' on Au / SiO2

/ 29

on Au / SiO2/ 76

Figure 6.4. In situ Raman band measurements of pentacene (1.5 nm) forming different interfaces when

deposited on Au and SiO2 substrates.

Table 6.1 summarizes the vibratioal bands of ML and bulk pentacene. For comparison, reported

vibrational values and those measured in the present work are presented.

The Raman band at 1163 cm-1 and measured in the z(xy)z´ geometry for the sub-monolayer (1.5 nm)

coverage is difficult to detect when the sample is exposed to the atmosphere. Additionally, this band

coincides with a broader feature observed in thicker pentacene layers (30 nm). The sub-monolayer

regime is governed by the Davydov splitting, whereas for the layers the bulky effect of the film

contributes, thereby hiding the Davydov splitting. This band has been assigned to the B2u symmetry

[Ross2002]. Experimental measurements based on generalized infrared spectroscopy ellipsometry

gIRSE have delivered the same symmetry [Schu2004].

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.9

The last column in table 6.1 shows the alteration of the Raman shift, indicating the deformation of

pentacene monolayers in comparison with bulk pentacene. An interesting feature of the OBL is that it

exhibits a compressive deformation, demonstrating the phenomenon of strain due to the pristine

coverage in closest contact to the inorganic substrate (Au or SiO2).

For the spectral region indicated in table 6.1, the elastic potential energy per molecule is modified,

presumably leading to an increase in the number of molecules per unit volume. Therefore, a higher

layer compactness in conjunction with an increase of the molecular orbital overlapping is likely.

Table 6.1. Raman shift of monolayer and bulk pentacene thin films

Symmetry Calculated

/ cm-1

Reported

/ cm-1

Symmetry

/ cm-1

ML

(Exp.)

/ cm-1

Bulk

(Exp.)

Elasticity

(ML-Bulk)

Ag 996a 996.4 995.7 compressive

B1g 1125a

Ag 1156.6 1155 compressive

Ag 1158a Ag 1159.4 1157 compressive

Davidov 1159b (pellets)

B1g (IR B2u

c?)

1173 1163a,c B1g 1165.4 1163 compressive

? 11778 1177.8 compressive

Ag 1178a Ag 1179.5 1179 compressive

1180.9 1180 compressivea Experimental values of Matteheus [Matt2002] on 1.54 nm pentacene bMeasured in pellets IR and theoretical calculation based on mean-field theory and HF/6-31G(d)

B3LYP / 631G(d) [Ross2002] cMeasured in pentacene thin films [Schu2004]

6.4. Characteristic regions of the organic layer in OFETs

The analysis of the experimental Raman bands in the previous sections revealed a stratified type of

pentacene growth. Submonolayer Raman features below 5 nm are different than the thicker overlayers,

demonstrating a coexistence of more than one crystal structure. Similar behavior has been observed in

other organic layers, such as VOPc and perfluorinated -VOPc [Hash1999a-b]; the different layer

arrangement might be due to the molecular dipole moment.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.10

The organic layer building up the channel can be divided into three main regions (cf. inset Figure

6.5(a)). The first one is described by the metal/organic interface, where pentacene layers are

influenced by the Au contact to an extent of 10 nm [Amy2005]. The second region is the effective

OBL, with a thickness below 10 nm. Raman measurements have proved that the OBL is formed by the

strained layer of about 1.5 nm [Paez2005a-b].

In addition, reported HREELS measurements by S. D. Wang and colleagues of pentacene growth on

SiO2 [Wang2005] demonstrated two stages of the 4.2 nm pentacene film. The first one was composed

of a layer with a thickness of ∼2.6 nm, while the second one was ∼1.6 nm thick.

A similar finding was observed by Ruiz et al. [Ruiz2005, Maye2004] through in situ synchrotron x-ray

scattering. Further investigations evidenced that interactions between the organic molecules and the

substrate during the deposition have dramatic effects on the crystallinity of the thin film and thus on

the resulting electronic properties [Vere2004]. The third region can be referred to as the bulk material

that contributes, to a minor extent, to the drain current.

Figure 6.5. Principal regions of the

organic layer forming the channel in

OFETs

6.5. Organic – insulator electrodynamics

Assuming that the organic material is already deposited onto the gated dielectric, a first consideration

is to assume the molecular arrangement as a domain composed of a compact layer (cf. Figure 6.6(a))

and with a nominal thickness below 10 nm as determined by the combined in situ Raman and

electrical measurements. The dynamic molecular arrangement of this layer is also given by

intermediate states. This means that molecules in closest contact with the substrate initially form

clusters (below 1.5 nm), then islands, and finally percolates, forming a compact layer.

From the point of view of charge transport, the organic/insulator interface acquires a capacitive charge

which in first approximation and under steady state conditions follows the Mott equation [Mott1938],

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.11

indicating an inverse square root law of the capacitive induced charge density in the channel and the

applied field through the isolator. This sequence of ideas leads to the charge density given by

⎟⎟⎠

⎞⎜⎜⎝

⎛ ϕ−ϕ=′=ρ

Tk)(q

expnqnq sss , (6.1)

where

erfacetheatdensityechnpotentialerface

s

s

intarg:int:ϕ

The charge density is determined by solving the Poisson equation

ερ−=ϕ∇ /2 , (6.2)

with ϕ the electrostatic potential and related to the electric field )( E perpendicular to the channel

plane

ϕ−∇=E , (6.3)

the induced charge density n is determined as a function of the applied gate voltage gV and the

location x of each monolayer parallel to the channel, 2

Ds

2i

s xL2

q1Tk2

)V(Cn

−

⎟⎟⎠

⎞⎜⎜⎝

⎛+

ε=′ g , (6.4)

with iC the gate capacitance per unit area and DL the Debye´s length. The surface charge density of

each layer as a function of its location parallel to the gated dielectric is depicted in Figure 6.6(b).

In section 6.2, it was experimentally determined that 1.1 ML are enough to have charge transport in

the channel. Below this coverage, the drain current is negligible. Within the framework of the induced

charge density by the gate voltage, it is also plausible to assume that the clusters are affected by the

applied gate field. Therefore, for the steady state conditions and for a channel thickness of about 7

ML, the charge dependence illustrated in Figure 6.6(b) gives, as a first approximation, an estimate of

the two- dimensional charge carrier gas density. One must be careful with eq. (6.4), since the two-

dimensional charge density for layers thicker than 7 ML exhibit a terrace-like morphology (suggested

in sect. 6.2).

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.12

Monolayersorganic

0Gateddielectric

X a b

OBL

1 2 3 4 5 6 7 8 9 10

Cha

rge

dens

ity /

a.u.

monolayer position / x

Gate voltage -1 V -2 V -3 V -4 V -5 V

Figure 6.6. Charge density distribution induced in the organic layers by applying different gate

voltages. (a) Monolayer stratification and (b) dependence of the charge density as a function of the

layer location parallel to the gated dielectric with the gate voltage as parameter.

6.6. Vibrational band profiling of the active layer

In situ Raman scanning spectroscopy (RSS) measurements between the organic/Au and organic/SiO2

structures were carried out. The addressed experiment allowed the influence of the substrates and

interfaces on the organic channel formation to be identified.

The sample was RSS scanned as shown in Figure 6.7(a), where the laser spot on the organic/inorganic

interface is sketched. A detailed behavior of the measured Raman bands is shown in Figure 6.7(b),

where the spatial scale is basically divided into two regions. The first one corresponds to pentacene

molecules on the gold substrate, ranging from 0 μm to 900 μm. The second one corresponds to

pentacene on the SiO2 substrate, starting from 900 μm to 1100 μm. For better recognition of the

organic/inorganic interface, it was subtracted from the 3D plot; the intensity of the mode at 1179 cm-1,

which corresponds to the C-H outer ring molecular in plane vibration, it is shown as middle up half-

filled circles in Figure 6.7(c). Additionally, on this plot, to highlight the metal / SiO2 boundary

position, the derivative of the signal was taken, which is shown as the middle down half filled circles,

resulting in an asymmetric Gaussian like profile. The derivative permits better localization of the

metal/insulator interface. From this profile, the semi halfwidth at full maximum (SFWHM) of the

Raman signal from molecules on Au was found to be 41.1 μm, and for those on SiO2 34.7 μm.

Deconvoluting these SFWHMs with the laser line profile, one finds the effective SFWHMs of the

Raman signal from pentacene on Au and on SiO2 of about 14.6 μm and 8.2 μm, respectively.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.13

Excitation energy1.91 eV

Phonon profiling

SiO2

Metalcontact

Organiclayer

n-Si

(a)

(b)

0.0 0.2 0.4 0.6 0.8 1.00

2

4

6

8

10

-120

-100

-80

-60

-40

-20

0

dA / dxλ = 647.1 nm

130 mW.cm-2

Pentacene(30 mn) / Au / SiO2

Spot position / mm

Pen

tace

ne(3

0 m

n) /

SiO

2

Area of the bandat 1179 cm-1

Ram

an A

rea

Figure 6.7. Raman spectra of the

organic/inorganic interface. (a) Sketch of

the swept interface in pentacene-based

OFETs, (b) 3D plot of the Raman signal as a

function of the vibronic bands and spatial

position in the organic/inorganic interface,

and (c) Profile of the Raman intensity

(middle up half filled circles) and derivative

(middle down filled circles) as a function of

scanning position.

The determined quantities indicate the influence of the boundary electrode upon molecules deposited

on Au and those on SiO2.

It must be noted that the reported values are limited by the lateral scanning resolution of the setup.

Another effect exhibited by the molecules on the metal contact is the enhancement of the Raman

signal, promoted by the roughness of the metallic surface. As a consequence, the local electric field at

the metal/organic interface is tremendously enhanced about 31 times, leading to the surface-enhanced

Raman spectroscopy (SERS) effect (described in chapter 2).

The difference between the two SFWHMs indicates that the organic layer on the metal is more

influenced by the Au contact boundary than those molecules on the SiO2 substrate. The discrepancy is

also an indication that the way the organic layer grows in the vicinity of the metal / gate insulator

interface is not uniform [Kell2006]. Complementary measurements based on scanning Kelvin probe

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.14

microscopy (SKPM) have shown a polynomial dependence of the electric potential through the

channel [Scho2005]. Therefore, a non-constant electric field parallel to the gated region was

determined, which might additionally cause variation of the charge density distribution along the

channel.

It is important to consider that the non-steady state potential behavior can be affected by technical and

physical factors, such as the tip apex, cone cantilever and the limited spatial resolution governed by

the long-range property of the Coulomb force [Jaco1998]. In this respect, experimental measurements

based on electrostatic force microscopy phase mode (EFMPM) would give more accurate values of the

local electrostatic surface potential [Lei2004].

Performing a correlated fitting of the spectra by using the algorithm introduced in chapter 3, one

obtains the resulting relative FWHM shown in Figure 6.8 with reference to molecules on the SiO2

substrate and away from the metal interface. From Figure 6.8(a), it can be seen that for molecules on

the metal, the relative change of the FWHM remains constant for each band, while those on the

isolator present higher deviations. An interesting and non-simple dependence of the FWHM close to

the metal/dielectric interface is observed.

0.0 0.2 0.4 0.6 0.8 1.0

-10

-5

0

5

10

15

1155 cm-1

1158 cm-1

1162 cm-1

1177 cm-1

1179 cm-1

1180 cm-1

Pen

tace

ne(3

0 m

n) /

SiO

2

OFET: FWHM of the Raman bandsas a function of the scanning position

Pentacene(30 mn) / Au / SiO2

Spot position / mm

Rel

ativ

e FW

HM

(a)

0.6 0.8 1.0-8

-6

-4

-2

0

2

OFET: FWHM of the Raman bandsas a function of the scanning position

1155 cm-1

1158 cm-1

1162 cm-1

1177 cm-1

1179 cm-1

1180 cm-1

Pent

acen

e(30

mn)

/ Si

O2

Pentacene(30 mn) / Au / SiO2

Spot position / mm

Rel

ativ

e FW

HM

(b)

Figure 6.8. Fitted FWHM of the Raman bands at different spot scanning positions (a), and inset of the

FWHM beneath the organic/inorganic interface (b).

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.15

6.7. Bias-stress effects and multi-exponential current relaxation

Within the framework of electrical characterization, in situ measurements of the current voltage

characteristics ( dd V)t(I − ) were carried out. The investigation was performed in two situations, in

darkness and under illumination. A series of successive drain current-voltage ( dd V)t(I − )

characteristics measurements were done for a fixed gate voltage.

The resulting experimental data are plotted in Figure 6.9. The dd V)t(I − characteristics taken in

darkness (cf. Figure 6.9(a)) and under illumination (cf. Figure 6.9(b)) exhibit a decrease of the Id

current over a long period of time. Consequently, this non-steady state behavior of the dd V)t(I −

characteristics might point to a time-dependent charge carrier density. On this occasion, the

drain current as a timedependent quantity prompted an estimate of the dominant time scales to

reach the steady state condition of the drain current. Due to the tendency of numerous

excitation-initiated natural phenomena to relax via exponential decays, the Id is considered to

be a linear combination of single-exponential (Debye) functions.

Given a well-behaved function in a certain interval (continuous and differentiable), in general it can be

linearly decomposed to an exponential basis set. Following the idea formulated above, from the

)t(V)t(I dd − characteristics shown in Figure 6.9, two time constants scaling the dynamic behavior of

the drain current were determined.

(a) (b)

Figure 6.9. Effect of bias stress on the dd V)t(I − characteristics of OFETs (a) in darkness and (b) under

illumination. The red contour lines indicate the time profile behavior of the Id current for a fixed Vd voltage.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.16

A first attempt to describe the dd V)t(I − characteristics is given by [Paez2005a]

)/texp(A)/texp(AII 2211od τ−+τ−+= , (6.5)

with )V,t(I do the steady-state contribution to the current and 1A and 2A the coefficients weighting

the exponential decay with time constants 1τ and 2τ , respectively. These time decays are related to

relaxation phenomena developed in the OFET device, as confirmed by combined Raman spectroscopy

and QTS measurements (chapters 7-8).

The variable t (time) is assumed to have a discrete value for each )t(V)t(I dd − characteristic. The

devices were swept at 10 V/min; accordingly, the drain current intensity given in eq. (6.5) can be

correlated with the sweep drain voltage by td/dVV dd =& , and by replacing

K&

,3,2,1jwithVV

jtd

d =⎟⎟⎠

⎞⎜⎜⎝

⎛= , (6.6)

where j represents the number of repetitions of the )t(V)t(I dd − measurement. It should be noted

that when 0Vd →& then dI is reduced to the expressions already discussed in the previous chapter.

Under stationary conditions, one obtains the drain current in the form

21od AAI)0t(I ++== . (6.7)

This expression is reduced to the non-trap formulation given in chapter 5. The addition of the

coefficients gives the initial drain current of the device.

The decomposition of an dd V)t(I − curve using eq. (6.5) is shown in Figure 6.10. The time

dependence of the exponential functions is closely related to the physical properties of the organic

layer, i.e., charge carrier interaction, charge density distribution, mobility, and structural relaxation.

The average time constants in darkness and under illumination were quite similar min591 =>τ< and

min62 =>τ< differing by less than 5%. The time constants determined from the dd V)t(I −

measurement taken in darkness and under illumination indicate that for this excitation photon energy

the trap density is preserved but not the charge carrier density as expected. The electrical and

vibrational experimental observations suggest that a modified model to describe the output

characteristics in OFETs must be considered.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.17

Figure 6.10. Simulation of one of the dd V)t(I − characteristics shown in Figure 6.9. Experimental

curves are plotted with circles. The quantities Io, A1, A2, τ1, τ2, were determined from the red

lineprofiles shown in Figure 6.9. To reproduce the experimental curve here (uppermost curve), the

time t related to the drain sweep rate of 10 V/min.

The continuous decay of the output characteristics indicates electrical instability of the device. This is

still an issue concerning not only small organic molecules but also polymers and a-Si based devices. In

the field of pentacene-based transistors, the phenomenon has been attributed to shifts in the threshold

voltage (VT) which might cause the dynamic behavior of the drain current (not only decay).

The issue is addressed in more detail in chapters 7 and 8, where traps induced by the applied electric

fields in the organic layer are considered. It should be noted that at last two types of degradation

processes should be considered, one related to the natural aging of the structure, and the other one

when the device is in operation. In most cases, the former phenomenon is irreversible, while the latter

can be reversed after leaving the structure un-biased for a certain time. The mobility of this structure

under darkness conditions was determined to be between 1123 sVcm100.1 −−−× and

1123 sVcm109.3 −−−× for the sublinear and saturation regions, respectively. After long bias stressing,

the mobility is reduced about two orders of magnitude

Time scales lower than >τ< 1 can be measured by means of more refined experimental techniques. In

order to achieve this requirement, charge transient spectroscopy (QTS) (chapter 3) was found to be

suitable for measuring in the range between s1 μ and ms10 . Several time constants were registered

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.18

in this range and a relaxation envelope of about min63=>τ< was estimated after the data

evaluation (chapter 8).

Conclusions

Pentacene-based field effect transistors were characterized by in situ electrical and Raman

spectroscopy measurements. It was demonstrated that for monolayer coverages between 1.1 ML and

∼7 ML with a nominal thickness of about 1.5 nm and 10 nm, respectively, a noticeable increase in the

drain current starting at the pentacene coverage of 1.1 ML indicates that the charge transport in the

channel is governed by a two-dimensional charge carrier gas.

This result, in conjunction with the compressive signature of the Raman bands (table 6.1), proves that

the conductive channel of the OFET behaves in a manner similar to the boosting channel of strained

FETs. In addition, it has been observed that this layer thickness is not enough to produce stable

devices under normal atmospheric conditions. Therefore, either a thicker layer or a capping material is

required.

The profiling of the Raman bands indicates a strong effect of the electrode edges upon the

organic/inorganic interface formation, provoking a non-uniform growth of the active layer. This effect

is quite likely to promote an electric field gradient in the channel accompanied by a spatially

dependent distribution of the charge carrier density in a given molecular plane of the organic active

layer.

Additionally, drain current – drain-voltage ( dd V)t(I − ) characteristics as a function of time showed

that there are relaxation phenomena promoting the reduction of the drain current intensity and the

mobility. The phenomena have an influence on both threshold voltage and on field effect mobility.

The mobility under darkness conditions was estimated to be between 1123 sVcm100.1 −−−× and

1123 sVcm109.3 −−−× for the sublinear and saturation regions, respectively; while for long term bias

stressing, it is reduced by a factor of 210 .

Complementary dynamic dd V)t(I − measurements were carried out in two situations, in darkness and

under illumination (1.916 eV). From both experiments, similar sets of (two) time constants were

determined, indicating that for this excitation photon energy the charge accumulation kinetics may be

preserved, while the injected charge carrier density is enhanced under the illumination.

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.19

The dd V)t(I − characteristics were modeled considering a bias-dependent steady-state current

combined with a linear superposition of two exponential current decays. Regarding the tremendous

amount of the total transient charge (in mC range), it is important to consider not only trap states of

charge carriers at the organic/inorganic interface but also traps induced into the organic material by the

externally applied field.

