Study of low refractive index homogeneous thin film for application on metamaterial

Fig 4.23: Transmittance spectrum of PVA, PVP solution with and without existence of silver nanoparticles. The fabricated thin films by drop-coating method are light brown and [r]

VIETNAM NATIONAL UNIVERSITY OF HANOI

VIETNAM JAPAN UNIVERSITY

PHAM DINH DAT

STUDY OF LOW REFRACTIVE INDEX HOMOGENEOUS THIN FILM FOR APPLICATION ON METAMATERIAL

MASTER’S THESIS

VIETNAM NATIONAL UNIVERSITY OF HANOI

VIETNAM JAPAN UNIVERSITY

PHAM DINH DAT

STUDY OF LOW REFRACTIVE INDEX HOMOGENEOUS THIN FILM FOR APPLICATION ON METAMATERIAL

MAJOR: NANOTECHNOLOGY CODE: PILOT

RESEARCH SUPERVISOR: Ph.D PHAM TIEN THANH

i

Acknowledgement

First and foremost, I want to express my appreciation to my supervisor, Pham Tien Thanh Ph.D for his patient guidance and encouragement during my study and research at Vietnam Japan University

I would like to thank Prof Kajikawa Kotaro and his students at Kajikawa Lab, Faculty of Electrical and Electronics Engineering, Tokyo Institute of Technology who helped us facilities to perform calculation, experiments and measurements

I also would like to send my sincere thanks to the lecturers of Nanotechnology Program, Vietnam Japan University, who have taught and interested me over the past two years

Besides, I am grateful to my family and my friends who are always there to share their experiences that help me overcome the obstacles of student’s life

Hanoi, 17 June, 2019 Author

ii

TABLE OF CONTENTS

Acknowledgement i

LIST OF FIGURES, SCHEMES iv

LIST OF ABBREVIATIONS vi

CHAPTER 1: INTRODUCTION

1.1 Metamaterial

1.2 Optical material relate to refractive index

CHAPTER 2: FUNDAMENTAL THEORY

2.1 Effective Medium Theory

2.1.1 Effective medium

2.1.2 Permittivity calculation

2.2 Transfer Matrix for multilayer optics 10

2.3 Finite Difference Time Domain (FDTD) 14

CHAPTER 3: EXPERIMENTS 19

3.1 Silver nanoparticles synthesis 19

3.1.1 Chemicals 19

3.1.2 Process 19

3.2 Thin films fabrication 20

3.2.1 Chemicals 20

3.2.2 Process 20

3.3 Optical properties determination 21

3.4 Thin films thickness determination 21

CHAPTER 4: RESULTS AND DISCUSSION 22

iii

4.1.1 Index of refraction and index of extinction depend on element of particles

22

4.1.2 Index of refraction and index of extinction depend on volume fill fraction of silver nanoparticles on polymer matrix 25

4.1.3 Calculation for thin film following EMT using TMM 28

4.1.4 Calculation for thin film using FDTD method 31

4.1.5 Neighbor particles interaction 34

4.2 Experiment results 37

4.2.1 Properties of silver nanoparticles 37

4.2.2 Properties of thin films 40

iv

LIST OF FIGURES, SCHEMES

Fig 1.1: Multilayer structure and nanowires embedded structure metamaterial (A: metal-dielectric layered, B: wires in dielectric host) Fig 2.1: A material model of UEM Fig 2.2:Three simple model of UEM material classified following topology _ Fig 2.3: A simple model for assumption limitation of volume fill fraction _ Fig 2.4: Considered system of TMM problem 11 Fig 2.5: The arrangement of electric- and magnetic-field nodes in space and time 17 Fig 4.1: The index of refraction of PVP including 3% volume fill fraction of silver, gold and copper 22 Fig 4.2: The index of extinction of PVP including 3% volume fill fraction of silver, gold and copper 23 Fig 4.3: The index of refraction of PVA including 3% volume fill fraction of silver, gold and copper 24 Fig 4.4: The index of extinction of PVA including 3% volume fill fraction of silver, gold and copper 24 Fig 4.5: The index of refraction of PVP including 2%, 3%, 4% and 5% volume fill fraction of silver _ 25 Fig 4.6: The index of refraction of PVA including 2%, 3%, 4% and 5% volume fill fraction of silver _ 26 Fig 4.7: The index of extinction of silver and PVP including 2%, 3%, 4% and 5% volume fill fraction of silver 27 Fig 4.8: The index of extinction of silver and PVA including 2%, 3%, 4% and 5% volume fill fraction of silver 27 Fig 4.9: Transmittance spectrum of 30 nm PVP-based films corresponding to

different Ag fill fraction _ 28 Fig 4.10: Transmittance spectrum of 30 nm PVA-based films corresponding to different Ag fill fraction _ 29 Fig 4.11: The calculated transmittance spectrum of 200 nm PVP-based films

v

Fig 4.12: The calculated transmittance spectrum of 200 nm PVA-based films

corresponding to different Ag fill fraction using TMM _ 31 Fig 4.13: The FDTD domain for calculation of 200nm film by x, y, z direction and 3D visions 32 Fig 4.14: The calculated transmittance spectrum of 200 nm PVP-based films

corresponding to different Ag fill fraction using FDTD method 33 Fig 4.15: The calculated transmittance spectrum of 200 nm PVA-based films

corresponding different Ag fill fraction using FDTD method 33 Fig 4.16: The simple model for consider neighbor-particles interaction 35 Fig 4.17: Calculated extinction spectra of two neighbor-particles with distance equal 3nm in medium that has refractive index equal 1.5 using FDTD 36 Fig 4.18: Calculated extinction spectra of neighbor-particles with distance equal 3nm in medium that has refractive index equal 1.5 using DDA _ 37 Fig 4.19: The images of silver nanoparticles solution after synthesis(a), after

centrifugation(b) and after re-disperse on water(c) _ 38 Fig 4.20: SEM image of self-synthesis silver nanoparticles 39 Fig 4.21: Transmittance spectrum of self-synthesis and commercial silver

vi

LIST OF ABBREVIATIONS

DDA: Discrete Dipole Approximation EMT: Effective Medium Theory EM: Effective Medium

E-field: Electric field

LSPR: Localized Surface Plasmon Resonance MGG: Maxwell Garnet geometry

MGT: Maxwell Garnett theory

FDTD: Finite Different Time Domain H-field: Magnetic field

PVP: Poly Vinyl Pyrrolydone PVA: Poly Vinyl Alcohol PML: Perfect Match Layer

1

CHAPTER 1: INTRODUCTION 1.1 Metamaterial

Electromagnetic metamaterial is a class of material using for engineering electromagnetic space and controlling light propagation Metamaterials have shown their promise for the next generation optical materials with electromagnetic behaviors almost can’t be obtained in any conventional materials They have a plenty of application including cloaking [11,15,26], imagining [12,29,41], sensing [18,23,36], wave guiding [13,22,38], absorber [5], etc

The metamaterial is fabricated based on the composite structures including inclusions that have sub-wavelength structures The inclusions have designed structure They can be totally artifact or emulate based on nature structure The inclusions are arranged on a host medium that is normally dielectric Due to the small size and distance of inclusion, the metamaterials can be considered as the homogeneous mediums The properties of material are represented through permittivity and permeability By changing shape and size of inclusion, permittivity and permeability of metamaterial can be adjusted to very high or low (even negative) value Under the consideration for permittivity and permeability, the material can be classified into groups [31] They are epsilon-negative material (ENG), mu-negative material (MNG), double positive material (DPS) and double negative material (DNG) The metamaterial is in class of ENG, MNG and DNG materials Besides that, the metamaterial includes band gap material but it will not be considered in this research

2

Fig 1.1: Multilayer structure and nanowires embedded structure metamaterial (A: metal-dielectric layered, B: wires in dielectric host)

The metamaterials structuring as in Fig are called as hyperbolic metamaterial In this class of metamaterial, the refractive indexes and arrangement of components play a significant role to properties of metamaterial The below equations is used to calculate the anisotropic dielectric function of layered metamaterial

