EFFECTIVE NONLINEAR SUSCEPTIBILITIES OF METAL-INSULATOR AND
METAL-INSULATOR-METAL NANOLAYERED STRUCTURES
Dissertation
Submitted to
The School of Engineering of the
UNIVERSITY OF DAYTON
In Partial Fulfillment of the Requirements for
The Degree of
Doctor of Philosophy in Electro-Optics
By
Mallik Mohd Raihan Hussain
Dayton, Ohio
May, 4242 EFFECTIVE NONLINEAR SUSCEPTIBILITIES OF METAL-INSULATOR AND
METAL-INSULATOR-METAL NANOLAYERED STRUCTURES
Name: Hussain, Mallik Mohd Raihan
APPROVED BY:
Imad Agha, Ph.D. Andrew Sarangan, Ph.D. Advisory Committee Chairman Committee Member Associate Professor, Department of Professor, Department of Physics, and, Department of Electro- Electro-Optics and Photonics Optics and Photonics
Partha Banerjee, Ph.D. Michael Scalora, Ph.D. Committee Member Committee Member Professor and Department Chair, Research Physicist, Charles M. Department of Electro-Optics and Bowden Research Facility, Photonics AMRDEC, US Army RDECOM
Robert J. Wilkens, Ph.D., P.E. Eddy M. Rojas, Ph.D., M.A., P.E. Associate Dean for Research Dean, School of Engineering and Innovation Professor, School of Engineering
ii © Copyright by
Mallik Mohd Raihan Hussain
All rights reserved
4242 ABSTRACT
EFFECTIVE NONLINEAR SUSCEPTIBILITIES OF METAL-INSULATOR AND
METAL-INSULATOR-METAL NANOLAYERED STRUCTURES
Name: Hussain, Mallik Mohd Raihan University of Dayton
Advisor: Dr. Imad Agha
Nonlinear electromagnetic radiation (second and third harmonic) from the metal-insulator and metal-insulator-metal structures were measured and compared against predictions from the hydrodynamic models of plasmonics. This model incorporated higher order terms stem- ming from electron tunneling and nonlocality. This study shows that, besides the linear optical parameter like permittivity, conductivity etc, changes in the nonlinear optical pa- rameters, namely, second and third order susceptibilities (χ(2) and χ(3), respectively) can also be used to probe and compare the higher-order terms of the hydrodynamic model of plasmonics. Two insulator materials (ZnO and Al2O3) were used in two separate sets of experiments, and atomic layer deposition was used to cover the gold substrate with variable thicknesses of these insulator films (nanometer to sub-nanometer range). Large reduction in second and third harmonic signals was measured after the insulator film was deposited over the gold substrate revealing the spilled-out electronic states in the insulator region at the vicinity of the metal-insulator interface, which are dubbed metal insulator gap states.
Then, the metal-insulator samples were spin-coated with Au-nanoparticle solution to pre- pare a metal-insulator-metal structures. For these structures, saturation and quenching of the third harmonic efficiencies were observed which were indicative of the capping of E-field enhancement due to the existence of higher order terms in the hydrodynamic model that accounts for nonlocality and quantum tunneling of electrons. A generalized 4 × 4 matrix
iii method was utilized to calculate the effective χ(2) and χ(3) parameters that confirm the changes of effective material properties for ultra-thin films. These nonlinear coefficients, besides the linear permittivity and conductivity σ, can be a useful material parameter to study the effects of higher-order terms of hydrodynamic model.
iv To
my mentor,
Dr. Joseph W. Haus,
my parents,
Nasima Khatun
&
Mohd Shakhawat Hussain Mallik.
v ACKNOWLEDGMENTS
A lot has changed since I started my journey as a Ph.D. student. My ex-advisor, Dr.
Joseph Haus suddenly passed away on January 33, 423;. I would not have come this far without him. He used to challenge me with hard problems and give me the space to work on novel ideas. I always admired his openness and humble, close-to-earth demeanor during technical/nontechnical discussions. After Dr. Haus’s death, I was kindly picked up by my current advisor Dr. Imad Agha. I would like to thank him from the deepest of my heart for supporting me till the end. I appreciate his patience and encouragements during every meeting. I have truly found a friend in him and I am looking forward to grow this friendship into a life-long research collaboration tackling the frontiers of optics. I would like to extend my sincere gratitude towards Dr. Andrew Sarangan. His support through the Laser-assisted Si wet etch project has helped me immensely. I have learned a lot from his teaching style and attitude towards solving a problem. In October 423;, Dr. Partha
Banerjee suggested me to work with Dr. Behzad Bordbar on the ONR project and he finally offered me a position as a ’Research Engineer’. That opportunity has been a blessing to me and I can never thank Dr. Banerjee enough for trusting in me. I will keep working hard. I would like to thank Drs. Michael Scalora, Domenico de Ceglia, Maria Vincenti,
Parag Banerjee and Zhengning Gao for making me a part of a wonderful collaboration. I look forward to contribute more to this collaboration in future. I would like to thank Dr.
Shekhar Guha for promptly replying to my emails and queries. I found his book titled
“Laser Beam Propagation in Nonlinear Optical Media” to be very rigorous and detailed.
The book helped me with many concepts. I would also like to thank Dr. Qiwen Zhan for his insightful suggestions on how to incorporate evanescent/inhomogeneous waves in TMM.
I have had many friends in the department over the years. I would like to thank all of them for being my well-wishers. I would like to thank the University of Dayton for creating such
vi a warm sense of community and being my home away from home. Lastly, I would like to thank my dear wife Syeda Prem Ara Bahar for supporting me through thick and thin. The world is going through a pandemic. In such situation, I can not defend my dissertation in person and thank and shake hands with everyone who congratulates me. Yet, I feel everyone’s blessings are with me.
