Development of nanosphere lithography and its applications 2

1.6 Nanosphere lithography used to create various nanostructures with a negatively charged functional group, such as carboxylate or sulfate that is electrostatically repelled by the nega

Chapter 1 Introduction

1.1 Introduction

Nanostructures and nanosystems have attracted much attention in modern science and technology due to their unique physical and chemical properties, which results in not only improving performance of current devices and processes but also potentially generating many new applications in physical [1, 2], chemical [3, 4] and bio technologies [5, 6]

Since the introduction of the theory of quantum mechanics in the twenty of last century, it is a well known fact, that objects change their behavior if they approach a certain lower size limit Below this size limit, certain energies are allowed or forbidden, sharp steps in energy spectra arise and the general physics of objects change as shown in Fig.1.1 This quantum effects not only greatly improve some device performance, such as high speed transistors [7] but also result in some totally new devices like the semiconductor laser [8] or the superconducting nanowires [9]

Fig 1.1 Quantum effects of matter

So far it has been demonstrated that nanostructures exhibit particularly peculiar and interesting characteristics, for example: quantized excitation [8], Coulomb blockade [10], single-electron tunneling [11], and metal-insulator transition [12] These phenomena occur in structures small enough for quantum mechanical effects to dominate

Besides the quantum effects, new phenomena result from the nanostructures or nanosystems also occur and applied in physical, chemical and biotechnologies For example, ballistic movement of an electron in a semiconductor [13], near- and far-field diffraction of visible light [14], diffusion of an active species close to an electrode [15], excitation of collective resonance by light [16] Fabrication and study

of these systems have become active areas of research in physics, material science, chemistry and biology

In addition, photonic crystals are periodic optical nanostructures that affect the motion of photons in much the same way that ionic lattices affect electrons in solids Combining the quantum dots and the photonic crystals has already attracted much attention [17-20]

However, how to fabricate such periodic nanostructures efficiently is still a big challenge

1.2 Review of nanofabrication technologies

The ability to fabricate structures from the micro- to the nanoscale with high precision in a wide variety of materials is of crucial importance to the advancement of micro- and nanotechnology and the nanosciences The semiconductor industry has been pushing high-precision nanoscale lithography to manufacture ever-smaller transistors and higher-density integrated circuits (ICs) Critical issues, such as resolution, reliability, speed, and overlay accuracy, all need to be addressed in order

to develop new lithography methodologies for such demanding, industrially relevant

processes On the other hand, less stringent conditions are found in many other areas, for example, photonics, micro- and nanofluidics, chip-based sensors, and most biological applications Beside traditional lithographical technologies, such as photolithography, e-beam lithography, several alternative approaches towards nanostructure fabrication have been exploited in the past 15 years These techniques include microcontact printing (or soft lithography) [21], nanoimprint lithography (NIL) [22], scanning-probe-based techniques (e.g., atomic force microscope lithography) [23], dip-pen lithography [24], and nanosphere lithography In the thesis,

I will focus on the development of nanosphere lithography (NSL), and its various applications, such as fabricating surface nanostructures, forming templates and growing nanostructures through the templates

Basically, lithography is a chemical process to pattern parts of a thin film or the bulk of a substrate These patterns can be formed on a mask and be transferred to other thin film which can be used to form various small devices, such as integrated circuit, MEMS and small devices including light emitting diodes Traditional lithographic techniques include photolithography [25], in which light is used as an energy source to change the photoresist; e-beam lithography [26], in which electrons are used to change the chemical properties of the resist Recently various lithography techniques, for example, nanoimprinting [22], and nanosphere lithography [27], have been developed to overcome the problems arising from traditional lithographic techniques

1.2.1 Lithography with photons

The well known lithography is photolithography, which uses light to transfer a geometric pattern from a photomask to a light-sensitive chemical "photoresist" on the substrate After development, a deposition or etch process is applied to form the pattern on a film or a substrate as shown in Fig 1.2

Fig 1.2 Basic outline of optical lithography processes The diagram shows the optical radiation entering the system, which is then filtered by the chromium mask The image is then projected on to the resist, and any non-exposed material is removed during developing

Photons have been used for many years to induce chemical reactions in photographic materials or resist polymers The lithographic technology is an invaluable tool for micro-fabrication in a broad range of applications in science and technology and one of the most widely used and highly developed technologies now practiced [25] In this process, the mask is placed in physical contact with, or in close proximity to, the resist Most fabrication in the integrated circuit industry uses such lithography

Photolithography is a fast technique to form patterns due to its large area expose The minimum feature size that can be obtained by this process is primarily determined by diffraction that occurs as light passes through the gap between the mask and the resist Even with the use of elaborate vacuum systems to pull the mask and substrate together, it is still difficult in practice to reduce the gap between a

conventional rigid mask and a rigid flat substrate to less than ~1 µm over large areas

As a result [28], the resolution of contact mode photolithography is typically 0.5-0.8

µm when UV light (360-460 nm) is used The resolution of photolithography

increases as the wavelength of the light used for exposure decreases Feature sizes of

250 nm can be obtained when 248 nm UV light is used However, there are big problem with transparency of optical parts when wavelength of light is further decreased [29]

Although photolithography was demonstrated with soft EUV and X-rays many years ago, to fabricate the masks and optics capable of supporting a robust, economical method still provides significant unsolved challenges [30]

1.2.2 Lithography with Particles

will absorb them From the quantum mechanical principle of wave-particle duality [31] and the de Broglie equation , it is found that an electron with an energy of 10 keV has a wavelength of around 12 pm This obviously represents a huge reduction in wavelength compared to X-Ray radiation, and therefore electron beam lithography has the possibility at a better resolution than any of the electromagnetic methods previously considered

Electron beam lithography replaces the photons with an electron beam, and utilizes a different system with image formation between the source and the resist with no mask in the system (Fig 1.3)

