laser microprocessing and fabrication of structures … · mechanical properties 4 1.2. methods for...

LASER MICROPROCESSING AND FABRICATION OF

STRUCTURES ON GLASS SUBSTRATES

HUANG ZHIQIANG

NATIONAL UNIVERSITY OF SINGAPORE

2010

LASER MICROPROCESSING AND FABRICATION OF

STRUCTURES ON GLASS SUBSTRATES

HUANG ZHIQIANG

(B. Eng. (Hons), National University of Singapore)

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2010

Acknowledgement

i

ACKNOWLEDGEMENTS

I would like to express my earnest gratefulness to my supervisor, A/Prof. Hong

Minghui, for his guidance and support during my course of study. The useful and invaluable

advices have made it possible for me to complete this thesis so smoothly. I also appreciate

the help and the personal lessons he has given me along the way.

I would also like to thank Dr. Lap Chan and Dr. Ng Chee Mang for believing in me

and giving me a chance to do this project. I have benefitted and learnt much during the

weekly student meeting. I would also like to thank Dr. Lin Qun Ying who so readily

accepted to be my mentor in the company and has given me many opportunities along the

way.

I would also like to thank my friends and colleagues in ECE-DSI Laser

Microprocessing Lab and DSI for the countless help and useful discussion they have given

me. They are always a ready source of ideas and solutions to my problems, both work and

non-work related. I cherished my time with them.

I thank my wife for her great encouragement, understanding and moral support

during these years. Her constant assurance gives me strength to carry on. My heartfelt

thanks to my dad and mum and my family members too, who showed their support in subtle

yet encouraging ways.

Lastly, I would like to thank and acknowledge with gratitude the scholarship

GLOBALFOUNDRIES Singapore Pte Ltd has provided me during the course of my Ph.D

candidature.

Table of contents

ii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS i

TABLE OF CONTENTS ii

SUMMARY v

LIST OF TABLES viii

LIST OF FIGURES ix

NOMENCLATURE xvii

CHAPTER 1 INTRODUCTION 1

1.1. Properties of glass 1

1.1.1. Optical properties 1

1.1.2. Mechanical properties 4

1.2. Methods for glass processing 5

1.2.1. Mechanical scribing and breaking 6

1.2.2. CO2 laser cutting 6

1.2.3. Waterjet cutting 7

1.2.4. Photolithography 7

1.2.5. Direct laser ablation 8

1.2.6. Laser Induced Plasma Assisted Ablation (LIPAA) 9

1.2.7. Laser Induced Backside Wet Etching (LIBWE) 9

1.2.8. Large area micro-/nano-processing 10

1.3. Applications of periodic arrays of micro-/nano-structures 11

1.4. Objectives and motivation 13

1.5. Research contribution 14

1.6. Thesis outline 16

1.7. Reference 18

Table of contents

iii

CHAPTER 2 HIGH QUALITY MICRO-PROCESSING OF GLASS

SUBSTRATES BY

LIBWE 25

2.1. Introduction 25

2.2. Etching mechanism of LIBWE 25

2.3. Experimental procedure for LIBWE 27

2.3.1. Type of laser used and raster scan by galvanometer 27

2.4. Results and discussion: Organic absorbing solution 29

2.4.1. Effect of absorbing liquid 29

2.4.2. By-product formation and removal by laser cleaning 34

2.4.3. Applications in micro-fluidic chips 38

2.5. Results and discussion: Inorganic absorbing solution 42

2.5.1. Direct etching of glass using copper (II) sulphate solution 42

2.5.2. Mechanism of the glass etching by laser irradiation of copper (II) sulphate

solution 47

2.5.3. Dicing structures of various shapes from glass substrate 48

2.6. Reference 50

CHAPTER 3 FABRICATION OF NANOHOLES ARRAY ON GLASS BY LASER

INTERFERENCE LITHOGRAPHY 54


3.2. Experimental 56

3.2.1. Process flow 56

3.2.2. LIL setup and exposure 58

3.2.3. Transfer of patterns from photoresist to Cr layer 61

3.2.4. Transfer of nanoholes array patterns from Cr layer to quartz substrate 63

3.3. Applications 65

3.3.1. Nanoholes arrays as a phase mask 65

3.3.2. Potential applications of the patterned nanoholes array 69

3.4. Reference 77

CHAPTER 4 FABRICATON OF MICRO-OPTICS ON GLASS 81


4.2. Fabrication of microlens array 82

4.2.1. Sample preparation 86

4.2.2. Expose of photoresist and reflow 86

4.2.3. Transferring the photoresist shape to the substrate 89

Table of contents

iv

4.3. Introduction to solid immersion lens 93

4.4. Fabrication of the micro-solid immersion lens 97

4.5. Applying the h-µSIL to improve the resolution of a MLA 101

4.5.1. Spot size analyses 103

4.5.2. Intensity profile of projected lines 105

4.5.3. Projection of ‘+’ shape 107

4.6. Reference 110

CHAPTER 5 PARALLEL MICRO-/NANO-PATTERNING BY

LASER MLA LITHOGRAPHY 112


5.1.1. Microlens array lithography 112

5.2. Experimental 114

5.2.1. Sample preparation 114

5.2.2. Experimental setup 115

5.3. Results and discussion 116

5.3.1. Arbitrary patterning 116

5.3.2. Arbitrary angle patterning 118

5.3.3. Resist trimming 121

5.3.4. Patterning and etching for phase mask fabrication 125

5.3.5. Laser MLA lithography using the fabricated MLA 134

5.4. Reference 137

CHAPTER 6 CONCLUSIONS AND FUTURE WORK 139

6.1. Conclusions 139

6.2. Recommendation for future work 142

APPENDIX 144

LIST OF PUBLICATIONS

Summary

v

SUMMARY

Glass is a useful and important material which has many uses in modern society,

ranging from microelectronics, solar cells, optical engineering, biomedical to everyday

appliances. It has excellent physical properties, such as high transparency in a wide

wavelength range, strong resistance to thermal and mechanical stress, and high chemical

stabilities. Failure in glass usually starts from a crack, which then progresses to eventual

failure of the whole glass surface. Therefore, this makes the quality processing of glass

challenging. On the other hand, advances in technology have created the need for micro-

and nano-structures. Conventional methods of processing glass are insufficient to fabricate

these micro- and nano-structures on glass. Semiconductor processing techniques, such as

photolithography and e-beam lithography, are costly tools to own and operate. Therefore,

the goal of this thesis is to investigate alternative methods to achieve micro-/nano-

processing of glass by the use of advanced lasers technologies. The methods are aimed to be

cost-effective and easy to be implemented.

Laser direct writing of glass with the assistance of absorbing liquids is investigated.

The effects of the laser and the absorbing liquid on the etching process are studied. By using

a mixture of organic substances, the absorption of the laser light is improved. This helps to

increase the etch rate and decrease the threshold to initiate etching. By-products are formed

during the process and dry laser cleaning using the same laser source is successfully

developed to remove the by-products. The method is then employed to fabricate high quality

micro-structures for micro-fluidics applications. Two different approaches are investigated.

The first approach forms the micro-fluidic channels at the laser written areas while the

Summary

vi

second approach forms the structures by removing away the unwanted areas. The technique

is further extended to the use of an inorganic liquid and a cost-effective near infra-red laser.

An etching mechanism is proposed based on the deposition of metal during the process.

Miniature arbitrary shapes are diced out from glass substrates, which are challenging to be

accomplished by other conventional methods.

Laser interference lithography (LIL) is investigated as a maskless and parallel

processing method to structure glass substrates. Large array of periodic patterns can be

fabricated. The LIL technique is successfully used to process quartz with an array of

nanoholes. The nanoholes array is of high quality and uniformly over a large area. The

processed quartz is then used for phase mask applications. Using the processed quartz as a

phase mask, nanoholes array can be replicated in a single exposure. Defect engineering

capabilities are also demonstrated. Simulation is carried out and the results match the

experimental results very well.

Laser microlens array lithography is explored for patterning glass with parallel and

direct writing capabilities as well. Arrays of arbitrary patterns can be fabricated rapidly with

this technique. It is used to fabricate arbitrary phase shift structures on glass. By etching to a

depth with 180° phase difference, destructive interference is introduced. Using a 365 nm

UV light for exposure of the phase shift structures, array of patterns with much smaller

feature sizes are obtained. Through the control of the exposure time, different sets of results

can be obtained using the same array of patterns. This shows that there is flexibility in the

design and patterning process. Simulation results also match the experimental results very

well.

Summary

vii

Micro-optics is formed on glass substrates using photoresist melt and reflow method.

A MLA is successfully fabricated on glass substrate and characterized. The same technique

is used as a novel method to fabricate an array of micro-solid immersion lens array. It is

then applied to a MLA and demonstrated to be functional and able to increase the resolution

of the MLA.

List of tables

viii

LIST OF TABLES

Table 4.1. The etching cycle used to transfer the photoresist onto the fused silica

substrate.

Table 4.2 Comparison between the magnification enhancements of the two SILs.

Table 5.1. Comparison between the lateral and vertical etching rates in the resist

trimming process.

.

List of figures

ix

LIST OF FIGURES

Fig. 2.1. The chamber for LIBWE experiment. The inlet is for filling up with the

absorbing liquid.

Fig. 2.2. UV-Vis spectra of toluene and pyrene dissolved in toluene at a concentration

of 0.4 M.

Fig. 2.3. Comparison between the etched depths achieved by different liquids. The

graph of pyrene/toluene solution shows a lower etching threshold and a

deeper etched depth, implying higher etching rate.

Fig. 2.4. Various depth profiles obtained in a single laser scan at varying laser

fluencies.

Fig. 2.5(a). The binding energy of carbon at 284.65 eV, 286.19 eV and 288.94 eV, which

correspond to C-C / C-H, C-O and C=O bonds, respectively.

Fig. 2.5(b). The binding energy of oxygen at 532.51 eV and 533.17eV, which correspond

to O-Si / O-C and O=C bonds, respectively.

Fig. 2.5(c). The binding energy at 103.18 eV is due to Si-O bonds, from the quartz

composition.

Fig. 2.6. Comparison between two substrate surfaces along the micro-fluidic channel

with and without laser cleaning.

Fig. 2.7(a). Profile scans of the microfluidics channels. Uniform cross-sections are

obtained.

Fig. 2.7(b). Zoomed in 3D view of the bifurcating junctions. The depth is uniformed

etched and the edges are crack-free.

List of figures

x

Fig.2.7(c). Grooves fabricated on glass surface. The surrounding unwanted areas are

milled away using LIBWE.

Fig. 2.8. The absorption spectrum of copper (II) sulphate solution. There is

transmission in the 300 nm ~ 600 nm range, and it absorbs strongly in the

NIR range.

Fig. 2.9(a). Survey scan of the surface after the irradiation of 1064 nm laser light. The

survey scan shows the presence of Cu, C, O, and Si.

Fig. 2.9(b). Narrow scan of the Cu2p peak. The position of the Cu2p1/2

and Cu2p3/2

shows

the presence of metallic Cu.

Fig. 2.10. Microscopic view of the glass surface in contact with CuSO4 solution after

the laser irradiation. There are materials deposited along the line edges.

Fig. 2.11. A schematic illustrating the process of glass material removal by continuous

irradiation of laser.

Fig. 2.12. (Top and bottom left) Glass substrates with a star and a circle diced out. The

glass substrates have remained intact after the laser processing. (Top right)

Circles of diameters 2 mm, 5mm, and 1.5 mm, from left to right, diced out

from glass substrates. (Bottom right) Different structures cut out from the

glass substrate, showing the star shape and the alphabets ‘C’ and ‘G’.

Fig. 3.1. Schematic illustration of a standing wave formed by the interference of two

light beams.

Fig. 3.2. Process flow for patterning the quartz surface.

Fig. 3.3. Schematic drawing of a Lloyd’s mirror setup for laser interference

lithography of periodic structures on photoresist.

List of figures

xi

Fig. 3.4. Illustration of the reflection angle of the reflected ray and the angle of

interference.

Fig. 3.5(a). Line array fabricated by a single exposure on negative photoresist.

Fig. 3.5(b). Nanohole array obtained after a LIL double exposure on negative photoresist.

Fig. 3.6. AFM profile of the photoresist pattern after a double LIL exposure and RIE

to open the SiO2 capping layer.

Fig. 3.7. Nanoholes array patterns transferred to the Cr hard mask layer.

Fig. 3.8. The processed quartz substrate with the nanoholes array etched transferred to

the substrate.

Fig. 3.9. A schematic illustration of the phase delay introduced to the wave front after

passing through the surface with the nanoholes. This results in a modulation

of the wave front.

Fig. 3.10. Simulation result of the light distribution after light passes through the

surface with the nanoholes array. The modulation of the phase of the light

causes the redistribution of the light energy, resulting in a periodic array of

enhanced light intensities.

Fig. 3.11. Periodic array of nanoholes patterned uniformly on the photoresist.. The top

right inset shows the profile of the nanoholes formed. The bottom inset shows

the exposure method.

Fig. 3.12. The nanoholes array patterned on photoresist transferred to the silicon

substrate with good fidelity and uniformity.

Fig. 3.13. A defect formed on photoresist. The right inset shows the corresponding

defect on the quartz phase mask.

List of figures

xii

Fig. 3.14. Simulation result of a defect among the nanohole array. The defect reduces

the light intensity of that region.

Fig. 3.15. SEM image of the line of defects created on the phase mask by FIB.

Fig. 3.16. Light intensity distribution at various Z planes. This 3D distribution of light

has interesting applications in 3D photonic crystals patterning.

Fig. 4.1. Fabrication steps for fabricating a microlens array using the photoresist

reflow method.

Fig. 4.2. The plot of f versus D at three different photoresist thicknesses.

Fig. 4.3(a). Photoresist pattern array formed after the double exposure by a chrome-on-

glass mask.

Fig. 4.3(b). After the melting, the photoresist patterns are reflowed into a spherical

profile.

Fig. 4.4. Illustration showing the need to totally transfer the photoresist pattern profile

to fused silica.

Fig. 4.5. AFM profile of the MLA transferred from the photoresist onto the fused

silica substrate.

Fig. 4.6. The fabricated MLA focused the incident light into an array of focused spots.

The spot size is measured as ~ 660 nm.

Fig. 4.7. The hemispherical configuration, where the incident light focused at the

centre of the sphere.

Fig. 4.8. Illustration showing how a half sphere can be used as a hemispherical solid

immersion lens for higher resolution imaging.

List of figures

xiii

Fig. 4.9. The aplanatic point of the sphere, where the incident light focuses at the point

Z0. It has a virtual focus at Z1.

Fig. 4.10. A super-hemispherical solid immersion lens used for the sub-surface imaging.

Fig. 4.11. The concept of the fabrication of a h-µSIL. The hemispherical centre of the

h-µSIL is offset to the back of the glass substrate.

Fig. 4.12. SEM image of the fabricated h-µSIL.

Fig. 4.13. Plot of the measured sag height of the h-µSIL. The sag height is ~ 3.4 µm

while the diameter is ~ 72 µm.

Fig. 4.14. The schematic drawing of the alignment setup used to align the h-µSIL to the

MLA.

Fig. 4.15. Alignment process showing the intentional offset between the h-µSIL and the

MLA.

Fig. 4.16(a). Spot size produced by the MLA before the attachment of h-µSIL. The spot

size is ~ 1.9 µm.

Fig. 4.16(b). The spot size produced by the MLA after the attachment of the h-µSIL. The

spot size reduced to ~ 1.3 µm.

Fig. 4.17(a). An array of 4 lines projected by the MLA. The same array of 4 lines is

projected by the MLA before and after the attachment of the h-µSIL.

Fig. 4.17(b). The width of the lines projected by the MLA without the h-µSIL and

measured by the beam analyzer.

Fig. 4.17(c). The same array of 4 lines projected by the MLA with the h-µSIL attached.

The width of the lines decreases to ~ 3 µm.

Fig. 4.18(a). An array of ‘+’ sign projected by the MLA without the h-µSIL array attached.

List of figures

xiv

Fig. 4.18(b). An array of ‘+’ sign projected by the MLA with the h-µSIL array attached.

The ‘+’ shapes are smaller due to the reduction effect of the h-µSIL.

Fig. 5.1. Illustration of the concept of laser MLA lithography.

Fig. 5.2. An array of dots formed after laser light irradiation through a MLA.

Fig. 5.3. Experimental setup for the laser MLA lithography.

Fig. 5.4. Microscope image of the MLA used in the experiment. The period of the

MLA is 50 µm.

Fig. 5.5(a). A logo array pattern fabricated on fused silica by the laser MLA lithography

and RIE.

Fig. 5.5(b). Generic IC design fabricated by laser MLA lithography on fused silica. The

inset shows the 3D AFM image of an unit cell after the patterns were

transferred to the fused silica.

Fig. 5.6. Schematic drawing of laser MLA patterning by moving the nanostage at an

angle to the horizontal axis.

Fig. 5.7. Line array patterns with different periods obtained by changing the angle of

the traverse direction.

Fig. 5.8. The left figure shows the feature size obtained after the laser MLA

lithography. The right figure shows the further reduced feature size obtained

after the resist trimming.

Fig. 5.9. Process flow to incorporate the resist trimming.

Fig. 5.10. Destructive interference occurs at the edges of the phase shift structures [5],

resulting in lower light intensities at the edges.

List of figures

xv

Fig. 5.11. An array of ‘L’ shapes patterned using the laser MLA lithography and RIE.

The inset shows the 3D AFM image of a ‘L’ shape. The surface is uniform,

showing good control of the height uniformity.

Fig. 5.12. By using the fabricated phase shift structures for UV exposure, periodic

arrays of ‘double L’ are obtained.

Fig. 5.13. Simulated result showing the regions of intensity minimum at the edges of

the structures (white dashed lines), due to the destructive interference of light.

Fig. 5.14. An array of square shape phase shift structures fabricated on fused silica. The

inset shows the 3D AFM image of one of the structure. It showed the through

hole at the centre and the uniform height.

Fig. 5.15. Two concentric square patterns formed on photoresist. The smaller square is

due to the interior edge of the square phase structure while the bigger square

is due to the exterior edge.

Fig. 5.16. Simulated result showing the intensity under the square phase shift structure

as light passes through. There are two regions of low light intensities (dotted

lines), one at the exterior edge and the other at the interior edge. The through

hole concentrates the light at the centre of the square.

Fig. 5.17. By increasing the exposure time, the interior smaller square patterns on

photoresist can be deliberately exposed, leaving behind only the exterior

bigger square.

Fig. 5.18. An array of ‘L’ patterned by laser MLA lithography using the fabricated

MLA. The feature size is ~ 580 nm.

List of figures

xvi

Fig. 5.19. An array of 3 ‘L’s patterned parallel to one another. The minimum feature

size is ~ 440 nm. The space between the ‘L’s can still be distinguished.

Nonmenclature

xvii

NONMENCLATURE

T

0 Transmission κ Extinction coefficient

λ Wavelength α Absorption coefficient

M Molar F Laser fluence

p Period θ Angle of interference / Angle of

incidence

n Refractive index h Height of the nanholes / phase shift

structures

φ Phase difference d Diameter of nanoholes

ZT Talbot length c Lattice constant

x Distance γ Angle of traverse (Parallel patterning)

s Sag height of photoresist

/ microlens

D Diameter of lens / photoresist islands

t Thickness of photoresist J1 Bessel function of the first kind

I Intensity I0 Maximum intensity

k Wavenumber β Angle of observation

a Radius of aperture l Resolution limit

R Radius of curvature f Focal length

T Temperature P Pressure

Z0 Aplanatic point of sphere NA Numerical aperture

Chapter 1: Introduction

1

CHAPTER 1

INTRODUCTION

Glass is one of the important industrial materials today. It is used in many different

fields, such as building, automotives, electronics, lighting and optics. Glass is a uniform

amorphous solid material, usually produced when a viscous molten material cools rapidly to

below its glass transition temperature, without sufficient time for a regular crystal lattice to

form. Common glass contains about 70 ~ 72 weight % of silicon oxide (SiO2). The major

raw material is sand that contains almost 100% of crystalline silica in the form of quartz.

Even though it is an almost pure quartz, it still contains a little (<1%) iron oxides that would

color the glass, so this sand is usually enriched in the factory to reduce the iron oxide

amount to <0.05%. Large natural single crystals of quartz are purer silicon dioxide, and

upon crushing, are used for the high quality specialty glasses. Synthetic amorphous fused

silica (almost 100% pure) is the raw material for the most expensive specialty glasses. As

science and technology progress, there is an increasing demand for glass in new areas of

applications, such as flat panel display, hard disk manufacturing, biomedical, solar cell

manufacturing and micro-optics. Glass has many advantageous properties which make it

such an invaluable material for manufacturing.


2

1.1. Properties of glass

The properties of glass can be varied and regulated over an extensive range by

modifying the composition, production techniques, or both. These procedures fix the

molecular structure of the glass and determine its interaction with the electromagnetic

waves. The important properties of glass are its mechanical, optical, and thermal properties.

These properties of glass are briefly discussed in the follow sections.

1.1.1. Optical properties

When a beam of light falls on a piece of glass, some of the light is reflected from the

glass surface, some of the light passes through the glass, and the rest is absorbed by the

glass. Optical components make use of these properties of glass to absorb, emit or direct

light. The important optical properties of glass are its refractive index, dispersion, and

transmission.

The refractive index of glass is the measure of the speed of light as it travels through

glass. It is expressed as a ratio of the speed of light in vacuum relative to that in the glass.