References

[Abst1985] G. Abstreiter, H. Brugger, and T. Wolf, H. Jorke and H. J. Herzog, Strain-induced two-dimensional electron gas in selectively doped Si/SixGe1-x superlattices”, Phys. Rev. Lett. 54, 2441–2444 (1985).

[Amy2005] F. Amy, C. Chan, A. Kahn, “Polarization at the gold/pentacene interface”, Organic Electronics 6, 85-91 (2005).

[Brab2001] Ch. J. Brabec, N. S. Sariciftci, and J. C. Hummelen, “Plastic solar cells”, Adv. Funct. Mater. 11, 15-26 (2001).[Cell2005] G. Celler, “Strained Silicon could extend limits of CMOS technology”, APS online May (2005).

[Chen2003] D. Chenery and H. Bowring, “Infrared and Raman spectroscopic imaging in biosciences”, Spectroscopy Europe 15, 8-14 (2003).

[Colo2003] P. Colomban, “Raman analyses and smart imaging of nanophases and nanosized materials”, Spectroscopy Europe 15 (6), 8-16 (2003).

[Coll2003] L. Collins, “Some people work better under stress: major players in the semiconductor industry are betting on the same being true for Silicon, Silicon takes T”, IEE Review, 46-49 (2003).

[Dine2004] F. Dinelli, M. Murgia, P. Levy, M. Cavallini, F. Biscarini, and Dago M. de Leeuw, “Spatially correlated charge transport in organic thin film transistors”, Phys. Rev. Lett. 92, 116802 1-4 (2004).

[Drak2003] T. S. Drake, C. N. Chleirigh, M. L. Lee, A. J. Piteraa, E. A. Fitzgerald, and D. A. Antoniadis, D. H. Anium, J. Li, R. Hull, N. Klymko and J. L. Hoyt, “Effect of rapid thermal annealing on strain in ultrathin strained silicon on insulator layers”, Appl. Phys. Lett., 83, 875-877 (2003).

[Gámi2002] F. Gámiz, J. B. Roldán, and A. Godoy, “Strained-Si/SiGe on insulator inversion layers: The role of strained-Si layer thickness on electron mobility”, Appl. Phys. Lett., 80, 4160-4162 (2002).

[Gaus1998] Gaussian 98 (Revision A.1), M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G., Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A., Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M. W. Gill, B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. Head-Gordon, E.S. Replogle and J.A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.

[Hash1999a] S. Hashimoto, T. Ogawa, S. Isoda, and T. Kobayashi, “Electron diffraction analysis of polymorph structures in ultra thin film of vanadyl phthalocyanine on KBr (001)”, Journal of Electron Microscopy 48, 731-738 (1999).

[Hash1999b] S. Hashimoto, S. Isoda, H. Kurata, G. Lieser and T. kobayashi, “Molecular orientation of perfluoro-nanadyl-phthalocyanine examined by electron energy loss spectroscopy”, Journal of Porphyrins and Phthalocyanines 3, 585–591 (1999).

[Jaco1998] H. O. Jacobs, P. Leuchtmann, O. J. Homan, and A. Stemmer, “Resolution and contrast in Kelvin probe force microscopy”, J. Appl. Phys. 84, 1168 (1998).

[Kell2006] T. Kelley, in “Organic Electronics”, chap. 2, edited by H. Klauk, Wiley-VCH, Weinheim 2006. [Kim2005] H. -K. Kim, D. -G. Kim, K. -S. Lee, M. -S. Huh, S. H. Jeong, K. I. Kim, and T. –Y. Seong, “Plasma damage-

free sputtering of indium tin oxide cathode layers for top-emitting organic light-emitting diodes”, Appl. Phys. Lett. 86, 183503 1-3 (2005).

[Lamm2003] D. Lammers, “Intel, TI 'straining' to improve 90-nm silicon”, EE Times (2003). [Lei2004] C. H. Lei, A. Das, M. Elliot, and J. E. Macdonald, “Quantitative electrostatic force microscopy-phase

measurements”, Nanotechnology 15, 627-634 (2004). [Liab2005] C. N. Liab, C. Y. Kwonga, A. B. Djurisicc, P. T. Laia, P. C. Chuia, W. K. Chand, S. Y. Li, “Improved

performance of OLEDs with ITO surface treatments”, Thin Solid Films 477, 57– 62 (2005). [Matt2002] C. C. Matteheus, “Polymorphism and electronic properties of Pentacene”, PhD work (Print Partners Ipskamp

B.V., Enschede 2002). [Maye2004] A. C. Mayer, R. Ruiz, R. L. Headrick, A. Kazimirov, G. G. Malliaras, “Early stages of pentacene film growth

on silicon oxide”, Organic Electronics 5, 257–263 (2004). [Mott1938]

N. F. Mott, “Note on the contact between a metal and an insulator or semiconductor”, Proc. Cambridge Philos. Soc. 34, 568-572 (1938).

[Muck2004]

T. Muck, V. Wagner, U. Bass, M. Leufgen, J. Geurts, L.W. Molenkamp, “In situ electrical characterization of DH4T field-effect transistors”, Synth. Met. 146, 317–320 (2004).

Chapter 6 B. A. Paez-Sierra, Combined Raman spectroscopy and electrical characterization…

6.20

[Paez2005a] B. A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F 1-9 (2005).

[Paez2005b] B. A. Paez, G. Salvan, R. Scholz and D. R.T. Zahn, “Interface formation of Mg with DiMe-PTCDI studied by Raman spectroscopy”, phys. stat. sol. (c) 2, 4048-4052 (2005).

[Rees2004] C. Reese, M. Roberts, M. -M. Ling, and Z. Bao, “Organic thin film transistors”, Materials Today, 7, 20-27 (2004).

[Ridl1999] B. A. Ridley, B. Nvi and J. M. Jacobson, “All organic field effect transistors fabricated by printing”, Science 286, 746-748 (1999).

[Ross2002] D. Ross and R. Aroca, “Efective medium theories in surface enhanced infrared spectroscopy: the Pentacene example”, J. Chem. Phys. 117, 8095-8103 (2002).

[Ruiz2004] R. Ruiz, A. C. Mayer, and G. G. Malliaras, B. Nickel, and G. Scoles, A. Kazimirov, H. Kim, R. L. Headrick, Z. Islam, “Structure of pentacene thin films”, Appl. Phys. Lett., 85, 4926-4928 (2004).

[Ruiz2005] R. Ruiz, A. Mayer, A. Papadimitratos, and G. Malliaras, “Petacene growth for transistor applications”, SPIE 2005 (Optics & Photonics San Diego) (Pictures taken during the talk).

[Salv2005] G. Salvan, B. A. Paez¸ S. Silaghi, and D. R. T. Zahn, “Deposition of silver, indium, and magnesium onto organic semiconductor layers: Reactivity, indiffusion, and metal morphology”, Microelectronic Engineering, 82, 228-235 (2005).

[Sche2005] S. Scheinert, G. Paasch, I. Hörselmann, and A. Herasimovich, “Influence of source/drain contacts on sub-micrometer organic field-effect transistors”, phys. stat. sol. (a) 202, R82-R84 (2005).

[Scho2005] R. Scholz, A. -D. Müller, F. Müller, I. Thurzo, B. A. Paez, L. Mancera, D. R. T. Zahn, “Comparison between the charge carrier mobilities in Pentacene OFET structures as obtained from electrical characterization and potentiometry”, Proc. of SPIE 5940, 59400I 1-7 (2005).

[Schr2004] F. Schreiber, “Organic molecular beam deposition: Growth studies beyond the first monolayer”, phys. stat. sol. (a) 201, 1037–1054 (2004).

[Shan2006] H. Shang, M. M. Frank, E. P. Gusev, J. O. Chu, S. W. Bedell, K. W. Guarini, M. Ieong, “Germanium channel MOSFETs: Opportunities and challenges”, IBM J. RES. & DEV. 50, 377-386 (2006).

[Schu2004] M. Schubert, C. Bundesmann, G. Jacopic, H. Maresch, and H. Arwin, “Infrared dielectric function and vibrational modes of Pentacene thin films”, Appl. Phys. Lett. 84, 200-2002 (2004).

[Stre2005] R. A. Street, J. E. Northrup, and A. Salleo, “Transport in polycrystalline polymer thin-film transistors”, Phys. Rev. B 71, 165202 1-13 (2005).

[Sun2005] Y. Sun, S. Kim, I. Adesida, and J. A. Rogersa, “Bendable GaAs metal semiconductor field-effect transistors formed with printed GaAs wire arrays on plastic substrates”, Appl. Phys. Lett. 87, 083501 1-3 (2005).

[Tiba2003] M. V. Tiba , O. Kurnosikov, C. F. J. Flipse, H. J. M. Swagten, J. T. Kohlhepp, R. A. J. Janssen, P. M. Koenraad, W. J. M. de Jonge and B. Koopmans, “Growth of organic molecules on ferromagnetic metallic substrates -a combined STM/AFM study”, CP696, Scanning tunneling microscopy and related techniques: 12th International Conf., 521-528 (2003).

[Vere2004] J. Veres, S. ogier, G. Lloyd, and D. de Leeuw, “Gate insulators in organic field-effect Transistors“, Chem. Mater. 16, 4543-4555 (2004).

[Wang2005] S. D. Wang, X. Dong, C. S. Lee, and S. T. Lee, “Molecular orientation and film morphology of pentacene on Native Silicon Oxide Surface”, J. Phys. Chem. B 109, 9892-9896 (2005).

[Wu2004] Ch. -H Wu, Y. -K. Su, Sh. -J. Chang, Y. -Sh. Huang, and H. -P. Hsu, “Device characteristics of GaAs-based heterojunction bipolar transistors using an InGaAs/GaAsP strain-compensated layer as a base material”, Semicond. Sci. Technol. 19, 828–832 (2004).

[Yoo2001] S. Yoo, B. Domercq, and B. Kippelen, “Efficient thin-film organic solar cells based on Pentacene/C60 heterojunctions”, Appl. Phys. Lett., 85, 5427-5429 (2004).

[Zahn2001] D. R. T. Zahn. “Raman Studies of Molecular Thin Films”, phys. stat. sol. (a) 184, 41-50 (2001). [Zahn2005] D. R. T. Zahn, G. Salvan, G. Gavrila, and B. A. Paez, “Chemistry and morphological properties of metal

interfaces to organic semiconductors”, Adv. sol. sta. phys. 45, 313-324 (2005).

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.1

Chapter 7

Influence of electric fields and illumination in OFETs

In this chapter, the modification of the vibronic-spectrum in molecular systems under external electric

fields is discussed. The resulting combined Raman measurements with applied electric fields gave

valuable information about the molecular structure alteration, and the correlation between charge

carriers scattering and the relaxation phenomena developed at intra- and inter-molecular levels.

The investigated structures revealed a structural relaxation with a time constant of approximately 94 s.

This was estimated from the Raman spectroscopy measurements performed after switching off the

external electric field. Therefore, an extended recovery time for the molecular structure of about 9 h

was estimated. The induced molecular modification partially explains the drain current decrease when

the OFETs are continuously operated.

7.1. Introduction The interaction of matter with external electric fields leads to alterations at electron energy levels,

changes in the molecular dipole moment, structural configuration [Bish1990, Feyn1963] and

molecular redistribution [Ren2006]. The electric field induces charge displacement of the electrons in

the outer shell of the atom, then seen as induced dipole promoting shifts of the Raman bands and/or

modifications of the mode intensity [Auss1986, Jeon2003, Paez2003, Ralp1990].

Alteration of the normal molecular absorption and distortion in the scattered light under optical

excitation is likewise a phenomenon able to demonstrate the interaction between matter and the

applied fields. The electronic or geometric changes produced by the electric field are understood as a

breakdown of the molecular symmetry. Despite these issues, some theoretical calculations have been

carried out on organic structures and assuming periodic boundary conditions [Tóbi2004].

In a fashion similar to that of inorganic semiconductors, it is expected to have intraband and interband

transitions reflected in the Fröhlich and Franz-Keldysh type interaction matrix elements, respectively.

The first type of interaction involves intermediate excitation states promoted by the non-local

molecular polarized discrete and continuous electron-hole pair states [Shan1972]; the second one is a

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.2 consequence of the Fröhlich interaction. On the other hand, the probability of the resonant Raman

transition is affected by the field. Therefore, the established rules by the non-perturbed Franck-Condon

effect are modified as a consequence. Further detailed analysis shows the influence of the electric

field on the Franz-Condon effect.

Investigations of combined Raman and applied external electric fields cover a broad area. By using

Pico-Raman spectroscopy, the drift velocity in heavily degenerated semiconductors [Gran1994] has

been proved. An interesting and apparent antagonistic feature of the Raman effect is that in spite of the

fact that the scattering cross section is inversely proportional to the square of the effective mass, it

selectively proves electron distribution functions even if holes are present. This experimental

procedure works nicely for high charge carrier concentrations.

As shown in the previous chapter, the conductive channel thickness extends up to a few nanometers

(approx. three monolayers), indicating that the enhanced charge density of the organic layer is

confined to a two-dimensional gas [Ando1982]. Recent novel investigations using surface-sensitive

IR, visible sum frequency generation (SFG), nonlinear optical spectroscopy on interfaces of OFETs

during operation [Ye2006], and absorption infrared active vibrational (IRAV) modes spectroscopy

restate the capability of vibrational techniques to scope the depleted organic material by gate fields in

OFETs [Li2006]. In order to undertake the combined vibronic-electrical research of the active

material, one aspect to be considered is the geometrical configuration of the leads in the OFET. The

bottom contacts arrangement gives greater advantages over other geometrical contact patterns.

7.2. Raman bands and external electric fields

From the phenomenological point of view, first-order IR absorption and the first-order Raman

scattering tensor arise from conditions in the electric dipole moment vector (per unit cell) μ . The

Raman line intensities are determined by differentiating the classical polarizability tensor (α ) with

respect to the normal molecular coordinate ( iQ ).

It should be noted that α contains elements equal to the derivatives of the μ components with respect

to the applied electric field (Ε ) components. Therefore, rewriting the expressions found in chapter 2,

one obtains Raman intensities related to

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.3

oij

i2

oji

i2

k

ij

QQQ ⎟⎟⎠

⎞⎜⎜⎝

⎛

∂∂μ∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛

∂∂μ∂

=∂

α∂

EE, (7.1)

therefore, depending on the molecular structure orientation and for different applied external fields,

the polarizability scales as the inverse of the square of the molecular band gap [Kozi2006].

7.2.1. Band gap modification by external electric fields

As an example, consider the benzene molecule fixed in a coordinate system and exposed to external

electric fields, as illustrated in Figure 7.1. For the applied electric fields, the HOMO and LUMO states

were evaluated at the B3LYP / 3-21G level in Gaussian 98 [Gaus1998] and the resulting molecular

orbitals were plotted in ChemOffice. The picture indicates the changes due to the excitation field.In

the same figure, the band gap alterations in comparison with the relaxed structure are summarized. It is

observed that the highest modification of the band gap occurs when the electric field is applied along

the z axis; the peculiarity of this direction in comparison with the other two is that it involves the

highest delocalization of the electronic cloud.

Figure 7.1. Benzene molecule under different external electric fields. (a) Molecular structure, (b) chart

summarizing the band gap modification by the applied fields, and (c) HOMO and LUMO

configuration in response to the applied electric fields. (The unit H in the table means Hartree = 27.2

eV).

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.4 In a similar manner, the interaction of the external electric field with pentacene was simulated. It has

been found that the main electronic alteration is along the y axis according to the orientation depicted

in Figure 7.2. It was also observed that the “y” axis corresponds to the direction of the active layer

thickness in the OFET.

In chapter 5, the molecular arrangement of the pentacene layers on the gate-dielectric substrate was

discussed. Moreover, it was estimated that the molecules are in an upright position. This means that

for molecules oriented in this direction, the influence of the gate field is decisive for the charge

carriers transport.

For isolated molecules or interlayer structures without electrodes, the polarized charge density is given

by

α⋅−∇=ρpol , (7.2a)

or, in the event that the spin electron contribution is considered, eq.(7.2a) reads [Park1986],

α⋅∇−ΨΨ−=ρ +epol , (7.2b)

where α is given through eq.(7.1) andΨ is the spinor describing the quantum state of the electron. It

must be noted that this expression keeps the proportionality between the Raman signal and the charge

induced by the external field. Then theoretically, as expected, any external field will produce

modification of the Raman bands.

A particular feature of the pentacene molecule and similar acenes is the higher number of C-H bonds

than C-C bonds. This makes the induced polarization in the direction of the hydrogen atoms more

favorable and delocalization of the molecular orbitals broader.

Figure 7.2. DFT simulation of the influence of an applied electric field on the molecular band gap and

HOMO-LUMO states in pentacene. (The unit H in the table means Hartree = 27.2 eV).

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.5 7.2.2. Raman bands alteration by external electric fields

7.2.2.1. Pentacene

Similar calculations with electric fields were made to investigate the Raman shifts. In Figure 7.3, the

orientation of the pentacene molecule and the relative Raman shift of the electrically excited molecule

are shown. In order to quantify the effect of the applied electric field on the Raman bands of the

molecule, the Raman shifts of the excited structure were subtracted from those of the free molecule

(cf. Figure 7.3(b)). The experiment was primarily focused on the shaded region indicated in Figure 7.3,

where the in-plane C-H vibrations are involved. The discrepancy between the experimental Raman

shifts upon applied electric fields and those calculated by DFT, proved the necessity to include the

electron-electron correlation and delocalized basis sets in the calculations.

The non-shaded spectral regions were experimentally monitored under the influence of the electric

field as well, and minor or absent structural modifications were detected. Therefore, attention was

mainly targeted to the shaded zone depicted in Figure 7.3(b).

In chapter 6, the characteristics of the organic boosting layer, which holds most of the gate induced

charge density for the charge transport in the OFET, was discussed. In analogy to investigations of the

depletion region in highly doped ( 319 cm105.2 −× ) inorganic semiconductors with Raman

spectroscopy [Fuka1988], the method appears very sensitive to comparable doping modifications

where the electronic susceptibility is assumed to be modulated by the atomic relative displacement and

the macroscopic electric field associated with the LO phonons propagating along the inter-atomic

sites.

Organic materials are characterized as having a much lower charge carrier density. This makes the

analysis of the accumulation region of the OFET more challenging, since lower charge density

distributions for both inorganic and organic semiconductors are more complex.

Other factors involved are the conductive channel thickness and the photon excitation energy. The

OBL maintains the field-modulated charge density with non-degenerated distribution of charge

carriers. Therefore, despite the long penetration depth λ relative to the accumulation region width, the

contribution to the Raman spectrum of surface phonons is dominant compared to the bulk

contribution.

