Optical Waveguides of Conjugated Polymers for All-Optical … · 2005-12-28 · Optical Waveguides...
166
Optical Waveguides of Conjugated Polymers for All-Optical Switching Devices Dissertation zur Erlangung des Grades “Doktor der Naturwissenschaften” am Fachbereich Physik der Johannes Gutenberg-Universität in Mainz vorgelegt von Ayi Bahtiar geboren in Indonesien Mainz 2004
-
Upload
truongkhuong -
Category
Documents
-
view
226 -
download
1
Transcript of Optical Waveguides of Conjugated Polymers for All-Optical … · 2005-12-28 · Optical Waveguides...
Optical Waveguides of Conjugated Polymers for
All-Optical Switching Devices
Dissertation zur Erlangung des Grades
“Doktor der Naturwissenschaften”
am Fachbereich Physik
der Johannes Gutenberg-Universität
in Mainz
vorgelegt von
Ayi Bahtiar
geboren in Indonesien
Mainz 2004
Tag der mündlichen Prüfung: 06.07.2004
Dekan: Prof. Dr. K. Binder
1. Berichterstatter: Prof. Dr. C. Bubeck
2. Berichterstatter: Prof. Dr. E. W. Otten
Die vorliegende Arbeit wurde im Zeitraum August 2000 bis April 2004 am Max-Planck
Institut für Polymerforschung in Mainz unter Betreuung von Herrn Prof. Dr. C. Bubeck
und Mitbetreuung von Herrn Prof. Dr. E. W. Otten angefertigt.
For my family
Content
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Scientific and Technological Background . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Task of the Work . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .
5
2 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1 Planar Waveguide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Transverse Electric Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.2 Transverse Magnetic Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.3 Losses in Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Nonlinear Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
3 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1 Spin Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2 Thickness Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 UV-Vis-NIR Transmission and Reflection Spectroscopy . . . . . . . . . . . . 21
3.4 Photostability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5 FTIR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Third-Harmonic Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6.1 Experimental Setup and Measurement Procedure . . . . . . . . . . . . . . 26
3.6.2 Evaluation of Third-Order Nonlinear Optical Susceptibility . . . . . 28
3.7 Attenuation Loss of Slab Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.8 Intensity Dependent Prism Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.8.1 Procedure of Prism Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.8.2 Numerical Model of the Prism-Film Coupler . . . . . . . . . . . . . . . . . . 32
3.8.3 Determination of Air-Gap Thickness . . . . . . . . . . . . . . . . . . . . . . . . 34
3.8.4 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 37
3.9 Fluorescence Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.9.1 Steady-State Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.9.2 Multiphoton Excited Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . 38
4 Materials Properties of PPV Derivatives . . . . . . . . . . . . . .. . . . . . . . . 40
4.1 Materials and Film Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.2 Linear Optical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.3 Waveguide Propagation Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4 Third Harmonic Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.5 Stability Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.5.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.5.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.6 Materials Comparison and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . .
60
5 Nonlinear Refractive Index and Multiphoton Absorption of MEH-PPV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.1 Nonlinear Prism Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.2 Multiphoton Excited Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.3.1 Control Experiments of Prism Coupling Data . . . . . . . . . . . . . . . . . 73
5.3.2 Interpretation of Third-Order Nonlinear Optical Spectra . . . . . . . . 73
5.3.3 Three-Photon Absorption Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.3.4 Figures of Merit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
6 Influence of Molecular Weight on the Properties of MEH-PPV 82
6.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.2 Spin Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.2.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.3 Linear Optical Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.3.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.3.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.4 FTIR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.4.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.5 Third-Harmonic Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.5.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.5.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.6 Waveguide Propagation Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.6.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.6.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.7 Photostability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.7.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.7.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
104
7 Microstructuring of MEH-PPV Waveguides: Towards All-Optical Switching Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.1 Photoablation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.1.1 Damage Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.1.2 Microstructuring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.1.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
7.2 Hot Embossing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
7.2.1 Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3 Solvent-Assisted Microcontact Molding . . . . . . . . . . . . . . . . . . . . . . . . . . 113
7.3.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
7.4 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
115
8 General Consequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
10 Zusammenfassung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 120
11 References . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Appendix A: Studies of Oligomer OPEs and conjugated polymer PPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Appendix B: Full Chemical Names of PPV Derivatives . . . . . . . . . . . 136
Appendix C: THG Data of PPV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Appendix D: Properties of MEH-PPVs . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Appendix E: Studies of Cyano-Ether-PPV . . . . . . . . . . . . . . . . . . . . . . . . 146
Appendix F: Studies of Polyfluorene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Appendix G: Studies of NRS (Super-Yellow) Material . . . . . . . . . . . 153
List of Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
_____________________________________________________________________1
1 Introduction
1.1 Scientific and Technological Background
In the telecommunication networks, the data are transmitted by using optical
fibers. Optical fibers provide the data transmission over a long distance with an enormous
bandwidth and high speed. However, processing and switching of the data are still carried
out by electronic circuits. Electronics/Optics (E/O) and Optics/Electronics (O/E)
conversions are needed, and the speed of the entire system is limited by the speed of the
electronic circuits. Therefore, it is necessary to develop integrated all-optical switching
devices in order to increase the speed of data processing. Various concepts for integrated
all-optical switching devices have been studied [Stegeman’93, Stegeman’97]. One of the
simplest examples of integrated all-optical switching devices is a nonlinear Bragg grating
that induces a photonic bandgap in planar waveguides [Scalora’94] as depicted in Fig.
1.1.
Fig. 1.1: Schematic view of a nonlinear optical Bragg grating waveguide (top) and optical response of the device (bottom). At high intensities the transmission minimum is shifted to longer wavelengths due to the intensity dependent refractive index. The dashed vertical line indicates a wavelength where the switching between states of low and high transmission occurs.
Substrate
Polymer waveguide
Bragg reflector
Substrate
Polymer waveguide
Substrate
Polymer waveguide
Bragg reflector
λ0
Wavelength
Tran
smis
sion
lowintensity
highintensity
0
1
_____________________________________________________________________2
In this structure, the Bragg grating acts as a perfect reflector for laser pulses
whose carrier wavelength is at or near the Bragg wavelength λ0 = 2 neff Λ, where neff
denotes the effective refractive index of the waveguide and Λ the grating period.
However, if the refractive index of the waveguiding material is no longer constant but
intensity dependent, light with sufficient intensity can shift the position of the Bragg
wavelength. Therefore, at the wavelength marked with the dashed vertical line in the
lower panel of Fig 1.1, an all-optical switching from low to high transmission will occur
with increasing incident light intensity. The realization of an optical switch as shown
schematically in Fig. 1.1 requires suitable materials which must be multifunctional.
The materials must be suitable for thin film processing and waveguide fabrication.
Optical waveguides are advantageous for nonlinear optical devices because they can be
easily integrated and miniaturized and offer large electric field intensity due to light
confinement in the waveguides [Stegeman’89, Kajzar’96]. An optical waveguide consists
of a region with a refractive index larger than that of the surrounding media. The light
could be coupled into waveguides to excite the modes by using a prism coupler or other
means [Tien’77]. Waveguides with attenuation propagation losses αgw < 1 dB/cm are
required in order to realize devices with high throughput [Stegeman’89, Bubeck’00].
Furthermore, the materials should be suitable for microstructuring. The periodic
structure or grating in the waveguide shown in Fig. 1.1 is considered as the simplest form
of a photonic crystal. Photonic crystals are formed by periodic modulation of refractive
index and posses a photonic bandgap [Joannopoulos’95], in which light in a certain
frequency range can not propagate through the structure. In recent years, photonic crystals
have found much interest for integrated optical waveguides because of their possibility to
guide light with large efficiency along a straight path [Joannopoulos’95] or along a sharp
corner [Mekis’96, Lin’98].
Photostability of the materials is always of a major concern. The material for
nonlinear waveguide applications must be relatively stable at UV-Vis wavelengths and
additionally at high intensity laser pulses. The UV-Vis light and sample handling at
ambient air may cause photo-oxidation which lead to the decrease of the optical constants
of materials [Rothberg’96, Bader’02]. For nonlinear optical waveguides, materials with
high optical damage threshold at near infrared (NIR) laser wavelengths are needed
because high intensities in the waveguides are required to achieve sufficiently large
_____________________________________________________________________3
changes of the intensity dependent refractive index. A damage threshold intensity Idt >
100 MW/cm2 is usually needed [Bubeck’00].
Materials with cubic optical nonlinearities allow switching of light by light by
means of the intensity dependence of refractive index n(I) or absorption coefficient α(I)
which are described by
Inn)I(n 20 += (1.1)
I)I( 20 α+α=α (1.2)
where n0 and α0 are linear refractive index and absorption coefficient, respectively. The
nonlinear refractive index n2 and nonlinear absorption coefficient α2 are proportional to
the real and imaginary part of the complex of third-order nonlinear optical susceptibility
χ(3) (−ω;ω,−ω,ω), respectively [Butcher’90, Boyd’92]. A large n2 is required to achieve
fully reversible and ultrafast changes of the refractive index at intensities I < Idt.
Furthermore, nonlinear propagation losses due to two-photon absorption must be
negligibly small which is expressed by the requirement of a small value of α2.
The basic material requirements for all-optical waveguide switching have been
formulated in terms of figures of merit (FOMs) [Stegeman’93]
( ) ,1InW0
2 ≥λα
=λ (1.3)
( ) 1n
2T2
2 ≤λα
=λ (1.4)
at the working wavelength λ of the device. Although many nonlinear optical materials
have been studied, quantitave data of W(λ) and T(λ) are still rare because of many
experimental difficulties in measuring the dispersions of all required linear and nonlinear
optical coefficients. Therefore, published data of FOMs often contain extrapolations and
estimates rather than data actually measured at the wavelength λ.
The demonstration of all-optical switching in planar waveguides would become
the breakthrough in integrated nonlinear optics. The bottleneck for the realization of such
devices is still the problem of identifying the material that satisfies all above
requirements. In general, three different classes of materials are of potential interest, i.e.
special glasses, semiconductors, and organic materials. Fused silica fibers were among
_____________________________________________________________________4
the best reported materials with appropriate figures of merit. However, long interaction
lengths in the order of tens of centimeters to meters are required due to the small
nonlinearity of glass (n2 ~ 10-16 cm2/W), making an integrated optics design impossible
which limits the application possibilities of glasses. In semiconductors, a large nonlinear
refractive index induced by absorption is obtained just below the energy bandgap
[Stegeman’89]. However, in this range the absorption is also large which causes a small
value of W. Furthermore, the recovery time of the excited electrons which cause the large
nonlinearity is low, typically in the order of nanoseconds (10-9 seconds). This would limit
the repetition rate of the switching. By tuning the wavelength far from the band gap, the
value of W could be increased, but the two-photon absorption begins to dominate
resulting in a large value of T [Stegeman’89, Stegeman’93].
Organic materials are advantageous as compared with inorganic materials because
of their relatively low cost and tailorability which allows to tune the chemical structures
and materials properties for the envisioned applications [Brédas’94]. Conjugated
polymers that posses a delocalized π-electron system have been considered to be the most
promising organic material candidates for all-optical switching applications because of
their high cubic nonlinearity and fast response times in the order of picoseconds (10-12
seconds) or less, and relative ease of waveguide preparation [Messier’89, Kajzar’96,
Stegeman’97]. In particular, poly(p-phenylenevinylene) (PPV) (see its chemical structure
in Fig. 1.2) was identified as a promising material for nonlinear optical applications
because of large cubic nonlinearities with fast response times and high damage thresholds
[Bubeck’91, Samoc’95, Mathy’96, Uberhofen’99]. However, the waveguide propagation
losses of PPV is relative high due to its well-known polycrystalline morphology
[Bradley’87, Ueberhofen’99]. Here, we have focussed on newly synthesized and solution
processable PPV derivatives, because they have incurred considerable interest due to their
good combination of large nonlinearity and superior waveguide properties [Bartuch’92,
Gabler’97, Gabler’98, Bubeck’00, Koynov’02, Bader’02, Fitrilawati’02].
nn
Fig. 1.2: Chemical structure of poly(p-phenylenevinylene) (PPV)
_____________________________________________________________________5
1.2 Task of the Work
The general aim of this work is to find a suitable material that fulfills the
application requirements for all-optical switching and to understand the relationships
between the molecular structure and the optical properties of conjugated polymers that
yield guidelines for future materials development. The following tasks had to be
performed to achieve these goals.
Thin films and planar waveguides of several newly synthesized derivatives of
poly(p-phenylenevinylene) were prepared by use of the spin coating technique. Spin
coating is a well known and suitable method for preparing polymeric thin films and
planar waveguides from the solutions [Washo’77]. Thin films with good optical quality
and optical waveguides with propagation losses < 1 dB/cm are required. The waveguide
propagation losses depend mainly on intrinsic absorption and scattering losses in the bulk
and at the interfaces of the films. The surface scattering is related to the surface roughness
which can be minimized by optimization of the spin coating parameters [Fitrilawati’02].
Therefore, numerous spin coating experiments of several conjugated polymers were
performed to find the optimum condition for preparing homogeneous films with
minimum surface roughness.
The knowledge of optical constants like refractive index and absorption
coefficient is crucial for designing any optical device. Here, we have used transmission
and reflection spectroscopy [Mathy’96, Penzkofer’98] to measure the dispersion of
optical constants in thin films. In addition, the prism coupler technique [Tien’77] was
employed to measure the refractive index of planar waveguides.
Large cubic nonlinearity of the materials is very important for all-optical
switching applications. As a screening test to estimate the purely electronic contribution
to the cubic nonlinearity χ(3) of the PPV derivatives, third-harmonic generation at variable
wavelengths using the Maker fringe technique [Kajzar’85, Neher’89] was used. Third-
harmonic generation is a very attractive method to characterize newly developed
materials because of its large detection sensitivity of χ(3).
The photooxidation of conjugated polymers at ambient air is always a severe
problem. Appropriate sealing and encapsulation of the samples from UV-Vis light could
solve this problem. We have studied the photostability of the polymers by irradiating the
thin films with UV light and measuring the absorption changes after certain exposure
times.
_____________________________________________________________________6
Materials for nonlinear waveguide applications must have a good combination of
large third-order optical nonlinearity and low waveguide propagation losses and relatively
high photostability at UV irradiation. Therefore, we had to compare the relevant
properties of several newly synthesized PPV derivatives and select the best-suited
material for all-optical switching applications. These comparative studies are described in
Chapter 4. The following Chapters 5 - 7 will describe an in-depth characterization of this
polymer only.
The sensitive and accurate measurement of nonlinear refractive index n2 and
nonlinear absorption coefficient α2 directly in polymer slab waveguides is an important
task. In particular their absolute values and dispersion in the near infrared (NIR) range
need to be known, because in this range most of materials show minima of their
waveguide propagation losses. Here, we have used nonlinear prism coupling
[Ueberhofen’99, Koynov’02] to measure the absolute values and signs of n2 and α2 of the
selected material over a broad spectral interval. In addition, multiphoton excited
fluorescence was performed in order to improve the accuracy of the α2 data from the
prism coupling method and to study the multiphoton absorption resonances and their
possible influence on the nonlinear refractive index spectra. By using both methods, the
dispersions of figures of merit W(λ) and T(λ) were evaluated and the spectral range
which fulfills the figures of merit requirements was identified.
Surprisingly, in the course of this work it turned out that the molecular weight of
the polymer has crucial impact on the morphology and the optical properties of thin films.
Therefore, we present in Chapter 6 a study of the birefringence, optical constants,
waveguide propagation loss and third-harmonic generation of thin films of the best-suited
polymer with a large variation of molecular weight.
Microstructuring is one of the most challenging tasks in realizing integrated all-
optical switching devices. We have applied several techniques for fabrication of sub-
micrometer gratings in slab waveguides of the selected material. The aim is to find the
best method for the preparation of photonic structures in the polymer waveguides. This
study is described in Chapter 7.
_____________________________________________________________________7
2 Theoretical Background
2.1 Planar Waveguide
Optical waveguides are the key element in integrated optics. They consist of a
region with high refractive index surrounded with regions of lower refractive index. They
can be classified as planar, channel and optical fibers. Optical waveguides are also very
interesting for nonlinear optical effects because they provide strong light confinement
over long propagation distances. Basically, the propagation of light in waveguide is
described by Maxwell’s equations. The origin of Maxwell’s equation can be found in the
excellent book of Jackson [Jackson’83]. In this thesis, we only give the basic concepts of
Maxwell’s equation in the waveguide. Maxwell’s equations can be written in the form
[Jackson’83]
0H.tEnHx
0E.tHEx
20
0
=∇∂∂
ε=∇
=∇∂∂
µ−=∇
rr
rrr
rr
rrr
(2.1)
where Ev
and Hr
are the electric and magnetic fields vectors, and n is refractive index
which is a function of position, )r(nr
. The quantity ε0 and µ0 are the permittivity and
permeability of free space, respectively. In this thesis, we do not consider magnetic
materials, therefore, µ = µ0. The form of Maxwell’s equation given in Eq. (2.1) applies
only in source-free region of space, where there is no free current or charge ( 0j =r
and σ
= 0). The Maxwell’s equation (2.1) can be written in the form
0tE
c)r(nE 2
2
2
22 =
∂
∂−∇
rr
(2.2)
If we assume a harmonic wave traveling in positive z direction with propagation constant
β, [ ])zt(iexp)y,x(E)r(E β−ω=rrr
, Eq. (2.2) could be written in the form
_____________________________________________________________________8
[ ] 0)y,x(E)r(nky
)y,x(Ex
)y,x(E 2222
2
2
2=β−+
∂∂
+∂
∂ rrrr
(2.3)
where k = (ω/c) = (2π/λ) and β to be determined from Maxwell’s equations. We will
restrict ourselves to the description of the planar waveguides, by assuming that there is no
variation of refractive index in y direction. This means that 0y/ =∂∂ and )x(n)r(n =r
.
Therefore, the wave equation can be simplified in the form
[ ] 0)x(E)x(nkdx
d 2222 =
β−+
r (2.4)
We consider now a planar dielectric waveguide displayed in Fig. 2.1. It consists of
a film of thickness d and refractive index n2 surrounded by media with refractive indices
n1 and n3. In order to support the guided modes in the waveguide, it is necessary that n2
must be larger than n1 and n3. We will show later that under these conditions, light can be
confined in the waveguide by total internal reflection at the interfaces between the high
and low index materials. We have a symmetric planar waveguide, if n1 = n3, while for n1
≠ n3 the waveguide is asymmetric. In this thesis, we study the asymmetric planar
waveguides only.
Fig. 2.1: A planar dielectric waveguide
Referring to Fig. 2.1, the refractive index profile n(x) is described by
=
,n,n,n
)x(n
3
2
1
dx0xd
0x
−<<<−
> (2.5)
n1
n2
n3
propagation
x = -d
x = 0
x
y
z
d
_____________________________________________________________________9
We assume that 132 nnn >> . The solution of Eq. (2.4) for this refractive index profile
depends on the value of )nk( 22i
2 β− , where i = 1,2,3. Fig. 2.2 shows the illustration of
electric field distribution for different values of propagation constant β at fixed
frequency ω [Yeh'88].
Fig. 2.2: Electric field distribution in a planar waveguide for different values of propagation constant β [Yeh’88].
x = -d x = 0
_____________________________________________________________________10
The electric field distribution E(x) for β > kn2 is exponential at all three regions,
as illustrated in Fig. 2.2(a). It increases exponentially away from the waveguide.
Physically, this solution is not realizable because it does not correspond to a real wave.
For kn3 < β < kn2, the solution is sinusoidal in the waveguide (-d < x < 0), but it
decays exponentially in the surrounding media. It is possible to have solutions E(x) that
satisfy the boundary conditions at the interfaces. One of the solutions is shown in Fig.
2.2(b). The energy carried by these modes is mostly confined in the guiding layer, only a
small fraction is flowing into surrounding media. Therefore, these modes are referred as
confined or guided modes. The confined modes are possible only if n2 > n1, n3. By
applying the boundary conditions at interfaces, the values of allowed β are discrete. The
number of confined modes depends on the thickness of waveguide, frequency or
wavelength, and the refractive indices n1, n2, and n3. At a given wavelength, the number
of confined modes increases with increasing film thickness.
Solutions of Eq. (2.4) for kn1 < β < kn3 correspond to exponential behavior in the
region x > 0 and to sinusoidal in the region x < 0, as illustrated in Fig. 2.2(c). These
modes are referred as substrate radiation modes. Finally, the solution for 0 < β < kn1 is
sinusoidal in all three regions [Fig. 2.2(d)]. These are called as radiation modes of the
waveguides.
We would like to discuss the derivation of the guided modes, which according to
Fig. 2.2(b), have a propagation constant β (kn3 < β < kn2) for n1 < n3. The guided modes
of the planar waveguide can be classified as TE and TM modes. TE or transverse electric
modes do not have a component of electrical field in the direction of wave propagation,
while TM or transverse magnetic modes do not have a longitudinal magnetic field
component. We will consider TE and TM modes separately.
2.1.1 Transverse Electric Modes
TE modes have their electric field perpendicular to the plane of incidence. They
have only three field components: Ey, Hx, and Hz. The field component of Ey can be taken
in the form [Taylor’74, Marcuse’74, Yeh’88]
[ ])zt(iexp)x(E)t,z,x(E my β−ω= (2.6)
_____________________________________________________________________11
y0
x E)t,z,x(Hωµ
β−= (2.7)
xEi)t,z,x(H y
0z ∂
∂
ωµ= (2.8)
By substituting Eq. (2.6) into (2.4), the function Em(x) can be written as
++−
−=
)],dx(pexp[)]hdsin()h/q()hd[cos(C)],hxsin()h/q()hx[cos(C
),qxexp(C)x(Em
dx0xd
0x
−<<<−
> (2.9)
where C is the normalization constant and h, q and p are given by
23
22
222
2
21
22
)kn(p
)kn(h
)kn(q
−β=
β−=
−β=
(2.10)
The boundary conditions require that the tangential E and H fields be continuous at the
interfaces x = 0 and x = -d. It means that Ey and ( )( )x/E/iH y0z ∂∂ωµ= are continuous
at x = 0 and x = -d. By applying these requirements into Eq. (2.9), we obtain
)pqh()qp(h)hdtan( 2 −
+= (2.11)
This equation is used to obtain the eigenvalues β for the confined TE-modes and it can be
solved graphically or numerically.
2.1.2 Transverse Magnetic Modes
For transverse magnetic or TM modes, the magnetic field vector is perpendicular
to the plane of incidence. The derivation of the confined TM modes is similar to that of
TE modes. The nonzero field components Hy, Ex, and Ez are [Taylor’74, Marcuse’74,
Yeh’88]
_____________________________________________________________________12
[ ])zt(iexp)x(H)t,z,x(H my β−ω= (2.12)
y0
2x Hn
)t,z,x(Ei ωε
β= (2.13)
x
H
ni)t,z,x(E y
02i
z ∂
∂
ωε−= (2.14)
The function Hm(x) for all three regions
[ ]
−−+−
++−=
),qxexp(C)q/h()],hxsin()hxcos()q/h([C
)],dx(pexp[)hdsin()hdcos()q/h(C)x(Hm
0x0xd
dx
><<−
−< (2.15)
where C is a normalization constant and h, q, and p are given by Eq. (2.10). The
continuity of Hy and Ez components at the interfaces x = 0 and x = -d, leads to the
eigenvalue equation for confined TM-modes
)qph()qp(h)hdtan( 2 −
+= (2.16)
where
,pnnp
2
3
2
= q
nnq
2
1
2
= (2.17)
Again, similar to that of TE, this eigenvalue equation can only be solved graphically or
numerically.
We can also solve the eigenvalue equation for both TE- and TM-modes by
applying geometrical or ray optics. This is possible because the waveguide consists of
layers of homogeneous dielectric materials. Wave propagation in each region can be
represented by the superposition of two plane waves. One of the plane waves may be
considered as the incident wave, while the other is viewed as reflected one. The total
internal reflection is sufficient to assure the confinement of energy in the guiding layer.
The phenomenon of total internal reflection can be found in literature [Hecht’87,
Möller’88, Yeh’88]. Consider a ray of light A propagates in the film toward the film-
substrate interface (x = -d) with an incident angle θ, as displayed in Fig. 2.3.
_____________________________________________________________________13
Fig. 2.3: Ray trajectory of a guided wave in a thin film waveguide. The coordinate system and wave vector are also given.
If θ is larger than the critical angle sin-1(n3/n2), the ray A will be totally reflected into ray
B at film-substrate interface. The phase of B at this interface is
=Φ −
hptan 1
23 (2.18)
for the TE modes and
=Φ −
hp
nn
tan2
3
2123 (2.19)
for the TM modes. Similarly, the B is reflected into A’ at film-air interface (x = 0). In this
case the phase changes are
=Φ −
hqtan 1
12 (2.20)
for TE modes, and
=Φ −
hq
nn
tan2
1
2112 (2.21)
for TM modes, respectively. Here, 0 < Φij < π/2, the constants p and h are given by Eq.
(2.10).
θ θ
substrate (n3)
film (n2)
air (n1)
z
x
y
A
B A‘ kn2
β
b1θ
_____________________________________________________________________14
We will use the zigzag model to derive the mode equation. As can be seen in Fig.
2.3, the ray A’ follows the ray A after one zigzag path. Because the total reflection at both
film-substrate and film-air interfaces, the amplitude of rays A and A’ differs only by a
phase ∆. After subsequent zigzags, the ray has phase differences 2∆, 3∆ …relative to A.
The superposition of such of rays is zero, except when ∆ = 2mπ with integer m. The total
phase difference between A and A’ is [Tien’70, Tien’77, Kogelnik’79]
π=Φ−Φ− m222db2 12231 m = 0, 1, 2, .... (2.22)
This equation is called as mode equation, where m represents the number of the mode.
Equation (2.22) is valid for both TE and TM modes, but Φij differ. According to the
vector diagram in Fig. 2.3, we have
θ=β
θ=sinkncosknb
2
21 (2.23)
For each allowed mode, the corresponding propagation constant and velocity are given by
β
=
θ=β
m
m2m
kcv
sinkn. (2.24)
Effective refractive index can be defined as
kv
cn meff
β== (2.25)
which is bounded by n3 < neff < n2.
2.1.3 Losses in Waveguides
An important aspect of guided wave propagation in thin films is attenuation of the
intensity of light along the propagation direction. Three different mechanisms can lead to
propagation losses in the waveguide. The first-mechanism is absorption of the
electromagnetic energy by the molecules of the film. It occurs in the UV-Vis region due
_____________________________________________________________________15
to electronic absorption and in the near infrared (NIR) region due to overtones of
molecular vibrations. The second-mechanism is volume scattering which is caused by
imperfections, such as density variations, impurities, and defects, within the volume of
waveguide. These losses depend on the relative size of the imperfections as compared to
the wavelength of the light and number of scattering centers (imperfections). The third-
mechanism is surface scattering loss or reflection loss. Tien [Tien’71] has derived an
expression for scattering due to surface roughness, based on the Rayleigh criterion. The
scattering loss coefficient is defined as [Tien’71]
++
θθ
=α
q1
p1d
1sin
cos2
A
m
m32
s (2.26)
21
)(4A 223
212 σ+σ
λπ
= (2.27)
where d is the thickness of waveguide, θm is the incident angle, p and q are constants
given by Eq. (2.10), σ23 and σ12 are the variances of surface roughness at film-substrate
and film-air interfaces, respectively. We can understand from Eqs. (2.26) and (2.27) that
the loss coefficient is reduced for small roughness at interfaces, thick films, long
wavelengths and small of mode number (large θm). The scattering loss of the fundamental
mode (zero mode) is the smallest, because the electromagnetic fields concentrate mainly
in the center of the waveguide. Theferore, the influence of the roughness at the film-
substrate and film-air interfaces is not so large as compared to higher modes. A more
detailed theory of surface scattering in planar waveguide has been developed by Marcuse
[Marcuse’74].
To describe quantitatively the magnitude of optical loss, the attenuation
coefficient is generally used. In that case, the intensity of light at any point along the
length of the waveguide is given by
)zexp(I)z(I 0 α−= (2.28)
where I0 is the initial intensity. In the waveguides, the total propagation loss, αgw is
frequently expressed by
αgw [dB/cm] = 4.3 α [cm-1] (2.29)
_____________________________________________________________________16
2.2 Nonlinear Optics
Nonlinear optical phenomena can arise if materials are exposed to intense
electromagnetic radiation. Consider a dielectric material in the presence of an electric
field. The electrical field distorts the distribution of charged particles. Therefore, it
induces a polarization. The polarization Pr
responds linearly to the incident electrical
field Er
, if the field strength is low. At high electrical fields, the polarization of a medium
becomes nonlinearly related to the field strength. The material polarization can be
expanded in a Taylor expansion of electric field strength [Butcher’90, Boyd’92]
( )...EEEEEEP )3()2()1(0 +χ+χ+χε=
rrrrrrr. (2.30a)
This expansion can be written in the form
( )...EEE),,;(KEE),;(KE);(P )3()3()2()2()1(0 +ωω−ωω−χ+ωωω−χ+ωω−χε=
rrrrrrr
(2.30b)
if the polarization is considered only at the fundamental frequency ω. The tensors χ(n) are
the macroscopic linear and nonlinear optical susceptibility of the order n of the medium,
respectively and ε0 is the vacuum permittivity. Equation (2.30b) differs from (2.30a) by
numerical factors K(n) which are related to the kind of the nonlinear optical process and to
the number of distinguishable permutations of frequencies [Butcher’92]. We will discuss
later the consequences of these numerical factors, if we compare the measured values of
nonlinear optical susceptibilities by means of different methods. The χ(n) is complex,
which can be written in term of real part Re[χ(n)] and imaginary part Im[χ(n)] as
]Im[i]Re[ )n()n()n( χ+χ=χ (2.31)
Next, we insert the electrical field for a plane wave which propagates in the z-
direction and has frequency ω and wave vector k = 2π/λ (λ is the wavelength)
)kztcos(EE 0 −ω= (2.32)
into Eq. (2.30a). By using appropriate trigonometric identities, one can obtain
_____________________________________________________________________17
[ ] −ω+ωωω−χ+−ωωωχε= )kz2t2cos(1
21E),;(K)kztcos(E);(P 2
0)2()2(
0)1(
0
−ω+−ωωωωω−χ+ )kz3t3cos(
41)kztcos(
43E),,;(K 3
0)3()3( (2.32)
As can be seen in Eq. (2.32), the polarization contains not only the component that
oscillates at the fundamental frequency ω, but also new frequencies 2ω and 3ω. The
polarization in Eq. (2.32) can be arranged into three components regarding to the
frequencies ω, 2ω, and 3ω, respectively
)kztcos(EE),,;(43K);()(P 0
20
)3()3()1(0 −ω
ωωωω−χ+ωω−χε=ω (2.33a)
[ ])kz2t2cos(1E),;(K21)2(P 2
0)2()2(
0 −ω+ωωω−χε=ω (2.33b)
)kz3t3cos(E),,;(K41)3(P 3
0)3()3(
0 −ωωωωω−χε=ω (2.33c)
The first term in P(ω) is related to the linear refractive index and the second term leads to
an intensity-dependent refractive index n(I). The P(2ω) term gives rise to many important
effects such as frequency doubling or second-harmonic generation (SHG), and sum- and
difference-frequency generation. There is also a frequency-independent contribution
which is referred to as optical rectification. Because of symmetry reasons,
centrosymmetric materials will not have second-order polarization and all subsequent
even-orders. The P(3ω) term corresponds to third-harmonic generation (THG).