Light travels more slowly through materials with a refractive index greater that of vacuum,

which has a value of 1. As a light wave enters the glass material, it encounters repetitive

scattering and rescattering by the atoms of the materials. Among atoms, light is traveling at

the light speed, c = 3 × 108 m/s. However, the scattering and rescattering events introduce a

phase shift into the light field and this eventually results in an apparent slow down of the

light as it travels through the glass [1]. In glass, the apparent light speed is about 2×108 m/s.


3

Light undergoes refraction at the interface between vacuum and glass according to

the Snell‟s law:

1

2

2

1

sin

sin

n

n

(1.1)

where θ1 and θ2 are the angle to the normal for the light rays in air and glass respectively,

and n1 and n2 the refractive index of air and glass. This refraction of light makes it possible

to use glass for making lenses, by shaping the lens to the proper shape to focus light.

The complex refractive index can be defined as:

inn ~

(1.2)

where n is the refractive index and κ the extinction coefficient, which describes loss in the

light energy as it passes through the glass. Both n and κ are dependent on the wavelength of

the light, and because n varies with the wavelength of light, this causes dispersion in glass.

Blue light sees a larger value of n as compared to red light. Upon refraction, blue light bends

more than red light and this is the property which makes a prism able to split the white light

into its constituent colours. French mathematician, Augustin-Louis Cauchy, was the first to

formulate the dispersion with respect to the refractive index:

42

cb

an (1.3)

with an accuracy within the visible spectrum of the order of 10-4

[2]. Coefficients a, b and c

are then determined experimentally by measuring the refractive index at different

wavelengths. There are various other equations which relate the dispersion to the refractive

index, such as the Sellimeier equation:


4

3

2

2

3

2

2

2

2

1

2

2

12 1C

B

C

B

C

Bn

(1.4)

where B1, B2, B3 and C1, C2, C3 are experimentally determined Sellmeier coefficients. It is

therefore an empirical formula that relates the refractive index to the wavelength of the

light, but offers higher degrees of accuracy [2].

The extinction coefficient κ in Eq. 1.2 is associated with loss of the light energy due

to absorption and scattering [3] as it propagates through the medium, and therefore directly

affects the transmission properties of glass. The transmission of light is given by the Beer-

Lambert Law:

x

i

t eI

IT 0 (1.5)

where T0 is the transmission, It the transmitted light, Ii the incident light, α the absorption

coefficient and x the distance the light has travelled. The composition which makes up the

glass material determines the amount of absorption and scattering at different wavelengths.

Therefore, a wide range of different glasses can be chosen for transmitting/absorbing light at

different wavelengths, for different applications, such as filters and windows.

1.1.2. Mechanical properties

Glass is an unusual material when subjected to mechanical stress because it returns

exactly to its original shape when the bending or stretching force is removed. This

characteristic of glass classifies it as a perfect elastic material [4]. If more force is applied,


5

the glass breaks when the force reaches the ultimate strength of the glass. But at any small

pieces of breakage, it can be found that the glass does not deform permanently.

There are three types of forces to be considered:

1. A tensile force exerts a pull on the material. A mild tensile force exerts a pull on

the material but a severe tensile force can pull the glass material apart.

2. A compressive force acts to squeeze the glass material.

3. A shear force acts on the material to slide one part of the material in one

direction and the other in the opposite direction.

Tensile force is the most important in glass because it gives rise to tensile strain

within the glass and glass breaks only from tensile tension. The strength of glass is only

slightly affected by composition but is highly dependent on surface conditions. This

explains why mechanical scribing of glass, where a scratch mark is made on the surface of

the glass, can easily aid the breakage of the glass. Strength is measured in the laboratory by

applying a load to glass. This stretches the lower surface of the glass material so that it is in

tension and squeezes the top surface, causing it to be in compression. The load is increased

until the glass breaks. The break originates in the lower surface since glass always fails from

tension.

1.2. Methods for glass processing

As discussed in the previous section, glass has excellent properties, making it an

excellent candidate for many applications. However, failure in glass usually starts from a

crack, which then progresses to eventual failure of the whole glass surface. Therefore, it is


6

challenging to process glass, especially high precision engineering. In the following

sections, methods used in the industry and research field are briefly discussed.

1.2.1. Mechanical scribing and breaking

Mechanical scribing and breaking are the classic and most prevalent glass separation

technologies. This is a process that involves the mechanical scribing of a vent in the upper

surface of the glass. This is usually accomplished with a diamond or tungsten carbide wheel.

Mechanical scribing and breaking perform reasonably well when the operational parameters

are optimized. High performance scribing requires optimization to minimize the damage to

the glass surface while maintaining a vent crack that is sufficient for a good break.

Unfortunately, maintaining control of operational parameters is often somewhat of an art,

not well understood by process engineers, and causes yield loss. The vent is created in only

the top 10 ~ 20 percent of the glass thickness. Applying a pressure on the opposite side of

the glass mechanically breaks the glass. This is generally done through an impact, roller or

guillotine breaker. The major limitation of this technology is that particulates are generated

during the scribing process along with surface damage and lateral cracks. These defects

generated during the scribing process often cause failure to the glass surface.

1.2.2. CO2 laser cutting

CO2 laser at a wavelength of 10.6 μm is opaque to glass as glass absorbs strongly at

this wavelength. Shining the glass surface with the laser beam precisely heats the glass at


7

the localized region. This is followed by a cold jet of air or an air/liquid mixture. This

thermally induced tension causes a precise fissuring of the glass which results in a high

quality cut. There are two kinds of cut: full body cut for materials of thickness from 50 μm

to 1 mm and the scribe and break for materials of thickness from 200 μm to 10 mm.

However, this cutting method requires a high power CO2 laser [5].

1.2.3. Water jet cutting

Water jet cutting makes use of a highly pressurized jet of a mixture of water and

other abrasives for cutting and processing materials [6]. It is a controlled and accelerated

erosion process which removes the materials through physical abrasion. However,

expensive pumps are needed to continuously stir the water in order to keep the abrasive

particles in suspension. Furthermore, the use of abrasives tends to clog up the jet outlet,

which results in frequent maintenance.

1.2.4. Photolithography

Photolithography can be used to define the desired patterns on photoresist and then

etching [7], either dry or wet etching, used to transfer the patterns from the photoresist down

to the glass substrate. Photolithography can offer high resolution surface patterning.

However, the photolithography system is costly and the expensive mask sets used to print

the desired patterns further add to the cost of the system. Furthermore, photolithography

systems are typically optimized to handle standard silicon wafers and could have difficulties


8

and limitations when used for patterning glass substrates. These restrictions limit the use of

photolithography for patterning glass.

1.2.5. Direct laser ablation

Another method to process glass is by laser direct ablation [8-10]. This is a direct

writing method whereby a laser, with laser fluence F higher than the ablation threshold Fth

of the glass material, irradiates on the sample. This deposition of high energy to the glass

surface causes the surface to heat up rapidly, resulting in subsequent melting of the glass

materials. Evaporation, vaporization or ejection of the materials then take place, resulting in

the ablation of the glass materials.

Various lasers have been used for laser direct ablation of glass. UV lasers [11-12]

and femtosecond lasers [13] have been used successfully to ablate glass. Usually, pulsed

laser is used due to the high peak power delivered during the short pulse duration τ, which is

typically in the range of 10 ~ 100 ns. The short pulse duration also minimizes the amount of

thermal damage distributed into the substrate. In order to use lasers for ablation of glass, the

glass must absorb the laser light for ablation to occur. However, glass is highly transparent

and this limits the types of lasers which could be used for direct ablation of glass. High

power deep UV lasers emitting light with photon energy larger than the bandgap of glass are

used for ablation of glass [11-12]. Ultrafast laser has very short pulse width and hence high

peak power. This enables it to cause the ablation of glass due to nonlinear processes [13],

such as two-photon and multiphoton absorption.


9

1.2.6. Laser Induced Plasma Assisted Ablation (LIPAA)

The LIPAA process was first described by Zhang et al. [14], where laser-induced

plasma helps in the process of ablation of the glass materials. In the LIPAA process, the

glass substrate is transparent to the incident laser beam. The laser hits the metallic target

below the glass substrate, which is less than 1 mm away, producing plasma that propagates

towards the glass surface. Diagnostic investigations revealed that the laser-induced plasma

can directly contribute to the ablation when the substrate-to-target distance is sufficiently

small (less than a few hundred microns), and its influence increases with decreasing distance

[14-17]. After the process, a thin metal film is often deposited on the ablated glass surface.

Different target materials produce different colors of metallic films. However, an additional

post-process cleaning step is needed to remove the deposited metallic film.

1.2.7. Laser Induced Backside Wet Etching (LIBWE)

Laser-induced backside wet etching (LIBWE) [18-27] is a process that can etch

transparent materials by laser ablation of a highly laser-absorbent organic solution. It was

first discovered by Wang et al. for the processing of fused silica [18]. This process has been

effectively used for micromachining a variety of transparent materials, including fused

silica, and CaF2 glass. LIBWE is an effective method to process glass with good quality and

this method will be further investigated in Chapter 2.


10

1.2.8. Large area micro-/nano-processing

Nanoimprint lithography [28] is a method of fabricating nanometer scale patterns. It

is a simple nanolithography process with low cost, high throughput and high resolution. It

creates patterns by the mechanical deformation of an imprint resist. The imprint resist is

typically a monomer or polymer formulation that is cured by heat or UV light during the

imprinting process. It is usually coated onto a substrate by spin coating. The mold, which is

pre-defined with topological patterns, is then brought into contact with the sample and

pressed together under a pressure. Subsequently, heat is applied to increase the temperature

of the polymer to its glass transition temperature. The pattern on the mold is pressed into the

softened polymer film. After cooling down, the mold is separated from the sample and the

pattern of the mold is left imprinted on the polymer. However, the major concern of this

method is the defects formed on the mold. Repairing the defects could incur high

maintenance and repair cost. Furthermore, any defects are reproduced on subsequent wafers.

This could affect the yield and increase the operation cost.

Nanosphere lithography (NSL) has developed into a very useful tool as a bottom-up

approach for creating a periodic array of self-assembled nanospheres. Typically,

nanospheres suspended in a solution were dispensed onto the substrate. By controlling the

charges of the substrate to be oppositely charged [29] with respect to those of the

nanospheres suspended in the solution, the colloids can be dispersed and subsequently self-

assembled on the surface of the wafer. Spin coating can also be used to produce

microcrystalline arrays of colloidal particles [30]. After creating the layer of self-assembled


11

nanospheres, it can then be used as a mask to deposit metals between the gaps of the

hexagonally packed nanospheres [30]. However, it is challenging to obtain a large array of

well-ordered nanospheres because defects, such as dislocations and grain boundaries, are

present during the process of self-assembly. This limits the area of patterns which are

usable.

Soft lithography [31-32] represents a non-photolithographic strategy for generating

patterns and structures with feature sizes ranging from 30 nm to 100 µm. In soft lithography,

an elastomeric stamp with patterned relief structures on its surface is used as the stamp for

replicating its patterns onto another material. This technique provides a convenient,

effective, and low-cost method for the formation and manufacturing of micro- and nano-

structures. Many variants have been developed, such as microcontact printing (µCP) [33],

replica molding (REM) [34], and solvent-assisted micromolding (SAMIM) [35]. However,

these methods also have disadvantages, such as damages and contamination introduced to

the master mold.

1.3. Applications of periodic arrays of micro-/nano-

structures

In recent years, periodic arrays of patterns have been used in an increasing number

of applications. Nanoholes array has been used widely in optics investigation ever since the

discovery of extraordinary light transmission through small holes by Ebbesen and his co-

workers [36]. Since that discovery, the field of plasmonics research explodes as numerous


12

plasmonics related investigations are conducted, such as optical bandpass filter due to the

selective wavelength transmission of light [37]. Moreover, an enhanced electrical field is

generated when incident light excites an active plasmonic mode in the nanoholes [38], and

this enhanced electrical field can be exploited for surface enhanced Raman scattering

(SERS) [39]. By exciting surface plasmon polaritons with a Kretschmann configuration,

bio-analyses based on refractive index sensing has been demonstrated [40]. Similarly, based

on refractive index sensing, ring resonators are emerging in the field of chemical analyses

[41-42].

Periodic array of nanostructures has also played an important role in the preparation

of anti-reflection surfaces [43-48]. Subwavelength structures for antireflection are inspired

by the moth-eye corneas where they comprise of an array of protuberant structures [49-51].

Because its periodicity is smaller than the wavelength of the incident light, the

nanostructured array surface behaves like an effective homogeneous medium with

continuous gradient of refraction index. By patterning the surface with arrays of

nanopatterns, the reflectivity of the surface can be reduced. These antireflection surfaces

have applications ranging from military to high performance solar cells.

Recently there is increased research interest in super-hydrophobic surfaces for self-

cleaning functions. Fundamentally, the super-hydrophobic surface mimics the surface of

the lotus leaf, which has arrays of micro- and nano-structures on the surface. This creates

micro-pockets of air among the structures and the water in contact with the surface, thereby

making the surface super-hydrophobic. Controlling the wettability of the surface by


13

controlling the surface geometries has aroused a lot of interest for both fundamental and

practical applications [52-54]. The super-hydrophobicity of these surfaces is enabled by the

creation of periodic nanostructures [55-56].

1.4. Objectives and motivation

As mentioned in Section 1.2, glass has various uses in many different applications.

However, the advantageous properties of glass, such as its resistance to mechanical stress,

also mean that it is challenging to process glass with good quality, especially if micro- and

nano-structures are to be fabricated on glass. The demand to fabricate smaller structures

increases with the advance of nanotechnology, and this has made conventional methods of

processing glass quite inadequate to meet this demand. On the other hand, photolithography

is a well established method to define the desired patterns. However, it is a very expensive

tool to own and operate. The photomask, which defines the desired patterns, is very costly.

Furthermore, any changes to the design would mean a new photomask is needed, which

increases the cost further. This makes direct writing techniques attractive because the

desired designs can be directly patterned without the need of expensive photomasks. E-

beam and focused ion beam lithography techniques are useful tools to perform nano-

patterning. However, they have a few drawbacks which could constraint their usage in the

patterning of glass. They are expensive tools to own and operate. In additional, they are

serial processes that have to operate in a high vacuum. This makes the patterning process

slow and tedious. These various limitations in the methods of processing glass provide the

objectives and motivation of this thesis:


14

1. To investigate alternative methods of glass processing. These methods should

provide high quality processing of glass by laser, and be cost effective.

2. To investigate how these methods can be used to fabricate high quality functional

micro- and nano-structures on glass. Various important process steps and parameters

are optimized. The fabricated micro- and nano-structures are then used for various

applications.

3. To investigate maskless and direct writing techniques that can create the patterns

without the need of a photomask with pre-defined patterns. This adds flexibilities to

the processing methods. Design changes can be done easily without the need to

change the photomask.

4. To investigate large area parallel processing techniques. Many novel applications are

exploiting the unique advantages of arrays of periodic structures. Therefore, there is

a demand to fabricate arrays of periodic structures quickly and with good quality.

The techniques investigated enable large arrays of periodic structures to be formed

easily and rapidly.

1.5. Research contribution

The main contribution of this thesis can be summarized as follows:

1. A laser direct writing technique that uses the assistance of a laser absorbing

liquid is studied and applied to the processing of glass by using a different type

of laser. Important relationship between the absorption and the etching rate of

the glass is established. It is shown, through the measurement of the absorption


15

spectra of the solutions, that a better absorption of the laser light results in a

lower etching threshold and a higher etch rate. By-product formation is

observed, studied in detail and removed by laser cleaning using the same laser

for in-situ cleaning.

2. The technique is then further extended to the use of an inorganic liquid and a

cost-effective near infra-red laser. The by-product formation is also analyzed.

An etching mechanism based on this formation of the by-product is proposed.

3. Using the technique, high quality microstructures can be fabricated for

microfluidics applications. Various delicate miniature structures can be diced

out from glass substrates using this method.

4. Maskless and parallel processing by laser interference lithography is

investigated for patterning periodic nanostructures on glass. A process flow is

formulated and optimized to produce a large array of periodic nanoholes on the

glass substrates. After fabricating the nanoholes on the glass, it is applied to

replicate the nanoholes array by a single exposure step. Defects can also be

deliberately introduced into the periodic nanoholes, and this has potential

applications in photonic crystals defect engineering.

5. Micro-optics is fabricated on glass substrates using a photoresist reflow

technique. MLA is successfully fabricated on glass using this technique. The

technique is also applied as a new process to fabricate an array of micro-solid

immersion lens array. It is successfully applied to increase the resolution of a

MLA.


16

6. To further extend the maskless and parallel processing capabilities, laser

microlens array lithography is investigated to fabricate large area arbitrary

patterns on glass. It is a parallel processing technique with direct writing

capabilities. Multiple arbitrary patterns can be directly written simultaneously at

the same time. Various phase shift structures are patterned using this technique

and their optical properties are studied and characterized. It is found that

features sizes smaller than the original structures can be patterned by using the

phase shift of light.

7. The fabricated MLA can also be used for laser microlens array lithography.

Smaller feature sizes can be obtained using the fabricated MLA.

1.6. Thesis outline

The outline of the thesis is as follows:

Chapter 2 presents the results of using a UV laser for liquid assisted etching of glass.

The experimental setup, the optimized parameters of the etching process, as well as the

experimental results will be discussed. It also shows how the technique can be extended to

the use of another inorganic liquid and a near infra-red laser for the etching of glass. The

technique is then demonstrated for the fabrication of functional micro-structures.


17

Chapter 3 shows the results of using laser interference lithography for the nano-

patterning of glass. It discusses the process flow from the patterning of the photoresist to the

final transfer of the patterns to the substrate by reactive ion etching. It also shows how the

patterned glass can be used as a mask for an one-step exposure to reproduce the nanoholes

array on photoresist. Defects formed on the mask are also duplicated on the photoresist

during photolithography. Simulation is also carried out to verify the experimental results

Chapter 4 shows the fabrication of micro-optics on glass substrates. Microlens array

and micro-solid immersion lens array are successfully fabricated on glass substrates. The

micro-solid immersion lens is shown to increase the resolution of a MLA.

Chapter 5 shows how direct writing with parallel processing capabilities can be

achieved through the use of microlens array lithography. Large area arbitrary patterns can be

directly written by moving the nanostage during the exposure. It demonstrates how the

technique can be used to fabricate phase shift structures on glass. The optical behavior and

characteristics of the phase shift structures are also presented. The results are further

supported by simulation results.

Chapter 6 summarizes the various glass processing techniques investigated in this

thesis. It shows that glass can be processed by either serial direct writing methods or parallel

patterning methods. Each method has its unique advantages and applications. Important

results and ideas are highlighted. Possible future works are also proposed.


18

1.7. Reference

1. E. Hecht, Optics, 4th

ed., (Addison Wesley, San Francisco, 2002).

2. H. Bach, and N. Neuroth, The properties of optical glass, (Springer-Verlag, Berlin,

1995).

3. C. F. Bohren and D. R. Huffman, Absorption and scattering of light by small

particles, (Wiley, New York, 1983).

4. M. J. Weber, Handbook of optical materials, (CRC Press, Boca Raton, 2003).

5. W. M. Steen, Laser material processing, 3rd

ed., (Springer, London, 2003).

6. J. Lorincz, “Waterjets: Evolving from macro to micro,” Manuf. Eng. 143, 47-53

(2009).

7. I. Steingoetter, A. Grosse, H. Fouckhardt, “Very deep fused silica etching”, SPIE

Proc., 4984 234-243 (2003).

8. D. Bäuerle, Laser processing and chemistry (Springer, New York, 2000).

9. J. Ihlemann, B. Wolff, and P. Simon, "Nanosecond and Femtosecond Excimer Laser

Ablation of Fused-Silica," Appl. Phys. A 54, 363-368 (1992).

10. J. Ihlemann, and B. WolffRottke, "Excimer laser micro machining of inorganic

dielectrics," Appl. Surf. Sci. 106, 282-286 (1996).

11. P. R. Herman, R. S. Marjoribanks, A. Oettl, K. Chen, I. Konovalov, and S. Ness,

"Laser shaping of photonic materials: deep-ultraviolet and ultrafast lasers," Appl.

Surf. Sci. 154-155, 577-586 (2000).

12. K. Sugioka, S. Wada, H. Tashiro, K. Toyoda, and A. Nakamura, "Novel Ablation

of Fused Quartz by Preirradiation of Vacuum-Ultraviolet Laser-Beams Followed


19

by 4th Harmonics Irradiation of Nd-Yag Laser," Appl. Phys. Lett. 65, 1510-1512

(1994).

13. H. Varel, D. Ashkenasi, A. Rosenfeld, M. Wahmer, and E. E. B. Campbell,

"Micromachining of quartz with ultrashort laser pulses," Appl. Phys. A 65, 367-373

(1997).

14. J. Zhang, K. Sugioka and K. Midorikawa, “Direct fabrication of microgratings in

fused quartz by laser-induced plasma-assisted ablation with a KrF excimer laser”,

Opt. Lett. 23, 1486–1488 (1998).

15. J. Zhang, K. Sugioka and K. Midorikawa, “High-speed machining of glass

materials by laser-induced plasma-assisted ablation using a 532-nm laser”, Appl.

Phys. A 67, 499–501 (1998).

16. J. Zhang, K. Sugioka and K. Midorikawa, “Laser-induced plasma-assisted ablation

of fused quartz using the fourth harmonic of a Nd+:YAG laser”, Appl. Phys. A 67,

545–549 (1998).

17. J. Zhang, K. Sugioka, and K. Midorikawa, “High-quality and high-efficiency

machining of glass materials by laser-induced plasma-assisted ablation using

conventional nanosecond UV, visible, and infrared lasers”, Appl. Phys. A 69,

S879–S882 (1999).