Because of using a beam of electrons, whose direction can be controlled by a magnetic field, there is no need for a mask in the lithographic system A computer controls the strength of the magnetic field, whilst there is very little diffraction from the electrons, so the patterns produced on the resist are extremely accurate, even though it suffers from scattering in the resist Less than 10 nm features has been obtained by this technique [32] However, electron beam lithography still accounts a big problem The system has a very low throughput due to its series nature of writing,

only one point on the resist can be exposed at any given time

1.2.3 Nanoimprinting

The principle of nanoimprinting is very simple Figure 1.4a shows a schematic

of the originally proposed NIL process [33, 34]

A hard mold that contains nanoscale surface-relief features is pressed into a polymeric material cast on a substrate at a controlled temperature and pressure, thereby creating a thickness contrast in the polymeric material A thin residual layer

of polymeric material is intentionally left underneath the mold protrusions, and acts

as a soft cushioning layer that prevents direct impact of the hard mold on the substrate and effectively protects the delicate nanoscale features on the mold surface

Fig.1.4 (a) Schematic of the originally proposed NIL process (b) Scanning electron microscopy (SEM) image of a fabricated mold with a 10 nm diameter array (c) SEM image of hole arrays imprinted in poly(methyl methacrylate) by using such a mold [34]

Advantages of the NIL are that it demonstrated ultrahigh resolutions soon after its inception Figures 1.4 b and c show SEM images of a mold with a pillar array (pillar diameter 10 nm) and an imprinted 10 nm hole array in poly(methyl methacrylate) (PMMA) that were obtained almost a decade ago [35] NIL is inherently high-throughput, because of parallel printing, and it requires only a simple equipment, leading to low-cost processes A variation of the NIL technique that uses

a transparent mold and UV-curable precursor liquid to define the pattern flash imprint lithography) has been demonstrated [36], allowing the process to be carried out at room temperature and making it attractive for IC semiconductor device manufacturers

(step-and-However, it still has some challenges in meeting the stringent requirements of various applications, such as mold fabrication, mold surface preparation, NIL resist, and residual layer problem [34]

1.3 Nanosphere lithography

Nanosphere lithography (NSL) is an inexpensive, simple to implement, inherently parallel, high throughput general nanofabrication technique capable of producing an unexpectedly large variety of nanostructures and well-ordered 2D nanostructural arrays

The inception of “natural lithography” dates from the seminal work of Fischer and Zingsheim with the introduction of “naturally”-assembled polystyrene latex nanospheres as a mask for contact imaging with visible light in 1981 [37] In 1982, Deckman and co-workers greatly extended the scope of Fischer’s approach by demonstrating that a self-assembled nanosphere monolayer could be used as both a material deposition and etch mask Deckman coined the term “natural lithography” to describe this process Deckman and co-workers continued to explore various fabrication parameters and possible applications of natural lithography but always employed a single layer (SL) of nanospheres as the mask [38] The third stage in the evolution of natural lithography, renamed nanosphere lithography (NSL) to be more operationally descriptive, is represented by the work of Van Duyne et al [39], who extended the SL methodology with (1) the development of a double layer (DL) nanosphere mask, (2) atomic force microscopy (AFM) studies of SL and DL periodic particle arrays (PPAs) of Ag on mica, and (3) fabrication of defect-free SL and DL PPAs of Ag on mica with areas of 10-100 µm2 that were large enough to permit microprobe studies of nanoparticle optical properties [39] As an efficient nanofabrication technique, NSL is being used in laboratories around the world to

study the size-dependent optical, magnetic, electrochemical, thermodynamic, and catalytic properties of materials

Fig 1.5 (a) side and (b) top-views of self-assembly of nanospheres

It has been demonstrated that the self-assembly process to form 2D ordered arrangement of the nanospheres starts from a nucleation [40] The nucleus formation

is governed by attractive capillary forces appearing between spheres partially immersed in a liquid layer Then, crystal growth occurs through convective particle flux caused by the water evaporation from the already ordered array (Fig 1.5) In principle, the monolayer of hexagonally close-packed (hcp) spheres can be used as a mask to form nanoparticles by depositing other materials through the holes between the spheres Actually, nanostructures can also be created on the substrate by dry etching or infiltrating process as seen in Fig 1.6

Methods to deposit a nanosphere solution onto the desired substrate include spin coating [41], drop coating [42], template-directed growth [43], angled cooling plates and Langmuir-Blodgett techniques [44] All these deposition methods require that the

nanospheres are able to freely diffuse across the substrate, seeking their lowest energy configuration This is often achieved by chemically modifying the nanosphere surface

Fig 1.6 Nanosphere lithography used to create various nanostructures

with a negatively charged functional group, such as carboxylate or sulfate that is electrostatically repelled by the negatively charged surface of a substrate such as mica

or glass Following the self-assembly of the nanosphere mask, a metal or other material is then deposited by thermal evaporation, electron beam deposition (EBD),

or pulsed laser deposition from a source normal to the substrate through the

nanosphere mask to a controlled mass thickness dm After the metal deposition, the nanosphere mask is removed by sonicating the entire sample in a solvent, leaving behind the material deposited through the nanosphere mask on the substrate

In the simplest NSL case, only a monolayer of hcp nanospheres with a diameter

of D is self-assembled onto the substrate When one deposits metal through the

monolayer mask, the three-fold interstices allow deposited metal to reach the

substrate, creating an array of triangular shaped nanoparticles with P 6mm symmetry

[Figs 1.7(a) and (b)] Simple geometric calculations define the relationship between

the perpendicular bisector of the triangular nanoparticles a and the interparticle spacing d ip to the nanosphere diameter D as shown in Fig 1.7(e)

(1)

(2)