In this thesis, we focus only on third-order nonlinear effects. The terms in brackets
of Eq. (2.33a) are treated as en effective susceptibility effχ , which is related to the
refractive index n through
( ) ( ) 20
)3()3()1(eff
2 E,,;K43;11n ωω−ωω−χ+ωω−χ+=χ+= . (2.34)
If the intensity 2000 Ecn
21I ε= is inserted into Eq. (2.34), we obtain
( ) ( )I,,;Kcn2
3;1n )3()3(
00
)1(2 ωω−ωω−χε
+ωω−χ+= . (2.35)
_____________________________________________________________________18
By introducing the linear refractive index n0 into Eq.(2.35) via ( ) 1n; 20
)1( −=ωω−χ and
with the approximation )nn(n2nn 0020
2 −≅− , we obtain the intensity-dependent
refractive index
Innn 20 += . (2.36)
To determine the relation between the real and imaginary part of χ(3) and the
nonlinear refractive index n2 and nonlinear absorption coefficient α2, the expression of
the complex of the refractive index of Eq. (2.36) has to be squared, yields
220
2 )In~n~(n~ += (2.37)
where
000 inn~ κ+= (2.38)
222 inn~ κ+= (2.39)
ii4
κλπ
=α . (2.40)
where κi is the imaginary part of refractive index ni. By using the combinations (2.35) and
(2.31) and the comparison with (2.37), the following relations are obtained
( )];Re[1n )1(20
20 ωω−χ+=κ− (2.41)
( )0
)1(
0 n2];Im[ ωω−χ
=κ (2.42)
( )],,;Re[Kcn4
3nn )3()3(
002020 ωω−ωω−χ
ε=κκ− (2.43)
( )],,;Im[Kcn4
3nn )3()3(
002020 ωω−ωω−χ
ε=κ+κ (2.44)
The first two equations describe the linear refractive index and linear absorption. The last
two equations describe the nonlinear refractive index and the two-photon absorption. As
the two-photon effects denoted by n2 and α2 are only interesting outside a one-photon
resonance, the linear loss term can be assumed to be zero (κ0 ≈ 0). Consequently, the
_____________________________________________________________________19
nonlinear refractive index n2 and the two-photon absorption coefficient α2 are directly
proportional to the real and imaginary part of the third-order susceptibility χ(3) through
( )],,;Re[Kcn4
3n )3()3(
020
2 ωω−ωω−χε
= (2.45)
( )],,;Im[Kcn
3 )3()3(
020
2 ωω−ωω−χλε
π=α (2.46)
Here the appropriate K(3) = 3/4 should be inserted, if χ(3) is defined as in Eq. (2.30b).
However, if χ(3) contains the numerical factor implicitly, as in Eq. (2.30a), K(3) = 1 has to
be used. The values of K(3) for several processes are listed in Tab. 2.1. The insertion of
this numerical factor is very important, if we compare the value of χ(3) obtained from
different methods. In this thesis, for example, we have used two different methods to
measure χ(3), i.e. third-harmonic generation (THG) and intensity-dependent prism
coupling. By using THG, we measure χ(3)(−3ω;ω,ω,ω), while n2 ~ Re[χ(3)(−ω;ω,−ω,ω)].
Therefore, we can not calculate n2 from THG measurements. Additionally, different
resonance enhancements can occur for χ(3)(−ω;ω,−ω,ω) and χ(3)(−3ω;ω,ω,ω).
Table 2.1: Selected third-order nonlinear optical process, abbreviations, frequency arguments of χ(3) and their numerical factor K(3) [Butcher’90].
Process χ(3) K(3)
Electric field induced second-harmonic generation (EFISH)
χ(3)(-2ω;ω,ω,0) 3/2
Third-harmonic generation (THG) χ(3)(-3ω;ω,ω,ω) 1/4
Generate four-wave mixing χ(3)(-ω4;ω1,ω2,ω3) 3/2
Degenerate four wave mixing (DFWM), χ(3)(-ω;ω,−ω,ω) 3/4
intensity-dependent refractive index χ(3)(-ω;ω,−ω,ω) 3/4
Cross-phase modulation, non-degenerate TPA χ(3)(-ω1;ω1,−ω2,ω2) 3/2
Quadratic electro-optic effect (Kerr) χ(3)(-ω;0,0,ω) 3
_____________________________________________________________________20
3 Experimental Methods
3.1 Spin Coating
Spin coating was used to prepare thin films from solutions. All studied films were
prepared on fused silica substrates (Spectrosil 2000, Saint-Gobain Quartz GmbH,
Germany) or silicon wafers (Si-Mat Silicon Materials, Germany) which had a thickness
of 1 mm. The substrates were cleaned by the following procedures:
• Washing with liquid soap and 10 x rinsing in purified water (Milli-Q, Millipore)
• 15 minutes cleaning in an ultrasonic bath with a solution of 1% detergent
(Hellmanex, Hellma) in milli-Q water
• 10 x rinsing in purified water
• Cleaning with ethanol and subsequently drying the substrate in a flow of nitrogen
The polymers were dissolved in appropriate organic solvents such as toluene and
chlorobenzene. The solutions were filtered by means of micro-syringe filters (0.5-1 µm).
Thin films were deposited on fused silica substrates by use of a spin coater (Headway
Research Type: ECD101D) from freshly prepared and filtered solutions under a laminar
flow hood to minimise dust particles. The rotation speed was varied from 500 to 9000
rpm (rounds per minute). The rotation time was set at 60 s. The film thickness d can be
varied by adjusting of concentration of the solution Cw and spinning speed ω, according
to the empirical relation [Ziegler’00, Fitrilawati’02]
βα
ωω
=0
1
0
101 Cw
Cwdd . (3.1)
The coefficients α and β were determined experimentally. The exponent α was -0.5
which is typical for the spin coating process.
Subsequent to the spin coating process, the films were annealed under vacuum for
at least 6 hours at the elevated temperature Ta to let the residual solvent evaporate. Ta was
set below the glass transition temperature, Tg of polymers, frequently at 500C. The
cooling process was done slowly (about 200C/h) to avoid stress of the polymer films.
_____________________________________________________________________21
3.2 Thickness Measurement
The thickness and the surface roughness of the polymer films were measured with
a step-profilometer (KLA Tencor Coorporation Model P 10) by scanning the surface
profile with a diamond tip (radius = 2 µm, force = 2 mg and sampling rate = 50 Hz). The
film thickness was measured at a scratch on the film which was made with a sharp needle.
The film thickness was evaluated as the height difference between film surface and
substrate surface at the scratch. The resolution of the thickness measurement was about 2
nm. The film roughness Ra is quantitatively determined from the average of absolute
values of the profile height deviations [Tencor’98].
3.3 UV-Vis-NIR Transmission and Reflection Spectroscopy
The dispersion of intrinsic absorption coefficient α(λ) and refractive index n (λ)
of thin polymer films (d ≈ 50 – 70 nm) were determined from the transmission and
reflection spectra measured by using the UV-Vis-NIR spectrophotometer (Perkin Elmer
model Lambda 900). The configurations for transmission and reflection measurements
are shown in Fig. 3.1. The light was s-polarized which means that the electrical field
vector was parallel to the film plane. Transmission spectra of thin films were measured at
normal incidence [see Fig. 3.1(a)]. Part of light is reflected at interfaces of air-film-, film-
substrate- and substrate-air-interfaces, respectively. However, the total reflection R can
not be measured at this transmission configuration. The reflection R was measured using
a reflection unit [Fig 3.1(b)] at an incidence angle of 150.
I0
TR
Film Substrate
I0R
SubstrateFilm
(a) (b)
150
I0
TR
Film Substrate
I0
TR
Film Substrate
I0R
SubstrateFilm
(a) (b)
150
Fig. 3.1: The configuration for measurement of (a) Transmission and (b) Reflection of thin films on fused silica substrates.
_____________________________________________________________________22
All measurements were performed using a bare fused silica substrate as reference. A
typical measured transmission and reflection spectrum is shown in Fig. 3.2.
400 600 800 1000 1200-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6 Transmission Reflection
A
[a.u
.]
λ [nm]
Fig. 3.2: Transmission and reflection spectra of a thin film of MEH-PPV (d = 54 nm) on a fused silica substrate expressed in absorbance units (A).
The dispersion of refractive index n(λ) and absorption coefficient α(λ) are
calculated by means of the available computer program that was developed earlier by
Ueberhofen [Ueberhofen ’96]. The calculations are based on the Fresnel equations and a
matrix formalism taking into account the refractive index dispersions of fused silica
substrate and polymer films. Typical results are shown in Fig. 3.3. The dashed line
(spectrum 1) refers to external absorption of film measured relative to the bare fused
silica substrate. The solid line (spectrum 2) refers to the intrinsic absorption, which was
obtained after correction of reflection losses at the air-film and substrate-air interfaces,
respectively. This leads to a blue shift of λmax relative to uncorrected spectra. The λmax-
value has an estimated uncertainty of ± 2 nm because of the broad absorption bands. The
experimental error of αmax is in the order of 5% which is caused mainly by the uncertainty
of thickness measurement by means of the step profilometer. The dispersion of refractive
index n is displayed in Fig. 3.3(b).
_____________________________________________________________________23
400 600 800 10000
5
10
15
20
2
1
(a)
n
α [1
04 cm
-1]
400 600 800 10001.2
1.4
1.6
1.8
2.0
2.2 (b)
λ [nm]
Fig. 3.3: Spectra of absorption coefficient α and refractive index n of a thin film of MEH-PPV at TE polarization. (a) Absorption spectra of a thin film (d = 54 nm) on fused silica substrate measured relative to fused silica (dashed line) and after correction of reflection losses at interfaces. (b) Dispersion of refractive index n.
3.4 Photostability
The photostability of the thin polymer films (d ≈ 70 nm) was studied by two
experiments: the wavelength dependence of photostability and the sensitivity of the
materials to UV photobleaching. All studies were done in ambient air. The wavelength
dependence of photostability experiment was performed in the following way: A high
pressure 450 Watt Xenon arc lamp (AMCO Model XBO 450 OFR) combined with water
filter and a variety of edge filters (Schott Glass) was used to irradiate the polymer films.
_____________________________________________________________________24
The transmission spectra of the edge filters are shown in Fig. 3.4(a). First, a longer
wavelength edge filter (λe > 610 nm) was used and then the shorter wavelength filters
were employed. For each filter, the absorption of the film was measured after every 10
minutes of exposure. The aim of this experiment is to find the cut-off wavelength (λc),
where no absorption changes after irradiation at λ > λc were observed.
500 550 600 650
0
20
40
60
80
100
λ = 361 nm
λ [nm]
Tram
smis
sion
[%]
(a)
OG
-610
OG
-590
OG
-570
OG
-550
OG
-530
Tram
smis
sion
[%]
λ [nm]200 400 600 800
0
20
40
60
80UG-1(b)
λ = 747 nm
Fig. 3.4: Transmission of several filters which were used in photostability experiments. Edge filters (a) and UG-1 filter (b).
The UV photobleaching experiment was performed in the following way: A high
pressure 100 Watt mercury lamp (OSRAM Model HBO 100 W/2) combined with a water
filter and a filter UG-1 was used [see Fig 3.4(b) for the transmission of UG-1 filter]. The
absorption of the film was measured after increasing exposure times. The aim of this
experiment is to measure the photosensitivity of the polymer films to UV-light (SUV). The
SUV value was determined from the initial decrease of αmax of the polymer films as a
function of absorbed intensity IA, which is given by
)101(II A0A
−−= (3.2)
where A the absorbance of the film at 365 nm, which refers to the peak of the irradiation
spectrum, caused by the strong mercury emission line. I0 is the irradiation intensity at the
sample position (typically I0 = 56 W/cm2).
_____________________________________________________________________25
3.5 FTIR Spectroscopy
Infrared spectra were recorded with a Nicolet Model Magna 850 FTIR
spectrometer at two configurations: Transmission and Grazing Incidence Reflection
(GIR). For transmission measurements, thin films (d ≈ 70 nm) were spin coated onto
silicon wafer substrates. The transmission spectra were measured at normal incidence (the
electrical field vector oriented parallel to the plane of film, E). Reflection spectra at
grazing incidence (electrical field vector is perpendicular to the film plane, E⊥) were
measured with thin films (d ≈ 70 nm) on 50 nm thick gold layers that were prepared
before by thermal evaporation onto glass substrates.
Film
E
SubstrateAu
E
Film Substrate
(a) (b)
Film
EE
SubstrateAu
EE
Film Substrate
(a) (b)
Fig. 3.5: FTIR spectroscopy at (a) Transmission and (b) Grazing incidence reflection (GIR) geometries.
3.6 Third-Harmonic Generation
Third-harmonic generation (THG) was used as screening test of the third-order
nonlinear optical susceptibility χ(3)(-3ω;ω,ω,ω) of the polymers. It measures only the
purely electronic contribution to the χ(3) of the polymers. THG can also be used for
spectroscopy if the fundamental laser frequency ω is varied. Because of two- and three-
photon resonances, THG can give additional information about electronic states that are
inaccessible with linear optical spectroscopy. Moreover, THG can also be used as a tool
to study the orientation of conjugated polymer chains [Kajzar’92]. However, THG can
not be used to measure the nonlinear refractive index n2 and nonlinear absorption
coefficient α2.
_____________________________________________________________________26
3.6.1 Experimental Setup and Measurement Procedure
The frequency doubled pulses of the laser (energy: 8-10 mJ, repetition rate: 10
Hz) of an actively/passively mode-locked Nd:YAG laser with a double pass amplifier
(EKSPLA model PL 2143B) were used to pump an optical parametric generator-amplifier
configuration (OPG - EKSPLA model PG 501). The OPG produced picosecond laser
pulses with typical energies of 200 - 500 µJ which were wavelength tuneable in the range
of 680 - 2200 nm. The temporal pulse profile was determined by background free
autocorrelation of second-harmonic generation (SHG) and was well fitted by a Gaussian
line-shape function. A pulse duration of τFWHM = 16 ± 2 ps was obtained which did not
depend significantly on the wavelength. The setup of the THG experiment is shown in Fig. 3.6. The laser beam was split
into two beams by means of a beam splitter. The first laser beam was focused by use of a
lens L1 on the sample (film thickness d ≈ 50 – 70 nm) which was mounted on a rotation
stage and placed in an evacuated chamber. The focal length of the lens L1 was 30 cm.
This procedure is necessary in order to avoid the contribution of air to the THG signal.
The contribution of air to the THG signal was studied by Kajzar and Messier [Kajzar’85].
Nd:YAG,
SHG 532 nm
OPG
690-2200 nm
θ M1 PM
M2 PM
Boxcar
BS L1
L2
RG850 KG3
quartz
sample
PC
Nd:YAG,
SHG 532 nm
OPG
690-2200 nm
θ M1 PMM1 PM
M2 PMM2 PM
Boxcar
BS L1
L2
RG850 KG3
quartz
sample
PC
Fig. 3.6: Setup of the THG experiment. OPG: optical parametric generation, BS: beam splitter, L1, L2: lenses, M1, M2: monochromator, PM: photomultiplier, PC: personal computer.
_____________________________________________________________________27
Because the focal length of the lens is wavelength dependent, the lens L1 was
mounted on a step-motor driven translation stage which was controlled by computer. The
focusing of the beam on the sample was adjusted at every incident wavelength. The other
beam was focused by means of a lens L2 (focal length = 5 cm) on a bare fused silica
substrate. The third-harmonic signal from this substrate was used as a reference. The
harmonic signal generated before the sample and reference were filtered by use of Schott
RG 850 filter. Behind the sample and reference, the fundamental light was absorbed by a
KG3 filter. After passing the monochromator, the harmonic light was detected with a
photomultiplier tube and measured with a Boxcar integrating amplifier (Stanford
Research Systems model SR 250). The signals of the sample and reference were
measured simultaneously to eliminate the effect of laser intensity fluctuation.
The experiment was performed by the following procedures: First, the intensity of
the third-harmonic signal generated from the film on substrate was measured as a
function of the incidence angle θ. A typical result of a thin film of TPD4-PPV on quartz
and a bare quartz measured at laser wavelength λL = 1350 nm is shown in Fig. 3.7. This is
commonly known as Maker fringe pattern. The oscillations at incident angle around 0
degree are caused by multiple reflections at the interfaces.
-30 -20 -10 0 10 20 30
0
20
40
60
80
100 λL = 1350 nm TPD4-PPV (film+quartz)
Quartz 5X
I 3ω [a
.u.]
Angle, θ [0]
Fig. 3.7: Maker fringe patterns of a thin film of TPD2-PPV (d ≈ 70 nm) and a quartz substrate measured at 1350 nm and TE-polarization. The symbols are measured data and lines are the theoretical fits.
_____________________________________________________________________28
Second, part of the film was removed from the substrate and then the Maker fringe
pattern of the substrate was measured. Two methods were used: First, for the substrate
with very good thickness homogeneity (d = 1 mm ± 1µm) (Hellma GmbH & Co. KG,
Germany), the Maker fringe of the film and fused silica substrate were measured at
different places. The homogeneity of the quartz was confirmed by obtaining the similar of
Maker fringes at different position of the quartz. For quartz substrates with a larger
variation of the thickness of ± 5 µm, the Maker fringe pattern of quartz was measured
after removal of the film without displacement of the sample (the same position as
measured for the film) [Neher’89].
3.6.2 Evaluation of Third-Order Nonlinear Optical Susceptibility
The third-order nonlinear optical susceptibility of thin films was calculated using
the available computer program that was developed earlier by Neher [Neher’90]. The
evaluation was carried out by the following procedures: First, the Maker fringe pattern of
the substrate was calculated using three input parameters: laser wavelength, refractive
index of the substrate at fundamental and harmonic wavelengths. Two fitting parameters
were used: the thickness and the modulus of third-order susceptibility of the substrate.
Second, the Maker fringe pattern of the film was fitted using the following input
parameters: laser wavelength, intrinsic absorption coefficient, refractive indices of the
substrate and film at fundamental and harmonic wavelengths, film thickness, the
thickness and the modulus of third-order susceptibility of the substrate which were
obtained from the fitting of the Maker fringes of the substrate. The modulus of χ(3) and
the phase angle Φ of the complex third-order susceptibility χ(3) = χ(3)exp iΦ were used
as fitting parameters for the film. The χ(3) value of the film was determined with respect
to the reference value of χs(3) = 3.11 x 10-14 esu for fused silica at laser wavelength λ =
1064 nm [Kajzar’85].
The experimental error of the χ(3) values was estimated in the order of 10 %. It is
mainly caused by laser pulse fluctuations, film thickness errors and experimental errors of
absorption coefficient and refractive index of film. A systematic error of χ(3) is also
possible, because the χ(3) value of the film was evaluated relative to the χ(3)-value of the
fused silica substrate which is not precisely known.
_____________________________________________________________________29
3.7 Attenuation Loss of Slab Waveguides
The measurement of waveguide losses was performed by the prism coupling
technique. The experimental setup is shown in Fig. 3.8. HeNe (λ = 633 nm) and cw-
Nd:YAG (λ = 1064 nm) lasers were used. The laser beams were coupled into waveguide
using a high refractive index glass prism LaSF18. The film was clamped onto the half-cut
prism mounted on a precision rotation table. The lens L1 with focal length of 30 cm was
used to focus the laser beam at the coupling edge of the prism. The coupling angle was
adjusted until the guided mode was launched in the waveguide.
HeNe,Cw Nd:YAG
PCSi-Diode array
S
L2
SubstrateFilm
P
L1AHeNe,
Cw Nd:YAG
PCSi-Diode array
S
L2
SubstrateFilm
P
L1A
Fig. 3.8: Experimental setup for waveguide attenuation loss measurement. A: polarizer, L1, L2: lenses, P: prism, S: shield, PC: computer.
Waveguiding in the film occurs if two conditions are fulfilled. First, the wave
velocities in the two coupled media (prism base and waveguide) must be the same and
second, the length along the coupled boundary must be adjusted according to the strength
of the coupling. The principle of a prism coupler is displayed in Fig. 3.9 [Tien’77].
_____________________________________________________________________30
Fig. 3.9: General principle of a prism coupler. The incident light is totally reflected at the base of the prism and the light wave in the prism is coupled into waveguide through the evanescent fields [Tien’77].
In the vector diagram of Fig. 3.9, the refractive index of the prism is n3 and the
incidence angle of the light in the prism is θ3. The wave vector in the prism has x-
component kx3 = k n3 sin θ3. On the other hand, the x-component of the wave vector of
the waveguide mode is β = k n1 sin θ1, where n1 is the refractive index of the film. When
both wave vectors are equal, an effective coupling is satisfied. Second condition: The
coupling length depends on the width of the light beam. Furthermore, the coupling
strength varies with the size of the air gap (the spacing between prim and film). Both
quantities can be adjusted. The air gap thickness can not be measured but it can be
evaluated from coupling experiments. This will be discussed later.
The scattered light from the waveguide was imaged onto a diode array by means
of lens L2 (focal length = 50 mm, diameter = 40 mm). To suppress the stray light of the
coupling prism the shield of black paper S was used. The distance calibration was
performed with a sheet of millimetre paper at the sample position imaged on the diode
array. Attenuation loss coefficients αgw were determined from the scattered light intensity
as function of distance from the coupling prism
[ ] ( ))x(Ilogx
10cm/dBgw =α (3.3)
where I(x) is the measured intensity of scattered light as a function of distance x. At
wavelength 1064 nm, the propagation of light in the waveguide can be observed by
means of an IR-camera (Micronviewer from Polytech Model 7290). The detection limit
of this method is in the order of αgw ≈ 0.5 dB/cm.
_____________________________________________________________________31
3.8 Intensity Dependent Prism Coupling
3.8.1 Procedure of Prism Coupling
Prism coupling of waveguides modes is a well-known and very accurate technique
to measure the refractive index of thin films. The theory of prism coupler was developed
by Tien and Ulrich [Tien’70, Ulrich’70, Ulrich’73]. The schematic cross section through
a prism-film coupler is shown in Fig. 3.10. Film and substrate are pressed against the
prism base by means of a spring loaded clamping screw. An air-gap forms between prism
and film with a thickness that depends on the clamping force. The refractive indices of
prism, air-gap, film, and substrate are denoted np, na, nf and ns, respectively. The incident
laser beam has intensity I0 and enters the prism at incidence angle θ. It is totally reflected
at the prism base. Its evanescent field can extend across the air-gap and is able to excite a
waveguide mode which is possible at discrete coupling angles θm, where m = 0, 1, 2, ...
describes the mode number. Consequently, the reflected intensity IR has relative minima
at θm because of absorption and scattering losses in the film (see Fig. 3.11).
IRθ
γ
I0
nf
na
np
da
db
Fig. 3.10: Prism coupling of a planar waveguide. See text for definition of symbols.
Nonlinear prism coupling occurs when the incident intensity I0 is so large that the
refractive index of the film becomes intensity dependent. Because the film is the only
nonlinear medium considered here, the intensity dependent refractive index is written as
_____________________________________________________________________32
nf(I) = n0 + n2I and the intensity dependent absorption coefficient of the film is αf(I) = α0
+ α2I. The intensity dependencies of nf and αf will lead to intensity dependent changes of
the coupling angles θm and the minimum values of the reflected intensities IR,
respectively. The intensity dependent changes of the coupling curves IR(θ) can be
evaluated to obtain the values and signs of n2 and α2. The model calculations were
developed earlier by Ueberhofen [Ueberhofen’99]. A more refined and accurate model
has been developed by Koynov to estimate the parameters which can not be measured
directly [Koynov’02].
3.8.2 Numerical Model of the Prism-Film Coupler
The well-established theory of Tien and Ulrich treats the linear case of prism
coupling in the limit of low intensities [Tien’70, Ulrich’70, Ulrich’73]. It is based on a
plane wave analysis and the use of Fresnel coefficients. In order to take into account
deviations from a uniform incident beam profile as well as nonlinear effects, the coupled-
wave equations for the spatial distributions of the electric fields of the guided wave
agw(y,z,t) and reflected wave aR(y,z,t) were used. They were derived by Carter and Chen
[Carter’83] and extensively studied by Assanto et al [Assanto’88]:
[ ]z)k)I(k(iexp)t,z,y(at)t,z,y(adzd
zpgwmoiningw −β−=
)t,z,y(a2
)I(t gw
gwgwout
α+− (3.4)
)t,z,y(a)t/r(2)t,z,y(ar)t,z,y(a gwinpfinpfR −= (3.5)
where the wave vector at the prism base kzp is related to the angle of incidence θ on the
entrance face of the prism [Ulrich’70]
])(sinn)sin()sin()[cos(k=k 22p0zp θ−γ+θγ (3.6)
and k0 = 2π/λ. The angles θ and γ are indicated in Fig. 3.10 and λ denotes the vacuum
wavelength. Basically, Eqs. (3.4) and (3.5) are rate equations for the electric fields
transmitted into and out of the film, where ain is the electric field of the incident light at
_____________________________________________________________________33
the base of the prism. βm(Igw) is the mode propagation constant, which, in the most
general case, depends on the intensity Igw of the guided wave and αgw(Igw) denotes the
attenuation coefficient of the waveguide that describes light scattering losses and intensity
dependent absorption. The coupling coefficients tin and tout describe the transfer of the
electric fields into and out of the waveguiding film. Coefficient rpf denotes the reflected
electric field strength for light incident from the prism to the prism-air-film boundary
with the film-substrate boundary removed. These three coefficients have already derived
and described in detail by Ueberhofen [Ueberhofen’96, Ueberhofen’99].
The function ain(y,z,t) is assumed to have Gaussian shape and Eqs. (3.4) and (3.5)
are integrated numerically. Then the reflected intensity is calculated using
2R
0pR )t,z,y(a
2cn
)t,z,y(Iε
= (3.7)
where c is the speed of light and ε0 is the vacuum permittivity. The result of numerical
calculations for IR(θ) can be compared with the experimentally measured coupling curves.
The quantitative comparison needs all parameters of the prism-film coupler. The guided
wave intensity Igw is a function of space and time which is defined as
)t(j)z(h)y(g)x(fI)t,z,y(I)x(f)t,z,y,x(I 0
gwgwgw == (3.8)
2
gw0f
gw )t,z,y(a2cn
)t,z,y(Iε
= (3.9)
where f(x) is the distribution of the intensity of the TE-waveguide mode perpendicular to
the film. 0gwI denotes the peak intensity is calculated using
tyx2/3
p0
)2(
EI
gw σσσπ= (3.10)
where Ep is the incident pulse energy, σx, σy and σt are the half widths of the beam size in
x, y direction and pulse duration, respectively.
The functions g(y) and j(t) are normalized Gaussian distribution functions determined by
the width and the duration of the incident laser pulse. The normalized distribution of the
intensity in the z direction, h(z), where agw(z) = agw(y = 0, z, t = 0) is
_____________________________________________________________________34
∫
=∞+
∞−dz)z(a
)z(a)z(h
2gw
2gw
. (3.11)
The average guided wave intensity in the film can be defined as
∫ ∫∫ ∫= ∞∞−
fd
0gw
fgw dydzdt)t,z,y(I)t(j)z(h)y(gdx)x(f
d1I (3.12)
It is important to point out that while most of the parameters can be directly measured,
this is not possible for the air-gap, which needs special attention.
3.8.3 Determination of Air-Gap Thickness
If the film is pressed against the prism, an air-gap is formed between them as
shown in Fig. 3.10. In general, this air-gap is not uniform but has some kind of parabolic
profile with minimum thickness at the center of the prism and increasing width towards
the edges. As an approximation in the numerical model, it can be replaced by a uniform
gap with thickness da and length la over which efficient coupling exists (see Fig. 3.10). To
determine la, the position of the incident beam at the prism base was shifted and the
coupling curves were measured at every position. When the incident beam deviated from
the center of the prism, a decrease in coupling efficiency was observed. For wide shifts,
the coupling disappeared entirely. Using this procedure, la can be estimated. In most cases
la was in the order of 2 mm.
The air-gap thickness da can be varied by increasing or decreasing the clamping
pressure. Changes in da strongly influence the minimal reflection Rmin at the center of the
coupling curve (θ = θm) and slightly influence the resonance position (θm) itself. This
effect is caused by modification of the free waveguide modes by the presence of the
prism and was described in the early papers of Tien and Ulrich [Tien’70, Ulrich’70]. An
example of coupling curves measured at three-different clamping pressures and their
numerical fit are shown in Fig. 3.11.
_____________________________________________________________________35
-3.0 -2.8 -2.6 -2.4 -2.2 -2.0
0.7
0.8
0.9
1.0
p1 p2 p3
θ [deg]
I R [a. u
.]
Fig. 3.11: Prism coupling of a MEH-PPV film on a fused silica substrate (film thickness 590 nm, laser wavelength λ = 806 nm). The TE0 modes were excited at low incident intensities I0 (linear case) and various clamping pressures p1 < p2 < p3. The lines represent numerical simulations calculated for n = 1.68555, α = 3.8 cm-1 and air-gap thicknesses da(p1) = 315 nm, da(p2) = 270 nm and da(p3) = 180 nm, respectively.