18. J. Wang, H. Niino and A. Yabe, “One-step microfabrication of fused silica by laser

ablation of an organic solution”, Appl. Phys. A 68, 111–113 (1999).

19. J. Wang, H. Niino and A. Yabe, “Micromachining of quartz crystal with excimer

lasers by laser-induced backside wet etching”, Appl. Phys. A 69, S271–S273

(1999).


20

20. J. Wang, H. Niino, and A. Yabe, “Micromachining by laser ablation of liquid:

Super-heated liquid and phase explosion”, SPIE Proc. 3933, 347–354 (2000).

21. J. Wang, H. Niino and A. Yabe, “Micromachining of transparent materials by laser

ablation of organic solution”, SPIE Proc. 4088, 64–69 (2000).

22. J. Wang, H. Niino, and A. Yabe, “Micromachining of transparent materials with

super-heated liquid generated by multiphotonic absorption of organic molecule”,

Appl. Surf. Sci. 154–155, 571–576 (2000).

23. Y. Yasui, H. Niino, Y. Kawaguchi and A. Yabe, “Microetching of fused silica by

laser ablation of organic solution with XeCl excimer laser”, Appl. Surf. Sci. 186,

552–555 (2002).

24. X. Ding, Y. Yasui, Y. Kawaguchi, H. Niino and A. Yabe, “Laser induced backside

wet etching of fused silica with an aqueous solution containing organic molecules”,

Appl. Phys. A 75, 437–440 (2002).

25. X. Ding, Y. Kawaguchi, H. Niino and A. Yabe, “Laser-induced high quality

etching of fused silica using a novel aqueous medium”, Appl. Phys. A 75, 641–646

(2002).

26. X. Ding, Y. Kawaguchi, H. Niino and A. Yabe, “Fabrication of 1 μm patterns on

fused silica plates by laser-lnduced backside wet etching (LIBWE)‟, SPIE

Proc.4830, 156–161 (2003).

27. H. Niino, Y. Yasui, X. Ding, A. Narazaki, T. Sato, Y. Kawaguchi and A. Yabe,

„Surface micro-fabrication of silica glass by laser-induced backside wet etching

with toluene solution‟, J. Photochem. Photobiol. A: Chem. 158, 179–182 (2003).


21

28. S. Y. Chou, P. R. Krauss and P. J. Renstrom, "Imprint Lithography with 25-

Nanometer Resolution," Science 272, 85–87 (1996).

29. H. W. Deckman and J. H. Dunsmuir, "Natural lithography," Appl. Phys. Lett. 41,

377-379 (1982).

30. J. C. Hulteen and R. P. van Duyne, "Nanosphere lithography: A materials general

fabrication process for periodic particle array surfaces," J. Vac. Sci. Technol. A 13,

1553-1558 (1995).

31. Y. N. Xia, and G. M. Whitesides, "Soft lithography," Annu Rev Mater Sci 28, 153-

184 (1998).

32. J. A. Rogers, and R. G. Nuzzo, "Recent progress in soft lithography," Mater Today

8, 50-56 (2005).

33. A. Kumar and G. M. Whitesides, “Features of gold having micrometer to

centimeter dimensions can be formed through a combination of stamping with an

elastomeric stamp and an alkanethiol „„ink‟‟ followed by chemical etching,” Appl.

Phys. Lett. 63, 2002–2004 (1993).

34. Y. Xia, E. Kim, X. M. Zhao, J. A. Rogers, M. Prentiss and G. M. Whitesides,

“Complex Optical Surfaces Formed by Replica Molding Against Elastomeric

Masters,” Science 273, 347–349 (1996).

35. X. M. Zhao, Y. Xia and G. M. Whitesides, “Fabrication of three-dimensional

micro-structures: Microtransfer molding,”Adv. Mater. 8, 837–840 (1996).

36. T. W. Ebbesen, H. J. Lezec, H. F. Ghaemi, T. Thio, and P. A. Wolff,

"Extraordinary optical transmission through sub-wavelength hole arrays," Nature

391, 667-669 (1998).


22

37. B. Gauvreau, A. Hassani, M. F. Fehri, A. Kabashin and M. Skorobogatiy,

“Photonic bandgap fiber-based Surface Plasmon Resonance sensors,” Opt.

Express 15, 11413 (2007).

38. S. H. Chang, S. K. Gray and G. C. Schatz, “Surface plasmon generation and light

transmission by isolated nanoholes and arrays of nanoholes in thin metal films,”

Opt. Express 13, 3150 (2005).

39. J. T. Bahns, F. N. Yan, D. L. Qiu, R. Wang and L. H. Chen, “Hole-Enhanced

Raman Scattering,” Appl. Spectrosc. 60, 989 (2006).

40. J. Homola, S. S. Yee and G. Gauglitz, “Surface plasmon resonance sensors:

review,” Sens. Actuators B 54, 3 (1999).

41. H. Y. Zhu, I. M. White, J. D. Suter, P. S. Dale and X. D. Fan, “Analysis of

biomolecule detection with optofluidic ring resonator sensors,” Opt. Express 15,

9139 (2007).

42. H. Y. Zhu, I. M. White, J. D. Suter, M. Zourob and X. D. Fan, “Integrated

Refractive Index Optical Ring Resonator Detector for Capillary Electrophoresis,”

Anal. Chem. 79, 930 (2007).

43. Y. F. Huang, S. Chattopadhyay, Y. J. Jen, C. Y. Peng, T. A. Liu, Y. K. Hsu, C. L.

Pan, H. C. Lo, C. H. Hsu, Y. H. Chang, C. S. Lee, K. H. Chen, and L. C. Chen,

“Improved broadband and quasi-omnidirectional anti-reflection properties with

biomimetic silicon nanostructures,” Nat. Nanotechnol. 2, 770 (2007).

44. J. Y. Huang, X. Wang, and Z. L. Wang, “Controlled Replication of Butterfly

Wings for Achieving Tunable Photonic Properties,” Nano Lett. 6, 2325 (2006).


23

45. H. Xu, N. Lu, D. Qi, J. Hao, L. Gao, B. Zhang, and L. Chi, “Biomimetic

Antireflective Si Nanopillar Arrays,” Small 4, 1972 (2008).

46. J. Y. Huang, X. D. Wang, and Z. L. Wang, “Bio-inspired fabrication of

antireflection nanostructures by replicating fly eyes,” Nanotechnology 19, 025602

(2008).

47. J. Q. Xi, M. F. Schubert, J. K. Kim, E. F. Schubert, M. Chen, S. Y. Lin, W. Liu,

and J. A. Smart, “Optical thin-film materials with low refractive index for

broadband elimination of Fresnel reflection,” Nat. Photonics 1, 176 (2007).

48. S. L. Diedenhofen, G. Vecchi, R. E. Algra, A. Hartsuiker, O. L. Muskens, G.

Immink, E. Bakkers, E. W. Vos, and J. E. Rivas, “Broad-band and

Omnidirectional Antireflection Coatings Based on Semiconductor Nanorods,”

Adv. Mater. 21, 973 (2009).

49. S. J. Wilson and M. C. Hutley, “The optical properties of „moth eye‟ antireflection

surfaces,” Opt. Acta. 29, 993–1009 (1982).

50. P. B. Clapham and M. C. Hutley, “Reduction of lens reflexion by the „moth eye‟

principle,” Nature 244, 281 (1973).

51. D. G. Stavenga, S. Foletti, G. Palasantzas, and K. Arikawa, “Light on the moth-

eye corneal nipple array of butterflies,” Proc. R. Soc. London, Ser. B 273, 661

(2006).

52. W. Chen, A. Y. Fadeev, M. C. Hsieh, D. Oner, J. Youngblood, and T. J.

McCarthy, “Ultrahydrophobic and Ultralyophobic Surfaces: Some Comments and

Examples,” Langmuir 15, 3395 (1999).


24

53. W. L. Min, B. Jiang, and P. Jiang, “Bioinspired Self-Cleaning Antireflection

Coatings,” Adv. Mater. 20, 3914 (2008).

54. W. L. Min, P. Jiang and B. Jiang, “Large-scale assembly of colloidal nanoparticles

and fabrication of periodic subwavelength structures,” Nanotechnology 19,

475604 (2008).

55. Y. R. Lin, H. P. Wang, C. A. Lin and J. H. He, “Surface profile-controlled close-

packed Si nanorod arrays for self-cleaning antireflection coatings,” J. Appl. Phys.

106, 114310 (2009).

56. S. G. Park, S. Y. Lee, S. G. Jang and S. M. Yang, “Perfectly hydrophobic surfaces

with patterned nanoneedles of controllable features,” Langmuir 26, 5295–5299

(2010).

Chapter 2: Laser direct writing of glass substrate with the assistance of liquid absorbers

25

CHAPTER 2

HIGH QUALITY MICRO-PROCESSING OF GLASS

SUBSTRATES BY LIBWE

2.1. Introduction

In this chapter, laser direct writing of glass substrates through the assistance of a

liquid absorber is investigated. The laser light irradiates and passes through the transparent

glass and interacts with the liquid absorber at the backside of the glass. This results in the

etching of the glass at the backside. Therefore, the process is called Laser Induced Backside

Wet Etching (LIBWE). Various parameters, such as the effect of the absorbing liquid and

the formation of by-products, are investigated.

2.2. Etching mechanism of LIBWE

LIBWE is a novel method of etching and micropatterning glass, discovered by Wang

and his coworkers [1]. This method is gaining prominence and many research works have

been done by this technique, such as investigation on its etching mechanisms [2-5],

application of different lasers [6-10], and using this technique for various types of

applications. This technique is a laser direct writing method. However, the major difference

is that laser energy is not absorbed by the glass. The energy is absorbed by a liquid absorber,

which is placed in contact with the backside of the glass. Laser light irradiates on the front

side of the glass, passes through the transparent glass and then is absorbed by the liquid

absorber at the glass-liquid interface. After absorbing the laser energy, the solution reaches a


26

high temperature through a cyclic multi-photon absorption mechanism [1, 5, 11-12]. The

absorbing molecules absorb the laser energy and are excited. Upon relaxation, heat is

released to the solution. Upon numerous cycles of photon absorption and release of heat

energy, the liquid is super-heated. This heats up the surface of the fused silica substrate,

eventually melting and softening the glass surface. At the same moment, the shockwave

[13-15] released upon the irradiation of the laser causes the removal of the molten glass

materials, resulting in the final etching of the glass. After the laser pulse, the irradiated area

cools immediately and this prevents further thermal damage to the glass surface. Therefore,

although the end result is the micro-patterning of glass, the LIBWE mechanism is different

from laser direct ablation. The main advantage of this method is that lasers which are

transparent to glass can be used to microprocess the glass by this method, as long as the

liquid can absorb the light. This method provides an alternative to expensive deep UV lasers

or ultrafast lasers processing. Conventional visible and UV lasers, such as diode pumped

solid state (DPSS) lasers at wavelengths of 355 nm and 532 nm, can be employed to

microprocess glass, which are transparent to these wavelengths.

In this chapter, research work done using this technique to microprocess glass is

presented. Research efforts are spent to understand this technique and apply it by a UV laser

to fabricate microstructures on glass substrates. This technique is then applied to another

liquid absorber, thus enabling a more cost-effective laser to be used to process glass.


27

2.3. Experimental procedure for LIBWE

Quartz substrates purchased from Photonik Singapore are used as the samples. The

liquid absorber used is toluene from Merck & Co, Inc (Singapore). Besides using pure

toluene, pyrene (Merck & Co) is also dissolved in toluene at different concentrations. The

pyrene is totally miscible in toluene. Figure 2.1 shows the experimental setup. An aluminum

chamber is custom-made to fit the size of the fused silica wafer so as to hold it firmly in

place. The liquid absorber is injected through the inlet until the chamber is filled up and the

liquid reaches the fused silica interface. Laser light from a Coherent diode pumped solid

state (DPSS) laser, at a wavelength of 355 nm is directed into a galvanometer, which is

capable of moving the light beam at high speeds along X and Y directions through the

rotation of its mirrors. The control of the galvanometer is connected to a PC pre-loaded with

the CAD software Mastercam, which is used to design the desired patterns. A

programmable file is then exported by Mastercam and imported into the software Scanlab to

control the galvanometer by computer numeric control (CNC). This allows various intricate

shapes to be designed and drawn out by the laser beam. After the laser processing, the fused

silica wafer is cleaned in acetone, isopropyl alcohol and then de-ionized water.

2.3.1. Type of laser used and raster scan by galvanometer

Different types of UV lasers have been used for LIBWE, such as UV excimer lasers

and solid state lasers. In this chapter, a UV DPSS laser, wavelength of 355 nm from the 3rd

harmonic generation of a fundamental wavelength of 1064 nm, is used in the study. Laser at

this wavelength has not been widely investigated for LIBWE as compared to other types of


28

lasers, such as excimers lasers at shorter wavelengths. Compared to the excimers lasers, it

has the advantages of not using toxic gases of fluorine or chlorine to operate.

The use of a high speed galvanometer enables rapid direct writing. In some research

work, a mask is used to generate the patterns on the glass surface by LIBWE [1, 5, 9, 13].

By using a galvanometer, patterns can be direct written without the need of a mask. This is

accomplished through the use of high speed raster scanning and the precise control of the

scanned paths through the use of CNC. Complicated shapes can be easily designed through

the use of CAD design software.

Fig. 2.1. The chamber for LIBWE experiment. The inlet is for filling up with the absorbing liquid.

Aluminum

chamber

Inlet


29

2.4. Results and discussion: Organic absorbing solution

2.4.1. Effect of absorbing liquid

In the original work [1], Wang and coworkers used acetone solution with pyrene

dissolved as the absorbing solution. Since then, various other liquids [13-14, 16-18] have

been investigated for their suitability for LIBWE. Toluene has been successfully used for

LIBWE by various groups [2, 14-15, 19]. Mixtures containing different kinds of aromatic

solvents and solutes have also been tested. Cheng et al. investigated how different solutions

and solids [6] affect the etching threshold of the material. All these research works show

that the absorbing liquid plays an important role on the etching process. Therefore, efforts

are taken to investigate the effect of the absorbing liquid on the etching process.

Pure toluene is compared against a solution of toluene and pyrene. Three different

concentrations of pyrene dissolved in toluene are prepared; 0.2 M, 0.4 M and 0.6 M. All

three solutions have higher etching rates than that using pure toluene alone, and it is found

that the 0.4 M solution gives the optimal etching rate [20]. At a higher concentration of 0.6

M, the etching rate (depth etched per unit time) decreases, suggesting a saturation of the

enhancement in the etching rate. Therefore, the 0.4 M concentration is chosen for

comparison with pure toluene.


30

The UV-Vis spectra of pure toluene and 0.4M of pyrene dissolved in toluene are

compared and shown in Fig. 2.2. The graph shows that toluene absorbs weakly in the near-

UV range to the visible range. The cut-off absorption is about 300 nm, beyond which there

is little absorption. This poses a challenge to the application of 355 nm light for LIBWE,

since the absorbing liquid has to absorb the incident laser light in order for LIBWE to take

place. This also suggests that although toluene has been used in other studies [2, 14-15, 19],

the usage of shorter wavelength UV light is crucial. In comparison, by dissolving pyrene in

toluene, the absorption spectrum of the solution extends to 500 nm. This also greatly

enhances the absorption of 355 nm light, which signifies that the absorption coefficient is

enhanced. This is beneficial because the etch rate increases with the absorption coefficient

[3]. Hence, by adding pyrene to toluene, two advantages are achieved: 1) to make the

solution more suitable for LIBWE at the 355 nm wavelength by extending the absorption

spectrum, and 2) to increase the etch rate by increasing the absorption of the light.


31

Fig. 2.2 The UV-Vis spectra of toluene and pyrene dissolved in toluene at a concentration of 0.4 M.

To validate the point that increasing the absorption enhances the etch rate,

experiments are carried out to compare the results of LIBWE on fused silica by the two

different solutions. Figure 2.3 plots the etched depth of the fused silica versus the laser

fluence when pure toluene and the pyrene/toluene solution are used. The laser repetition rate

is 10 kHz and the galvanometer scanning speed is 100 mm/s. The laser fluence is defined as

the laser pulse energy divided by the laser spot size, which is measured with a piece of

thermal paper.


32

Fig. 2.3 Comparison between the etched depths achieved by different liquids. The graph of

pyrene/toluene solution shows a lower etching threshold and a deeper etched depth, implying higher

etching rate.

There are two points to notice in the graph. Firstly, both liquid absorbers show a

power relationship between the laser fluence and the etched depth, as the curve fitting shows

a good fit for the data points. At the same laser fluence F and using the pyrene/toluene

mixture, a much deeper etched depth is achieved as compared to using toluene only. A laser

fluence of F = 1.5 J/cm2 attains an etched depth of 250 nm when using toluene alone. At the

same laser fluence, an etched depth of 1350 nm is achieved with the mixture of pyrene and

toluene. This is about five times increase in the etch rate, illustrating the point that the etch

rate is enhanced when the absorption is increased. The second point is that the threshold

fluence Fth needed to initiate the laser etching is lowered with the addition of pyrene. From

the extrapolation of the fitted curves in Fig. 2.3, the threshold fluence in toluene is about Fth


33

= 1.1 J/cm2 while the threshold fluence in the pyrene/toluene mixture is only Fth = 0.4 J/cm

2.

The lower threshold fluence is because the pyrene/toluene solution has a higher absorption

of 355 nm light than toluene alone. A liquid with a higher absorption coefficient has a

shorter light penetration depth, resulting in decreased heated liquid volume and higher

temperature [16]. The resulting higher temperature could start to melt the fused silica even

when the laser fluence is relatively low. This result shows that by using LIBWE to process

glass, much lower etching threshold fluence can be achieved, as compared to using lasers

for direct ablation of glass.

Figure 2.4 shows the various etched profiles obtained at a laser fluence of F = 0.90,

0.96, 1.18, 1.42, 1.52 and 1.70 J/cm2 in toluene alone. At F = 0.90 J/cm

2, there is almost no

modification to the surface but as the laser fluence is increased to F = 0.96 J/cm2 and then to

1.18 J/cm2, shallow but observable trenches are etched into the surface. The increase of laser

fluence from F = 1.18 to 1.42 J/cm2, from 1.42 to 1.52 J/cm

2, and from 1.52 to 1.70 J/cm

2

results in an increase of 20 nm, 120 nm and 184.6 nm in the etched depth, respectively.

With an increase of 0.1 J/cm2 in laser fluence from 1.42 to 1.52 J/cm

2, it leads to an increase

over 100 nm in the etched depth. In comparison, an increase of 0.24 J/cm2 from 1.18 J/cm

2

to 1.42 J/cm2 results in only 20 nm increase in the etched depth. This further illustrates the

non-linear relation between the etched depth and laser fluence.


34

Fig. 2.4. Various depth profiles obtained in a single laser scan at varying laser fluences.

2.4.2. By-product formation and removal by laser cleaning

It is well reported that the LIBWE process results in the formation of by-products

that are deposited on the etched surfaces [21-25]. The by-products have been identified as

carbon by-products. In the case here, it is observed that the by-products are also formed

during the LIBWE process and they are deposited on the etched surfaces. The by-products

adhere firmly to the surfaces, and cannot be removed by ultrasonic cleaning and cleaning

solvents. This affects the usability of the etched structures. To determine the nature of these


35

contaminants, XPS analyses are carried out to examine the chemical compositions of these

contaminants. A survey scan is done on the contaminated surface and the elements C, O and

Si are detected. The individual narrow scans are shown from Figs. 2.5(a) to 2.5(c).

278280282284286288290292294296298

500

1000

1500

2000

2500

3000

3500

4000

4500channel005.spe

Binding Energy (eV)

c/s

Pos. Sep. %Area284.65 0.00 53.32286.19 1.54 38.92288.94 4.28 7.76

Fig. 2.5(a). The binding energy of carbon at 284.65 eV, 286.19 eV and 288.94 eV, which correspond

to C-C / C-H, C-O and C=O bonds, respectively.

524526528530532534536538540542544546

0

2000

4000

6000

8000

10000

12000

14000

16000

channel005.spe

Binding Energy (eV)

c/s

Pos. Sep. %Area

531.07 0.00 5.24

532.51 1.44 58.92

533.17 2.10 35.84

Fig. 2.5(b). The binding energy of oxygen at 532.51 eV and 533.17eV, which correspond to O-Si /

O-C and O=C bonds, respectively.


36

949698100102104106108110112114116

0

500

1000

1500

2000

2500

3000

channel005.spe

Binding Energy (eV)

c/s

Pos. Sep. %Area 103.18 0.00 100.00

Fig. 2.5(c). The binding energy at 103.18 eV is due to Si-O bonds, from the quartz composition.

The carbon detected has three states and it is deduced1 that the carbon has a C-C / C-

H bond at a binding energy of 284.65 eV, a C-O bond at 286.19 eV and a C=O bond at

288.94 eV. O has three states. The binding energy of 531.07 eV could be due to the OH

group. The binding energy of 532.51 eV corresponds to the O-Si bond found in the quartz

and the O-C bond, and it is the O=C bond at 533.17 eV. Si has a binding energy of 103.18

eV and this corresponds to the Si-O bonds in quartz. From all these data, it can be

interpreted that the contamination is of carbon nature as carbon is the foreign element. The

carbon is most probably due to the decomposition of the absorbing liquid. Upon continual

irradiation of laser pulses, the absorbing liquid repeatedly goes through cycles of laser

absorption and heat release [11]. This causes the decomposition of the organic molecules,

producing the carbon by-products which are subsequently deposited onto the surface during

etching.