vii TABLE OF CONTENTS
ABSTRACT ...... iii
DEDICATION ...... v
ACKNOWLEDGMENTS ...... vi
LIST OF FIGURES ...... x
LIST OF TABLES ...... xv
CHAPTER I. INTRODUCTION ...... 3
3.3 Problem statement and novel contributions ...... 5 3.4 Organization of ideas ...... 7
CHAPTER II. LIGHT PROPAGATION IN STRATIFIED MEDIA ...... 9
4.3 Notation convention ...... 32 4.3.3 Subscripts and superscripts ...... 32 4.3.4 Vectors and scalars ...... 33 4.3.5 Tensors ...... 34 4.3.6 Matrices ...... 34 4.4 Material birefringence and layer geometry ...... 35 4.5 Maxwell’s equations ...... 37 4.6 Free mode calculation ...... 43 4.6.3 Free K-mode calculation ...... 44 4.6.4 Free E-mode calculation ...... 46 4.6.5 Boundary conditions for free modes ...... 4; 4.6.6 Reflectance and transmittance ...... 56 4.7 Bound mode calculation ...... 64 4.7.3 Bound K-mode calculation ...... 64 4.7.4 Bound E-mode calculation ...... 66 4.7.5 Boundary conditions with free and bound modes ...... 67 4.8 A interim summary of the TMM algorithm ...... 74 4.9 What is effective susceptibility? ...... 76 (2),2ω 4.9.3 deff calculation ...... 77 (3),3ω 4.9.4 deff calculation ...... 78 4.: Theoretical models of susceptibility ...... 79 4.:.3 Herman-Hayden model ...... 79 4.:.4 Effective medium model ...... 79 4.:.5 Nonlocality and tunneling ...... 82
CHAPTER III. EXPERIMENTS AND RESULTS ...... 83
5.3 Experimental setup ...... 83 5.4 Results ...... 86 5.4.3 Effective d(2),2ω calculation ...... 87 5.4.4 Effective d(3),3ω calculation ...... 8: 5.5 Electron density in metal-induced-gap-states ...... 92
viii 5.6 Effects of nonlocality and quantum tunneling ...... 95 5.7 Summary of findings ...... 96
CHAPTER IV. CONCLUSION ...... 98
BIBLIOGRAPHY ...... 9:
APPENDICES
A. Surface Characterization ...... :8
A.3 Thickness and growth rate ...... :8 A.4 Surface roughness ...... :9 A.5 Surface coverage ...... :; A.6 Chemical composition ...... :; A.7 Au nanoparticle distribution ...... ;2
B. Singular Value Decomposition ...... ;4
C. MATLAB Codes ...... ;5
ix LIST OF FIGURES
3.3 Miniaturization of electronics has opened a new paradigm of nanoplasmonics that can serve as an interface between electronics and photonics [38]...... 4
4.3 A stratified layered structure with arbitrary numbers of layers and layer thick- nesses...... 9
4.4 (a)Resonant metamaterial structures (e.g. split-ring resonators show resonance in dispersion relation) are used to design left-handed-materials, whereas, (b) a non-resonant metamaterial structure is designed to engineer exotic anisotropic behaviour...... :
4.5 The transformation between the laboratory axes system and the principle axes system of the crystal. The symbols of Euler angles (Φ, Θ, Ψ) are uppercase and color coded to green, red and green respectively. Symbols for angle of incidence and azimuthal angle (θ, φ) are lowercase...... 36
4.6 Vector diagram of EM fields (not to scale and color coded for easy correspondence). 38
4.7 An example case of the normal surface of a positive uniaxial medium with an arbitrary direction of oˆ with respect to the laboratory axes system (xˆyˆzˆ). Four (1),Ω (1),Ω free K-modes (Kl ) along with their angles (θl ) with respect to zˆ is marked. 44
4.8 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of an isotropic material for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 47
4.9 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of a positive uniaxial material (arbitrary orientation of optic axis, oˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 47
4.: Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of a negative uniaxial material (arbitrary orientation of optic axis, oˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 48
4.; Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of a type-3 hyperbolic uniaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 48
4.32 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of a type-4 hyperbolic uniaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 49
x 4.33 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of positive biaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 49
4.34 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of negative biaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 4:
4.35 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of type-3 hyperbolic biaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 4:
4.36 Index ellipsoid indicating the directions of free K-modes and free E-modes (scaled for easy visualization) of type-4 hyperbolic biaxial material (for optic axis, oˆ k zˆ) for 0◦ ≤ θ < 90◦ and 0◦ ≤ φ < 360◦...... 4;
4.37 Generalized Snell’s law for (linear) free K-modes. The components of K parallel to the interface remains the same for all free K-modes. An isotropic/uniaxial (1),Ω interface is chosen for an example case. Both normal surfaces and Klk are marked for reference...... 52
4.38 Boundary condition of free E-modes for linear cases. The D¯, P¯andE¯ are pictori- ally represented...... 55
4.39 The reflectance (R) and transmittance (T ) of glass(n = 1.5)/Air(n = 1) for (a) s-polarized and (b) p-polarized light...... 59
4.3: The reflectance (R) and transmittance (T ) of Air(n = 1)/Calcite(no = 1.655, ne = 1.485) for (a) s-polarized and (b) p-polarized light. Optic axis direction is 45◦ from zˆ in the zˆyˆ plane. The results match with transmittance/reflectance found in Figure 5.5 of Reference [59]...... 5:
4.3; The reflectance (R) and transmittance (T ) of Air(n = 1)/calcite (no = 1.655, ne = 1.485)/Air(n = 1) for (a) s-polarized and (b) p-polarized light. The thick- ness of calcite is 1500 nm. The Euler angles of rotation to transform oˆ into the the direction of wˆ is [30◦, 60◦, 0◦]...... 5;
4.42 The reflectance (R) and transmittance (T ) of glass(n = 1.5)/Chiolite(no = 1.349, ne = 1.342) for (a) s-polarized and (b) p-polarized light. Optic axis direction is 45◦ from zˆ in the zˆyˆ plane. The results match with transmittance/reflectance found in Figure 5.7 of Reference [59]...... 62
xi 4.43 The reflectance (R) and transmittance (T ) of glass(n = 1.5)/Chiolite(no = 1.349, ne = 1.342)/air(n = 1) for (a) s-polarized and (b) p-polarized light. Optic axis direction is 45◦ from zˆ in the zˆyˆ plane. The thickness of Chiolite crystal is 1500 nm...... 63
4.44 Generalized Snell’s law for bound (nonlinear) cases. A uniaxial layer is chosen (1),Ω (2),Ω for an example case. The normal surfaces and Klk , Klk are marked for reference...... 65
4.45 Maker fringe pattern from a Quartz layer for second harmonic. Quartz layer thickness is 4477.4µm. The blue line is the simulated SH signal, black line is the envelope of the simulated SH signal and the red circles are experimental SH signal collected from Reference [3:, 63]...... 6;
4.46 Maker fringe pattern from a ZnO layer for SH signal. ZnO layer thickness is 511µm. The blue line is the simulated SH signal, black line is the envelope of the simulated SH signal and the red circles are experimental SH signal collected from Reference [3;]...... 6;
5.3 Experimental setup: Mode-locked Ti:Sapphire laser, chopper; HWP: half wave plate L3: focusing lens; L4: collimating lens; P3 and P4: filtering prisms; SPF: short-pass filter; L5: detector lens; PD: (silicon) photodiode [94]...... 83
5.4 The SH and TH signals are collected, first, from an ultra-flat Au surface, then, thin film of insulator (on top of the Au surface), and lastly, from a thin film of insulator coated with AuNPs...... 85
5.5 SH efficiencies measured for samples with (a) Al2O3 and Al2O3 + AuNP , and (b) ZnO and ZnO + AuNP as insulator layer...... 86
5.6 TH efficiencies measured for samples with (a) Al2O3 and Al2O3 + AuNP , and (b) ZnO and ZnO + AuNP as insulator layer...... 86