Numerical calculations of minimal reflection versus air-gap thickness for various
coupling parameters were performed by Koynov [Koynov’02]. The results are displayed
in Fig. 3.12. For any set of parameters, a minimum of Rmin(da) was observed and it was
called the optimal coupling situation. The following procedure to determine da was used:
The clamping pressure, and consequently also da, was gradually changed and the coupling
curves were measured at very low input intensities (linear case) until optimal coupling
was achieved. This curve at optimal coupling was fitted by use of the numerical model
described above, with two fit parameters only: linear refractive index n0 and linear loss
coefficient αgw of the film. Now with known n0 and αgw, it is straightforward to determine
da for any clamping pressure just by fitting the corresponding coupling curves measured
at low intensity as in Fig. 3.11.
_____________________________________________________________________36
0.7
0.8
0.9
1.0
(a)
db/la = 1/4
α=2cm-1
α=3cm-1
α=4cm-1R m
in
da [nm]
R min
200 300 400 500 6000.2
0.4
0.6
0.8
1.0
(b)
db/la = 3/4
α=2cm-1
α=3cm-1
α=4cm-1
Fig. 3.12: Numerical calculation for the minimal reflection Rmin (at the minimum of the coupling curve, θ = θm) as a function of air-gap thickness da for different absorption coefficients. The ratios between incident beam diameter db and air-gap length la, db\la are shown [Koynov’02].
Another important question connected with the air-gap is: Which thickness da is
the most suitable for observations of intensity-dependent changes of the coupling curves
and the evaluations of n2 and α2? It is clear from Fig. 3.12(a) that in case of a too large
air-gap (the so-called undercoupled situation), the changes of the absorption coefficient α
practically do not change the minimal reflection Rmin, the parameter that is measured
experimentally. If da is small, however (the so-called overcoupled situation), even small
changes in α lead to significant changes of Rmin. Of course too strong overcoupling is also
not desirable, as it leads to a decrease of the intensity of the guided wave. A comparison
of Fig. 3.12(a) and 3.12(b) shows that another important parameter is the ratio between
incident beam diameter db at prism base and air-gap length la. The minimal reflection Rmin
_____________________________________________________________________37
depends strongly on absorption, if the diameter of the incident beam db is considerably
smaller than air-gap length la. Therefore, slight overcoupling situation and incident beam
size db few times smaller than air-gap length la are preferable conditions for sensitive
detection of intensity-dependent changes in the coupling curves.
3.8.4 Experimental Setup
The experimental set-up for intensity dependent prism coupling is shown in Fig.
3.13. The laser source was the same laser as for the THG experiment. The absolute pulse
energy in front of the prism was measured with a pyroelectric detector (Laser probe
model RjP-735). The beam profile was measured with a beam-profiler (Exitech). The
intensity of the laser pulses was varied by means of a polarisation attenuator consisting of
a Fresnel rhombus and a polarizer. A spatial filter was used to improve the spatial profile
of the pulses and to focus them to the prism base as shown in Fig. 3.10. Typical beam
diameters at the prism base were around 0.3 - 0.5 mm. Two types of prisms were used,
i.e., LsF18 and SF59 glass which had a prism angle γ = 60°. The prisms were mounted on
the θ-arm of a precision 2θ-rotation stage. Its angular resolution was 0.01° and the
reproducibility was better than 0.005°. The reflected beam was detected by a photodiode
(PD2) mounted on the 2θ-arm of the rotation stage.
Nd:YAG, SHG: 532 nm
OPG: 680-2200 nm
L1 D L2 AFR
PD1
PD2
BOXCAR
PCNd:YAG, SHG: 532 nm
OPG: 680-2200 nm
L1 D L2 AFR
PD1
PD2
BOXCARBOXCAR
PC
Fig. 3.13: Experimental setup of nonlinear prism coupling. SHG: second harmonic generation, OPG: optical parametric generator, L1, L2: lenses, D: spatial filter, FR: Fresnel Rhombus, A: polarizer, BS: beam splitter, PD1, PD2: photodiode, PC: computer.
_____________________________________________________________________38
Thin films on fused silica substrate were pressed against the base plane of the
prism by means of a spring-loaded clamping screw. The signal of PD2 was measured with
a Boxcar integrated amplifier (Stanford Research Systems Model SR 250).
Simultaneously the signal of PD1, mounted in front of the prism to monitor the incident
intensity of the beam, was measured with the same Boxcar device. To reduce the problem
of intensity fluctuations of the OPG output, only such pulses that were located within a
narrow energy window at PD1 were selected and amplified. The ratio of the reflected to
the incident pulse energies at PD2 and PD1 was evaluated for each pulse and averaged
over 30 - 50 pulses. In this way, the normalized reflected intensity IR was obtained.
3.9 Fluorescence Spectroscopy
3.9.1 Steady-State Fluorescence
Steady-state fluorescence spectra of diluted solutions of polymers were recorded
with a SPEX Fluorolog 2 (F212) spectrometer, which was equipped with double
monochromators for excitation and emission. The light beam was incident to the cuvette
which was filled with the solution. The fluorescence signal was detected at 900 from the
incident beam.
3.9.2 Multiphoton Excited Fluorescence
The experimental setup for multiphoton excited fluorescence is shown in Fig.
3.14. As excitation source, the same Nd:YAG laser – OPG system as in THG experiment
was used. The incident beam was focused by means of lens onto a fused silica cuvette
containing the solution (internal path length: 2 mm). Lenses with focal length of 500 mm
and 300 mm were used for two-photon and three-photon fluorescence experiments,
respectively. The pulse energy was measured with a pyroelectric detector (Laser probe
model RjP-735). The beam profile was measured with a beam-profiler (Exitech). The
intensity of the laser pulses was varied by means of polarisation attenuator consisting of a
Fresnel rhombus and a polarizer. A spatial filter was used to improve the spatial profile of
the pulses and to control the laser beam size. The beam size was kept constant at about
0.5 mm. A part of the fluorescent light was collected and guided by an optical fiber to a
monochromator. Then the signal was amplified by photomultiplier and recorded by
Boxcar.
_____________________________________________________________________39
MC
PMBOXCAR
EM
LBSAFR Sample
OPG
680-2200 nm
Nd:YAG, SHG 532 nm PC
D
MC
PMBOXCAR
EM
LBSAFR Sample
OPG
680-2200 nm
Nd:YAG, SHG 532 nm PC
D
Fig. 3.14: Experimental setup of multiphoton excited fluorescence. SHG: second harmonic generation, OPG: optical parametric generator, D: spatial filter, L: lens FR: Fresnel Rhombus, A: polarizer, BS: beam splitter, OF: optical fiber, MC: monochromator, PM: photomultiplier, PC: computer.
_____________________________________________________________________40
4 Materials Properties of PPV Derivatives
High-speed photonic switching and all-optical signal processing require materials
that have multifunctional properties like high third-order nonlinearity with fast response
times, low absorption losses, high photostability and easy fabrication of slab waveguides.
Derivatives of poly(p-phenylenevinylene), PPV have been considered the most promising
organic material candidates for these applications [Bartuch’92, Bubeck’95, Michelotti’95,
Mathy’96, Gabler’97, Gabler’98, Ueberhofen’99, Bubeck’00, Koynov’02,
Fitrilawati’02]. In this chapter, we present a study of the linear and nonlinear optical
properties of several PPV derivatives with the aim to select the best-suited material for
all-optical switching applications.
4.1 Materials and Film Preparation
The investigated materials were provided by Prof. H.-H. Hörhold from the
University of Jena, Germany. Their chemical structures are displayed in Fig. 4.1. The
synthesis of the polymers has been referred to in earlier reports (see Appendix B for their
full chemical names): MEH-PPV [Pfeiffer’99a, b, Hörhold’02], MEH-DOO-PPV
[Hörhold’01, Hörhold’02], M3EH-PPV [Pfeiffer’99a, b], MEH-M3EH-PPV
[Hörhold’01], TPD2-PPV, and TPD4-PPV [Hörhold’01]. These dialkoxy-substituted PPV
derivatives were synthesized via the polycondensation route by the use of the Horner-
carbonylolefination as the step growth polymerization process that yields well-defined
conjugated polymers with excellent solubility in organic solvents. This chemical route is
advantageous over the frequently used dehydrohalogenation process as it avoids both
branching (crosslinking or gel formation) and incomplete double-bond formation,
resulting in strictly linear fully conjugated defect-free phenylene vinylene backbone
structures [Hörhold’02]. With appropriate side chains (e.g. long or branched dialkoxy
groups attached to p-phenylene) the materials are completely soluble in toluene or
chlorobenzene. The weight-average molecular weight (Mw) and number-average
molecular weight (Mn) of all PPVs were determined with gel permeation chromatography
(GPC) using polystyrene standards and tetrahydofuran (THF) as eluent. The glass
transition temperature Tg was measured by means of differential scanning calorimetry
(DSC). Their values are displayed in Table 4.1 as provided by Prof. H.-H. Hörhold.
_____________________________________________________________________41
O
O
CH3
nCH=CH
MEH-PPV
nCH CH CH CH
O
CH3O
OC8H17
C8H17O
MEH-DOO-PPV
CH CH
OCH3
CH3O
O
CH3On
CH CH
M3EH-PPV
CH CH CHCH
CH3O
O
CH3O
O
0.5nCH CH CH
CH3O
O
CH3O
OCH3
0.5nCH
MEH-M3EH-PPV
N CH CH CH CH
O
CH3O
CH3
N
CH3
n
TPD-4M-MEH-PPV (TPD2-PPV)
N CH CH CH CH
O
CH3O
CH3
N
CH3
0,5n 0,5nCH CH CH CH
OCH3
CH3O CH3O
O
TPD-4M-MEH-M3EH-PPV (TPD4-PPV)
Fig. 4.1: Chemical structures of poly-(p-phenylenevinylene) (PPV) derivatives. Their full chemical names are given in Appendix B.
_____________________________________________________________________42
Table 4.1: Properties of PPV-derivatives
Polymer Mw [g/mol] Mn [g/mol] PDI Tg [0C] Solvent
MEH-PPV 4.03 x104 1.41 x104 2.86 65 Toluene
MEH-DOO-PPV 3.43 x 104 1.30 x 104 2.64 (*) Toluene
M3EH-PPV 4.40 x 104 1.20 x 104 3.67 113 Chlorobenzene
MEH-M3EH-PPV 3.23 x 104 1.14 x 104 2.83 81 Toluene
TPD2-PPV 5.31 x 104 1.46 x 104 3.64 188 Toluene
TPD4-PPV 3.57 x 104 1.31 x 104 2.73 165 Toluene
(*). In the temperature range up to 2000C, no glass-transition or melting peak was
observed
Two common organic solvents like toluene and chlorobenzene were used to
dissolve the polymers. Most of the polymers were dissolved in toluene (see Table 4.1)
and processed to thin films by spin coating at room temperature as described in Chapter 3.
However, M3EH-PPV was treated differently. It could not be dissolved in toluene.
Therefore, this polymer was dissolved in chlorobenzene. The solution was heated at
approximately 1000C and stirred for 1 hour in order to obtain complete solubility. The
films of this polymer were prepared by spin coating at high temperature (~ 1000C). All
films were annealed under vacuum for at least 6 hours at an elevated temperature Ta to let
the residual solvent evaporate. Ta was set below Tg of polymers, typically Ta ≈ 500C.
4.2 Linear Optical Properties
4.2.1 Results
Transmission and reflection spectra of thin films (d ≈ 50 nm) were measured by
using a UV-Vis-NIR spectrophotometer (Perkin Elmer Model Lambda 900) with s-
polarized light (electrical field vector oriented parallel to the film plane, TE-polarization).
_____________________________________________________________________43
The dispersions of the intrinsic absorption coefficient α(λ) and the refractive index n(λ)
of thin films were evaluated from the transmission and reflection spectra using the
procedure described in Chapter 3. The absorption spectra and the dispersions of linear
refractive index of thin films of the PPVs are displayed in Figs. 4.2 and 4.3.
200 300 400 500 600 7000
5
10
15
20
25
n
λ [nm]
(a) M3EH-PPV
MEH-M3EH-PPV
MEH-DOO-PPV
MEH-PPV
α [1
04 cm
-1]
λ [nm]
200 400 600 800 10001.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
(b) MEH-PPV MEH-DOO-PPV MEH-M3EH-PPV M3EH-PPV
Fig. 4.2: Spectra of (a) the absorption coefficient after correction of reflection losses and (b) the refractive index of thin films of PPV derivatives at transverse electric (TE) polarization.
_____________________________________________________________________44
200 300 400 500 6000
10
20
30
(b)
n
λ [nm]
TPD4-PPVTPD2-PPV
(a)
α [1
04 cm-1]
λ [nm]
200 400 600 800 1000
1.4
1.6
1.8
2.0
2.2 TPD2-PPV
TPD4-PPV
Fig. 4.3: Spectra of (a) the absorption coefficient after correction of reflection losses and (b) the refractive index of thin films of TPD containing PPVs at transverse electric (TE) polarization.
The maxima of absorption αmax in the UV and visible range, their wavelengths
λmax, and the onsets of absorption at λ0 for all PPVs studied are given in Table 4.2. The
wavelength λ0 is defined as the intercept between the baseline and the tangent to the
absorption edge. The value of λ0 will be used later in the discussion of the photostability
studies.
_____________________________________________________________________45
Table 4.2: Linear optical constants of PPV-derivatives.
Polymer αmax (UV)
[104 cm-1] ± 5%
λmax (UV)
[± 2 nm]
αmax (Vis)
[104 cm-1] ± 5%
λmax (Vis)
[± 2 nm]
λ0
[± 5 nm]
MEH-PPV 12.3 210 16.6 489 570
MEH-DOO-PPV 12.2 208 16.0 486 590
MEH-M3EH-PPV 16.1 210 20.4 485 590
M3EH-PPV 18.8 209 22.0 486 600
TPD2-PPV 30.1 208 18.4 426 500
TPD4-PPV 26.7 208 16.9 432 520
For quantitative specification of the refractive index dispersions, we have fitted the
spectra in the range of 550 to 1600 nm by the Sellmeier equation [Lines’91]
∑λ−λ
λ+=λ
=
2
1i 2i
2
2i2 a1)(n (4.1)
using two poles. The fit parameters a1, λ1, a2, and λ2 are given in Table 4.3.
Table 4.3: The Sellmeier fit parameters of PPV-derivatives.
Polymer a1 λ1 [nm] a2 λ2 [nm]
MEH-PPV 1.416 200.0 0.208 528.9
MEH-DOO-PPV 1.357 180.0 0.242 518.7
MEH-M3EH-PPV 1.543 200.0 0.265 532.4
M3EH-PPV 1.652 200.0 0.295 538.0
TPD2-PPV 1.571 102.5 0.305 445.4
TPD4-PPV 1.587 185.4 0.224 486.4
_____________________________________________________________________46
In addition to reflection spectroscopy, we used prism coupling i.e. the m-line
technique [Ulrich’73] to measure the refractive index of PPV slab waveguides at several
wavelengths between 633 nm and 1064 nm. The typical thicknesses of waveguides were
in the range of 400 - 800 nm. The results are displayed in Figs. 4.4 and 4.5, together with
the results of transmission-reflection measurements. The refractive indices at TE-
polarization, nTE of prism coupling and transmission-reflection measurements agree very
well which shows that nTE does not significantly depend on the film thickness, at least for
film thicknesses in the range of 50 to 800 nm. A slight disagreement of the refractive
index of thin films and waveguides of MEH-DOO-PPV might be due to the error in the
refractive index calculation from transmission-reflection data. We have also measured the
refractive index at TM polarization, nTM of several slab waveguides of PPVs. The results
are displayed in Fig. 4.5 with open symbols. The values of nTM for these three polymers
are nearly identical.
500 600 700 800 900 1000 1100
1.6
1.7
1.8
1.9
2.0
2.1
TPD4-PPV
MEH-DOO-PPV
TPD2-PPV
n
λ [nm]
Fig. 4.4: Dispersions of the refractive indices n at TE polarization of MEH-DOO-PPV (dotted line), TPD2-PPV (dashed line) and TPD4-PPV (solid line). Lines are from transmission-reflection experiments of thin films. Data points are from prism coupling experiments [MEH-DOO-PPV (full squares), TPD2-PPV (full circles), and TPD4-PPV (open triangles)].
_____________________________________________________________________47
500 600 700 800 900 1000 1100
1.6
1.8
2.0
2.2
2.4
M3EH-PPV
MEH-M3EH-PPV
MEH-PPV
n
λ [nm]
Fig. 4.5: Dispersions of refractive indices of MEH-PPV, MEH-M3EH-PPV and M3EH-PPV. Lines are from transmission-reflection experiments of thin films at TE-polarization. Data points are from prism coupling experiments at TE polarization (full symbols) and TM polarization (open symbols).
4.2.2 Discussion
The absorption spectra of PPV derivatives consist of two strong peaks at 430 - 500
nm and 210 nm and two shoulders at 255 nm and 330 nm (see Figs. 4.2 and 4.3). The
assignment of the absorption band of PPV and its derivatives has been discussed in the
past [Chandross’94, Garstein’95, Chandross’97, Martin’99]. The maximum at around 430
- 500 nm is caused by an electronic transition between a molecular orbital delocalized
along the polymer backbone (π−π* transition). The strong UV peak (~ 210 nm) and the
shoulder at 255 nm are due to transitions between localized and delocalized states
(σ−π∗ transition), which are originating from the phenyl ring. The origin of the shoulder
at 330 nm (3.7 eV) is still unclear. Gartstein et al [Garstein’95] have assigned this peak to
charge conjugation symmetry (CCS) breaking due to asymmetry substitution which
allows a transition that is forbidden in unsubstituted PPV. They have used a microscopic
model for photoexcitation of polyphenylenes which allowed to calculate the effects of
energy correlations between charge-neutral excitons and dipole-dipole interactions
between neighboring monomers. They have found that the CCS breaking induced a peak
at 3.7 eV (~ 330 nm) which is a general characteristic of substituted PPVs. By using
_____________________________________________________________________48
Coulomb correlated theoretical models, however, Chandross et al [Chandross’97] have
suggested that the effect of CCS breaking alone on the absorption spectrum is rather
weak. They have argued that the peak at 3.7 eV is caused by the influence of short PPV
chains that exist in all real polymers, which lead to the finite size effects. By comparing
the absorption spectra of unsubstituted PPV and MEH-PPV, Martin et al [Martin’99]
have suggested that the CCS breaking is the most likely explanation for the appearance of
peak at 330 nm.
TPD containing PPVs have stronger absorption in the UV region (~ 210 nm) as
compared to the other PPVs. This is caused by the larger number of phenyl-rings per
repeat unit (see their chemical structures in Fig. 4.1). The stronger UV maxima of TPD
containing PPVs lead to a larger refractive index n but reduced dn/dλ in the near infrared
(NIR) as compared with MEH-DOO-PPV (see Fig. 4.5). This structure-property relation
indicates a promising way to tailor the dispersion of the refractive index by appropriate
chemical synthesis of the polymers.
M3EH-PPV has the strongest absorption maximum αmax as compared to the other
PPVs studied. It can be understood from the chemical structures (see Fig. 4.1) that
M3EH-PPV has long alkyl substituents at only every second phenyl-ring, whereas MEH-
PPV contains alkyl-chains at every phenyl-ring. Their copolymer MEH-M3EH-PPV has
three alkyl-chains in every four phenyl-rings. Thus, the reduced amount of alkyl-chains
leads to an increase of the number π-electrons per unit volume and consequently αmax
increases. It can be understood in a similar way that nTE and the average refractive index,
nave [nave = (nTE + 2 nTM)/3] increases in the sequence from MEH-PPV, MEH-M3EH-PPV
to M3EH-PPV.
4.3 Waveguide Propagation Losses
4.3.1 Results Waveguide propagation loss experiments were performed at 1064 nm by use of a
cw-Nd:YAG laser and the setup described in Chapter 3. The laser beam was s-polarized
to excite TE modes. The waveguide propagation loss was determined by measuring the
stray light of TE modes as a function of distance from the coupling prism. Fig. 4.6 shows
examples of the stray light at the diode array for MEH-DOO-PPV and MEH-M3EH-PPV
slab waveguides. The slopes of the lines fitted to the experimental data yield the loss
_____________________________________________________________________49
coefficients of the guided waves αgw. The values of αgw of all PPVs are presented in
Table 4.4.
0.2 0.4 0.6 0.8 1.0 1.2
100
1000 (b)
Inte
nsity
[a.u
.]
αgw = 12.6 dB/cm
Inte
nsity
[a.u
.]
distance [cm]
0.0 0.5 1.0 1.5 2.0
100
1000 (a)
αgw = 0.8 dB/cm
Fig. 4.6: Stray light of the TE0 mode of MEH-DOO-PPV waveguide (a) and MEH-M3EH-PPV waveguide (b) versus distance from coupling prism at laser wavelength λ = 1064 nm. The solid lines represent the waveguide propagation loss coefficient αgw. Table 4.4: Waveguide propagation loss coefficient αgw of PPV-derivatives, measured at 1064 nm and TE polarization, and the relative surface roughness Ra/d of the films.
MEH-PPV
MEH-DOO-PPV
MEH-M3EH-PPV
M3EH-PPV
TPD2-PPV
TPD4-PPV
αgw (TE0) [dB/cm]
0.5 ± 0.3 0.8 ± 0.3 12.4 ± 1.4 > 20 > 20 3.0 ± 1.2
Ra/d [%] 0.4 0.6 0.5 1.2 0.6 0.5
_____________________________________________________________________50
4.3.2 Discussion
Materials with a small waveguide propagation loss coefficient (αgw < 1 dB/cm)
are required for all-optical waveguide applications in order to realize devices with high
throughput [Stegeman’89, Bubeck’00]. We have succeeded in preparing the required
waveguides with losses of less than 1 dB/cm by using MEH-PPV and MEH-DOO-PPV
polymers. However, this was not yet feasible with the other PPVs studied here.
The data for αgw, which are presented in Table 4.4, contain the contributions of
intrinsic absorption and the scattering losses, which depend on the surface roughness of
the waveguide. Although the values of the relative surface roughness (Ra/d) of all PPVs
are comparable, they exhibit different values of αgw. This might be related to aggregate
formation caused by a different solubility of the polymers. The substitution with the
branched 2-ethylhexyloxy group causes very good solubility of the PPVs. We have
observed that the solubility is reduced in going from MEH-PPV to MEH-M3EH-PPV and
finally to M3EH-PPV due to the decrease of the relative number of this “solubility
providing” substituent (see Fig. 4.1). Consequently, an increased tendency of aggregate
formation is imaginable which would cause an increase of light scattering in the sequence
of these three PPVs. As a consequence, αgw increases. Another factor is probably related
to the morphology of the films. We will show later in Chapter 6, that large changes of αgw
can be caused by different morphology, in particular on the arrangement of polymer
chains in the films.
.
4.4 Third Harmonic Generation
4.4.1 Results
We have performed the THG experiments as a screening test to elucidate the
purely electronic contribution to the third-order susceptibility χ(3) of the PPV-derivatives.
The experimental setup and evaluation procedure were described in Chapter 3. The
typical film thickness was approximately 50 - 70 nm. We have measured the modulus of
χ(3) in a broad spectral range. The dispersions of the modulus of χ(3) at 1/3 of the
fundamental wavelength compared with linear absorption spectra for all PPVs studied are
displayed in Figs. 4.7 and 4.8. The χ(3) values exhibit a strong spectral dependence on the
laser wavelength λL because of three-photon resonances with electronic states of the
_____________________________________________________________________51
polymers. The spectrum of the χ(3) resembles the linear absorption coefficient: it has a
maximum, denoted χ(3)max, at the laser wavelength λL ≈ 3λmax of the absorption
coefficient. The values of χ(3) of all PPVs studied are listed in Appendix C.
300 400 500 600
0
5
10
15
20
0
2
4
6
8MEH-PPV
300 400 500 600
0
5
10
15
20 MEH-M3EH-PPV
0
2
4
6
8
10
300 400 500 600
0
5
10
15
20
25M3EH-PPV
χ(3) [1
0-11 e
su]
χ(3) [1
0-11 e
su]
χ(3) [1
0-11 e
su]
α [1
04 cm
-1]
α [1
04 cm
-1]
α [1
04 cm
-1]
λ, λL/3 [nm]
0
5
10
15
Fig. 4.7: Spectra of the modulus of χ(3) at λL/3 in comparison with linear absorption spectra of thin films of MEH-PPV, MEH-M3EH-PPV and M3EH-PPV. Both types of spectra were measured at TE-polarization.
_____________________________________________________________________52
300 400 500 600
0
5
10
15MEH-DOO-PPV
300 400 500 600
0
5
10
15
20TPD2-PPV
0
1
2
3
4
300 400 500 600
0
5
10
15
20TPD4-PPV
λ, λL/3 [nm]
0
1
2
3
4
α [1
04 cm
-1]
α [1
04 cm
-1]
α [1
04 cm
-1]
χ(3) [1
0-11 e
su]
χ(3) [1
0-11 e
su]
χ(3) [1
0-11 e
su]
0
3
6
9
Fig. 4.8: Spectra of the modulus of χ(3) at λL/3 in comparison with linear absorption spectra of thin films of MEH-DOO-PPV, TPD2-PPV and TPD4-PPV. Both types of spectra were measured at TE-polarization.
_____________________________________________________________________53
4.4.2 Discussion
Many theoretical models have been developed to interpret the THG spectra,
namely non-interacting electrons, excitonic and strongly correlated electrons [Abe’97].
Among them, the excitonic model was successfully used to explain the THG spectra of
unsubstituted PPV as sketched in Fig 4.9 [Bubeck’95, Mathy’96]. It consists of a valence
band (VB), a conduction band (CB) and an exciton state (EX) located below the CB.
VB
EX
CB
1 2 3 4VB
EX
CB
1 2 3 4 Fig. 4.9: Possible multiphoton resonances in third-harmonic generation and energy states of conjugated polymers; VB: valence band; CB: conduction band; EX: exciton. The processes 1, 2, 3 and 4 are explained in the text.
The strong maximum of χ(3)max is ascribed to a three-photon resonance with states
located at the top of the valence band and the exciton state (process 1). It occurs at
wavelength λL(χ(3)max) ≈ 3λmax of linear absorption coefficient. The values of χ(3)
max and
their laser wavelengths λL (χ(3)max) of all PPVs studied are given in Table 4.5. However,
the peaks of χ(3) spectra of all-PPVs are red shifted as compared to their linear absorption
maxima (see Tables 4.1 and 4.5). These shifts were also observed in other compounds
and were explained as a consequence of the statistical distribution on the effective π-
conjugation length [Kurihara’91, D’Amore’02]. Longer conjugation lengths are more
_____________________________________________________________________54
effective in the THG process than the shorter ones. As consequence, the red-shift was
always observed in all PPVs.
Process 2 is a three-photon resonance with the continuum of states located at the
bottom of the conduction band. Process 3 is related to a two-photon resonance with a two-
photon state located energetically below the exciton state (EX). As sketched in Fig. 4.9,
both processes 2 and 3 could be visible at the same laser wavelength λL. Therefore, it is
not possible to distinguish between them only by means of a THG experiment. Two-
photon absorption spectroscopy or other nonlinear optical techniques could help to
identify and distinguish these processes. Processes 2 and 3 will lead to an additional
resonance at λL < 3λmax in the THG spectra [Mathy’96]. However, as shown in Figs. 4.7
and 4.8, no additional resonance was observed in all PPVs studied here.
The two-photon absorption state of PPV and its derivatives were studied by
several groups [Baker’93, Leng’94, Mathy’96, Liess’97, Meyer’97]. It was shown that
the two-photon state lies energetically above the lowest one-photon allowed optical
transition. Martin et al [Martin’99] have reported that the two-photon state of MEH-PPV
is located at 2.9 eV. This state should be visible as a two-photon resonance in THG
(process 4) at laser wavelength λL around 850 nm, which is not accessible with our
present experimental setup.
Table 4.5: Nonlinear optical properties of PPV-derivatives measured by means of THG. The values of αmax were taken from Table 4.1.
MEH-PPV
MEH-DOO-PPV
MEH-M3EH-
PPV
M3EH-PPV
TPD2-PPV
TPD4-PPV
)3(maxχ
[10-11 esu] ± 10 %
6.50 8.60 9.65 13.80 3.70 3.43
λL ( ))3(maxχ [nm] 1530 1560 1500 1530 1350 1350
)3(maxχ /αmax
[10-16 [esu cm] ± 10%
3.9 5.4 4.7 6.3 2.0 2.0
_____________________________________________________________________55
As can be seen in Table 4.5, the conjugated polymer MEH-PPV and its related
structures have larger )3(maxχ as compared to the TPD containing PPVs. This might be
related to the different conjugation lengths of the polymers. MEH-PPV and its related
structures have longer conjugation length (λmax ≈ 486 nm), whereas TPD containing PPVs
have shorter conjugation length (λmax ≈ 430 nm). This is also roughly in line with scaling
behavior of 1D conjugated system between αmax, λmax and )3(maxχ through [Mathy'96,
Bubeck'98]
( ) xmaxmax
)3(max ~,,;3 λαωωωω−χ (4.2)
with an exponent x = 10 ± 1. The TPD-PPVs show smaller value of )3(maxχ /αmax, which
indicates shorter conjugation length, as compared with MEH-PPV and its related
structures (see Table 4.5).
4.5 Stability Investigation
4.5.1 Results
Photostability of the polymer becomes an important issue for photonic device
applications. It is well known that PPVs can be easily photooxidized if they are exposed
to UV/VIS light at ambient air [Yan’94, Cumpston’95, Rothberg’96, Xing’97, Atreya’99,
Fitrilawati’02]. In this study, we have measured the sensitivity of the polymers to
photodegradation by UV-light, denoted SUV, as defined in Chapter 3. We have used a
high-pressure mercury lamp in combination with a glass filter with a high transmission at
365 nm (UG1 from Schott) to irradiate thin films (d ≈ 70 nm) of the PPV-derivatives at
ambient air. The UV-photobleaching was measured by recording the transmission and
reflection spectra after certain times of exposure. An example result for MEH-PPV is
shown in Fig. 4.10.