1 XPS peaks analysis sourced from www.lasurface.com


37

The carbon adheres stubbornly on the substrate surface and is difficult to be

removed. Laser cleaning is employed to remove the carbon contamination. The

contaminated surface is cleaned by overlapping laser pulses on the quartz surface in air

(without the absorbing liquid) by the same 355 nm laser source, but at a lower laser fluence

of 0.50 J/cm2. The carbon particles are held to the surface by adhesion forces, such as van

der Waals’ force or chemisorbed onto the surface by forming chemical bonding with the

surface. Upon the irradiation of laser cleaning pulses, the particles absorb laser energy and

cause thermal stresses in the particles [25]. The resulting stress could be large enough for

the physical displacement, and hence removal of the particles from the surface.

Furthermore, besides the above photo-thermal removal of the particles, photo-chemical

reaction could also occur, resulting in the ablation and elimination of the particles by the

laser irradiation, giving a clean and undamaged surface. Figure 2.6 shows the comparison of

a micro-fluidic channel surface with and without the laser cleaning. The cleanliness of the

quartz surface shows significant improvement after the laser cleaning, and demonstrates that

the laser cleaning process can be incorporated into the micro-fluidic channel fabrication to

achieve in-situ dry laser cleaning by the same laser source.


38

Fig. 2.6 Comparison between two substrate surfaces along the micro-fluidic channel with and

without laser cleaning.

2.4.3. Applications in micro-fluidic chips

In recent years, there is a rapidly growing interest in the technologies and market

trends of micro-fluidic chips. The attraction of such chips is small size and high throughput.

Microscopic channels have different properties from macroscopic counterparts, such as

laminar flow properties and fluidic resistances. There are various types of materials suitable

for the channel fabrication, with glass, silicon and polymer as the key substrate materials.

The stability of glass to chemical reactions and the transparency of glass for visual

observation of processes make it an attractive candidate to manufacture micro-fluidics.

Laser cleaned surface


39

In this research work, application of LIBWE in micro-fluidics is demonstrated.

LIBWE is used to etch micro-sized trench structures on quartz. The designs are based on the

idea of droplet manipulation and droplet fission [26], and groove structures for enhanced

mixing in micro-fluidic channels [27]. In the first case, as a droplet flows towards a

bifurcating junction, the fluidic flow towards the left and right channels creates a fluidic

pressure on the droplet. Based on the geometries of the spilt channels, the amount of fluid

flowing to the left and the right can either be similar or different. The fluidic pressure on the

droplet can thus be varied through different geometries of the channels. Droplets of smaller

sizes can be split off from the main droplet. The sizes of the droplets depend on the channel

geometries. For example, if both the left and right channels are of the same size, the amount

of fluid following into both channels would be similar. The pressure exerted on the droplet

is thus split evenly and two droplets of similar sizes can be created, with one going to the

right channel and the other to the left. This allows the control of the fission of the droplets.

In the second case, grooves are placed in the path of the fluidic flow to increase the mixing

of the fluids. Generally, liquid travelling through micro-fluidic channels are difficult to mix

and interact due to the laminar flow nature of microfluidic channels. This gives rise to the

categories of active and passive mixing. In the case of passive mixing, 3 broad categories of

mixing can be used [27]. One category is the use of static structures, such as grooves, inside

the micro-fluidic channels to introduce perturbations to the fluid flow, creating rotating

vortices and thereby increasing the amount of mixing. In these cases, LIBWE is used to

demonstrate its capabilities as an alternative and potential tool for microfluidics

applications. In the first case, LIBWE is used to direct write the desired microfludics


40

channels. In the second case, LIBWE is used to mill away the unwanted areas to leave

behind the grooves and protrusions on the microfluidics channels.

Figure 2.7(a) shows the 3D profiles of a microfluidic structure created by high speed

LIBWE processing with the galvanometer, at a repetition rate of 10 kHz and a fast scanning

speed of 100 mm/s. The surface profile is obtained using a surface profiler which probes

progressively in X and Y directions. The measured area is 2.00 mm by 1.80 mm. The cross-

sectional view horizontally across the surface is displayed and shows the Gaussian profiles

of the trenches. The Gaussian surface profile is due to the Gaussian beam profile of the

laser. In the magnified view of the respective four trenches, they have similar color-coded

profiles, showing that all of them are similar in the etched depth. Furthermore, the widths of

the four channels are 30 ± 1 μm. This shows good geometrical control of the etched

channels in achieving equally sized channels. The similarly sized channels would allow

similar volume of liquid to flow, hence splitting off similarly sized daughter droplets. The

left picture of Fig. 2.7(b) shows the 3D zoom-in view of the bifurcating channel junction.

The picture is tilted to show the bottom edges of the channels. The bottom is color-coded

blue throughout, showing the good uniformity of the etched depth. The 3D scan also shows

that the edges are straight, crack-free and of low distortion and roughness. The right picture

shows the optical microscope image of the fabricated micro-fluidics channel, showing the

crack-free and good quality surface. Figure 2.7(c) shows the types of grooves that can be

fabricated onto the quartz surface to enhance the mixing of reagents in microfluidics. They

are fabricated by using LIBWE to mill away the surrounding area to leave behind the

grooves and protrusions on the quartz surface.


41

Fig. 2.7(a) Profile scans of the microfluidics channels. Uniform cross-sections are obtained.

Fig. 2.7(b) Zoomed in 3D view of the bifurcating junctions. The depth is uniformly etched and the

edges are crack-free.


42

2.5. Results and discussion: Inorganic absorbing solution

2.5.1. Direct etching of glass using copper (II) sulphate solution

In the previous section, organic solvents are used for the direct-writing and etching

of glass. Toluene is one of the main solvents used and it is quite toxic and poses a potential

health hazard if it is used for a long period of time. Therefore, a much safer alternative was

considered. Copper (II) sulphate, CuSO4, is an inorganic metal salt that dissolves in water to

form copper (II) sulphate solution. Compared to toluene, it is of low toxic level, much less

volatile than toluene and hence does not give off hazardous fumes. To study its suitability as

an absorbing liquid for LIBWE, its absorption spectrum is studied. Figure 2.8 shows the

absorption spectrum of CuSO4 solution. There is a transmission band from 300 ~ 600 nm. It

absorbs strongly outside of this transmission band. This could also explain why CuSO4

solution is blue in colour. Since the CuSO4 solution absorbs strongly in the near infra-red

Fig. 2.7(c). Grooves fabricated on glass surface. The surrounding unwanted areas are milled

away using LIBWE.


43

(NIR) range, it could function as an absorbing liquid for lasers with wavelengths in the NIR

range.

Fig. 2.8. The absorption spectrum of copper (II) sulphate solution. There is transmission in the 300

nm ~ 600 nm range, and it absorbs strongly in the NIR range.

Copper salt solution has been used in various works [28-31] for the deposition of

copper by laser irradiation, and various mechanisms for the deposition of copper have been

proposed, including heat assisted reduction of copper due to photothermal effect [28-29,

31]. In the experiment, a Nd:YVO4 laser, which emits laser light at a wavelength of 1064

nm, a pulse width of 15 ns, a repetition rate range of 15 ~ 300 kHz and power output up to

20 W, is utilized for LIBWE using CuSO4 solution as the absorbing liquid. Glass is

transparent to 1064 nm light. Therefore, 1064 nm light cannot be used to process the glass


44

substrates. By absorbing the laser light, localized heating at the regions of laser irradiation

causes the deposition of copper onto the surface in contact with the solution [28-29, 31]. To

study whether there is indeed deposition of copper on the surface after the absorption of

laser light by the CuSO4 solution, an XPS analysis is carried out on the backside surface

after the laser irradiation. Prior to the XPS analysis, the samples are washed in an ultrasonic

bath for 30 minutes to dissolve and remove any copper sulphate crystals residue. Figures

2.9(a) and (b) plot the survey XPS scan and the narrow XPS scan of the surfaces after the

laser irradiation. In the survey scan, Cu peaks are picked up while S peaks are absent. This

shows that the Cu peaks are due to the presence of Cu deposition and are unlikely to be due

to residual CuSO4 crystals or solution. The narrow scan in Fig. 2.9b shows the binding

energies of the Cu electrons detected. The binding energy of the 2p1/2

electrons is at 932.61

eV and the 2p3/2

peak is separated from the 2p1/2

peak by 19.80 eV. This relates to the

element copper in the metallic state. From these analyses, it can be concluded that after the

CuSO4 solution is irradiated with the 1064 nm laser, copper is deposited on the surface of

the glass substrate in contact with the solution at the areas irradiated by the laser.


45

Fig. 2.9(a). Survey scan of the surface after the irradiation of 1064 nm laser light. The survey scan

shows the presence of Cu, C, O, and Si.

925930935940945950955960965970975

4000

5000

6000

7000

8000

9000

sample035.spe

Binding Energy (eV)

c/s

Pos. Sep. %Area932.61 0.00 72.73952.40 19.79 27.27

Fig. 2.9(b). Narrow scan of the Cu2p peak. The position of the Cu2p1/2

and Cu2p3/2

shows the

presence of metallic Cu.

Figure 2.10 shows the microscopic view of the glass surfaces processed by the 1064

nm laser irradiation. The glass is scanned in circular paths 10 times. There are materials

deposited along the line edges, which is attributed to copper according to the XPS analyses.

In addition, the shiny appearance on the top right corner of the path, with a tint of light

brown, is due to the reflection from the deposited copper. The scanned path has a width of ~


46

60 μm and in the middle of the path, etched craters can be observed. This illustrates that

there is interaction between the laser and the glass to cause the removal of the glass

materials. Under normal circumstances, the irradiation of the glass substrates with 1064 nm

laser light would not cause the removal of the glass materials. However, in this case, by

placing the glass substrates in contact with CuSO4 solution and irradiating the substrates

from the front side with 1064 nm laser light, copper deposition and glass etching are

observed on the backside surface. This clearly demonstrates that the CuSO4 solution plays

an important role in the removal of the glass material.

Fig. 2.10. Microscopic view of the glass surface in contact with CuSO4 solution after the laser

irradiation. There are materials deposited along the line edges.


47

2.5.2. Mechanism of the glass etching by laser irradiation of copper (II)

sulphate solution.

In order to explain the etching of glass by the irradiation of 1064 nm laser, a

mechanism based on the deposition of metallic copper after the irradiation of laser is

proposed. Figure 2.11 illustrates the proposed mechanism of glass etching by the laser

irradiation of 1064 nm light to the copper sulphate solution. During the first pulse, laser

light passes through the glass substrate, irradiates and heats up the CuSO4 liquid. During the

first few nanoseconds of the pulse, the temperature rises up by a few hundreds of degrees

Celsius. As the pulse ceases, the temperature starts to decrease in a millisecond time scale

(Fig. 2.11a). The initial high temperature causes the CuSO4 to decompose and initiate

copper deposition (Fig. 2.11b). Upon the arrival of the second pulse, one part of the laser

energy is absorbed by the CuSO4 liquid, causing the liquid to heat up again while the other

part of the energy is absorbed by the deposited copper. Upon continuous irradiation, the

copper absorbs more energy and heats up (Fig. 2.11c). At a high laser fluence, the deposited

copper can be boiled off, producing a recoil force [32] which then removes the molten glass

materials (Fig. 2.11d), and thus produces the etched craters as shown in Fig. 2.10. Hence, it

is a continuous two-step process of copper deposition and subsequent removal of the glass

with the assistance of the deposited copper film. This process has the advantage of a

continuous supply of copper from the copper sulphate solution, which, in principle, enables

the etching process to continue until the glass substrate is etched through.


48

Fig. 2.11. A schematic illustrating the process of glass material removal by continuous irradiation of

laser.

2.5.3. Dicing structures of various shapes from glass substrate

To demonstrate the etching capabilities, various structures are designed by the

Scanlab software and the exact shapes are scanned out by the laser beam. Employing this

capability, a range of intricate shapes and designs are drawn out in the Scanlab software and

used to carry out etching. Figure 2.12 shows various arbitrary structures that are diced out

from an 1 mm thick glass slide by using the CuSO4 solution as the absorbing liquid to etch

the glass substrate. The top and bottom left pictures show a star and a circle shaped

structures that have been diced out from the glass slides. It is observed that the glass slides

have remained intact and do not suffer cracking. The top right picture of Fig. 5 shows circles


49

at different diameters diced out from the glass slides. The circles (from left to right) are of

diameters 2 mm, 5 mm and 1.5 mm, respectively. The bottom right picture shows some of

the structures diced out from glass slides, the star and the alphabets ‘C’ and ‘G’. Other

delicate structures can also be processed and cut out by this technique.

Fig. 2.12. (Top and bottom left) Glass substrates with a star and a circle diced out. The glass

substrates have remained intact after the laser processing. (Top right) Circles of diameters 2 mm,

5mm, and 1.5 mm, from left to right, diced out from glass substrates. (Bottom right) Different

structures cut out from the glass substrate, showing the star shape and the alphabets ‘C’ and ‘G’.

In summary, the works presented in the current chapter focuses on using laser and a

suitable absorbing liquid to enable laser direct writing on glass, which is usually transparent

to lasers of those wavelengths. These direct writing methods can be used for various

applications, such as micro-fluidics channel for biomedical, and precision dicing of delicate

structures.


50

2.6. References

1. J. Wang, H. Niino, and A. Yabe, "One-step microfabrication of fused silica by

laser ablation of an organic solution," Appl. Phys. A 68, 111-113 (1999).

2. H. Niino, Y. Yasui, X. M. Ding, A. Narazaki, T. Sato, Y. Kawaguchi, and A.

Yabe, "Surface micro-fabrication of silica glass by excimer laser irradiation of

organic solvent," J. Photochem. Photobiol. A: Chem. 158, 179-182 (2003).

3. C. Vass, B. Hopp, T. Smausz, and F. Ignacz, "Experiments and numerical

calculations for the interpretation of the backside wet etching of fused silica," Thin

Solid Films 453-54, 121-126 (2004).

4. C. Vass, T. Smausz, and B. Hopp, "Wet etching of fused silica: a multiplex study,"

J. Phys. D: Appl. Phys. 37, 2449-2454 (2004).

5. J. Wang, H. Niino, and A. Yabe, "Micromachining of transparent materials with

super-heated liquid generated by multiphotonic absorption of organic molecule,"

Appl. Surf. Sci. 154, 571-576 (2000).

6. J. Y. Cheng, M. H. Yen, C. W. Wei, Y. C. Chuang, and T. H. Young, "Crack-free

direct-writing on glass using a low-power UV laser in the manufacture of a

microfluidic chip," J. Micromech. Microeng. 15, 1147-1156 (2005).

7. J. Y. Cheng, M. H. Yen, and T. H. Young, "Crack-free micromachining on glass

using an economic Q-switched 532 nm laser," J. Micromech. Microeng. 16, 2420-

2424 (2006).

8. C. Vass, D. Sebok, and B. Hopp, "Comparing study of subpicosecond and

nanosecond wet etching of fused silica," Appl. Surf. Sci. 252, 4768-4772 (2006).


51

9. J. Wang, H. Niino, and A. Yabe, "Micromachining of quartz crystal with excimer

lasers by laser-induced backside wet etching," Appl. Phys. A 69, S271-S273

(1999).

10. K. Zimmer, A. Braun, and R. Bohme, "Etching of fused silica and glass with

excimer laser at 351 nm," Appl. Surf. Sci. 208, 199-204 (2003).

11. H. Fukumura, and H. Masuhara, "The Mechanism of Dopant-Induced Laser-

Ablation - Possibility of Cyclic Multiphotonic Absorption in Excited-States,"

Chem. Phys. Lett. 221, 373-378 (1994).

12. H. Fukumura, N. Mibuka, S. Eura, and H. Masuhara, "Mass spectrometric studies

on laser ablation of polystyrene sensitized with anthracene," J. Phys. Chem. 97,

13761-13766 (1993).

13. X. Ding, Y. Kawaguchi, H. Niino, and A. Yabe, "Laser-induced high-quality

etching of fused silica using a novel aqueous medium," Appl. Phys. A 75, 641-645

(2002).

14. X. M. Ding, Y. Kawaguchi, T. Sato, A. Narazaki, R. Kurosaki, and H. Niino,

"Micron- and submicron-sized surface patterning of silica glass by LIBWE

method," J. Photochem. Photobiol. A: Chem. 166, 129-133 (2004).

15. Y. Kawaguchi, X. Ding, A. Narazaki, T. Sato, and H. Niino, "Transient pressure

induced by laser ablation of liquid toluene: toward the understanding of laser-

induced backside wet etching," Appl. Phys. A 79, 883-885 (2004).

16. R. Bohme, A. Braun, and K. Zimmer, "Backside etching of UV-transparent

materials at the interface to liquids," Appl. Surf. Sci. 186, 276-281 (2002).


52

17. J. Wang, H. Niino, and A. Yabe, "Microfabrication of a fluoropolymer film using

conventional XeCl excimer laser by laser-induced backside wet etching," Jpn. J.

Appl. Phys. 38, L761-L763 (1999).

18. K. Zimmer, R. Bohme, A. Braun, B. Rauschenbach, and F. Bigl, "Excimer laser-

induced etching of sub-micron surface relief gratings in fused silica using phase

grating projection," Appl. Phys. A 74, 453-456 (2002).

19. Y. Kawaguchi, X. Ding, A. Narazaki, T. Sato, and H. Niino, "Transient pressure

induced by laser ablation of toluene, a highly laser-absorbing liquid," Appl. Phys.

A 80, 275-281 (2005).

20. Z. Q. Huang, "Laser Etching of Glass using Absorbing Liquids," (National

University of Singapore, Singapore, 2006). Refer to appendix for relvant

information.

21. R. Bohme, and K. Zimmer, "Effects of halogenated organic solvents on laser-

induced backside wet etching of fused silica," Appl. Phys. A 83, 9-12 (2006).

22. H. Niino, Y. Yasui, X. Ding, A. Narazaki, T. Sato, Y. Kawaguchi, and A. Yabe,

"Surface micro-fabrication of silica glass by excimer laser irradiation of organic

solvent," J. Photochem.Photobiol. A: Chem. 158, 179-182 (2003).

23. R. Böhme, D. Spemann, and K. Zimmer, "Surface characterization of backside-

etched transparent dielectrics," Thin Solid Films 453-454, 127-132 (2004).

24. R. Böhme, and K. Zimmer, "The influence of the laser spot size and the pulse

number on laser-induced backside wet etching," Appl. Surf. Sci. 247, 256-261

(2005).


53

25. Y. F. Lu, W. D. Song, B. W. Ang, M. H. Hong, D. S. H. Chan, and T. S. Low, "A

theoretical model for laser removal of particles from solid surfaces," Appl. Phys.

A 65, 9-13 (1997).

26. Y. C. Tan, J. S. Fisher, A. I. Lee, V. Cristini, and A. P. Lee, "Design of

microfluidic channel geometries for the control of droplet volume, chemical

concentration, and sorting," Lab Chip 4, 292-298 (2004).

27. P. B. Howell, D. R. Mott, S. Fertig, C. R. Kaplan, J. P. Golden, E. S. Oran, and F.

S. Ligler, "A microfluidic mixer with grooves placed on the top and bottom of the

channel," Lab Chip 5, 524-530 (2005).

28. L. Nanai, I. Hevesi, F. V. Bunkin, B. S. Lukyanchuk, M. R. Brook, G. A. Shafeev,

D. A. Jelski, Z. C. Wu, and T. F. George, "Laser-Induced Metal-Deposition on

Semiconductors from Liquid Electrolytes," Appl. Phys. Lett. 54, 736-738 (1989).

29. R. J. V. Gutfeld, E. E. Tynan, R. L. Melcher, and S. E. Blum, "Laser Enhanced

Electroplating and Maskless Pattern Generation," Appl. Phys. Lett. 35, 651-653

(1979).

30. L. Mini, C. Giaconia, and C. Arnone, "Copper Patterning on Dielectrics by Laser

Writing in Liquid Solution," Appl. Phys. Lett. 64, 3404-3406 (1994).

31. J. C. Puippe, R. E. Acosta, and R. J. Vongutfeld, "Investigation of Laser-Enhanced

Electroplating Mechanisms," J. Electrochem. Soc. 128, 2539-2545 (1981).

32. B. Hopp, T. Smausz, and M. Bereznai, "Processing of transparent materials using

visible nanosecond laser pulses," Appl. Phys. A 87, 77-79 (2007).

Chapter 3: Fabrication of nanoholes array on glass by laser interference lithography

54

CHAPTER 3

FABRICATION OF NANOHOLES ARRAY ON GLASS

BY LASER INTERFERENCE LITHOGRAPHY

3.1. Introduction

With the advance of nanotechnology, micro- and nano-structures are finding

increasing applications in different fields. Nanohole array is such an example, with

applications ranging from microelectronics [1, 2], periodic holes for photonic crystal

applications [3, 4], sub-wavelength holes for plasmonics studies and applications [5, 6], to

chemical and biomedical analyses [7, 8]. There are a variety of methods to fabricate the

nanohole arrays, with serial methods, such as electron-beam lithography and focused ion

beam lithography, and parallel methods, such as photolithography, soft lithography [9, 10]

and colloidal lithography [11]. Another parallel and maskless technique is laser interference

lithography (LIL), which has attractive advantages of: 1) cost effectiveness; 2) maskless

technique; 3) large area parallel patterning; and 4) flexible adjustment of the period and

feature sizes of the nano-patterns. In this chapter, LIL is applied to process quartz substrates.