It was shown in chapter 6 that the accumulated charge density scales as the square inverse of the OBL

thickness, being higher for molecular planes closer to the gate dielectric. This has advantages for the

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.6 experiment combining Raman measurements and electric fields. Applying drain-source fields, the

contribution to the Raman cross section is negligible or at least below the setup resolution.

(a)

900 1050 1200 1350 1500-40

-30

-20

-10

0

10

20

30

40

Freq

uenc

y sh

ift /

cm-1

Frequency / cm-1

Ex - Eo

Ey - Eo

Ez - Eo

Influence of an applied electricfield on the vibronic bands

experiment

B3LYP / 321 - G

(b)

Figure 7.3. Calculated vibrational states of a pentacene molecule under the influence of external electric

fields. (a) Orientation of the pentacene molecule, and (b) variation of the frequency assignment as a

function of the vibrations without an applied electric field.

The experimental results of Raman spectroscopy with electric fields are shown in Figure 7.4. For the

investigations, the drain and source contacts were grounded, and step-wise voltages were applied to

the gate contact. The gate voltages were varied between 0 V and –24 V.

The most significant modification of the Raman bands was found to occur at the molecular in-plane

C-C ring and C-H vibrations with Ag symmetry. For the experiment, the Raman intensity followed a

square law dependence with the applied gate field; this is summarized in the correlated fitting

illustrated in Figure 7.5. Additionally, no changes in bands positions were observed.

The Raman signal is basically dependent on the polarizability modulation (chapter 2) of the

investigated media Consequently, one expected signature is the change of the Raman intensity (cf.

Figure 7.4 and 7.5 (a)).

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.7

Figure 7.4. Experimental measurements of the Raman bands in pentacene (30 nm) for different applied

fields. The involved modes of this spectral region belong to in-plane vibrations of C-H bonds of the

outer ring and antiphase vibrations of the C-H bonds parallel to the main axis of the molecule.

The gradual enhancement of the Raman bands in response to the increase of the gate voltage provides

an indication as to the formation of the conductive channel in the pentacene-based FET. The increase

of the Raman intensity as a function of the electric field (cf. Figure 7.5) is developed up to a critical

value; afterwards a saturation region appears, followed by a decrease in the intensity.

The first part to be considered is the charge state formation at the OBL. The saturation corresponds to

the influence of molecular polarizability inducing dipole-anion sites trapping the induced charge by

the gate field. A further increase of the electric field enhances the dipole-anion sites population, while

a decrease of the Raman cross section is observed.

Combining the equations (7.1, 7.2), one can describe the proportionality between the induced charge

density and the Raman intensity as

enIRaman l∝ , (7.3)

with ln the charge density per layer and e the electron charge. It should be noted that by

accumulating charge, the Raman intensity becomes proportional to the square of gV as shown

eq.(6.4) and experimentally indicated in Figure 7.5. Nevertheless, additional considerations should be

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.8 taken into account, since the higher the electric field the higher the probability of generating artificial

flaws (chap. 8) in the organic layer.

0 -5 -10 -15 -20 -250

15

30

45

60

75

90

Voltage / V

λ = 676.4 nm / 130 W cm-2

Raman intensiy in dependence

of the applied electric field

Pentacene(30 nm) based FET

Ram

an In

tens

ity (a

rea)

/ c

ts.m

W-1.s

-1.c

m-1

1155.5 cm-1

1158.6 cm-1

1164.0 cm-1

1176.5 cm-1

1179.0 cm-1

1180.2 cm-1

Figure 7.5. Correlated fitting of the Raman intensity as a function of the applied gate voltage (the

digits behind the floating point were delivered after the correlated fitting procedure).

The enhancement of the Raman signal is an indication that the Raman cross section ( RSσ ) is

increased. Typical values of RSσ are between 10-27 and 10-14 cm2, where the lower RSσ is the normal

value of isolated molecules interacting with the photon beam, while the higher one scales with the

typical values quoted for surface enhancement Raman spectroscopy and mediated by electromagnetic

fields (Chapters 2 and 4). On the other hand, by combining the Raman cross section and the charge

carrier capture cross section determined from complementary QTS measurements (chap. 8), it is

possible to evaluate the charge density appearing in eq. (7.3), if a coupled electron-phonon system

with impurities is assumed [Itai1992]; hence a value of about 1012 cm-2 is obtained.

The experiment was then carried out by switching off the gate field and in situ monitoring of

the vibrational bands relaxation, as illustrated in Figure 7.6. The time decay constant was

found to be about 94 min.

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.9

0 30 60 90 120 150 18020

22

24

26

28

30

32

Relaxation of the bandat 1179 cm-1 after switching offthe electric field (Vg = - 24 V )

Inte

nsity

/ ct

s. m

W -1

s-1

time / min

Decay time: 94 min

λ = 676.4 nm / 130 W cm-2

Pentacene(30 nm) based FET

(b)

Figure 7.6. In situ measurements of the Raman bands relaxation after switching off the electric field, and

(b) time decay of the Ag band at 1179 cm-1 after switching off the electric field.

7.2.3. The C60 fullerene

In this work, the charge induced in C60 was produced by changing the gate voltage in the OFET

structure. In Figure 7.7, the effect of the electric field on the Raman spectra is presented. Theoretical

calculations performed with Gaussian 98 [Gaus1998] at B3LYP [Beck1993, Koch2002, Salo2002]

level and the 6-21G basis set reveal that, of the 46 symmetry operations, many of the vibrational

modes are two-fold and even five-fold degenerated, giving 174 vibrational modes. The most

representative Raman modes of the isolated C60 molecule theoretically evaluated are: squeezing (271

cm−1), breathing (491 cm−1), and pentagonal pinch (1495 cm−1) modes.

In the experiment, the in situ macro-Raman signal was measured at very low power density (25.4

W/cm-2 and laser power of 2.0 mW), approximately 50 times lower than that used to investigate the

perylene derivatives (chapter 4) and about five times lower in comparison with the micro-Raman

experiments performed on bias-stressed pentacene FETs. The obtained Raman intensity is about 100

times lower than that registered from pentacene-based OFETs. The light source and intensity were set

to such low values in order to avoid additional features in the spectrum induced by the photo-

polymerization phenomenon. The investigated spectral region was selected where the pentagonal

pinch mode is located, being the only region where signal detection was possible for such thin film.

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.10 The electrical charging basically leads to two results. The first one is symmetry breakdown, which is

reflected in the splitting of the degenerate modes, and the second one is the significant change of the

pentagonal pinch mode, i.e., the change in the Raman intensity and the center of mass of this peak, is

shifted to lower frequencies, in agreement with the theoretical results. Experimentally, the advantage

of the OFET structure is utilized to induce charge in the C60 molecules.

The differential Raman spectra, in addition to the spectra with and without an electric field, are shown

in Figure 5. There are clear effects of the voltage on the Raman signal which are fully reversible. One

is the change of the FWHM and the other is the decrease of the Raman signal, producing quenching of

the resonance inter-band transitions. Consequently, on the basis of the Franck-Condon effect, it is

more probable that higher stretching in and out of the pentagonal pinch mode will be promoted. Then

attenuation of the Raman signal is expected, as experimentally observed and shown in Figure 7.7.

Figure 7.7. Raman spectra (upper plot) and differential Raman spectra (lower plot) of 3 nm C60 layer,

the latter obtained by subtracting the spectra without an electric field from the ones under the influence

of the gate voltage.

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.11 7.3. Illumination and charge transport in OFETs

The samples were also illuminated by different wavelengths, for which the maximum drain current in

two situations was monitored, one taking into account the saturation drain-gate voltages of the

saturation regime (illumination), and the second one with zero gate voltage (cf. Figure 7.8). Both

situations illustrated in Figure 7.8 show a similar line-profile with a difference in the enhancement

drain-current intensity. The photon-molecule interaction promotes an increase of free charge carriers,

due to ionization of neutral traps and photodetachment of free excess carriers at anion-dipole sites. The

first mechanism is strongly dependent on the excitation photon energy, while the second one is

understood as the emission of trapped charge carriers at molecular dipole domains. Therefore, these

two effects have an impact on the measured drain current.

The measured Id-max as a function of the excitation photon energy looks like a mirror image of the

absorption spectrum of pentacene (30 nm) deposited on quartz. This result corresponds to a large

extent with the vibronic characterization at energies close to the first absorption peak.

Figure 7.8. Photocurrent of pentacene (30

nm) based FET. The spectra were taken for

a drain voltage of –40 V and gate voltage of

0 V (blue) and –10 V (green). The red

spectrum corresponds to the absorption of

30 nm pentacene deposited on quartz.

7.3.1. Persistent effects and multi-exponential kinetics

A complementary investigation, consisting of measurement of the drain current after suppressing the

photon source, is displayed in Figure 7.9. The drain current relaxes monotonously, obeying

exponential kinetics of at least two relaxation times. Similar dynamics have been addressed elsewhere

[Dutt2003, Palm1984, Quei1985, Schi1995]. The quoted relaxation time constants indicate the

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.12 evolution of two charge carrier scattering mechanisms, one scenario with s771 =τ and the other as a

collective phenomenon with a time constant of s3372 =τ . Both time constants were stimated with an

accuracy better than 3 %.

0 100 200 300 400 500 600

-6.5

-7.0

-7.5

-8.0

-8.5

-9.0

-9.5

-10.0

Vg = -12 V Vds = -40 V

I ds /

nA

t / s

Off: λ = 676.4 nm 0.6 W.cm-2

τ1 = 77 s

τ2 = 337 s

Pentacene (30 nm) based FET

Figure 7.9. Persistent behavior of the

drain current in pentacene (30 nm)-

based FET after switching off the

photon source. The decay follows

multi-exponential kinetics.

A similar photocurrent experiment was performed by switching on and off the coherent photon source

(Kr+ laser). The dynamics during and after illumination are well described by multi-exponential

kinetics. It is important to note that the first effect of light exposure is the recovery of the molecules.

Subsequent light doses induce an envelope function for the drain current with sample illumination and

a second envelope develops following the minima of the drain current under dark conditions. The time

constant of the upper envelope is about 209 days (ON state), while the lower envelope has a relaxation

constant of approximately 96 days (OFF state). In both situations (light ON and OFF), the device

operates under quasi-stationary conditions.

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.13

Figure 7.10. Photocurrent and

dark current in 30 nm pentacene

FET. The sample was illuminated

by coherent light of 676.4 nm.

0 1000 2000 3000 40000

-20

-40

-60

-80

-100

-120

ON OFF OFFON

Non first order exponential process

I d / n

A

t / s

Induced Photocurrent inPentacene OFETsd x W x L = 16 μm x 100 μm

λ = 676.4 nmVg = -12 VVds = -40 V

Conclusions

The influence of electric fields on molecular properties has been investigated by density functional

theory (DFT), and experimentally with a combination of Raman spectroscopy and electrical

characterization. The theoretical calculations showed an alteration of the molecular band-gap, vibronic

and electronic states and removal of degeneracy when the investigated structure had a multiplicity of

states at a given energy level.

The experimental results of Raman spectroscopy with electric fields in OFETs demonstrated a

proportionality between the Raman signal and the applied field, which for pentacene (30 nm)-based

FETs increases until applied voltages of about –20 V, followed by an attenuation for higher fields. The

behavior suggests considering competing phenomena between induced charge density and the creation

of dipole-anion sites induced by the external field. The enhanced Raman signal in pentacene-based

FETs indicates that the cross section of inelastic scattered photons is enhanced as well.

The subsequent vibronic measurements after switching-off the electric field indicate a dynamic

relaxation of the Raman cross section with a time constant of about 94 min. The experiment predicted

a long recovery time for the device (confirmed indirectly in chap. 8 by QTS measurements).

The pentacene FETs were illuminated with different photon energies, resulting in selection rules for

the enhanced drain current. The photocurrent followed a mirror-like profile of the absorption

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.14 spectrum. Further experiments with chopped photon light gave additional hints about the charge

relaxation transport phenomena with temporal envelope profiles for the drain current during sample

illumination and darkness conditions. After illumination, the drain current evolves in a persistent

multi-exponential decay, described by two time constants of about 77 s and 337 s.

The second model system was a C60 (3 ML)-based FET which had to be carefully handled in order to

avoid additional spectral features promoted by the photon source. Therefore, an optimum set of

experimental parameters, such as photon energy, power density and layer thickness, were tuned to skip

artifacts during the measurments. Under these conditions, the Raman intensity was approximately 100

times lower than that obtained from pentacene (30 nm)-based FETs. The C60 molecule is characterized

as being totally symmetric and the first consequence of the applied electric field is the removal of this

degeneracy, as demonstrated experimentally and theoretically.

The experimental Raman measurements of the C60 films were delimited to the pentagonal pinch mode

(PPM) region, where the signal was significant for extracting the phenomena induced by the applied

field. In this molecule, the Raman band of the PPM had a decrement of its intensity together with a

blue shift. In the framework of the Franck-Condon effect, this experimental behavior is likely to occur,

since the PPM can be decomposed as a combination of breathing modes; consequently, the

internuclear distances are modified, producing a reduction of the Raman transition probability.

Another factor is the splitting of some bands close to the PPM; due to the setup resolution, however, it

was not possible to observe this.

References

[Ando1982] T. Ando, A. B. Fowler and F. Stern, “Electronic properties of two-dimensional systems”, Rev. Mod. Phys. 54, 437–672 (1982).

[Auss1986] F. R. Aussenegg, M. E. Lippitsch, “On the electric field dependence of Raman scattering”, Journal of Raman Spectroscopy 17, 45 – 49 (1986).

[Bish1990] D. Bishop, “Molecular vibrational and rotational motion in static and dynamic electric fields”, Rev. Mod. Phys. 62, 343-374 (1990).

[Beck1993] A. D. Becke, “Density-functional thermochemistry. III. The role of exact exchange“, J. Chem. Phys. 98, 5648–5652 (1993).

[Dutt2003] S. Dutta and K. S. Narayan, “Nonexponential relaxation of photoinduced conductance in organic field effect transistors”, Phys. Rev. B 68, 125208 1-4 (2003).

[Feyn1963] R. P. Feynman, R. B. Leighton, and M. L. Sands, “The Feynman lectures on physics”, V. 3 Ch. 9, (Reading, Mass.: Addison-Wesley Pub. Co., 1963).

[Fuka1988] R. Fukasawa, S. Katayama, A. Hasegawa, and K. Ohta, “Analysis of Raman spectra from heavily doped p-GaAs”, J. Phys. Soc. Jap. 57, 3632-3640 (1988).

[Gaus1998] Gaussian 98 (Revision A.1), M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G., Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A., Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M. W. Gill, B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. Head-Gordon, E.S. Replogle and J.A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.

[Gran1994] E. D. Grann, S. J. Sheih, C. Chia, K. T. Tsen, and O. F. Sankey, “Picosecond Raman studies of electric-field-induced nonequilibrium “, Appl. Phys. Lett. 64, 1230-1232 (1994).

[Itai1992] K. Itai, “Theory of Raman scattering in coupled electron-phonon systems”, Phys. Rev. B 45, 707–717 (1992).

Chapter 7 B. A. Paez-Sierra, Influence of electric fields… 7.15

[Jeon2003] T. S. Jeong, C. J. Youn, M. S. Han, J. W. Yang, and K. Y. Lim, “Electric-field-induced quenching effect of Raman scattering in Mg-doped p-GaN”, App. Phys. Lett. 83, 3483-3485 (2003).

[Koch2002] W. Koch, M. C. Holthausen. “A chemist´s guide to density functional theory”. Pp. 78-85, (Wiley-VCH, Germany, 2002).

[Kozi2006] B. Kozinsky, and N. Marzari, “Static dielectric properties of carbon nanotubes from first principles”, Phys. Rev. Lett. 96, 166801 1-4 (2006).

[Li2006] Z. Q. Li, G. M. Wang, N. Sai, D. Moses, M. C. Martin, M. Di Ventra, A. J. Heeger, and D. N. Basov, “Infrared imaging of the nanometer-thick accumulation layer in organic field-effect transistors”, Nano Lett., 6, 224-228 (2006).

[Paez2003] B. A. Paez, M. Bartzsch, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Combined electrical and Raman characterization of C60 based organic field effect transistors”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 63 (2003).

[Palm1984] R. G. Palmer, D. L. Stein, E. Abrahams, and P. W. Anderson, “Models of hierarchically constrained dynamics for glassy relaxation”, Phys. Rev. Lett. 53, 958–961 (1984).

[Park1986] G. W. Parker, “Derivation of the lectric dipole-dipole interaction as an electric hyperfine interaqction”, Am. J. Phys. 54, 715-717 (1986).

[Quei1985] H. J. Queisser, “Nonexponential relaxation of conductance near semiconductor interfaces”, Phys. Rev. Lett. 54, 234–236 (1985).

[Ralp1990] S. E. Ralph and G. J. Wolga, “Field-induced nonequilibrium carrier distributions in GaAs proved by electronic Raman scattering”, Phys. Rev. B 42, 1353–1363 (1990).

[Ren2006] H. Ren, Y. H. Lin, and S.T. Wua, “Adaptive lens using liquid crystal concentration redistribution”, App. Phys. Lett. 88, 191116 1-3 (2006).

[Salo2002] O. Salomon, M. Reiher, and B. A. Hess, “Assertion and validation of the performance of the B3LYP* functional for the first transition metal row and the G2 test set”, J. Chem. Phys. 117, 4729-4737 (2002).

[Schi1995] H. Schiessel, R. Metzlrt, A. Blumen, and T. F. Nonen, “Generalized viscoelastic models: their fractional equations with solutions”, J. Phys. Math. Gen. 28, 6567-6584 (1995).

[Shan1972] M. L. Shand, W. Richter, E. Burstein and J. Gay, “Electric field induced Raman scattering in CdS”, J. Nonmetals 1, 53-62 (1972).

[Tóbi2004] J. Tóbik and A. Dal Corso, “Electric fields with ultrasoft pseudo-potentials: Applications to benzene and anthracene”, J. Chem. Phys. 120, 9934-9941 (2004).

[Ye2006] H. Ye, A. Abu-Akeel, J. Huang, H. E. Katz, D. H. Gracias, “Probing organic field effect transistors in situ during operation using SFG”, J. Am. Chem. Soc. 128, 6528-6529 (2006).

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.1

Chapter 8

Trap distribution in OFETs and anomalous QTS

In this chapter, anomalies of the charge-transient spectroscopy (QTS) signal in pentacene-based

OFETs are presented. The anomalous phenomenon is evidenced by deviation of the QTS spectra from

those which are in close accordance with the polarities of the applied bias and the relaxed signal. The

anomalous QTS spectrum results with combination of peaks with positive and negative components.

This phenomenon demosntrates the presence of trap centers for minority and majority charge carriers.

The unusual behavior is attributed to intrinsic and extrinsic trap sources of the OFET device. The first

trap source is developed at the boundaries of the organic material, while the second one comes from

dipole sites in the organic layer prompted by the external electric field. In addition, the negative

transconductance, or so-called "drain current collapse", provides strong indications of trapping from

both types of charge carriers at the pentacene film. A description of the anomalous QTS measurements

is complemented by a theoretical model. For the simulation, two relaxation processes are assumed,

one dielectric and the other governed by decay of the induced dipoles in the local field.