When a second layer of nanospheres assembles onto the first, every other fold hole is blocked, and a smaller density of six-fold interstices results After depositing metal through the double layer (DL) nanosphere mask, a regular pattern of hexagonal nanoparticles forms on the substrate [Figs 1.7(c) and (d)] This array is referred to be a DL PPA As in the SL PPA, the size of the DL PPA nanoparticles can

three-be tuned by the deposited nanosphere size and the d m Similarly as illustrated in Fig 1.7(f),

(3)

(4)

There are two possible outcomes when a third layer of nanospheres assembles onto a DL: (1) if the nanospheres pack in an ABAB sequence, the regular pattern of hexagonal holes remains or (2) if the nanospheres pack in an ABCABC sequence, all the mask holes are blocked Therefore, a nanosphere mask containing three or more layers in the ABCABC stacking sequence does not allow any deposited materials to reach the substrate

Beside nanodots, other nanostructures can also be formed through the NSL, for example nanoring [39] During deposition of nickel by EBD, nanorings have been observed as shown in Fig 1.8 In this case, it seems that bimodal kinetic energy distribution of gas-phase atoms are produced by EBD Low kinetic energy (~0.1 eV) atoms that travel along a direct line of sight from the EBD target to the substrate and stick where they strike the substrate form the triangular nanodots In contrast, high kinetic energy (~1-10 eV) atoms that travel along off-normal trajectories strike the

substrate within the “footprint” of the triangular interstices of the nanosphere mask, and then continue to travel, because of excess kinetic energy, adhering to the substrate underneath the nanosphere to form the nanorings These nanorings can also

be used as masks to create thin-walled nanocylinders

Fig 1.7 Schematic illustration (a) and representative AFM image (b) of SL PPA The

AFM image was captured from a SL PPA fabricated with D = 542 nm nanospheres and d m = 48 nm thermally evaporated Ag metal after removing the nanospheres; Schematic illustration (c) and representative AFM image (d) of DL PPA The AFM

image was captured from a SL PPA fabricated with D = 400 nm nanospheres and d m=

30 nm thermally evaporated Ag metal after removing the nanospheres [39] (e) and (f)

show the definition of the parameters of D, a and d ip for single and double layer arrangement, respectively

In addition, some modified NSLs, such as Angle-Resolved Nanosphere Lithography, have also been developed to form unique shaped nanostructures [46] For example, they are nanooverlaps, nanogaps and nanoparticle chains as seen in Fig 1.9

In the aforementioned nanoparticle architectures, all materials were deposited from a collimated source along a line perpendicular to the plane of the nanosphere mask A new class of NSL structures has been fabricated by varying the angle between the nanosphere mask and the beam of material being deposited The size and shape of the three-fold interstices of the nanosphere mask change relative to the

deposition source as a function of , and accordingly, the deposited nanoparticles’ shape and size are controlled directly by (Figure 1.9)

Fig 1.8 Schematic illustration (a) and representative AFM image (b) of nanoring and

SL PPA fabrication The AFM image was captured from a sample fabricated with D

= 979 nm nanospheres and d m = 50 nm e-beam deposited Ni metal after removing the nanospheres [39]

Although the nanostructures created through the NSL have many applications in nanophotonics, catalysis, and biotechnologies [39-46], as in all naturally occurring crystals, nanosphere masks include a variety of defects that arise as a result of nanosphere polydispersity, site randomness, point defects (vacancies), line defects (slip dislocations), and polycrystalline domains Especially, it is difficult to obtain a single layer over a large area, which limits its applications because nanostructures created by different layer(s) are totally different Physical and chemical properties of the nanostructures depend seriously on the shapes Most nanostructures created by the NSL are hexagonally arranged and controlling the shape of the nanostructures in the plane and in the vertical direction is also a big challenge

Fig 1.9 Schematic of the angle resolved deposition process (a) Samples viewed at 0° (a), 30°, (b) and 45°, (c), respectively [46]

1.4 Motivation and objectives

Although nanosphere lithography (NSL) has been recognized as an inexpensive, high throughput, and flexible technology to fabricate nanostructures, there are two main problems limit its further applications - 1) how to obtain pure single or double layer of the nanospheres over a large area It is known that nanostructures fabricated through single and double layered arrays of the nanospheres are totally different So this technology is hardly to use in device applications if a pure single or double layered array of the nanospheres cannot be obtained in a device level 2) how to control the shapes of the nanostructures fabricated through the NSL So far, only nanostructures with limited shapes can be fabricated by the NSL due to the nature of the spherical particles and its 2D hexagonal arrangement of the nanospheres So, the objectives of my work are:

1 To develop techniques to overcome above mentioned problems, which include: controlling arrangement of the nanospheres in areas as larger as the device required; controlling shapes of the nanostructures fabricated by the NSL in both vertical and lateral directions

2 To investigate applications of the nanostructures, as templates, fabricated by the NSL, which include: MOCVD nano-growth of III-V compound; fabricating metal nanostructures and applying nanostructures created by NSL to enhance LED performances

1.5 Scope of thesis

Chapter 1 summarizes some technologies for nanofabrication The working principle, advantages and disadvantages of nanosphere lithography are discussed Focusing on overcoming the weaknesses of nanosphere lithography, a method for obtaining either a single layer or double layer of nanospheres on an entire area as large as a light emitting diode (LED), is described in Chapter 2 Some techniques, such as multi-cycle-etching, used to create 3D surface nanostructures; 3D masks, used to fabricate more complex nanostructures and a one step method to fabricate nanostructures with multi-size features, have been demonstrated in this chapter In addition, RIE behaviors in controlling shapes of nanostructures in vertical and lateral directions are also systemically investigated in the chapter 2