The absorption coefficient α(λ) decreases and its maximum shifts towards shorter
wavelengths with increasing exposure time. This effect is caused by photochemical
conversion of the π-electron of the chromophores that is combined with shortening of the
π-electron delocalization length [Rothberg’96]. The refractive index of the bleached
_____________________________________________________________________56
polymer also decreases significantly. In the example of MEH-PPV shown in Fig. 4.10(b),
we have observed a refractive index change ∆n = 0.17 at λ = 1064 nm. The film thickness
after UV-bleaching, however, did not change as we have verified by surface-profilometry.
Even after extended photobleaching, the films remained stable. Mechanically unstable
films can be easily identified because the diamond needle of the profilometer can leave
traces on the surface of the films. This was not observed in the case of UV-bleached films
of PPVs.
200 300 400 500 600 700
0
5
10
15
λ [nm]
n
(a)
α [1
04 cm
-1]
300 500 700 900 1100
1.4
1.6
1.8
2.0
2.2 (b)
Fig. 4.10: Spectra of (a) the absorption coefficient and (b) the refractive index of a fresh MEH-PPV film with thickness 58 nm (solid line) and after UV-irradiation at 365 nm with an intensity of 33 mW/cm2 at sample position for 1 hour (dashed line) and 2 hours (dotted line).
_____________________________________________________________________57
We have observed a linear increase of ∆αmax/αmax with the number of absorbed
UV photons IA (Eq. 3.2), without induction period or threshold, as can be seen in Fig.
4.11. The corresponding initial slope SUV is used as a quantitative measure for the
sensitivity to the UV bleaching of the material. All polymers studied here can be
bleached. However, they show different sensitivities SUV (see Table 4.6). For practical
reasons, it is desirable to use a more stable polymer (smaller SUV), if the film preparation
and optical investigations are performed at ambient air condition.
0 10 20 30
0.0
0.1
0.2
0.3
SUV
MEH-PPV
MEH-DOO-PPV
∆α
max
/αm
ax(t=
0)
IA [a.u.]
Fig. 4.11: Comparison of the UV sensitivity of MEH-PPV and MEH-DOO-PPV.
The wavelength dependence of photobleaching was studied by illumination of the
polymer films by use of a high pressure Xe arc lamp combined with a water filter and a
variety of edge filters as described in Chapter 3. The cut-off wavelength λc is defined in
such the way that no change of absorption could be observed when we irradiated the
samples with λ > λc. The values of λc are presented in Table 4.6. A comparison with
Table 4.1 shows that λc agrees well with λ0, which indicates that the photodegradation
problems do not occur in the near-infrared at working wavelengths λ >> λc.
_____________________________________________________________________58
Table 4.6: The UV photosensitivity SUV and cut-off wavelength λc of PPV-derivatives.
MEH-
PPV
MEH-
DOO-PPV
MEH-
M3EH-PPV
M3EH-
PPV
TPD2-
PPV
TPD4-
PPV
SUV [a.u.] 0.007 0.011 0.005 0.004 0.013 0.011
λc [± 5 nm] 550 590 590 590 500 520
1800 1700 1600 1500 1000 900 8000.00
0.01
0.02
0.03(a)
Cha
nge
in e
xtin
ctio
n
Exposure fluence [J/cm2]
Exposure Fluence: 0 J/cm2
25.9 J/cm2
51.7 J/cm2
77.5 J/cm2
Extin
ctio
n [a
.u.]
Wavenumber [cm-1]
0 20 40 60 80
-0.005
0.000
0.005
0.010(b) 970 cm-1
1503 cm-1
1600 cm-1
1672 cm-1
1740 cm-1
Fig. 4.12: (a) FTIR spectra of MEH-PPV film after irradiation with different UV light intensities given in the inset. (b) Changes of major infrared bands as a function of UV exposure.
_____________________________________________________________________59
In order to study the influence of UV light on the chemical structure, we have
measured the FTIR spectra of thin film of MEH-PPV as an example with different
exposure times. The film (d ≈ 70 nm) was spin coated onto a silicon substrate. The
changes of infrared spectra as a function of exposure time are presented in Fig. 4.12. The
intensities of the bands at 970 cm-1 and at 1503 cm-1 decrease with increasing exposure
time. At the same time, new bands (1600 cm-1, 1672 cm-1 and 1740 cm-1) appear and
increase with exposure time.
4.5.2 Discussion
The photostability of PPVs have been widely studied in the last decade [Yan’94,
Scott’96, Rothberg’96]. It is known that the UV light causes the formation of carbonyl
C=O groups in the PPV backbone, which leads to the shortening of the conjugation
length, as illustrated in Fig. 4.13 [Rothberg’96].
CC
H
H n
CC
O
nH
HPhoto oxidation
Fig. 4.13: Mechanism of photooxidation in PPV [Rothberg’96].
Five major bands are used to study the primary changes in the chemical structure
of PPVs due to photooxidation. They are listed in Table 4.7 [Bradley’87, Cumpston’95,
Scott’96, Rothberg’96]. The intensity of trans-vinylene CH band at 970 cm-1 decreases
with increasing exposure time, reflecting the breaking of the backbone double-bond. At
the same time the carbonyl C=O band at 1672 cm-1 increases. The reduction of intensity
at 1503 cm-1 indicates the redistribution of the oscillator strength due to incorporation of
C=O. The formation of carbonyl bond leads to the shortening of the conjugation length
[Rothberg’96]. As consequences, αmax, λmax and also refractive index decrease. The band
at 1600 cm-1 is assigned to a conjugation defect [Rothberg’96]. At longer exposure time, a
higher oxidation product like ester or carboxylic acid at 1740 cm-1 becomes more
pronounced.
_____________________________________________________________________60
Table 4.7: The infrared assignments of unsubstituted PPV.
Wavenumber [cm-1] Assignment References
970 trans-vinylene CH-wag [Bradley’87]
1503 semicircular phenyl stretch [Scott’86]
1600 C-C ring stretch [Bradley’87]
1672 C=O aromatic aldehyde [Bradley’87, Scott’86]
1740 ester or carboxylic acid [Cumpston’95]
Though the sensitivity of PPVs to UV light is a practical problem, it can be used
favorably, on the other hand, for structuring purposes. The changes of the refractive index
and mechanical stability of the film after UV irradiation give the opportunity for
microstructuring of the polymer slab waveguides by means of UV-photobleaching
technique. This will be discussed later in Chapter 7.
4.6 Materials Comparison and Conclusion
On the basis of the above studies, we summarize in Table 4.8 the important
properties of all PPVs studied to select the most promising polymer for nonlinear
waveguide devices. Clearly, the materials for high-speed all-optical switching based on
slab waveguides need to be multifunctional. They have to exhibit large cubic nonlinearity
to achieve fully reversible and ultrafast changes of the refractive index. Furthermore, they
must be suitable for thin film processing and should form waveguides with propagation
losses of the guided waves αgw < 1 dB/cm. In addition, they should also show high
photostability.
_____________________________________________________________________61
Table 4.8: Summary of the important properties of PPV-derivatives studied for all-optical switching applications.
)3(maxχ [10-11 esu]
± 10 %
αgw (TE0) at 1064 nm [dB/cm]
SUV [a.u.]
MEH-PPV 6.50 0.5 ± 0.3 0.007
MEH-DOO-PPV 8.60 0.8 ± 0.3 0.011
MEH-M3EH-PPV 9.65 12.4 ± 1.4 0.005
M3EH-PPV 13.80 > 20 0.004
TPD2-PPV 3.70 > 20 0.013
TPD4-PPV 3.43 3.0 ± 1.2 0.011
As can be seen in Table 4.8, the conjugated polymers MEH-PPV and its related
structures are more appropriate candidates for all-optical switching applications, because
they exhibit larger value of cubic nonlinearity )3(maxχ as compared with TPD containing
PPVs. For waveguide applications, the waveguide propagation loss coefficient αgw < 1
dB/cm is also required. This could only be satisfied by MEH-PPV and MEH-DOO-PPV.
However, MEH-DOO-PPV is more sensitive to the UV light than MEH-PPV. Therefore,
we conclude that MEH-PPV is the most appropriate polymer studied here for all-optical
switching applications because it shows the superior combination of high cubic
nonlinearity χ(3), low waveguide propagation losses αgw and relatively low sensitivity to
UV photobleaching. The other polymers still appear to be quite useful if special material
properties need to be optimized. In the next Chapter, we will present detailed studies of
the nonlinear optical properties of MEH-PPV by means of intensity dependent prism
coupling and multiphoton excited fluorescence.
_____________________________________________________________________62
5 Nonlinear Refractive Index and Multiphoton Absorption of MEH-PPV
It was shown in Chapter 4 that MEH-PPV appears to be the most suitable
candidate for all-optical switching applications. Therefore, the nonlinear refractive index
of MEH-PPV was measured in order to obtain the figures of merit W and T, as defined in
Chapter 1, in a broad spectral region. The polymer was provided by Prof. H.-H. Hörhold
(University of Jena, Germany). It was synthesized by using the Horner-type
polycondensation route. The molecular weights were Mw = 4.03 x 104 g/mol and Mn =
1.41 x 104 g/mol as determined with gel permeation chromatography (GPC) using
polystyrene standards and THF as eluent [Pfeiffer’99a,b]. We have used the intensity
dependent prism coupling method to measure the nonlinear absorption coefficient α2 and
nonlinear refractive index n2 of MEH-PPV films in the near infrared (NIR) range.
Furthermore, we have performed multiphoton excited fluorescence experiments of MEH-
PPV solutions in order to improve the accuracy of the α2 data from prism coupling and to
study the two- and three-photon absorption resonances and their possible influences on
the nonlinear refractive index spectra. Finally based on the measured values of α2(λ),
n2(λ), and αgw(λ), we have evaluated the dispersion of W(λ) and T(λ) to find spectral
regions that fulfill the requirements for all-optical switching applications.
5.1 Nonlinear Prism Coupling
A typical set of intensity dependent prism coupling curve is shown in Fig. 5.1. The
angular dependence of the light reflected at the prism base shows a minimum (θm) at a
certain coupling angle if a waveguide mode is excited (see Chapter 3). The laser beams
were s-polarized to excite the TE0 waveguide mode. The reflected light intensity IR(θ)
shown in Fig. 5.1 was normalized and measured as a function of angle incidence θ. The
following procedure was used. First, the measurement was performed at low intensity to
evaluate the coupling parameters: air-gap thickness da, air-gap length la and incident beam
diameter db which are introduced in Chapter 3. Second, for sensitive detection of
intensity-dependent changes of absorption ∆αf and refractive index ∆nf, the air-gap da was
adjusted for a slightly overcoupled condition (see Chapter 3 for the definition of the
_____________________________________________________________________63
coupling conditions). The incident laser beam was focused on the prism base to a
diameter db few times smaller than air-gap length la. Finally, the normalized reflected
intensity IR(θ) was measured with gradually increasing energy of incident laser pulses.
The minimum of the resonance coupling angle θm shifts with increasing energy of the
incident laser pulses. This is directly related to a change of refractive index nf of the film.
The resonance curve also becomes deeper due to an increase of the absorption coefficient
αf of the film. The intensity-dependent shifts of the coupling curves were fully reversible
at the input laser pulse energy below a certain limiting value which was approximately 3
µJ for 1064 nm < λ < 1600 nm and < 1 µJ for λ < 1064 nm.
-10.2 -10.0 -9.8 -9.6 -9.4 -9.20.6
0.7
0.8
0.9
1.0λ = 1064 nm
0.1 µJ 1.0 µJ 2.0 µJ
I R [a
.u.]
θ [0]
Fig. 5.1: Prism coupling of a MEH-PPV film on a fused silica substrate (film thickness 590 nm, laser wavelengths λ = 1064). The TE0 mode was excited at different pulse energies given in the inset. Symbols represent the experimental data and lines are the numerical fits. The air-gap thickness da = 340 nm was held constant. Only nf and αf of the film were varied to fit the measurements at different energies (0.1 µJ: nf =1.65925, αf = 0.2 cm-1; 1.0 µJ: nf =1.65960, αf = 14 cm-1; 2.0 µJ: nf =1.6603, αf = 17.4 cm-1)
Intensity induced changes of refractive index and absorption coefficient are not
uniform everywhere in the film, but follow a bell-shaped distribution of the guided wave
intensity in the mode-propagation direction and a cos2 distribution (for the TE0 mode) in
the direction perpendicular to the plane of the film, respectively. Nevertheless, it is
_____________________________________________________________________64
possible, as an approximate approach, to replace this distributed change of nf and αf in the
case of small nonlinearities and moderate input intensities with uniform changes of the
form
gw2f Inn =∆ (5.1)
gw2f Iα=α∆ (5.2)
The average guided wave intensity gwI is calculated by means of Eq. (3.12).
0 1 2 3 4 5 6 7-0.3
0.0
0.3
0.6
0.9
1.2 (a)
λ = 830 nm
λ = 1064 nm
λ = 1064 nm
∆αf [
cm-1]
<Igw> [GW/cm2]
∆nf [
10-3]
0 1 2 3 4 5 6 70
4
8
12
16
20 (b)
λ = 830 nm
Fig. 5.2: Changes in (a) refractive index ∆nf and of (b) absorption coefficient ∆αf at laser wavelengths 830 nm (open circles) and 1064 nm (full squares) as a function of the average intensity in the waveguide gwI .
_____________________________________________________________________65
The last step for determination of n2 and α2 is to make numerical fits to
experimentally measured coupling curves. In this approximate approach the intensity
dependencies of βm(Igw) and αgw(Igw) in Eq. (3.4) are neglected, and the fits in Fig. 5.1 are
calculated by use of two parameters only, nf and αf. Finally, we calculate n2 and α2 from
the intensity-dependent changes in ∆nf and ∆αf using Eq. (5.1) and Eq. (5.2). Fig. 5.2
shows examples of the changes in refractive index ∆nf and absorption coefficient ∆αf at
wavelengths 830 nm and 1064 nm as a function of average intensity in the film gwI . A
remarkably large ∆nf up to 10-3 with complete reversibility of the coupling curves was
observed at 1064 nm.
The intensity dependent prism coupling experiments were performed in a broad
interval of laser wavelengths 680 nm < λ < 1600 nm. The spectral dependence of the
linear waveguide losses αgw, nonlinear absorption coefficient α2 and nonlinear refractive
index n2 are shown in Fig. 5.3. The spectrum of αgw that results from linear absorption α0
plus light scattering losses, is shown in Fig. 5.3(a). The full squares represent results
calculated from low-intensity coupling curves of prism coupling experiments. Additional
data (open diamonds) are shown for comparison that were obtained by separate
waveguide attenuation loss measurements at the wavelengths of HeNe and Nd:YAG
lasers [Fitrilawati’02]. The results from both experimental methods are in very good
agreement. The value of αgw is large at shorter wavelengths due to the tail of the
electronic absorption band and it decreases towards the NIR and becomes as low as 0.5
dB/cm for λ > 1050 nm. At longer wavelengths αgw ≤ 0.5 dB/cm was always observed.
But αgw cannot be specified more precisely because the detection limit of the
experimental method was approached.
The spectral dependence of nonlinear absorption coefficient α2 is shown in Fig.
5.3(b). Very high values of α2 were observed at wavelengths around 800 - 850 nm that
are caused by strong two-photon absorption. We were not able to measure α2 for λ < 800
nm, because of the increasing linear absorption, which makes it difficult to determine α2.
At λ > 1000 nm, α2 decreases strongly and does not show any other significant resonance.
At longer wavelengths than 1000 nm the value of α2 is smaller than 10-8 cm/W.
The spectrum of nonlinear refractive index n2 is shown in Fig. 5.3(c). Negative n2
values were observed at λ < 980 nm. Around 980 nm, n2 is zero and becomes positive
towards longer wavelengths. The maximum of n2 = (2.2 ± 0.4) x 10-13 cm2/W was
_____________________________________________________________________66
observed at λ = 1080 nm. At wavelengths larger than 1100 nm, the n2 values decrease
monotonically. The same spectra for λ ≥ 900 nm are displayed in Fig. 5.4 in an expanded
scale. At λ ≥ 1200 nm, the values of α2 are zero because no changes in absorption with
increasing incident laser intensities were detectable.
0
4
8
12
16
20(a)
αgw
[cm
-1]
600 800 1000 1200 1400
-8
-6
-4
-2
0
2 (c)
n 2 [10-1
3 cm2 /W
]
λ [nm]
0.0
0.1
0.2 (b)
α2 [1
0-6cm
/W]
Fig. 5.3: Comparison of linear and nonlinear optical spectra of MEH-PPV measured at laser wavelengths λ. (a) Linear attenuation loss αgw(λ) determined from coupling curves at low intensity (full squares) and from waveguide loss experiments (open diamonds). Dispersions of α2 and n2 are shown in (b) and (c).
_____________________________________________________________________67
0.0
0.5
1.0
1.5
2.0(a)
αgw
[cm
-1]
1000 1100 1200 1300 1400 1500
0
1
2(c)
n 2 [10-1
3 cm2 /W
]
λ [nm]
0.00
0.01
0.02
0.03(b)
α2 [1
0-6cm
/W]
Fig. 5.4: Comparison of linear and nonlinear optical spectra of MEH-PPV measured at the laser wavelengths λ in the range of 900 to 1500 nm. (a) Linear attenuation loss αgw(λ) determined from coupling curves at low intensity. Dispersions of α2 and n2 are shown in (b) and (c).
_____________________________________________________________________68
5.2 Multiphoton Excited Fluorescence
The nonlinear prism coupling method is not very sensitive to measure the
nonlinear absorption coefficient α2. In order to control and refine the α2 data of films, we
have performed multiphoton excited fluorescence experiments of MEH-PPV in toluene
solution. Furthermore, we have used this method to gain additional information on the
observed resonance in the α2 and n2 spectra of MEH-PPV from prism coupling
experiment. The linear absorption and fluorescence spectra of MEH-PPV in highly
diluted solution are displayed in Fig. 5.5. Both spectra were measured by means of SPEX
Fluorolog 2 (F212) spectrometer. The fluorescence spectra were measured for different
concentrations of the solutions. The fluorescence spectra consist of two peaks: the main
peak at 560 nm and a shoulder at IFL (600 nm). The spectra were normalized at 600 nm.
The main peak at 560 nm decreases with increasing concentration, which indicates the
reabsorption effect due to strong overlap between the absorption and emission spectra at
wavelengths 500 - 560 nm.
400 450 500 550 600 650 7000.0
0.2
0.4
0.6
0.8
1.0
I FL [
a.u.
]
A [a
.u.]
λ [nm]
0
1
2
3
C1 < C2 < C3
C1 C2 C3
Fig. 5.5: Linear absorption and fluorescence spectra of MEH-PPV in toluene solution excited at 532 nm. The fluorescence spectra were measured at different concentrations shown in the inset. Multiphoton excited fluorescence experiments were performed using a solution of
MEH-PPV in toluene (concentration: 3% by weight). The experimental setup was shown
_____________________________________________________________________69
in Fig. 3.14 (Chapter 3). The fluorescence signal was detected at the front face of the
cuvette. Typical spectra of one-, two- and three-photon excited fluorescence are shown in
Fig. 5.6. For one-photon fluorescence (λExc = 532 nm), we have observed two peaks: the
main peak at 560 nm and a shoulder at 600 nm, similar to that of highly diluted solution
(Fig. 5.5). For two- and three-photon excited fluorescence, however, only the shoulder at
600 nm was observed, while the main peak at 560 nm was not visible. This can be
explained by the reabsorption effect as mentioned above. For one photon excited
fluorescence, the exciting light does not penetrate deeply into the solution and it is
absorbed quickly. The fluorescence signal is generated only in the small volume of the
cell close to the front face and it could reach the detector without being completely
reabsorbed. Therefore, the main peak at 560 nm still can be observed. For two- and three-
photon excited fluorescence, on the other hand, the penetration depth of exciting light is
large, resulting in large volume of excited chromophores. As consequence, the
reabsorption effect is very strong and therefore, the main peak at 560 nm disappears.
500 550 600 650 7000.0
0.2
0.4
0.6
0.8
1.0
λ [nm]
Excitation: 1-photon λExc= 532 nm 2-photon λExc= 825 nm 3-photon λExc= 1500 nm
I Fl [a
.u.]
Fig. 5.6: Multiphoton fluorescence spectra of a 3 % of concentration by weight of MEH-PPV in toluene: one-photon (full squares), two-photon (full circles) and three-photon excited fluorescence (open triangles). The excitation wavelengths are given in the inset.
The fluorescence intensity was measured as a function of excitation laser
intensity. In order to avoid the reabsorption effect, we have detected the fluorescence
_____________________________________________________________________70
intensity at 600 nm. Fig. 5.7 shows the typical results for different excitation wavelengths
λExc. At laser excitation wavelengths between 750 nm and 1200 nm, which is about two
times of the linear absorption peak (λmax = 498 nm), the fluorescence intensity scales with
the square of the intensity of the exciting pulses. This is a clear confirmation of a two-
photon absorption process. For λExc ≈ 3λmax [Fig. 5.7(b)], the cubic dependence was
observed which proofs that three-photon absorption occurs.
Fig. 5.7: Fluorescence intensity as a function of excitation intensity at different excitation wavelengths for (a) two-photon excitation and (b) three-photon excitation.
In Fig. 5.8(a), the fluorescence intensity is plotted as a function of the square of
the incidence laser intensity. A clear linear dependence is observed with the line slope
1 2 4 6 8 10 20
0.1
1
I Fl, 6
00nm
[a.
u.]
IExc [a.u.]
(a)
1.951.92
1.861.82
850 nm 900 nm 950 nm 1000 nm
1 2 3 4 5 6
0.1
1(b)
IExc [a.u.]
I Fl, 6
00nm
[a.
u.]
2.94
2.84
3.12
2.90
1500 nm 1520 nm 1530 nm 1600 nm
_____________________________________________________________________71
proportional to the imaginary part of the cubic nonlinearity Im{χ(3)}. The spectra of
Im{χ(3)} of solutions are displayed in Fig. 5.9(b) together with the α2 data of films from
prism coupling experiments. The scale of Im{χ(3)} was adjusted for good comparison
with α2. Both results are in very good agreement.
Fig. 5.8: Fluorescence intensity versus square of excitation intensity (a) and cubic excitation intensity (b). In the case of three-photon absorption, the imaginary part of “quintic”
nonlinearity Im{χ(5)} is obtained from the slope of fluorescence intensity as function of
cubic excitation intensity [Fig. 5.8(b)]. The dispersion of Im{χ(5)} is displayed in Fig.
5.9(c). We have no data for λ ≤ 1300 nm because of the strong influence of two-photon
absorption, which is difficult to be separated due to simultaneous occurrence of two- and
0 20 40 600.0
0.5
1.0
1.5(b)
~ Im χ(5)
λ=1510 nm
I Fl, 6
00nm
[a.
u.]
I3Exc [a.u.]
0 5 10 15 20 250.0
0.3
0.6
0.9
1.2
(a)
~ Im χ(3)
λ=850 nm
I2Exc [a.u.]
I Fl, 6
00nm
[a.
u.]
_____________________________________________________________________72
three-photon processes. At excitations with λ > 1600 nm, the fluorescence intensity was
too weak to be detected. Two peaks at approximately λ1 = 1450 nm and λ2 =1520 nm
were observed.
350 400 450 500 550 6000.0
0.2
0.4
0.6
0.8
1.0 (a)
Im {χ
(5) } [
a.u.
]
α2 [
10-6 c
m/W
]
Im { χ
(3) } [
a.u.
]
Abso
rban
ce [a
.u.]
700 800 900 1000 1100 1200
0.00
0.05
0.10
0.15
0.20 (b)
0.00
0.05
0.10
0.15
0.20
1050 1200 1350 1500 1650 18000.00
0.02
0.04
0.06(c)
λ [nm]
Fig. 5.9: Comparison of linear and nonlinear optical absorption spectra of MEH-PPV in toluene. (a) Linear absorption coefficient α(λ) measured by means of UV-Vis-NIR transmission spectroscopy. (b) Dispersions of Im{χ(3)}of solution (full circles) and α2 of the film (open squares). (c) Dispersion of Im{χ(5)}.
_____________________________________________________________________73
5.3 Discussion
5.3.1 Control Experiments of Prism Coupling Data
The waveguide loss coefficients αgw from prism coupling experiments at low-
intensity are compared with the stray light detection experiments of Fitrilawati et al
[Fitrilawati’02]. Both results are in very good agreement. This confirms the validity of
our numerical model of the prism-film coupler in general and the appropriate description
of the air-gap in particular. Moreover, it shows that prism coupling can be used
successfully as a method for the determination of waveguide losses. The αgw of MEH-
PPV waveguides in the NIR region are unprecedented low (≤ 0.5 dB/cm). This means,
first, the linear absorption is low at these wavelengths and, second, we were able to
process MEH-PPV to high-quality waveguides with the required low attenuation losses
for practical applications.
The spectrum of α2 of films as evaluated from prism coupling experiment agree
very well with the spectrum of Im{χ(3)}of solution. This means that the chromophores of
MEH-PPV behave very similar in solid state (film) as compared to diluted solution. No
strong spectral changes due to aggregate formation in films are observed. Furthermore,
because the experimental error of Im{χ(3)} is smaller than that of α2, these values can be
used to evaluate the figures of merit with reduced experimental error as shown in Chapter
5.3.4.
5.3.2 Interpretation of Third-Order Nonlinear Optical Spectra
We interpret the main spectral features of the dispersions of α2 and n2 as a result
from two-processes: saturable absorption and two-photon absorption. The α2 and n2
spectra are correlated. Fig 5.10 shows the calculated spectra of α2 and n2 for three-level
system which consists of one even-parity ground state (g) and two excited states (u, g’)
with odd and even parities [Ueberhofen’99].
One-photon absorption is a transition from the ground-state g to the excited state
u. At high intensity, the excited state becomes more populated, resulting in the decrease
of the absorption coefficient. As consequence, α2 is negative. It has maximum at
wavelength λ1 = hc/Eug. The n2 is positive at the higher energy side (λ < λ1) and negative
_____________________________________________________________________74
at the lower energy side (λ > λ1) of the resonance. The second process is two-photon
absorption, which is the transition from the ground state g to the lowest excited state with
even parity g’[see Fig. 5.10(b)]. It is characterized by α2 > 0 and a maximum at λ2 =
2hc/Egg’. The n2 is negative at the higher energy side and positive at lower energy side of
the resonance. In the non-resonant region at long wavelengths, both processes compete.
Therefore, the sign of n2 is determined by the dominant mechanism.
Fig. 5.10: Calculated spectra of α2 and n2 of (a) one-photon and (b) two-photon resonances with a three-level system. The energies of the states g, u and g’ were chosen arbitrary [Ueberhofen’99].
The broad maximum of the α2 spectrum of MEH-PPV at approximately λ = 830
nm, which corresponds to an energy of 3.0 eV, is caused by two-photon absorption. This
energy level is in very good agreement with results of Martin et al, who reported the
energy of the lowest two-photon excited state of MEH-PPV at 2.9 eV [Martin’99].
A schematical plot of the n2 dispersion of a three-level system over a broad energy
range is shown in Fig. 5.11(a) [Stegeman’97]. At larger energies, the spectral features of
n2 are caused by saturable absorption. At lower energies, both the saturable absorption
and the two-photon absorption play an important and competing role. For comparison, we
have plotted our n2 spectrum of MEH-PPV as a function of energy in Fig. 5.11(b).
_____________________________________________________________________75
Clearly, both spectra are similar which indicates that MEH-PPV behaves in general as a
three-level system. The negative n2 values in the range of 1.3 to 1.8 eV are assigned to
saturable absorption. At E ≤ 1.2 eV, the positive sign of n2 is caused by the dominance of
two-photon absorption. No additional resonance of n2 at lower energy than 1 eV was
observed.
Fig. 5.11: (a) Schematical plot of the dispersion of nonlinear refractive index n2 for three-level system [Stegeman’97]. (b) Dispersion of n2 of MEH-PPV film obtained from intensity dependent prism coupling. 5.3.3 Three-Photon Absorption Effects
Three-photon absorption could result from transitions between states with
opposite parity only [Pantell’67]. Therefore, it occurs as a transition from the ground state
0.8 1.0 1.2 1.4 1.6 1.8
-8
-6
-4
-2
0
2 (b)
Energy [eV]
n 2 [1
0-13 cm
2 /W]
_____________________________________________________________________76
(g) to the lowest excited state (u), similar to one-photon absorption. The three-photon
absorption coefficient α3 is proportional to the imaginary part of fifth-order nonlinear
optical susceptibility, Im{χ(5)}, which can be described in a with very simplified way
[Stegeman’00]:
,...),(f)i3(
Im~ gugugu
6gu
3 ωω
Γ−ω−ω
µα . (5.3)
where f(ω, ωgu,…) is a slowly varying function of its variables. Clearly, α3 has a
maximum at 3ω = ωgu, where ωgu is the corresponding transition frequency between the
ground state g and excited state u. Based on the linear absorption of MEH-PPV, three-
photon absorption occurs at wavelengths around 3 x λmax of linear absorption (λ =1500
nm. However, two resonances at λ1 = (1452 ± 2) nm and λ2 = (1532 ± 2) nm were
observed (see Fig. 5.12). This splitting corresponds to the wavenumber difference PA3~ν∆
= (1079 ± 38) cm-1. The energy of this splitting is typical for molecular vibrations.
Therefore, the shape of absorption spectrum of MEH-PPV is inspected. It is shown in Fig.
5.13.
1300 1400 1500 1600 17000.00
0.01
0.02
0.03
0.04
0.05
Im {χ
(5) } [
a.u.
]
λ [nm]
Fig. 5.12: Dispersion of Im χ(5) of MEH-PPV in toluene solution (full squares). Dashed lines indicate the Gaussian fits for determining the peaks.
_____________________________________________________________________77
In order to resolve the spectral positions of the maximum and the shoulder at
longer wavelengths, we have calculated the second derivative of the linear absorption
spectrum. The result is displayed also in Fig. 5.13. Two minima at wavelengths 492 ± 2
nm and 526 ± 2 nm were observed. The difference between them is A~ν∆ = (1314 ± 110)
cm-1. As the value of A~ν∆ is close to PA3
~ν∆ , we assume that the splitting in the three-
photon absorption spectrum could be related to a vibronic state. In principle we have to
expect that the peak positions of the linear and α3-absorption are related. Therefore, we
expect that the peaks of α3 occur at 1476 ± 6 nm (3 x 492 nm) and 1578 ± 6 nm (3 x 526
nm). However, we have observed these peaks at 1452 ± 2 nm and 1532 ± 2 nm. This
small disagreement could be interpreted by an ageing of the solution in the 3-photon
absorption experiment, causing a blue shift.