ϵ𝑝𝑎𝑟𝑎𝑙𝑙𝑒𝑙 =𝑑𝑑+ 𝑑𝑚 𝑑𝑚 𝜖𝑚 +

𝑑𝑑 𝜖𝑑 ϵ𝑝𝑒𝑟𝑝𝑒𝑛𝑑𝑖𝑐𝑢𝑙𝑎𝑟 =𝑑𝑚𝜖𝑚+ 𝑑𝑑𝜖𝑑

𝑑𝑑 + 𝑑𝑚

withϵ𝑝𝑎𝑟𝑎𝑙𝑙𝑒𝑙 and ϵ𝑝𝑒𝑟𝑝𝑒𝑛𝑑𝑖𝑐𝑢𝑙𝑎𝑟are dielectric function following directions those are parallel and perpendicular with surface of multilayer structure; dd and dm

are thickness; 𝜖𝑚 and 𝜖𝑑are dielectric function of dielectric material and metal Following it, the very low refractive index n = √𝜖 can be achieved by this way [38] The problem is that the fabrication is very complex and expensive The distance between wires, the thickness of each layer must be very precise

3

metamaterial focus on optimizing structure So, it is lacking in the studies which develop the constituent material of metamaterial The second is the difficulty in fabrication that mentioned above As an impact of the second, the limitation of working wave length also is an issue The most common topic about metamaterial relates to terahertz region that corresponds to long wavelengths where demand inclusion in micrometer level We need more research about metamaterial that works in shorter wavelength region So, it is necessary to study a material which is easy to fabricate and can be applied to metamaterial working in visible wavelength

1.2 Optical material relate to refractive index

The refractive index is very important parameter describing optical material properties It relate to all optical phenomena such as refraction, reflection, transmission By changing the refractive index of material, we can create new materials that can be to various fields There has been many researches related to high refractive index material and negative refractive index material The high refractive index materials are very useful for application of solar cell due to anti-reflection property of them [1,6,7] The negative index material is new class of material that is promising for many applications [11-13] However, it has a lack of research for low refractive index material They play a significant role in application relate to the reflection materials and metamaterials It has some types of low refractive index including metal nano-rod or metamaterial used nano-wires as inclusion [11,12] They are hard to fabricate and only work in IR wavelength region I want to make a material that is easy to fabricate and work in visible region It is possible based on the effective medium theory

4

detail later The point is that both of these topology demand simpler than the layered structure It suggests a composite material that can achieves properties as like as layered metamaterial but easier to fabricate This material can be based on a polymer host material with metal nanoparticles as inclusion It can be used for thin films, metamaterial application

In this study, my purpose is making a type of nano-composite material that has low index of refraction and low index of extinction Based on the idea of hyperbolic metamaterial, it is able to create the low refractive index and low loss medium by the combination of low refractive index but loss material as metals and low loss but high refractive index as polymers I fabricated the nanocomposite based on nano silver particles embedded on polymers This type of material was considered in about absorption [49], high refractive index region [33], etc In this study, I used calculation to orient and predict about object material and experiment to verify my prediction

The research contents include:

- Calculation refractive index of PVP-based and PVA-based material with Ag nanoparticles as inclusion

- Calculation transmittance of thin films based on calculated materials - Fabricate the thin films using object materials and compare with

5

CHAPTER 2: FUNDAMENTAL THEORY 2.1 Effective Medium Theory

2.1.1 Effective medium

Consider a type of material that is presented in Fig 2.1, it has some length scales which are presented (a and b), are well-defined and all much less than the wavelength of light This condition means that the scattering cause by the inhomogeneity resulting from the composite natural can be negligible In this case, the real composite material, with host dielectric constant (𝜖ℎ) and inclusion dielectric constant (𝜖𝑖), can be replaced by a Uniform Effective Medium (UEM) with a dielectric constant (𝜖𝑒𝑓𝑓) [41]

Fig 2.1: A material model of UEM

Fig 2.2 shows three simple models of this type of material that are classified based on their topology The first that is called the Maxwell Garnet composite geometry, including well-defined spherical inclusions in host background [1] The next topology is disordered where the constituent materials are more or less than inclusion The last is the ordered, layered composite geometry [41]

a

6

Fig 2.2: Three simple model of UEM material classified following topology

The object of research is the material that following the Maxwell Garnet composite geometry for applying to metamaterial as the third type of geometry introduced above For predictable by Effective Medium Theory (EMT), the material should considered following some conditions At first, the scattering should be neglect able, at least with theoretical view It means that the size of metal particles must be much smaller than the working wavelength This study mostly consider characteristic of material on the visible wavelength region of light that around 300 – 800 nm So, the particles radius should be smaller than about 30nm (about tenth times compare with the shortest wavelength) In this study, the nanoparticle 20nm in diameter is chosen

7

Fig 2.3: A simple model for assumption limitation of volume fill fraction

In this model, the medium can be divided to cube cells which include a part of space that is occupied by one particle (8 pieces x 8) If we call that the mean distance between each particle and the nearest approximately is a, the volume fill fraction f of 20 nm diameter particles on polymer matrix should be limited depend on a The distance b should much less than wavelength of light As the size of particles condition, the distance a should be less than 30nm Hence,

f =𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠

𝑡𝑜𝑡𝑎𝑙 𝑣𝑜𝑙𝑢𝑚𝑒 =

𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑎 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑎 𝑐𝑒𝑙𝑙 =

4 3𝜋(

𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟 )

3

(𝑎+𝑑𝑖𝑎𝑚𝑒𝑡𝑒𝑟)3 ≥ 3𝜋(

20 2)

3

(30+20)3 ≈ 0.0335 For predictable by Maxwell Garnett Topology (MGT), the fill fraction of metal nanoparticles on material should be less than 0.0335 The size and fill fraction of nanoparticles which will be used for fabricate material could be roughly considered and limited following above There is an upper limitation of fill fraction related to index of extinction but we will consider it later

8 2.1.2 Permittivity calculation

The index of refraction and extinction of material following MGT can be predicted through calculation There are two approaches to deriving the calculation ways The first is to examine, at some level of approximation, the nature of mesoscopic fields in material and perform spatial averages over them to identify the values of the macroscopic fields [7,34] The second is based on the expression for internal energy of the material and comparing it with expression for an effective medium [8] For easy to understanding, the first way will be used to introduce the calculation method

At first, we can refer to the particles as “molecules” in a region which include amount of particles much more than one [29] So, we can consider “particles” as an atom which is characterized by polarizability (α) In a space consisting of atoms that are arranged in defined lattice, the atomic polarizability links to the dipole moment p by the local field that due to Maxwell electrical field (E) and dipole respond field that can be expressed by local field corrections Hence:

p = α(E + local field corrections) (2.1.1)

We assume that the integral of the microscopic electric field e over a sphere around a charge distribution with a dipole moment p is given in electrostatic limit [16,24] This condition is represented by the below equation This assumption means that the electromagnetic interaction between dipoles (particles or atom in this consideration) should be neglected The reason is that the averages of the fields due to dipoles come to zero in case of medium contain a large amount of dipole

∫ 𝑒𝑑𝑉 = −4𝜋

3 𝑝 (2.1.2)

With N atoms in a lattice that has volume V, the local field correction can be estimated as −−

4𝜋 3𝑝

9

p = α [E − −

4𝜋 3𝑝 𝑉 𝑁

] (2.1.3)

Then, the dipole moment per unit volume P is given following:

P =𝑁

𝑉P = 𝑁

𝑉α (E + 4𝜋

3 P) (2.1.4)

Meanwhile, there is a relation between dipole moment per unit 𝑃, electrical field E and dielectric constant 𝜖 following:

E + 4πP = ϵE (2.1.5)

So, we can derive an equation describing relation of dielectric constant 𝜖 and polarizability α, which is as known as Claudius – Mossotti relation:

ϵ−1 ϵ+2=

4π

N

Vα (2.1.6)

The above equation is derived in case the “atoms” are in vacuum In our case, we consider the inclusion as sphere not atom Under effective of electrical field, there is an internal electrical field inside the sphere that occurred by external field and depolarization field So, we have the dipole moment p𝑖 of the inclusion sphere within the host medium is:

p𝑖 = a3 ϵ𝑖−ϵℎ

ϵ𝑖+2ϵℎ𝐸0 (2.1.7)