It is shown that under illumination the transconductance can be switched from negative to positive

values. Thus, the trapped charge can be ionized by well-defined photon energies. Therefore, a

photodetachment process at the (anion, cation)-dipole sites might be developed.

Furthermore, subsequent light doses increase the mobile charge density, while the removal of the

photon source has a persistent effect, a phenomenon registered by the slow decay of the drain current.

Experiments revealed that the drain current measurements, either with illuminated channels or in

darkness, follow time-dependent multiexponential kinetics [Paez2005].

8.1. Introduction

In organic electronics, the organic/inorganic interface formation is the main object responsible for

several interactions affecting the device performance. The most widely accepted conclusions of these

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.2 interface studies is that metal/organic contacts generally do not follow the Schottky-Mott model and

exhibit large dipole barriers [Amy2005]. Here it is assumed that the local dynamics of charge

distribution contributes to the electric dipole moment of the total active layer in spite of not being in

direct contact with the inorganic material.

In order to identify the charge carrier interaction with traps, ex situ charge transient spectroscopy

(QTS) measurements were performed on the OFETs. These results provide valuable information about

the influence of dipolar traps on the majority and minority charge carrier energetic distributions

[Thur2004].

The anomalous behavior of the QTS signal observed in pentacene-based OFETs is similar to results

that have been reported for GaAs MESFETs [Zhao1990], current DLTS in AlGaAs/GaAs FETs

[Cava2003], and capacitance-DLTS measurements in pentacene MIS structures [Yang2002].

Experimental measurements of drain current relaxation for single active layers of pentacene, CuPc,

and PT FETs as a gas sensor like device [Tane2005] show minority charge carrier trapping and dipole-

anion formation (a phenomenon not discussed in the publication by Tanese et al. [Tane2005]). In a

similar context of minority charge carrier trapping, the capacitance as a function of the applied bias in

polymer light-emitting diodes has proved this phenomenon of sign inversion maintaining the same

pulse polarity [Shro2005].

In the present investigation, transient current Id(t) measurements were performed to determine the

condition required to achieve steady state of charge transport at the conductive channel, and to

determine the relaxation time constants. These measurements revealed a non-steady state behavior of

the )t(V)t(I dd − characteristics for short time scales, i.e., less than 5 min. This means that there is a

strong influence of the stress biases Vg and Vd, respectively, on the charge transport.

A first attempt to simulate these transient effects was made by considering a model based on two

exponential contributions to the Id-Vd characteristics. The mathematical formalism relies upon the

basis of the fractional calculus [Hilf2004]. Although the present discussion is not devoted to an

analysis of the well-known transport equations, fractional calculus is considered the expedient

mathematical tool to model the multiexponential behavior of the )t(V)t(I dd − characteristics.

Therefore, the relaxation process is expressed as a linear combination of Debye exponential functions

rather than a single stretched exponential [Schi1993].

Non-exponential relaxation has been observed in the persistent behavior of the photo-induced Id

current after the photon source is switched off [Dutt2003]. Moreover, persistent photoconductivity in

inorganic semiconductors [Quei19886, Theo1982], meta-stable transport effects in pentacene single

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.3 crystals due to bias-dependent generation and the quenching of defects [Chi2004] have been

experimentally confirmed; drain current relaxation of printed RFID tags based on pentacene and

oligothiophenes [Subr2005] have proved the phenomenon as well. However, these developments are

lacking of discusions in terms of multiexponential kinetics.

8.2. Traps and charge density distribution

Previous analyses of photo-enhanced current in organic-based devices have been founded on the

assumption of Gaussian-like trap distributions [Godl2001]. Further experimental evidence suggests a

combination of such distribution and the time dependencies of the traps, i. e., they might be transitory

and/or static, as has been found in the present investigation.

The first kind of trapping appears when the device is stressed gradually by an external field. When the

device is continuously switched on and off, it preserves part of the previous history (memory effects),

in this way modifying the physical conditions of the traps and thereby acquiring a series of quasi-

equilibrium states.

A model to augment the comprehension of the trapping phenomena in organic electronics is based on

the extension of earlier theoretical investigations by Fermi and Teller in 1947 [Ferm1947]. The idea

was to determine the minimum electric dipole moment required to form a dipole-anion state.

8.2.1. Effect of the electric field

A specificity of weakly bound complexes is that they can undergo large deformations, as has been

experimentally evidenced on para-amino-benzoic-acid dimers (PABA dimers) [Vasc1999]. Because of

the vibrational motion. Despite the molecular dipole being zero in equilibrium, the average of the

square dipole moment for the dimer is different from zero, which leads to a high electric susceptibility.

To verify this, the vibration-induced dipole moment in the gas phase of a molecular complex was

determined [Comp2002]. Therefore, the total dipole moment, with all likely contributions ( iμ ), is

∑μi

i . (8.1)

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.4 In chapter 2, the nature of the probability of having a sort of band structure in organic solids was

discussed. Earlier experiments on several molecules investigated by means of x-ray excited

photoelectron valence band spectroscopy have demonstrated this dispersion relation for the organic

solid [Pire1984].

The quasi-band diagram depicted in Figure8.1 as a modification of a similar band diagram reported

elsewhere [Pire1984], shows the normal organic solid with intercalated dipole states. Consequently,

the charge carrier has three possibilities: to be driven, transmitted, reflected or trapped (bound) by the

dipolar center. It should be noted that the possibility of back-scattering is not rejected; this might play

a fundamental role at the metal/organic interface. Thus, The transmission is understood as tunneling

through the dipole-wall potential, and the charge trapping is seen as a quantum confinement. For the

last situation, a detailed description of the dipole-bound anion with spherical symmetry can be found

elsewhere [Rone2003].

Then it is also quite plausible to think about a molecular cluster arrangement with zero dipole, which

in an external field will produce sides with zero and nonzero dipole moments (cf. Figure 8.1). This

means that not all molecules would contribute to the charge transport, hence, the free charge carriers

face the presence of dipoles induced by the electric field, and will tend to be trapped (bonded),

forming dipolar anions. This is what is labeled as artificial traps in this research.

The total number of flaw states in the organic material is higher than that for isolated systems.

Therefore, it is convenient to consider a time-dependent total dipole moment following a linear

combination of exponential relaxations

))/texp(1())/texp(1()t( 22o11o τ−−μ+τ−−μ=μ −− , (8.2)

with 1o−μ and 2o−μ the expected maxima of dipole moments for the time constants 1τ and 2τ ,

respectively.

Figure 8.1. Schematic view of combined

artificial flaws or traps induced by the

external electric field dipole, and

intermolecular potential acting on the two-

dimensional charge carrier gas. The z

coordinate gives the strength of the

potentials, while the xy plane correponds

to spatial coordinates

The number of dipoles that can bond the free charge carriers is

)t()t()t( Ts μ=μ′+μ , (8.3)

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.5

with )t(sμ the static dipole moment and )t(μ′ the dipole moment, including finite effects and degree

of freedom (rotation, vibration). In this way, the number of (anion, cation)-dipole sites arN can be

calculated as follows

minsmin

sar D

)t(D

)t(N

′μ′

+μ

=−

, (8.4)

with sminD − the minimum static dipole moment and minD′ the minimum dipole moment, including

internal coordinates. Both quantities are required in order to form a dipole-anion bond configuration.

The calculation of both dipoles is done on the basis of solving the Schrödinger equation for a quasi-

free particle subjected to a dipolar potential,

⎟⎠⎞

⎜⎝⎛

′′+

′−+∇−=

r1

r1q

m2H 22

2h. (8.5)

The time-independent Schrödinger equation is

−

h2

2m∇2 + eq −

1′ r +

1′ ′ r

⎛ ⎝ ⎜

⎞ ⎠ ⎟

⎡

⎣ ⎢

⎤

⎦ ⎥ ψ = εψ ; (8.6)

by introducing the variables indicated in Figure 8.2,

the situation becomes azimuthally symmetric, where

the particle is confined in two dimensions. The

numerical solutions of eq. (8.6) thus result in

D625.1D smin =− [Byer1967, Ferm1947, Garr1971-

1982, Lapi1980, Turn1968-1977], approximately

50 % lower than the classical estimation, and

D]5.2;0.2[Dmin =′ .

Figure 8.2. Particle subjected to electric

dipole potential.

In the framework of the present discussion, the applied external electric field that can promote dipole-

anion bond states will effect an artificial trap (artificial flaw) distribution for the charge carriers.

Therefore, the total density of flaws is given by

tarT N)t(NN += , (8.7)

with Nar (t) the artificial flaw density determined by eq. (8.4).

Measurements of the current as a function of time provide important information in order to estimate Nar (t) ; this issue is discussed further in the chapter.

For this simple model, dipole-dipole bonds are not assumed. The same holds true for the transition rate

contribution between Nar (t) and Nt . Otherwise, the creation and annihilation of "particles" at the site

should be considered.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.6 The external electric field produces two effects: one is the polarization of the organic molecules and

the other is the driving of free charge carriers. If the molecular solid is not affected by the field, then

the traps are independent of that contribution, at least for moderate applied fields. What it is observed

is the appearance of "anomalous features”, related to creation of dipoles in the organic structure.

The dipoles act as scattering centers for the charge carriers. One special case is when the magnitude of

the dipole moment equals or surpasses a certain critical value [Ferm1947], increasing the probability

of dealing with electron-dipole bound states, responsible for the induced artificial flaw density.

Therefore, it is expected that within the framework of the standard trapping and detrapping modeling

the estimated trap densities will exhibit strong deviations from those expected from non-stressed

device.

Some consequences of the dipole-electron bound states on charge transport are

- negative differential conductance,

- anomalous contributions to the QTS or DLTS signal: enhancement of minority-charge carrier

trapping,

- reduction of the current,

- dynamics of the capture cross section, and

- threshold voltage dynamics.

An excess electron can be bound to many molecules and into a very diffusive molecular orbital as a

result of the long-range contributions of the molecular multipole electrostatic fields. In this context, further theoretical analysis requires diffuse molecular orbitals, in order to model the delocalization of

the electronic charge density.

An electron interacting with a molecule or molecular cluster which possesses a large dipole moment

may bond to the molecule or cluster to form a dipolar anion. Compagnon and colleagues reported the

first electric deflection measurement on the PABA dimer. It is characterized as being a weakly bound

molecular complex. The PABA molecule has a strong permanent dipole moment (3:9 D [Vasc1999])

but the dimer has a symmetrical structure, which is bound by a pair of hydrogen bonds [Meij1990],

and the two dipoles cancel each other at equilibrium.

A specificity of weakly bound complexes is that they can undergo large deformations. Because of the

vibrational motion, the average of the square of the dipole moment is non-zero, which leads to a high

electric susceptibility. This is the first observation of a vibration induced dipole moment of a

molecular complex in gas phase. The spectacular effect was predicted by Whitehouse [Whit1993], and

never observed for an isolated system.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.7 The results are interpreted by a simple and general model within the framework of the linear response

theory. It constitutes an original example of a system with an electric susceptibility that does not

follow the well-known Curie law. This proves the probability of observing electric-field induced

modification in the Raman spectra of structures with zero or negligible dipole moment under

equilibrium conditions.

8.3. Anomalous behavior of the QTS signal

In chapter 3, the description of the QTS method was addressed on the basis of interfacial phenomena.

In general, the applied electric field produces modifications not only in the charge carriers and or

trapped particles densities, but also in the molecular structure properties [Paez2003].

In terms of relaxations, the fast processes are achieved by those charge carriers with high mobility,

while for carriers with lower mobility, the relaxation time constant can be hundreds of times higher in

comparison with those filling up the faster capture process. Each time the electric field is applied the

"reservoir” (molecules) is modified and the new state is a quasi-equilibrium state.

8.3.1. Advantage of floating gate configuration in QTS measurements

The experimental spectra presented in this section were obtained by applying a bias UDS between the

source and drain, while keeping the gate electrode floating. It may be clear that the goal is to

characterize the properties of the channel material between the source and drain of an operating OFET.

Yet it is really important to explain why the floating gate is profitable. If one applies a dc bias between

the gate and source, the excitation pulse ΔU would appear across the gate oxide with the following

consequences:

– transient charging of the relatively high oxide capacitance through the Si gate electrode,

– populating Si/SiO2 interface states,

– populating SiO2/organics interface states,

- transient charging of the oxide capacitance through the organics (dielectric relaxation) [Agar1974],

and ion movements in the oxide.

Under such conditions, it is difficult to resolve the transient charge across the tiny capacitance of the

channel. Applying a bias pulse ΔU (cf. Figure 3.8(b)), there is a net bias UDS + ΔU during the pulse; in

most experiments, the reference bias potential was set to zero. After the pulse, the three-channel

correlator was activated to provide a ΔQ signal for each discrete delay t1, further details of the

experimental procedure are given in sect. 3.4.2.

The respective results for high- and low-power architectures will be presented separately.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.8

8.3.2. OFET devices with interdigitated source-drain electrodes

The QTS spectra measured in the interdigitated source-drain electrodes OFETs already mentioned in

chapter 3 are partially shown in Figure 8.3. The stressing pulse 1U (ΔU) was set between the drain and

source in the floating gate configuration. The pulses were applied according to the scheme depicted in

Figure 8.3(a), starting with –1.5 V and being gradually increased to –10 V.

For the first pulse amplitude of –1.5 V, an inverted satellite feature appears, peaking at around

s10 25.3− in the logarithmic time scale. This is attributed to the minority charge carrier trapping

[Zhao1990].

In order to confront bulky organic and the organic/metal interface contributions to the QTS signal, an

analysis of the QTS peak sign is given in the next section. The trap filling by minority charge carriers

is considered to be one of the mechanisms responsible for allowing the inversion of the QTS intensity

against the stressing potential.

The dynamic monitoring of the QTS signal during the experiment was achieved by recording the

average time per measurement. Therefore, the extracted 1τ determined by the correlated fitting

algorithm (Chap. 3) allows the temporal 1τ dependence illustrated in Figure 8.3(b) to be evaluated.

Each point of the panel (b) corresponds to the fitted spectra partially described in Figure 8.3(a) and

using the relationship defined in eq. (3.7). There is a fluctuating behavior of 1τ around the quasi-

equilibrium profile described by the red exponential decay.

-5.5 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5

0

3

6

9

12

15

Pulse U1

-9.0

-7.5

-5.5

-5.0

-4.5-4.0

-3.0

-10.0

-1.5

-3.0

-2.5-2.5-2.0-1.5

log10( t / s)

Transitory QTS Spectrain Pentacene (30 nm) FETs

QTS

sig

nal Δ

Q /

pC

(a)

0 20 40 60 80 1000

10

20

30

40

50Transitory QTS trappingparameter ( τ

1 ) in

Pentacene (30 nm) FETs

τ1 /

μ s

< τ = 63 min >

time / min (b)

Figure 8.3. Transitory QTS spectra (a), and

evolution of the trapping time constant of

pentacene (30 nm) FETs (b).

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.9

The time constant 1τ is independent of the stressing pulse. The advantage of gradually increasing the

pulse amplitude is that the temporal QTS evolution can be identified more precisely. The average time

relaxation constant of about min63=>τ< (cf. Figure 8.3(a) ) indicates that 1τ develops towards

lower values and indicates the gradual ionization from shallow to deeper traps as long as the device is

pulsed. Therefore, one consequence in the working OFET device is the dynamics of the threshold

voltage [Gome2004].

This memory phenomenon influences the Id-Vd characteristics, promoting a correlated hierarchical

scattering of charge carriers. The first signature of the 1τ dynamics is the decrease of the effective

charge carrier density, inducing the so-called current collapse (sect. 8.4), where the channel

conductance becomes negative due to the charge-carrier optical phonon interaction [Pop2005].

Afterwards, the previous dynamics is followed by a continuous decrease in the drain current (sect. 5.5)

with positive channel conductance.

In order to recover the pristine QTS signals, the sample should be relaxed. This means it should be

without any stress bias potential, since polarity inversion of the former applied pulse does not fully

reverse the structure to the initial state.

Using the experimental data points of Figure 8.3(b) for 1τ , and the reported effective mass for

pentacene [Gill2003] of o* m7.1m = and o

* m5.5m = , the dynamics of the capture cross section

( σ ) can be estimated, as is effectively described in Figure 8.4. The results coincide with typical values

reported in similar molecular structures [Ménd2006, Thur2005a-b].

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.10

0 100 200 300 4000

3

6

9

12

15

effective mass m* = 1.7 mo

m* = 5.5 mo

σ

x 10

-17 c

m 2

time / min

Dynamics of the capture crosssection in Pentacene OFETs

Figure 8.4. Temporal evolution of the capture

cross section ( σ ) in bias-stressed OFETs with

floating gate configuration. The estimate of σ

was carried out by fitting the capture probability

1/1 τ shown in Figure 8.3 and assuming

reported charge carrier effective masses of

o* m7.1m = and o

* m5.5m = [Gill2003].

On the other hand, the time dependence of σ is confirmed by the drain current collapse and drain

current relaxation in OFETs. Another important feature of this finding is the scattering of the charge

carriers, which is an important factor in the differential conductance of the conductive-organic

channel.

The time constant >τ< found by the transitory QTS measurement coincides in order of magnitude

with those determined from dd V)t(I − measurements (chap. 5), and from the relaxation of the

Raman band after removing the externally applied electric field. Therefore, the underlying mechanism

is the structural modification of the molecular solid by external fields [Paez2003].

8.3.3. Single-channel OFET devices

In order to skip multi-interfacial configurations, the QTS measurements of the single channel OFET

gives more direct information of the organic/inorganic interface. Here the investigated structures have

a channel geometry with width (W) and length (L)a of m16m100LW μ×μ=× . Previous

investigations into similar structures with current-voltage measurements and monitored during a

period of about 9 months, revealed a dynamics of the threshold voltage with a decrease in the mobility

by about 3 orders [Pann2004].

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.11 In contrast to the interdigitated geometry, here the amplification of the transitory QTS discussed in the

previous section is lost; the advantage of this geometry is the improvement in probing more local

charge relaxations in the organic layer. The experiment was carried out between 140 K and 350 K at

intervals of 10 K ± 0.5 K; the resulting measurements for easier recognition are partially addressed in

Figure 8.5(a).

The QTS spectra basically exhibit two main features: one is related to the expected peak polarity in

agreement with the applied negative pulse (negative peaks), while the other feature corresponds to the

positive peaks. Here again, the "anomalous behavior of the QTS signal" is revealed. Mechanisms

promoting the inverted peak sign contribution come from trap filling by minority charge carriers and

induced artificial flaws, i.e., local dipole distribution in the organic material. The first contribution is

due to higher mobility distribution of the minor charge carriers, while the second one is mediated by

the induced electric dipoles promoted by the applied field. In chapters 6 and 7, it was demonstrated by

time-dependent current-voltage and vibrational relaxation experiments, respectively, that longer

relaxation time scales for the organic based device are also possible.