400 450 500 550 6000.0
0.2
0.4
0.6
0.8
1.0
1.2
A [a
.u.]
λ [nm]
526 nm
492 nm d2 A/dλ
2 [a.u
.]
Fig. 5.13: Spectrum of linear absorption coefficient of MEH-PPV in toluene solution (solid line) and its second derivative (dashed line). Two minima at 492 nm and 526 nm were observed.
_____________________________________________________________________78
Recently, Yoshino et al [Yoshino’03] have measured the spectra of three-photon
absorption coefficient α3 of 200 µm thick single crystals of PTS [polymer poly (bis para-
toluene sufonate] by means of Z-scan. Similar to our results for MEH-PPV, they observed
a strong oscillation in the α3 spectra at 1900 < λ < 2000 nm, which is about three times of
its linear absorption (λmax = 620 nm). However, the reason for this oscillation is still
unclear. Further theoretical and experimental studies are necessary to interpret and
understand the observed splitting in the three-photon absorption spectrum of MEH-PPV.
5.3.4 Figures of Merit
The measured values of αgw, α2 and n2 from intensity dependent prism coupling
were used to calculate the spectral dependences of the figures of merit (FOMs) W and T
as defined in Chapter 1. We have used the total waveguide loss coefficient αgw instead of
α0 and a light intensity I = 1 GW/cm2 as in earlier estimates of W [Stegeman’93]. The
values of Im{χ(3)}were used to refine the calculation of T in the NIR region, because their
experimental error is smaller than that of α2. The dispersions of W and T are shown in
Fig. 5.14.
At short wavelengths, W is close to unity due to strong absorption. At longer
wavelengths, the value of W increases due to the increase of n2. At 1080 nm, the
maximum value W = 22 is observed, according to the maximum of the nonlinear
refractive index. At λ > 1100 nm, W decreases monotonically with increasing
wavelength. The dispersion of T is displayed in Fig. 5.14(b). It shows high values at short
wavelengths owing to the strong two-photon absorption. At 980 nm, the value of T grows
to infinity since n2 is zero at this wavelength [see Fig. 5.3(c)]. Towards longer
wavelengths T decreases and becomes < 1 at 1100 nm < λ < 1200 nm. No data exist at λ
> 1200 nm, because the α2 values were below the detection limits of our experimental
method.
_____________________________________________________________________79
800 1000 1200 1400
0.1
1
10
100
1000 (b)
W
T
λ [nm]
800 1000 1200 1400-5
0
5
10
15
20
25(a)
Fig. 5.14: Spectral dependences of FOMs (a) W and (b) T of MEH-PPV waveguides. The arrow in (b) indicates the divergence of T because of n2 = 0 at λ = 980 nm.
It is interesting to compare the figures of merit W and T of MEH-PPV with other
possible candidates for all-optical switching applications. A survey of FOMs of various
organic materials has been reported [Stegeman’93, Stegeman’96]. Although many
organic materials have been investigated by numbers of techniques, such as third
harmonic generation, degenerate four-wave mixing, nonlinear prism coupling, Z-scan,
nonlinear transmission etc., only several of them are studied well enough to evaluate the
figures of merit. Moreover, different definitions of W and T were used in the literature,
which leads to some confusion [Stegeman’93, Bredas’94, Gubler’02]. A factor of 2
appears sometimes either in the nominator or denominator of W and T. In order to have a
better comparison, therefore, we have recalculated the reported FOMs W and T using a
_____________________________________________________________________80
specific definition as given in Chapter 1 [Stegeman’93]. The values of W and T are listed
in Table 5.1.
Table 5.1: Recalculated of figures of merit (FOMs) of various organic materials by means of Eqs. (1.1) and (1.2). The intensity was assumed by I = 1 GW/cm2.
Material λ [nm]
n2
[cm/W] α
[cm-1]
α2
[cm/GW]
W T Ref.
PTS (crystal) 1300
1600
3.2 x 10-12
2.2 x 10-12
-
< 0.8
20
-
> 10
1.6
< 0.1
1
2
Poly-4BCMU 1319 4.8 x 10-14 1.8 < 0.25 0.2 ~ 1.4 3
DANS 1064
1319
7.0 x 10-14
8.0 x 10-14
~ 3
0.4
0.8-2
< 0.08
0.2
1.5
2-6
< 0.26
4
5
DAN2 1064 1.9 x 10-13 4 1-2 ~ 0.4 1-2 4
PPV 800
925
> 10-11
5.0 x 10-12
-
9
80
-
-
6
~1.3
-
6
7
DPOP-PPV 880-920
970
~ 1.5 x 10-14
1.0 x 10-14
< 0.35
< 0.35
~ 0.14
0.02
~ 0.5
0.3
~ 1.7
~ 0.3
8
PPV(RCo 52) 880-920
970
3 - 4x10-14
3.0 x 10-14
< 0.35
< 0.35
~ 0.3
0.1
> 1
0.9
> 1
~ 0.6
8
MEH-PPV 1080
1130
1200
2.2 x 10-13
8.0 x 10-14
2.0 x 10-14
0.1
0.1
0.1
1.4
0.13
0.1
20.4
7.1
1.7
1.4
0.4
0.12
This work
Reference numbers refer to: 1- Kim’94, 2- Lawrence’94, 3- Rochford’91, 4- Marques’91, 5- Kim’93, 6- Samoc’95, 7- Ueberhofen’99, 8- Gabler’98.
The figures of merit of single crystals of PTS [polymer poly(bis para-toluene
sulfonate) have been reported in the broad wavelength region between 1000 nm and 1600
nm [Kim’94, Lawrence’94, Stegeman’00]. The domination of two-photon absorption at
shorter wavelengths leads to T > 1. At 1600 nm the values of W > 1 and T < 1 were
_____________________________________________________________________81
obtained. The nonlinearities of solution processed films of poly-4BCMU were reported
two orders of magnitude smaller than that of PTS. The large scattering losses from
waveguide imperfections of this polymer lead to the small value of W [Rochford’91].
The n2 and α2 of side-chain polymers DANS (4-dialkylamino-4′-nitro-stilbene)
and DAN2 (4-dialkylamino-4′-nitrodiphenylbutadiene) waveguides were measured by
means of nonlinear grating coupling at wavelength 1064 nm [Marques’91]. The value of
T for both materials was above one because of the strong two-photon resonance. At λ =
1319 nm, however, the FOMs requirements for all-optical switching (W > 1 and T < 1)
were fulfilled [Kim’93].
The nonlinearities of unsubstituted PPV were well studied in the past [Samoc’95,
Ueberhofen’99]. It exhibits high values of n2, but its linear losses caused by the well-
known polycrystalline morphology are too large for waveguide applications
[Ueberhofen’99]. The soluble PPV-derivatives, however, have stimulated much interest
due to their good waveguide-forming properties and high third-order nonlinearities
[Bartuch’92, Michelotti’95, Gabler’97, Gabler’98, Bubeck’00, Fitrilawati’02,
Koynov’02, Bader’02]. By Mach-Zehnder interferometer and spectral broadening
experiments, the FOMs of DOO-PPV and copolymer RCo 52-PPV were measured in the
range of 780 to 980 nm. T > 1 was observed at 780 ≤ λ ≤ 920 nm, due to strong two-
photon absorption. T decreases significantly, however, when λ was detuned from two-
photon resonance [Gabler’97, Gabler’98]. These trends are consistent with our results for
MEH-PPV.
As shown in Table 5.1, only three materials fulfill the FOMs requirement for all-
optical switching applications, i.e. PTS crystal, DANS and MEH-PPV. The difficulties to
prepare large size of single crystals PTS with high optical quality, however, prevents this
material for any waveguide application. Our MEH-PPV exhibits better combination of
both W and T as compared to that of DANS polymer. Therefore, we conclude that MEH-
PPV is the best presently known candidate for all-optical switching applications, in
particular at wavelengths around 1100 nm.
_____________________________________________________________________82
6 Influence of Molecular Weight on the Properties of MEH-PPV
We have already shown in Chapter 4 and 5 that MEH-PPV is the most suitable
candidate for all-optical switching applications. Furthermore, MEH-PPV is frequently
used as prototype material for the development of photonic devices such as light-emitting
diodes [Braun’91, Scott’96, Friend’99], plastic lasers [McGehee’00, Turnbull’03,
Gaal’03] and photovoltaic cells [Sariciftci’93, Yu’95, Petrisch’99]. These applications
require thin films with thicknesses in the range of 50 - 800 nm, which are commonly
prepared by spin coating. Such films of conjugated polymers frequently exhibit uniaxial
anisotropy due to preferred alignment of the polymer chains parallel to the film plane
[McBranch’95, Becker’97]. Because the main electric polarizability and transition dipole
moment of the conjugated π−electron system are parallel to the chain direction, the
anisotropic orientation of chain segments is strongly correlated with significant
birefringence of the films, i.e., refractive indices at transverse electric (TE) and transverse
magnetic (TM) polarizations differ considerably with nTE > nTM. The design of devices
clearly needs precise knowledge of the optical constants of thin films, in particular their
birefringence.
However, reports on basic optical properties like refractive index of thin films of
MEH-PPV show significant disagreements, as can be seen in Fig. 6.1 [Boudrioua’00,
Kranzelbinder’02, Tammer’02, Bader’02, Fitrilawati’02]. This problematic inconsistency
might be caused by the large variations of the number-average molecular weight (Mn) and
weight-average molecular weight (Mw) of the different MEH-PPV samples studied
[Bahtiar’02, Koynov’02]. Most reports of optical properties of MEH-PPV did not pay
appropriate attention to the influence of Mn and Mw and frequently, their values are not
provided at all. It is well recognized that the optoelectronic properties of thin polymer
films are quite sensitive to chain packing morphology [McGehee’00]. In this Chapter we
provide experimental evidence that molecular weight of MEH-PPV has profound
influence on morphology, in particular on the alignment of the polymer chains in thin
films and consequently on the optical birefringence, waveguide propagation losses, and
nonlinear optical properties.
_____________________________________________________________________83
550 600 650 7001.6
1.8
2.0
2.2
2.4 Kranzelbinder et al. Boudrioua et al. Tammer et al. Fitrilawati et al. Bader et al.
n TE
λ [nm]
Fig. 6.1: Disagreements of the reported refractive indices of MEH-PPVs reported at TE- polarization by Kranzelbinder et al [Kranzelbinder’02], Boudrioua et al [Boudrioua’00], Tammer et al [Tammer’02], Fitrilawati et al [Fitrilawati’02] and Bader et al [Bader’02].
6.1 Materials
A large variation of the molecular weight of MEH-PPV can be accomplished by
different synthetic routes. The frequently used, so-called Gilch-type dehydrohalogenation
route [Gilch’66] yields high molecular weight polymers with Mw in the order of 105 to
106 g/mol. However, Mw can be reduced by appropriate choice of end-cappers in the
synthetic process. Another synthetic approach to MEH-PPV with very well defined chain
structure was realized by using the Horner-type polycondensation route which yields
polymers with Mw in the typical order of several 104 g/mol [Pfeiffer’99a, b]. We have
studied several MEH-PPVs with Mw in the range of 104 to 1.6 x 106 g/mol. Their
molecular weights were measured by gel permeation chromatography (GPS) using
polystyrene standards and THF as eluent. We obtained the polymers from different
sources (see Table 6.1).
_____________________________________________________________________84
Table 6.1: MEH-PPVs from different synthesis pathways and molecular weights were obtained from different sources. PDI is polydispersity index = Mw/Mn.
Polymer Route Mw [g/mol] Mn [g/mol] PDI Source
1 Gilch 9.3x103 4.8 x103 1.9 ADS, Canada
2 Horner 1.3 x104 6.4 x103 2.0 Ahn, MPIP
3 Horner 2.5 x104 9.1 x103 2.7 Hörhold, Jena
4 Horner 4.03 x104 1.41 x104 2.8 Hörhold, Jena
5 Gilch 1.28 x 105 2.46 x 104 5.2 Ahn, MPIP
6 Gilch 2.65 x105 8.71 x104 3.0 ADS, Canada
7 Gilch 2.76 x105 1.05 x105 2.6 Ahn, MPIP
8 Gilch 4.2 x105 1.08 x105 3.9 ADS, Canada
9 Gilch 1.6 x106 1.3 x105 12.3 Covion, Germany
6.2 Spin Coating
6.2.1 Results
All polymers were dissolved in toluene and processed to thin films by spin coating
of filtered solutions (0.45-2 µm syringe filters) onto fused silica substrates. They were
subsequently placed in a vacuum oven for about 6 hours at 50 0C to remove residual
solvent. The film thickness d was measured with a Tencor Model P10 step-profiler. Thin
films with thickness d ≈ 50 – 70 nm were used for spectroscopic studies, and thick films
(400 - 800 nm) for optical waveguide experiments. The film thickness was controlled by
variation of concentration Cw and spinning speed ω using the spin coating parameters as
defined by Eq. (3.1). The spin coating parameters for all polymers studied, except for
polymers 5 and 9, are listed in Table 6.2. We did not have the data for polymer 5 due to
small quantity of this polymer and we were not able to dissolve the polymer 9 at higher
concentration than 0.5 % by weight, because its extremely large Mw causes solubility
problems.
_____________________________________________________________________85
Table 6.2: Spin coating parameters of MEH-PPVs. The thickness d0 was calculated at concentration of 5% by weight and spinning speed 2000 rpm. (a) As reported by Fitrilawati et al [Fitrilawati’02]. The exponents α and β refer to Eq. 3.1.
Polymer 1 2 3(a) 4 6 7 8
α -0.52 -0.52 -0.50 -0.50 -0.49 -0.51 -0.50
β 1.72 1.50 1.61 1.57 1.79 1.80 2.57
d0 [nm] 334.4 235.7 542 380.0 1200.5 1485 -
As shown in Table 6.2, the parameter α is similar for all polymer studied.
However, β and d0 depend on the molecular weight. Fig. 6.2 shows the molecular weight
dependence of the calculated film thickness using the spin coating parameters listed in
Table 6.2 at 1% concentration and spinning speed 1000 rpm. For Mw < 3 x 105 g/mol, the
film thickness increases linearly with the molecular weight with the slope of 0.38. The
film thickness at Mw = 4.2 x 105 g/mol, however, deviates from the slope.
104 105 106101
102
0.38
d [n
m]
MW [g/mol]
Fig. 6.2: Dependence of film thickness on the molecular weight of MEH-PPVs. The film thicknesses were calculated using the parameters listed in Table 6.2 at 1% concentration by weight and spinning speed 1000 rpm.
_____________________________________________________________________86
6.2.2 Discussion
The spin coating process involves several factors including the hydrodynamics
and rheology of the polymer solutions, solvent mass transfer, surface and interfacial
phenomena, heat transfer and the interplay of these processes. A number of semi-
empirical models have been proposed to rationalize the experimental data with theory
[Meyerhof’78, Yonkoski’92]. The final film thickness can be varied by spinning speed ω
and concentration of the solution Cw, according to Eq. (3.1). It was also found that the
molar mass of polymer as measured by intrinsic viscosity η, has an effect on the film
thickness at constant rotation speed through the Eq. (6.1) with a constant γ [Pethrick’99].
γη~d (6.1)
For a given polymer-solvent system at a certain temperature, the intrinsic viscosity
is related to the viscosity-average molecular weight Mv through the Mark-Houwink-
Sakurada equation [Cowie’91]
av )M(K=η (6.3)
where K and a are constants, which can be established by calibrating with polymer
fractions of known molar mass. The value of “a” lies between 0.5 for a polymer dissolved
in a poor solvent and 0.8 in very good solvent. Because all MEH-PPVs studied here are
polydisperse, where the polydispersity of the polymer is characterized by the ratio of
Mw/Mn, the viscosity-average molecular weight Mv depends on the molecular weight
distribution of the polymer as defined by [Cowie’91]
( )[ ] a/1aiiv MwM ∑= (6.4)
where Mi is the molecular weight of the ith component and wi is the weight fraction of that
component. The value of Mv lies between Mn and Mw, but often close to Mw [Cowie’91].
Therefore, the final thickness of spin coated film becomes a function of the spinning
speed, concentration and viscosity through Eq. (6.5).
γβαω )M()C(~d ww (6.5)
_____________________________________________________________________87
A clear indication that the film thickness depends on the molecular weight is shown in
Fig. 6.2, with a constant γ = 0.38 for Mw < 4 x 105 g/mol. However, at Mw = 4.20 x 105
g/mol, the calculated thickness deviates from the slope, which might be related to the
polydispersity of the polymers. The polydispersity index (PDI) of polymer 8 is higher
than that of polymers 1 - 7 (see Tab. 6.1).
6.3 Linear Optical Constants
6.3.1 Results
The dispersions of the refractive index n(λ) and the intrinsic absorption coefficient
α(λ) of thin (d ≈ 50 nm) MEH-PPV films were determined by evaluation of transmission
and reflection spectra of thin films on quartz substrates (see Chapter 3). Spectra of the
intrinsic absorption coefficient α(λ) of all polymers 1 – 9 and selected polymers are
shown in Fig. 6.3(a) and 6.3(b), respectively. In the selected spectra [Fig. 6.3(b)], the
spectrum of α(λ) of polymer 6 is omitted for clarity, because it is very close to polymer 7.
The spectra of polymers 5 and 9 are also removed because of their different molecular
weight distribution or larger PDI as compared to the others (see Tab. 6.1). A systematic
trend to larger values of αmax with increasing Mw in spectra of polymers with similar PDI
is observed [Fig. 6.3(b)]. This trend is probably related to changes of the average polymer
chain orientation as will be discussed below. Respective spectra of the refractive index for
all polymers 1 – 9 and selected spectra are shown in Fig. 6.4(a) and (b), respectively.
These data correspond to nTE because light polarized parallel to the film plane was used in
the measurements. Similar to the intrinsic absorption coefficient, nTE increases for higher
molecular weight MEH-PPVs. The values αmax, wavelengths λmax of the maximum of the
main absorption band and refractive index at 633 nm are listed in Table 6.3 for all
investigated polymers.
_____________________________________________________________________88
Fig. 6.3: Spectra of linear absorption coefficient α(λ) of thin films of all MEH-PPVs 1 - 9 (a) and selected MEH-PPVs (b), measured at TE-polarization.
300 400 500 6000
4
8
12
16
20 (a) 8 9 76
53 4
21
α
[104 c
m-1]
300 400 500 6000
4
8
12
16
20
λ [nm]
(b) 8
7
3
421
α
[104 c
m-1]
_____________________________________________________________________89
Fig. 6.4: Dispersions of linear refractive index of thin films of all MEH-PPVs 1 – 9 (a), and selected MEH-PPVs (b), measured at TE-polarization.
500 600 700 8001.6
1.8
2.0
2.2
2.4 (a)
12
35
64
78
9
n
500 600 700 8001.6
1.8
2.0
2.2
2.4 (b)
12
34
78
n
λ [nm]
_____________________________________________________________________90
Table 6.3: Linear optical constants of thin film MEH-PPVs measured at TE polarization. (a) Data was taken from Fitrilawati et al [Fitrilawati’02].
Polymer λmax [± 3 nm] αmax [104 cm-1] ± 5 % nTE (at 633 nm) ± 5 %
1 472 13.2 1.7313
2 474 13.9 1.7099
3(a) 477 15.0 1.7782
4 489 16.6 1.8061
5 494 14.6 1.8021
6 489 18.2 1.8297
7 491 18.4 1.8611
8 495 19.1 1.9007
9 495 19.4 1.9010
In order to study a possible anisotropy in the plane of the film (lateral orientation),
the absorption spectra of the films were measured by means of UVVIS-NIR transmission
spectroscopy for different polarizations of the incident beam (0 to 900). No lateral
anisotropy, however, was observed for all films in the central area of the samples.
Prism coupling (m-line technique) was used to measure the refractive index at
both transverse electric polarization, nTE and transverse magnetic polarization, nTM by
changing the polarization of the incident beam (see Chapter 3). MEH-PPV waveguides
with typical thicknesses ranging from 400 - 800 nm were studied. The dispersions of nTE
and nTM at laser wavelengths between 633 nm and 1100 nm, together with results of
transmission-reflection measurements for the MEH-PPVs 4, 6 and 8 are shown in Fig.
6.5. A very pronounced influence of Mw on the refractive index and its birefringence (nTE
- nTM) is observed. Furthermore, the results of prism coupling and transmission-reflection
measurements agree very well. This shows that nTE does not significantly depend on the
film thickness, at least for 50 nm < d < 800 nm.
_____________________________________________________________________91
600 700 800 900 1000 1100
1.6
1.8
2.0
8 6 4
n
λ [nm]
Fig. 6.5: Dispersions of refractive indices of MEH-PPVs. Data points are from prism coupling experiments at TE polarization (full symbols) and TM polarization (open symbols). Lines are from transmission-reflection experiments at TE polarization. The numbers refer to the kind of MEH-PPVs and their molecular weight (4: Mw = 4.03 x 104 g/mol, 6: Mw = 2.65 x 105 g/mol, 8: Mw = 4.2 x 105 g/mol).
Fig. 6.6 shows nTE and nTM values at λ = 633 nm as a function of Mw for all
polymers studied. As mentioned above, we were not able to prepare a waveguide (film
thickness larger than 150 nm) of polymer 9 due to solubility problems. Therefore, nTE of
this polymer was obtained by reflectometry only [see Fig. 6.4(a)]. The nTM data of this
polymer was extrapolated from recent report of Tammer et al, [Tammer’02] who
measured the birefringence of the same polymer by means of variable angle spectroscopic
ellipsometry (VASE). The nTE of polymer 3 was extracted from Fitrilawati et al
[Fitrilawati’02].
_____________________________________________________________________92
104 105 106
1.5
1.6
1.7
1.8
1.9
TM0
TE0λ = 633 nm
n
MW [g/mol]
Fig. 6.6: Molecular weight dependence of refractive index and birefringence of thin films of MEH-PPVs 1 – 9 measured by prism coupling at 633 nm. nTE of polymer 3 was extracted from the recent report of Fitrilawati et al [Fitrilawati’02]. The data of nTE of polymer 9 is measured by transmission-reflection spectroscopy and its nTM is extrapolated from Tammer et al [Tammer’02].
As can be seen in Fig. 6.6, at high molecular weights (Mw > 5x105 g/mol), the
birefringence is very large and seems to approach a saturation limit. The values of nTE and
nTM in this region are in good agreement with those reported recently [Boudrioua’00,
Wasey’01, Tammer’02]. With the decrease of the molecular weight the birefringence is
significantly reduced and at Mw < 1.5x104 g/mol nearly complete loss of birefringence is
observed with a remaining birefringence nTE - nTM less than 0.005. Although MEH-PPVs
were synthesized in different laboratories via different routes, the results displayed in Fig.
6.6 show good reproducibility and a continuous dependence of the refractive index data
on the molecular weight.
_____________________________________________________________________93
6.3.2 Discussion
We discuss the molecular weight dependence of the absorption coefficient and
refractive index of spin cast MEH-PPV films in the context of the average orientation of
polymer chain segments with respect to the layer plane. The spin cast film exhibit
uniaxial anisotropy due to preferred alignment of polymer chains in the film plane, as
indicated at the beginning of this Chapter. The average chain orientation is strongly
related to the optical constants of thin films. The main electronic π-π* transition at λmax,
and the major electric polarizability which is related to refractive index n, are both
polarized in the chain direction of PPV. Consequently, αmax and n are largest if the
electric field vector Er
of incident light is parallel to the chain direction. If the PPV chain
segments become increasingly aligned parallel to the film plane, clearly αmax which is
measured at Er
parallel to the film plane, and nTE will increase, and correspondingly nTM
will decrease. The experimental results show this dependence of nTE, nTM, and αmax on
Mw quite well. The values of αmax, λmax and refractive indices nTE and nTM of polymer 5
deviate from this tendency. This deviation is probably related to its very different the
molecular weight distribution as it can be seen in the index of polydispersity (see Table
6.1). Moreover, this polymer was synthesized by using the Gilch route at a temperature
00C (ice bath condition), which was the first attempt to reduce the molecular weight of the
Gilch-type polymer [Ahn’03]. Obviously, the molecular weight distribution can also have
influence on the structure and optical properties of the films, in addition to the dominant
influence of Mw.
In this thesis, however, we had to restrict the investigations to the polymer
samples that have a similar polydispersity index. The study of the influence of
polydispersity remains an interesting task for further studies which are beyond the scope
of this thesis.
The strong correlation between optical properties with molecular weight indicates
that spin cast films of MEH-PPVs with larger Mw have an increasing amount of PPV
chain segments aligned parallel to the film plane. In order to confirm this phenomenon,
we have performed additional experiments using polarized FTIR spectroscopy.
_____________________________________________________________________94
6.4 FTIR Spectroscopy
6.4.1 Results
Fourier Transform Infrared (FTIR) spectroscopy with polarized light is a well-
known technique for determining molecular orientation. As the transition dipole moment
vector of a vibrational mode Mr
is oriented at a specific angle ϕ relative to the chain axis,
as shown in Fig. 6.7(a), the IR spectra can be used to obtain information on the chain
orientation by comparing the relative strength of vibration bands with different ϕ. We
have measured the IR spectra of thin films of 4, 6 and 8 in two configurations:
transmission and grazing incidence reflection [see Fig. 6.7(b) and (c)].
ϕ
(a)
M
E
Film Substrate
(b)
EE
Film Substrate
(b)
Film
SubstrateAu
(c)
E
ϕ
(a)
MM
E
Film Substrate
(b)
EE
Film Substrate
(b)
Film
SubstrateAu
(c)
E
Film
SubstrateAu
(c)
EE
Fig. 6.7: (a) Schematic of the transition dipole moment Mr
, oriented at a specific angle ϕ relative to the polymer chain axis. Scheme of FTIR experiments at perpendicular incidence (b), and grazing incidence reflection (c).
In the transmission at perpendicular incidence, the electrical field vector is
oriented parallel to the plane of film, E. Therefore, if the polymer chains are oriented
parallel to the film plane, it will lead to strong absorption for the vibration bands with
angle ϕ ≈ 0 ( Mr
parallel to Er
) and a small absorption for ϕ ≈ 900 ( Mr
perpendicular to
_____________________________________________________________________95
Er
). On the other hand, the electrical field vector is perpendicular to the film plane, E⊥ in
the reflection at grazing incidence. Then, the vibrational bands with angle ϕ ≈ 900 will
have strong absorption and the absorption will be small for ϕ ≈ 0. The FTIR spectra of
thin films of 4, 6 and 8 are displayed in Fig. 6.8.
3100 2900 1800 1600 1400 1200 1000 8000.0
0.5
1.0
1.5
2.0
Wavenumber [cm-1]
A (n
orm
aliz
ed)
(a)
861
(83°
)
968
(84°
)
1415
(64°
)
3058
(30°
)
E parallel
A (n
orm
aliz
ed)
3100 2900 1800 1600 1400 1200 1000 8000.0
0.5
1.0
1.5
2.0
2.5
64
8
E perpendicular
3058
(30°
)
1415
(64°
)
968
(84°
)
861
(83°
)
(b)
Fig. 6.8: FTIR spectra of MEH-PPVs 4 (solid line), 6 (dashed line) and 8 (dotted line). Absorbance A is normalized with respect to the band at 1415 cm-1. Wavenumbers and angles ϕ are shown for selected bands only. (a) Transmission spectroscopy. (b) Reflection spectroscopy.
_____________________________________________________________________96
6.4.2 Discussion
When an oriented polymer is investigated with a linearly polarized light, the
absorbance A of a single group in the polymer chain is proportional to the square of the
scalar product of its transition moment vector Mr
and the electrical vector Er
of the
incident light through [Siesler’84]
ϕ= 222 cos)E.M()E.M(~Arr
(6.6)
where ϕ is the angle between the transition moment and the electric vector. Thus, a
maximum absorption occurs when the electric vector is parallel to the transition moment
of the vibration band, while it shows a minimum when the electric vector is perpendicular
to the transition moment. The measured absorbance A is equal to the sum of absorbance
contributions of all molecules
∫
n
2 dn)E.M(~Arr
(6.7)
where the integral refers to the summation of all molecules.
The IR bands listed in Table 6.4 were used to evaluate the chain orientation
because their assignment to molecular vibrations and angles ϕ is known from earlier work
[Bradley’86]. As the spectra were measured from films with different thicknesses, the
band at 1415 cm-1 was used for normalization of the spectra.
Table 6.4: Assignment of IR band of unsubstituted PPV [Bradley’86]
Wavenumber [cm-1] Assignment Angle, ϕ [0]
861 out of plane phenyl CH wag 83
968 trans-vinylene CH wag 84
1415 semicircular phenyl stretch 64
3058 trans-vinylene C-H stretch 30
_____________________________________________________________________97
As shown in Fig. 6.8, the relative intensities of the IR bands with ϕ = 83° and ϕ =
84° are small for E and much larger for E⊥. This is a clear evidence for a preferred
orientation of the polymer chains in the plane of the film. The same conclusion can be
derived from the behavior of the band at 3058 cm-1 (ϕ = 30°) which is stronger for E as
compared to E⊥. The degree of chain orientation is strongly dependent on the Mw of the
MEH-PPVs which is especially evident at the bands at 861 cm-1 (ϕ = 83°) and 968 cm-1
(ϕ = 84°) for E⊥ in Fig. 6.8(b). These bands become stronger with growing Mw from 4 to
6 and 8, which indicates an increasing alignment of chains parallel to the film plane. This
orientation effect can be described more quantitatively by means of the ratio R of the
absorbances of the bands at 968 cm-1 (ϕ = 84°) and 3058 cm-1 (ϕ = 30°): R = A968 / A3058.
The results are shown in Table 6.5. For E⊥, R increases with Mw, whereas R decreases
with Mw in the case of E, which shows that the polymer chain segments become aligned
more parallel to the layer plane at higher molecular weights.
Table 6.5: Comparison of the ratio R of the IR absorbances of the bands at 968 cm-1 (ϕ = 84°) and 3058 cm-1 (ϕ = 30°) measured for incident electric field perpendicular and parallel to the film plane, respectively.