Some nanostructures created by methods discussed in Chapter 2 are used as templates to investigate nano-epi-growth of semiconductor nanostructures, which are discussed in Chapter 3 An ordered array of InGaAs/GaAs nanobars has been obtained by MOCVD through a SiO2 template created by nanospheres lithography Especially, combining photolithography, such template can be formed so that the nanobars can be grown on selected regions, or such arrays of InGaAs/GaAs nanobars with multi-sized features can be grown in one step MOCVD In addition, growth of GaN films on Si nanopillars formed on a Si substrate is also presented in the chapter

3

In Chapter 4, metal nanostructures created by the templates described in Chapter

2 are presented together with their optical properties The nanostructures are ordered

3D Au nanostructures formed through a 2D polymer nanosphere array and ordered

Ag nanoparticles with excellent uniformity formed by a template guided annealing process Mechanism of forming the uniform, well arranged Ag nanoparticles is also discussed in this chapter

In Chapter 5, nanostructures created through nanosphere arrays are applied to LEDs to enhance light extraction Red LEDs grown on a GaAs substrate and blue LEDs grown on a sapphire substrate are used In addition, Au honeycomb nanostructures fabricated through nanosphere lithography on a red LED is also presented

In Chapter 6, I have summarized the results obtained mainly from the developed techniques

[2] P Bhattacharya, S Ghosh, A D Stiff-Roberts, “Quantum Dot Opto-electronic Devices”, Annu Rev Mater Res 34, 1 (2004)

[3] Haiping M Chen, Lin Pang, Andrew King, Grace M Hwang and Yeshaiahu Fainman, “Plasmonic coupled nanotorch structures leading to uniform surface enhanced Raman scattering detection”, Nanoscale 4, 7664 (2012)

[4] Paul L Stiles, Jon A Dieringer, Nilam C Shah, and Richard P Van Duyne,

“Surface-Enhanced Raman Spectroscopy”, Annu Rev Anal Chem 1, 601 (2008)

[5] Marinus A Otte, M.-Carmen Estevez, Laura G Carrascosa, Ana B Guerrero, Laura M Lechuga, and Borja Sepulveda, “Improved Biosensing Capability with Novel Suspended Nanodisks”, J Phys Chem C 115, 5344 (2011)

Gonzalez-[6] Mrinmoy De, Partha S Ghosh, Vincent M “Applications of Nanoparticles in Biology”, Adv Mater 20, 4225 (2008)

[7] T Ashley, L Buckle, S Datta, M.T Emeny, D.G Hayes, K.P Hilton, R Jefferies, T Martin, T.J Phillips, D.J Wallis, P.J Wilding and R Chau,

“Heterogeneous InSb quantum well transistors on silicon for ultra-high speed, low power logic applications”, Electronics Letters 43, 0013-5194 (2007)

[8] S Fafard, K Hinzer, S Raymond, M Dion, J McCaffrey, Y Feng, S Charbonneau, “Red-Emitting Semiconductor Quantum Dot Lasers”, Science 274,

1350 (1996)

[9] S Dubois, A Michel, J P Eymery, J L Duvail and L Piraux, “Fabrication and properties of arrays of superconducting nanowires”, Journal of Materials Research 14, 665 (1999)

[10] Effendi Leobandung, Lingjie Guo, Yun Wang, and Stephen Y Chou,

“Observation of quantum effects and Coulomb blockade in silicon quantum-dot transistors at temperatures over 100 K”, Appl Phys Lett 67, 938 (1995)

[11] L I Glazman, “Single Electron Tunneling”, Journal of Low Temperature Physics 118, 247 (2000)

[12] Evgheni Strelcov, Yigal Lilach and Andrei Kolmakov, “Gas Sensor Based on Metal−Insulator Transition in VO2 Nanowire Thermistor”, Nano Lett 9, 2322 (2009)

[13] Martin J Stevens, Arthur L Smirl, R D R Bhat, Ali Najmaie, J E Sipe, and H

M van Driel, “Quantum Interference Control of Ballistic Pure Spin Currents in Semiconductors”, Phys Rev Lett 90, 136603 (2003)

[14] Dmitri K Gramotnev & Sergey I Bozhevolnyi, “Plasmonics beyond the diffraction limit”, Nature Photonics 4, 83 (2010)

[15] Ying Wang, Jin Lu, Longhua Tang, Haixin Chang and Jinghong Li, “Graphene Oxide Amplified Electrogenerated Chemiluminescence of Quantum Dots and Its Selective Sensing for Glutathione from Thiol-Containing Compounds”, Anal Chem 81, 9710 (2009)

[16] J Alan Creighton, Christopher G Blatchford and M Grant Albrecht, “Plasma resonance enhancement of Raman scattering by pyridine adsorbed on silver or gold sol particles of size comparable to the excitation wavelength”, J Chem Soc., Faraday Trans 2, 790 (1979)

[17] Peter Lodahl, A Floris van Driel, Ivan S Nikolaev, Arie Irman, Karin Overgaag, Daniël Vanmaekelbergh & Willem L Vos, “Controlling the dynamics of spontaneous emission from quantum dots by photonic crystals”, Nature 430, 654 (2004)

[18] Susumu Noda, Masayuki Fujita & Takashi Asano, “Spontaneous-emission control by photonic crystals and nanocavities”, Nature Photonics 1, 449 (2007) [19] Yurii A Vlasov, Nan Yao, David J Norris, “Synthesis of Photonic Crystals for Optical Wavelengths from Semiconductor Quantum Dots”, Adv Mater 11, 165 (1999)

[20] G C La Rocca, “Organic photonics: Polariton lasing”, Nature Photonics 4, 343 (2010)

[21] Ravi S Kane, Shuichi Takayama, Emanuele Ostuni, Donald E Ingber, George M Whitesides, “Patterning proteins and cells using soft lithography”, Biomaterials

[24] R D Piner, J Zhu, F Xu, S Hong, C A Mirkin, “" Dip-pen" nanolithography”, Science 283, 661 (1999)