Where a is the radius of sphere, ϵ𝑖 is the dielectric constant of inclusion, ϵℎ is the dielectric constant of host medium and 𝐸0 is the electrical field applied far from inclusion Thus, we can identify an effective polarizability as:

α = a3 ϵ𝑖−ϵℎ

ϵ𝑖+2ϵℎ (2.1.8)

10

ϵ𝑒𝑓𝑓−ϵℎ

ϵ𝑒𝑓𝑓+2ϵℎ=

4𝜋

𝑁 𝑉a

3 ϵ𝑖−ϵℎ

ϵ𝑖+2ϵℎ = 𝑓

ϵ𝑖−ϵℎ

ϵ𝑖+2ϵℎ (2.1.9)

The dielectric constant of effective medium ϵ𝑒𝑓𝑓 can be calculated from dielectric constant of constituents and fill fraction of inclusion f Here, we can see that the dielectric constant which is calculated following Maxwell Garnett equation is just depend on material of host, inclusion and fill fraction of inclusion However, the properties of real thin film depend on some other factors, i.e size of particles, distance and distribution of them, etc More detail consideration will be given in Chapter

2.2 Transfer Matrix for multilayer optics

11

Fig 2.4: Considered system of TMM problem

Our problem is illustrated in Fig 2.4 It is designed to simulate the real measurement of samples We consider the propagation of incident light from air medium (medium 1) through material (medium 2) and glass (medium 3) to air medium (medium 4) The complex dielectric constants of air, glass and material are 𝜖𝑎𝑖𝑟 = 1, 𝜖𝑔𝑙𝑎𝑠𝑠 = 1.5 and 𝜖𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙, respectively The dielectric constants of air and glass are almost unchanged following the wavelength Meanwhile, dielectric constant of material is considered as a function of wavelength due to dependence of dielectric constant of medium on dielectric constant of nanoparticles inclusion The used constants is taken from the available database [24] The thickness of air is assumed as infinity because it’s external medium The thickness of glass can be determined but it not very necessary because the neglected extinction on glass The thickness of material is important to calculate so it must be known for calculation

The consideration about propagation of light is processed by considering forward and backward propagating electric fields through mediums The E-field in medium E1 is represented by two-component vector:

Air

𝜖𝑎𝑖𝑟

Material

𝜖𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙

Glass

𝜖𝑔𝑙𝑎𝑠𝑠

Air

𝜖𝑎𝑖𝑟

L

IGHT

SOUR

C

E

𝐸1+ 𝐸2+ 𝐸3+ 𝐸4+

𝐸1− 𝐸2− 𝐸3− 𝐸4−

12

E1 = [𝐸1

+

𝐸1−] (2.2.1)

With 𝐸1+is forward-propagating E-field and 𝐸1− is backward-propagating E-field in medium By the same way, the E-E-field in medium is represented by:

E4 = [𝐸4

+

𝐸4−] (2.2.2)

Following matrix representation of for polarized light, the relation between E-field in medium and medium is showed by matrix multiplication:

T E1 = E4 (2.2.3)

With T is 2x2 matrix The matrix T is the overall transfer matrix that describe the effect of medium to incident light as an operator that change vector of E-field from incident medium to measuring medium Call that:

T = [𝑇11 𝑇12

𝑇21 𝑇22] (2.2.4)

Then, we have expression of multiplication:

[𝑇11 𝑇12 𝑇21 𝑇22] [

𝐸1+ 𝐸1−] = [

𝐸4+

𝐸4−] (2.2.5)

We can assume that 𝐸4− = because it has no incident light from medium Then, if divide both side to 𝐸1+we have:

[𝑇11 𝑇12 𝑇21 𝑇22] [

1

𝐸1−/𝐸1+] = [𝐸4

+/𝐸 1+

0 ] (2.2.6)

It’s easily to see that 𝐸1−/𝐸1+is the overall reflection amplitude r and 𝐸4+/ 𝐸1+is the overall transmission amplitude t So, the reflectance R and transmittance Tr

can be evaluated following:

R = r2 = (−𝑇21 𝑇22)

2

13

𝑇𝑟 = t2 = (𝑇11 + 𝑇12 𝑟)2 (2.2.8)

Now, the problem is finding the overall transfer matrix The propagation of light through mediums includes two ingredients The first is the propagation at the interface of medium i and medium j those have different refractive index For isotropic media, the transfer matrix for interface is a 2x2 matrix defined by:

Mij =

𝑡𝑖𝑗[

1 𝑟𝑖𝑗

𝑟𝑖𝑗 1] (2.2.9)

With 𝑡𝑖𝑗 and 𝑟𝑖𝑗 are the transmission and reflection amplitudes for light come from medium i to medium j Call that Ni and Nj is the reduced wave vector on

propagation direction and 𝜖𝑖 and 𝜖𝑗 are complex dielectric constant of medium i and medium j The reflection and transmission amplitudes for s-polarized light are following [9]:

𝑟𝑖𝑗 =𝑁𝑖−𝑁𝑗

𝑁𝑖+𝑁𝑗 and t𝑖𝑗 =

2𝑁𝑖

𝑁𝑖+𝑁𝑗 (2.2.10)

And the corresponding expressions for p-polarized light are:

𝑟𝑖𝑗 =𝜖𝑖𝑁𝑗−𝜖𝑗𝑁𝑖

𝜖𝑖𝑁𝑗+𝜖𝑗𝑁𝑖 and t𝑖𝑗 =

2(𝜖𝑖𝜖𝑗)1/2𝑁𝑖

𝜖𝑖𝑁𝑗+𝜖𝑗𝑁𝑖 (2.2.11)

The second ingredient of propagation of light on our problem is the propagation on each medium The propagation matrix Pi of medium i is expressed

following:

P𝑖 = [𝑝𝑖

0 𝑝̅𝑖] (2.2.12)

With 𝑝𝑖 = exp (𝑖𝑁𝑖𝑘0𝑑𝑖) and 𝑝𝑖 = 𝑝̅𝑖−1 (k0 is vacuum wave vector

magnitude and di is thickness of material) Now, we can evaluate overall transfer

matrix from two ingredients applying for propagation of light from medium to 3:

14

By this way, the transmittance Tr13 can be calculated with determined

complex dielectric constant and thickness of thin film Following Fressnel’s equation [14] the transmittance Trglass of light propagating from glass to To

compare with experiment results, the final calculated transmittance is evaluated by:

Tr =𝑇𝑟13 × 𝑇𝑟𝑔𝑙𝑎𝑠𝑠

𝑇𝑟𝑔𝑙𝑎𝑠𝑠 (2.2.14)

This result correspond to measured transmittance of thin film on glass substrate with reference is glass

2.3 Finite Difference Time Domain (FDTD)

The Finite-Difference Time-Domain (FDTD) method is the simplest full-wave techniques used to solve problems in electromagnetics The FDTD method can solve complicated problems, but it consumes a lot of computation resource Solutions may demand a large amount of memory and computation time The FDTD method loosely fits into the category of “resonance region” techniques, i.e., ones in which the characteristic dimensions of the domain of interest are somewhere on the order of a wavelength in size If an object is very small compared to a wavelength, quasi-static approximations generally provide more efficient solutions Alternatively, if the wavelength is exceedingly small compared to the physical features of interest, ray-based methods or other techniques may provide a much more efficient way to solve the problem [36]

The FDTD method is mainly based on the central-difference approximation This approximation can be applied to both the spatial and temporal derivative in Maxwell’s equation Now, we consider the Taylor series expansions of the function f(x) expanded about the point x0 with an offset of ±𝛼

2:

f (x0+𝛼

2) = f(x0) + 𝛼 2f

′(x 0) +

1 2!(

𝛼 2)

2

f′′(x

15

f (x0−𝛼

2) = f(x0) − 𝛼 2f

′(x 0) +

1 2!(

𝛼 2)

2

f′′(x0) − ⋯ (2.3.2)

where the primes indicate differentiation Subtracting the equation (2.3.1) to the equation (2.3.2), we have:

f (x0+𝛼

2) − f (x0− 𝛼

2) = 𝛼f ′(x

0) + ⋯ (2.3.3)