(2) 5.7 |d15 | for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. . . . 87
(2) 5.8 |d31 | for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. . . . 88
(2) 5.9 |d33 | for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. . . . 88
(2) 5.: Re{deff } for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. 89
(2) 5.; Im{deff } for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. 89
(3) 5.32 |d16 | for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. . . . 8:
xii (3) 5.33 Re{deff } for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. 8;
(3) 5.34 Im{deff } for (a) Al2O3 and (b) ZnO as insulator for MI and MIM structures. 8;
5.35 The insulator layer (Al2O3) is considered as effective medium. The permittivity of Al2O3 is modified by the MIGS spillage and surface scattering due to the Au surface and, then, by the surface coverage ratio [:;]...... 92
5.36 The reflected SH efficiency (left axis) as a function of ALD Al2O3 cycles. Corre-
spondingly, the coverage factor, fAl2O3 (right axis) varies as a function of ALD cycles from 2, i.e. no Al4O5 coverage to 3, i.e. full Al4O5 coverage at cycle 69. The MIGS factor α has been shown to vary from a = 0.0 to a = 0.7...... 94
5.37 For different values of MIGS factor α, the RMS error between simulated and measured harmonic power is calculated. The minimum RMS error is found for MIGS factor α = 0.13...... 95
5.38 (a) Comparison between four different simulations. The Classical results omit nonlocality and quantum effects; the Quantum effects incorporate electron tun- neling and the nonlocal calculation incorporates spatial dispersion effects. The nonlocal + quantum curve includes both effects. (b) Field enhancement distri- bution near the gap region for an MIM with a gap size of 0.2nm for nonlocal + quantum effects...... 96
A.3 The growth rate (change of thickness per unit cycle of ALD) was measured for each insulator...... :8
A.4 The AFM image showing the topography of bare Au surface and Au surface
coated with (a) Al2O3 and (b) ZnO. Each image is labeled with the number of ALD cycles. The inset shows a length scale of 100nm. The bar on the right shows the heigt range...... :9
A.5 The schematic of nucleation and growth model. (a) pill box, (b) nucleation inside pill-boxes, The yellow represents Au and the white represents the deposited insulator, (c) critical point 3: when the nucleation circles touch each other and the sides of the pill box, (d) critical point 4: when the nucleation circles touch the corners of the pillbox. At this point the growth envelop is marked by the black curve...... ::
A.6 (a) The measured surface roughness of Al2O3 in the context of (b) the expected surface roughness from nucleation and growth model...... ::
A.7 XPS valence band spectra showing the Au valence band transitioning to the Zn (left) or Al (right) as the number of ALD cycle increases...... ;2
xiii A.8 The SEM image of Au surface spin-coated with Au nanoparticle solution. Au nanoparticles are (marked in green) randomly distributed and sparse enough to not cause any internal coupling of E-fields. The density of Au nanoparticles. . . ;3
xiv LIST OF TABLES
4.3 List of subscripts/superscripts...... 32
4.4 List of directions (unit vectors), angles, distances and planes related to the ge- ometry of the layers and material anisotropy (see Figure 4.5 for reference). . . 35
4.5 List of constants...... 37
4.6 List of vectors, directions, magnitude and angles related to EM fields...... 37
4.7 List of material tensors...... 39
4.8 List of matrices in the TMM engine [c() = cos(), s() = sin()]...... 72
xv CHAPTER I
INTRODUCTION
After the invention of laser in 3;82, the field of nonlinear light-matter interaction blos- somed to its fullest potential. Since then, researchers have used lasers to investigate the higher-order susceptibilities of materials in the optical frequency region. The linear prop- erties of material e.g. linear permittivity and conductivity provide insight into emerging properties like dispersion, absorption, electronic band structure etc. Similarly, the non- linear material properties i.e. susceptibilities carries the signature of material resonances that can only be probed at high intensity. Comprehensive understanding of different non- linear optical processes has created a whole range of applications [3–5], from academic research oriented tools to widespread industrial devices. For example, the second-order nonlinear effects are utilized in frequency doubled lasers [6], optical parametric oscillators
(OPOs) [7], optical parametric amplifiers (OPAs) [8], and quantum light sources [9]. The third-order nonlinear effects has given rise to self-phase and cross-phase modulation and intensity induced changes in refractive indices [3]. The process of four-wave mixing [:] is proven to be hugely beneficial in optical signal processing, optical switching [;], electro- optic modulators-demodulators [32], and optical correlators [33]. Also, second-harmonic imaging microscopy [34, 35] and surface second-harmonic generation to probe material in- terfaces [36, 37] are emerging as advanced characterization tools. While these applications readily justify further research in the field of nonlinear optics, the conjugation of nanoplas- monics with nonlinear optics needs further discussion to instigate completely new ideas of applications. Due to the high absorption coefficient of plasmonic materials in the optical frequency range, they may seem unsuitable at first. In spite of this apparent demerit, plas- monics offer several advantages. The benefits of plasmonic materials will be sought in light of the latest nanofabrication technology.
3 Figure 3.3: Miniaturization of electronics has opened a new paradigm of nanoplasmonics that can serve as an interface between electronics and photonics [38].
Plasmonics can be a natural interface for electronics and photonics in terms of size and operational speed (Figure 3.3), resulting into smaller, less costly and more efficient devices.
The advantages of introducing plasmonics to achieve nonlinear optical effects are two-fold:
(3) light confinement (i.e. large wavevectors along the plane of interface and large electric
field normal to the plane of interface may exist) and (4) surface or bulk plasmon-polariton modes. Although the high intensity of laser light has enabled photons of different frequencies to interact with each other through the bulk material, the efficiency of power exchange is limited by the phase-matching conditions, dispersion relations and degree of anisotropy of the material. Usually, the nonlinear optical devices that utilizes bulk material properties are much thicker then the wavelength of the light. With the advent of nanotechnology, photonic
4 crystals and metamaterials can be easily fabricated with thicknesses in range of or much smaller than the quarter-wavelength. Such length scales are not enough for efficient phase- matching. So, electric field localization i.e. large wavevector and high field enhancements are required. The introduction of metal or metal-like layers enables propagation modes with large wavevectors to exist causing highly confined light and, thus, more efficient light-matter interaction. Also, in the field of electronics, miniaturization has reached the nano-scale range. But the operation frequency e.g. the clock frequency is much slower then the optical frequency. The worlds of optics and electronics can be bridged by utilizing the physics of plasmonics. Further research in this field, namely, integrated optics, can enable nonlinear optical devices to be put into photonic integrated circuits reducing the footprint of regular nonlinear optical devices.