[25] W M Moreau, “Semiconductor Lithography: Principles and Materials”, Plenum, New York, (1988)

[26] R.E Fontana, J.Katine, M.Rooks, R.Viswanathan, J.Lille, S.MacDonald, E.Kratschmer, C.Tsang, S.Nguyen, N.Robertson, P.Kasiraj, “E-beam writing: a next-generation lithography approach for thin-film head critical features”, Magnetics, IEEE Transactions on 38, 95 (2002)

[27] J C Hulteen, R P Van Duyne, “Nanosphere lithography: A materials general fabrication process for periodic particle array surfaces”, J Vac Sci Technol A

13, 1553 (1995)

[28] John E Bjorkholm, “EUV Lithography—The Successor to Optical Lithography?”, Intel Technology Journal Q3’98

[29] Takashi Ito & Shinji Okazaki, “Pushing the limits of lithography”, Nature 406,

[33] S Y Chou, P R Krauss, P J Renstrom, “Imprint of sub‐25 nm vias and trenches in polymers”, Appl Phys Lett 67, 3114 (1995)

[34] L Jay Guo, “Nanoimprint Lithography: Methods and Material Requirements”, Adv Mater 19, 495 (2007)

[35] Stephen Y Chou, Peter R Krauss, Wei Zhang, Lingjie Guo, and Lei Zhuang,

“Sub-10 nm imprint lithography and applications”, J Vac Sci Technol B 15, 2897(1997)

[36] T Bailey, B J Choi, M Colburn, M Meissl, S Shaya, J G Ekerdt, S V Sreenivasan, and C G Willson, “Step and flash imprint lithography: Template surface treatment and defect Analysis”, J Vac Sci Technol B 18, 3572 (2000) [37] U C Fischer, H P Zingsheim, “Submicroscopic pattern replication with visible light”, J Vac Sci Technol 19, 881 (1981)

[38] H W Deckman, J H Dunsmuir, “Natural lithography“, Appl Phys Lett 41,

377 (1982)

[39] Christy L Haynes and Richard P Van Duyne, “Nanosphere Lithography: A Versatile Nanofabrication Tool for Studies of Size-Dependent Nanoparticle Optics”, J Phys Chem B 105, 5599 (2001)

[40] Helmut Colfen and Stephen Mann, “Higher-Order Organization by Mesoscale Self-Assembly and Transformation of Hybrid Nanostructures”, Angew Chem Int Ed 42, 2350 (2003)

[41] P Jiang, T Prasad, M J McFarland, V L Colvin, “Two-dimensional packed colloidal crystals formed by spincoating”, Appl Phys Lett 89, 011908 (2006)

nonclose-[42] C L Haynes, A D McFarland, M T Smith, J C Hulteen, and R P Van Duyne, “Angle-Resolved Nanosphere Lithography: Manipulation of Nanoparticle Size, Shape, and Interparticle Spacing”, J Phys Chem B 106,

[46] Christy L Haynes, Adam D McFarland, Matthew T Smith, John C Hulteen, and Richard P Van Duyne, “Angle-Resolved Nanosphere Lithography: Manipulation of Nanoparticle Size, Shape, and Interparticle Spacing”, J Phys Chem B 106, 1898 (2002)

Chapter 2 Development of nanosphere lithography

2.1 Introduction

Although nanosphere lithography (NSL) has been used to create nanostructures

on various substrates [1-6], it still lacks the control of shapes and arrangements of the nanostructures due to the nature of NSL There are three main progresses in the development of the technology On the first stage [1], only a single layer of nanospheres was used to create pillar structures by dry etching, and triangle-shaped metal nanodots by depositing through the voids of the hexagonally close-packed array

of the nanospheres On the second stage [2], a double layered hexagonally packed array of the nanospheres was used to create the nanostructures The shape and the arrangement of the nanostructures formed through the double layered nanospheres differ to that formed by the single layered nanospheres Obviously, such nanostructures created by normal single and double layered nanospheres cannot satisfy various demands in modern nanosciences and nanotechnologies Therefore, some modified NSLs, as third stage, have been proposed, such as angle-resolved NSL [5] and Shadow Nanosphere Lithography [7], which are able to create more complex nanostructures, such as nanooverlaps, nanocontacts, nanogaps, and nanochains These nanostructures also show unique optical properties [5] As we know, physical and chemical properties of nanostructures strongly depend on its shape not only in lateral, but also in vertical directions and its arrangement However, the current techniques of the NSL are difficult to engineer the shapes and the arrangement of the nanostructures

close-It is desired to develop new techniques

On the other hand, how to arrange the nanospheres uniformly is also a big challenge Due to the nature of the self-assembling of the nanospheres, the perfect

arrangement of a hexagonally close-packed array of the nanospheres can be obtained

in some micrometers If the area increases to tens of micrometers, some line and point defects appear If the area increases to hundreds of micrometers, pure single layer cannot be obtained over the large area In this case, single layer is accompanied almost with double layer or multi-layers which distribute randomly This phenomenon limits the applications of the nanosphere lithography because the nanostructures fabricated by single and double layered nanospheres are totally different Although some techniques have been used to improve the uniformity of the arrangement of the nanospheres, all of them need extremely tight conditions

In this chapter, some techniques were proposed to overcome the problems of the current nanosphere lithography One is in the aspect of controlling position and uniformity of self-assembly process A simple method has been demonstrated to obtain a single layer over areas as large as 400 µm2, which are large enough for device application, for example, LEDs One is in the aspect of shape engineering in both of vertical and lateral directions A simple method was invented to create 3D nanostructures only by a 2D nanosphere array The concept of 3D masks has been demonstrated, which can be used to fabricate more complex nanostructures In addition, a simple method has been developed to create nanostructural arrays with different size features over a wafer by one-step lithography