Previous investigations by capacitance DLTS in pentacene organic films assembled as an interlayer in

MIS capacitors with a n+-Si/SiO2/pentacene/Au structure revealed positive and negative C-DLTS

features. These indicate a contribution of majority and minority trapping centers [Yang2002].

The anomalous QTS signal is indirectly observable not only for small molecules, but has also been

proved for long polymer chains, despite the reported literature lacks of discussion about the

multiexponential dynamics and dipole bound states [Tane2005]. Particular signatures of the relaxing

drain current - not discussed in the publication - are its anomalous nature and substantial decay time

constants [Tane2005].

A simple transformation of the QTS spectra displayed in Figure 8.5(b) can restore the relaxed charge

signal as a function of time. In this way, the time dependence of the charge dynamics resembles the

behavior of the drain current as a function of time in organic-gas-sensor FETs [Tane2005].

In previous chapters, the long time scale relaxation processes in organic-based FETS have been

experimentally demonstrated. Measurements of the drain current with and without illumination, and

Raman bands after switching off the external field in OFETs have yielded time scales ranging from

seconds to several hours. These time scales are a strong indication to devise new setups with a

modified rate window concept in order to undertake DLTS characterization on a broader time scale.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.12

(a)

(b)

Figure 8.5. Anomalous behavior of

the QTS signal in pentacene-based

OFETs. (a) QTS measurements

performed on pentacene (30 nm)-

based FET. The sample was

stressed by a negative pulse of -

10V between the source and drain

in the floating gate configuration.

Similar anomalies have been

observed in resistance-DLTS (R-

DLTS) measurements on GaAs-

based MESFETs [Zhao1990], and

from capacitive-DLTS (C-DLTS)

in organic interlayers of MIS

capacitors with structure n+-

Si/SiO2/pentacene/Au [Yang2002].

(b) Charge relaxation QTS

spectrum measured at 293 K.

Figure 8.6 shows the activation energies obtained after fitting the anomalous QTS spectra presented in

Figure 8.5. Results indicate a distributed trap of states and some of them are already ionized at room

temperature. It can be observed that the tiny feature appearing at room temperature resembles the one

found in the interdigitated geometry. The other peaks indicate the organic/metal interface.

For a given applied pulse, a change in the QTS signal preserving the sign of the stressing bias is

expected. The measured spectra show apparently additional components with inverted signs,

indicating the influence of the minority charge carriers, which have lower effective mass and higher

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.13 mobility. Therefore, they are the most probable to fill/escape from the occupied traps after removing

the applied pulse.

Figure 8.6. Activation energies of

pentacene (30 nm) based FETs

determined from the anomalous

QTS spectra sketched in Figure 8.5

8.4. Approach to modeling of the "anomalous QTS signal"

What has been found is that in transitory conditions the probability of minor charge carriers being

trapped is higher; therefore, the anomalous peak appears in the QTS spectra. Another finding is the

formation of anion-dipole sites, which become dominant when the minimum dipole required to bind a

charge carrier is reached.

Let’s consider the charge dynamics in the form

∑ τ−=i

ii )/texp()t(Q a , (8.8)

the coefficients ia are related to the sign of the net charge involved in the process with time constant

iτ .

As an example, a charge dynamics phenomenon composed of three processes with 031 >= aa and

02 <a is assumed. A likely temporal dependence of the charge Q is shown in Figure 8.7(a); in a

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.14 similar manner, the current and capacitance can be obtained. It is noted that )t(Q should not be

strictly positive; it is allowed negative values or even intervals with 0)t(Q = . The QTS spectra are

displayed in Figure 8.7 (b), where the time constants are marked by arrows and correspond to the peak

maximum for positive components and to the minimum for negative contributions. The charge

dynamics described in Figure 8.5(b) were determined with similar multiexponential kinetics.

Figure 8.7. Simulation of multiexponential kinetics of charge relaxation (a) and the corresponding

QTS spectrum indicating the relaxation time constants (b).

A partial experimental and theoretical description of the anomalous QTS signal was published by us

[Thur2006]. Some key features and extended insights into the model were given in the previous

sections and complemented in this paragraph.

As soon as an electric field ( )t(E ) is applied, the initial number of dipoles is modified. In linear

approximation, then,

)t(E)T,t()T,t(td

)T,t(dsidip

i χ=μ+τμ

, (8.9)

here the term )T,t(iμ means the average value of dipoles oriented along the "i" direction, T is

temperature, and )T,t(χ is the molecular susceptibility.

Introducing tx = , )T,t(y iμ= , 1dip)x(p −τ= , and )t(E)x(s sχ= , one has to solve a differential

equation of the type

( ) )x(syxp'y =+ . (8.10)

The initial internal electric field oE)t(E = is determined from the space charge sheet oσ and is

defined as roo / εεσ . The relaxation of the local field is assumed to be described as

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.15

)/xexp(E)x(E Do τ−= , with Dτ the dielectric relaxation time constant. Then the solution to eq.

(8.9) takes the form

( )⎥⎥⎦

⎤

⎢⎢⎣

⎡+

τχ

τ−= Caxexpa

E)/xexp(y

dip

0sdip , (8.11)

with )(a 1D

1dip

−− τ−τ= and C being a constant.

Considering the initial condition 0~)0(y i =>μ=< due to the in-plane anisotropy (parallel to the gate

dielectric substrate) molecular distribution, then a/EC dipos τχ−= . The DFT calculations

demonstrate that the single pentacene molecule has a non-zero electric dipole moment in equilibrium

conditions, as depicted in Figure 8.8.

Figure 8.8. Effective electric dipole

in pentacene. The calculations were

performed in Gaussian 98

[Gaus1998] and at the B3LYP/3-21G

level with and without an electric

field (dipole in D).

In this way, one obtains the time-dependent dipole moment )t(μ

( ) ( ) ( )[ ]dipDDdip

os /texp(/texp/1

Et τ−−τ−

ττ−χ

=μ , (8.12)

when simulating the QTS response component. Due to the polarization, the susceptibility is replaced

by the Langevin-Debye formula [Comp2002] and multiplied by the initial local electric field oE ;

therefore, the saturated polarization is given by

0e

2dip

os EkT3

~NE~

⎟⎟

⎠

⎞

⎜⎜

⎝

⎛α+

μ=χ=μ

∞, (8.13)

where k Boltzmann’s constant (1.38x10-23 JK-1), ]m[N 3dip

− is the concentration of dipoles, eα is the

static electronic polarizability, and the dipole moment is in C-m.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.16

For non-interacting harmonic oscillators and expanding 2~μ in terms of the normal coordinates, one

finds a linear proportionality between 2~μ and the temperature. Finally, eq.(8.12) reads

( ) ( ) ( )[ ]dipDedip

Ddip

o /texp(/texpk3

NB/1

Et τ−−τ−⎟⎟

⎠

⎞⎜⎜⎝

⎛α+

ττ−=μ , (8.14)

with B a constant to be determined.

In order to recover the charge dynamics given in eq.(8.8), one can readily describe the dipole in terms

of a center of dipole and then use the expression for the QTS signal addressed in chapter 3. The

simulation of the dipolar contribution to the QTS spectrum is described in Figure 8.9.

Figure 8.9. Simulation of the dipolar

contribution to the QTS spectrum for

three different dielectric relaxation

constants and one dipolar relaxation.

The dielectric relaxation of the local field was assumed to be a single Debye-like decay. As has been

seen in complementary current-voltage and Raman measurements, there are still contributions by

much lower time-relaxation rates. Despite the QTS measurements in a set-up with a limited rate

window range, the temporal envelope decay described in Figure8.3(b) predicts the existence of

features outside the minimum available rate window. Another interesting result is given by eq.(8.9),

which represents a key link between the vibronic and QTS measurements, predicting electrical and

vibronic signal detection in the lower and higher time-relaxation scales.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.17

8.5. Current collapse

8.5.1. Negative conductance and photodetachment

The I-V measurements done on previously unstressed samples exhibit negative behavior in the

channel transconductance (cf. Figure 8.8) This has been evidenced by the character of the temporal

variation of the Id-Vd characteristics, also known as current collapse [Klei2003]. The effect has been

observed in similar organic-based structures but not discussed [Hepp2003, Panz2005]; reports on

inorganic-based structures discussed these sorts of characteristics in more detail, indicating

phenomena involving a thermal-electric effect and leading to the power loss of the device [Paul1994].

Some authors argue a reversibility when the incident photon energy

exceeds the AlGaAs band gap near 1.8 eV [Klei2003], contrary to OFETs, where the effect of

illumination plays a selection role (sec. 6.3), i.e., the negative conductance remains even if the channel

is illuminated by some photon energies higher than the band-gap. For comparison, the drain current

characteristics for some intermediate photon energies are displayed in Figure 8.10(b). It can be

observed that the most efficient photodetachment process is achieved with photon energies of about

2.40 eV (cf. Figure 8.10(c)).

Recent investigations of molecular quantum dots with metallic contact electrodes and suspended

single-wall carbon nanotubes (SWCNTs) have shown a highly negative differential conductance

[Peng2006, Pop2005]. For the first system [Peng2006], the authors considered the Schottky barrier

formation at the metal/carbon nanotube to be responsible for the negative differential conductance,

while for the molecular quantum dots, polaronic effects mediated by the electron-phonon coupling

were identified. The latter proposal applies better to the output characteristics described in Figure 8.10

(a), since the vibronic-charge carrier interaction has been proved to be one of the signatures of current

relaxation. This permits determination of the loss of power of the device as proposed elsewhere for

metal-oxide FETs (MOSFETs) [Paul1994]. The Id(t)-Vd(t) measurements indicate that the condition

0d <g is transitory and this state remains for no longer than 5 minutes, despite the continuous decrease

of the drain current at a given drain voltage.

The time scale of the first Id-Vd during the experiments coincides with the time of measuring the QTS

bottom-most spectrum in Figure 8.3(a), taken at U1 = -1.5 V. This peak, presenting a maximum

distribution at 3.3)s/t(log10 −≈ ( )Tk/Eexp( BtΔ∝τ ), did not appear when the device was biased

again by the same pulse. The phenomenon is associated with a fast trap filling due to high mobility of

minority charge carriers.

Additionally, one has to take into account the artificial traps mediated by dipole-anion states which

can relax slower than the detrapping and trapping dynamics, thus hampering the restoration of the

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.18 initial condition stated for the bottom-most spectrum of Figure 8.3(a). In order to restore the initial

condition prior to the next pulsed biasing, the device was left to relax for about 8 h.

(a) (c)

Figure 8.10. Experimental evidence of "drain

current collapse" in OFETs, and negative

differential conductance ( 0g d < ) in OFETs.

Measurements in (a) darkness, and (b) under

illumination. (c) Drain current intensity as a

function of the excitation photon energy. The

photodetachment is more efficient with photon

energy of about 2.40 eV.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.19

8.5.1.1. Charge conservation and photodetachment

After applying a bias to the structures, the drain current drops temporarily, the reduction being an

indication of relaxation effects in the structure. Usually the current density is given as the transported

charge per unit time (I = dq/dt). If the stressing electric field and the sample conditions remain the

same, then the current intensity will remain constant and then the charge flow is not modified.

Therefore, in a given time interval, the charge in a control volume is zero. Now the variation of charge

in terms of the current density can be defined:

∫ Δ=Δ→Δ

tdIlimQ0t

. (8.15)

This means that the current is evaluated twice and the difference will tell us about the modification of

charge flow in the control volume. In ideal situations 0Q =Δ .

The modification of QΔ does not by any means imply that the charge conservation is violated. On

the contrary, within the framework of this research, the artificial flaw formation was considered by

applying external fields. Accordingly, if there are dipoles in the control volume, there is a probability

that at the output the number of particles is lower, leading to 0Q ≠Δ .

8.6. Photodetachment

The electron affinities of dipole-bound anions can also be inferred by detaching the excess electron via

application of an external electric field [Comp2001, Desf1994-1996, Hamm2003] or by

photodetachment [Cher2005, Dike2004, Hamm2004, Rau1971, Sind2004] i.e., by applying well-

defined photon energies to the structure Figure 8.7(c). During illumination, matrix elements are

formed which contribute to the photodetachment cross section [Cher2005, Rau1971, Sind2004] and

are proportional to the square matrix element of the dipole operator μ~ .

The differential cross section for an electron transition from the bound state χηλ to k , due to

interaction with a photon with frequency ω and electrically polarized along ε is

( ) ( ) kkk Ω⋅χηλπ

ω=ω→χηλσε dˆ~

c2km;d

2*

εμh

, (8.16)

with kΩ the solid angle subtended by the scattered electrons of effective mass *m . Following

Chernov [Cher2005] and doing additional algebra on the cross section, one can readily find the

effective cross section for the artificial traps in organic electronic devices by complementing with the

trap time constant

h1

1 Nk=τ − . (8.17)

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.20 As far as the Id recovery after illumination is concerned, it is merely a first order effect. This does not

mean that the interaction of photons with the traps in organic-based transistorsresults in the total

recovery of the charge transported in the channel, as stated by some authors for organic solids

[Stre2004] and inorganic-based transistors [Klei2003]. In the work of Street and colleagues

[Stre2004], long-term recovery of the devices in about two days was observed. The bipolaron model

can be extended, considering an additional charge trapping effect. The organic layer is divided into

two regions, one connected with the channel OBL and the second with the rest of the organic material.

This view can be utilized when admitting a polaronic interaction between the two regions, leading to

the phenomena already observed: bi-exponential decay of the drain current (chap. 6), negative

conductance, and the dynamics of the capture cross section. The effect can be explained in the

framework of slow carrier trapping, implying states with small and/or varying capture cross sections,

as experimentally found in the previous sections. The short time scales below 10-1 s, in which the QTS

measurements were done, might be explained by considering faster effects like first-order dipole

formation in the organic material. Additionally, both vibrational (Raman) and electrical

characterization techniques (Id(t)-V and QTS) give different time scales, prompting us to consider a

polaronic distribution in either time or energy.

Conclusions

The pentacene molecule possesses a natural tendency to form artificial traps induced by external

applied fields, in this way forming (cation, anion)-dipole sites with an identity similar to polaronic

systems, reducing the number of charge carriers transported through the channel. These traps can be

partially removed by applying well-defined photon energies.

The bulky features are considerably enhanced in OFETs with interdigitated source-drain electrodes,

while the interfacial contributions are much better observed in single channel devices. The

measurements performed in OFETs with interdigitated source-drain contacts considerably hinder the

QTS signal coming from the monolayers in contact with or close to (less than 10 nm as determined by

RSS) the metal electrode. For this reason, the device is more suitable to enhance bulky contributions

and the minority charge carrier dynamics participating in the detrapping and trapping filling process.

The complementary measurements on the single channel device allow for more scoping of the metal /

organic interface than the bulk contribution in comparison with the former device. This conclusion is

based on a simple picture of the metal/organic band diagram with two electrodes. It is found that the

difference in enhancement is about 103 which becomes appreciable when sensing the QTS signal in

the range from 10-6 s to 0.1 s.

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.21 The computed capture cross sections explain the threshold voltage shift of the working device.

The complementary photodetachment measurements with selected photon energies highlight the fact

that the channel conductance can be modulated to enable freezing or the dynamics of the charge

carrier capture cross section. Finally, in table 8.1, the main time scales of the organic channel in

OFETs involving vibrational-, charge transport, and trap-dynamics are summarized.

Table 8.1. Time-dependent phenomena in OFETs determined by complementary experimental

techniques (based on the setups used in this research).

Raman Current-Voltage DLTS

Time scale 20 min – days 1s – 59 min or days s1 μ - s1 and temporal averages

of hours

Involved

molecular and/or

transport

property

Structural, slow

polaronic effects.

Structural, internal molecular

domains. Current collapse,

negative differential

conductance

Traps, fast polaronic effect,

influence of free charge carriers

and dipoles (anomalous

response), charge carriers hot

optical phonons scattering

References

[Agar1974] S. C. Agarwal, “Analysis of the thermally stimulated capacitor-discharge method for characterizing localized states in amorphous semiconductors”, Phys. Rev. B 10, 4340-4350 (1974).

[Amy2005] G. Abstreiter, H. Brugger, and T. Wolf , H. Jorke and H. J. Herzog, “Strain-induced two-dimensional electron gas in selectively doped Si/SixGe1-x superlattices”, Phys. Rev. Lett. 54, 2441–2444 (1985).

[Byer1967] W. B. Brown and R. E. Roberts, “On the critical binding of an electron by an electric dipole“, J. Chem. Phys. 46, 2006 (1967).

[Cava2003] A. Cavallini, G. Verzellesi, A. F. Basile, C. Canali, A. Castaldini, and E. Zanoni, “Origin of hole-like peaks in current deep level transient spectroscopy of n-channel AlGaAs/GaAs heterostructure field-effect transistors”, J. Appl. Phys. 94, 5297-5301 (2003).

[Cher2005] V. E. Chernov, A. V. Dolgikh, and B. A. Zon, “Analytic description of dipole-bound anion photodetachment”, Phys. Rev. A 72, 052701 1-10 (2005).

[Chi2004] D. V., X. Chi, T. Siegrist, A. M. Sergent, and A. P. Ramirez, “Bias-dependent generation and quenching of defects in Pentacene”, Phys. Rev. Lett. 93, 076601 1-4 (2004).

[Comp2001] R. N. Compton and N. I. Hammer, in Advances in Gas-Phase Ion Chemistry, Vol. 4, edited by N. Adams and L. Babcock (Elsevier, New York, 2001), pp. 257–259.

[Comp2002] I. Compagnon, R. Antoine, D. Rayane, M. Broyer, and Ph. Dugourd, “Vibration induced electric dipole in a weakly bound molecular complex”, Phys. Rev. Lett. 89, 253001(1-4) (2002).

[Desf1994] C. Desfrancois, H. Aboul-Carime, N. Khelifa, and J. P. Schermann, “From 1/r to 1/r2 Potentials: Electron exchange between Rydberg atoms and polar molecules”, Phys. Rev. Lett. 73, 2436-2439 (1994).

[Desf1995] C. Desfrancois, “Determination of electron binding energies of ground-state dipole-bound molecular anions”, Phys. Rev. A 51, 3667 (1995).

[Desf1996] C. Desfrancois, H. Aboul-Carime, and J. P. Schermann, “Ground-state dipole-bound anions”, Int. J. Mod. Phys. B 10, 1339-13395 (1996).

[Dike2004] E. G. Diken, N. I. Hammer, and M. A. Johnson, “Preparation and photoelectron spectrum of the glycine molecular anion: Assignment to a dipole-bound electron species with a high-dipole moment, non-zwitterionic form of the neutral core”, J. Chem. Phys. 120, 9899-9902 (2004).

[Dutt2003] S. Dutta, and. K.S. Narayan, “Nonexponential relaxation of photoinduced conductance in organic field effect transistors”, Phys. Rev. B 68, 125208 (2003).