Dividing both side to 𝛼 :

f(x0+𝛼2)− f(x0−𝛼2)

𝛼 = f

′(x

0) + ⋯ (2.3.4)

Here, we see that with 𝛼 is very small, the parts including high derivative of f(x) are neglect able So, we have an approximation following:

df(x) dx x=x0

=f(x0+

𝛼

2)− f(x0− 𝛼 2)

𝛼 (2.3.5)

This is the central-difference approximation Since the lowest power of 𝛼

being ignored is the second order, the central difference is said to have second-order accuracy or second-order behavior This implies that if 𝛼 is reduced by a factor of 10, the error in the approximation should be reduced by a factor of 100 (at least approximately) In the limit as 𝛼 goes to zero, the approximation becomes exact

The FDTD algorithm as first proposed by Kane Yee in 1966 employs second-order central differences The algorithm can be summarized as follows [36]:

1 Replace all the derivatives in Ampere’s and Faraday’s laws with finite differences Discretize space and time so that the electric and magnetic fields are staggered in both space and time

2 Solve the resulting difference equations to obtain “update equations” that express the (unknown) future fields in terms of (known) past fields

16

4 Evaluate the electric fields one time-step into the future so they are now known (effectively they become past fields)

5 Repeat the previous two steps until the fields have been obtained over the desired duration

Here, let’s consider dimension problem of FDTD method We assumed that the E-field only has a z component and there are only variations in x direction Following Maxwell’s equation, we can derive two scalar equations corresponding to Faraday’s law and Ampere’s:

μ𝜕𝐻𝑦

𝜕𝑡 = 𝜕𝐸𝑧

𝜕𝑥 (2.3.6)

ϵ𝜕𝐻𝑦

𝜕𝑡 = 𝜕𝐸𝑧

𝜕𝑥 (2.3.7)

where μ and ϵ are permeability and permittivity of medium, respectively Then, we could replace the derivatives in (2.3.6) and (2.3.7) with finite differences To convenient, the below notation will be used to indicate the location in space and time that the fields are considered:

E𝑧(x, t) = E𝑧(m∆x, q∆t) = E𝑧𝑞[m] (2.3.8)

H𝑦(x, t) = H𝑦(m∆x, q∆t) = H𝑦𝑞[m] (2.3.9)

17

Fig 2.5: The arrangement of electric- and magnetic-field nodes in space and time

Now, let consider the space-time point ((m + 1/2)∆x, q∆t) by equation (2.3.6):

μ𝜕𝐻𝑦

𝜕𝑡 (𝑚 + 𝟏/𝟐)∆𝑥,𝑞∆𝑡 = 𝜕𝐸𝑧

𝜕𝑥 (𝑚 + 𝟏/𝟐)∆𝑥,𝑞∆𝑡 (2.3.10)

Using the central approximation, we can see that it possible to derive unknown value H𝑦𝑞+1/2[m + 1/2] from available values H𝑦𝑞−1/2[m + 1/2] ,

E𝑧𝑞[m] and E𝑧𝑞[m + 1] following:

μH𝑦

𝑞+1/2

[m+12]−H𝑦𝑞−1/2[m+1/2]

∆𝑡 =

E𝑧𝑞[m+1]−E𝑧𝑞[m]

∆𝑥 (2.3.11)

Solving this to take H𝑦𝑞+1/2[m + 1/2]:

H𝑦𝑞+1/2[m +1

2] = H𝑦 𝑞−1/2

[m +1

2] + ∆𝑡 𝜇∆𝑥E𝑧

𝑞[m + 1] − E 𝑧 𝑞[m]

(2.3.12)

This is known as an update equation, specifically the update equation for the Hy field And by the same way applying for equation (2.3.7), we can derive the

18

electric-field node in the grid, the dividing line between what is known and what are unknown moves forward another one-half temporal step They would be updated again, then the electric fields would be updated, and so on

It is often convenient to represent the update coefficients ∆t/ϵ∆xand ∆t/μ∆x

in terms of the ratio of how far energy can propagate in a single temporal step to the spatial step The maximum speed electromagnetic energy can travel is the speed of light in free space c = 1/√ϵ0𝜇0 and hence the maximum distance energy can travel

in one time step is c∆t (in all the remaining discussions the symbol c will be reserved for the speed of light in free space) The ratio c∆t/∆x is often called the Courant number which we label Sc It plays an important role in determining the stability of a simulation

The more detail consideration about 3D problem and the boundary condition is important to understand clearly about FDTD method but it’s not suitable to discus in here The deeper discussions are provide in many the other relation document [8,17,21,28,30,37]

19

CHAPTER 3: EXPERIMENTS 3.1 Silver nanoparticles synthesis

3.1.1 Chemicals

Silver nitrate: AgNO3 (Sigma Aldrich)

Poly Vinyl Pyrrolidone (PVP) powder

Sodium borohydride: NaBH4 (Sigma Aldrich)

Distilled water

3.1.2 Process

Step 1: Take 0.51 g Poly Vinyl Pyrrolidone and dissolve in 20 mg distilled water, stirring in 60 minutes (solution M2)

Step 2: Take 0.05 mg NaBH4 and dissolve in 50 ml distilled water, stirring in

20 minutes (Solution M3)

Step 3: Take 0.01 g AgNO3 and dissolve in 10 ml distilled water, stirring in 20 minutes (Solution M1)

Step 4: After all the solutes are completely dissolved, Add solution M1 to solution M2, string in 15 minutes (Mixture M)

Step 5: After that, drop slowly solution M3 to mixture M The drop flow is less than 1l/second

Step 6: After that, keep the string in hour or more

Step 7: Purification by centrifugation at 11000 rounds per minutes The centrifugation is repeated times with 20 minutes each time

20 3.2 Thin films fabrication

3.2.1 Chemicals

Poly Vinyl Pyrrolidone (PVP) powder Poly Vinyl Alcohol (PVA) powder

Silver nanoparticles solution (Sigma Aldrich) Distilled water

3.2.2 Process

Step 1: The solution used to film fabricate is prepared from 20nm diameter silver nanoparticles solution (Sigma Aldrich) with sodium citrate is used as stabilizer and water as solvent It has two types of solution:

 Solutions made by PVP and silver nanoparticles were prepared by adding nanoparticles solution to PVP powder The mass ratios of nanoparticles solution and PVP powder correspond to 3%, 4% and 5% fill fraction of silver nanoparticles on thin films

 Solutions made by PVA and silver nanoparticles were prepared by add nanoparticles solution into prepared PVA 10%w.t solution with water as solvent The mass ratios of nanoparticles solution and PVA correspond to 3%, 4% and 5% fill fraction of silver nanoparticles on thin films

Step 2: Before film making, the solutions are sonicated for 30 minutes using ultrasonicator bath The thin film was fabricated on glass substrate following methods:

 Drop coating: 10l prepared solution were dropped into side of glass substrates Then, samples were dried on vacuum at 60oC for more than

hours

21

rpm on 10 seconds Then, samples were dried on vacuum at 60oC for more

than hours

3.3 Optical properties determination

The optical property of thin films is determined by UV – VIS spectrophotometer The measurement investigates transmittance of thin films and solutions on wavelength region from 300nm to 800nm The glass substrates which have thin films are placed directly into measuring chamber The solutions are packaged in cuvettes The reference is glass substrate in case of thin films or distilled water in case of solution The scan speed is 40nm/minute

3.4 Thin films thickness determination

22

CHAPTER 4: RESULTS AND DISCUSSION 4.1 Calculation results

4.1.1 Index of refraction and index of extinction depend on element of particles

In this study, it has two host materials which are Poly Vinyl Pyrrolidone (PVP) and Poly Vinyl Alcohol (PVA) Their dielectric functions are considered as constants because they are stable over visible wavelength The considered inclusions are by copper, gold and main object – silver The complex dielectric constants of these elements are functions of wavelength The dielectric constants used for calculations are taken from the available database [30]

The calculated index of refraction and index of extinction of types of material based on PVP as host medium following UEM are described in Figure 4.1 and Figure 4.2, respectively The calculation is processed using Maxwell Garnett expression for effective dielectric constant ϵ𝑒𝑓𝑓 The host is PVP with index of refraction as 1.5523 and neglected index of extinction [24] The inclusions are gold, copper and silver with volume fill fraction 3%