Polymer R at E⊥ R at E||
4 8 1.8
6 14 1.5
8 22 1.0
6.5 Third-Harmonic Generation
6.5.1 Results
Third-harmonic generation (THG) technique was used to study the influence of
the molecular weight on the third-order nonlinear optical susceptibility χ(3) of MEH-
PPVs. The film thicknesses of all materials studied were in the range of 50 - 70 nm. The
spectra of linear absorption and the dispersions of the modulus of χ(3) of selected MEH-
_____________________________________________________________________98
PPVs 1, 3, 4, 7 and 9 as a function of 1/3 of the fundamental wavelength, λL are shown in
Fig. 6.9. The values of χ(3) for all MEH-PPVs studied are listed in Appendix D. A
systematic trend of increasing χ(3) with Mw for all measured wavelengths was observed.
Fig. 6.9: (a) Spectra of linear absorption coefficient of thin films MEH-PPV (1, 3, 4, 7, 9). (b) Dispersions of third order nonlinear susceptibility χ(3) for polymer 1 (full squares), 3 (open circles), 4 (full stars), 7 (open squares) and 9 (full diamonds). Both experiments were performed at TE polarization.
300 400 500 600
0
4
8
12
16
20 (a) 97431
λ [nm]
α
[104 c
m-1]
300 400 500 6000
2
4
6
8
10(b) 9
7431
χ(3) [1
0-11 e
su]
λL/3 [nm]
_____________________________________________________________________99
The maximum value χ(3)max of all polymers studied as a function of Mw is
displayed in Fig. 6.10. We did not measure the third-order nonlinear optical susceptibility
of polymer 5, due to small amount of this polymer. Similar to nTE, the χmax(3) value also
increases with increasing Mw and it tends to saturate for higher Mw.
104 105 1060
5
10
χ(3) m
ax [1
0-11 es
u]
MW [g/mol]
Fig. 6.10: Molecular weight dependence of the resonance of third-order nonlinear optical susceptibility χ(3)
max of thin films of MEH-PPVs.
6.5.2 Discussion
The 3-photon resonance of the third-order nonlinear optical susceptibility χ(3)max
occurs at the wavelength λL ≈ 3λmax of their linear absorption coefficient. The peaks of
χ(3) of all MEH-PPVs studied are always red shifted relative to their λmax. These effects
have been observed earlier and explained as a consequence of the statistical distribution
on the effective π-conjugation length [Kurihara’91, D’Amore’02]. In the THG process,
longer chain segments have a much larger molecular hyperpolarizability than short chain
segments. Therefore, the relative contribution of the long chain segments dominates in the
THG process and we observe these shifts in all polymers studied.
_____________________________________________________________________100
The increase of χ(3)max of MEH-PPV films with molecular weight can be
explained with the polymer chain orientation in the film plane. The polymer chains in the
films of high Mw are more oriented parallel to the film plane than low Mw polymers. This
argument is in agreement with the birefringence results and FTIR spectroscopy as
discussed above. The average macroscopic third-order nonlinear optical susceptibility is
given by Eq. (6.1) [Kajzar’92]
Θχ=χ 4)3(xxxx
)3( cos (6.1)
where Θ is the angle between the polymer chain segments and the incident light
polarization. In this case one has [Kajzar’92]
for a mono-orientation (all-polymer chain parallel to a given direction)
for a bidimensional disorder (all-polymer chain parallel to (6.2) a plane and randomly disoriented within this plane) for a three-dimensional disorder
If we assume that the polymer chain orientation in the polymer 9 is describes by a
bidimensional disorder, we will have <cos4Θ>9 = 3/8. On the other hand, <cos4Θ>1 = 1/5,
if we assume that the polymer chain orientation in the polymer 1 is perfectly disordered.
Then, according to Eq. (6.2) we should have the ratio <cos4Θ>9/<cos4Θ>1 equal to 1.9.
However, based on our results of χ(3)max, we obtain a ratio of 3.7, which is higher than the
expected value. This shows that a direct application of this model for evaluation of the
chain orientation is not feasible. Other parameters like conjugation length, ordering, local
field factor and density of the molecules have also to be taken into account.
The increase of third-order nonlinear optical susceptibility χ(3) with molecular
weight suggests that higher Mw are favorable for all-optical switching applications.
However, the waveguide propagation loss coefficient αgw need also to be considered. We
will see below that αgw also increases with Mw.
=Θ
51
83
1
cos 4
_____________________________________________________________________101
6.6 Waveguide Propagation Losses
6.6.1 Results
The morphology of thin film waveguides has decisive influence on the total
waveguide loss coefficient αgw which is the sum of intrinsic absorptions and scattering
losses. The waveguide losses were measured at wavelength 1064 nm at both TE and TM
polarizations as described in Chapter 3.7 and 4.3. Figure 6.11 shows the data of αgw(TE0)
and αgw(TM0) at 1064 nm for all MEH-PPVs studied, except for polymer 9. We were not
able to prepare a waveguide of polymer 9 due to solubility problems.
0 1x105 2x105 3x105 4x105
0
10
20
30 TE0 TM0
λ = 1064 nm
αgw
[dB
/cm
]
MW [g/mol]
Fig. 6.11: Molecular weight dependence of waveguide losses measured at λ = 1064 nm for TE-polarized light (full circles) and TM-polarized light (open triangles). Waveguides prepared from Mw < 5 x 104 g/mol have unprecedented low
attenuation losses αgw(TE0) < 1 dB/cm, whereas higher Mw MEH-PPVs 5 - 8 have
significantly larger values of αgw(TE0) which increase strongly with Mw. Interestingly, all
MEH-PPVs studied at TM polarization have nearly identical data of αgw(TM0) in the
range of 0.5 – 1.0 dB/cm.
_____________________________________________________________________102
6.6.2 Discussion
Obviously, the molecular weight of MEH-PPV has major impact on the
waveguide propagation loss coefficient. However, it is not possible to understand how the
observed increase of the TE mode losses with molecular weight could be caused by the
averaged orientation of the polymer chains. Other important factor is the different
solubility of the MEH-PPVs studied. For preparing of optical waveguides with a typical
thickness in the range of 400 to 800 nm, high concentrations of the MEH-PPV solutions
were needed. Low molecular weight MEH-PPVs 1 – 4 were easily dissolved in
concentrations up to 7 % by weight and their solutions were filtered using 0.45 µm filters.
However, we were not able to dissolve higher molecular weight 5 – 9 in such
concentrations due to microgel formation. The higher concentration solutions of all 5 – 8
could not be filtered with 0.45 µm filter, and therefore a 1 µm filter was used. It is
possible that aggregates may still exist in such solutions and can be preserved through the
casting process and persists in the films [Nguyen’99, Nguyen’00]. In this way, different
amount of aggregation domains may exist in the different MEH-PPVs. We can assume
that the size and relative amount of these domains are related to Mw of the MEH-PPVs. In
case of very low Mw these domains may not exist at all. This assumption allows to
explain the variation of the waveguide propagation losses by the light scattering which is
caused mainly by local changes of the refractive index in the boundary regions between
the domains. This should depend on the polarization of the waveguide mode.
TE waveguide modes have the electric field vector Er
oriented in the film plane.
The local refractive index will change strongly, if the electric field propagates through the
boundary region of the domains. The local component of the anisotropic refractive index
of a domain will be largest (smallest) for Er
parallel (perpendicular) to the directions of
the PPV backbones, respectively. Consequently, the loss coefficient αgw(TE0) will depend
strongly on the relative amount and the size of these aggregation domains which both
may grow with increasing Mw.
As discussed above, the polymer chain segments are preferably aligned parallel to
the film plane. For TM polarization, the electrical field is perpendicular to the PPV chain
directions, even when the PPV chains lie in a domain. Therefore, the local changes of the
refractive index at the boundary regions will be rather small and independent on the
different lateral orientations of domains and polymer backbones, respectively. This model
_____________________________________________________________________103
explains why nearly the same and very small loss coefficients αgw(TM0) for all MEH-
PPVs independent on their Mw were observed. Clearly, this hypothesis needs verification
by further studies of the morphology of MEH-PPV which, however, is already known to
be nematic-like [Chen’02].
6.7 Photostability
6.7.1 Results
For photonic devices applications, the photostability of the polymers is a very
important parameter. Therefore, we have measured the UV-photosensitivity SUV (defined
in Chapter 3) of thin films MEH-PPVs 1 - 9 by means of a high pressure mercury lamp
combined with a water filter and UG-1 filter as described already in Chapters 3.4 and 4.5.
The film thicknesses of all films were kept constant at approximately 70 nm. The values
of SUV for all polymers studied are listed in Table 6.6. The films of MEH-PPVs
synthesized by the Gilch route show higher SUV than the Horner polymers. This might be
related to the amount of the chemical structure defects caused by the synthetic routes.
Table 6.6: Photosensitivity SUV values of MEH-PPVs from different synthetic routes. (*) No data for polymer 3 because this polymer was not available any more.
Polymer Synthetic Route SUV [a.u.]
1 Gilch 0.010
2 Horner 0.005
3 Horner (*)
4 Horner 0.007
5 Gilch 0.013
6 Gilch 0.012
7 Gilch 0.012
8 Gilch 0.014
9 Gilch 0.013
_____________________________________________________________________104
6.7.2 Discussion
As already discussed in Chapter 4, the UV irradiation causes defects in the
chemical structure which leads to shortening of the conjugation length. The better
photostability of the MEH-PPV films from Horner than Gilch route indicates a strong
relation between the photostability and amount of defects in the polymer. The Gilch-type
polymer shows several structural defects, as the result of side-reactions in the
dehydrohalogenation process. The most important defects are branching or crosslinking
with partial gel or microgel formation and an incomplete formation of vinylene groups
(Tolane-bis-benzyl moieties, TBB) which result in an interrupted backbone conjugation
[Becker’00, Hörhold’02]. On the other side, the Horner-type polymer exhibits a strictly
linear backbone with very well defined chain structure and extremely few chain defects
[Hörhold’02]. By means of 13C/1H-NMR spectroscopies, Holzer et al [Holzer’04] have
shown that Gilch-type MEH-PPV contains only 70% of the regular vinylene group,
whereas Horner-type MEH-PPV has ~ 97% of vinylene double-bonds in the polymer
backbone. Therefore, we believe that the poor photostability of MEH-PPV films from
Gilch route is mainly due to a large amount of defects in this polymer.
6.8 Conclusions
We have studied the properties of spin cast films of MEH-PPVs with a large
variation of the molecular weight which was caused by different synthetic routes. We
have shown that the basic optical properties, such as absorption coefficient, refractive
index and its birefringence, waveguide propagation losses, and third-order nonlinear
optical properties are related to the polymer chain orientation which depends significantly
on the molecular weight, especially in the range Mw < 5x105 g/mol. Thin films of high
molecular weight MEH-PPV have most chain segments oriented parallel to the film plane
(large birefringence), in contrast to low molecular weight samples which have nearly
random orientation of chain segments (small birefringence). By using the birefringence
method, the increase of polymer chain orientation parallel to the film plane with
increasing molecular weight were also observed in drop-cast polystyrene films [Prest’80],
as-cast polyamide films [Li’97], and spin-cast poly-3BCMU films [Grando’03]. These
observations seem to be rather general and can have strong impact on the device
applications using thin films of conjugated polymers because the polymer chain
_____________________________________________________________________105
orientation relates directly to their optoelectronic properties. Therefore, any report of
optoelectronic data of conjugated polymers must refer to details of synthesis and
molecular weight in order to avoid inconsistencies in scientific literature of these
materials. On the other side the UV photostability does not depend on the molecular
weight but on the amount of defects which are caused by the synthetic route. The films
from Horner-type MEH-PPV show better photostability than the defect-rich Gilch
polymers. Therefore, the Horner-type MEH-PPV is favourable for photonic applications.
Finally, we conclude that lower molecular weight MEH-PPVs are optimally suited
for the realization of all-optical switching devices, because of the ease of thin film
preparation, and good combination of high nonlinearity and ultimately low waveguide
propagation losses.
_____________________________________________________________________106
7 Microstructuring of MEH-PPV Waveguides: Towards All-Optical Switching Devices
In this Chapter, we present our studies of gratings fabrication in MEH-PPV planar
waveguides by use of UV-Photoablation, hot embossing and solvent-assisted
microcontact molding (SAMIM) techniques. 7.1 Photoablation
The UV photoablation technique is frequently used for the microstructuring of
thin films and surfaces. The mechanism of photoablation is a very complex processes
involving a variety of competing phenomena such as photochemical process with
subsequent thermal degradation of fragmented particles and molecules microstructure of
material [Schmidt’98]. The materials are removed or ablated at intensities larger than a
threshold which depends on the absorption coefficient of the polymers, laser wavelength
and laser pulse duration. This threshold is called as ablation or damage threshold.
Therefore, the knowledge of the damage threshold of the polymer is very useful for
microstructuring purposes. Additionally, the damage threshold at NIR laser wavelength is
an important limiting factor for any nonlinear optical device application.
7.1.1 Damage Threshold We have studied the photoablation of MEH-PPV waveguides by using the
fundamental and harmonic wavelengths of Nd:YAG laser at 1064 nm (pulse duration of
τFWHM ≈ 30 ps), 532 nm (τFWHM ≈ 21 ps) and 335 nm (τFWHM ≈ 17 ps). In all cases the
laser was focused by a telescope with focal distance of approximately 2 meter on the
polymer film. The focused spot size in the Gaussian beam waist was measured directly at
low intensities with the beam profiler placed at the sample position. For the specification
of the peak intensities I0 of the damage thresholds, we have used the definition of the
Gaussian beam profile
σ
−σ
−σ
−= 2t
2
2x
2
2x
2
0 2t
2y
2xexpI)t,y,x(I , (7.1)
_____________________________________________________________________107
where
tyx2/3
p0
)2(
EI
σσσπ= (7.2)
is the peak intensity and Ep is the pulse energy. The films were exposed to series of 100
shots with defined intensity. For intensities larger than damage threshold, well-defined
holes down to the substrate were ablated into the thin films and observed by means of a
microscope. The diameters of these holes were proportional to the laser intensities. Below
the threshold intensity, however, no damage of the film surface was observed even after
exposure to more than 1000 shots. We have obtained the following threshold intensities:
I0 = 21 GW/cm2 at λ = 1064 nm, I0 = 2 GW/cm2 at λ = 532 nm and I0 = 4 GW/cm2 at λ =
355 nm. These data provide limits for photoablation purposes as will be discussed below
but show, on the other hand, that MEH-PPV has excellent photostability at the working
wavelengths of nonlinear switching devices in the NIR.
7.1.2 Microstructuring
Microstructuring of MEH-PPV waveguides by means of UV-photoablation
technique was done in collaboration with Laser Laboratory Göttingen. Two different
methods were used as shown in Fig. 7.1: the direct illumination method and the imaging
method using a Schwarzschild objective [Bader’02]. The films were prepared in Mainz
and investigated in Göttingen.
The third-harmonic of a Nd:YAG laser (λ = 355 nm, pulse duration of 40 ps) was
used to fabricate gratings by the direct illumination method. This method uses a grating
phase mask with twice the period of the grating to be generated. The phase mask is
brought as close to the polymer film as possible; typical spacing is approximately 300
µm, including a protective quartz plate of 200 µm. The advantage of this setup is twofold:
one can generate a non-homogeneous amplitude profile of the grating, e.g. an apodized
grating profile, by applying a laser beam with an appropriate beam profile, and one can
achieve grating areas in the range of square millimeters or even square centimeters.
The second setup for grating fabrication in polymer films applies a Schwarzschild
objective to image a chromium mask onto the polymer surface. Two excimer lasers, ArF
(λ = 193 nm, pulse duration of 20 ns) and KrF (λ = 248 nm, pulse duration of 500 fs)
_____________________________________________________________________108
were used. The objective blocks out the zeroth diffraction order of the grating, the two
first diffraction order signals interfere on the grating surface, resulting in a grating period
reduced by a factor of 15 to 50 compared to the chromium mask depending on the
objective in use. The amplitude profiles of the gratings are homogeneously reflecting the
beam profile of the excimer lasers in use.
-1. +1.
0.
phase mask
polymer film600 nm
quartz plate200 mµ
substrate
(a)
+1
-1
Schwarzschildobjective
polymerfilm
beamblockgrating mask
(b)
200 µm
-1. +1.
0.
phase mask
polymer film600 nm
quartz plate200 mµ
substrate
(a)
+1
-1
Schwarzschildobjective
polymerfilm
beamblockgrating mask
(b)
200 µm
Fig. 7.1: Schematic diagrams of the experimental setups for writing sub-micrometer gratings in polymer films. (a) Direct illumination of a phase mask on top of the sample surface. (b) Imaging of a chromium mask by a Schwarzschild objective. 7.1.3 Results
Sub-micrometer gratings in thin films of MEH-PPV on quartz substrates have
been fabricated by using both setups described above. The grating thickness was
controlled by the applied fluence and could be varied from very shallow for fluences just
above the ablation threshold to a thickness that corresponds to the thickness of the
_____________________________________________________________________109
polymer film. Scanning electron microscope (SEM) pictures of the gratings generated in
MEH-PPV films of 590 nm thickness are shown in Fig. 7.2.
a)
b)
c)
Fig. 7.2: Grating structures in MEH-PPV waveguides with a thickness of 590 nm on quartz substrates and generated by UV photoablation. The applied wavelength, pulse duration, fluence, number of pulses, and grating period, respectively, were (a) 248 nm, 500-fs, 11.2 mJ/cm2, 50, and 386 nm; (b) 193 nm, 20-ns, 350 mJ/cm2, 2, and 1.0 µm; (c) 355 nm, 40-ps, 7 mJ/cm2, 100, and 720 nm.
_____________________________________________________________________110
Very good results of sinusoidal grating structures have been achieved by use of
500-fs pulses at 248 nm [Fig. 7.2(a)] and 20-ns pulses at 193 nm [Fig. 7.2(b)]. Both
gratings were fabricated by imaging a chromium grating mask onto the polymer surface
using a Schwarzschild objective. The generated grating periods are 386 nm [Fig. 7.2(a)]
and 1.0 µm [Fig. 7.2(b)]. The applied fluences and numbers of pulses were 11.2 mJ/cm2
and 50 pulses [Fig. 7.2(a)] and 350 mJ/cm2 and 2 pulses [Fig. 7.2(b)]. The debris
observed in Fig. 7.2(b) could have been caused by the grating fabrication setup. The film
was ablated in a horizontal position instead of a vertical position as in Fig. 7.2(a), so some
of the ablated particles could have fallen back down onto the sample. All the gratings
done with this setup show a highly homogeneous amplitude profile; the grating area is in
the order of 2.5 × 103 µm2, corresponding to a grating length of approximately 50 µm.
A grating generated by direct illumination using a phase mask is shown in Fig.
7.2(c). The third-harmonic of a 40-ps Nd:YAG laser at 355 nm was used, a grating period
of 720 nm has been generated by applying a fluence of 7 mJ/cm2 and 100 pulses. The
nearly Gaussian laser beam profile that was used in this case induced an apodized
amplitude profile in the grating. This approach is the most promising for applications like
gap soliton switching where the sharp edge of the photonic band gap in the waveguide
transmission spectrum of an apodized grating favors the observation of nonlinear optical
switching effects based on χ(3) nonlinear materials [Sipe’94]. Numerical simulations of a
switching device have been performed based on photonic bandgap formation in a
nonlinear periodic structure with the obtained material data and the sub-micrometer
structure above. We have shown that photonic bandgap all-optical switching can be
expected at wavelength 1064 nm in MEH-PPV planar waveguide [Bader’02].
7.2 Hot Embossing
As an alternative way for microstructuring of MEH-PPV waveguides, we have
applied the hot embossing method in collaboration with K. Petersen (MPIP Mainz). Hot
embossing, also known as nanoimprint lithography (NIL) has recently gained much
interest, in particular as a flexible and low cost method for structuring of polymers with
high reproducibility [Chou’95, Torres’03]. Three basic components are required: (i) a
stamp or master with a suitable size prepared by using standard photolithography
_____________________________________________________________________111
technique, (ii) polymer films to be printed, and (iii) equipment for printing with
controlled temperature, pressure and the parallelism of the stamp and the substrate.
A schematic view of the hot embossing process and a typical cycle of the sample
temperature during the embossing process are shown in Fig. 7.3 [Petersen’03]. First, the
master (silicon wafer or glass) was patterned by means of standard photolithography. In
order to minimize the sticking effects between the master and the polymer film, the
master was coated with a monolayer of n-octadecyltrichlorosilane. The polymer films
were spin coated onto glass or quartz substrates. The film and the master are heated to
approximately 800C which is above the glass transition temperature of MEH-PPV (Tg =
650C) so that the polymer film is soft to be imprinted. This is referred to as the embossing
temperature (Timp). After the master and the film reach the embossing temperature, the
master is pressed into the polymer film and the polymer flows around the master
conforming to its structure. The typical embossing time was 30 minutes. Afterwards, the
polymer was cooled to the room temperature and then the master was peeled off resulting
in a structured polymer film. The entire procedure was carried out in ambient air.
Fig. 7.3: Scheme of the hot embossing process (left) and a typical cycle of the sample temperature during the embossing process (right). A → B: heating (T > Tg), B: master lowered onto polymer film, B → C: embossing (T = Timp), C → D: cooling and E: removal of the master (T < Tg).
0 50 100 150 2000
20
40
60
80
100
120
140
160
E
C
D
B
A
T [0 C
]
t [min]
master
master
master
film
substrate
film
substrate
T > Tg
T = Timp
filmmaster
T < Tg
master
mastermaster
master
film
substrate
film
substrate
T > Tg
T = Timp
filmmastermaster
T < Tg
_____________________________________________________________________112
7.2.1 Result
The quality of the gratings was evaluated by use of an atomic force microscope
(AFM). Fig. 7.4 shows the AFM images of the silica master and the embossed MEH-PPV
film. The master grating had a period of 309 nm and a peak-to-peak height of 22 nm. The
embossed film replicates the master [Fig. 7.4(b)]. It has a period of 307 nm, which
matches that of the silicon master but its depth is shallower (17 nm). The surface of the
imprinted structure is very rough, which might be due to the sticking problems between
the master and the polymer film.
(b)
(a)
-20.
0
0
2
0.0
[n
m]
-25.
0
0
2
5.0
[nm
]
0 1.00 2.00 3.00 [µm]
(b)(b)
(a)
-20.
0
0
2
0.0
[n
m]
-25.
0
0
2
5.0
[nm
]
0 1.00 2.00 3.00 [µm]
Fig. 7.4: AFM images and surface trace analysis of (a) the master, (b) the embossed MEH-PPV film. The master has a period of 309 nm and a depth of 22 nm. The period, a peak-to peak height of the embossed film is 307 nm and 17 nm, respectively.
_____________________________________________________________________113
7.3 Solvent-Assisted Microcontact Molding Soft lithography using soft elastomeric mold to pattern soft materials has been
recently developed as a tool for micro- and nanofabrication [Xia’98]. Poly(di-
methylsiloxane) (PDMS) is used as a stamp to pattern the surface instead of a rigid master
as in the hot embossing technique. PDMS is mechanically flexible and it has a relatively
low surface energy resulting in weak adhesion to other materials. This property allows a
large number of PDMS molds to be produced from one master. Furthermore, it is also
ease to release the master from the molded polymer films without damage.
Therefore, we have applied a solvent-assisted microcontact molding (SAMIM)
technique to pattern the MEH-PPV waveguides. In principle, this technique is similar to
the hot embossing, but it uses a solvent instead of heat to soften the polymer film. The
process consists of four steps, as displayed in Fig. 7.5.
master
PDMS solutions
PDMS
master
stampfilm
substratefilm
substratemaster
PDMS solutions
master
PDMS solutions
PDMS
master
PDMS
master
stampfilm
substrate
stampfilm
substratefilm
substratefilm
substrate
Fig. 7.5: Schematic process of solvent assisted microcontact molding (SAMIM) technique for microstructuring MEH-PPV films.
First, the master (silica or glass) was patterned by means of standard lithography
technique. We used elastomeric PDMS (Sylgard 184, Dow Corning, Midland, MI) as a
stamp for imprinting the polymer film. A liquid silicon rubber base (vinyl-terminated
PDMS) and its curing agent (a mixture of a platinum complex and copolymers of
methylhydrosiloxane and dimethylsiloxane) were mixed with the ratio of 10:1. The mixed
solution was then put in an evacuated decicator for approximately 15 minutes to remove
air bubbles in the solution. Second, the solution was poured over the master and cured at
600C for about 4 hours. After cooling, the cured elastomer was separated from the master
resulting in a flexible stamp. The MEH-PPV films (d ≈ 600 nm) were prepared by spin
coating of chlorobenzene solution. The stamp was inked with a small amount of solvent
_____________________________________________________________________114
and then the stamp was brought into contact with the polymer film. The solvent dissolved
the polymer, and the liquid polymer is molded against the structures of the stamp. Once
the solvent has evaporated, the stamp was peeled off resulting in a patterned polymer
film.
7.3.1 Results
AFM images of the master and the imprinted grating in MEH-PPV film are shown
in Fig. 7.6. The master has a period of 330 nm and a peak-to-peak height of 80 nm. The
period of imprinted grating matches exactly with the master, but its depth is shallower (50
nm). The structure of the imprinted grating inverted its master.
Fig. 7.6: AFM images and cross sections of the master (top) and the imprinted grating on a MEH-PPV spin cast film (bottom). The master has a period of 330 nm and a peak-to-peak height of 80 nm. The period and depth of the imprinted film is 330 nm and 50 nm, respectively.
0.0 0.5 1.0 1.5 2.0
20
40
60
80
100
Hei
ght [
nm]
x [µm ]
0 nm
132 nm
0 2 4 [µm]
0 nm
132 nm
0 2 4 [µm]
0.0 0.5 1.0 1.5 2.00
20
40
60
Hei
ght [
nm]
x [µm ]0 2 4 [µm ]
79.1 nm
0 nm
0 2 4 [µm ]
79.1 nm
0 nm
_____________________________________________________________________115
The choice of the solvent is very important for the SAMIM technique. In general,
the solvent should have a higher vapor pressure and moderate surface tension to ensure
rapid evaporation of the solvent and minimal swelling of the PDMS stamp [Xia’98]. The
solvent should rapidly dissolve or swell the surface of the polymer. It should not, however
destroy the surface of the PDMS stamp. In our study, we have tried two common solvents
for MEH-PPV, i.e. toluene and chlorobenzene. However, toluene swells the PDMS stamp
resulting in poor quality of the imprinted grating as observed by an AFM. A good result
was obtained by using chlorobenzene [Fig. 7.6(b)].
7.4 Discussion and Conclusions We have applied several methods for fabrication of microstructures in MEH-PPV
films including photoablation and soft lithography. However, several problems still
remain to produce the prototype of a nonlinear Bragg reflector. Homogeneous sub-
micrometer gratings in MEH-PPV have been generated by application of the UV-
photoablation method. However, the UV lasers have caused additional photodegradation
of the films which leads to a decrease of the optical constants, i.e. absorption coefficient,
linear as well as nonlinear refractive index. Therefore, our attempts to demonstrate all-
optical switching with gratings generated by the UV-photoablation method were not
successful. Further investigations are necessary to find the optimum condition in order to
solve this problem, for example by use of 1064 nm laser pulses for the ablation process. As alternative way, the hot embossing technique is a quite promising method for
microstructuring of polymer films, as it allows easy serial production of grating
waveguide structures from a single master mask. However, the sticking between the
silicon master and the embossed film is a serious problem in the hot embossing method
resulting in a very rough surface of the grating. It seems that the anti-sticking monolayer
did not work properly. Moreover, this approach involves high temperature to soften the
polymer film that is known to affect the optical properties of the polymer film which are
caused by thermal degradation at T > Tg. We have shown that good quality of gratings in MEH-PPV films were prepared
by means of the SAMIM technique. The use of PDMS instead of a rigid master prevents
the sticking between master and imprinted polymer film, which is a serious problem in
the hot embossing technique. Recently, several groups have used this method for
_____________________________________________________________________116
microstructuring of MEH-PPV films for distributed feedback laser [Pissignano’03,
Lawrence’03, Gaal’03]. However, we still do not succeed to prepare a nonlinear Bragg
reflector in order to demonstrate all-optical switching experimentally. The major problem
is the poor resolved of discontinuity at the transition region between the waveguide and
the imprinted grating structures which is very crucial for the mode propagation. A
continuous increase of the grating depth would be required in order to minimize reflection
and scattering losses of the waveguide mode. The solution of this lithography problem
was not feasible with the presently available means. The principal feasibility of grating
fabrication by means of the SAMIM technique was confirmed for MEH-PPV.
_____________________________________________________________________117
8 General Consequences
This Chapter describes several general consequences of the results reported in this
thesis and gives a brief outlook for possible future developments. First of all, we have
shown that the conjugated polymer MEH-PPV satisfies the figures of merit requirements
for all-optical switching in a certain spectral window, located at the lower energy tail of
the two-photon absorption. In this region, the losses due to two-photon absorption are
already negligibly small but the nonlinear refractive index n2 is still enhanced by the two-
photon absorption resonance. This result can provide a guideline for future materials
development for nonlinear optical applications. The spectral region that fulfills the figures
of merit requirements could be shifted to the telecommunication windows at 1300 nm or
1550 nm by an appropriate design of molecular structures.
Secondly, we have shown that the molecular weight of rigid conjugated polymers
such as MEH-PPV has strong impact on the morphology of thin films, in particular on the
polymer chain alignment. Thin films of polymers with high Mw have most chain
segments oriented parallel to the film plane, in contrast to films of low Mw polymers
which have nearly random orientation of chain segments. These results have major
consequences on device applications using thin films of conjugated polymers because the
polymer chain orientation directly relates to their optoelectronic properties. Therefore,
any report of optoelectronic data of conjugated polymers must include the molecular
weight in order to avoid inconsistencies in the scientific literature of these materials.
Clearly, the reproducibility of optical properties of thin polymer films needs the control of
molecular weight. Furthermore, our results suggest an easy way for fine-tuning of optical
constants by appropriate choice of the molecular weight.
Finally, micro- and nanostructuring is a very important step in the development of
integrated optical devices for all-optical signal processing. We have shown that sub-
micrometer gratings in polymer planar waveguides is possible by using UV-
photoablation, hot embossing and solvent-assisted microcontact molding (SAMIM)
methods. It turned out that the SAMIM technique in particular seems to be the most
suitable method for preparing large area waveguide-grating structures with good quality.