2.2 Self-assembly of colloidal crystals

local interaction among the components themselves, without an external direction For the system of spherical colloidal particles, the process is driven by the lateral capillary force and the hexagonal closed-packed (hcp) structure is formed on the surface of a film or a substrate Uniform monolayer or bilayer of the latex spheres is a base of NSL To carry out the self-assembly process, a hydrophilic surface is the first requirement To obtain the uniform monolayer or bilayer, a right concentration and uniform distribution of the colloidal spheres is the second requirement Meanwhile, several elements also affect the self-assembly process, such as environment (temperature and humidity), uniformity of the colloidal spheres, and purity of the colloidal solution In this section, I describe some techniques that were used to obtain whole single or double layer on a large area and to control the location for self-assembly

2.2.2 Control of self-assembled colloidal crystals

The simplest method to obtain the monolayer of colloidal crystals is by putting

a drop of the colloidal solution on a horizontal hydrophilic surface After spreading the colloidal solution on the surface to form a film, the nucleation of the colloidal spheres begins when the thickness of the colloidal film reduces by evaporation to around the thickness of the spheres The crystal is grown from the nucleus due to the lateral capillary force as shown in Fig 1.5 If a small drop of the colloidal solution is put on the surface, the nucleation starts along the edge of the drop because the thickness of the film at the edge is lesser than that at the center due to the surface tension effect Then the crystal grows towards the center In this case, it is difficult to control the arrangement of the crystal at the center and it depends on the exact concentration of the colloidal particles If the wafer is large enough, we can assume that the thickness of the colloidal solution is almost the same at all the locations on the surface and this results in nucleation taking place at many positions because of the

thickness fluctuations In this case, the growth of a monolayer or a bilayer also occurs simultaneously at many positions Therefore, many boundaries are formed when the front of the grown monolayer crystals meet together

Although there are techniques to obtain a monolayer or bilayer arrays of the colloidal spheres, we just use the simplest and fastest method to form the arrays, such

as self-assembling on an inclined wafer or by spin-coating

Figure 2.1 shows a schematic of the setup With this method, we can put a drop

of the colloidal solution on a hydrophilic surface fixed at an angle θ and thereby a film of thickness h containing the spheres is formed due to the gravity force

Nucleation of the spheres at the top edges of the film occurs and then grows to form the monolayer or bilayer arrays

Although a similar setup has been reported by Antony [13], and a theoretical analysis has been provided to describe that self-assembled process, the situation of

their setup is different from mine In their process, the supply of the colloidal solution

is infinite Therefore, the evaporation of the water plays an important role They also ignore the effect of the tilt angle of the sample However, in my system, the self-assembly process is mainly determined by the thickness of the film which is associated with the tilt angle and the concentration of the colloidal solution In this case, a simple relation can describe this situation:

(1)

where is shear stress between the substrate and the colloidal solution, ρ is the

density of the solution, g is the acceleration due to the gravity force and is the tilt

angle of the wafer Here, the density of the solution is proportional to the concentration (φ) of the colloidal particles If we assume the shear stress is constant, the self-assemble process depends on the thickness of the colloidal solution

where is spin speed, and is initial solution viscosity In this relation, evaporation

is not considered Actually, the self-assembling may take place after the spin-coating

It is worth to note that, in real situation, the initial solution viscosity of and

shear stress of depend on the particle size, so these effects have to be considered in practice In principle, a monolayer or a bilayer can be obtained just by changing the concentration However, in practice, nucleation also plays an important role, especially for the bilayer formation In most cases, the positions of bilayer nucleation are random

2.2.2.1 Self-assembly of a monolayer

The experiments were carried out on a Si substrate which was treated by standard cleaning process To increase the hydrophility of the Si surface, the Si surface was treated by O2 RIE for 30s before depositing the colloidal solution The colloidal solution of polystyrene (PS) spheres was purchased The concentration is 12 wt% After distributing the colloidal solution on the whole wafer, the wafer was tilted

on a holder as indicated in Fig 2.1 at an angle, which is varied from 10o to 300 Due

to the gravity force, a thin film can be formed on the wafer except at the bottom edge where some solution accumulated After removing the excess solution by blotting a paper, the holder with the wafer is put into a chamber to minimize the environment effects on the self-assembly process

Figure 2.2(a) shows a photograph of the self-assembled microspheres with a diameter of 2µm on a Si substrate In this case, the tilted angle is 100, the concentration was adjusted to be ~20 wt% Rainbow-like pattern is clearly observed, due to the photonic crystal effect, indicating good arrangement of the microspheres

In addition, there is a half-moon shaped dark area, where no PS spheres are assembled because the spheres are depleted at the final stage of the self-assembly process Microscopic examination of the film [Fig 2.29b)] confirms that almost a monolayer was formed under these conditions However, for the smaller spheres, for example, 400 nm, the concentration of the solution should be lower and the tilted angle should be larger to obtain a monolayer as seen in the SEM image of Fig 2.2(c)

My experiments show that the optimized conditions were: concentration at 8 wt% and the tilted angle at 300

2.2.2.2 Self-assembly of a bilayer

For the NSL application, a bilayer or a trilayer array can also be used as a template to create nanostructures However, there are two kinds of trilayer which is

arranged as ABC and ABAB Only ABAB arrangement of the crystal can be used as

a mask to form nanostructures through the opened voids It is hard to control the crystal arrangement of a trilayer in practice So, only bilayer crystals are focused in the work

Fig 2.2 (a) Photograph and (b) SEM images of the monolayered PS spheres of 2 μm diameter self-assembled on a Si substrate (c) A SEM image of the monolayered PS spheres of 400 nm diameter self-assembled on a Si substrate