Fig 4.1: The index of refraction of PVP including 3% volume fill fraction of silver, gold and copper

0.5 1.5 2.5

0.3 0.4 0.5 0.6 0.7 0.8

Inde

x

of r

efr

ac

tion

Wavelength (m)

23

Fig 4.2: The index of extinction of PVP including 3% volume fill fraction of silver, gold and copper

Following Fig 4.1, it’s easily to see that silver nanoparticles should be the most suitable element for low refractive index material It shows the index of refraction about on wavelength region from 390 to 410 nm Comparing with index of refraction of PVP (approximate 1.55), the existence of silver, gold or copper also can decrease index of refraction But, the purpose is fabricating a material which has index of refraction approximate refractive index of air So, silver nanoparticle is chosen as inclusion of material Just composite materials including polymer and silver nanoparticles are considered on later part of thesis

The other considered host medium is PVA We also calculated refractive index for material with PVA as host material and inclusion as like as in case of PVP (material and fill fraction) The index of refraction and index of extinction by wavelength are illustrated on Fig 4.3 and Fig 4.4 Actually, both index of refraction and index of extinction of PVA based materials are quite similar as case of PVP The index of refraction of PVA – silver material is about on wavelength region from 385 to 400 nm However, the confirmation that PVP based and PVA based

0 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8

0.3 0.4 0.5 0.6 0.7 0.8

Index

of ex

tinction

Wavelength (m)

24

material has similar indexes of refraction is necessary The experimental discussion about two types of material will be showed in more detail later

Fig 4.3: The index of refraction of PVA including 3% volume fill fraction of silver, gold and copper

Fig 4.4: The index of extinction of PVA including 3% volume fill fraction of silver, gold and copper

0.5 1.5 2.5

0.3 0.4 0.5 0.6 0.7 0.8

Inde

x

of r

efr

ac

tion

Wavelength (m)

PVA_Ag3% PVA_Au3% PVA_Cu3%

0 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8

0.3 0.4 0.5 0.6 0.7 0.8

Index

of ex

tinction

Wavelength (m)

25

For both PVP and PVA cases, the main challenge is shown in Fig 4.2 and Fig 4.4 Although material including silver could have low index of refraction, the index of extinction of it is much higher than the others However, the extinction relate to both index of extinction and thickness of thin film For application for metamaterial, the thickness of thin film should much less than working wavelength that about 40nm in case of silver based material In this case, the extinction is acceptable

4.1.2 Index of refraction and index of extinction depend on volume fill fraction of silver nanoparticles on polymer matrix

The index of refraction and index of extinction not only depend on type of material but also relate to fill fraction of inclusion on material The wavelength depend index of refraction for PVP-based and PVA-based material in number of fill fraction are shown in Fig 4.5 and Fig 4.6

Fig 4.5: The index of refraction of PVP including 2%, 3%, 4% and 5% volume fill fraction of silver

0.3 0.8 1.3 1.8 2.3 2.8 3.3

0.35 0.4 0.45 0.5 0.55

Index

of r

efraction

Wavelength (m)

26

Fig 4.6: The index of refraction of PVA including 2%, 3%, 4% and 5% volume fill fraction of silver

The fill fraction approximate is achieved in case fill fraction of silver about 3% and higher for both case PVP-based and PVA-based materials Consistent with results in the previous section, the deep of index of refraction spectrum of PVP-based material appears at longer wavelengths than PVA’s The trend of results for PVP-based materials are similar as available results that confirmed by experiment in [33] However, the confirmed results are just limited for wavelength over 500nm The reliability of prediction should be considered by comparison with experiments

0.3 0.8 1.3 1.8 2.3 2.8 3.3

0.35 0.4 0.45 0.5 0.55

Index

of r

efraction

Wavelength (m)

27

Fig 4.7: The index of extinction of silver and PVP including 2%, 3%, 4% and 5% volume fill fraction of silver

Fig 4.8: The index of extinction of silver and PVA including 2%, 3%, 4% and 5% volume fill fraction of silver

The dependence of index of extinction on fill fraction is described in Fig 4.7 and Fig 4.8 corresponding to PVP-based and PVA-based materials The position of peak on index of extinction spectrum is quite similar as confirmed result in [15] As inferring naturally, the higher fill fraction is, the greater extinction will be There is

0 0.5 1.5 2.5 3.5

0.35 0.4 0.45 0.5 0.55

Index of ex tiction Wavelength (mm) Ag PVP_Ag2% PVP_Ag3% PVP_Ag4% PVP_Ag5% 0.5 1.5 2.5 3.5

0.35 0.4 0.45 0.5 0.55

Index

of ex

tinction

Wavelength (m)

Ag

28

the unnatural point that the indexes of extinction in case about 5% fill fraction or more are higher than silver’s This problem is being considered and will be discuss more detail in other research

4.1.3 Calculation for thin film following EMT using TMM

Applying Maxwell Garnett equation, the dielectric constant of material can be calculated in case of known fill fraction and host and inclusion material Using these data, the transmittance of thin films can be calculated using TMM Fig 4.9 and Fig 4.10 illustrate transmittance spectrum of different silver fill fraction thin film films that thick 30nm corresponding to PVP-based and PVA-based materials The glass substrate is though as reference The incident light is Transverse Electric and comes perpendicular to the surface of thin film The calculation is processed following section 2.2

Fig 4.9: Transmittance spectrum of 30 nm PVP-based films corresponding to different Ag fill fraction

0 0.2 0.4 0.6 0.8

0.3 0.4 0.5 0.6 0.7 0.8

T

ra

nm

ittan

ce

(un

it)

Wavelength (m)

29

Fig 4.10: Transmittance spectrum of 30 nm PVA-based films corresponding to different Ag fill fraction

For both two cases, the trends of spectrum are similar The deeps of spectrum of PVP-based material are red-shift (approximate 5nm in case of 5% fill fraction) compare with PVA-based material’s There is a point that the transmittance is higher than 100% in some region The reason is that the calculated indexes of refraction in those regions are higher than (assumed index of refraction of air) and lower than 1.5 (assumed index of refraction of glass substrate) In this case, the transmittance of system air/material/glass/air should be higher than the transmittance of system air/glass/air that considered as reference Here, the problem is that the regions of wavelength where material has low index of refraction (lower than 1) are in extinction region Although these regions not include deep of spectrum (the strongest extinction region), there is an obstacle for index of refraction determination The very common method for refractive index determination is determining the Brewster angle [14] using reflectometer In this case, the extinction can make the Brewster angle determination become inaccurate

Because of the lack of tool, the comparison between calculated transmittance spectrum and experimental results might be only way to preconceive the index of

0 0.2 0.4 0.6 0.8

0.3 0.4 0.5 0.6 0.7 0.8

T

ransm

ittance

(unit)

Wavelength (m)

30

refraction of material To compare with transmittance of fabricated film that will be discussed later, the transmittance spectrum of 200nm film were calculated and shown in Fig 4.11 and Fig 4.12 The overall trend of calculated transmittance spectrum of 200nm films is quite similar as in case of 30nm films However, the role of extinction becomes clearer There are some variations of calculated transmittance due to the variation of predicted index of extinction

Fig 4.11: The calculated transmittance spectrum of 200 nm PVP-based films corresponding to different Ag fill fraction using TMM

0 0.2 0.4 0.6 0.8

0.3 0.4 0.5 0.6 0.7 0.8

T

ransm

ittance

(unit)

Wavelength (m)

31

Fig 4.12: The calculated transmittance spectrum of 200 nm PVA-based films corresponding to different Ag fill fraction using TMM