Nanohole arrays are fabricated on the quartz substrates. They are subsequently used as a

phase mask for patterning large area nanohole arrays by photolithography. Upon irradiation

of UV light, a large area of nanohole array is exposed and produced on photoresist.

When two waves overlap, the superposition of the waves occurs [12]. The

superposition of the waves causes constructive interference when the two waves are in

phase, and destructive interference when the two waves are out of phase. This principle of


55

superposition applies to light as well because light is an electromagnetic wave. When the

light sources are coherent, the phenomenon of interference generates a standing wave

pattern, producing a stable pattern of alternating bright and dark fringes. Laser light is an

excellent coherent light source because it is monochromatic. The LIL technique [13-14] is

based on this interference phenomenon. The laser beams interfere to produce the alternating

bright and dark fringes. A light sensitive material, such as photoresist, can be used to record

this pattern. The pattern then forms on the photoresist after development. The LIL technique

is a maskless technique. It is a superior method to fabricate periodic patterns over a large

area by a single exposure [15-16]. Figure 3.1 shows a schematic illustration of the formation

of fringes by the interference of two light beams that forms a standing wave, with

alternating bright and dark fringes.

Fig. 3.1. Schematic illustration of a standing wave formed by the interference of two light beams.


56

The period p of the fringes depends on the angle of interference of the light beams θ

and the wavelength of the light λ, as governed by Eq. (3.1):

sin2p (3.1)

When this standing wave pattern is used to expose photoresist, a series of gratings is

achieved.

3.2. Experimental

Quartz is used as the substrate in the experiment to explore how LIL can be used to

structure it with nanoholes array. Quartz is highly transparent even in the deep UV region

[17]. This makes it an excellent candidate for UV light investigations and applications.

Furthermore, it has high resistance to thermal shock and mechanical stress, making it more

durable when subjected to various other processing steps.

3.2.1. Process flow

Quartz substrates (from Photonik Singapore) are cleaned in solvents, such as

acetone, IPA and de-ionized water, to remove contamination prior to usage. The quartz

substrates are then placed into an e-beam evaporation system for coating thin films on the

substrates. Figure 3.2 illustrates the process steps. A chromium thin film is firstly coated on

the quartz at a thickness of ~ 50 nm. Subsequently, a 10 nm thin layer of SiO2 is coated on

the top of the Cr thin film. After taking the substrates out from the chamber, negative

photoresist (Ma-N 1407 from Microchem) is coated onto the quartz substrate by spin


57

coating at a rotational speed of 5000 ~ 6000 rpm to obtain a film thickness of 400 ~ 500

nm. This is then followed by a post bake at 110 ˚C for 1 minute. LIL exposure is then

carried out on the photoresist. The patterns are obtained after developing the photoresist.

Reactive ion etching (RIE) is then used to etch the thin layer of SiO2 which is not covered

by the photoresist. The etching gas used is CF4 at a flow rate of 50 sccm. The chamber

pressure is 20 mTorr and the power is set to 250 W. Using these parameters, the etch rate of

SiO2 is found to be ~ 1 nm/s. Therefore, ~ 10 seconds is needed to etch through the exposed

areas of the SiO2 layer. Chromium etchant is used to wet etch the areas not being protected

by the SiO2 layer. The Cr etchant contains NaOH, which can readily dissolve photoresist.

This explains why a SiO2 capping layer is needed to define the areas to be etched by the Cr

etchant, instead of using the photoresist patterns obtained in step 3. After etching the Cr

layer, the remaining Cr layer is used as a hard mask to etch the underlaying quartz substrate.

The Cr hard mask is needed because: 1) it is used to enable deep anisotropic etching by RIE;

and 2) the photoresist patterns obtained in step 3 is not ideal to be used as the etching mask

for RIE because RIE is an anisotropic process. The surface profile of the photoresist patterns

obtained is sinusoidal. If it is used for etching, the final etched surface profile of the quartz

would be in a sinusoidal profile as well. This is undesirable for the phase mask fabrication

and the details will be explained in the next section. After the etching of the quartz substrate

by RIE, Cr etchant is used again to remove the remaining Cr layer.


58

Fig. 3.2. Process flow for patterning the quartz surface.

3.2.2. LIL setup and exposure

There are different LIL setups, with their advantages and disadvantages. The Lloyd’s

mirror setup is applied for LIL in the experiment. Figure 3.3 shows the schematic drawing

of the Lloyd’s mirror setup for LIL. A laser beam is directed to a spatial filter by a set of

Quartz

Cr

E-beam evaporation of Cr 1.

Quartz

Cr

E-beam evaporation of SiO2 2.

SiO2

Cr

3.

SiO2

Photoresist

Quartz

4.

SiO2

RIE

Cr

Quartz

5.

SiO2

Cr

Chrome Etch

Quartz

6.

SiO2

Cr

RIE

LIL exposure

Quartz


59

mirrors. The spatial filter consists of an objective lens and a pinhole with a diameter slightly

larger than the focused laser beam size. The function of the spatial filter is to filter out the

high frequency noise [18] so that a uniform Gaussian beam profile can be obtained.

Meanwhile, it also acts as a beam expander whereby after passing through the pinhole, the

beam diameter expands, yielding a more uniform beam intensity at the central area.

Furthermore, the expansion of the laser beam diameter by the spatial filter also gives a

closely approximated plane wave over the exposure area.

Fig. 3.3. Schematic drawing of a Lloyd’s mirror setup for laser interference lithography of periodic

structures on photoresist.

After the expanded beam travels a distance of ~ 1 m, it impinges onto a rotating

stage with a mirror mounted perpendicularly to the sample holder. A mirror surface is held

at 90˚ to the stage. The substrate for exposure is mounted onto the stage by a suction force

from a vacuum pump. Upon reaching the stage, one half of the beam reflects off the mirror

and interferes with the other half of the beam, creating the interference patterns on the

sample placed on the stage. As the mirror is always perpendicular to the stage, the reflected


60

beam arrives at the substrate at the same angle α as the incident beam, as illustrated in Fig.

3.4. Hence, the angle of interference θ (ref. Fig. 3.1) is the same as the angle of rotation of

the stage. According to Eq. 3.1, the period of the gratings can be adjusted by varying θ,

which is then achieved by rotating the stage by the same angle θ. This offers a simple and

flexible way to vary the period of the gratings.

Fig. 3.4. Illustration of the reflection angle of the reflected ray and the angle of interference.

The laser light source used is a Helium Cadmium (He-Cd) continuous wave (CW)

laser (Kimmon, Japan), at a wavelength of λ = 325 nm and a coherent length of 10 cm,

making it suitable for interference lithography. The substrate is exposed by LIL at an angle

θ ≈ 15°. This gives a grating series with a period of ~ 630 nm during the first exposure of

the negative photoresist. After the first exposure, the substrate is rotated 90° and the second

exposure is carried out. Figures 3.5(a) and (b) show the patterns obtained on the negative

photoresist after single and double exposures, respectively. Single exposure results in a

grating while double exposure gives a nanoholes array.

Stage roatation

angle θ


61

Fig. 3.5(a). Line array fabricated by a single exposure on negative photoresist.

Fig. 3.5(b). Nanohole array obtained after a LIL double exposure on negative photoresist.

3.2.3. Transfer of patterns from photoresist to Cr layer

After producing the patterns on the negative photoresist, the patterns are transferred

to the SiO2 capping layer by RIE, using CF4 as the etching gas. This step also exposes the Cr

layer beneath. Figure 3.6 shows the photoresist pattern after the etching of the SiO2 capping


62

layer, while the inset shows the cross-sectional profile of the patterns (taken across the white

dotted line). The cross-sectional profile shows that the photoresist is in a sinusoidal profile.

The height of the photoresist is ~ 135 nm. But it is too thin for deep etching, which explains

why a Cr hard mask is needed for deeper etching. The bottom of the nanoholes is flat, which

shows that the bottom of the photoresist is exposed through to the SiO2 layer. This is

important as it means that the RIE could uniformly etch the SiO2 capping layer and expose

the Cr layer.

Fig. 3.6. AFM profile of the photoresist pattern after a double LIL exposure and RIE to open the

SiO2 capping layer.

After etching the SiO2 capping layer, the Cr etching solution is used to etch the

exposed areas of Cr which are not covered by the SiO2 capping layer. Figure 3.7 shows the

Cr layer etched, with the nanoholes array patterns being transferred from the photoresist

layer to the Cr hard mask layer.

135 nm


63

Fig. 3.7. Nanoholes array patterns transferred to the Cr hard mask layer.

3.2.4. Transfer of nanoholes array patterns from Cr layer to quartz

substrate

After obtaining the nanoholes array on the Cr hard mask, the patterns are transferred

to the quartz substrate by RIE, illustrated as step 6 of the the process flow shown in Fig. 3.4.

Figure 3.8 shows an AFM image of the nanohole array being successfully transferred to the

quartz substrate. Due to the anisotropic etching nature of the RIE process, the nanoholes are

transferred to the quartz substrate with good fidelity. They are well defined and the

sidewalls of the nanholes are vertical.


64

Fig. 3.8. The processed quartz substrate with the nanoholes array etched transferred to the substrate.

The use of the Cr hard mask is crucial to achieve a vertical sidewall profile. The

photoresist patterns formed by LIL is in a sinusoidal profile, and the etching selectivity of

the photoresist to the quartz is ~ 1:1. As RIE is an anisotropic etching process, if the

photoresist is used as the etching mask, the etched profile of the processed substrate follows

the sinusoidal profile of the photoresist (inset of Fig. 3.6). Hence, if vertical sidewall

profiles are desired, the photoresist patterns formed by LIL are not ideal to be used as a

mask for RIE. The Cr hard mask has a good etching selectivity against the quartz substrate.

The Cr can withstand a much longer etching time, and this helps to achieve a more vertical

sidewall profile. Therefore, LIL is used to define the nanohole patterns on photoresist, and

chrome etchant is used to etch the Cr layer, effectively transferring the LIL patterns from the

photoresist onto the Cr hard mask. The Cr hard mask then defines the patterns on the quartz

substrate for the final RIE step to fabricate the nanoholes array on the quartz substrate.


65

3.3. Applications

3.3.1. Nanoholes array as a phase mask

After fabricating the nanoholes array on the quartz substrate, phase mask

applications is investigated. The surface of the quartz is periodically etched with circular

nanoholes and this causes the refractive index of the etched region to change from n = 1.5

(refractive index of quartz) to n = 1 (refractive index of air). When light passes through the

surface, it experiences the modulation of its phase front due to these periodic refractive

index changes, as illustrated in Fig. 3.9. As light travels at different speeds in different

media, there is a phase delay after light emerges from the surface of the periodic nanoholes

array. The phase delay between light travelling through the etched and unetched surfaces

depends on the refractive index of the materials and the height of the holes, according to:

360

hnhnairquartz

(3.2)

where n is the refractive index, h the height of the nanoholes and φ the phase difference.


66

Fig. 3.9. A schematic illustration of the phase delay introduced to the wave front after passing

through the surface with the nanoholes. This results in a modulation of the wave front.

If the phase delay between the waves is 180˚, destructive interference occurs. This

interference of the waves between the alternating nanoholes causes the re-distribution of the

light energy after it emerges from the surface. To investigate this effect, theoretical

simulation is carried out to study the distribution of light as it passes through the quartz with

the fabricated nanoholes array. The simulation is done using Computer Simulation

Technology (CST) software, which uses the Finite Integral Technique. In the simulation, the

wavelength of the light is chosen as 365 nm because this UV light would be used later for

the exposure of photoresist (using the nanholes array as the phase mask). The diameter of

the holes d is set to 330 nm and the period p 660 nm. The height of the nanoholes h is

calculated using Eq. 3.2. With λ = 365 nm, φ = 180°, nquartz = 1.475 and nair = 1, h is

calculated to be ~ 385 nm. These parameters reflect the physical parameters of the

processed quartz substrate. Figure 3.10 plots the calculation result of the light intensity in

Incident Wave

n=1.5

n=1


67

the XY plane at Z = 2 µm from the surface of the quartz as light passes through it. The inset

at the bottom right corner of Fig. 3.10 represents the structure used for the simulation. The

plot shows that there is a periodic distribution of regions with an enhanced light intensity,

equally spaced out at a period of 660 nm. These regions occur at the nodes among four

neighboring etched nanoholes. The ratio of the enhanced light intensity at these regions with

respect to the background light is ≥ 6 times. This increase in light intensity occurs due to the

modulation and re-distribution of the incident light energy. Furthermore, the enhanced light

intensity is in a Gaussian shape, with the peak intensity confined in a very small area, which

can be estimated as ~ 100 nm (FWHM). In contrast, the background light irradiating

through the quartz is a minimum at the other transparent areas. This simulation result shows

that the processed quartz substrate can be used as a phase mask to pattern periodic

nanoholes by making use of the periodic arrangement of the enhanced light intensities.


68

Fig. 3.10. Simulation result of the light distribution after light passes through the surface with the

nanoholes array. The modulation of the phase of the light causes the redistribution of the light

energy, resulting in a periodic array of enhanced light intensities.

The processed quartz substrate with the nanoholes is subsequently used as a phase

mask for the patterning of nanoholes. 365 nm light (mercury arc lamp, 8 mW) is irradiated

through the mask to expose the positive photoresist coated on a silicon wafer placed in

contact with the phase mask. After exposing and developing the photoresist, an AFM scan is

carried out on the exposed photoresist and the results are shown in Fig. 3.11. The bottom

right inset shows the exposure method. Periodic nanoholes array are produced on the

photoresist uniformly over a scanned area of 10 µm × 10 µm. The right inset cross-section

shows that the hole sizes are ~ 300 nm (at the bottom) and the depth is 270 nm. The period

of the holes is ~ 650 nm, which agrees well with the periodic arrangement of enhanced light

intensities of the phase mask. The nanoholes are of Gaussian shapes, concurring with the

Gaussian profile of the enhanced light intensity. Therefore, by using LIL and RIE to pattern


69

a quartz substrate with a nanohole array, a phase mask can be fabricated. The phase mask

can be used for patterning of nanoholes array by 365 nm light exposure. In a single

exposure, a large area of periodic nanoholes can be obtained.

Fig. 3.11. Periodic array of nanoholes patterned uniformly on the photoresist.. The top right inset

shows the profile of the nanoholes formed. The bottom inset shows the exposure method.

3.3.2. Potential applications of the patterned nanholes array

Using the processed quartz phase mask for patterning, nanoholes array can be

created flexibly and with high speed. It takes ~ 10 minutes to fabricate a 500 μm × 500 μm

nanoholes array using the phase mask. In contrast, it would take a long time for e-beam

lithography or focused ion beam lithography to create such an array of nanoholes. For

example, it would take ~ 1 hour to fabricate a 50 μm × 50 μm array of nanoholes by e-beam

lithography. However, e-beam lithography has the advantages of much higher precision and

365 nm light

Phase mask

Photoresist


70

better uniformity control of the sizes and shapes of the nanoholes, which is crucial in some

applications.

The fabricated periodic nanoholes can be used in various potential applications. An

example is to use the uniform nanoholes array to improve the external quantum efficiency of

LEDs [19-21]. As it is difficult to extract light from a LED due to internal reflection, one

way to enhance the external quantum efficiency is by using photonic crystals to couple light

out. However, for viable production, e-beam lithography and focused ion beam lithography

are too slow to create the large arrays of nanoholes. Using the processed quartz as a phase

mask for photolithography offers a rapid and cost-effective alternative to fabricate the 2D

photonic crystals over a large area. The patterns can subsequently be transferred onto the

wafer by RIE to fabricate a 2D photonic crystal slab. Figure 3.12 shows the nanoholes array

transferred from photoresist to silicon with high fidelity by RIE. Uniform and well-patterned

nanoholes array are produced on the silicon substrate. This result demonstrates that a 2D

photonic crystal slab can be fabricated flexibly by this approach.


71

Fig. 3.12. The nanoholes array patterned on photoresist transferred to the silicon substrate with good

fidelity and uniformity.

Meanwhile, defects on the nanoholes array on the processed quartz phase mask are

correspondingly patterned onto the photoresist during the exposure. The defects are some

defective nanoholes on the quartz substrate, which scatter and re-distribute the incident

light, causing a drop in light intensity incident over the local regions. The decrease in the

light intensity is unable to expose the photoresist, resulting in a corresponding defect in the

nanoholes array patterned on the photoresist. To demonstrate this concept, a quartz phase

mask with a defect on the nanoholes array is used for exposure. Figure 3.13 shows an AFM

image of a patterned nanoholes array on positive photoresist. There is a missing hole in the

periodic nanohole array over an area of ~ 5 m × 5 m. The inset shows a microscope

image of the corresponding defect pattern on the quartz phase mask. Such patterning of


72

defects has interesting and useful applications, for example, to create 1D localized defects in

photonic crystals.

Fig. 3.13. A defect formed on photoresist. The right inset shows the corresponding defect on the

quartz phase mask.

To verify this result, simulation is carried out. A circle is used to represent the

defect. The defect is located at the node among four neighbouring nanoholes. Figure 3.14

shows the simulation result. The presence of the defect redistributes the light, causing the

region to experience lower light intensity than the photoresist exposure threshold. The

decreased light intensity is not strong enough to expose the photoresist, and therefore forms

the defect.


73

Fig. 3.14. Simulation result of a defect among the nanohole array. The defect reduces the light

intensity of that region.

Besides forming a corresponding single defect, a 2D line defect could also be

introduced on the phase mask. Then, after exposure, the 2D line defects are reproduced on

the photoresist. The defects on the phase mask can be accomplished by using FIB to create

the defects on the phase mask. Figure 3.15 shows a SEM image of the line defect created

onto the phase mask, while the inset shows the close-up image of the defects. The materials

surrounding the holes are the remaining metal sputtered onto the mask to aid in the

conduction of the focused ions. The defects are created at the nodes with 4 adjacent

nanoholes, at the point where the strong light intensities occur (ref. Fig. 3.10). In this way,

the enhanced light intensities at these nodes are re-distributed by the defects, causing a drop

in the light intensities and hence leaving the corresponding defects on the photoresist to

form 2D defects.

Low light intensity

Defect


74

Fig. 3.15. SEM image of the line of defects created on the phase mask by FIB.

Besides the 2D patterning of nanoholes array, the processed quartz with the nanohole array

has a 3D distribution of light intensities as light passes through it. Figure 3.16 shows the

simulation results of the light intensity variation at various Z planes from the surface of the

quartz. As the Z distance increases, the region of strong light intensity shifts and changes.

This phenomenon is due to the Talbot effect, where the image of the diffraction grating is

repeated at regular distances away from the grating plane. The Talbot length ZT is given by:

22 pZT (3.3)

where p is the period of the grating and λ the wavelength. Furthermore, at a half of the

Talbot length, an image of the grating also occurs, but phase-shifted by a half of a period,


75

which means that the image is laterally shifted by a half of the width of the grating period.

At smaller regular fractions of the Talbot length, sub-images are also observed. At one

quarter of the Talbot length, the image is halved in size, and appears with a half of the

period of the grating. This increases the images observed by 2 times. At one eighth of the

Talbot length, the period and size of the images are halved again, and so forth. This creates a

fractal pattern of sub-images, often referred to as Talbot carpet. The nanoholes array

fabricated on the quartz substrate acts like a grating. Therefore, upon light irradiation, the

nanohole array phase mask displays a 3D distribution of periodic light intensity that changes

at different Z distances, repeating after a Talbot length. This would create alternating

periodic patterns within the volume of a photosensitive polymer, constructing complicated

repeating 3D patterns from the Talbot effect. An attractive application of this 3D patterning

is to use it for 3D photonic crystals structuring [22-24]. 3D photonic crystals are ideal to

create complete photonic bandgaps in 3 dimensions. Using the nanoholes array phase mask

could offer a cost effective method to fabricate the 3D photonic crystals.


76

Fig. 3.16. Light intensity distribution at various Z planes. This 3D distribution of light has interesting

applications in 3D photonic crystals patterning.

The processed quartz can have other extensive applications, such as the fabrication

of nanoholes array for the nucleation of catalytic nanoparticles in a periodic and orderly

manner to grow nanowires [1-2], and for the formation of 2D hole arrays on metallic films

to structure metallic lens [25].


77

3.4. Reference

1. W. K. Choi, T. H. Liew, H. G. Chew, F. Zheng, C. V. Thompson, Y. Wang, M. H.

Hong, X. D. Wang, L. Li, and J. Yun, "A combined top-down and bottom-up

approach for precise placement of metal nanoparticles on silicon," Small 4, 330-

333 (2008).

2. J. Yun, R. Wang, W. K. Choi, J. T. L. Thong, C. V. Thompson, M. Zhu, Y. L.

Foo, and M. H. Hong, "Field emission from a large area of vertically-aligned

carbon nanofibers with nanoscale tips and controlled spatial geometry," Carbon

48, 1362-1368 (2010).

3. S. John, "Strong Localization of Photons in Certain Disordered Dielectric

Superlattices," Phys Rev Lett 58, 2486-2489 (1987).

4. E. Yablonovitch, "Inhibited Spontaneous Emission in Solid-State Physics and

Electronics," Phys Rev Lett 58, 2059-2062 (1987).

5. T. W. Ebbesen, H. J. Lezec, H. F. Ghaemi, T. Thio, and P. A. Wolff,

"Extraordinary optical transmission through sub-wavelength hole arrays," Nature

391, 667-669 (1998).

6. J. A. Porto, F. J. Garcia-Vidal, and J. B. Pendry, "Transmission resonances on

metallic gratings with very narrow slits," Phys Rev Lett 83, 2845-2848 (1999).