3.3 Problem statement and novel contributions
The discussion in the above section asserts the significance of combining nonlinear optics with the field of nanoplasmonics. With that motivation in mind, this dissertation attempts to address the topic of measuring effective nonlinear optical susceptibilities of nanostructures that contains metallic/plasmonic materials as components using a “transfer matrix method” technique. These nonlinear susceptibilities, as measured from the experiment, will also be derived from a detailed physical model known as hydrodynamic model. The novel contribution of this work is as follows:
• The effective values of second and third order susceptibilities were calculated from
the measured second and third harmonic efficiencies, respectively. In the literature,
the second and third order susceptibilities are measured using a well-known technique
called “Maker fringes technique” [39–42]. But this technique has not been developed
5 for measurement in reflection direction. Also, the analytic results in Reference [39]
only accounts for the first reflection from any interface. This dissertation addresses
both reflection and transmission measurements and can handle multiple reflections
at once. A generalized transfer matrix method (TMM) technique that incorporates
nonlinearity was employed to calculate these susceptibility values. Inside the TMM
engine, the propagation modes are sorted based on the real and imaginary part of the
component of wavevector along surface normal. Singular value deposition (SVD) was
utilized to calculate the electric field vectors. This work will soon be published.
• Effective medium theory was used to model the second order susceptibility due to
metal-interfaces. In comparison to the TMM technique (briefly introduced above;
complete explanation in Chapter II) where the susceptibility parameters were directly
fitted, a phenomenological parameter was fitted which is derived from a physical model
of susceptibility. This parameter, dubbed as metal-induced-gap-state parameter, in
turn, provided the areal density of free electrons in the vicinity of metal-insulator
interface. The work was published in Reference [43].
• A more detailed model was utilized to predict the second and third harmonic effi-
ciencies of metal-insulator-metal interfaces. The model, as mentioned above, is called
the hydrodynamic model and it accounts for the resonance terms from bound and
free electrons, convection terms, pressure terms or nonlocal terms and tunneling cur-
rent terms. The experimental confirmation of the comparative effects of these terms
in harmonic generation is also one of the novel contribution of this project. These
results were published in Reference [44].
6 3.4 Organization of ideas
This research offers three novel contributions as discussed above. The chapters of this dissertation has been organized to explain the theoretical background of each contribution.
In chapter II, light propagation in stratified media will be discussed. A generalized 4 × 4
transfer matrix method (TMM) is utilized to calculate Fresnel coefficients for any combi-
nation of isotropic and birefringent layers with arbitrary orientation of optic axes. The
method is then modified to include optical nonlinearity and generate Maker-fringe patterns
for thin layers. The formulae to calculate the second and third order susceptibilities for
(2),2ω (3),3ω second and third harmonic generation, respectively, (i.e. deff and deff , the symbols
are formally introduced in Chapter II, see Table 4.7 for quick reference) using the TMM
engines are introduced. Then, the hydrodynamic model is introduced to predict the sec-
ond and third harmonic generation. The Lorentz-Drude model is discussed that extends
to the hydrodynamic model including several higher order terms stemming from different
physical phenomena at nanometer length scales. Emphasis is given on electron Fermi pres-
sure terms and other diffusion-like terms that can result into a shift of resonance frequency
peaks and line broadening. Chapter III discusses the experiment in details. Then, the
(2),2ω (3),3ω deff and deff are calculated from the measured second and third harmonic radiation
originated from a thin-film of bare insulator (Al2O3 and ZnO) deposited on Au (metal-
insulator sample) and from the same insulator with metal-nanoparticles dispersed on top
(metal-insulator-metal sample). Measurements are done for different thicknesses of the
insulators. Then we predict the areal density of delocalized electrons in metal-induced-
gap-states from the SH efficiencies while taking the surface roughness and nucleation and
growth model of atomic layer deposition into account. Then, the measured second and
third harmonic efficiencies emanating from the thin-insulator films are compared against
7 the prediction of hydrodynamic model. Finally, in Chapter IV, conclusions were drawn and the scope of future work is discussed. Details of sample thickness measurement using ellipsometry, surface roughness characterization using atomic force microscopy, chemical composition characterization using X-ray photoelectron spectroscopy and Au nanoparticle distribution characterization using scanning electron microscopy are given in Appendix A.
The singular value decomposition is briefly described in Appendix B. The MATLAB codes are also appended at the end (Appendix C) for easy reference.
8 CHAPTER II
LIGHT PROPAGATION IN STRATIFIED MEDIA
The analysis of the propagation of light through a stratified media is crucial for designing
filters, waveguides, cavities, mirrors and anti-reflection/high-reflection coating in a required wavelength range [45, 46]. Figure 4.3 shows an arbitrary layered structures where the ith
layer has a thickness of ai. if the layers repeat periodically and the period thickness, a, is comparable to the quarter of the wavelength, Λ, inside that layer medium i.e. a ≈ Λ/4, the periodic layered structures are known as photonic crystals. For much thinner layers, the photons traversing the layers can not register them as separate medium. In stead, the layers can be combined into an “effective medium” with an “effective refractive index”.
The “effective medium theory” takes the geometry and symmetry properties of the layered structures into account [47] and calculates the corresponding “effective permittivitty”.
Figure 4.3: A stratified layered structure with arbitrary numbers of layers and layer thick- nesses.
9 In some case, if the layered structure shows resonance peaks in the dispersion curve for a certain frequency band, the structure is considered as a “resonant” metamaterial for that frequency range. These resonance peaks may be caused by the presence of specific geometry of metallic structures with the critical dimension, a, such that a ≈ Λ/10. Such structures are used to design materials with less-then-one and negative refractive index (left-handed materials) [48].
(a) (b)
Figure 4.4: (a)Resonant metamaterial structures (e.g. split-ring resonators show resonance in dispersion relation) are used to design left-handed-materials, whereas, (b) a non-resonant metamaterial structure is designed to engineer exotic anisotropic behaviour.
If a Λ/10, the stratified structures can be thought of as “non-resonant metamaterial”.
They are designed to achieve materials with effective anisotropy not found in the nature.
Also, exotic features like hyperbolic [49, 4:], epsilon-near-zero [4;] and zero-index [52] dispersion regions can be engineered by properly designing non-resonant metamaterials.