4.1.4 Calculation for thin film using FDTD method

To predict transmittance spectrum of films, the FDTD method is used to calculate the transmittance of 200 nm film for both PVP-based and PVA-based material including 20 nm silver nanoparticles The description of domain is shown in Fig 4.13 The length of domain correspond to x, y, z direction are 200 nm, 200 nm and 600nm, respectively The grid size for all direction is 1.5nm The boundary condition for x and y direction is periodic and for z direction is 25nm thick PML The host material dielectric constants are defined following [20,31] The inclusions are spheres which are sized 20 nm diameter They dielectric constants are defined as Ag that following [30] Each sphere is distributed randomly and spacing the other at least nm The launch field is typed plane wave and expose incident light that is Transverse Electric and comes perpendicular to the surface of thin film Time step is auto optimized following Courant condition [34] The monitor is sized 200x200 nm2 The output is the power of light depends on wavelength in region from 300 nm

to 800 nm The estimated memory is about 133.1 megabytes The calculation processes are stable This model is different compare with real samples The

0 0.2 0.4 0.6 0.8

0.3 0.4 0.5 0.6 0.7 0.8

T

ransm

ittance

(unit)

Wavelength (m)

32

disappearance of glass substrate on the calculation model could be a factor that affect to accuracy of calculation The more accurate model will be considered on later research because limitation of time and calculation condition

Fig 4.13: The FDTD domain for calculation of 200nm film by x, y, z direction and 3D visions

Launch field Monitor Monitor

Launch field

100nm

100nm 200nm

33

Fig 4.14: The calculated transmittance spectrum of 200 nm PVP-based films corresponding to different Ag fill fraction using FDTD method

Fig 4.15: The calculated transmittance spectrum of 200 nm PVA-based films corresponding different Ag fill fraction using FDTD method

The calculate transmittance spectrum of 200 nm films using FDTD method are shown in Fig 4.14 and Fig 4.15 It can be seen that the main deeps of

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

0.395 0.495 0.595 0.695 0.795

T

ransm

ittance

(unit)

Wavelength (m)

PVP_Ag3%_FDTD PVP_Ag4%_FDTD PVP_Ag5%_FDTD PVP only_FDTD 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

0.3 0.4 0.5 0.6 0.7 0.8

T rans m ittan ce (un it)

Wavelength (m)

34

transmittance spectrum are similar both in case applying TMM for UEM and FDTD method However, there are a few of fluctuations of transmittance in region of wavelength about 450 nm to 650 nm It has two reasons:

- The first reason may be is related to grid As above mentioned, the grid volume of domain is 1.5x1.5x1.5 nm3 It is should be enough small compare with size of particles but not with distance between particles However, the decrease of grid size affects a lot to memory demand and calculation time The more precise calculation will be processed in the future

- The second reason is the neighbor-particles interaction Although the calculation of FDTD is very time-consuming, it has a plus point is that we can make the model which illustrate the real samples Experimentally, the distribution of silver particles in polymer is uncontrollable Assumed that the distribution is random, we can hope the above model can represent the real thin films In this case, the distance of particles can’t precisely comply with MGT Then, it has two consequences One is in the scattering signals can be clearer and the other is the existence of some bigger particles, rod-likes particles… The second consequence will be roughly considered in next section The FDTD can’t provide result exactly similar as measurement result but it still shows us the trend of results And it also can provide a reasonable explanation for experimental results in experimental parts

4.1.5 Neighbor particles interaction

This section is a bonus part about FDTD calculation Here, I tried building a simple but reasonable model called as model L that illustrated in Fig 4.16 The length of domain correspond to x, y, z direction are 280 nm, 280 nm and 400 nm, respectively The grid size for all direction is 1nm The boundary condition for all direction is 20nm PML The dielectric constant of constituents, the setting of launch field is defined similar as for thin films The monitor is sized 280x280 nm2 The

35

“medium slab” that has two particles inside The purpose is finding out the signal of neighbor-particles interaction through considering the changing of transmittance following particles’ distance change

36

Fig 4.17: Calculated extinction spectra of two neighbor-particles with distance equal 3nm in medium that has refractive index equal 1.5 using FDTD

The results for case of distance of particles is nm using FDTD method is showed in Fig 4.17 It can be seen that have extinction signals The peak at wavelength 410 nm is the ordinary signal of silver nanoparticles However, the other peak at wavelength 500 should be the signal of neighbor-particles interaction This thing can be explanation for fluctuation in mentioned transmittance spectra calculated by FDTD This problem is considered more detail by Discrete Dipole Approximation method (DDA) in Fig 4.18 It shows that with different number of neighbor-particles the extinction signals are different The location of peaks in results of two methods is quite similar The red shift of spectrum by FDTD method compared with DDA method is due to the difference of index of refraction of surrounding medium The calculation by FDTD method was processed using refractive index of medium about 1.55 but it’s 1.5 in case of DDA This is a phenomenon that hasn’t reported as I know For avoid any careless conclusion, I stop with just introduce about this phenomenon It’s necessary to make more calculation and experiment for more detail discussion about this interesting problem

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

0.3 0.4 0.5 0.6 0.7 0.8

Log

T

(un

it)

37

Fig 4.18: Calculated extinction spectra of neighbor-particles with distance equal 3nm in medium that has refractive index equal 1.5 using DDA

4.2 Experiment results

4.2.1 Properties of silver nanoparticles

The solution of silver nanoparticles has dark brown color that is shown in Fig 4.19 This figure also shows images of silver nanoparticles solution in three steps on cleaning process It can be seen that the concentration of nanoparticles solution is decreased a lot due to cleaning The good point is that it has a large amount of particles which are should have smaller size than participate particles, still in leachate after centrifugation It means that the synthesis is successful But, the bad point is that the centrifugation speed looks not very effective for cleaning process The lost still continue in case using very low concentration solution Maybe, the higher centrifugation speed should be better if it possible

0 10 12 14 16 18 20 22

300 350 400 450 500 550 600

QEX

T

(un

it)

Wavelength (nm)

38

Fig 4.19: The images of silver nanoparticles solution after synthesis(a), after centrifugation(b) and after re-disperse on water(c)

The SEM image in Fig 4.20 confirmed that particles size is approximate 20nm Besides that, the transmittance spectrum of solution in Fig 4.21 also shows that the size of self-synthesis silver nanoparticles is comparable with commercial sample used as standard sample The concentration of self-synthesis samples should be evaluated through comparing amplitude of deep that due to LSPR of silver nanoparticles in spectrum of very low concentration self-synthesis samples and commercial sample which has known concentration There is a point that the commercial sample has one more deep which not appear in case of self-synthesis samples It could be due to the existence of a family of particles that bigger than 20nm diameter It suggests a problem that may be the self-synthesis solution has less participate because of the remaining of PVP on solution This problem will cause errors for fill fraction of nanoparticles in materials So, this section is just for introduce a promise able synthesis process and general quality of products The optimization will be considered later In this study, the commercial Ag nanoparticle is used for all the other experiments

39

Fig 4.20: SEM image of self-synthesis silver nanoparticles

Fig 4.21: Transmittance spectrum of self-synthesis and commercial silver nanoparticles solution

75 80 85 90 95 100 105

300 400 500 600 700 800

T

ra

nsm

ittan

ce

(un

it)

Wavelength (nm)

40 4.2.2 Properties of thin films

Two polymers used as the host material are PVP and PVA which have molecular formula shown in Fig 4.22 On visible region, the index of refraction of PVP gradually decreases from 1.5606 to 1.5207 and the index of extinction decrease from 0.0045 to 0.0014 [24] The index of refraction of PVA also decreases from 1.5338 to 1.4702 on 300 to 800nm wavelength region Assumed that the extinction of polymers can be neglected, the transmittance of polymers can be evaluated using Fresnel’s equation In glass substrate, the transmittance of PVP film is about 0.9077 to 0.9127 and PVA film’s is about 0.9111 to 0.9182 compare with vacuum Meanwhile, transmittance of glass is about 0.9091 Theoretically, the transmittance of substrate with film is higher than glass substrate in some regions of wavelength In some cases, the results of real samples can be different due to scattering of ununiformed surfaces of film

PVP PVA

Fig 4.22: Molecular formula of PVP and PVA

41

Fig 4.23: Transmittance spectrum of PVA, PVP solution with and without existence of silver nanoparticles