[Ferm1947] E. Fermi and E. Teller, “The capture of negative mesotrons in matter”, Phys. Rev. 72, 399–408 (1947).

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.22

[Garr1971] W. R. Garrett, “Critical binding of an electron to a rotationally excited dipolar system”, Phys. Rev. A 3, 961–972 (1971).

[Garr1980] W. R. Garrett, “Critical binding of electron–dipole rotor systems; electronically excited states”, The J. Chem. Phys. 73, 5721-5725 (1980).

[Garr1982] W. R. Garrett, “Excited states of polar negative ions”, The J. Chem. Phys. 77, 3666-3673 (1982). [Gaus1998] Gaussian 98 (Revision A.1), M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.

Cheeseman, V.G., Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A., Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M. W. Gill, B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. Head-Gordon, E.S. Replogle and J.A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.

[Gill2003] Gilles A. de Wijsa, Christine C. Mattheusa, R. A. de Groota, T. T.M. Palstraa, “Anisotropy of the mobility of Pentacene from frustration”, Synthetic Metals 139, 109-114(2003).

[Godl2001] J. Godlewski, G. Jarosz, and R. Signersky, “Photoenhanced current in thin organic layers”, App. Phys. 175-176, 344-350 (2001).

[Gome2004] H. L. Gomes, P. Stallinga, F. Dinelli, M. Murgia, F. Biscarini, D. M. de Leeuw, T. Muck, J. Geurts, L. W. Molenkamp, V. Wagner, “Bias-induced threshold voltages shifts in thin-film organic transistors”, App. Phys. Let.t, 84, 3184 (2004).

[Hamm2003] N. I. Hammer, K. Diri, K. D. Jordan, C. Desfrancois, and R. N. Compton, “Dipole-bound anions of carbonyl, nitrile, and sulfoxide containing molecules”, J. Chem. Phys. 119, 3650-3660 (2003).

[Hamm2004] N. I. Hammer, R. J. Hinde, R. N. Compton, K. Diri, K. D. Jordan, D. Radisic, S. T. Stokes, and K. H. Bowen, Jr., “Dipole-bound anions of highly polar molecules: Ethylene carbonate and vinylene carbonate”, J. Chem. Phys. 120, 685-690 (2004).

[Hepp2003] A. Hepp, H. Heil,W. Weise, M. Ahles, R. Schmechel, and H. von Seggern, “Light-emitting field-effect transistor based on a tetracene thin film”, Phys. Rev. Lett. 91,157406 1-4 (2003).

[Hilf2004] R. Hilfer, “Applications of fracional calculus in physcis”, chap. 1, World Scientific, Singapore, 2000. [Klei2003] P. B. Klein and S C Binari, “Photoionization spectroscopy of deep defects responsible for current collapse in

nitride-based field effect transistors”, J. Phys.: Condens. Matter 15, R1641–R1667 (2003). [Lapi1980] I. R. Lapidus, “Interaction of a charge and an electric dipole in one dimension”, Am. J. Phys. 48, 51 (1980). [Meij1990] G. Meijer, M. S. de Vries, H. E. Hunziker, and H. R. Wendt, " Laser desorption jet-cooling spectroscopy of

para-amino benzoic acid monomer, dimer, and clusters”, J. Chem. Phys. 92, 7625-7635 (1990). [Ménd2006] H. Méndez, “Organic modification of metal/semiconductor Schottky contacts (organisch modifiziert

Metall / Halbleiterkontakte)”, PhD Dissertation, Technische Universität Chemnitz (2006). [Panz2005] M. J. Panzer, Ch. R. Newman, and C. D. Frisbiea, “Low-voltage operation of a pentacene field-effect

transistor with a polymer electrolyte gate dielectric”, App. Phys. Lett 86, 103503 1-3 (2005).

[Paez2003] B. A. Paez, M. Bartzsch, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Combined electrical and Raman characterization of C60 based organic field effect transistors”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 63 (2003).

[Paez2005] B.A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F (2005).

[Pann2004] Ch. Pannemann, T. Diekmann, and U. Hilleringmann, “Degradation of organic field-effect transistors made of Pentacene”, J. Mater. Res. 19, 1999-2002 (2004).

[Paul1994] R. Paul, „MOS-Feldeffekttransistoren”, Halbleiter-Elektronik 21, pp. 392-397, Springer-Verlag, Berlin, 1994.

[Peng2006] H. Peng, and A. Zettl, “Negative differential conductance in suspended semiconducting carbon nanotubes”, 2006 APS March Meeting, March 13–17, Baltimore, MD 2006.

[Pire1984] J. J. Pireaux and R. Caudano, “Experimental picture of the band structure formation in a solid”, Am. J. Phys. 52, 821-826 (1984).

[Pop2005] Eric Pop, David Mann, Jien Cao, Qian Wang, Kenneth Goodson, and H. Dai, “Negative differential conductance and hot phonons in suspended nanotube molecular wires”, Phys. Rev. Lett. 95, 155505 1-3 (2005).

[Quei1986] H. J. Queisser and D. E. Theodorou, “Decay kinetics of persistent photoconductivity in semiconductors”, Phys. Rev. B 33, 4027-4033 (1986).

[Rau1971] A. R. P. Rau and U. Fano, “Theory of photodetachment near fine-structure thresholds“, Phys. Rev. A 4, 1751–1759 (1971).

[Rone2003] S. Ronen, “Electron structure of a dipole-bound anion confined in a spherical box”, Phys. Rev. A 68, 12106 1-7 (2003).

[Schi1993] H. Schiessel and A. Blumen, "Hierarchical analogues to fractional relaxation equations“, J. Phys. A: Math. Gen. 26, 5057-5069 (1993).

[Shro2005] V. Shrotriya and Y. Yang, “Capacitance–voltage characterization of polymer light-emitting diodes”, J. Appl. Phys. 97, 054504 1-6 (2005).

[Sind2004] M. Sindelka, V. Spirko, and P. Jungwirth F. Wang, S. Mahalakshmi, and K. D. Jordan, “Calculation of the photodetachment cross sections of the HCN– and HNC– dipole-bound anions as described by a one-electron

Chapter 8 B. A. Paez-Sierra, Trap distribution in OFETs… 8.23

Drude model”, J. Chem. Phys. 121, 1824-1829 (2004). [Subr2005] V. Subramanian, J. M. J. Frechet, P. C. Chang, D. C. Huang, J. B. Lee, S. E. Molesa, A. R. Murphy, D. R.

Redinger, and S. K. Volkman, “Progress toward development of all-printed RFID Tags: Materials, processes, and devices”, Proc. IEEE 93, 1330-1338 (2005).

[Stre2004] R.A. Street, A. Salleo, and M.L. Chabinyc, “Biplaron mechanism for bias-stress effects in polymer transistors” Phys. Rev. B 68, 085316 1-7 (2004).

[Tane2005] M. C. Tanese, D. Fine, A. Dodabalapur, L. Torsi, “Interface and gate bias dependence responses of sensing organic thin-film transistors”, Biosensors and Bioelectronics 21, 782–788 (2005).

[Theo1982] D. E. Theodorou, H. J. Queisser, and E. Bauser, “Profiling of deep impurities by persistent photocurrent measurements”, Appl. Phys. Lett. 41, 628-629 (1982).

[Thur2004] I. Thurzo, G. Pham, and D. R. T. Zahn, “A time-domain analysis of dipolar effects in copper phthalocyanine thin films on indium–tin–oxide substrates”, Semicond. Sci. Technol. 19, 1075-1080 (2004).

[Thur2005a] I. Thurzo, H. Méndez and D. R. T. Zahn, “Analysis of charge transient spectroscopy data originating from Gaussian densities of electron states in organics”, phys. stat. sol. (a) 202, 1994–2007 (2005).

[Thur2005b] I. Thurzo, H. Méndez and D. R. T. Zahn, “Dielectric relaxation in a hybrid Ag/DiMe-PTCDI/GaAs device”. J. non crystal. Sol. 351, 2003-2008 (2005).

[Thur2006] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, “Anomalous charge relaxation in channels of pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 203, 2326-2340 (2006).

[Turn1968] J. E. Turner and V. E. Anderson and Kenneth Fox, “Ground-state energy eigenvalues and eigenfunctions for an electron in an electric-dipole field”, Phys. Rev. 174, 81-89 (1968).

[Turn1977] J. E. Turner, “Minimum dipole moment required to bind an electron-molecular theorists rediscovered phenomenon mentioned in Fermi-Teller paper”, Am. J. Phys. 45, 758-766 (1977).

[Vasc1999] M. E. Vaschetto, B. A. Retamal, and A. P. Monkman, “Density functional studies of aniline and substituted anilines”, J. Mol. Struct. (Theochem) 468, 209-221 (1999).

[Whit1993] D. B. Whitehouse and A. D. Buckingham, “The vibrational contribution to the polarisability of endohedral [C60M]n+ complexes (where M = metal atom)”, Chem. Phys. Lett. 207, 332-338 (1993).

[Yang2002] Y. S. Yang, S. H. Kim, J.-I. Lee, H. Y. Chu, L.-M. Do, H. Lee, J. Oh, T. Zyung, M. K. Ryu and M. S. Jang, “Deep-level defect characteristics in pentacene organic thin films”, App. Phys. Lett. 80, 1595-1597 (2002).

[Zhao1990] J. H. Zhao, “Modeling the effects of surface states on DLTS spectra of GaAs MESFET’s”, IEEE Transactions on Electron Devices, 37, 1235-1244 (1990).

Chapter 9 B. A. Paez-Sierra, Summary

9.1

Chapter 9

Summary

Within the framework of the present research, the formation of top metal contacts, consisting of

indium and magnesium deposited onto two Perylene derivatives, 3,4,9,10-perylene tetra-carboxylic

dianhydride (PTCDA) and N, N' dimethyl-3,4,9,10-perylene tetracarboxylic diimide (DiMe-

PTCDI), were investigated in situ and under UHV conditions by resonant Raman spectroscopy. The

metal/organic structures were assembled on sulphur-passivated Si-doped GasAs(100) substrates (S-

GaAs). The experimental results have proved that all metals deposited onto the organic layers of

PTCDA or DiMe-PTCDI promote enhancement of the normally Raman-active internal vibrational

mode intensities, accompanied by the activation of normally infrared-active modes. The promotion

phenomenon is called surface enhancement Raman spectroscopy (SERS).

The magnification of the scattering Raman cross section is basically due to two factors. The first one

is associated with the close proximity of the molecules and the metal atoms, realizing in this way

activation of former non-Raman active modes assisted by the metal-induced molecular distortion. This

signature is also referred to as chemical enhancement, because an interface formation at the level of

organic-metallic mono-layers is achieved. The second factor is the enhancement achieved by the huge

local electric field developed at the metal clusters and extending towards the organic film.

Consequently, further electronic and nuclear distortions appear on the molecular side, leading to

enhancement of the molecular dipole and alterations of the polarizability. In this context, the

phenomenon is understood as a long-range interaction. It has been shown that metal coverage of

several nanometers of about 40 nm or above still allow the identification of vibrational bands, taking

into account the roughness of the metallic layer. This contribution to the SERS effect is referred to as

electromagnetic enhancement. The enhancement factors are estimated to be 101-102 for the chemical

contribution or first monolayer effect and 104 for the electromagnetic contribution. Contrary to

standard SERS investigations, where the molecules are deposited on metal substrates, or single

molecules embedded in metal-like solvent media, here the metal atoms were deposited onto the

organic structure, resulting in a more efficient enhancement of the related vibrational features at the

interface. The break-down of the selection rules is characteristic of molecules in the very near vicinity

of a metal surface and can be induced by several mechanisms: structural deformation of the molecule,

charge transfer from the molecule into the metal or vice versa, or formation of new chemical bonds.

Thus the spectral changes induced by SERS can be used to extract information about the chemical

reactions at the interface, as well as the morphology of the metal film.

Chapter 9 B. A. Paez-Sierra, Summary

9.2

Indium deposition onto PTCDA and DiMe-PTCDI caused only molecular distortion, with a

remarkable metal diffusion into In/PTCDA structures, as demonstrated for the metal coverage on

organic monolayers and onto thicker (15 nm) films. The phenomenon was much less obvious in

DiMe-PTCDI thin films.

The deposition of Mg on both perylene derivatives was accompanied by much lower diffusion of the

metal into the organic layers compared to In, as demonstrated by the preservation of the external

modes upon metal coverage. The Mg/PTCDA structures are formed in two stages; one is the formation

of a new molecular structure at the interface, which continues until a nominal metal coverage of about

2.8 nm is reached, which is attributed to the removal of the oxygen atom from in the anhydride groups,

and the second one is surface-enhanced Raman spectra of the former structure by further depositions

of Mg.

In the case of Mg/DiMe-PTCDI, it has been found that the molecule exhibits a break-down of

selection rules as well, and there is no formation of new molecular species, contrary to the case with

Mg/PTCDA interfaces. This structure is characterized by the coupling between discrete molecular

states of the organic DiMe-PTCDI material and the electronic continuum states of the Mg metal

contact. The phenomenon was observed by the asymmetric broadening of the modes at 221 cm-1, 1291

cm-1 and 1606 cm-1 upon the metal deposition. These features amount to energy gap states above the

HOMO of 30 meV, 160 meV and 200 meV, respectively. Their corresponding energy line-shapes are

well described by the Breit-Wigner-Fano function. The structure is well preserved by identification of

the vibrational band of the imide groups.

The investigations on the previous heterostructures helped to experimentally analyze the channel

formation of pentacene-based field effect transistors. The organic channel was formed gradually by

molecular beam deposition of the organic molecules under UHV conditions, and at an evaporation rate

of 0.65 Å/min. After each deposition of the organic molecule (82 times), the resulting interlayer was

vibrationally and electrically characterized in situ in order to determine the minimum nominal

thickness of the organic material required for efficient charge transport through the OFET channel.

At a thickness of around 1.5 nm nominal coverage (1.1 ML), the first percolation paths through the

first organic monolayer develop, resulting in a sharp rise in the drain current. Up to a nominal film

thickness of about 32 nm, a subsequent slower increase of the drain current can be observed,

revealing that the percolation of the first monolayer continues at a slower pace up to thick organic

layers. The extracted areas of the decomposed vibrational spectra peaks as functions of the layer

thickness undergo a four-stage process with a well-defined staircase-like tendency, where the first step

is extending to about 10 nm, the second one going on up to 20 nm, the third one to 25 and the last one

to 32 nm or above.

Chapter 9 B. A. Paez-Sierra, Summary

9.3

This thickness dependence is closely related to the layer compactness; therefore, the thicker the layer

the smaller the terrace-like feature. The earlier organic coverages below 1.5 nm (1.1 ML) are seen to

be composed of isolated molecular structures, since no drain current was observed, while for the

thicker layers below 5 nm the channel becomes compact and the highest contribution to the

transported charge through the channel is reached. Above the quoted first-stage thickness, the

contribution of the overlayers to the current is not significant and their compactness is reduced.

These two correlations indicate that the charge induced by the gate voltage is initially confined to the

first compact channel with a thickness of about 1.5 nm.

Signatures of the first monolayers are well defined in comparison to thicker films, where the 1.1 ML

exhibits a strained nature due to its direct contact with the gate insulator, resembling in this way

inorganic-based transistors with a channel formed by a strained material at the gated dielectric and

followed by a graded active material towards outer layers away from the insulator, with recovery of

the bulk-structural nature.

It was found that the conductive channel, here referred to as –the Organic Boosting Layer (OBL)

exhibits a compressive deformation, demonstrating the phenomenon of strain due to the pristine

coverage in closest contact to the inorganic substrate. This type of deformation is related to an increase

of the atomic bond strength, as was proved by comparison of the 1.1 ML coverage (1.5 nm) with the

thicker pentacene films (32 nm).

The output characteristics of the OFETs were measured after the final coverage. It was found that the

drain current undergoes a relaxation process with two decay time constants, one in the order of 101

min and the other below 102 min. A similar experiment, involving illumination of the channel with a

676.4 nm laser source, enhances the drain current density, leaving the relaxation time constants

unmodified. Additionally, the devices were characterized ex situ by charge deep-level transient

spectroscopy QTS. The spectra showed positive and negative peaks of the relaxed charge with respect

to the unique bias pulse polarity. The phenomenon has been called "anomalous behavior of the Q-

DLTS signal" and seen for the first time in OFETs. Consecutive QTS measurements of the device over

a period of several hours showed a slow relaxation of the anomalous detrapping time constant

around s5 μ ; its temporal envelope function decayed with a time constant of ca. 63 min < 102 min.

Based on these results and assuming the effective masses o* m7.1m = and o

* m5.5m = for

pentacene, the capture cross sections were 217 cm10]15;1[ −× and 217 cm10]5.1;1[ −× for the smaller

and bigger effective masses, respectively. Additionally, calculations of the two-dimensional charge

density were performed in an algorithm based on the Gauss quadrature and developed in the frame of

this thesis. This computing procedure is of great advantage to solve the Fermi-Dirac integrals without

significant approximations.

Chapter 9 B. A. Paez-Sierra, Summary

9.4

Further investigations of the OFETs with a combination of Raman spectroscopy and applied electric

fields revealed magnification of the Raman cross section comparable to the chemical enhancement in

experiments based on SERS upon In or Mg deposition onto the perylene derivatives PTCDA and

DiMe-PTCDI. The Raman intensity was found to be proportional to the applied field followed by a

saturation and then by a decrease. The Raman bands were monitored after switching off the electric

field, pointing to a structural relaxation time constant of about 94 min.

Based on the time-dependent current-voltage characteristics and QTS measurements, the phenomenon

is interpreted as artificial generation of dipoles acting as traps for the charge carriers. Therefore, the

total number of traps is due to the existing defects at the interface and those induced by the applied

electric field in the organic material. This formulation was extensively explored and theoretically

substantiated in this work.

The reported experiments on the formation of organic/metal heterostructures and the combined Raman

spectroscopy with electrical characterization of OFETs suggest that refinements of the transport

models in organic-based structures should be pursued.

B. A. Paez-Sierra, List of figures 10.1

List of figures Figure 1.1. Organic materials forming different interfaces, i.e.metal / organic semiconductor/

inorganic semiconductor heterostructure.

1.3

Figure 2.1. Some representative molecules of the arene family. (a) APentacene and (b) Perylene

derivatives. (*) Main structures investigated in the present research.

2.1

Figure 2.2. σ−hybrids and π−molecular orbitals in benzene. a) Localized σ−orbitals, b) pz atomic

orbital, and c) delocalised π−orbitals with highest densities above and below the plane ring.

(Thanks to G. Gavrila).

2.2

Figure 2.3. Molecular states of an organic solid [Ishii1999, Ménd2006, Pire1974-1984]. 2.3

Figure 2.4. Spectrum of scattered light showing the Raman Stokes, Rayleigh, and Raman anti-

Stokes bands.

2.8

Figure 2.5. Molecular Morse potential of the (a) ground electronic and bound vibrational states.

(b) Infrared activity described by transitions between vibrational states at a given electronic state

mljneE .