Our goal in this chapter is to develop a simulation tool, namely a 4 × 4 transfer matrix method (TMM), to study the electromagnetic wave propagation through layered structures
: containing any sort of birefrengent layers (with arbitrary direction of the optic axes) or isotropic layers. Then, we will modify this tool to include nonlinearity and calculate the second and third harmonic radiation traveling in both reflection and transmission directions.
The benchmarks used to check the accuracy of the TMM are:
• benchmark 3, Free mode calculation: Conservation of momentum and energy
gives rise to the boundary conditions for EM propagation interfaces. In benchmark 3,
the conservation of momentum is used to derive the generalized Fresnel equation [4]
and generalized Snell’s law [53]. All propagation modes were calculated and the shape
of the the indicatrix were assessed for accuracy. Next, Fresnel reflection and trans-
mission coefficients were calculated and reflectance and transmittance were checked
for conservation of energy. The reflectance and transmittance spectrum were also
compared against the reported values in the literature.
• benchmark 4, Bound mode calculation: While employing nonlinearity in our
TMM engine a significant milestone is to benchmark the results against the well-
known Maker’s fringe patterns [39, 3:]. The fringe patterns show the interference of
SH/TH reflected/transmitted from multiple interfaces.
These benchmarks will confirm the dependability of our TMM engine. We will use our
TMM engine in Chapter III to calculate effective nonlinear susceptibility. The algorithm used in this TMM engine is discussed in great details in Reference [54–57]. In the following sections the algorithm is revisited.
; 4.3 Notation convention
Before beginning the discussion on TMM, the notation conventions will be introduced.
In this section, the focus is put on the features (fonts, bold/unbold, capital/non-capital, hats, bars etc) of the symbols and their subscripts or superscripts, but the symbols of the specific parameters are not introduced. They were introduced later in the appropriate contexts. Also, Table 4.4, 4.5, 4.6, 4.7 and 4.8 enlists all the symbols used in the TMM
engine and can be used for quick reference. The symbols are classified as constants, scalars,
vectors, unit vectors (directions), tensors, matrices, angles and planes.
4.3.3 Subscripts and superscripts
Table 4.3 discusses the symbols and meanings of sub/superscripts of vectors and scalars
related to those vectors. The same subscripts and superscripts are also used to describe
tensors and matrices later. If a superscript is omitted in any equation then that equation
holds for all values of that superscript. The subscripts follow Einstein summation convention
in which repeated summation implied over terms that carry the same subscripts index in a
formula.
Table 4.3: List of subscripts/superscripts.
Symbol Meaning Set of choice 2: Rectified or dc ω: Fundamental frequency (FF), Ω Harmonic number 4ω: Second harmonic (SH), 5ω: Third harmonic (TH). 2 or vacuum, Layer number l 3, 4, 5, ... or, material’s name, or interface number l : l + n : lth to (l + n)th interface 3: Linear, E-mode: free, 4: Three wave mixing, E-mode: bound, h Number of wave mixed 5: Four wave mixing, E-mode: bound, 5+: Not studied.
32 Table 4.3 Continued from previous page Symbol Meaning Set of choice 3 to 6: if h = 1, K-modes k 3 to 32: if h = 2, see Section 4.6.3 3 to 42: if h = 3, c: Crystal axes system, a Axes system l: Laboratory axes system. 1 to 3: components along x,ˆ y,ˆ zˆ Components or u,ˆ v,ˆ wˆ respectively, c or, degree of freedom k: component in the xˆyˆ plane or plane of interface.
4.3.4 Vectors and scalars
Symbols used for vectors are uppercase letters with bold fonts (e.g. E). Unit vectors
(or, directions) are symbolized by the same letter in lowercase and bold fonts with a hat (
ˆ) on top (e.g. eˆ). Symbols used for scalar quantity are not in bold font (e.g. E). If the
scalar represents the magnitude of a field vector, the same uppercase letter is used. The
convention used for subscripts and superscripts for an example field vector, E and its scalar components E, is as follows:
(h),Ω,l E(h),Ω,l Ek,ac ∈ k,a , (4.3) v u 3 (h),Ω,l uX (h),Ω,l2 |Ek | = t Ek,ac , for a = l or c, (4.4) c=1 (h),Ω,l E(h),Ω,l (h),Ω,l eˆk,a = k,a /|Ek |, (4.5)
(h),Ω,l (h),Ω,l eˆk,ac ∈ eˆk,a . (4.6)
The angles (e.g. θ) have the sub/superscript as mentioned in Equation 4.3 but without the
(h),Ω,l component number (i.e. θk,a ). The angular frequency, ω, has no sub/superscript.
33 4.3.5 Tensors
Symbols used for tensors are not in bold font. They have a bar (¯) on top. The elements of a C1 × C2 × ... × Cn-dimensional tensor (e.g. ¯) is as follows:
(h),Ω,l (h),Ω,l ac1c2...cn ∈ ¯a , (4.7) where h, Ω, l are discussed in Table 4.3. Also, cn = 1, 2, 3, ..., Cn and Cn is the total degree of freedom in the nth dimension. The tensor notations can be contracted based on symmetry relations. For this study, all tensors are contracted into a 4-D matrix format. Therefore,
(h),Ω,l (h),Ω,l ac1c2 ∈ ¯a , (4.8) where, 1 to 3, if h = 1, c1 = 1 to 3, and c2 = 1 to 6, if h = 2, (4.9) 1 to 10, if h = 3.
The values of c2 depends on the number of waves mixed (i.e. h) which in turn gives the number of K-modes available. Please find more details in Section 4.6.3. The tensors are listed in Table 4.7 in the context of Maxwell’s equation and the constitutive relations.
4.3.6 Matrices
The matrices are represented by uppercase letters in double-struck font with a bar above
(e.g. M¯ ). All matrices are 4-dimensional. Thus,
M(h),Ω,l M¯ (h),Ω,l c1c2 ∈ C1×C2 , (4.:) where, c1 = 1, 2, ..., C1, c2 = 1, 2, ..., C2, and C1,C2 are the total degree of freedom. Also, h,
Ω, l represents the sub/superscript as discussed in Table 4.3. Abstract vectors (e.g. an array of values which does not represent any spatial components) are considered as 3-D matrices
(C2 = 1). The matrices are listed in Table 4.8 in the context of boundary conditions.
34 4.4 Material birefringence and layer geometry
Figure 4.3 represents the geometry of the layered structures studied for this research.
The thicknesses of layers and the special directions, planes and angles introduced by the anisotropy of materials are enlisted in Table 4.4.
Table 4.4: List of directions (unit vectors), angles, distances and planes related to the geometry of the layers and material anisotropy (see Figure 4.5 for reference).