The fabricated thin films by drop-coating method are light brown and transparent Fig 4.24 shows transmittance depending on wavelength of 3% silver nanoparticles fill fraction PVP-based and PVA-based films compare with the transmittance of PVP, PVA without silver and silver sample in glass substrate The spectra of PVP and PVA without silver don’t show any deep while spectra of polymer including silver show the deeps in wavelength about 400nm corresponding extinction Those deep are due to LSPR of silver nanoparticles that also exist in spectrum of dropped silver sample The red shift of LSPR signal in PVP and PVA host compare with dropped silver particles is related to index of refraction of host medium In case of dropped silver particle, the host medium can be assumed as the air which has refractive index equals It is smaller than refractive index of both PVP and PVA in considered region of wavelength The spectra have many fluctuations in region 300 – 400nm because of used device The thickness of films made by drop-casting method is almost could not determine by Alpha step profile because of non-uniform surface There is unnatural point that transmittance of the films is much higher than the solution used in above case although the concentration of silver in films should much higher than in solution (at least in case of applying

0.7 0.75 0.8 0.85 0.9 0.95

0.3 0.4 0.5 0.6 0.7 0.8

T

ransm

ittance

(unit)

Wavelength (m)

PVA solution

PVA_Ag3% solution PVP solution

42

drop-casting method) The explanation demands more experiments that I could not complete on this research

Fig 4.24: Transmittance spectrum of drop-coating PVP, PVA films corresponding 3% fill fraction of silver nanoparticles

As showed in Fig 4.25, the transmittance spectrum of spin-coating PVP-based thin film with different fill fraction of silver nanoparticles has no signal of extinction due to silver nanoparticles The explanation could be that the viscosity of PVP solution is too low When sample is rotated, almost the amount of solution is almost wiped out The remaining solution is just enough for a thin film with a mount of silver particles too little to make a signal in spectra To avoid this problem, I tried decrease rotation speed But, the results are not good The fabricated films are as thick as drop coating and non-uniform

0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1.01 1.02

0.3 0.4 0.5 0.6 0.7 0.8

T

rans

m

ittan

ce

(un

it)

Wavelength (m)

43

Fig 4.25: Transmittance spectrum of PVP-based films different fill fraction of silver nanoparticles

The transmittance spectra of PVA-based films are showed in Fig 4.26 The thicknesses of films are approximate 164nm, 132nm and 136nm corresponding to 3%, 4% and 5% fill fraction of silver particles, respectively Here, we can expect the deeps to be LSPR signals, especially the deep in case of 3% silver fill fraction Now, it has two issue of this case

 The first issue is the huge difference between the extinction occurring in real samples and calculation Actually, this problem was already reported in [10] The actual index of extinction is usually lower than calculated by Maxwell Garnett’s equation It has some people have developed different model than MGT to get more accurate approximation, such as [39] This work also will be processed on continuing research

 The second issue is the blue shifts of deep from 3% fill fraction case to 5% fill fraction case It has a supposition that the LSPR signal of each particle is affected by the decreasing dielectric constant of overall medium due to effect from them The existence of silver nanoparticles decreases index of

0.99 0.995 1.005

0.3 0.4 0.5 0.6 0.7 0.8

T

ransm

ittance

(unit)

Wavelength (m)

44

refraction Then this decrease makes blue shift of LSPR signal For consider this phenomena, the range of particles fill fraction will be expanded later

Fig 4.26: Transmittance spectrum of PVA-based films different fill fraction of silver nanoparticles

On wavelength region about 310nm to 400nm, the transmittance of PVA-based films with silver particles (≥ 98.5%) is higher than calculated transmittance of films without particles ( ≥96%) It suggests that the index of refraction of PVA including silver is lower than bare PVA’s and higher than in this wavelength region of light This trend of index of refraction is quite similar as predicted index by EMT So, the decrease of index of extinction due to the nanoparticles is confirmed for PVA thin film As an initial research, this research phase can be considered as completed The next phase will focus to optimize the calculation approximation and to determine precisely both of index of refraction and index of extinction of material

0.975 0.98 0.985 0.99 0.995 1.005 1.01 1.015 1.02

0.3 0.4 0.5 0.6 0.7 0.8

T

rans

m

ittan

ce

(un

it)

Wavelength (m)

45

CONCLUSION

In conclusion, the initial study about the low refractive index and low loss material based on PVP and PVA including silver nanoparticles has conducted The numerical calculation was run using Wolfram Mathematica software and FullWAVE software Applying the approximation following Maxwell Garnett topology for uniform effective medium, the index of refraction and index of extinction of materials was calculated The silver is predicted to be a better element of inclusion for object material compare with gold and copper The suitable size of using silver particles is 20 nm diameter The suitable fill fraction of silver nanoparticles is about 3% to less than 5% Theoretically, the indexes of refraction of those materials are lower than in region of wavelength about 380 to 400 nm The transmittance of thin films based on two types of material was calculated using two methods those are TMM apply refractive index predicted by EMT and FDTD method They verify the existence of LSPR signal at wavelength about 400 nm The predicted extinction is stronger than in real samples The FDTD method also introduces the problem of neighbor-particles interaction affect to transmittance of films It illustrates the picture that should be nearly similar as in real films

46

REFERENCES

[1] Abdellatif, S., Sharifi, P., Kirah, K., Ghannam, R., Khalil, A S G., Erni, D., & Marlow, F (2018a) Refractive index and scattering of porous TiO films

Microporous and Mesoporous Materials, 264, 84–91 https://doi.org/10.1016/j.micromeso.2018.01.011

[2] Acquaroli, L.N., Urteaga, R., & Koropecki, R R (2010b) Innovative design for optical porous silicon gas sensor Sensors and Actuators B: Chemical, 149(1), 189–193 https://doi.org/10.1016/j.snb.2010.05.065

[3] Acquaroli, Leandro N., Urteaga, R., Berli, C L A., & Koropecki, R R (2011c) Capillary Filling in Nanostructured Porous Silicon Langmuir, 27(5), 2067– 2072 https://doi.org/10.1021/la104502u

[4] Acquaroli, Leandro N (2018d) Matrix method for thin film optics Retrieved from https://arxiv.org/pdf/1809.07708.pdf

[5] Agarwal, S., & Prajapati, Y K (2019e) Multifunctional metamaterial surface for absorbing and sensing applications Optics Communications, 439, 304–307 https://doi.org/10.1016/J.OPTCOM.2019.01.020

[6] Al-Kuhaili, M F., Durrani, S M A., El-Said, A S., & Heller, R (2017f) Enhancement of the refractive index of sputtered zinc oxide thin films through doping with Fe2O3 Journal of Alloys and Compounds, 690, 453–460

https://doi.org/10.1016/j.jallcom.2016.08.165

[7] AL-Rjoub, A., Rebouta, L., Costa, P., Barradas, N P., Alves, E., Ferreira, P J., … Pischow, K (2018g) A design of selective solar absorber for high

temperature applications Solar Energy, 172(April), 177–183 https://doi.org/10.1016/j.solener.2018.04.052

47

https://doi.org/10.1016/0370-1573(78)90009-1

[9] Bethune, D S (1989i) Optical harmonic generation and mixing in multilayer media: analysis using optical transfer matrix techniques Journal of the Optical Society of America B, 6(5), 910 https://doi.org/10.1364/JOSAB.6.000910 [10] Boyd, R W., & Sipe, J E (1994j) Nonlinear optical susceptibilities of layered

composite materials Journal of the Optical Society of America B, 11(2), 297 https://doi.org/10.1364/JOSAB.11.000297

[11] Choi, M., Choe, J H., Kang, B., & Choi, C G (2013k) A flexible

metamaterial with negative refractive index at visible wavelength Current Applied Physics, 13(8), 1723–1727 https://doi.org/10.1016/j.cap.2013.06.028

[12] Cox, P M (2013l) DESIGN AND FABRICATION OF LOW LOSS AND LOW

INDEX OPTICAL METAMATERIALS 2(SGEM2016 Conference Proceedings,

ISBN 978-619-7105-16-2 / ISSN 1314-2704), 1–39

[13] EMLab (n.d.-m) EE 5303 Electromagnetic Analysis Using Finite-Difference Time-Domain Periodic Structures in FDTD Retrieved from

http://www.emlab.utep.edu/ee5390fdtd/Lecture 19 Periodic structures in FDTD.pdf

[14] Fowles, G R (1975n) Introduction to Modern Optics.pdf (p 333) p 333 https://doi.org/10.1119/1.1975142