2.11

Figure 2.6. Non-resonant Raman effect involving (a) elastic light scattering or Rayleigh

scattering, and the inelastic processes: (b) Stokes scattering with the scattered photon energy

higher than the incoming one, and (c) anti-Stokes where the scattered photon possesses lower

energy than that of the excitation source.

2.12

Figure 2.7. Resonant Raman effect (a) vibronic transitions pointing to the Raman-Stokes,

Raman-anti-Stokes, and Rayleigh scattering.

2.13

Figure 2.8. Examples of SERS from a (a) pentacene (30 nm)-based OFET, and a (b) In (15 nm)/

PTCDA (15 nm)/S-GaAs heterostructure.

2.15

Figure 3.1. Experimental setup for combined Raman spectroscopy, scanning Raman

spectroscopy, and current-voltage (I-V) characteristics measurements.

3.2

Figure 3.2. Substrate passivation: ex situ chemical treatment and in situ annealing and material

deposition.

3.3

Figure 3.3. Molecular structure of perylene derivatives with the associated symmetry group of

(a) PTCDA and (b) DiMe-PTCDI [Kobi2004, Salv2003]. (c) UV-vis absorption spectra of

organic layers deposited on quartz.

3.5

Figure 3.4. Molecular structure of (a) pentacene (C22H14) and its 102 internal vibrational modes

divided into Raman active, IR active, and silent bands, belonging to the D2h symmetry group

[Ross2002], (b) Fullerene C60, which belongs to the symmetry group of the truncated

icosahedron [Kost1994]. (c), (d) Absorption spectrum of a 30 nm pentacene and 30 nm C60 film

deposited on quartz, respectively [Kolo2005]. The excitation energies for resonant Raman

spectroscopy measurements are indicated on the spectra.

3.7

B. A. Paez-Sierra, List of figures 10.2 Figure 3.5. Field effect structures used for the fabrication of organic-based field effect

transistors (OFETs). (a) Interdigitated structures and (b) single channel structures.

3.8

Figure 3.6. Field effect structure and formation of the channel by organic molecules of

pentacene or C60. The polarity of the Vg and Vd depends on the charge carrier type i.e. n or p

(chapter 5).

3.9

Figure 3.7. Band diagram for a semiconductor with a single deep level trap (recombination

processes between HOMO-LUMO or intermediate states are excluded).

3.10

Figure 3.8. Charge transient spectroscopy (QTS) based on the rate window concept. (a) Sample

wiring, and (b) applied bias pulse. (c) Triple boxcar integrator and; (d) output signal displayed

as a function of rate window time, the QTS maximum coincides with the relaxation time

constant τ of the trap (see sample spectra below for three different time constants) [Thur1994,

2005-2006].

3.13

Figure 3.9. Normalized QTS responses to three exponential decays with different time constants

τ are peaking when the processing starts at t1 = τ. It should be noted that both the height ΔQm

and the FWHM are invariant against τ; approximation of the fastest charge relaxation by a

Gaussian is shown by circles; w* stands for the variance of the Gaussian.

3.14

Figure 3.10. Flux diagram of the correlated fitting algorithm. 3.16

Figure 4.1. Raman spectra of In (5nm), Ag (4.5 nm) and Mg (5 nm) coverages on 15 nm thick

PTCDA films, compared with the spectrum of the bare PTCDA film in the spectral region of the

internal breathing mode (left) and in the spectral region of HC− deformation and C=C

stretching modes (right). (The Raman spectra involving Ag do not belong to this work, and are

addressed elsewhere [Salv2003]. They are presented here for comparison of metal contact

formation on similar molecular structures).

4.4

Figure 4.2. Raman spectra of In (5nm), Ag (4.5 nm) and Mg (6 nm) coverages on 15 nm thick

DiMe-PTCDI films, compared with the spectrum of the bare DiMe-PTCDI film.

4.6

Figure 4.3. Enhancement factors of the Bu mode (1243 cm-1 in PTCDA and 1246 cm-1 in DiMe-

PTCDI) and of the C-C stretch Ag mode (1572 cm-1 in PTCDA and 1570 cm-1 in DiMe-PTCDI)

for PTCDA (a), and DiMe-PTCDI (right) as a function of the metal coverage (b).

4.7

Figure 4.4. AFM topographic images of a 30 nm thick In film on PTCDA. (a) (right part

showing PTCDA covered by In clusters) and of a 113 nm thick Mg film on PTCDA (b).

4.8

Figure 4.5. Spectra of external Raman modes from 15 nm thick PTCDA films capped with 0.4

nm metal layers, i.e., Ag, In, and Mg. The spectral Raman shift between 25 cm-1 and 125 cm-1

corresponds to the libronic or collective modes of the interacting molecules in the unit cell

[Salv2003].

4.10

B. A. Paez-Sierra, List of figures 10.3 Figure 4.6. Raman monitoring in the external mode region upon metal deposition: (a) Ag, (b)

Mg, (c) In. The experimental spectra are shown by open symbols and the fitted spectra by red

lines. The Lorentzian functions used for the fitting of the Raman spectrum of the pure PTCDA

film are shown by lines in the lower parts of the figures. The spectra of Ag/PTCDA are

normalized for a better resolution of the phonons.

4.11

Figure 4.7. Evolution of the FWHM of the external mode at

41 cm-1 as a function of the metal coverage relative to the initial values before the metal (Mg,

Ag) deposition. The dashed lines are visual guides.

4.12

Figure 4.8. Raman spectra of Mg/DiMePTCDI in the region of: (a) external modes and the

breathing molecular vibration mode. (b) C-C and C-H modes. The spectra in (a) and (b) are

normalized with respect to the intensity of the breathing mode and to that of the C-C stretching

mode at 1570 cm-1, respectively. An asymmetric broadening develops for the three modes

marked with stars upon the Mg deposition.

4.14

Figure 4.9. Comparison between the Raman spectra of bare DiMe-PTCDI, Mg (2.8 nm) / DiMe-

PTCDI and the IR for the organic

4.15

Figure 4.10. Fitted Raman spectra of Mg / DiMe-PTCDI: from bottom to top: bare 15 nm

DiMe-PTCDI covered with 34 nm Mg and 122 nm Mg. The peaks fitted with the BWF function

are represented by thick black lines.

4.16

Figure 4.11. Energy level alignment of the DiMe-PTCDI / Mg heterostructure. The Fano

resonances indicated in the band diagram were obtained from the resonant Raman

measurements, while the other energy levels were quoted from NEXAFS spectroscopy

measurements on a similar sample [Gavr2006].

4.17

Figure 5.1. Scheme of a field effect transistor. The drain and source terminals serve to drive the

modulated current through the channel shown in blue color. The charge in the channel is

modulated by the capacitive effect of the third (isolated) terminal named gate.

5.2

Figure 5.2. Argument of the Fermi-Dirac integral as a function of the reduced energy ε , and the

reduced chemical potential η (evaluated in MatLab [MATL2003]). Orders (j) of the arguments

(a) 21 , (b) 2

1− , and (c) 0.

5.5

Figure 5.3. Charge density distribution as a function of the reduced chemical potential and

reduced transfer integral. The effective masses for pentacene are o* m7.1m = and

o* m5.5m = [Wijs2003]

5.6

Figure 5.4. Triclinic crystal structure to estimate the number of molecules per 2cm . 5.7

Figure 5.5. Organic field effect transistor 5.8

B. A. Paez-Sierra, List of figures 10.4 Figure 5.6. Energy band diagram between the gate and the organic film with an isolating

interlayer, (a) under equilibrium conditions. (b) Accumulation mode of the organic field energy

bands for negative and (c) positive (c) gate voltages ( gV ) respectively. (d) Band diagram

between the organic and the contacts drain and source (UDS = 0 V).

5.9

Figure 5.7. Output characteristics of a field-effect transistor and maximum drain current where

the saturation starts. (a) Simulation of an ideal FET and (b) experimental drain current - drain

voltage of a pentacene (30 nm)-based FET; the solid lines correspond to the fitted output

characteristics.

5.13

Figure 5.8. Field effect mobility in a pentacene (30 nm)-based FET. The values (half filled

circles) were extracted from the output characteristics, while the solid curve corresponds to a

power law with the applied gate field.

5.14

Figure 5.9. Threshold voltage of a single channel pentacene (30 nm)-based FET. The values

were extracted after fitting the output characteristics depicted in Figure 5.8(b)

5.17

Figure 6.1. Simultaneous monitoring of Raman bands (black spectra) and drain current (red

curve on the right hand side of the vertical plane) during the pentacene deposition. The pointed

Raman bands correspond to the in-plane C-C ring and in-plane C-H vibrations.

6.4

Figure 6.2. Extracted intermediate Raman spectra of pentacene based OFETs at different

organic layer thicknesses, together with the corresponding fitting line-shapes (a); and

comparison between the area of the Raman band at 1179 cm-1 and the drain current as a function

of the organic film thickness (b).

6.5

Figure 6.3. Fitting parameters in dependence on the molecular layer thickness in pentacene-

based OFETs (a) Raman band intensities and (b) FWHM broadening.

6.7

Figure 6.4. In situ Raman band measurements of pentacene (1.5 nm) forming different

interfaces when deposited on Au and SiO2 substrates.

6.8

Figure 6.5. Principal regions of the organic layer forming the channel in OFETs 6.10

Figure 6.6. Charge density distribution induced in the organic layers by applying different gate

voltages. (a) Monolayer stratification and (b) dependence of the charge density as a function of

the layer location parallel to the gated dielectric with the gate voltage as parameter.

6.12

Figure 6.7. Raman spectra of the organic/inorganic interface. (a) Sketch of the swept interface in

pentacene-based OFETs, (b) 3D plot of the Raman signal as a function of the vibronic bands

and spatial position in the organic/inorganic interface, and (c) Profile of the Raman intensity

(middle up half filled circles) and derivative (middle down filled circles) as a function of

scanning position.

6.13

Figure 6.8. Fitted FWHM of the Raman bands at different spot scanning positions (a), and inset

of the FWHM beneath the organic/inorganic interface (b).

6.14

B. A. Paez-Sierra, List of figures 10.5 Figure 6.9. Effect of bias stress on the dd V)t(I − characteristics of OFETs (a) in darkness and

(b) under illumination. The red contour lines indicate the time profile behavior of the Id current

for a fixed Vd voltage.

6.15

Figure 6.10. Simulation of one of the dd V)t(I − characteristics shown in Figure 6.9.

Experimental curves are plotted with circles. The quantities Io, A1, A2, τ1, τ2, were determined

from the red lineprofiles shown in Figure 6.9. To reproduce the experimental curve here

(uppermost curve), the time t related to the drain sweep rate of 10 V/min.

6.17

Figure 7.1. Benzene molecule under different external electric fields. (a) Molecular structure,

(b) chart summarizing the band gap modification by the applied fields, and (c) HOMO and

LUMO configuration in response to the applied electric fields. (The unit H in the table means

Hartree = 27.2 eV).

7.3

Figure 7.2. DFT simulation of the influence of an applied electric field on the molecular band

gap and HOMO-LUMO states in pentacene. (The unit H in the table means Hartree = 27.2 eV).

7.4

Figure 7.3. Calculated vibrational states of a pentacene molecule under the influence of external

electric fields. (a) Orientation of the pentacene molecule, and (b) variation of the frequency

assignment as a function of the vibrations without an applied electric field.

7.6

Figure 7.4. Experimental measurements of the Raman bands in pentacene (30 nm) for different

applied fields. The involved modes of this spectral region belong to in-plane vibrations of C-H

bonds of the outer ring and antiphase vibrations of the C-H bonds parallel to the main axis of

the molecule.

7.7

Figure 7.5. Correlated fitting of the Raman intensity as a function of the applied gate voltage

(the digits behind the floating point were delivered after the correlated fitting procedure).

7.8

Figure 7.6. In situ measurements of the Raman bands relaxation after switching off the electric

field, and (b) time decay of the Ag band at 1179 cm-1 after switching off the electric field.

7.9

Figure 7.7. Raman spectra (upper plot) and differential Raman spectra (lower plot) of 3 nm C60

layer, the latter obtained by subtracting the spectra without an electric field from the ones under

the influence of the gate voltage.

7.10

Figure 7.8. Photocurrent of pentacene (30 nm) based FET. The spectra were taken for a drain

voltage of –40 V and gate voltage of 0 V (blue) and –10 V (green). The red spectrum

corresponds to the absorption of 30 nm pentacene deposited on quartz.

7.11

Figure 7.9. Persistent behavior of the drain current in pentacene (30 nm)-based FET after

switching off the photon source. The decay follows multi-exponential kinetics.

7.12

Figure 7.10. Photocurrent and dark current in 30 nm pentacene FET. The sample was

illuminated by coherent light of 676.4 nm.

7.13

B. A. Paez-Sierra, List of figures 10.6 Figure 8.1. Schematic view of combined artificial flaws or traps induced by the external electric

field dipole, and intermolecular potential acting on the two-dimensional charge carrier gas. The

z coordinate gives the strength of the potentials, while the xy plane correponds to spatial

coordinates

8.4

Figure 8.2. Particle subjected to electric dipole potential. 8.5

Figure 8.3. Transitory QTS spectra (a), and evolution of the trapping time constant of pentacene

(30 nm) FETs (b).

8.8

Figure 8.4. Temporal evolution of the capture cross section (σ ) in bias-stressed OFETs with

floating gate configuration. The estimate of σ was carried out by fitting the capture probability

1/1 τ shown in Figure 8.3 and assuming reported charge carrier effective masses of

o* m7.1m = and o

* m5.5m = [Gill2003].

8.10

Figure 8.5. Anomalous behavior of the QTS signal in pentacene-based OFETs. (a) QTS

measurements performed on pentacene (30 nm)-based FET. The sample was stressed by a

negative pulse of -10V between the source and drain in the floating gate configuration. Similar

anomalies have been observed in resistance-DLTS (R-DLTS) measurements on GaAs-based

MESFETs [Zhao1990], and from capacitive-DLTS (C-DLTS) in organic interlayers of MIS

capacitors with structure n+-Si/SiO2/pentacene/Au [Yang2002].

(b) Charge relaxation QTS spectrum measured at 293 K.

8.12

Figure 8.6. Activation energies of pentacene (30 nm) based FETs determined from the

anomalous QTS spectra sketched in Figure 8.5

8.13

Figure 8.7. Simulation of multiexponential kinetics of charge relaxation (a) and the

corresponding QTS spectrum indicating the relaxation time constants (b).

8.14

Figure 8.8. Effective electric dipole in pentacene. The calculations were performed in Gaussian

98 [Gaus1998] and at the B3LYP/3-21G level with and without an electric field (dipole in D).

8.15

Figure 8.9. Simulation of the dipolar contribution to the QTS spectrum for three different

dielectric relaxation constants and one dipolar relaxation.

8.16

Figure 8.10. Experimental evidence of "drain current collapse" in OFETs, and negative

differential conductance ( 0g d < ) in OFETs. Measurements in (a) darkness, and (b) under

illumination. (c) Drain current intensity as a function of the excitation photon energy. The

photodetachment is more efficient with photon energy of about 2.40 eV.

8.18

B. A. Paez-Sierra, List of tables 10.7

List of tables Table 2.1. Parameters of the electron-phonon coupling in pentacene-based FETs. 2.5

Table 4.1. Skin depth of smooth metallic films, apparent penetration depth of 488 nm light in In,

Ag and Mg films grown on DiMe-PTCDI and PTCDA layers

4.9

Table 5.1. Parameters reported for the Fermi-Dirac integral [Blak1987] 5.5

Table 5.2. Lattice constants of the triclinic cell for pentacene with cell parameters °≈α 978.76 ,

°≈β 136.88 , °=γ 415.84 , and density of molecules per 2cm .

5.7

Table 5.3. Lattice constants of the triclinic cell for SML and MML Pentacene, with cell

parameters °≈α 978.76 , °≈β 136.88 , °=γ 415.84 . The last column indicates the density of

pentacene molecules per 2cm .

5.7

Table 6.1. Raman shift of monolayer and bulk pentacene thin films 6.9

Table 8.1. Time-dependent phenomena in OFETs determined by complementary experimental

techniques (based on the setups used in this research).

8.21

B. A. Paez-Sierra, Erklärung 10.8

Erklärung

Ich erkläre, dass ich die vorliegende Arbeit selbständig und nur unter Verwendung der

angegebenen Literatur und Hilfsmittel angefertigt habe.

06. March 2007

MSc. Phys. Beynor Antonio Paez-Sierra

B. A. Paez-Sierra, Curriculum Vitae 10.9

Curriculum Vitae

Beynor Antonio Paez-Sierra Date of birth Place of birth Nationality Gender Marital status Languages

24 September 1971 Florián - Santander, Colombia Colombian Male Married Spanish (native), German, English, Russian, French

EDUCATION AND AWARDS

1984 – 1989 High School: Colegio Vecinal Policarpa Salavarrieta Suba, Bogota, Colombia 1990 - 1996

B. Sc. in Physics Thesis work awarded with Merit Mention Thesis title: Thermoelectric power in SnO2 thin films Universidad Nacional de Colombia, Bogotá Advisor: Prof. Gerardo Gordillo

1997 - 1999 Msc. in Physics Thesis title: Transport properties of CdTe thin films Universidad Nacional de Colombia, Bogotá Advisor: Prof. Dr. Gerardo Gordillo-Guzmán

2001 14. May, Award for my teaching labor, Science Faculty, Pontificia Universidad Javeriana, Bogotá – Colombia

2002 - 2005 Phd in Physics Thesis title: ,,Raman Spectroscopy of Metal/Organic/Inorganic Heterostructures and Pentacene-Based Organic Field Effect Transistors” Semiconductors department, Technical University of Chemnitz, Germany Prof. Dr. Dr. h.c. Dietrich R. T. Zahn

WORK AND RESEARCH EXPERIENCE

1997 - 2002 As a physics teacher and as an assistantship researcher, physics department, Pontificia Universidad Javeriana. Bogotá – Colombia

1994 - 1999 Characterization of thin films for solar cells (SnO2, CdTe, ZnO) through thermoelectric power, Hall effect and electrical conductivity measurements to study transport properties in semiconductors, (Prof. Dr. Rer Nat. Gerardo Gordillo Guzmán), Universidad Nacional de Colombia

1998 19. January - 3. April, stay research. Work : Thermoelectric power measurements in quasicrystals, (Prof. Dr. Roberto Escudero Derat). Universidad Nacional Autónoma de México, UNAM, México D.F. - México

B. A. Paez-Sierra, Curriculum Vitae 10.10 1998 - 1999 Design and construction of an equipment to measure electrical conductivity.