From the equation (2), either the concentration or the angle can be adjusted to obtain the bilayer crystal In the experiments, it is easier to adjust the angle Therefore, for the 400 nm spheres, the concentration is kept the same to obtain the monolayer, but with the wafer tilted at an angle of 150 Under these conditions, over 80% of the area was covered with bilayer crystals Figure 2.3(a) shows a cross-section view of SEM image of the bilayer crystal However, it is noted that sometimes the bilayer is not continuous as seen in the micrograph image [Fig 2.3(b)], because of concentration fluctuations It is possible that the local concentration is lowered by a fast assembly process if the diffusion rate of the spheres in the solution is slower than the rate of self-assembly

Well arranged PS spheres show strong photonic crystal effect, whereby light reflection or transmission properties are affected by the period and layer thickness Therefore, if the period which is equal to the diameter of the spheres, is kept fixed, the color change reflects different numbers of layers of the sphere as seen in Fig 2.4

(the microscopy images of monolayer, bilayer and trilayer, respectively) These optical features can also be used to judge the structures quickly

Fig 2.3 (a) A cross-section viewed SEM image showing the bilayer arranged PS spheres with 400 nm diameter (b) A top viewed microscopy image showing the distribution of the bilayer array (green color) and the single layer (purple color) formed by the 400 nm spheres

Fig 2.4 Microscopic images (up panel) of different layered nanospheres assembled

on glass substrates (as indicated in bottom panel)

2.2.2.3 Selective self-assembly

A) Selective self-assembly inside wells

As discussed above, the self-assembly of a bilayer array often suffers the problem of local concentration fluctuations, resulting in the bilayer being not

continuous, and the position also randomly distributed Actually, a monolayer array also faces the same problem This problem seriously limits its applications For example, it is desired to form a monolayer array on a LED wafer to create uniform surface nanostructures to enhance light extraction at the device level Based on these requirements, I proposed a method to control the location of the self-assembled arrays

of the nanospheres at the device-sized level This method is demonstrated by my experimental results

Our basic idea is to use a well formed by photolithography to confine the assembly process into a relatively small area, but large enough to make a device In addition, the well edges can accept the badly arranged nanospheres, such as multi-layered arrays, because the nucleation starts at the central region of the well where the thickness of the solution is the least The “bad” portions can serve as sacrificial regions during device processing

self-Fig 2.5 (a) Schematic of the self-assembly within the wells (b) Microscopic image

of the patterned Si substrate

In these experiments, spin-coating was used to form a uniform film of the colloidal solution on the patterned Si substrate, which is formed on photoresist by photolithography The squares are designed to be 400x400 µm2 and the strip width is 10µm and the height is 1.5µm as shown in Fig 2.5(b)

of concentration, ~15 wt%, spin speed, 900 rpm

Figure 2.6(a) shows a typical microscopic image of the 300 nm diameter PS spheres self-assembled on a Si substrate by spin-coating The concentration of the colloidal solution is ~7 wt% and the spin speed is 1900 rpm It is clearly observed that an almost perfect monolayer was covered inside the wells “Bad” areas were formed only along sides of strips The width of the “bad” areas is smaller than 5 µm

It is noted that there are veins from the central area of the wells radiating to the edges, implying the 2D crystal is grown from the central region to the edges As illustrated

in Fig 2.5(a), after spin-coating the film that is formed does not have a flat surface The surface is curved around the photoresist strips due to surface tension Because of the hydrophobic property of the photoresist, the solution on the top of the strips can

be pulled back quickly At this stage, the wells functioning, like tanks, to hold the solutions In addition, the thinnest portion of the film inside the well is in the central region because of the surface tension As the solution evaporates, the nucleation starts

at the center and then grows towards the edges of the well Finally, extra spheres form multi-layers along the edges if the concentration is higher Blank areas are formed if the concentration is lower When the spin speed lowers to 1800 rpm, the “bad” areas are increased as seen in Fig 2.6(b)

Figure 2.6(c) displays a microscopic image of the bilayer formed on the patterned substrate, which is obtained by using the following conditions: concentration at 15wt% and spin speed at 900 rpm A continuous bilayer is successfully obtained inside the well At this condition, the width of the “bad” area is around 30-50 µm

B) Self-assembly on a micro-patterned substrate

In some applications, it is also desirable to arrange the nanospheres on a step film For example, such an arrangement can be used to fabricate photonic crystals with different structural features in a one-step process or in forming a template with

multi-nanoholes for the selective growth of semiconductor nanodots (as discussed later)

Fig 2.7 (a) Schematic illustration of the patterns created on a SiO2 surface by photolithography from top (up) and side views (bottom) (b) A SEM image of the monolayer arrays of 300 nm PS spheres formed at a SiO2 film with micro-wells Figure 2.7(a) illustrates a schematic of patterns created on a SiO2 surface by photolithography combined with dry etching The thickness of the SiO2 film is 150

nm deposited on a GaAs substrate by PECVD Width and depth of the wells are 8µm and 80 nm with a separation of 5 µm Figure 2.7(b) shows a SEM image of the 300

nm spheres self-assembled on a patterned SiO2 surface with shallow wells It is clearly observed that the arrangement of the nanospheres is uniform over the whole surface both outside and inside the wells The arrangement of the array is continuous

at the edges of the wells Two conditions are required to achieve this arrangement: (1) the depth of the well must be much smaller than the diameter of the spheres, and (2) the whole wafer (inside and outside of the wells) must be hydrophilic

C) Self-assembly selectively in micro-wells

If hydrophilic regions are partially formed on a surface, it is possible to arrange the sphere arrays on the designed areas Figure 2.8 shows a SEM image of the spheres self-assembled inside a deep well (~1µm) formed on a GaAs substrate by photolithography and dry etching The clean GaAs surface is hydrophobic However,