[15] Gittleman, J I., & Abeles, B (1977o) Comparison of the effective medium and the Maxwell-Garnett predictions for the dielectric constants of granular metals Physical Review B, 15(6), 3273–3275

https://doi.org/10.1103/PhysRevB.15.3273

[16] Gric, T., Hess, O., Gric, T., & Hess, O (2019p) Metamaterial Cloaking

48

[17] Guenneau, S., & Ramakrishna, S A (2009q) Negative refractive index, perfect lenses and checkerboards: Trapping and imaging effects in folded optical spaces Comptes Rendus Physique, 10(5), 352–378

https://doi.org/10.1016/J.CRHY.2009.04.002

[18] Hamouche, H., Shabat, M M., & Schaadt, D M (2017r) Multilayer solar cell waveguide structures containing metamaterials Superlattices and

Microstructures, 101, 633–640 https://doi.org/10.1016/J.SPMI.2016.08.047 [19] Heavens, O S (1991s) Optical properties of thin solid films Retrieved from

https://books.google.com.vn/books/about/Optical_Properties_of_Thin_Solid_F ilms.html?id=yfJjiZBAc-AC&redir_esc=y

[20] Islam, S S., Iqbal Faruque, M R., & Islam, M T (2017t) A dual-polarized metamaterial-based cloak Materials Research Bulletin, 96, 250–253

https://doi.org/10.1016/J.MATERRESBULL.2017.02.039

[21] Jackson, J D., & Wiley, J (n.d.-u) Classical Electrodynamics Third Edition Retrieved from

https://cds.cern.ch/record/490457/files/9780471309321_TOC.pdf

[22] Johnson, S G (2007v) Notes on Perfectly Matched Layers (PMLs) Retrieved from http://math.mit.edu/~stevenj/18.369/pml.pdf

[23] Kabashin, A V., Evans, P., Pastkovsky, S., Hendren, W., Wurtz, G A., Atkinson, R., … Zayats, A V (2009w) Plasmonic nanorod metamaterials for biosensing Nature Materials, 8(11), 867–871

https://doi.org/10.1038/nmat2546

[24] König, T A F., Ledin, P A., Kerszulis, J., Mahmoud, M A., El-Sayed, M A., Reynolds, J R., & Tsukruk, V V (2014x) Electrically Tunable Plasmonic Behavior of Nanocube–Polymer Nanomaterials Induced by a Redox-Active Electrochromic Polymer ACS Nano, 8(6), 6182–6192

49

[25] König, T A F., Ledin, P A., Kerszulis, J., Mahmoud, M A., El-Sayed, M A., Reynolds, J R., & Tsukruk, V V (2014y) Electrically Tunable Plasmonic Behavior of Nanocube–Polymer Nanomaterials Induced by a Redox-Active Electrochromic Polymer ACS Nano, 8(6), 6182–6192

https://doi.org/10.1021/nn501601e

[26] Lei Kuang, & Ya-Qiu Jin (n.d.-z) Implementation of the Periodic Boundary Condition in FDTD Algorithm for Scattering from Randomly Rough Surface

2005 Asia-Pacific Microwave Conference Proceedings, 4, 1–3 https://doi.org/10.1109/APMC.2005.1606911

[27] Liu, J., Chen, H., Jing, X., & Hong, Z (2018aa) Guided mode resonance in terahertz compound metamaterial waveguides Optik, 173, 39–43

https://doi.org/10.1016/J.IJLEO.2018.08.001

[28] Liu, W., Fan, F., Chang, S., Hou, J., Chen, M., Wang, X., & Bai, J (2017ab) Nanoparticles doped film sensing based on terahertz metamaterials Optics Communications, 405, 17–21 https://doi.org/10.1016/J.OPTCOM.2017.07.086 [29] Markel, V A (2016ac) Introduction to the Maxwell Garnett approximation:

tutorial Journal of the Optical Society of America A, 33(7), 1244 https://doi.org/10.1364/josaa.33.001244

[30] McPeak, K M., Jayanti, S V., Kress, S J P., Meyer, S., Iotti, S., Rossinelli, A., & Norris, D J (2015ad) Plasmonic Films Can Easily Be Better: Rules and Recipes ACS Photonics, 2(3), 326–333 https://doi.org/10.1021/ph5004237 [31] Of, J., & Фізики, N.-A E P Ж Н.-Т Е (2016ae) Metamaterials: Theory,

Classification and Application Strategies (Review) Том, 8(4), 4088 https://doi.org/10.21272/jnep.8(4(2)).04088

50

[33] Pei, Y., Yao, F., Ni, P., & Sun, X (2010ag) Refractive index of silver nanoparticles dispersed in polyvinyl pyrrolidone nanocomposite Journal of Modern Optics, 57(10), 872–875

https://doi.org/10.1080/09500340.2010.491923 [34] Protégé-Frames (2000ah) FullWAVE - Users Guide

[35] Ramm, A G (2008ai) Does negative refraction make a perfect lens? Physics Letters A, 372(43), 6518–6520

https://doi.org/10.1016/J.PHYSLETA.2008.09.003

[36] Schneider, J B (2010aj) Understanding the Finite-Difference Time-Domain Method Retrieved from www.eecs.wsu.edu/~schneidj/ufdtd, 2010

[37] Schnepf, M J., Mayer, M., Kuttner, C., Tebbe, M., Wolf, D., Dulle, M., … Fery, A (2017ak) Nanorattles with tailored electric field enhancement

Nanoscale, 9(27), 9376–9385 https://doi.org/10.1039/c7nr02952g [38] Schwartz, B T., & Piestun, R (2003al) Total external reflection from

metamaterials with ultralow refractive index Journal of the Optical Society of America B, 20(12), 2448 https://doi.org/10.1364/JOSAB.20.002448

[39] Sheng, P (1980am) Theory for the Dielectric Function of Granular Composite Media Physical Review Letters, 45(1), 60–63

https://doi.org/10.1103/PhysRevLett.45.60

[40] Sipe, J E., & Boyd, R W (1992an) Nonlinear susceptibility of composite optical materials in the Maxwell Garnett model Physical Review A, 46(3), 1614–1629 https://doi.org/10.1103/PhysRevA.46.1614

[41] Sipe, John E., & Boyd, R W (2007ao) Nanocomposite Materials for Nonlinear Optics Based on Local Field Effects Optical Properties of

51

[42] Sugumaran, S., Jamlos, M F., Ahmad, M N., Bellan, C S., & Schreurs, D (2018ap) Nanostructured materials with plasmonic nanobiosensors for early cancer detection: A past and future prospect Biosensors and Bioelectronics,

100, 361–373 https://doi.org/10.1016/J.BIOS.2017.08.044

[43] Taflove Susan Hagness, A C., & London, B I (n.d.-aq) Computational Electrodynamics The Finite-Difference Time-Domain Method Third Edition

ARTECH HOUSE Retrieved from

https://cds.cern.ch/record/1698084/files/1580538320_TOC.pdf

[44] Tang, T., & Luo, L (2015ar) Enhancement of lateral shift in indefinite

metamaterial–insulator–metamaterial waveguides Optik, 126(20), 2547–2549 https://doi.org/10.1016/J.IJLEO.2015.06.040

[45] Urteaga, R, Marín, O., Acquaroli, L N., Comedi, D., Schmidt, J A., & Koropecki, R R (2009as) Enhanced photoconductivity and fine response tuning in nanostructured porous silicon microcavities Journal of Physics: Conference Series, 167(1), 012005

https://doi.org/10.1088/1742-6596/167/1/012005

[46] Urteaga, Raúl, Acquaroli, L N., Koropecki, R R., Santos, A., Alba, M., Pallarès, J., … Berli, C L A (2013at) Optofluidic Characterization of Nanoporous Membranes Langmuir, 29(8), 2784–2789

https://doi.org/10.1021/la304869y

[47] Wang, W., Yadav, N P., Shen, Z., Cao, Y., Liu, J., & Liu, X (2018au) Two-stage magnifying hyperlens structure based on metamaterials for super-resolution imaging Optik, 174, 199–206

https://doi.org/10.1016/J.IJLEO.2018.08.064

52