Symbol Meaning Laboratory axes: xˆ, yˆ, zˆ Stratified media has C∞v symmetry. xˆyˆ is the plane of symmetry and zˆ is the surface/interface normal. Principle axes: The principle axes used for this study are determined based on the electric permittivity tensor. The principle axes of the crystal uˆ, vˆ, wˆ are different from its crystallographic axis. Also, the principle axes based on magnetic permeability and conductivity can be different in a generalized scenario. Such generalization is out of scope for this study. uˆl, vˆl, wˆl are the principle axes of the lth layer. Biaxial: oˆ1 and oˆ2 are the two optic axes lying in the uˆwˆ plane. oˆ is the direction that bisects the smaller angle between oˆ1 and oˆ2. oˆ k wˆ. Uniaxial: oˆ is the optic axis and oˆ k wˆ. oˆ Isotropic: Degenerate k-modes (i.e. oˆ ) exist in all directions. oˆl represents the bisection axis (biaxial) or optic axis (uniaxial) lth of the layer. Euler angles: Φ, Θ, Ψ These angles transform laboratory axes into principle axes. Φl, Θl, Ψl are the Euler angles of the lth layer. Rotation axis (compound): rˆ The axis of rotation of second Euler angle, Θ (see Figure 4.5) is the rotation axis. rˆl is the rotation axis of the lth layer. Plane of incidence: Generally, if multiple input beams are mixed, plane of reflection/ zˆkˆ transmission may not be parallel to the multiple planes of incidence. For this study, when φ = 0 (see Figure 4.5 and Table 4.6), zˆkˆ k zˆxˆ. Propagation distance or layer thickness: a The layer thickness (unit: m) is measured along zˆ (see Figure 4.5). al is the thickness of lth layer.
35 Figure 4.5: The transformation between the laboratory axes system and the principle axes system of the crystal. The symbols of Euler angles (Φ, Θ, Ψ) are uppercase and color coded to green, red and green respectively. Symbols for angle of incidence and azimuthal angle (θ, φ) are lowercase.
xˆyˆ plane is the plane of interface and zˆ is the surface normal. θ is the angle of incidence between zˆ and kˆ. φ is the azimuthal angle between xˆzˆ and kˆzˆ planes. When φ = 0,
xˆzˆ k kˆzˆ. The transformation from laboratory axes system (xˆ − yˆ− zˆ) to the principle axes system of the crystal (uˆ − vˆ − wˆ) uses Euler angles (Ψ, Θ, Φ). Both axes systems are color coded to red-blue-green, respectively. The primed (0) and double-primed (00) xˆ − yˆ − zˆ are the intermediary versions of the axes system after each rotation operation. The order of rotation operation is: R¯ z(Φ) (around zˆ) → R¯ x(Θ) (around rˆ) → R¯ z(Ψ) (around wˆ). The xˆ −yˆ plane (plane of interface) is colored white and the uˆ −vˆ plane of the crystal is colored gray for comparison. The broken yellow line marking the intersection of xˆ − yˆ and uˆ − vˆ planes is the direction of the rotation axis, rˆ and xˆ0 k rˆ.
36 4.5 Maxwell’s equations
In this section, the preliminaries of Maxwell’s equations are concisely discussed. The constants, scalars, vectors and tensors that go into the Maxwell’s equation will be introduced in the following few tables. First, the constants used in the TMM engine are given in Table
4.5. Table 4.5: List of constants.
Symbol Value Meaning Unit (S.I.) c 2.997924 × 108 Speed of light ms−1 0 8.854 × 10−12 Vacuum permeability C2s2kg−1m−3 µ0 4π × 10−7 Vacuum permittivity kgmC−2 √ j −1 imaginary unit ∼
Next, the field vectors and the corresponding scalars and angles are listed in Table 4.6.
Table 4.6: List of vectors, directions, magnitude and angles related to EM fields.
Vectors, unit vectors and magnitudes related to EM fields Symbol Meaning Unit (S.I.) (Vector/Unit Vector/Magnitude) E, eˆ, |E| Electric field V m−1 D, dˆ, |D| Electric displacement Cm−2 H, hˆ, |H| Magnetic field Am−1 B, ˆb, |B| Magnetic displacement V s2m−2 P, pˆ, |P | Polarization Cm−2 J, ˆj, |J| Current density Am−2 K, kˆ, |K| Wavevector rad m−1 ˆ 2π Wavelength Λ, λ, |Λ| = |K| m S, sˆ, |S| Poynting vector W m−2 Angles related to EM fields Angle of K-modes: θ This angle is between interface degree normal, zˆ and wavevector, kˆ Azimuthal angle: This angle is between xˆzˆ plane φ degree and plane of incidence, zˆkˆ. φ = 0, unless otherwise stated. Walk-off angle: w This angle is between sˆ and kˆ, degree or, eˆ and dˆ.
37 Table 4.6 Continued from previous page Symbol Meaning unit (S.I.) Angular frequency: - Rectified/DC: 2, - Fundamental frequency: ω, Ω - Second Harmonic: 2ω, rad s−1 - Third harmonic: 3ω, . . - nth harmonic: nω. Charge density: ρ ρl is the charge density Cm−2 of the lth layer.
Figure 4.6: Vector diagram of EM fields (not to scale and color coded for easy correspon- dence).
The EM radiation is assumed to be a plane wave with wavevector K. The materials of all
the layers are nonmagnetic (¯µ = ¯I3×3) and have no free currents and charges (J = 0, ρ = 0).
For any angular frequency (Ω), the Maxwell’s equations can be written as follows:
K × E = −ΩB, (4.;)
K × H = ΩD, (4.32)
K · B = 0, (4.33)
K · D = 0. (4.34)
Also, the time averaged Poynting vector is,
1 hSi = Re{E × H∗}. (4.35) 2
38 Figure 4.6 describes the relative orientation of field vectors. E, D, K and S lie on the same
plane and B, H are perpendicular to this plane. Also, E and S are perpendicular to each
other and K and S are perpendicular to each other. The angle between E and D, or, S
and K is known as walk-off angle.
w = cos−1(eˆ.dˆ) = cos−1(sˆ.kˆ), and − π/2 ≤ w ≤ π/2. (4.36) K θ = cos−1 l3 , and 0 ≤ θ ≤ π. (4.37) |Kl|
The material tensors will be listed in Table 4.7. These tensors will be used in the constitutive relations next.
Table 4.7: List of material tensors.
Symbol Meaning Refractive index [3 × 3]: - In crystal axes (uˆ, vˆ, wˆ), (1) n¯ nu 0 0 (1) n¯c = 0 nv 0 , where nu, nv, nw ∈ {all complex number}. 0 0 nw Electric permittivity (relative) [3 × 3]: In any axes system, ¯(1) = 1 +χ ¯(1).