7. T. F. Kosar, C. Chen, N. L. Stucky, and A. Folch, "Arrays of Microfluidically-

Addressable Nanoholes," J Biomed Nanotechnol 1, 161-167 (2005).


78

8. A. De Leebeeck, L. K. S. Kumar, V. de Lange, D. Sinton, R. Gordon, and A. G.

Brolo, "On-chip surface-based detection with nanohole arrays," Anal Chem 79,

4094-4100 (2007).

9. Y. N. Xia, and G. M. Whitesides, "Soft lithography," Annu Rev Mater Sci 28,

153-184 (1998).

10. J. A. Rogers, and R. G. Nuzzo, "Recent progress in soft lithography," Mater Today

8, 50-56 (2005).

11. P. Hanarp, D. S. Sutherland, J. Gold, and B. Kasemo, "Control of nanoparticle

film structure for colloidal lithography," Colloids and Surfaces A:

Physicochemical and Engineering Aspects 214, 23-36 (2003).

12. M. Born and E. Wolf, Principles of optics, 7 ed., (Cambridge University Press,

UK, 1999).

13. H.I. Smith, Submicron- and Nanometer-Structures Technology, MA:

NanoStructures Press Sudbury, 1994.

14. Walsh M E , Nanostructuring magnetic thin films using interference lithography,

M.S. Thesis, Massachusetts Institute of Technology, 2000.

15. Yu F, Li P, Shen H, Mathur S, Lehn C L, Bakowsky U and Mücklich F, Laser

interference lithography as a new and efficient technique for micropatterning of

biopolymer surface, Biomaterials, 26, pp.2307-2312. 2005.

16. S.I. Nesterov, D.V. Myagjov and E.L. Portnoi, Nanoscale periodical structures

fabricated by interference photolithography, Int. J. Nanoscience, 3, pp.59-64.

2004.


79

17. H. Bach, and N. Neuroth, The properties of optical glass (Springer-Verlag, Berlin,

1995.

18. E. Hecht, Optics, 4th

ed., (Addison Wesley, San Francisco, 2002).

19. S. H. Fan, P. R. Villeneuve, J. D. Joannopoulos, and E. F. Schubert, "High

extraction efficiency of spontaneous emission from slabs of photonic crystals,"

Phys Rev Lett 78, 3294-3297 (1997).

20. H. Ichikawa, and T. Baba, "Efficiency enhancement in a light-emitting diode with

a two-dimensional surface grating photonic crystal," Appl Phys Lett 84, 457-459

(2004).

21. T. Kim, P. O. Leisher, A. J. Danner, R. Wirth, K. Streubel, and K. D. Choquette,

"Photonic crystal structure effect on the enhancement in the external quantum

efficiency of a red LED," Ieee Photonic Tech L 18, 1876-1878 (2006).

22. S. Jeon, J. U. Park, R. Cirelli, S. Yang, C. E. Heitzman, P. V. Braun, P. J. A.

Kenis, and J. A. Rogers, "Fabricating complex three-dimensional nanostructures

with high-resolution conformable phase masks," P Natl Acad Sci USA 101,

12428-12433 (2004).

23. Z. Poole, D. Xu, K. P. Chen, I. Olvera, K. Ohlinger, and Y. Lin, "Holographic

fabrication of three-dimensional orthorhombic and tetragonal photonic crystal

templates using a diffractive optical element," Appl Phys Lett 91, 251101-251103

(2007).

24. D. Chanda, L. E. Abolghasemi, M. Haque, M. L. Ng, and P. R. Herman, "Multi-

level diffractive optics for single laser exposure fabrication of telecom-band


80

diamond-like 3-dimensional photonic crystals," Opt Express 16, 15402-15414

(2008).

25. Y. Z. Chen, C. X. Zhou, X. G. Luo, and C. L. Du, "Structured lens formed by a 2D

square hole array in a metallic film," Opt Lett 33, 753-755 (2008).

Chapter 4: Fabrication of micro-optics on glass

81

CHAPTER 4

FABRICATION OF MICRO-OPTICS ON GLASS

4.1. Introduction

A microlens array (MLA) is one of the important micro-optics components used in

many different applications, such as beam shaping, interconnections and imaging [1]. The

beam shaping applications include using MLAs for beam homogenizing, and

interconnection applications include matching light into the cores of optical fibers and

concentrating light into the active areas of CCD arrays, while imaging applications include

image transfer used in photocopiers and integral photography.

The microlens of a MLA can be either diffractive or refractive [2-3]. A diffractive

microlens has grooves with stepped edges that approximate the ideal lens shape. It uses

diffraction to direct the incident light into the desired distribution. Since diffraction effects

are wavelength dependent, a diffractive microlens is optimized for only a specific

wavelength. It has the advantages in fabrication and replication by using standard

semiconductor processes, such as photolithography and RIE. A refractive microlens has a

curved convex profile. It focuses incident light by light refraction at the convex surface, and

hence can be used for a range of different wavelengths. A refractive MLA consists of

numerous refractive convex microlenses formed in a two dimensional array on a substrate.


82

Different types of substrates cater for different wavelengths of light, such as fused silica for

visible to near-IR wavelengths and silicon for mid-IR wavelengths.

4.2. Fabrication of microlens array

Microlenses made by melting photoresist have emerged as one of the most widely

adapted methods for fabricating microlenses due to the simplicity and low cost. Popovic et

al. [4] first demonstrated the formation of array microlenses on a substrate by thermal

heating of the photoresist patterns. Cylindrical islands of photoresist are produced using

standard photolithographic techniques. These are then melted to the glass transition

temperature of the photoresist so that surface tension pulls the melted photoresist islands

into a segment of a spherical shape. Figure 4.1 illustrates the fabrication process. Various

researchers [5-7] have extended this technique to make two dimensional arrays of

microlenses with circular, square, hexagonal and other shaped apertures. This method is also

adapted for fabricating the microlens array in this chapter.


83

Fig. 4.1. Fabrication steps for fabricating a microlens array using the photoresist reflow method.

The focal length f and the diameter of the lens D are interrelated via the following

equations. The focal length f of the microlenses is related to the radius of curvature R by:

1

n

Rf , (4.1)

where n is the refractive index of the microlens. R is related to D by [2,5]:

s

Ds

R2

2

2

2

, (4.2)

where s is the sag height of the photoresist islands (ref. Fig. 4.1). Combining Eqs. 4.1 and

4.2, we get:

UV irradiation

Chrome-on-glass

Photoresist

Fused silica

Photoresist

Fused silica

Exposure

Develop

Melt

t

s t: thickness of photoresist

s: sag height of microlens

D: diameter of photoresist

D


84

1

2

4

22

n

s

Ds

f . (4.3)

Therefore, in the design and fabrication of the microlens, changing the diameter of the

microlens affects the focal length, and vice versa.

When using the photoresist thermal reflow method to fabricate the microlenses, we

can find the relationship between the initial thickness on the photoresist t, and its focal

length f and diameter D, to have a better understanding of how the thickness of the

photoresist coated influences the relationship between f and D. After patterning to form tiny

islands on the photoresist (as shown in Fig. 4.1) but before melting and reflowing, the

photoresist volume can be calculated as the volume of the cylinder:

tD

Vcyl

2

2

, (4.4)

where t is the thickness of the photoresist. After melting, the volume of the reflowed

spherical photoresist microlens is like a segment of a sphere and can be found as:

sRsVresist 3 3

1 2 . (4.5)

In the simple case assumption, the volume of the photoresist does not change during the

reflowing. Therefore, Vcyl = Vresist and combining Eq. 4.2 with Eqs. 4.4 and 4.5, the relation

of t with s and D can be found as:


85

2

2

436 D

sst . (4.6)

Combining Eqs. 4.3 and Eq. 4.6, the relationship between f and D can be plotted for

different values of t. Figure 4.2 plots the dependence of f with D at three different

photoresist thickness of t = 1, 3 and 5 µm. Figure 4.2 gives a rough estimation of the

parameters of a fabricated MLA when different photoresists thickness are used.

Fig. 4.2. The plot of f versus D at three different photoresist thicknesses.


86

4.2.1. Sample preparation

Positive photoresist (S1822 from Shipley) is spin-coated on fused silica substrates

(Photonik Singapore), each 1.5 cm × 1.5 cm and 1 mm thick. The substrates are cleaned in

acetone, IPA and de-ionized water prior to usage. The photoresist is spin-coated onto the

substrates at a rotational speed of 3000 rpm to give a thickness of ~ 3 µm. After the coating,

the samples are baked at 100 ~ 110 ˚C to remove the solvents. A mercury arc lamp (10mW,

λ = 365 nm) is used to expose the photoresist and the developer MF319 is used to develop

the photoresist.

4.2.2. Exposure of photoresist and reflow

A simple chrome-on-glass mask with periodic line patterns is used to expose S1822

positive photoresist to fabricate the microlens array. The period of the line-space is 9 µm,

the line width 5 µm and the space width 4 µm. Firstly, a single exposure is carried out on

the S1822 positive photoresist using a mercury arm lamp for an exposure time of ~ 30

seconds. Subsequently, the substrate is rotated 90˚ and a second exposure is carried out for ~

30 seconds. After developing, a square array of photoresist remains on the substrate. Figure

4.3(a) shows the photoresist patterns after the double exposure. The photoresist islands are ~

5.4 µm in width while the space is ~ 3.4 µm. These measurements correspond to the period

and duty cycle of the chrome-on-glass mask. The edges of the resist are slightly rough.

However, this is not a big concern because during the photoresist reflow step to form the

microlenses, the rough edges are removed due to the reflow of the photoresist. To reflow the


87

photoresist, the photoresist-coated substrates are then placed on a hot plate to heat up and

melt at a temperature T = 150 °C. When the photoresist has melted to a liquid state, it

reflows into a spherical shape due to the surface tension among the photoresist, the

surrounding air and the glass substrate. The relationship between the internal pressure of the

liquid photoresist and the external atmospheric pressure is related to the surface tension by

the Young-Laplace equation:

21

11

RRP , (4.7)

where ΔP is the pressure difference between the intternal pressure of the photoresist and the

external atmospheric pressure, R1 and R2 are the principle radii of curvature. Since the

photoresist is spherically shaped, R1 = R2 and Eq. 4.7 reduces to:

RRRP

211

21

. (4.8)

Eq. 4.8 shows that the radius of curvature of the reflowed photoresist depends on the surface

tension and the pressure difference.


88

Fig. 4.3(a). Photoresist pattern array formed after the double exposure by a chrome-on-glass mask.

Figure 4.3(b) shows the photoresist patterns after reflowing the photoresist at a

temperature of T = 150 ˚C for 5 minutes to obtain the profile of spherical segments. The

picture shows that the rough edges in Fig. 5.2a have disappeared after the reflow. A uniform

array of microlenses with spherical segments is obtained.


89

Fig. 4.3(b). After the melting, the photoresist patterns are reflowed into a spherical profile.

4.2.3. Transferring the photoresist shape to the substrate

After fabricating and reflowing the photoresist, reactive ion etching is used to etch

and transfer the spherical profile of the photoresist onto the fused silica substrate. This step

fabricates the MLA on the fused silica substrate [8-9]. The reactive etching gas is CF4 at a

flow rate of 50 sccm. The forward bias is 250 W and the chamber pressure is 20 mTorr. The

etching cycle illustrated in Table 4.1 is repeated ~ 10 times to over-etch into the fused silica

substrate. 180 seconds of etching is carried out and then followed by a break of 60 seconds

before the cycle is repeated. This is necessary because fused silica is not a good thermal

conductor of heat. As the etching takes place, the ion bombardment causes the substrate to

heat up. Hence, the break is required to decrease the temperature of the substrate. Otherwise,

overheating of the sample occurs, causing the photoresist to carbonize before the patterns

are fully transferred to the substrate. This destroys the spherical profile of the photoresist.


90

The over-etching is done to make sure that the photoresist profile is fully transferred down

to the substrate. If a partial transfer occurrs, an unusable flat profile is obtained, as shown in

Fig. 4.4.

Table. 4.1. The etching cycle used to transfer the photoresist onto the fused silica substrate.

Time (s)

CF4 (50 sccm, 250 W, 20mTorr) 180

Break (to cool down substrate) 60

Fig. 4.4. Illustration showing the need to totally transfer the photoresist pattern profile to fused silica.

Spherical photoresist before etching

Photoresist profile after partial etch

Photoresist totally transferred to substrate to

form the MLA

Flat profile after removal of photoresist


91

Figure 4.5 shows the AFM profile of the MLA transferred successfully from the

photoresist onto the fused silica substrate. The scan shows that the surface is smooth and of

a high quality. The inset at the top right corner shows the cross-section profile of one

microlens. The measured sag height is ~ 1 µm while the diameter is ~ 7 µm. This shows that

a good spherical profile can be obtained after transferring the MLA from the photoresist

onto the fused silica substrate by RIE.

Fig. 4.5. AFM profile of the MLA transferred from the photoresist onto the fused silica substrate.

6.98 µm

1.03 µm


92

After fabricating the microlens array, a simple test is carried out to investigate its

light focusing ability. A semiconductor laser diode, with a wavelength of 632.8 nm, is shone

onto the backside of the fabricated MLA. The MLA then focuses the light into an array of

focused spots. The focused spots are observed using a microscope. The image is then

captured by a CCD. Figure 4.6 shows the array of focused spots. It shows that the MLA is

functional. Meanwhile, the illustration on the right shows how the focused spots are imaged.

The microscope first focuses on the MLA substrate (point A). Then the microscope stage is

moved up until the spots appear sharply focused (point B). The distance moved is the focal

length of the MLA. For the fabricated MLA, its focal length measured using this method is

~ 7 µm. Furthermore, the size of the focused spot gives a rough estimation of the resolution

of the microlenses. The spot size measured using the microscope is ~ 660 nm. The

fabricated MLA can then be used for MLA lithography, which will be shown in Chapter 5.

Fig. 4.6. The fabricated MLA focused the incident light into an array of focused spots. The spot size

is measured as ~ 660 nm.

Incident light

Microscope with CCD

Point A

Point B


93

4.3. Introduction to solid immersion lens

The previous section introduces the MLA and its fabrication steps to structure the

MLA on glass. This section introduces another important optics – the solid immersion lens.

Solid immersion lens (SIL) has been used to increase the resolution of optical systems.

However, from the viewpoint of micro-optics, few attempts are made to miniaturize the SIL

into a micro-SIL [10-11]. Just as the case of the MLA, transforming the SIL into micro-SIL

could potentially have benefits and other useful applications.

The concept of a solid immersion lens was first introduced when solid immersion

microscopy was invented by Mansfield and Kino in 1990 [12] and since then it has

generated an increasing level of interest. The basic idea of immersion technology is to fill

the object space with a higher refractive index material, so that the effective wavelength of

light is „reduced‟ when it passes through the material, resulting in enhanced resolution. The

solid immersion lens (SIL) is made of higher refractive index materials (with refractive

index greater than air) and there are two configurations of SIL used: the hemispherical solid

immersion lens and the super-hemispherical solid immersion lens.

If the focal point of the lens is directed towards the centre of the sphere, the

incoming rays arrive at normal incidence to the surface of the sphere. Therefore, there is no

refraction of the light rays at the surface of the sphere, resulting in spherical aberration free

imaging, as shown in Fig. 4.7. A transparent solid with a hemispherical shape is used as the

lens, with the centre of the flat surface placed in contact with the object of interest, as


94

illustrated in Fig. 4.8 [13]. This condition forms the basis for the design of a hemispherical

SIL.

Fig. 4.7. The hemispherical configuration, where the incident light focused at the centre of the

sphere [13].

Fig. 4.8. Illustration showing how a half sphere can be used as a hemispherical solid immersion lens

for higher resolution imaging [13].


95

The super-hemispherical solid immersion lens makes use of the aplanatic point of

the sphere. This concept is derived from the theory described by Born and Wolf [14]. The

theory states that sharp imaging by a spherical surface can be achieved by making use of the

aplanatic point of the sphere. Figure 4.9 shows the position of the aplanatic point, labeled

Z0. The construction implies that all rays focused at point Z1 refract at the surface of the

sphere and focus instead at point Z0. The point Z0 is located at a distance Z0 = (n1/n0)R from

the centre of the sphere [14], where R is the radius of curvature of the sphere, and n1 and n0

are the refractive index of air and the sphere, respectively.. Figure 4.10 illustrates the shape

of the super-hemispherical lens when it is used for imaging.

Fig. 4.9. The aplanatic point of the sphere, where the incident light focuses at the point Z0. It has a

virtual focus at Z1 [13].


96

Fig. 4.10. A super-hemispherical solid immersion lens used for the sub-surface imaging [13].

One of the major advantages of using solid immersion lenses is the enhanced

magnification. According to the Abbe sine condition, to have sharp imaging, the following

condition has to be fulfilled [14-15]:

o

i

i

o

NA

NA

y

y

, (4.9)

where yo and yi are the lateral sizes of the object and the image, respectively. NAo and NAi

the numerical apertures at the object and image side, respectively. The lateral image

magnification is Δyi / Δyo and in the case of the hemispherical SIL, a material with refractive

index n is being used to fill up the object space, therefore NAo = n∙sinθ, while NAi = sinθ.

Hence, from Eq. 4.9, we get the magnification factor of n due to the hemispherical SIL. As

for the case of a super-hemispherical SIL, there is additional refraction at the surface of the


97

sphere. Due to the Snell‟s law of refraction, an additional magnification factor of n is thus

introduced. Applying Eq. 4.9, a magnification factor of n2 can be obtained when the super-

hemispherical SIL is used. Table 4.2 compares the magnification enhancements of the two

SILs [13, 16], where n is the refractive index of the SIL.

Table 4.2. Comparison between the magnification enhancements of the two SILs.

Lateral

magnification

Longitudinal

magnification

Resolution gain

Hemispherical SIL n n 1/n

Aplanatic SIL n2 n

3 1/n

2

4.4. Fabrication of the micro-solid immersion lens array

The strategy adapted to fabricate the micro-solid immersion lens array is similar to

the photoresist melting and reflow method illustrated in Section 4.2. However, the crucial

point is to offset the focal plane of the micro-solid immersion lens array to the back of the

glass substrate, as illustrated in Fig.4.11, which illustrates a hemispherical configuration of

the µSIL array (h-µSIL).


98

Fig. 4.11 The concept of the fabrication of a h-µSIL. The hemispherical centre of the h-µSIL is

offset to the back of the glass substrate.

To fabricate the h-µSIL, positive photoresist (AZ9260 from AZ Electronic

Materials) is coated at a spin speed of 6000 rpm on a 1.5 cm × 1.5 cm fused silica substrate

at a thickness of 190 µm. The photoresist film thickness coated is ~ 5 µm. A chrome mask

of circular holes, each with a radius of 36 µm and a pitch of 150 µm, is used to expose the

photoresist, creating an array of photoresist islands. These photoresist islands are then

reflowed using the thermal reflow at T = 150°C for 5 minutes to transform the photoresist

islands into a spherical profile. Subsequently, RIE is used to etch the photoresist islands

anisotropically and transfer the patterns down to the fused silica. 50 sccm of CF4 and 5 sccm

of O2 are used as the etching gases and the forward power is set at 140 W. Inductively

coupled plasma (ICP) at a power of 420 W is used to increase the etching rate. The crucial

part of the fabrication step is to ensure that the hemispherical centre of the h-µSIL is offset

to the backside of the fused silica, as illustrated in Fig. 4.11. The focal plane of the MLA is

then aligned to the backside of the h-µSIL. The SEM image of the fabricated h-µSIL is

shown in Fig. 4.12, demonstrating that high quality h-µSIL can be fabricated.

Back of glass substrate Hemispherical centre offset to the

back of glass substrate


99

Fig. 4.12. SEM image of the fabricated h-µSIL.

Figure 4.13 plots the profile of the h-µSIL measured by a surface profiler. The sag

height is ~ 3.4 µm while the radius of the lens is ~ 36 µm. From Eq. 4.2, the radius of

curvature R can be found to be R ≈ 190 µm, which is the same as the thickness of the fused

silica. This means that after the etching, the hemispherical point of h-µSIL is located at the

back of the substrate.

The radius of curvature can also be found by measuring the focal length of the h-

µSIL, according to Eq. 4.1. To measure the focal length, the setup illustrated in Fig. 4.6,

which is used to observe the focused spots, can be adapted to measure the focal length. By

first focusing on the substrate and then moving the microscope stage to the plane of the

focused spots, the distance moved is approximately equal to the focal length of the h-µSIL.


100

A semiconductor laser diode at a wavelength of 632.8 nm is used to measure the focal

length of the h-µSIL. The focal length f is found to be ~ 420 µm. The refractive index of

fused silica at this wavelength is ~ 1.457 so R is calculated to be ~ 190 µm. This result

matches well with the thickness of the fused silica substrate.

Fig. 4.13. Plot of the measured sag height of the h-µSIL. The sag height is ~ 3.4 µm while

the diameter is ~ 72 µm.


101

4.5. Applying the h-µSIL to improve the resolution of a MLA

The h-µSIL is applied to a MLA to investigate and characterize its enhancement

effects. To attach the h-µSIL to a MLA, the h-µSIL is mounted on a stationary arm support.