However, there is still an unsolved problem to fabricate a continuous transition at the
borders between waveguide and imprinted grating structures which is very crucial to
_____________________________________________________________________118
avoid large propagation losses. Consequently, further improvements of this lithographic
method or the implementation of other techniques for microstructuring of polymer planar
waveguides are still needed to realize integrated all-optical switching devices.
_____________________________________________________________________119
9 Summary
We have studied and compared the linear and nonlinear optical properties of thin
films and planar waveguides of several new solution processable derivatives of poly(p-
phenylenevinylene) (PPV). We have found that MEH-PPV is the most promising
candidate for an integrated all-optical switching device because it shows a superior
combination of large third-order nonlinearity, ultimately low waveguide propagation
losses and sufficient photostability.
We have measured the dispersions of waveguide propagation loss αgw, nonlinear
refractive index n2 and nonlinear absorption coefficient α2 of MEH-PPV waveguides in
the range of 700 - 1600 nm. We have observed remarkably large values of intensity-
dependent changes of the refractive index in the order of ∆n = 10-3 at 1064 nm which are
fully reversible. We have found that the figures of merit requirements for all-optical
switching applications are satisfied in the wavelengths around 1100 - 1200 nm which is at
the lower energy tail of two-photon absorption. These figures of merit of MEH-PPV are
the best presently known values.
We have studied the optical properties of thin films of MEH-PPV with large
variations of molecular weight (Mw) in the range of 104 to 106 g/mol. We have shown that
the molecular weight has strong impact on the alignment of polymer chain segments in
thin films and consequently on the optical birefringence, waveguide propagation loss and
cubic nonlinearity. Thin films of high Mw have most chain segments oriented parallel to
the film plane (large birefringence), whereas low Mw films have nearly random
orientation of chain segments (small birefringence). We conclude that MEH-PPV with
lower Mw is most favorable for the realization of all-optical switching devices because of
its easiness of thin film processing and good combination of large nonlinearity and
ultimately low waveguide propagation losses at transverse electric polarization.
We have prepared sub-micrometer gratings in MEH-PPV waveguides by applying
UV-photoablation, hot embossing and solvent-assisted microcontact molding (SAMIM)
techniques. It turned out that the SAMIM technique is the only suitable structuring
technique to fabricate large area waveguide-grating structures. However, the fabrication
of a nonlinear Bragg reflector and the demonstration of all-optical switching still require
an improved solution to solve some problems related to lithography which are beyond the
possibility of our laboratory.
_____________________________________________________________________120
10 Zusammenfassung
Wir haben die linearen und nichtlinearen optischen Eigenschaften dünner
Schichten und planarer Wellenleiter aus neuen Derivaten des Poly(p-phenylenvinylen)
(PPV) studiert und verglichen, die aus Lösungen verarbeitbar sind. Wir haben gefunden,
daß MEH-PPV das vielversprechendste Material für einen integriert-optischen Schalter
ist, weil es eine am besten geeignete Kombination aus großer optischer Nichtlinearität 3.
Ordnung, geringen Wellenleiterdämpfungsverlusten und ausreichender Photostabilität
zeigt.
Wir haben die Dispersionen des Wellenleiterdämpfungsverlustes αgw, des
nichtlinearen Brechungsindex n2 und des nichtlinearen Absorptionskoeffizienten α2 von
Wellenleitern aus MEH-PPV im Bereich 700 - 1600 nm gemessen. Wir haben
außerordentlich große Brechungsindexänderungen im Bereich ∆n = 10-3 bei 1064 nm
beobachtet, die völlig reversibel sind. Wir haben gefunden, daß die Gütekriterien (Figures
of Merit) für rein optische Schalter im Wellenlängenbereich 1100 - 1200 nm erfüllt sind.
Dieser Bereich entspricht dem niederenergetischen Ausläufer der Zwei-Photonen
Absorption. Die Gütekriteren von MEH-PPV sind die besten momentan bekannten Werte.
Wir haben die optischen Eigenschaften der dünnen Schichten von MEH-PPV mit
einer großen Variation des Molekulargewichts (Mw) im Bereich 104 bis 106 g/mol
untersucht. Wir haben gezeigt, daß das Molekulargewicht starke Auswirkungen auf die
Ausrichtung der Polymerkettensegmente in den dünnen Schichten und infolgedessen auf
die optische Anisotropie des Brechungsindex, den Wellenleiterdämpfungsverlust und die
kubische optische Nichtlinearität hat. Die Polymerkettensegmente in dünnen Schichten
aus MEH-PPV mit hohem Mw orientieren sich bevorzugt parallel zur Schichtebene (große
Anisotropie des Brechungsindex), während die Schichten aus MEH-PPV mit niedrigem
Mw eine nahezu isotrope Ausrichtung der Kettensegmente haben (kleine Anisotropie des
Brechungsindex). Wir stellen fest, daß MEH-PPV mit niedrigem Molekulargewicht für
die Realisierung rein optischer Schalter höchst geeignet ist wegen der besseren
Filmherstellung und der guten Kombination aus großer Nichtlinearität und äußerst
geringen Wellenleiterdämpfungsverlusten bei transversal-elektrischer Polarization.
Wir haben Gitterstrukturen im Submikrometerbereich in Wellenleitern aus MEH-
PPV präpariert mit den Techniken UV-Photoablation, Heißprägen und Lösungs-
_____________________________________________________________________121
assistierter Mikrokontakt Abformung (Solvent-Assisted Microcontact Molding, SAMIM).
Es ergab sich, daß die SAMIM-Technik das einzige brauchbare Verfahren zur
Herstellung großflächiger Gitter-Wellenleiter-Strukturen ist. Jedoch erfordern die
Herstellung eines nichtlinearen Bragg-Reflektors und die Demonstration des rein
optischen Schalters noch verbesserte Lösungen einiger Lithographieprobleme, die
außerhalb der Möglichkeiten unseres Labors liegen.
_____________________________________________________________________122
11 References [Abe’97] A. Abe in “Primary Photoexcitations in Conjugated Polymers:
Molecular Exciton versus Semiconductor Band Model“, N. S. Sariciftci (Ed.), World Scientific Publ., Singapore (1997), 115.
[Ahn’03] T. Ahn, private communications. [Aramaki’92] S. Aramaki, G. Assanto, G. I. Stegeman, W. H. G. Horstuis, G. R.
Möhlman, Optics Commun. 94 (1992), 326. [Assanto’88] G. Assanto, R. M. Fortenberry, C. T. Seaton, G. I. Stegemann, J. Opt.
Soc. Am. B 5 (1988), 432. [Atreya’99] M. Atreya, S. Lia, E.T. Kang, K.G. Neoh, Z.H. Ma, K.L. Tan, W.
Huang, Polymer Degradation and Stability 65 (1999), 287. [Bader’02] M. A. Bader, G. Marowsky, A. Bahtiar, K. Koynov, C. Bubeck, H.
Tillmann, H.-H. Hörhold, S. Pereira, J. Opt. Soc. Am. B 19 (2002), 2250.
[Bahtiar’02] A. Bahtiar, K. Koynov, C. Bubeck, M. A. Bader, U. Wachsmuth, G.
Marowsky, in Nonlinear Guided Waves and Their Applications, OSA Trends in Optics and Photonics (TOPS) Vol.80, OSA Technical Digest, Postconference Edition (Optical Society of America, Washington DC, 2002), p. NLMD58-1.
[Baker’93] C. J. Baker, O. M. Gelsen, D. D. C. Bradley, Chem. Phys. Lett. 201
(1993), 127. [Bartuch’92] U. Bartuch, A. Bräuer, P. Dannberg, H.-H. Hörhold, D. Raabe, Int. J.
Optoelectronics 7 (1992), 275. [Becker’97] H. Becker, S. E. Burns, and R. H. Friend, Phys. Rev. B 56 (1997), 1893. [Becker’00] H. Becker, H. Spreitzer, W. Kreuder, E. Kluge, H. Schenk, I. Parker, Y.
Cao, Adv. Mater. 12 (2000), 42. [Boudrioua’00] A. Boudrioua, P. A. Hobson, B. Matterson, I. D. W. Samuel, W. L.
Barnes, Synth. Met. 111-112 (2000), 545. [Boyd’92] R. W. Boyd, “Nonlinear Optics“, Academic Press, Orlando, USA, 1987. [Bradley’86] D. D. C. Bradley, R. H. Friend, H. Lindenberger, S. Roth, Polymer 27
(1986), 1709. [Bradley’87] D. D. C. Bradley, J. Phys. D: Appl. Phys. 20 (1987), 1389. [Braun’91] D. Braun and A. J. Heeger, Appl. Phys. Lett. 58 (1991), 1982.
_____________________________________________________________________123
[Brédas’94] J. L. Bredas, C. Adant, P. Tackx, A. Persoons, Chem Rev. 94 (1994), 243.
[Bubeck’91] C. Bubeck, A. Kaltbeitzel, A. Grund, M. LeClerc, Chem. Phys. 154
(1991), 343. [Bubeck’95] C. Bubeck in “Nonlinear Optical Materials: Principles and
Applications”, V. Degiorgio, C. Flytzanis (Eds.), IOP Press, 1995, 359. [Bubeck’98] C. Bubeck in “Electronic Materials: The Oligomer Approach”, K.
Müllen, G. Wegner (Eds.), Wiley-VCH, Weinheim (1998), 449. [Bubeck’00] C. Bubeck, K. Ueberhofen, J. Ziegler, F. Fitrilawati, U. Baier, H.
Eichner, C. Former, K. Müllen, S. Pfeiffer, H. Tillmann, H.-H. Hörhold, Nonlinear Optics 25 (2000), 93.
[Butcher’90] P. N. Butcher, D. Cotter, “The Elements of Nonlinear Optics“,
Cambridge University Press, Cambridge, 1990. [Carter’83] G. M. Carter, Y. J. Chen, Appl. Phys. Lett. 42 (1983), 643. [Chandross’94] M. Chandross, S. Mazumdar, S. Jeglinski, X. Wei, Z. V. Vardeny, E. W.
Kwock, T. M. Miller, Phys. Rev. B 50 (1994), 14702. [Chandross’97] M. Chandross, S. Mazumdar, M. Liess, P. A. Lane, Z. V. Vardeny, M.
Hamaguchi, K. Yoshino, Phys. Rev. B 55 (1997), 1486. [Chen’02] S. H. Chen, A. C. Su, Y. F. Huang, C. H. Su, G. Y. Peng, C. A. Chen,
Macromolecules 35 (2002), 4229. [Chou’95] S. Y. Chou, P. R. Krauss, P. J. Renstrom, Appl. Phys. Lett. 67 (1995),
3114. [Cowie’91] J. M. G. Cowie, ”Polymers: Chemistry and Physics of Modern
Materials”, 2nd Edition, 1991, London, United Kingdom. [Cumpston’95] B.H. Cumpston, K. F. Jensen, Synth. Met. 73 (1995), 195. [D’Amore’02] F. D’Amore, A. Zappettini, G. Facchini, S. M. Pietralunga, M.
Martinelli, C. Dell’Erba, C. Cuniberti, D. Comoretto, G. Dellepiane, Synth. Met. 127 (2002), 143.
[DeLong’90] K. W. DeLong, G. I. Stegeman, Appl. Phys. Lett. 57 (1990), 2063. [Fitrilawati’02] F. Fitrilawati, M.O. Tjia, S. Pfeiffer, H. Tillmann, H.-H. Hörhold, A.
Deutesfeld, H. Eichner, C. Bubeck, Opt. Mater. 21 (2002), 511. [Friend’99] R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, R. N.
Marks, C. Taliani, D. D. C. Bradley, D. A. Dos Santos, J. L. Bredas, M. Lögdlund, W. R. Salaneck, Nature, 397 (1999), 122.
_____________________________________________________________________124
[Gaal’03] M. Gaal, C. Gadermeier, H. Plank, E. Moderegger, A. Pogantsch, G. Leising, E. J. W. List, Adv. Mater. 15 (2003), 1165.
[Gabler’97] Th. Gabler, R. Waldhäusl, A. Bräuer, U. Bartuch, R. Stockmann, and
H.-H. Hörhold, Opt. Commun. 137 (1997), 31. [Gabler’98] Th. Gabler, A. Bräuer, R. Waldhäusl, U. Bartuch, H.-H. Hörhold, and F.
Michelotti, Pure and Appl. Optics 7 (1998), 159. [Gartstein’95] Yu. N. Gartstein, M. J. Rice, E. M. Conwell, Phys. Rev. B 51 (1995),
5546. [Gilch’66] H. G. Gilch, W. L. Wheelwright, J. Polym. Sci., Part A-1: Polym. Chem.
4 (1966), 1337. [Grando’03] D. Grando, G. P. Banfi, D. Fortusini, R. Ricceri, S. Sottini, Synth. Met.
139 (2003), 863. [Gubler’02] U. Gubler, Ch. Bosshard, Adv. Poly. Sci. 158 (2002), 123 [Hecht’87] E. Hecht, “Optics“, 2nd edition, Addison Wesley Inc. USA, 1987. [Holzer’04] W. Holzer, A. Penzkofer, H. Tillmann, H.-H. Hörhold, Synth. Met. 140
(2004), 155. [Hörhold’01] H.-H. Hörhold, H. Tillmann, D. Raabe, M. Helbig, W. Elflein, A.
Bräuer, W. Holzer, A. Penzkofer, Proc. SPIE 4105 ( 2001), 431. [Hörhold’02] H.-H. Hörhold, H. Tillmann, C. Bader, E. Klemm, W. Holzer, A.
Penzkofer, Proc. SPIE 4464 (2002), 317. [Jackson’83] J. D. Jackson, “Klassische Elektrodynamik“, Walter de Gruyter, Berlin,
1983. [Joannopoulos’95] J. D. Joannopoulos, R. D. Meade, J. N. Winn, “Photonic Crystals :
Molding the Flow of Light“, Princeton University Press, New Jersey, 1995.
[Kajzar’85] F. Kajzar, J. Messier, Phys. Rev. A 32 (1985), 2352. [Kajzar’92] F. Kajzar, L. Le Moigne, A. Thierry, in “Electronic Properties of
Polymers; Orientation and Dimensionality of Conjugated Systems”, H. Kuzmany, M. Mehring (Eds.), Springer Series in Solid State Sciences, Springer-Verlag, 1992, 202.
[Kajzar’96] F. Kajzar and J. D. Swalen, eds., “Organic Thin Films for Waveguiding
Nonlinear Optics“, Gordon and Breach Publ. Amsterdam, (1996). [Kim’93] D. Y. Kim, M. Sundheimer, A. Otomo, G. I. Stegeman, W. H. G.
Horsthuis, G. R. Möhlmann, Appl. Phys. Lett. 63 (1993), 290.
_____________________________________________________________________125
[Kim’94] D. Y. Kim, B. L. Lawrence, W. E. Torruellas, G. I. Stegeman, G. Baker, J. Meth, Appl. Phys. Lett. 65 (1994), 1742.
[Kogelnik’79] H. Kogelnik in “Integrated Optics: Topics in Applied Physics Vol. 7”, T.
Tamir (Ed.), Springer Verlag Berlin (1079), 13. [Koynov’02] K. Koynov, N. Goutev, F. Fitrilawati, A. Bahtiar, A. Best, C. Bubeck,
H.-H. Hörhold, J. Opt. Soc. Am. B 19 (2002), 895. [Kranzelbinder’02] G. Kranzelbinder, E. Toussaere, J. Zyss, A. Pogantsch, E. W. J.
List, H. Tillmann, H.-H. Hörhold, Appl. Phys. Lett. 80 (2002), 716. [Kurihara’91] T. Kurihara, Y. Mori, T. Kaino, H. Murata, N. Takada, T. Tsutsui, S.
Saito, Chem. Phys. Lett. 183 (1991), 534. [Lawrence’94] B. L. Lawrence, M. Cha, J. Kang, W. E. Torruellas, G. I. Stegeman, G.
Baker, J. Meth, S. Etemad, Electron. Lett. 30 (1994), 447. [Lawrence’03] J. R. Lawrence, G. A. Turnbull, I. D. W. Samuel, Appl. Phys. Lett. 82
(2003), 4023. [Leng’94] J. M. Leng, S. Jeglinski, X. Wei, R. E. Benner, Z. V. Vardeny, Phys.
Rev. Lett. 72 (1994), 156. [Li’97] F. Li, K.-H. Kim, E. P. Savitski, J.-C. Chen, F. W. Haris, S. Z. D.
Cheng, Polymer 38 (1997), 3223. [Liess’97] M. Liess, S. Jeglinski, Z. V. Vardeny, M. Ozaki, K. Yoshino, Y. Ding
and T. Barton, Phys. Rev. B 56 (1997), 15712. [Lin’98] S.Y. Lin, E. Chow, V. Hietala, P. R. Villeneuve, J. D. Joannopoulos,
Science 282 (1998), 274. [Lines’91] M. E. Lines, "Physical properties of materials: Theoretical Overview,"
in Handbook of Infrared Optical Materials, P. Klocek (Ed), Marcel Dekker, New York 1991, 57.
[Marcuse’74] D. Marcuse, “Theory of Dielectric Optical Waveguides”, Academic
Press, New York, (1974). [Marques’91] M. B. Marques, G. Assanto, G. I. Stegeman, G. R. Möhlmann, E. W. P.
Erdhuisen, W. H. G. Horsthuis, Appl. Phys. Lett. 58 (1991), 2613. [Martin’99] S.J. Martin, D.D.C. Bradley, P.A. Lane, H. Mellor, and P.L. Burn, Phys.
Rev. B 59 (1999), 15133. [Mathy’96] A. Mathy, K. Ueberhofen, R. Schenk, H. Gregorius, R. Garay, K.
Müllen, C. Bubeck, Phys. Rev. B 53 (1996), 4367.
_____________________________________________________________________126
[McBranch’95] D. McBranch, I. H. Campbell, D. L. Smith, J. P. Ferraris, Appl. Phys. Lett. 66 (1995), 1175.
[McGehee’00] M. D. McGehee, A. J. Heeger, Adv. Mater. 12 (2000), 1655. [Mekis’96] A. Mekis, J. C. Chen, I. Kurland, S. Fan, P. R. Villeneuve, J. D.
Joannopoulos, Phys. Rev. Lett. 77 (1996), 3787. [Messier’89] J. Messier, F. Kajzar, P. N. Prasad, and D. Ulrich, eds., “Nonlinear
Optical Effects in Organic Polymers“, Kluwer, Dordrecht, (1989). [Meyer’97] R. K. Meyer, R. E. Benner, Z. V. Vardeny, M. Liess, M. Ozaki, K.
Yoshino, Y. Ding, T. Barton, Synth. Met. 84 (1997), 549. [Meyerhof’78] D. Meyerhofer, J. Appl. Phys. 49 (1978), 3993. [Michelotti’95] F. Michelotti, T. Gabler, H.-H. Hörhold, R. Waldhausl, A. Bräuer, Opt.
Commun. 114 (1995), 247. [Möller’88] K. D. Möller, “Optics”, University Science Books, USA, 1988. [Neher’89] D. Neher, A. Wolf, C. Bubeck, G. Wegner, Chem. Phys. Lett. 163
(1989), 116. [Neher’90] D. Neher, PhD Dissertation, University Mainz, Germany, 1990. [Nguyen’99] T.-Q. Nguyen, V. Doan, B. J. Schwartz, J. Chem. Phys. 110 (1999),
4068. [Nguyen’00] T.-Q. Nguyen, I. B. Martini, J. Liu, B. J. Schwartz, J. Phys. Chem. B
104 (2000), 237. [Pantell’67] R. Pantell, F. Pradere, J. Hanus, M. Schott, H. Puthoff, J. Chem. Phys.
46 (1967), 3507. [Petersen’03] K. Petersen, Ph.D Dissertation, University of Mainz, Germany, 2003. [Pethrick’99] R. A. Pethrick, K. E. Rankin, J. Mater Sci-Mater El. 10 (1999), 141. [Petritsch’99] K. Petritsch, R. H. Friend, Synth. Met. 102 (1999), 976. [Penzkofer’98] A. Penzkofer, E. Drotleff, W. Holzer, Opt. Commun. 158 (1998), 221. [Pfeiffer’99a] S. Pfeiffer, H.-H. Hoerhold, Synth. Met. 101 (1999) 109. [Pfeiffer’99b] S. Pfeiffer, H.-H. Hörhold, Macromol. Chem. Phys. 200 (1999), 1870. [Pissignano’03] D. Pisignano, L. Persano, F. Babudri, G. M. Farinola, F. Naso, R.
Cingolani, G. Gigli, Synth. Met. 139 (2003), 679.
_____________________________________________________________________127
[Prest’80] W. M. Prest, D. J. Luca, J. Appl. Phys. 51 (1980), 5170. [Rochford’91] K. Rochford, R. Zanoni, G. I. Stegeman, W. Krug, E. Miao, M. W.
Beranek, Appl. Phys. Lett. 58 (1991), 13. [Rothberg’96] L. J. Rothberg, M. Yan, F. Papadimitrakopoulos, M.E. Galvin, E.W.
Kwock, T.M. Miller, Synth. Met. 80 (1996), 41. [Samoc’95] A. Samoc, M. Samoc, M. Woodruff, B. Luther-Davies, Opt. Lett. 20
(1995), 1241. [Sariciftci’93] N. S. Sariciftci, D. Braun, C. Zhang, V. I. Srdanov, A. J. Heeger, G.
Stucky, F. Wudl, Appl. Phys. Lett. 62 (1993), 585. [Scalora’94] M. Scalora, J. P. Dowling, C. M. Bowden, M. J. Bloemer, Phys. Rev.
Lett. 73 (1994), 1368. [Schmidt’98] H. Schmidt, J. Ihlemann, B. Wolff-Rottke, K. Luther, J. Troe, J. Appl.
Phys. 83 (1998), 5458. [Scott’96] J. C. Scott, J. H. Kaufman, P. J. Brock, R. DiPietro, J. Salem, J. A.
Goitia, J. Appl. Phys. 79 (1996), 2745. [Siesler’84] H. W. Siesler, Adv. Polym. Sci. 65 (1984), 1. [Sipe’94] J.E. Sipe, L. Poladian, C.M. de Sterke, J. Opt. Soc. Am. A 11 (1994),
1307. [Stegeman’89] G. I. Stegeman, R. H. Stolen, J. Opt. Soc. Am. B 6, (1989), 652. [Stegeman’93] G. I. Stegeman, Proc. SPIE 1852 (1993), 75. [Stegeman’96] G. I. Stegeman, W. E. Torruellas, Phil. Trans. R. Soc. Lond. A 354
(1996), 745. [Stegeman’97] G. I. Stegeman in “Nonlinear Optics of Organic Molecules and
Polymers”, H. S. Nalwa and S. Miyata (Eds.), CRC Press Inc. 1997, 799.
[Stegeman’00] G. I. Stegeman, M. Liu, S. Polyakov, F. Yoshino, L. Friedrich, Nonlinear
Optics 24 (2000), 1. [Tammer’02] M. Tammer, A. P. Monkman, Adv. Mater. 14 (2002), 210. According to
private communication of Prof. A. P. Monkman, MEH-PPV used in this work had Mw = 1.6x 106 g/mol as specified by supplier.
[Tencor’98] Manual of KLA Tencor P-10 surface profilometer. [Tien’70] P. K. Tien, R. Ulrich, J. Opt. Soc. Am. 60 (1970), 1325.
_____________________________________________________________________128
[Tien’71] P. K. Tien, Appl. Opt. 10 (1971), 2395. [Tien’77] P. K. Tien, Rev. Mod. Phys. 49 (1977), 361. [Torres’03] C.M. Sotomayor Torres, S. Zankovych, J. Seekamp, A.P. Kam, C.
Clavijo Cedeno, T. Hoffmann, J. Ahopelto, F. Reuther, K. Pfeiffer, G. Bleidiessel, G. Gruetzner, M.V. Maximov, B. Heidari, Mat. Sci. Eng. C 23 (2003) 23.
[Turnbull’03] G. A. Turnbull, P. Andrew and W. L. Barnes, I. D. W. Samuel, Appl.
Phys. Lett. 82 (2003), 313. [Ulrich’70] R. Ulrich, J. Opt. Soc. Am. 60 (1970), 1337. [Ulrich’73] R. Ulrich, R. Torge, Appl. Opt. 12 (1973), 2901. [Ueberhofen’96] K. Ueberhofen, PhD Dissertation, University Mainz, Germany, 1996. [Ueberhofen’99] K. Ueberhofen, A. Deutesfeld, K. Koynov, C. Bubeck, J. Opt. Soc. Am.
B 16 (1999), 1921. [Wasey’01] J. A. E. Wasey, A. Safonov, I. D. W. Samuel, W. L. Barnes, Phys. Rev.
B 64 (2001), 205201. [Washo’77] B. D. Waso, IBM J. Res. Dev. 21 (1977), 190. [Xia’98] Y. Xia, G. M. Whitesides, Ann. Rev. Mater. Sci. 28 (1998), 153. [Xing’97] K. Z. King, N. Johansson, G. Beamson, D. T. Clark, J. L. Bredas, W. R.
Salaneck, Adv. Mater. 9 (1997), 1027. [Yan’94] M. Yan, L. J. Rothberg, F. Papadimitrakopoulos, M. E. Galvin, T. M.
Miller, Phys. Rev. Lett. 73 (1994), 744. [Taylor’74] H. F. Taylor, A. Yariv, Proc. IEEE 62 (1974), 1044. [Yeh’88] P. Yeh, “Optical Waves in Layered Media“, John Wiley & Sons Inc.,
USA, 1988. [Yonkoski’92] R. K. Yonkonski, D. S. Soane, J. Appl. Phys. 72 (1992), 725. [Yoshino’03] F. Yoshino, S. Polyakov, M. Liu, G. I. Stegeman, Phys. Rev. Lett. 91
(2003), 0639021. [Yu’95] G. Yu, J. Gao, J. C. Hummelen, F. Wudl, and A. J. Heeger, Science 270
(1995), 1789. [Ziegler’00] J. Ziegler, PhD Dissertation, University Mainz, Germany, 2001.
_____________________________________________________________________129
Appendix A
Studies of Oligomer OPEs and conjugated polymer PPE
A.1 Oligomer OPEs
The oligomer series of oligo(1,4-phenyleneethynylene)s (OPEs) was obtained
from the group of Prof. H. Meier (Institute of Organic Chemistry, University of Mainz).
Fig. A.1 shows the absorption spectra of diluted solutions of OPEs (1a – e) in CHCl3. The
wavelengths of the absorption maxima λmax show the strong bathochromic shift with
increasing number of repeat units, which is typical for short oligomers of conjugated
systems. The corresponding molar extinction coefficients εmax also increase with
increasing number of repeat units.
250 300 350 400 450 5000
20000
40000
60000
80000
100000
1200001e
1d
1c
1b
1a
ε [l/m
ol c
m]
λ [nm]
Fig. A.1: UV-Vis spectra of OPEs 1a - e in CHCl3
Thin films were prepared in the following way: Compounds 1a – 1e were mixed
with polystyrene (Mw = 1 x 105 g/mol) at (10.00 ± 0.03) % of concentration by weight,
dissolved in toluene using a concentration by weight of the mixture of 2.7 %. The
solutions were spin coated onto fused silica substrates at spinning speeds of 500 rpm and
9000 rpm. The final film thicknesses d were approximately 212 nm and 50 nm,
_____________________________________________________________________130
respectively. Thin films (d ≈ 50 nm) were used for linear optical studies and thick films (d
≈ 212 nm) for THG (third-harmonic generation) studies.
Fig. A.2: Spectra of (a) the absorption coefficient after correction of reflection losses at interfaces and (b) the refractive index of thin films of OPEs and polystyrene (PS), measured at transverse electric (TE) polarization. Transmission and reflection spectra of thin films (d ≈ 50 nm) were measured by
use of a spectrophotometer (Perkin Elmer Model Lambda 900) at TE-polarization. The
300 400 500
0.0
0.5
1.0
1.5
2.0
2.5 (a) 1a = OPE21b = OPE31c = OPE41d = OPE51e = OPE6
PS1c
1d
1e
1b
1a
α
[104 cm
-1]
λ [nm]
300 400 500 6001.58
1.60
1.62
1.64
1.66
1.68 (b)
1e1d
1c1b1a
1a = OPE21b = OPE31c = OPE41d = OPE51e = OPE6
PS
n
λ [nm]
_____________________________________________________________________131
dispersions of the refractive index n(λ) and the intrinsic absorption coefficient α(λ) of
thin (d ≈ 50 nm) were determined by evaluation of transmission and reflection spectra of
thin films on quartz substrates (see Chapter 3). The spectra of α(λ) and the dispersions of
n(λ) of thin films of OPEs are shown in Fig. A.2. The data of λmax and αmax are given in
Table A.1.
We have measured the nonlinear optical properties of thin films (d ≈ 200 nm) of
OPEs by means of THG at the three-photon resonance condition with laser wavelengths
λL ≈ 3λmax. This ensures sufficient detection sensitivity for the thin OPE films (1a - e).
The results for modulus )3(resχ and phase angle Φmax are given in Table A.1. Clearly, the
modulus of the third-order susceptibility increases with chain length L. The phase angle
also increases on going from 1a to 1c and then stays constant - at around 90o. The
changes of phase were interpreted in terms of different contributions of polystyrene (PS)
matrix and OPEs to the total sum of the complex χ(3) value, which is the vector sum of
χ(3)PS and χ(3)
OPE. Because we have performed the THG experiments always at λL ≈ 3λmax
of the OPEs, the contribution χ(3)OPE is imaginary and ΦOPE = 900, which is typically
observed at the peak of the three-photon resonance. On the other hand, χ(3)PS is
nonresonant and real because λL is much larger than 3λmax of PS, which implies ΦPS = 00.
As the modulus )3(OPEχ increases strongly with L, the contribution of PS to the total sum
of )3(resχ becomes negligible for the longer oligomers 1c – e and the total phase angle
approaches 900.
In order to understand the influence of chain length on nonlinear optical
properties, it is better to consider the second hyperpolarizability γ, which is a molecular
parameter, instead of the macroscopic susceptibility χ(3). In the cgs system of units, these
quantities are related by Eq. (A.1) where N is the number of oligomers per unit volume.