- In crystal axes (uˆ, vˆ, wˆ), (1) c11 0 0 (1) (1) (1) (1) 2 (1) ∆12 = Re{c11} − Re{c22}, ¯c = (¯nc ) = 0 c22 0 , and, (1) (1) . (1) ∆23 = Re{c22} − Re{c33}, 0 0 c33 Biaxial (1) (1) (1) : c11, c22, c33 are sorted such that (1) (1) (1) (1) (1) ¯ , χ¯ Re{c11} < Re{c22} < Re{c33}. Positive: ∆12 > ∆23, Negative: ∆12 < ∆23. Uniaxial (1) (1) (1) : c11, c22, c33 are sorted such that (1) (1) Re{c11} − Re{c22} = ∆12 = 0. (1) (1) Positive: ∆23 > 0 and Re{c22} ≥ 0 , or, ∆23 < 0 and Re{c33} < 0 , (1) (1) Negative: ∆23 > 0 and Re{c22} < 0 , or, ∆23 < 0 and Re{c33} ≥ 0 . Isotropic (1) (1) (1) : Re{c11} = Re{c22} = Re{c33}.
- In laboratory axes (xˆ, yˆ, zˆ), (1) A¯ (1) A¯ −1 A¯ ¯l = 3×3¯c ( 3×3) . (see Table 4.8 for )
39 Table 4.7 Continued from previous page Symbol Meaning Electric susceptibility (second order) [3, 4]: [unit: m/V ] [Expanded: 3 × 3 × 3, Contracted: 3 × 6] Contracted 3 × 6 version is used, unless otherwise specified. In any axes system, 1 optical rectification: d¯(2),0 = χ¯(2),0. 2 1 second harmonic generation: d¯(2),2ω = χ¯(2),2ω. 2 d¯(2), χ¯(2) - In crystal axes (uˆ, vˆ, wˆ), (2) (2) (2) (2) (2) (2) dc11 dc12 dc13 dc14 dc15 dc16 ¯(2) (2) (2) (2) (2) (2) (2) dc = dc21 dc22 dc23 dc24 dc25 dc26. (2) (2) (2) (2) (2) (2) dc31 dc32 dc33 dc34 dc35 dc36
- In laboratory axes (xˆ, yˆ, zˆ), ¯(2) A¯ ¯(2) A¯ −1 A¯ dl = 3×3dc ( 6×6) . (see Table 4.8 for ) Electric susceptibility (third order) [3, 58]: [unit: m2/V 2] [Expanded: 3 × 3 × 3 × 3, Contracted: 3 × 10] Contracted 3 × 10 version is used, unless otherwise specified. In any axes system , 1 optical Kerr effect: d¯(3),ω = χ¯(3),ω. 3 1 third harmonic generation: d¯(3),3ω = χ¯(3),3ω. 3 d¯(3), χ¯(3) - In crystal axes (uˆ, vˆ, wˆ), (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) dc11 dc12 dc13 dc14 dc15 dc16 dc17 dc18 dc19 dc1,10 ¯(3) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) dc = dc21 dc22 dc23 dc24 dc25 dc26 dc27 dc28 dc29 dc2,10. (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) dc31 dc32 dc33 dc34 dc35 dc36 dc37 dc38 dc39 dc3,10
- In laboratory axes (xˆ, yˆ, zˆ), ¯(3) A¯ ¯(3) A¯ −1 A¯ dl = 3×3dc ( 10×10) . (see Table 4.8 for ) Magnetic permeability (relative) [3 × 3]: Non-magnetic media is assumed. µ¯ In both crystal axes (uˆ, vˆ, wˆ) and laboratory ¯ ¯ axes (xˆ, yˆ, zˆ), µ¯c =µ ¯l = I3×3. (see Table 4.8 for I)
The constitutive relationships in laboratory axes system (xˆyˆzˆ) can be written as follows:
0 Dl = El + Pl, (4.38)
0 Bl = µ µ¯lHl. (4.39)
3: Here, Linear Nonlinear z }| { z }| { n (1),Ω X (h),Ω Pl = Pl + Pl , (4.3:) h=2,3 The values of Ω and h, along with their interpretation, are given in Table 4.3. Section 4.6.3 and 4.7.3 constitute the detailed discussion on the number of free and bound P modes. We assume no depletion at fundamental frequency (Ω = ω). Thus, only free E-modes (h = 1) exist and optical Kerr effects and other higher order mixing (i.e. h > 1) at Ω = ω is ignored.
In laboratory axis, at fundamental frequency (Ω = ω),
4 4 (1),ω X (1),ω (1),ω 0 X (1),ω (1),ω (1),ω Pl = Pk,l exp jKk,l3 z = χ¯l Ek,l exp jKk,l3 z , (4.3;) k=1 k=1
(1),ω where, χ¯l is given in Table 4.7.
In laboratory axis, at second harmonic frequency: (Ω = 2ω), - For free wave (h = 1):
4 4 (1),2ω X (1),2ω (1),2ω 0 X (1),2ω (1),2ω (1),2ω Pl = Pk,l exp jKk,l3 z = χ¯l Ek,l exp jKk,l3 z , (4.42) k=1 k=1
(1),2ω where, χ¯l is given in Table 4.7. - For bound wave (h = 2):
10 (2),2ω X (2),2ω (2),2ω Pl = Pk,l exp jKk,l3 z k=1 (1),ω (1),ω m · Ep,l1 Eq,l1 (1),ω (1),ω m · E E p,l2 q,l2 4 4 (1),ω (1),ω 0 X X (2),2ω (1),ω (1),ω m · Ep,l Eq,l = 2 d¯ exp j K + K z 3 3 , l p,l3 q,l3 (1),ω (1),ω (1),ω (1),ω p=1 q=p m · (Ep,l2 Eq,l3 + Eq,l2 Ep,l3 ) m · (E(1),ωE(1),ω + E(1),ωE(1),ω) p,l1 q,l3 q,l1 p,l3 (1),ω (1),ω (1),ω (1),ω m · (Ep,l2 Eq,l1 + Eq,l2 Ep,l1 ) (4.43)
(1, if p = q, where, d¯(2),2ω is given in Table 4.7 and m = l 2, if p 6= q.