The MLA has a pitch of 150 µm, a radius of 36 µm and a focal length of ~ 200 µm. The

numerical aperture (NA) of the MLA is about 0.18. It is mounted on the top of a beam-

splitter, which is then placed on a precision nanostage. The nanostage has 7 degrees of

freedom: X-Y-Z, piezo-Z, tilt and tip to level the MLA to the h-µSIL, and the theta axis for

rotation control. Figure 4.14 shows the schematic drawing of the alignment setup. Using the

nanostage ensures that the MLA can be moved relative to the h-µSIL with high precision in

all axes. A CCD camera and objective lenses mounted above the arm support monitor the

alignment process. Objective lenses ranging from 10× to 100× can be interchanged during

the alignment process so as to provide the necessary magnification to detect any

misalignment between the MLA and the h-µSIL. Light is irradiated on the beam-splitter,

which reflects light to the backside of the MLA, producing an array of focused spots. The

focused spots and alignment process are monitored closely through the CCD as the Z axis is

slowly raised to move the MLA towards the h-µSIL. The microscope objectives are focused

on the backside of the h-µSIL. When it is observed through the CCD that the focused spots

of the MLA appear on the back of the h-µSIL in sharp focus, it shows that the focal plane of

the MLA has reached the hemispherical centre of the h-µSIL.


102

Fig. 4.14. The schematic drawing of the alignment setup used to align the h-µSIL to the MLA.

After the alignment process, optical adhesive (Norland Optical Adhesive 63) is applied at

the peripheral areas of the h-µSIL and cured using a UV LED, which emits light at a

wavelength of 365 nm, and bonds the h-µSIL to the MLA. This step is to maintain and lock

the position of the h-µSIL to the MLA after the alignment. This is crucial since any slight

shift would cause the h-µSIL to misalign significantly with the MLA. Figure 4.15 shows an

optical image of the alignment process, where the h-µSIL is aligned to the MLA. The MLA

is intentionally moved slightly aside to evaluate the alignment achieved between the h-µSIL

(red dash circle) and the MLA (black dash circle).

Light source

CCD

h-µSIL

MLA

Beam-splitter

Arm

support

Nanostage

Interchangeable

microscope objectives


103

Fig. 4.15. Alignment process of the intentional offset between the h-µSIL and the MLA.

4.5.1. Spot size analyses

According to the Sparrow‟s criterion [17] by measuring the spot size at its full-width

half-maximum (FWHM), the resolution of an optical system can be obtained. Therefore, the

spot sizes produced by the MLA before and after the attachment of the h-µSIL are observed.

Figure 4.16(a) shows the spot size produced by the MLA before the attachment of the h-

µSIL when light from the semiconductor laser diode, with a wavelength of 632.8 nm, passes

through it and is focused at the focal plane. The spot size is ~ 1.9 µm. Figure 4.16(b) shows

the spot size produced by the MLA after the h-µSIL is attached. The spot size is reduced to

MLA

h-µSIL


104

~ 1.3 µm. By reducing the spot size from ~ 1.9 to ~ 1.3 µm, the h-µSIL is able to improve

the resolution of the MLA imaging by n = 1.5.

Fig. 4.16(a). Spot size produced by the MLA before the attachment of h-µSIL. The spot size is ~ 1.9

µm.

Fig. 4.16(b). The spot size produced by the MLA after the attachment of the h-µSIL. The spot size

reduced to ~ 1.3 µm.

1.9 µm

1.3 µm


105

4.5.2. Intensity profile of projected lines

Besides studying the spot size, the spread of lines projected by the MLA with and

without the h-µSIL is also investigated. A smaller spread of a point translates to a better

resolution. Similarly, this concept extends to the case of lines. A series of 4 lines are printed

on transparency and projected by the MLA. The intensities of the 4 lines, before and after

the attachment of the h-µSIL, are then analyzed by a beam profiler, which measures the

light intensities it receives at each individual pixel of its detector. Figure 4.17(a) shows the

images of the array of 4 lines projected by the MLA. The lines are spaced ~ 10 µm apart.

Figure 4.17(b) shows the intensity profile of the 4 lines projected by the MLA before the

attachment of the h-µSIL. The width of the 2nd

line is ~ 4.1 µm. The line spread is reduced

when the h-µSIL is attached. Figure 4.17(c) shows that the width of the 2nd

line decreases to

~ 3 µm after the attachment of the h-µSIL to the MLA. In Figs. 4.17(b) and (c), the 4

intensity peaks measured are not of the same amplitude and this could be due to non-

uniformity of the line patterns printed on the transparency. Nevertheless, it shows that with

the attachment of the h-µSIL, the spread of the lines is better controlled.


106

Fig. 4.17(a). An array of 4 lines projected by the MLA. The same array of 4 lines is projected by the

MLA before and after the attachment of the h-µSIL.

Fig. 4.17(b). The width of the lines projected by the MLA without the h-µSIL and measured by the

beam analyzer.

4.1 µm


107

Fig. 4.17(c). The same array of 4 lines projected by the MLA with the h-µSIL attached. The

width of the lines decreases to ~ 3 µm.

4.5.3. Projection of ‘+’ shape

Subsequently, a two dimensional „+‟ shape is projected by the MLA with and without

the h-µSIL to characterize the resolution enhancement. The „+‟ sign is printed onto a

transparency and projected by the MLA. The setup is further illustrated in [18]. Figure

4.18(a) shows the images of the „+‟ sign projected through the MLA only. The side of the

cross is ~ 55 µm long. Meanwhile, Fig. 4.18(b) shows the images after the h-µSIL is bonded

and attached to the MLA. The size of the „+‟ sign is reduced with the attachment of the h-

µSIL. The side of the cross is ~ 38 µm. This corresponds to a reduction of ~ 1.5 times. A

reduction, instead of a magnification, occurs because a projection setup is used. The ability

3 µm


108

of the h-µSIL to reduce the image by n, where the refractive index of the fused silica is ~ 1.5,

further indicates that the h-µSIL is properly aligned.

Fig. 4.18(a). An array of „+‟ sign projected by the MLA without the h-µSIL array attached.

55 µm


109

Fig. 4.18(b). An array of „+‟ sign projected by the MLA with the h-µSIL array attached. The „+‟

shapes are smaller due to the reduction effect of the h-µSIL.

In summary, this chapter shows how good quality micro-optics can be fabricated on

glass using the photoresist melting and reflow method. Besides the MLA, it is also shown

that the solid immersion lens can also be miniaturized to a hemispherical micro-solid

immersion lens array. It is then successfully applied to a MLA to increase its resolution.

38 µm


110

4.6. Reference

1. D. Daly, Microlens Array, (Taylor & Francis, London, 2001).

2. S. Sinzinger and J. Jahns, Microoptics, (Wiley-VCH, Weinheim, 1999).

3. N. F. Borrelli, Microoptics Technology, (Marcel Dekker, New York, 2005).

4. Z. D. Popovic, R. A. Sprague and G. A. Neville Connell, “Technique for monolithic

fabrication of microlens arrays,” Appl. Opt. 27, 1281-1284 (1998).

5. D. J. Daly, R. F. Stevens, M. C. Hutley and N. Davies, “The manufacture of

microlenses by melting photoresist,” Meas. Sci. Technol. 1, 759-766 (1990).

6. K. Mercereau, C. R. Nijander, W. P. Townsend, R. J. Crisci, A. Y. Feldblum and D. J.

Daly, “Fabrication and measurement of fused silica microlens arrays,” Proc. SPIE

1751, 229-233 (1992).

7. T. R. Jay, M. B. Stern and R. E. Knowlden, “Effect of refractive microlens array

fabrication parameters on optical quality,” Proc. SPIE 1751, 236-245 (1992).

8. M. B. Stern and T. R. Ray, “Dry etching for coherent refractive microlens arrays,” Opt.

Eng. 33, 3547-3551 (1994).

9. P. Savander, “Microlens arrays etched into glass and silicon,” Opt. Lasers Eng. 20, 97-

107 (1994).

10. D. A. Fletcher, K. B. Crozier, K. W. Guarini, S. C. Minne, G. S. Kino, C. F. Quate and

K. E. Goodson, "Microfabricated silicon solid immersion lens," J. Microeletromech.

Syst. 10, 450-459 (2001).

11. T. Kishi, S. Shibata and T. Yano, "Fabrication of SIL array of glass by surface-tension

mold technique," Proc. SPIE 6126, 61260P 1-8 (2006).


111

12. S. M. Mansfield and G. S. Kino, "Solid immersion microscope," Appl. Phys. Lett. 57,

2615-2616 (1990).

13. K. A. Serrels, E. Ramsay, P. A. Dalgarno, B. D. Gerardot, J. A. O'Connor, R. H.

Hadfield, R. J. Warburton and D. T. Reid, "Solid immersion lens applications for

nanophotonic devices," J. Nanophotonics 2, 021854 (2008).

14. M. Born and E. Wolf, Principles of optics, 7 ed., (Cambridge University Press, UK,

1999).

15. Q. Wu, "Imaging with solid immersion lenses, spatial resolution, and applications,"

Proc. IEEE 88, 1491-1498 (2000).

16. S. B. Ippolito, “High spatial resolution subsurface microscopy,” Ph.D Dissertation,

Boston College of Engineering, (2004).

17. T. R. Corle and G. S. Kino, Confocal scanning optical microscopy and related imaging

systems, (Academic Press, New York, 1996).

18. M. H. Wu, and G. M. Whitesides, "Fabrication of two-dimensional arrays of

microlenses and their applications in photolithography," J. Micromech. Microeng. 12,

747-758 (2002).

Chapter 5: Parallel micro-/nano-patterning by laser MLA lithography

112

CHAPTER 5

PARALLEL MICRO-/NANO-PATTERNING BY LASER

MICRO-LENS ARRAY LITHOGRAPHY

5.1. Introduction

Chapter 3 introduces LIL as a parallel, flexible and cost effective method to fabricate

large area periodic structures. Periodic lines, dots and nanoholes can be fabricated,

depending on the types of photoresist used. In addition, the periodicity of the patterns can be

varied flexibly by changing the interference angle. However, the LIL method is restricted to

create simple structures, such as lines, dots, holes, and ovals [1]. It is challenging to use the

LIL technique to pattern arbitrary shape designs. To address this limitation, the use of

micro-lens array (MLA) lithography to create arbitrary patterns is investigated in this

chapter.

5.1.1. Micro-lens array (MLA) lithography

Upon light irradiation, each micro-lens of the MLA focuses the light at their

respective focal points. This effectively transforms the incident light into an array of parallel

focused micro-light beams at the focal plane. Then, by using these focused micro-light

beams to perform surface patterning, an array of patterns can be created at the same time.


113

Figure 5.1 illustrates this concept. A laser light irradiates through the MLA. Due to the high

directionality of the laser beam, the light rays impinge parallel on the MLA, which focuses

the light at the focal points. As the light rays converge at the focal points, the light

intensities at the focal points are much stronger than the rest of the regions. A substrate

coated with photoresist is then brought to the focal plane of the MLA. The focused spots

expose the photoresist while other regions of the photoresist remain unexposed. This is

achieved by controlling the focused spots light intensity to be above the exposure threshold

to expose the photoresist while the other regions remain unexposed. Figure 5.2 shows the

dots array produced. The substrate can be moved with respect to the stationary MLA to

write out arrays of different arbitrary designs on the photoresist. This procedure forms the

basis of the laser MLA lithography throughout the work done in this chapter.

Fig. 5.1. Illustration of the concept of laser MLA lithography.

Substrate

Microlens

Arrays

Focal

Length

Incident laser source


114

Fig. 5.2. An array of dots formed after laser light irradiation through a MLA.

5.2. Experimental

5.2.1. Sample preparation

Fused silica substrates (Photonik Singapore), 1.5 cm × 1.5 cm and 1 mm thick are

used in the experiment. They are cleaned in acetone, IPA and de-ionized water to remove

surface contamination. The fused silica samples are coated with photoresist, either positive

photoresist (S1805, Shipley) or negative photoresist (Ma-N 1405, Microchem), depending

on the types of applications. The photoresist is spin-coated onto the fused silica substrates at

a rotational speed of 4000 ~ 6000 rpm. Immediately after the photoresist coating, the

samples are placed on a hotplate to bake the photoresist at a temperature of 100 ~ 110 ˚C to

remove the remaining solvents.


115

5.2.2. Experimental setup

Figure 5.3 shows the experimental setup of the MLA lithography [2-3]. The laser

used is a second harmonic femtosecond laser (Tsunami, Spectra Physics) with a wavelength

of 400 nm. It passes through a beam expander to enlarge the beam size before irradiating the

MLA. The MLA is held stationary on a substrate holder by a suction force from a vacuum

pump. The sample is placed on the nanostage and fixed by the suction force. The nanostage

has seven degrees of freedom: x-y-z control, fine-z, tilt and tip and rotation control. It levels

the sample with respect to the MLA and then brings the sample to the focal plane of the

MLA for laser MLA lithography. A computer controls the motion of the nanostage and the

shutter operation. Figure 5.4 shows the MLA used in the experiment. It is arranged in a

triangular array with a focal length of 28.8 µm, a diameter of 40 µm and a period of 50 µm.

Fig. 5.3. Experimental setup for the laser MLA lithography.

Fs

lase

r λ =

400

nm

X control

Y control

Z control

Fine Z control

Tilt

Tip

Rotation

Beam Expander

PC Control

Shutter Mirror Mirror

Nanostage

MLA

Sample


116

Fig. 5.4. Microscope image of the MLA used in the experiment. The period of the MLA is 50 µm.

5.3. Results and discussion

5.3.1. Arbitrary patterning

The computer controls the movement of the nanostage precisely during the laser

MLA lithography to achieve maskless direct-writing. Various arbitrary shapes can be

programmed and the shape paths would be traced accurately by moving the nanostage,

allowing a variety of intricate designs to be directly written and fabricated by the laser MLA

lithography. Figures 5.5(a) and (b) show some examples of the structures patterned using

this technique. Figure 5.5(a) shows a logo array. A circle with 3 diagonal lines is

successfully produced on negative photoresist and transferred to the fused silica substrate by

RIE. Figure 5.5(b) shows a generic IC design patterned using the laser MLA lithography


117

while the inset shows the 3D profile measured by AFM after the patterns are transferred to

the fused silica substrate by RIE. It shows that good quality patterns can be fabricated with

this technique. The area size of the IC design is 10 µm × 10 µm and it takes about 12

seconds to complete the patterning when the nanostage moves at a speed of 5 µm/s. These

results demonstrate the flexibility and high throughput of this technique for applications in

direct writing of large area periodic patterns.

Fig. 5.5(a). A logo array pattern fabricated on fused silica by the laser MLA lithography and RIE.


118

Fig. 5.5(b). Generic IC design fabricated by laser MLA lithography on fused silica. The inset shows

the 3D AFM image of an unit cell after the patterns were transferred to the fused silica.

5.3.2. Arbitrary angle patterning

Besides programming the nanostage to move in the regular Cartesian X-Y direction,

the nanostage can also be programmed to move in any arbitrary direction. By varying the X

to Y ratio, the nanostage can move at various angles in the diagonal direction. For example,

moving the nanostage simultaneously in the X and Y direction by 1 unit makes it moves 45°

to the horizontal direction instead. Therefore, by controlling the X to Y ratio, the nanostage

can move in other angles to the horizontal direction. Figure 5.6 illustrates the design. The

circles represent the individual microlenses arranged in a triangular array. The symbols c, x,

and γ represent the lattice constant, perpendicular distance between the lines and angle to the


119

horizontal axis, respectively. By moving the nanostage at an angle γ with respect to the

horizontal axis, the distance x between the adjacent rows can be varied according to:

x = c sin γ (5.1)

Fig. 5.6. Schematic drawing of laser MLA patterning by moving the nanostage at an angle to the

horizontal axis.

Using this patterning technique, parallel lines with varying distance between lines

can be patterned. Figures 5.7(a) to (d) show some examples of such line patterns at various

angles γ. The spacing between the lines decreases as γ decreases. This method can be used

to vary the period of the patterns.

x

Traverse

direction

c γ


120

(a)

(b)

(c)

(d)

Fig. 5.7. Line array patterns with different periods obtained by changing the angle γ.


121

5.3.3. Resist trimming

After forming the patterns on negative photoresist, the feature sizes of the patterns

can be further reduced by post-process resist trimming. RIE is used to perform resist

trimming on the patterns. The reactive gas used is O2 at a gas flow rate of 45 sccm.

Photoresist is generally hydrocarbon in nature so O2 reacts with the photoresist to form

volatile by-products, thereby etching and removing the photoresist. However, the bias power

and the chamber pressure affect the trim rate [4]. The etching rate increases with the

chamber pressure but decreases with the bias power. When the bias power is high, the

reactive ion species have higher ion energies. The bias power accelerates the ions towards

the substrate vertically, reducing the amount of ions available on the sidewalls of the

photoresist patterns. This results in more vertical etching and less lateral etching. By

reducing the power, the reactive ions have lower ion energies, making them less directional

and increasing the chances for the ions to be deflected to the photoresist sidewalls [4]. This

results in more lateral etching and less vertical etching.

When the chamber pressure increases, the ion energy decreases due to the increase in

the number of ion collisions. This reduces the mean free path of the ions, scattering more of

the ions onto the sidewalls of the photoresist. This in turn increases the lateral etching.

Furthermore, the chemical flux increases with the chamber pressure. Since the trimming

process depends more on the chemical flux rather than the ion energy [4], this also

contributes to an increase in the lateral etching rate. Table 5.1 tabulates the lateral and

horizontal etching rates at a chamber pressure of 40 mTorr and a bias power of 40 W for


122

resist trimming. The lateral etching rate is about twice as high as the vertical etching rate.

Figure 5.8 compares the feature size of a patterned „L‟ shape before and after resist

trimming. By the photoresist trimming, the dimension can be reduced from ~ 1 to ~ 0.5 µm.

Table 5.1. Comparison between the lateral and vertical etching rates in the resist trimming process.

Etching time of 60 seconds Lateral etching Vertical etching

Etched dimensions (nm) 150 75

Etched rates (nm/s) 2.5 1.25

Fig. 5.8. The left figure shows the feature size obtained after the laser MLA lithography. The right

figure shows the further reduced feature size obtained after the resist trimming.

1.09 µm

0.51 µm


123

Although the lateral etching rate is higher than the vertical etching rate, the vertical

etching causes the photoresist patterns to become thinner after the resist trimming. After the

photoresist trimming, the thickness of the photoresist is about 50 nm thick. This is not thick

enough to act as an etching mask if the patterns are to be transferred to the fused silica

substrate. To address this issue, a hard mask is used for deeper etching of the fused silica

substrate. Figure 5.9 illustrates the modified process flow when the resist trimming process

is incorporated to further reduce the feature sizes. The SiO2 capping layer is 10 nm thick.

Since the etching selectivity of the photoresist to the SiO2 layer is about 1:1, the 50 nm thick

photoresist is thick enough to transfer the patterns to the SiO2 layer. After patterning the

SiO2 layer, the Cr hard mask layer can then be etched using Cr etchant, which does not

attack the SiO2 layer.


124

Fig. 5.9. Process flow to incorporate the resist trimming.

Fused silica

Cr

E-beam evaporation of Cr

1.

Fused silica

E-beam evaporation of SiO2 2.

SiO2

5.

SiO2

Resist trimming

Fused silica

Photoresist

Fused silica

6.

SiO2

RIE

Cr

Fused silica

7.

SiO2

Cr

Cr etchant

Fused silica

8.

SiO2

Cr

RIE

Cr

4.

SiO2

Photoresist MLA lithography

Fused silica

Cr

Cr

Fused silica

Photoresist coating

3.

SiO2 Photoresist


125

5.3.4. Patterning and etching for phase mask fabrication

Various studies [5-8] have shown that by utilizing the phase shift of light, phase shift

masks have been used successfully to print smaller patterns. Figure 5.10 shows the concept

of using a phase shift structure to change the phase of the incident light. The light undergoes

a phase shift of 180° after passing through the phase shift structure. In order to introduce the

180° phase shift, the structure has a height of:

12

n

h

(5.2)

where h is the height of the phase shift structure, λ the wavelength of the light and n the

refractive index of the material. From Eq. 5.2, different wavelength of light requires

different height to introduce the phase shift. For example, at λ = 365 nm, glass has a

refractive index ~ 1.47 so the phase shift structure requires a step height of h ≈ 385 nm to

introduce the phase shift. When light passes through the phase shift structure (region A in

Fig. 5.10), the electric field is shifted 180˚ out of phase. Meanwhile, the phase of the light

passing through the normal region (region B in Fig. 5.10) is at 0°. Therefore, the electric

field of the light at the phase shifted region (A) is equal in magnitude but shifted 180° with

respect to the light at the normal region (B). At the edge of the phase shift structure, the

electric fields from the phase shifted region (A) and the clear region (B) interfere

destructively, cancelling out at the edge of the phase shift structure. The left plot in Fig. 5.10

shows that the electric field has zero magnitude at the phase edges. The intensity of the light

is given by the |E|2, the square of the magnitude of the electric field. The right plot of Fig.

5.10 shows the intensity profile. There is deep intensity nulls formed at the edges of the

phase shift structure. This is due to the destructive interference at the edges, which causes


126

the electric field at the edges to be at a minimum. When used for photolithography, the low

light intensity at the edges of the phase shift structure is not strong enough to expose the

photoresist. This forms patterns with small feature sizes.

Fig. 5.10. Destructive interference occurs at the edges of the phase shift structures [5], resulting in

lower light intensities at the edges.

Laser MLA lithography, combined with RIE, is used as a cost effective method to

fabricate periodic arrays of phase shift structures on fused silica. Firstly, laser MLA

lithography is used to direct write arbitrary patterns on photoresist. RIE is then applied to

transfer the patterns from the photoresist onto the fused silica. The RIE is controlled to etch

the fused silica to the height calculated according to Eq. 5.2. The wavelength of the light

used is 365 nm so the RIE is used to etch the fused silica to a depth of ~ 385 nm.

Subsequently, the phase shift structures are used as a phase mask for UV exposure of

photoresist. Making use of the low light intensities at the phase edges, arrays of patterns

with smaller dimensions can be formed.