[ ] ),,:3()(f)3(Nf),,:3( 3)3( ωωωωγωω=ωωωωχ (A.1)
We have calculated N as in Eq. (A.2), where NA is the Avogadro constant, MOPE and cOPE
are the molecular mass and the concentration by weight of the respective OPE compound
and ρ is the density of the film.
_____________________________________________________________________132
OPE
OPEAM
cNN ρ= (A.2)
The Lorentz local field factors f(ω) was calculated using f(ω) = [n2(ω) + 2]/3. For the
determination of the molecular hyperpolarizabilities |γres|, we have used the imaginary
part of the total third-order susceptibility of the film. The |γres| of different OPEs 1a – e
are listed in the last row of Table A.1. The |γres| increases with increasing L according to a
power law γres ~ L3.3 (see Fig. A.3). The details of these studies are published already
[H. Meier, D. Ickenroth, U. Stalmach, K. Koynov, A. Bahtiar, and C. Bubeck, Eur. J.
Org. Chem. 23 (2001), pp. 4431-4443].
0,8 0,9 1 2 3 41E-33
1E-32
1E-31
3.3
|γ res|
[esu
]
L [nm]
Fig. A.3: Double logarithmic plot of molecular hyperpolarizability γresof OPEs 1a – e versus chain length L, the dashed line represents an increase from 1a to 1b according to the power law γres ~ L3.3.
_____________________________________________________________________133
Table A.1: Linear and nonlinear optical properties of the oligomers 1a – e in highly diluted solutions (CHCl3) and in thin films of polystyrene (PS) with (10.00 ± 0.03) % (mass) of OPE.
Compound 1a 1b 1c 1d 1e
Repeat units n 1 2 3 4 5
Chain length L [nm] 0.952 1.633 2.314 2.995 3.676
Absorption in CHCl3
λmax [nm] (± 1 nm) 338 378 399 412 419
εmax [L mol-1 cm-1] (± 3 %) 15900 39000 54700 80600 112100
Absorption in PS films
λmax [nm] (± 3 nm) 335 377 394 410 417
αmax [104 cm-1] (± 3 %) 0.923 1.220 1.481 1.781 2.115
THG of films
χ(3)res [10-13 esu] 2.05 ± 0.4 4.82 ± 0.3 7.96 ± 0.6 12.1 ± 0.8 13.4 ± 1.2
Φ [°], (± 10°) 33 66 86 84 94
γres [10-32 esu] 0.134 ± 0.03 0.81 ± 0.1 2.05 ± 0.3 3.93 ± 0.5 5.57 ± 0.8
A.2 Poly[p-phenyleneethynylene] (PPE)
The conjugated polymer poly[p-phenyleneethynylene] (PPE) was obtained from
the group of Prof. K. Müllen (Max-Planck Institute for Polymer Research, MPIP Mainz).
The chemical structure is shown in Fig. A.4. The polymer was dissolved in toluene and
the solution was heated to approximately 500C in order to obtain a fully soluble solution.
The transparent and homogeneous films of PPE were prepared by spin coating at elevated
temperature (T = 500C). The spectrum of the linear absorption coefficient α(λ) and
dispersion of the refractive index n(λ) are displayed in Fig. A.4.
_____________________________________________________________________134
300 350 400 450 500
0
10
20
30(a)
C C
C6H13
C6H13n
α [1
04 cm
-1]
400 600 800 1000 12001.0
1.5
2.0
2.5(b)
thin film waveguide
λ [nm]
n
Fig. A.4: Spectrum of linear absorption coefficient α(λ) of a thin film of PPE and the chemical structure is shown in the inset (a) and dispersion of refractive index of thin film (line) and data of refractive indices of waveguide (b), measured at TE polarization. Data points of the waveguide are extracted from dissertation of Ziegler [Ziegler’00].
We have measured the nonlinear optical properties of thin film of PPE by means
of THG. The dispersion of the modulus of χ(3) at 1/3 of the fundamental wavelength
compared with the linear absorption spectrum is shown in Fig. A.5. The spectrum of χ(3)
resembles the linear absorption coefficient: it has a maximum at the laser wavelength λL ≈
3λmax. The χ(3) values at several wavelengths are listed in Table A.2.
_____________________________________________________________________135
300 350 400 450 500
0
10
20
30
χ(3) [1
0-11 e
su]
α [1
04 cm
-1]
λ, λL/3 [nm]
0
2
4
6
8
Fig. A.5: Dispersion of the modulus of χ(3) at λL/3 in comparison with the linear absorption spectrum of a thin film of PPE. Both types of spectra were measured at TE-polarization.
Table A.2: Values of the modulus of χ(3) and phase angles of PPE measured at several laser wavelengths.
Polymer λL/3 [nm] χ (3) [10-11 esu] Φ [o]
PPE
355 370 385 400 425 437 440 450 470
3.50 3.40 4.00 5.37 4.80 6.10 7.96 5.74 2.11
140 131 126 91 84 68 68 14 3
_____________________________________________________________________136
Appendix B
Full Chemical Names of PPV Derivatives
MEH-PPV: Poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene-vinylene].
MEH-DOO-PPV: Poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene-vinylene-2,5-
dioctyloxy-1,4-phenylene-vinylene].
M3EH-PPV: Poly[2,5-dimethoxy-1,4-phenylene-1,2-ethenylene, 2-methoxy-5-
(2-ethylhexyloxy)-1,4-phenylene-vinylene].
MEH-M3EH-PPV: Poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene-vinylene-2,5-
dimethoxy-1,4-phenylene-vinylene-2-methoxy-5-(2-
ethylhexyloxy)-1,4-phenylene-vinylene]
TPD2-PPV [TPD(4M)-MEH-PPV]: Poly[1,4-phenylene-(4-methylphenyl)imino-4,4’-
diphenylene-(4-methylphenyl)imino-1,4-phenylene-vinylene-2-
methoxy-5-(2-ethylhexyloxy)-1,4-phenylene-vinylene].
TPD4-PPV[TPD(4M)-MEH-M3EH-PPV]: Poly[1,4-phenylene-(4-methylphenyl)imino-
4,4’-diphenylene -(4-methylphenyl)imino-1,4-phenylene-vinylene-
2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene-vinylene]CO[2,5-
dimethoxy-1,4-phenylene-vinylene-2-methoxy-5-(2-
ethylhexyloxy)-1,4-phenylene-vinylene] (50:50).
_____________________________________________________________________137
Appendix C
THG Data of PPV-Derivatives
Table C.1: Values of the modulus of χ(3) and phase angles of several PPV-derivatives measured at several laser wavelengths.
No. POLYMER λL/3 [nm] χ(3) [10-11 esu] Φ [0]
1
MEH-PPV (HT-137)
λmax = 489 nm
400 420 440 460 480 490 500 510 520 540 560 580 600 620
1.75 2.13 2.68 3.40 4.00 4.40 4.93 6.50 6.16 4.70 3.56 2.11 1.30 1.10
220 190 164 138 120 113 103 88 55 46 36 17 1 0
2
MEH-DOO-PPV
(B 185) λmax = 490 nm
425 450 500 520 550 600 630
3.57 4.67 7.30 8.60 6.80 3.00 2.10
237 229 109 69 55 15 0
3
MEH-M3EH-PPV
(HT 143) λmax = 485 nm
367 400 433 467 490 500 510 533 567 600
2.47 3.50 5.23 7.89 8.89 9.65 9.63 8.39 7.53 2.82
202 191 179 176 121 119 89 76 50 0
_____________________________________________________________________138
No. POLYMER λL/3 (nm) χ (3) [10-11 esu] Φ [o]
4
M3EH-PPV
(B 249) λmax = 486 nm
433 467 500 510 550 567 600 633
8.78 11.55 13.34 13.80 10.90 9.40 4.70 3.10
218 196 146 119 96 40 0 0
5
TPD2-PPV (HT 132/1)
λmax = 424 nm
355 370 400 430 450 500 550
0.17 1.70 2.53 3.30 3.70 1.80 0.72
189 152 143 111 77 36 0
6
TPD4-PPV (HT 134/1)
λmax = 417 nm
367 400 430 450 500 550 600
2.10 2.35 2.86 3.43 2.49 1.59 0.78
156 195 144 103 53 10 0
_____________________________________________________________________139
Appendix D Properties of MEH-PPVs D.1 GPC Curves of MEH-PPVs
103 104 1050.0
0.2
0.4
0.6
0.8
1.0
Wei
ght f
ract
ion
(Wn)
Molecular weight (Mx)
Fig. D.1: Molecular weight distribution of Gilch route MEH-PPV (Mn = 4.8 x 103 g/mol, Mw = 9.3 x 103 g/mol, PDI = 1.9). The polymer was obtained from American Dye Source (ADS), Canada.
103 104 1050.0
0.2
0.4
0.6
0.8
1.0
Wei
ght f
ract
ion
(Wn)
Molecular weight (Mx)
Fig. D.2: Molecular weight distribution of Horner route MEH-PPV (Mn = 6.4 x 103 g/mol, Mw = 1.3 x 104 g/mol, PDI = 2.0). The polymer was synthesized by Dr. T. Ahn (MPIP-Germany).
_____________________________________________________________________140
0
1
2 3 4 5 6 7LOG M (PS standard)
Wn
(a.u
.)
Mn (GPC) 14.100Mw (GPC) 40.300
34.600
Fig. D.3: Molecular weight distribution of Horner route MEH-PPV (Mn = 4.03 x104 g/mol, Mw = 1.41 x 104 g/mol, PDI = 2.8). The polymer and the curve were provided by Prof. H.-H. Hörhold (University of Jena, Germany).
104 105 1060.0
0.1
0.2
0.3
0.4
0.5
0.6
Wei
ght f
ract
ion
(Wn)
Molecular weight (Mx)
Fig. D.4: Molecular weight distribution of Gilch route MEH-PPV (Mn = 2.46 x 104 g/mol, Mw = 1.28 x 105 g/mol, PDI = 5.2). The polymer was synthesized by Dr. T. Ahn (MPIP-Germany).
_____________________________________________________________________141
104 105 1060.0
0.2
0.4
0.6
0.8
1.0
Wei
ght f
ract
ion
(Wn)
Molecular weight (Mx)
Fig. D.5: Molecular weight distribution of Gilch route MEH-PPV (Mn = 8.71 x 104 g/mol, Mw = 2.65 x 105 g/mol, PDI = 3.0). The polymer was obtained from American Dye Source (ADS), Canada.
Fig. D.6: Molecular weight distribution of Gilch route MEH-PPV (Mn = 1.05 x 104 g/mol, Mw = 2.76 x 105 g/mol, PDI = 2.6). The polymer was synthesized and measured by Dr. T. Ahn (KAIST-Korea).
_____________________________________________________________________142
104 105 1060.0
0.2
0.4
0.6
0.8
1.0
Molecular weight (Mx)
Wei
ght f
ract
ion
(Wn)
Fig. D.7: Molecular weight distribution of Gilch route MEH-PPV (Mn = 1.08 x 105 g/mol, Mw = 4.2 x 105 g/mol, PDI = 3.9). The polymer was obtained from American Dye Source (ADS), Canada.
Fig. D.8: Molecular weight distribution of Gilch route MEH-PPV (Mn = 1.3 x 105 g/mol, Mw = 1.6 x 106 g/mol, PDI = 12.3). The polymer was obtained from Covion GmbH, Germany.
_____________________________________________________________________143
Table D.1: Waveguide propagation loss coefficients of MEH-PPV with different molecular weights and synthetic routes, measured at 1064 nm and both TE and TM polarizations.
Polymer Synthetic route αgw (TE0) [dB/cm] αgw (TM0) [dB/cm]
1 Gilch 0.5 ± 0.3 0.7 ± 0.3
2 Horner 0.9 ± 0.4 0.6 ± 0.3
3(a) Horner 0.5 ± 0.3 -
4 Horner 0.5 ± 0.3 0.5 ± 0.3
5 Gilch 5.1 ± 0.4 0.5 ± 0.3
6 Gilch 6.2 ± 1.6 0.5 ± 0.3
7 Gilch 6.4 ± 1.6 1.0 ± 0.3
8 Gilch 30 ± 5.0 0.6 ± 0.3 (a) Data was extracted from literature. [F. Fitrilawati et al, Opt. Mater. 21 (2002), 511]. (-) No data of αgw (TM0) was measured because polymer 3 was not available any more.
Table D. 2: Values of the modulus χ(3) and phase angles of MEH-PPVs with different molecular weights and synthetic routes, measured at several laser wavelengths.
Polymer Route Mw [103 g/mol] λL/3 [nm] χ (3) [10-11 esu] Φ [0]
1
Gilch
9.3
400 420 440 460 480 490 500 520 540 560 580 600 620
1.59 1.68 1.95 2.24 2.43 2.40 2.35 2.20 1.64 1.40 1.03 0.64 0.63
206 160 132 119 92 76 72 56 21 2 0 0 0
_____________________________________________________________________144
Polymer Route Mw [104 g/mol] λL/3 [nm] χ (3) [10-11 esu] Φ [0]
2
Horner
1.32
400 420 440 460 480 490 500 520 540 560 580 600 620
1.32 1.66 2.07 2.41 2.78 2.65 2.48 2.38 2.10 1.76 1.21 0.70 0.44
196 172 139 120 89 76 56 35 18 0 0 0 0
3
Horner
2.50
400 420 440 460 480 490 500 510 520 540 560 580 600 620
1.42 2.11 2.35 3.12 3.78 3.84 3.68 3.49 3.38 2.89 2.57 1.61 1.11 1.08
205 183 152 133 105 91 81 53 42 18 0 0 0 0
4
Horner
4.03
400 420 440 460 480 490 500 510 520 540 560 580 600 620
1.75 2.13 2.68 3.40 4.00 4.40 4.93 6.50 6.16 4.70 3.56 2.11 1.30 1.10
220 190 164 138 120 113 103 88 55 46 36 17 1 0
_____________________________________________________________________145
Polymer Routes Mw [105 g/mol] λL/3 [nm] χ (3) [10-11 esu] Φ [0]
6
Gilch
2.65 500
510 530
6.86 7.58 7.04
86 93 123
7
Gilch
2.76
400 420 440 460 480 490 500 520 540 560 580 600 620
2.93 3.72 4.08 4.50 5.62 6.61 7.67 6.97 6.22 5.60 4.07 2.84 2.42
221 190 160 140 104 97 84 61 29 11 0 0 0
8
Gilch
4.20
400 420 440 460 480 490 500 520 540 560 580 600 620
4.16 5.07 5.60 6.14 8.35 8.77 9.17 8.53 8.00 6.78 5.80 4.00 2.78
235 206 197 160 138 115 97 72 48 25 10 0 0
9
Gilch
16
400 420 440 460 480 490 500 520 540 560 580 600 620
3.50 4.08 5.16 5.65 6.71 7.86 8.95 8.18 7.42 6.48 4.61 2.24 3.06
220 189 164 138 124 113 92 71 34 14 0 0 0
_____________________________________________________________________146
Appendix E
Studies of Cyano-Ether-PPV
The polymer Cyano-Ether-PPV or CN-Ether-PPV (the chemical structure is
shown in Fig. E.1) was obtained from Prof. H.-H. Hörhold, University of Jena, Germany.
Its molecular weights are Mw = 2.81 x 104 g/mol, Mn = 1.36 x 104 g/mol, PDI = 2.1, and
Tg = 610C.
O CCN
CH CH CCN
n
OC8H17
C8H17O
Fig. E.1: Chemical structure of CN-Ether-PPV.
The polymer was dissolved in toluene and processed to thin films by spin coating
of filtered solutions (0.45 µm syringe filters) onto fused silica substrates. The films were
subsequently placed in a vacuum oven for about 6 hours at 50 0C to remove residual
solvent. We have studied the photostability of thin CN-Ether-PPV films by irradiating the
films with a high-pressure mercury lamp in combination with a glass filter with
transmission at 365 nm (UG-1 filter from Schott). Fig. E.2 shows the absorption
coefficient and refractive index of a fresh CN-Ether-PPV film and after different UV
irradiation times. The absorption coefficient α(λ) decreases and its maximum shifts
towards shorter wavelengths with increasing exposure time. The refractive index of the
bleached polymer also decreases significantly. We have compared the UV-
photosensitivity of this polymer with MEH-PPV (Fig. E.3). Its UV-photosensitivity
shows larger value than that of MEH-PPV. The FTIR spectra of thin CN-Ether-PPV film
on Si-substrate measured in transmission configuration (perpendicular incidence) after
different UV-irradiation times are shown in Fig. E.4. The changes of infrared spectra as a
function of exposure time are presented in Fig. E.5. The intensity of the bands at 1246
cm-1, 1502 cm-1, and 2210 cm-1 decreases with increasing exposure time. At the same
time, a new carbonyl band (1730 cm-1) appears and increases with increasing exposure
_____________________________________________________________________147
time. The assignment of these bands is shown in Tab. E.1. Surprisingly, the fluorescence
spectra of the film do not decrease with UV irradiation. The fluorescence intensity
(excited at 365 nm) increases with increasing number of measurements and the
fluorescence spectrum shifts to the shorter wavelengths, as shown in Fig. E.6. The
emission- and excitation spectra (excited at 365 nm) of fresh and bleached films are
shown in Fig. E.7. The structures in the excitation spectra are presumable artifacts due to
poor correction of the lines of the Xenon lamp.
300 400 500 6000
5
10
15(a)
α [1
04 cm
-1]
300 400 500 600 700 800 900
1.4
1.6
1.8
2.0 (b)
λ [nm]
n
Fig. E.2: Spectra of (a) the absorption coefficient α(λ) and (b) the refractive index n of a fresh CN-Ether-PPV film with a thickness of 67 nm (solid curve) and after UV irradiation at 365 nm with an intensity of 21.5 mW/cm2 at the sample position for 20 minutes (dashed curve), 40 minutes (dotted curve), and 60 minutes (dashed-dotted curve).
_____________________________________________________________________148
0 10 20 30
0.0
0.2
0.4
0.6
0.8 MEH-PPV CNE-PPV
SUV = 0.007
SUV = 0.034
∆α
max
/αm
ax (t
=0)
IA [a.u.]
Fig. E.3: UV photosensitivity of thin film of CN-Ether-PPV in comparison with thin MEH-PPV film (Mw = 4.03 x 104 g/mol).
2220 2200 1800 1600 1400 1200
0.00
0.02
0.04
0.06
17302210
1502 1246
Abs
orba
nce
[a.u
.]
Wavenumber [cm-1]
Fig. E.4: FTIR spectra of a fresh thin CN-Ether-PPV film on silicon substrate and after UV irradiation at 365 nm with an intensity of 21.54 mW/cm2 at the sample position for 20 minutes (dashed curve), 40 minutes (dotted curve), and 60 minutes (dashed-dotted curve).
_____________________________________________________________________149
0 20 40 60 80
-0.5
0.0
0.5
1.0
1.5
2.0 2210 cm-1
1730 cm-1
1502 cm-1
1246 cm-1
Cha
nge
in a
bsor
banc
e
Expose fluence [J/cm2]
Fig. E.5: Changes of absorbencies of several infrared bands of thin CN-Ether-PPV film as a function of UV intensity.
Table E.1: The infrared assignments of CN-Ether-PPV.
Wavenumber [cm-1] Assignment References
1246 ether O-C stretching [Cumpston’95]
1502 semicircular phenyl stretch [Bradley’87]
1730 carbonyl group [Cumpston’95]
2210 cyano triple band (C ≡ N) [Streitweiser’86]
References:
[Cumpston’95] B. H. Cumpston et al, Synth. Met. 73 (1995), 195.
[Bradley’86] D. D. C. Bradley, J. Phys. D: Appl. Phys. 20 (1987), 1389
[Streiweiser’86] A. Streitweiser and C. H. Heatschock,”Organische Chemie”, VCH-Weinheim, 1986.
_____________________________________________________________________150
500 550 600 650 7000
1x105
2x105
3x105
Inte
nsity
[a.u
.]
λ [nm]
Fig. E.6: Spectra of fluorescence of thin film of CN-Ether-PPV (excited at 365 nm) measured at several times by means of fluorescence spectrometer (SPEX Fluorog 2). The arrow indicates the increasing number of measurements.
300 400 500 600 7000.0
0.2
0.4
0.6
0.8
1.0
bleached film
bleached film
fresh film
fresh film
A, I
Fl [a
.u.]
λ [nm]
Fig. E.7: Spectra of absorption and fluorescence of fresh and bleached films of CN-Ether-PPV, measured by means of a SPEX Fluorog 2 (F212) spectrometer.
_____________________________________________________________________151
Appendix F
Studies of Polyfluorene Two different molecular weights of polyfluorene PF2/6 were studied:
(a). Mw = 6.01 x 104 g/mol, Mn= 2.64 x 104 g/mol. Source: Dr. T. Ahn (KAIST,
Korea)
(b). Mw = 2.29 x 105 g/mol, Mn= 9.8 x 104 g/mol. Source: C. Chi (MPIP, Mainz)
n
PF2/6
Fig. F.1: Chemical structure of polyfluoerene PF2/6
200 300 400 5000
5
10
15
20
25
30 Mw = 6.01 x 104 g/mol Mw = 2.29 x 105 g/mol
λ [nm]
α [1
04 cm
-1]
Fig. F.2: Spectra of linear absorption coefficient α(λ) of thin films (d ≈ 50 nm) with different molecular weights, measured at TE-polarization. Their molecular weights Mw are given in inset.
_____________________________________________________________________152
400 600 800 1000
1.4
1.6
1.8
2.0
2.2 Mw = 6.01 x 104 g/mol Mw = 2.29 x 105 g/mol
n TE
λ [nm]
Fig. F.3: Dispersions of linear refractive index at TE-polarization of thin films (d ≈ 50 nm) with different molecular weights, measured at TE-polarization. Their molecular weights Mw are given in inset.
600 700 800 900 1000
1.55
1.60
1.65
Mw = 6.01 x 104 g/mol Mw = 2.29 x 105 g/mol
TM
TE
n
λ [nm]
Fig. F.4: Refractive indices of spin-cast films (approximately 600 nm thick) prepared from PF2/6 of different molecular weights, measured by prism coupling at TE (full symbols) and TM (open symbols) polarizations.
_____________________________________________________________________153
Appendix G Studies of NRS (Super-Yellow) Material
According to provider (Covion GmbH, Germany), the polymer is called NRS or
Super-Yellow material. Its chemical structure is shown in Fig. G.1. [H. Becker et al, Adv.
Mater 14 (2000), 12]
OR
0.5
0.5
OR
OMe
n
R = *
Fig. G.1: The chemical structure of NRS polymer.
Table G.1: Molecular weights and index of polydispersity of NRS polymers.
Molecular weight Polymer A Polymer B Polymer C Polymer D
Mw [105 g/mol] 1.97 3.30 6.50 10.8
Mn [105 g/mol] 1.10 1.70 2.30 2.60
PDI 1.8 1.9 2.8 4.2 All polymers were provided as toluene solutions. They were processed to thin
films by spin coating of filtered solutions onto fused silica substrates. The films were
annealed in a vacuum oven for 6 hours at 500C. Thin films with thickness d ≈ 50 – 70 nm
were used for spectroscopic studies, and thick films (400 - 800 nm) for optical waveguide
_____________________________________________________________________154
experiments. Fig. G.2 shows the spectrum of the intrinsic linear absorption coefficient
and the dispersion of refractive index of polymer D (Mw = 1.08 x 106 g/mol). The optical
properties of polymers C and D are listed in Tab. G.1. We did not investigate thin film
properties of polymers A and B. The value of nTE and nTM at λ = 633 nm as a function of
Mw for all polymers A - D is shown in Fig. G.3.
400 600 800 10001.2
1.4
1.6
1.8
2.0
α [1
04 cm
-1]
nTE (thin film) nTE (waveguide) nTM (waveguide)
n
λ [nm]
300 400 500 6000
5
10
15
Fig. G.2: Spectrum of intrinsic linear absorption coefficient α(λ) at TE-polarization (top) and dispersion of refractive index of polymer D (Mw = 1.08 x 106 g/mol) (bottom). The line is from transmission-reflection spectroscopy at TE polarization. Data points are from prism coupling experiments at TE polarization (full squares) and TM polarization (open squares).
_____________________________________________________________________155
Table G. 2: Linear optical constants of NRS polymers, measured at TE polarization.
Optical Constants Polymer A Polymer B Polymer C Polymer D
λmax [± 3 nm] - - 465 465
αmax [104 cm-1 ± 5%] - - 15.5 15.8
αgw (TE0, 1064 nm) [dB/ cm] - - 8 ± 2 11 ± 3 (-) No data was available
0 2 4 6 8 10 121.5
1.6
1.7
1.8
λ = 633 nm
n
MW [105 g/mol]
nTE nTM
Fig. G.3: Molecular weight dependence of refractive index and birefringence of waveguides of polymers A - D measured by prism coupling at 633 nm.
_____________________________________________________________________156
List of Publications
1. H. Meier, D. Ickenroth, U. Stalmach, K. Koynov, A. Bahtiar, and C. Bubeck, “Preparation and Nonlinear Optics of Monodisperse Oligo(1,4-phenyleneethynylene)s,” Eur. J. Org. Chem. 23 (2001), pp. 4431-4443.
2. K. Koynov, N. Goutev, F. Fitrilawati, A. Bahtiar, A. Best, C. Bubeck, H.-H.
Hörhold, “Nonlinear prism coupling of MEH-PPV waveguides and their figure of merit for all-optical switching,” J. Opt. Soc. Am. B 19 (2002), pp. 895-901.
3. M. A. Bader, G. Marowsky, A. Bahtiar, K. Koynov, C. Bubeck, H. Tillmann, H.-
H. Hörhold, S. Pereira,”PPV-Derivatives: New Promising Materials for Nonlinear All-Optical Waveguide Switching,” J. Opt. Soc. Am. B 19 (2002), pp. 2250-2262.
4. K. Koynov, A. Bahtiar, T. Ahn, H.-H. Hörhold, C. Bubeck, "Molecular weight
dependence of birefringence of thin films of the conjugated polymer MEH-PPV", Appl. Phys. Lett. 84 (2004), 3792-3794.
5. K. Koynov, A. Bahtiar, T. Ahn, H.-H. Hörhold, C. Bubeck, "Molecular weight
dependence of waveguides birefringence of thin films and chain orientation of the conjugated polymer MEH-PPV", in preparation.
6. A. Bahtiar, K. Koynov, T. Ahn, H.-H. Hörhold, C. Bubeck, "Nonlinear Optical
Spectroscopy of the conjugated polymer MEH-PPV", in preparation. 7. A. Bahtiar, K. Koynov, T. Ahn, H.-H. Hörhold, C. Bubeck, ”Influence of
Molecular Weight on Third-Order Nonlinear Optical Susceptibility of MEH-PPV”, in preparation.
_____________________________________________________________________157
Acknowledgements
Experimental research is almost impossible without the close interaction of many people.
I am deeply indebted to all those involved in the research that I present in this thesis for
their help, guidance, friendship, patience and their constructive criticism. I am especially
grateful to:
Prof. Dr. C. Bubeck, my supervisor, for providing me this interesting and challenging
research topic. I would like also to express my gratitude and appreciation for his
guidance through the many scientific and non-scientific discussions. Discussions with
him have often led to find new perspectives or deeper insight into observations. I
benefited greatly from the opportunity he gave me to present my work at workshops
and conferences and his strive for improvement of such presentations. I am deeply
indebted also for his time and patience for reading and correcting the manuscript
completely.
Prof. Dr. W. Knoll for his kindly welcome to his multidisciplinary group and for
allowing me to work at the institute.
Dr. K. Koynov for his continuous motivation and encouragement to continue working
during troublesome and frustrating time. Discussions with him have been a crucial
step for a deeper understanding of all aspects of this thesis. Moreover, his extensive
knowledge of experimental technique has been of great help to me in many cases. I
would like to thank for his critical proof-reading the manuscript of this thesis.
Prof. H.-H. Hörhold for providing me many interesting conjugated polymers.
Prof. E. W. Otten, for his kindly agreeing to be my supervisor at the Faculty of
Physics, Johannes-Guttenberg-Universität Mainz.
DR. T. Ahn for his explanations, discussions about Chemistry and Synthesis. Special
thanks also for providing me some interesting materials. Playing table tennis with him
is always exciting.
_____________________________________________________________________158
Former and current members of AG-Bubeck: Dr. F. Fantinel, A. Best, Dr. T. Ahn,
Dr. K. Koynov, Dr. A. Kaltbeitzel, R. Cordeiro, W. Scholdei, G. Hermann, H. J.
Menges, who made great atmosphere and so many occasions to have fun together.
Dr. M. A. Bader and Dr. K. Petersen for their collaboration in fabrication of
microstructure in the polymer films.
G. Hermann, W. Scholdei, H. J. Menges, B. Menges, U. Rietzler and C. Schwind for
their technical supports.
S. Koynova and her family for their warmth and encouragement.
DAAD (Deutsche Akademischer Austauschdienst) for providing me the fellowship to
continue my study and research in Germany.
Prof. Dr. R. E. Siregar, Dr. Fitrilawati, and my colleges at the Department of Physics,
University of Padjadjaran Indonesien for their help and support during my study in
Germany.
Last but not least, I have to thank my fiancé and my family for their great support during
the time of my study in Germany and for their understanding for the little time I was able
to spend with them during the past years.
I hope for understanding that it is impossible to mention everyone individually.
_____________________________________________________________________159
Curriculum Vitae Personal Information ________________________________________________________________________ Name: Ayi Bahtiar Address: Department of Physics, University of Padjadjaran Bandung Jl. Raya Jatinangor km. 21 Sumedang 45363 Indonesien Date of birth: 29.10.1970 Pace of birth: Ciamis, Indonesien School Education ________________________________________________________________________ 1977-1983 Elementary School in Indonesien 1983-1986 Junior High School in Indonesien 1986-1989 Senior High School in Indonesien University Education ________________________________________________________________________ 1989-1994 Studies in Physics, University of Padjadjaran Bandung, Indonesien 1994 Bachelor in Physics 1995-1997 Postgraduate studies, Bandung Institute of Technology, Indonesien 1997 Master of Science in Physics 2000-2004 Dissertation under Prof. Dr. C. Bubeck, Max-Planck Institute for
Polymer Research, Mainz, Germany Additional Information ________________________________________________________________________ Since 1996 Staff member at Department of Physics, University of Padjadjaran
Bandung, Indonesien