3; In laboratory axis, at third harmonic frequency: (Ω = 3ω), - For free wave (h = 1):
4 4 (1),3ω X (1),3ω (1),3ω 0 X (1),3ω (1),3ω (1),3ω Pl = Pk,l exp jKk,l3 z = χ¯l Ek,l exp jKk,l3 z , (4.44) k=1 k=1
(1),3ω where, χ¯l is given in Table 4.7. - For bound wave (h = 3):
20 (3),3ω X (3),3ω (3),3ω Pl = Pk,l exp jKk,l3 z k=1 4 4 4 0 X X X ¯(3),3ω (1),ω (1),ω (1),ω = 3 dl exp j Kp,l3 + Kq,l3 + Kr,l3 z × p=1 q=p r=q (1),ω (1),ω (1),ω m · Ep,l1 Eq,l1 Er,l1 (1),ω (1),ω (1),ω m · E E E p,l2 q,l2 r,l2 (1),ω (1),ω (1),ω m · Ep,l3 Eq,l3 Er,l3 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l2 Eq,l3 Er,l3 + Ep,l3 Eq,l2 Er,l3 + ... m · (1),ω (1),ω (1),ω ...Ep,l3 Eq,l3 Er,l2 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l3 Eq,l2 Er,l2 + Ep,l2 Eq,l3 Er,l2 + ... m · (1),ω (1),ω (1),ω ...Ep,l2 Eq,l2 Er,l3 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l3 Eq,l3 Er,l1 + Ep,l3 Eq,l1 Er,l3 + ... m · (1),ω (1),ω (1),ω ...Ep,l1 Eq,l3 Er,l3 , (4.45) (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l1 Eq,l1 Er,l3 + Ep,l1 Eq,l3 Er,l1 + ... m · (1),ω (1),ω (1),ω ...Ep,l3 Eq,l1 Er,l1 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l1 Eq,l2 Er,l2 + Ep,l2 Eq,l1 Er,l2 + ... m · (1),ω (1),ω (1),ω ...Ep,l2 Eq,l2 Er,l1 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l1 Eq,l1 Er,l2 + Ep,l1 Eq,l2 Er,l1 + ... m · (1),ω (1),ω (1),ω ...Ep,l1 Eq,l1 Er,l2 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω Ep,l1 Eq,l2 Er,l3 + Ep,l1 Eq,l3 Er,l2 + ... (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω m · ...E E E + E E E + ... p,l2 q,l1 r,l3 p,l2 q,l3 r,l1 (1),ω (1),ω (1),ω (1),ω (1),ω (1),ω ...Ep,l3 Eq,l2 Er,l1 + Ep,l3 Eq,l1 Er,l2 1, if p = q = r, ¯(3),3ω where, dl is given in Table 4.7 and m = 3, if p = q, or, q = r, or, r = p, 6, if p 6= q 6= r.
42 4.6 Free mode calculation
In this section, we introduce the wave equation in matrix format. The wave equation can be derived from Equation 4.;, 4.32, 4.38 and 4.39 as:
2 0 0 K × K × E − Ω µ µˆl E = P , (4.46)
As discussed in Table 4.3, the superscript h represents the number of waves mixed. h = 1 represents the linear modes of propagation for any frequency Ω. All the K-modes for h = 1 are known as free K-modes. Each free K-mode corresponds to a free E-modes. Since we are assuming no depletion of power of the fundamental frequency (ω) compared to the power transferred into the mixed modes, we are only interested in the mixing of waves with fundamental frequency. Mixing only fundamental frequencies give rise to higher harmonics and AC Kerr effect. Only second and third harmonic generation are of interest for this study.
For the free modes of any frequency (Ω = ω, 2ω, or, 3ω), Equation 4.46 can be written as,
(1),Ω (1),Ω (1),Ω 2 0 0 (1),Ω (1),Ω Kl × Kl × El − Ω µ El = Pl , (4.47)
(1),Ω (1),Ω (1),Ω 2 0 0 (1),Ω (1),Ω =⇒ Kl × Kl × El − Ω µ ¯l El = 0, (4.48) | {z } | {z } U(1),Ω V(1),Ω 3×3 3×3 where,
(1),Ω 2 (1),Ω 2 (1),Ω (1),Ω (1),Ω (1),Ω (Kl2 ) + (Kl3 ) −Kl1 Kl2 −Kl1 Kl3 (1),Ω (1),Ω (1),Ω (1),Ω (1),Ω (1),Ω (1),Ω U = −K K (K )2 + (K )2 −K K , (4.49) 3×3 l2 l1 l1 l3 l2 l3 (1),Ω (1),Ω (1),Ω (1),Ω (1),Ω 2 (1),Ω 2 −Kl3 Kl1 −Kl3 Kl2 (Kl1 ) + (Kl2 ) and, Ω2 V(1),Ω = Ω2µ00¯(1),Ω = ¯(1),Ω. (4.4:) 3×3 l c2 l
The wave matrix, W¯ = U¯ − V¯. Thus, from Equation 4.48, 4.49, 4.4:, the following can be
written,
43 W¯ (1),Ω (1),Ω 3×3 El = 0, (4.4;) W¯ (1),Ω (1),Ω or, 3×3 eˆl = 0. (4.52)
4.6.3 Free K-mode calculation
Equation 4.4; is the matrix format of wave propagation equation derived from Maxwell’s equations and constitutive relations. For Equation 4.52 to have a nontrivial solution, the determinant of W¯ should be zero. Thus,
W¯ (1),Ω det( 3×3 ) = 0. (4.53)
(a)
Figure 4.7: An example case of the normal surface of a positive uniaxial medium with an arbitrary direction of oˆ with respect to the laboratory axes system (xˆyˆzˆ). Four free (1),Ω (1),Ω K-modes (Kl ) along with their angles (θl ) with respect to zˆ is marked.
44 For any ith layer, Figure 4.7 shows the solution of Equation 4.53 graphically for a certain
(1),Ω (1),Ω value of Kl|| . The unknowns are the Kl3 . A positive uniaxial medium is chosen as an example case. The four K-modes of propagation is marked. To increase the readability of the solution process, following symbols are chosen for the time being.
(1),Ω α := Kl1 , (4.54)
(1),Ω β := Kl2 , (4.55)
(1),Ω γ := Kl3 , and (4.56)
V(1),Ω Vpq ∈ 3×3 , (4.57)
where p, q are the row and column positions of V¯. To solve Equation 4.53, the matrix W¯ is
partitioned in to an unknown part (named, B¯) containing γ and a known part (named, B¯0)
containing the rest. Thus, −γ2 0 −αγ B¯ 2 = 0 −γ −βγ , and, (4.58) −αγ −βγ 0
0 0 0 2 B11 B12 B13 β − V11 −αβ − V12 −V13 B¯0 0 0 0 2 = B21 B22 B23 = −αβ − V21 α − V22 −V23 . (4.59) 0 0 0 2 2 B31 B32 B33 −V31 −V23 α + β − V33 Therefore, from Equation 4.53, the following can be written,
det B¯ + B¯0 = 0, (4.5:)
4 3 2 =⇒ c4γ + c3γ +c2γ + c1γ + c0 = 0. (4.5;)
Here,
0 2 2 c4 =B33 + α + β , (4.62)