Intensity plot Electric field plot

B

A


127

Figure 5.11 shows an array of „L‟ shapes structured on fused silica. Laser MLA

lithography formed the patterns and the patterns are transferred to the fused silica substrate

by the process flow illustrated in Fig. 5.9. The inset shows the 3D AFM image of the

structure, illustrating the uniform height of the structure obtained through the use of the hard

mask during RIE. This is crucial because the optimal phase shift is obtained when there is

an optical path length difference of λ/2 to introduce destructive interference.

Fig. 5.11. An array of „L‟ shapes patterned using the laser MLA lithography and RIE. The inset

shows the 3D AFM image of a „L‟ shape. The surface is uniform, showing good control of the

height uniformity.

The phase shift structures are used for an illumination wavelength of λ = 365 nm.

Using Eq. 5.2, the etched depth required is then h = λ/[2(n-1)] ≈ 385 nm, where n = 1.475 is

the refractive index of fused silica. The RIE process is carefully monitored to etch the fused

silica to the required depth. The „L‟ shape array structured onto the fused silica surface is


128

100% transparent, with a dimension of ~ 800 nm. When light passes through the structure,

the phase of the light is shifted out of phase by 180˚. Destructive interference between the 0°

and the 180° phase shifted light at regions near the edges of the structure causes a steep drop

in the light intensity. Due to this effect, when the phase shift structures are used for UV

exposure, areas at the phase edges would not expose the photoresist while other transparent

areas transmit the incident light and expose the photoresist. Figure 5.12 shows the result

obtained by using the fabricated „L‟ phase structures for UV exposure of positive

photoresist. The UV light is from a mercury arc lamp (8mW power, λ=365 nm), and the

exposure is done by contact printing. After developing the photoresist, an array of „double

L‟ lines are produced on the photoresist. They correspond to the left and right edges of the

„L‟ phase shift structure. The intensity minimum due to destructive interference at the edges

is not strong enough to expose the positive photoresist, leaving the narrow lines on the

photoresist. The optical microscope measurement shows that the L lines have a feature size

of ~ 290 nm, which is ~ 3 times smaller than the original L shape phase shift structure, and

in the order of the wavelength of the incident light, λ = 365 nm.


129

Fig. 5.12. By using the fabricated phase shift structures for UV exposure, periodic arrays of „double

L‟ are obtained.

On a closer inspection, the right L lines are broken at the perpendicular junction. To

further study and verify this result, theoretical simulation is carried out by Computer

Simulation Technology (CST) software. A similar „L‟ shape structure with the same

material properties is used in the simulation. 365 nm light is incident upon the structure and

after light passes through it, the light intensity distribution under the structure is

investigated. Figure 5.13 plots the simulated light intensity below the phase shift structure

upon irradiation of 365 nm light. There are two regions of intensity minimum, as

highlighted in the figure with dotted lines. These regions occur at the edges of the phase

shift structure and are due to the destructive interference of light. Furthermore, at the

perpendicular junction of the „L‟ shape, there is a coupling of light to the exterior light

waves. This causes the right „L‟ line to be exposed and dis-jointed at the area near the

corner, which is observed in the experimental result in Fig. 5.12. This theoretical result

290 nm


130

matches well with the experimental results presented in Fig. 5.12, where the regions of the

intensity minimum is not strong enough to expose the positive photoresist at the phase

edges, resulting in the printing of the fine double „L‟.

Fig. 5.13. Simulated result showing the regions of intensity minimum at the edges of the structures

(white dashed lines), due to the destructive interference of light.

To demonstrate the design flexibility of using the laser MLA lithography to fabricate

the phase shift structures, another phase shift design is adopted and shown in Fig. 5.14. An

array of square shape phase shift structures, each with a through-hole at the centre, is

fabricated on fused silica. The square has two edges; one exterior edge and a smaller interior

edge at the through hole area. The interior edge is ~ 2.85 µm away from the outer edge. The

inset 3D AFM image illustrates the surface of the square structure. It shows the through-


131

hole in the centre of the square. Furthermore, the surface of the square pattern is flat and

uniform in height. This is favorable for destructive interference to occur at the edges. In this

case, minimum light intensities are expected at the interior and exterior edges due to

destructive interference at those regions.

Fig. 5.14. An array of square shape phase shift structures fabricated on fused silica. The inset shows

the 3D AFM image of one of the structure. It shows the through hole at the centre and the uniform

height.

After using the fused silica structured with the square phase shift structures for the

exposure, Fig. 5.15 shows the patterns obtained. The exposure time is ~ 5 seconds. An array

of two concentric squares is obtained, with the smaller square inside the bigger square. This

result is expected as both the exterior and the interior edges of the square phase structure

experience low light intensities due to the destructive interference. This forms the two

concentric square images on the photoresist.


132

Fig. 5.15. Two concentric square patterns formed on photoresist. The smaller square is due to the

interior edge of the square phase structure while the bigger square is due to the exterior edge.

Simulation result of the square phase shift structure is shown in Fig. 5.16. The

simulation results show that there are two regions of low light intensities. One occurs at the

outer edge of the square phase shift structure while the other occurs at the inner edge. The

intensity minimum is not strong enough to expose the photoresist, yielding the results

observed in Fig. 5.15. In addition, it is noted that there is a region of enhanced light intensity

that is located at the centre of the square. The enhanced light intensity is most likely due to

the concentration of light as it passes through the through hole at the centre of the square

phase shift structure. The enhanced light intensity region is very close to the interior edge so

if the exposure time is increased, the inner square in Fig. 5.15 would be exposed, leaving

only the outer square.


133

Fig. 5.16. Simulated result showing the intensity under the square phase shift structure as light

passes through. There are two regions of low light intensities (dotted lines), one at the exterior edge

and the other at the interior edge. The through hole concentrates the light at the centre of the square.

To verify this phenomenon, the experiment is repeated with the exposure time

increased from 5 to 10 seconds. Figure 5.17 shows the patterns array obtained. An array of

single squares is obtained instead of an array of two concentric squares. In contrast to Fig.

5.15, the smaller inner square is exposed and removed after photoresist development. The

enhanced light intensity at the centre of the structure exposes the photoresist at the interior

edge more easily than the light at the exterior edge. Therefore, by controlling the exposure

time, the interior square photoresist can be deliberately exposed, leaving behind only the

exterior larger square. This demonstrates the flexibility of forming different types of

patterns by using the phase shift structures for nano-patterning and controlling the exposure

time. Furthermore, it also shows the design flexibility available by using the laser MLA


134

lithography to direct write the desired patterns, since the designs can be changed easily via

PC control.

Fig. 5.17. By increasing the exposure time, the interior smaller square patterns on photoresist can be

deliberately exposed, leaving behind only the exterior bigger square.

5.3.5. Laser MLA lithography using the fabricated MLA

The MLA fabricated in the previous chapter (Fig. 4.3) can also be adapted for the

laser MLA lithography using the same setup (as shown in Fig. 5.3, the experimental setup of

the laser MLA lithography). Figure 5.18 shows an array of „L‟ patterned by the MLA. A

feature size as small as ~ 580 nm is achieved. This feature is much smaller than the feature

sized obtained previously (~ 1 µm in Fig. 5.8). This shows that the fabricated MLA, which

has smaller microlenses, can achieve higher resolution patterning.


135

Fig. 5.18. An array of „L‟ patterned by laser MLA lithography using the fabricated MLA. The

feature size is ~ 580 nm.

In the next example, an array of three successive „L‟s is drawn. Figure 5.19 shows

the patterning of the three „L‟s. After patterning the first „L‟, the nanostage is moved in both

the X and Y directions by 600 nm and then the second identical „L‟ is patterned. The third

„L‟ is patterned in the similar way, and three parallel „L‟s are finally fabricated. The space

between each individual „L‟ can distinguished after the patterning. Therefore, this shows

that the dimensions of each individual „L‟ are smaller than 600 nm. The optical

measurement shows that the feature size of the „L‟ is ~ 440 nm. By controlling the exposure

dose and exposure time, the laser MLA lithography can be optimized to print smaller

patterns using the fabricated MLA.


136

Fig. 5.19. An array of 3 „L‟s patterned parallel to one another. The minimum feature size is ~ 440

nm. The space between the „L‟s can be distinguished clearly.

In summary, this chapter shows the parallel and direct writing capabilities of the

laser MLA lithography technique. Various arbitrary shapes can be fabricated using this

technique. Using the MLA fabricated in the previous chapter, patterns with feature sizes ~

440 nm can be obtained. Furthermore, the technique is used to fabricate an array of phase

shift structures on fused silica. Using the patterned fused silica substrate as a phase mask,

patterns with feature sizes ~ 290 nm can be obtained. The laser MLA lithography technique

has been demonstrated to be a very versatile technique for the parallel micro-/nano-

patterning of large area periodic structures.


137

5.4. Reference

1. W. K. Choi, T. H. Liew and M. K. Dawood, H. I. Smith, C. V. Thompson and M. H.

Hong, “Synthesis of Silicon Nanowires and Nanofin Arrays Using Interference

Lithography and Catalytic Etching”, Nano. Lett. 3799-3802 (2008).

2. Y. Lin, M. H. Hong, T. C. Chong, C. S. Lim, G. X. Chen, L. S. Tan, Z. B. Wang and L.

P. Shi, “Ultrafast-laser-induced parallel phase-change nanolithography,” Appl. Phys.

Lett. 89, 041108 (2006).

3. Z. C. Chen, M. H. Hong, C. S. Lim, N. R. Han, L. P. Shi and T. C. Chong, “Parallel

laser microfabrication of large-area asymmetric split ring resonator metamaterials and

its structural tuning for terahertz resonance,” Appl. Phys. Lett. 96, 181101 (2010).

4. C. Y. Sin, B. H. Chen, W. L. Loh, J. Yu, P. Yelehanka, A. See and L. Chan, "Resist

trimming in high-density CF4/O2 plasmas for sub-0.1 µm device fabrication,” J. Vac.

Sci. Technol. B 20, 1071-1023 (2002).

5. K. K. H. Toh, G. Dao, R. Singh and H. Gaw, “Chromeless phase-shifted masks: A new

approach to phase-shifting masks,” Proc. SPIE 1496, 27-53 (1991).

6. L. Bauch, J. Bauer, H. Dreger, B. Lauche, G. Mehlib and St. Rothe, “A New Chromeless

Phase Mask for the Photolithography,” Microelectron. Eng. 17, 87-91 (1992).

7. J. F. Chen, J. S. Petersen, R. Socha, T. Laidig, K. E. Wampler, K. Nakagawa, G.

Hughes, S. MacDonald and W. Ng, “Binary Halftone Chromeless PSM Technology for

λ/4 Optical Lithography,” Proc. SPIE 4346, 515-533 (2001).

8. M. Fritze, J. M. Burns, P. W. Wyatt, D. K. Astolfi, T. Forte, D. Yost, P. Davis, A. V.

Curtis, D. M. Preble, S. Cann, S. Denault, H. Y. Liu, J. C. Shaw, N. T. Sullivan, R.


138

Brandom and M. E. Mastovich, “Application of chromeless phase-shift masks to sub-

100 nm SOI CMOS transistor fabrication,” Proc. SPIE 4000, 388-407 (2000).

Chapter 6: Conclusions and future work

139

CHAPTER 6

CONCLUSIONS AND FUTURE WORK

6.1 Conclusions

The main theme of this thesis focuses on different methods of glass processing to

fabricate micro- and nano-structures on glass. The thesis starts by discussing how high

quality glass processing can be achieved by laser direct writing with the assistance of

absorbing liquids. It then describes how parallel processing of glass can be carried out

through the use of laser interference lithography and micro-lens array lithography. The

following summarize and conclude the research contributions of this thesis:

1. Laser direct writing of glass with the assistance of liquid absorbers is demonstrated.

The LIBWE process is studied with the use of a UV laser (λ=355 nm). The effect of

the absorbing liquid on the etching rate is analyzed. Through the analysis of the

absorption spectra, a relationship between the laser fluence and the etch depth is

established. It is found that by using a mixture of pyrene/toluene, a higher absorption

of the laser is obtained. This increases the etching rate and lowers the etching

threshold. By-products are formed during the LIBWE process and analyzed. Laser

dry cleaning method is demonstrated to be able to remove the by-products. This

method is proposed for in-situ cleaning of the surfaces after LIBWE. It is shown that

the LIBWE process can produce very high quality micro-fluidics channels. By the


140

use of a high speed galvanometer, good quality micro-structures can be produced

rapidly.

2. The principle of LIBWE is further extended to the use of an inorganic liquid, CuSO4,

and an infra-red solid state laser with a wavelength of 1064 nm. The absorption

spectrum of CuSO4 shows that it absorbs strongly in the near infra-red wavelengths.

Therefore, by using CuSO4 solution the laser can be used to process the glass

substrates. This is advantageous because the laser is more energy efficient and less

costly. The laser is also widely used in the industry so there are potential industrial

applications for this novel process. Metal deposition is observed from the process

and is characterized by XPS analysis to be copper metal deposition. Based on the

XPS analyses, the mechanism for the etching process is studied in detail to explain

the etching process. Using this novel process, miniature structures are diced out from

the glass substrates.

3. Moving from a serial direct laser writing mode to a parallel large area processing

mode, laser interference lithography is used to pattern quartz with an array of

nanoholes. A process flow is devised to transfer the patterns from the photoresist

onto the quartz substrate with good fidelity. High quality nanoholes array is obtained

uniformly over a large area. By controlling the etching parameters, the processed

quartz substrate can be used as a phase mask to replicate arrays of nanoholes by a

single UV exposure. Defects can also be introduced onto the phase mask so that the

corresponding defects are reproduced on the photoresist upon a single exposure. This


141

has potential applications in photonic crystal applications. Simulations are carried

out and the simulation results match the experimental results very well.

4. Microlens array lithography is employed to further add flexibility to the parallel

processing of glass substrates. The technique is capable of massively parallel direct

writing of arbitrary patterns. Using this technique, various arbitrary shapes are

successfully patterned on glass substrates. The shapes are then functionalized as

phase shift structures by using reactive ion etching to etch down to a depth that

achieved a phase shift of 180°. The phase shift structures are characterized

experimentally. Using the phase shift structures, arrays of smaller patterns are

fabricated, with features sizes ~ 290 nm, which is ~ 3 times smaller than the phase

shift structures. Different shapes of phase shift structures are also demonstrated and

it is shown that by controlling the exposure time, different sets of results can be

obtained from the same phase shift structures. This demonstrates the flexibilities in

the design and pattern process. Simulation results match the experimental results

very well. The fabricated MLA has also been demonstrated for laser MLA

lithography.

5. Micro-optics is fabricated on glass substrates using the photoresist reflow method.

The method is also used as a novel process to fabricate micro-solid immersion lens

array on glass. It is used successfully to enhance the resolution of a MLA.


142

6.2 Recommendation for future work

A few possible future work are suggested as follows:

1. Some researchers have demonstrated that laser interference can be used in

combination with the LIBWE process to produce high quality gratings via a two

beam interference scheme. However, a four beam interference scheme has not been

investigated much so far. It is believed that many other interesting patterning

capabilities can be realized by using four beam interference in combination with

LIBWE. For instance, nanodots and nanorods arrays can be fabricated directly on

glass substrate using a four beam interference scheme.

2. The introduction of defects in the array of nanoholes has been demonstrated

experimentally as planar 2D patterns. However, as shown by the simulation results,

the phase mask with the array of nanoholes exhibits a 3D light distribution. This can

be exploited for 3D photonic crystals structuring. The effect of 3D defects

engineering can also be studied in more details.

3. Microlens array lithography patterns a whole array of structures simultaneously in a

single exposure. However, sometimes it might be desirable to have more control

over which microlens is ‘on’ and ‘off’. By having the ability to address each

microlens, it is possible to select specific areas where patterns are not created. A

possible way to implement this scheme is to exploit the polarization of the laser light.

A layer of liquid crystal film could be coated on the back of the microlens array and

addressed by a matrix grid, in a similar fashion to that of a memory cell array. The


143

microlens can then be turned ‘on’ and ‘off’ by switching the polarization of the

liquid crystal cell. Meanwhile, the array of solid immersion lens currently fabricated

has a hemispherical configuration. Using the same fabrication technique, a super-

hemispherical micro-solid immersion lens array can be fabricated. It is expected that

the resolution can be further improved by a factor up to n. Furthermore, higher

refractive materials can also be used to increase the resolution since the enhancement

is related to the refractive index. Lastly, the practical applications of the μSIL system

should be explored through the collaboration with other industrial partners.

Appendix

144

APPENDIX

1. LIBWE

This section shows the effect of the concentration of pyrene/toluene mixture on the

relationship between the etch depth and the laser fluence. The following 3 graphs plot the

etched depth versus the laser fluence for 3 different concentrations of pyrene dissolved in

toluene; 0.2 M, 0.4 M and 0.6 M. A linear line can be used to best fit the points. (A power

relationship is not directly observable here because the laser fluence used is not extended to

the region of the threshold fluence.) From the comparison of the 3 graphs, it shows that the

pyrene/toluene solution to a concentration of 0.4 M produced the steepest gradient. This

implies that the optimal etching rate is obtained at a pyrene/toluene concentration of 0.4 M.

No additional benefits in terms of etching rate are obtained by increasing the concentration

to 0.6 M.

Appendix

145

Etched depth vs fluence using Pyrene in Toluene 0.2M

y = 2895.6x - 2344.4

0

1000

2000

3000

4000

5000

0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4

Fluence (J/cm2)

Etc

hed

dep

th (

nm

)

Laser pulse energy: 175uJ

Repetition rate: 1000Hz

Scanning rate: 100mm/min

Etched depth vs fluence by varying height using Pyrene

in Toluene 0.4M

y = 3724.2x - 2848

0

1000

2000

3000

4000

5000

6000

0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1

Fluence (J/cm2)

Etc

hed

dep

th (

nm

)




Appendix

146

Etched depth vs fluence at different heights using

Pyrene in Toluene 0.6M

y = 3062.7x - 2007.4

0

500

1000

1500

2000

2500

3000

3500

4000

4500

0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1

Fluence (J/cm2)

Etc

hed

dep

th (

nm

)




2. MLA and μSIL profile control

This section discusses how the required radius of curvature of a MLA or μSIL is

obtained. Referring to the table below, the required radius of curvature is first identified.

Next, the radius r of the lens is determined. This radius is the same as the radius of the

circular chrome mask used to fabricate the photoresist island (refer to Fig. 4.1 for the

schematic process flow). Using Eq. 4.2, the sag height s can then be found. This in turn

gives the volume of the segment of sphere, which is the volume of photoresist reflowed

from the photoresist island. During the photoresist melt process, there is solvent evaporation

Appendix

147

from the photoresist, which results in a reduction of volume. This reduction of volume is

input as a percentage, and is found to be about 20% in the experiments. Calculating

backwards, the volume of photoresist before reflow is then obtained and this volume

corresponds to the volume of a photoresist cylinder. Since the volume of a cylinder is:

trVcyl

2 ,

t, the height of the cylinder, can be found.

Radius of curvature

radius r

Sag height

s

Vol. of segment

of sphere

Reduction of photoresist

volume

Vol. of cylinder

Height of

cylinder t

190 36 3.25 6634.17 0.2 8292.71 2.04

Therefore, through this calculation process, the required radius of curvature can be

obtained by getting the required height of the photoresist cylinder. This is controlled by the

photoresist spin-coating process and can be easily fine-tuned during the coating procedure.

Using this calculation process, the required radius of curvature R = 190 μm is obtained in

section 4.4.

Publications

Journals

1. Z. Q. Huang, M. H. Hong, K. S. Tiaw, and Q. Y. Lin, Quality glass processing by

Laser Induced Backside Wet Etching, Journal of Laser Micro/Nanoengineering 2,

194-199 (2007).

2. Z. Q. Huang, M. H. Hong, T. B. M. Do. and Q. Y. Lin, Laser ablation of glass

substrates using 1064 nm IR radiation, Applied Physics A, 93 159-163 (2008).

3. Z. Q. Huang, Q. Y. Lin and M. H. Hong, Phase Shift Mask Fabrication by Laser

Microlens Array Lithography for Periodic Nanostructures Patterning, Journal of

Laser Micro/Nanoengineering 5, 233-237 (2010).

Conference Proceedings

1. M. H. Hong, F. Ma, C.S. Lim, Y. Lin, Z.Q. Huang, L.S. Tan, L.P. Shi and T.C.

Chong, “Multi-lens Array Fabrication and Its Applications in Laser Precision

Engineering”, 8th International Symposium on Laser Precision Microfabrication,

Vienna, Austria, 24-28 April, 2007.

2. M. H. Hong, Z. Q. Huang, Y. Lin, J. Yun, L. S. Tan, L. P. Shi, T. C. Chong, “Laser

Precision Engineering From Microfabrication To Nanoprocessing”, Conference on

Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference, San

Jose, CA, 04-09 May, 2008.

3. Z. Q. Huang, M. H. Hong, Q. Y. Lin, “Chromeless phase shift mask fabrication by

laser interference lithography”, 10th International Conference on Laser Ablation,

Singapore, 22-27 November, 2009.

4. Z. Q. Huang, M. H. Hong, Q. Y. Lin, “Phase shift mask fabrication by laser

microlens array lithography for periodic nanostructuring patterning”, 11th

International Symposium on Laser Precision Microfabrication, Stuttgart, Germany,

7-10 June 2010. (A#53, ON-049)

laser microprocessing and fabrication of structures … · mechanical properties 4 1.2. methods for...

Documents