Universidad Javeriana, Bogotá – Colombia

2000 - 2001 Design and construction of an equipment to measure thermoelectric power in a wide range of temperature. Pontificia Universidad Javeriana, Bogotá – Colombia

At the semiconductors department Technical University of Chemnitz, Germany

20.05.2002 05.11.2005

PhD student. Experimental and theoretical investigations of: • Metal/organic interface formation by vibrational spectroscopies (Raman and

infrared) • Combined vibrational and electrical techniques to characterize organic field

effect transistors This research was supported by the Deutsche Forschungsgemeinschaft (DFG) within the project Za 146/4-2 as part of SPP 1121: Organic field effect transistors: Structural and dynamical properties

2002 Investigation of GaN structures by Raman spectroscopy and surface enhancement Raman spectroscopy. In cooperation with Dr. Elena Konenkova, St. Petersburg, Russia

2003 Summer semester. Auger spectroscopy Practical training for the students in Physics VI semester

2002 - 2003 Co-orientation of a diploma work (Matthias Bartzsch)

2003 - 2004 Winter semester. Micro Raman spectroscopy. Practical training for the students in Physics VI semester

2003 Summer semester. Advanced laboratory practice: Photoluminescence in Organic materials. Practical training for the students in Physics VIII semester

2004 Raman Spectroscopy. Practical training for the PhD students of the ACCUMOL program

2004 Summer semester. Auger spectroscopy. Practical training for the students in Physics VI semester

2004 September. Experimental and theoretical spectroscopic investigations of the interfaces Si/Au/Tetracene/Au and Si/Tetracene/Au. Dipl. Phys. Aline Hepp Darmstadt. Cooperation with the group of Prof.Dr. Heinz von Seggern, Fachbereich Material- und Geowissenschaften, Technische Universität Darmstadt

2004 - 2005

Winter semester. Micro Raman spectroscopy. Practical training for the students in Physics VI semester

2005 January. Investigation of GaN nanotubes by Raman spectroscopy and surface enhancement Raman spectroscopy. In cooperation with Dr. Alexander Milekhin, Noborsibirsk – Russia

2005 Micro Raman spectroscopy. Practical training for the students in Physics VI semester

B. A. Paez-Sierra, Curriculum Vitae 10.11

SEMINARS AND CONGRESSES1

1994 August. First Workshop in Semiconductor Physics, Universidad Nacional de Bogotá Colombia, Bogotá

1995 August. Seminar “Photovoltaic Systems”, Universidad Nacional de Colombia, Bogotá

1995 July. XVI National Congress of Physics, Santiago de Cali - Colombia

1997 June. XVII National Congress of Physics, Medellín - Colombia

1998 11.-16. January. XIV Latin American Congress of Solid State Physics, Oaxaca - México

1998 21.-25. September. Third School of Condensed Matter Physics, Armenia - Colombia

1999 June. XVIII National Congress of Physics, Bogotá, D.C - Colombia

1999 1.-5. November. XV Latin American Symposium on Solid State Physics “SLAFES - XV”, Cartagena de Indias – Colombia

2001 3.-6. July. X CLACSA Latinamerican Congress of Surface Science and their applications, San José - Costa Rica

2001 24.-28. September. XIX National Congress of Physics, Manizales – Colombia,

2003 16.-18. January, XI International Workshop on Computational Physics and Material Science: Total Energy and Force Methods, Trieste – Italy,

2003

4.-22 February , Winter College on "Numerical Methods in Electronic Structure Theory" Trieste - Italy

2003 24.-28. March, Deutsche Physikalische Gesellschaft Tagung, Dresden - Germany

2003 1.-3. April. VIII Holzhau group Meeting, Carlovi Bary - Czech Republic

2003 3.-8. August. Organic Field Effect Transistors II, San Diego - USA

2003 1.-5. September. XX National Congress of Physics, Cali - Colombia

2003 15.-19. September, 9th International Conference on the Formation of Semiconductor Interfaces (ICFSI-9) Madrid - Spain

2003 14.-16. September. International Workshop on Semiconductor Surface Passivation, SSP, Ustrón - Poland

2003 1.-2. October. Kolloquium des DFG Schwerpunktprogramms Organische

1 Where I directly participated or where part of my research activity was involved

B. A. Paez-Sierra, Curriculum Vitae 10.12

Feldeffekt-Transistoren: strukturelle und dynamische Eigenschaften, Bremen 2003 2.-7. November. AVS 50th International Symposium, Baltimore - USA

2004 18.-22. January. 31st conference on the physics and chemistry of semiconductor

interfaces, Kailua-Kona - Hawaii

2004 8.-12. March. Deutsche Physikalische Gesellschaft Tagung, Regensburg - Germany

2004 22.-25. March. IX Holzhau group Meeting, Breslau – Poland

2004 19.-22. May. Bad Honnef - Germany

2004 7.-8. July. Kolloquium des DFG Schwerpunktprogramms Organische Feldeffekt-Transistoren: strukturelle und dynamische Eigenschaften in Kaiserslautern - Deutschland

2004 9-13 August. 14th "International Conference on Crystal Growth" 12th "International Conference on Vapor Growth and Epitaxy", Alpes Congrès, Grenoble – France

2004

16.-17. December. Kolloquium des DFG Schwerpunktprogramms Organische Feldeffekt-Transistoren:

Berichtskolloquiums in Braunschweig - Deutschland

2005 6.-9. March, Materials for Advanced Metallization (MAM), Dresden - Germany

2005 04.-9. March, Deutsche Physikalische Gesellschaft Tagung, , Berlin - Germany

2005 17. May. Semiconductors Department, Technical University of Chemnitz

2005 26.29. June. 13 Tagung Festkörperanalytik, Technical University of Chemnitz

2005 3.-8, July. 10th International Conference on the Formation of Semiconductor Interfaces (ICFSI-10) Aix-en-Provence, France

2005 31. July -4. August. Optics & Photonics, San Diego, California USA

B. A. Paez, List of publications 10.13

List of publications

2006 [39] L. C. Jimenez B., H. A. Méndez P., B. A. Paez S., M. E. Ramírez, and H. Rodríguez H, ,, Production and Characterization of Indium Oxide and Indium Nitride”, Br. J. Phys., 36, 1017-1020 (2006).

[38] A.G. Milekhin, R. Meijers,T. Richter, R. Calarco, H. Lüth, B. A. Paez Sierra, and D.R.T. Zahn, “Surface enhanced Raman scattering by GaN nanocolumns”, phys. stat. sol. (c) 3, 2065-2068 (2006).

[37] A.G. Milekhin, R. Meijers,T. Richter, R. Calarco, S. Montanari, H. Lüth, B. A. Paez Sierra, and D.R.T. Zahn, “Surface enhanced Raman scattering by GaN nanostructures obtained by bottom-up and top-down approach”, J.Phys.Cond.Matt. 18, 5825-5834 (2006).

[36] I. Thurzo, B. Paez, H. Méndez, R. Scholz, and D. R. T. Zahn, ,,Anomalous charge relaxation in channels of Pentacene-based organic field-effect transistors: a charge transient spectroscopy study”, phys. stat. sol. (a) 204, 1-15 (2006).

2005 [35] B. A. Paez S, I. Thurzo, G. Salvan, R. Scholz, Dietrich R. T. Zahn, and H. von Seggern, “Combined Raman spectroscopic and electrical characterization of the conductive channel in pentacene based OFETs”, Proc. of SPIE 5940, 59400F 1-9 (2005).

[34] R. Scholz, A.-D. Müller, F. Müller, I. Thurzo, B. A. Paez, L. Mancera, D. R. T. Zahn, “Comparison between the charge carrier mobilities in pentacene OFET structures as obtained from electrical characterization and potentiometry”, Proc. of SPIE 5940, 59400I (2005).

[33] B. A. Paez, G. Salvan, R. Scholz and D. R.T. Zahn, “Interface formation of Mg with DiMePTCDIsStudied by Raman spectroscopy”, phys. stat. sol. (c) 2, 4048-4052 (2005).

[32] G. Salvan, B. A. Paez¸ S. Silaghi, and D. R. T. Zahn, “Deposition of silver, indium, and magnesium onto organic semiconductor layers: Reactivity, indiffusion, and metal morphology”, Microelectronic Engineering, 82, 228-235 (2005).

[31] Salvan, S. Silaghi, B. Paez, G. Baumann, T.U. Kampen, R. Scholz, and D. R. T. Zahn, “Structural and morphological properties of N,N0-dimethyl-3,4,9,10-terylenetetracarboxylic diimide films on passivated GaAs(1 0 0) substrates”, Journal of Crystal Growth 275, e1155–e1162 (2005).

[30] D.R.T. Zahn, G. Salvan, G. Gavrila, and B. A. Paez, “Chemistry and Morphological Properties of Metal Interfaces to Organic Semiconductors”, Solid State Phys. 45, 299–310 (2005).

2004 [29] G. Salvan, S. Silaghi, B. Paez, T.U. Kampen, D. R. T. Zahn, “Structural and Morphological Properties of 3,4,9,10-PeryleneTetraCarboxylic Dianhydride Films on Passivated GaAs(100) Substrates”, ICSM-2004 special issue of Synthetic Metals.

[28] B. A. Paez, G. Salvan, S. Silaghi, R. Scholz, T. U. Kampen and D. R. T. Zahn, “Raman Monitoring of In and Ag Growth on PTCDA and DiMe-PTCDI Thin Films”, Appl. Surf. Sci. Vol 234/1-4 pp 168-172 (2004).

[27] D.R.T. Zahn, G. Salvan, B. A. Paez, R. Scholz, “Interaction between metals and organic semiconductors studied by Raman spectroscopy”, J. Vac. Sci. Technol. A 22, 1482-1487 (2004).

[26] G. Salvan, D.R.T. Zahn, and B. Paez, “Surface enhanced Raman scattering in organic thin films covered with silver, indium and magnesium”, Journal of luminescence 110, 296 (2004).

[25] V. N. Bessolov, E. V. Konenkova, Yu. V. Zhilyaev, B. A. Paez Sierra, D. R. T. Zahn, “The effect of Ag-coated of surface on Raman spectra of GaN nanocrystals”, Appl. Surf. Sci. 235, 274–278 (2004).

B. A. Paez, List of publications 10.14 [24] G. Salvan, S. Silaghi, B. Paez, T.U. Kampen, D. R. T. Zahn, “Modification of

GaAs(100) Surfaces Upon Adsorption of Perylene Derivatives”, Appl. Surf. Sci. 234, 178-184 (2004).

2003 [23] B. A. Paez, N. C. Forero, and L. Castañeda, Electron-positron plasma under an external electromagnetic field, Rev. Col. Fís. 35-2, 222 (2003).

[22] B. A. Paez, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Interaction of metals with perylene derivatives as a model system for contact formation in OFET structures”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 210-217 (2003).

[21] B. A. Paez, M. Bartzsch, G. Salvan, R. Scholz, T. U. Kampen, and D. R. T. Zahn, “Combined Electrical and Raman characterization of C60 based organic field effect transistors”, Proc. SPIE Int. Soc. Opt. Eng. 5217, 63 (2003).

2002 [20] L.C. Jiménez, H. Méndez, B. A. Paez and I. Morales, “Thermoelectric power measurements in indium oxide thin Films”, Rev. Col. Fís. 34 (2), 434 (2002) .

[19] B. A. Paez and J.I. Barrera, “Resonance and phase curves in oscillators forced by multi-harmonic signals”, Rev. Col. Fís. 34 (1), (2002) (in spanish).

[18] J.I. Barrera and B. A. Paez, “Impedance determination of a damped harmonic oscillator”, Rev. Col. Fís. 34 (1), (2002) (in spanish).

[17] B. A. Paez, H. Méndez and L.C. Jimenez, “Design and construction of an equipment to measure the thermoelectric power, α, in bulk and thin films materials”, Rev. Col. Fís. 33 (2), 89 (2002).

[16] B. A. Paez, Nanotech, “Current state and perspectives”, Quantum 2 (2002) (in Spanish).

[15] B. A. Paez, A. Moreno and H. Méndez, “Solution of the Quantum Boltzmann Equation in Linear Transport”, Surf. Rev. Lett. 9, 1761 (2002).

[14] B. A. Paez, H. Méndez and J.C. Giraldo, “Thermoelectric Power Coefficient, α, in a Quantum Well”, Surf. Rev. Lett. 9, 1765 (2002).

2001 [13] B. A. Paez, A. Moreno and H. Méndez, “Dirac delta function and step function, a study with oscillators”, Rev. Col. Fís. 33 (2), 89 (2001) (in spanish).

2000 [12] B. A. Paez, “Thermoelectric power and Hall effect measurements in polycrystalline CdTe thin films”, phys. stat. sol. (b), 220 (1), 233 (2000).

[11] B. A. Paez, J.M Flórez, C.E. Jácome and G. Gordillo, “Thermoelectric characterization of CdTe thin films”, Rev. Col. Fís. 32 (1), 29 (2000) (in spanish).

[10] L.C. Jiménez, H. Méndez and B. A. Paez, “Tin oxide thin films deposited by reactive sputtering”, Rev. Col. Fís. 32 (1), 63 (2000) (in Spanish).

[9] H. Méndez, H. Rodriguez and B. A . Paez, “Design and construction of an equipment for electrical characterization”, Rev. Col. Fís. 31 (2), 353 (1999) (in Spanish).

1999 [8] B. A. Paez and H. Méndez, “A pedagogical view of the Boltzmann transport equation”, Rev. Col. Fís. 31 (2), 345 (1999) (in Spanish).

[7] C.E. Jácome, J.M. Flórez, and Y.G. Gurevich, B. A. Paez and G. Gordillo, “Study of transport properties in semiconducting CdS thin films”, Rev. Col. Fís. 31 (2) 243 (1999). (in Spanish).

[6] G. Gordillo, B. A. Paez, C.E. Jácome and J.M. Flórez, “Thermoelectric power in SnO2 thin films”, Thin Solid Films 342, 160-166 (1999).

1998 [5] B. A. Paez, C.E. Jácome, J.M. Flórez and G. Gordillo, “Thermoelectric power measurements in CdTe thin films”, Rev. Mex. Fís.44 S3, 74 (1998) (in Spanish)

[4] B. A. Paez, C.E. Jácome, J.M. Flórez and G. Gordillo, “Thermoelectric power in SnO2 thin films”, Rev. Col. Fís. 30 (1), 97 (1998). (in Spanish).

1997 [3] G. Gordillo, B. A. Paez, C.E. Jácome, L.C. Hernández, J.M. Florez, and H. Méndez, “Characterization of SnO2 thin films through thermoelectric power measurements”, Conference Record Of The 26th IEEE Photovoltaic Specialists Conference, Anaheim CA. PSC, 26, 519 (1997).

[2] C.E. Jácome, B. A. Paez, J.M. Flórez and G. Gordillo, “Theoretical investigation of the thermoelectric power in SnO2 thin films. Rev. Col. Fís. 29 (2) 115 (1997)

B. A. Paez, List of publications 10.15

(in Spanish). 1995 [1] A. Ortiz, B. Paez, F. Fajardo, J. Quiñones and G. Gordillo, Study of

optoelectronic properties in ZnO thin films deposited by reactive and DC sputtering. Revista Colombiana de Física (Rev. Col. Fís). 27(1) 359 (1995) (in Spanish).

ESSAYS AND BOOKS

[1] Beynor A. Paez S., “Phasorial Algebra” Physics Department, Pontificia

Universidad Javeriana; Bogotá, D.C, August 1999 (http://newton.javeriana.edu.co) (essay in Spanish)

[2] Beynor A. Paez S., “Physics V.1, Mechanics”, Physics Department, Pontificia Universidad Javeriana, (2001) (Book in Spanish) http://unicornio.javeriana.edu.co/

B. A. Paez-Sierra, Acknowledgements 10.16

Acknowledgements

To the invaluable and tremendous moral and scientific support of my wife “Viktoriia”, for being along with me during those almost endless experiments, and actively joining this scientific venture despite she was running her own magneto-optical experiments. Special thanks to the Deutsche Forschungsgemeinschaft (DFG) in the frame of the “Schwerpunktprogramm 1121: Organische Feldeffekt-Transistoren: strukturelle und dynamische Eigenschaften”, for the grant which has financially supported me to conduct this research in TU Chemnitz at the group of Prof. Dr. Dr. h.c. Dietrich R. T. Zahn. Several cooperations grew up during the evolution of this work. In particular I want to mention the cooperation with Christoph Pannemann at the group of Prof. Dr.-Ing. Ulrich Hilleringmann, who provided me the single channel structures; with Aline Hepp at the group of Prof. Dr. H. von Seggern, from them we got the interdigitated structures. These interchanges seeded mutual benefits ranging from substrates, organic molecules, DFT calculations, Raman spectroscopy, QTS experiments; and to what is important, the collective work to improve our comprehension about nature. Other substrates for transistors were processed at the center for microtechnologies in cooperation with Dr. Christian Kaufmann in the group of Prof. Dr. Dr. Prof. h.c. mult. T. Gessner. Thanks to Matthias Bartzsch at the pmTUC in the group of Prof. Dr. -Ing. A. C. Hübler, who provided me some flexible substrates for plastic electronic applications (polymer-based transistors), where we run some experiments with small organic molecules. Many other cooperations different from the OFET project streamed during my research. In particular those about GaN nanocolums with Alexander Milekhin (Novosibirsk), in GaN nanocrystals with Elena V. Konenkova (St. Petersburg), on quantum dots of CdSe/ZnS in cooperation with Prof. Dr. C. von Borczyskowski (TU Chemnitz). Specially those experiments and the involved scientists help me to understand much better the nanoscience and nanotechnology that I was doing with the SERS effect in organic molecules. Although no one of my colleagues at the semiconductors group (may 2002-Nov. 2005) was directly involved in the OFET project, just let me say that I got any message from all of you, I learned something to take with. Specially, I want to thank Georgeta Salvan for her orientation on some SERS experiments in PTCDA and DiMe-PTCDI. I also want to thank Simona Silaghi for her patience when I asked her to give me more time to use the Raman lab during her experiments. Concerning the theoretical part, I want to thank the people at the “International Center for Theoretical Physics (ICTP), Trieste” in particular Prof. Dr. S. Baroni whom allow me to join the Institue (16-January. 4-February 2003) to learn about “Numerical Methods in Electronic Structure Theory”. After this wondeful course, I appreciate the openning of an account at the super computer by Reinhard Scholz in the group of Prof. M. Schreiber. I perfromed there several DFT calculations that helped me to understand much better the influence of electric fileds on vibrational properties of organic molecules, and the relation of dipole-bound states on the QTS signal. I want to express my gratitude to Ilja Thurzo with whom we came to the conclusion on the “anomalous QTS in organic field effect transistors” that we condensed in a very nice publication. I am also indebted with Ilja, and my wife, who independetly gave me very nice highlights to improve this thesis. To my friends Henry and Luis, with who not only interesting experiments were realized, specially those about synthesizing Mg with PTCDA, but also for the mutual support and help to spread the hispanic culture at the TU Chemnitz. Special gratitude to the members of the spanish club (http://www.tu-chemnitz.de/stud/club/hispano/), who kindly cooperate for the sucessfull meetings at the “Club der Kulturen (CdK)” Finally, the last but not least to my marvelous family in Colombia: Gotita, Segundiales, Peluca, and co.