O2 plasma treatment in the well after it was created results in the GaAs surface being hydrophilic inside the well Therefore, the self-assembly can occur only inside the wells as seen in Fig.2.8 However, a key issue, besides concentration effect, is the method to spread the liquid In this case, the well diameter is 8 µm and the separation

is about 150 µm [Fig 2.8(b)] The hydrophilic area is too small to spread the colloidal solution on the surface The setup as shown in Fig 2.1 cannot be used in this case To obtain the self-assembly in the array of small hydrophilic areas, the wafer was dipped vertically in a small tank containing the colloidal solution The front of the solution goes through the hydrophilic areas when the solution is evaporating The self-assembly can only occur on the hydrophilic areas and not on the hydrophobic area because of the contact angle being larger than 900 due to surface tension Nucleation

at this portion cannot take place

Similar results are also observed at a polymer substrate with small square wells fabricated through nanoimprinting as seen in Fig 2.9 The surface of the polymer substrate is hydrophobic, while the 2 µm wells were treated to be hydrophilic The depth of the wells is 2 µm It is clearly observed that 400 nm PS spheres were selectively arranged inside the wells with an excellent selectivity as seen in Fig 2.9 (b) The liquid front clearly shows that self-assembly takes place only in the wells

Fig 2.8 (a) A SEM image of the 300 nm PS spheres selectively formed inside a circular well created on a GaAs substrate (b) A microscopic image of the arrangement of the wells created by photolithography

Fig 2.9 (a) A microscopic image of the 400 nm spheres selectively formed inside micro-wells created on hydrophobic polymer substrate by imprinting (b) The front form of the colloidal solution indicating the selection of the self-assembly inside the micro-wells

2.3 Applications of colloidal crystal as templates

2.3.1 Introduction

Although arrays of nanosphere can form functional structures such as photonic crystals [15-18], it is mostly used as templates to create secondary nanostructures, which greatly widens its applications [1-7, 12] Furthermore, the secondary nanostructures also can be used as templates to create a third nanostructure [6] Various nanostructures with different arrangements, shapes and sizes are desired, but

it still lacks such techniques for fabricating more complex nanostructures, especially, controlling the nanostructure’s shape vertically In this section, some approaches to create unique nanostructures are presented, including creation of ordered spherical nanocavities with excellent structural uniformity, and 3D silica hollow nanostructures

I have also used the 3D hollow nanostructures as a mask to create various 2D and 3D surface nanostructures firstly A method to use a 2D array of PS spheres to create 3D nanostructures is also described In addition, a simple method to fabricate nanostructural arrays with multi-size features in one-step lithography has been demonstrated

2.3.2 Ordered spherical nanocavities

Infiltration process has been used to obtain reversed bulk nanostructures, such

as photonic crystals, through the self-assembly of nanospheres [19.20] However, for the application as masks or templates, uniformity of the nanostructures is key Thus the fabrication of such nanostructures with excellent uniformity is still a big challenge Here, we propose a solution by combining spin-coating and dry etching processes, which are widely used in electronic industry to overcome this problem Spherical nanocavities and 3D silica nano-opals with excellent uniformity have been successfully obtained

Principle and experiments

cleaned and then treated by O2 plasma for 30s to make the surface hydrophilic 300

nm, 600 nm and 900 nm PS spheres were self-assembled on the Si wafer [Fig 2.10(a)] O2 RIE was then used to etch the PS spheres, which reduces the diameter of the spheres to ~ 270 nm for the 300 nm spheres (60s etching) and ~500 nm for the

600 nm spheres (100s etching) Then commercial liquid silica was put on the wafers which would infiltrate into the array due to capillary force and form a thin film To ensure a film with excellent uniformity, spin-off process at a spin speed of 5000 rpm was used and followed by baking at 2500 for10 minutes in air SEM pictures of the solidified wafers show an almost flat surface of the silica film and a monolayer of PS spheres distributed uniformly The total thickness of the silica film is about 600 nm and the thickness between the surface of the silica and the top of spheres is around 50

nm for the wafer with an array of 600 nm spheres A RIE etching of the silica film was carried out to expose the top of the PS spheres Etchant of CH4 with flow rate of

25 sccm at a chamber pressure of 15 mTor was used The etching duration is varied from 2 to 3 minutes Finally, the PS spheres were removed by toluene to form the array of silica cavities

Results and discussion

Fig 2.11 A cross-section view of a SEM image of the silica nanocavities

Figure 2.11 shows a cross section SEM image of the solidified silica film containing 900 nm PS spheres Uniform flat surface of the film is clearly observed

The thickness of the film is estimated to be 820 nm, and the thickness from the surface of the film to the top of the spheres is around 140 nm It is noted that the diameter of spheres after etching is 737 nm in horizontal direction and 646 nm in vertical direction, implying that the etch rates in the two directions are different It is reasonable that the etch rate in vertical direction is faster than that in the horizontal direction because the plasma generated in the etching chamber is directed vertically After etching the top layer of the silica film above the PS spheres, the PS spheres can be removed by toluene to form the spherical cavities in the silica Controlling the etch depth can control the diameter of hole opening of the nanocavities Fig 2.12(a) shows a SEM image of the hole openings of the cavity array, where 150 nm silica is etched off When the etch depth reaches the 170 nm, the opening is enlarged as seen in Fig.2.12(b)

Fig 2.12 Top-view SEM images of the periodic ordered nanocavities To form the nanocavities the top of the silica film was etched down (a) 150 nm and (b) 170 nm to expose the PS spheres

When the etch depth was further increased to the point where the up-half of the sphere was etched away, the silica nanostructures left behind become a hexagonal lattice with protruding tip structures located at the interstitial position of the honeycomb lattice This surface morphology of the 2D silica nanostructures has been

confirmed by AFM measurements A similar experiment was carried out with 300 nm

PS spheres The same procedure as shown in Fig.2.10 is applied to form the structures The solidified silica film was etched down to the point below a half diameter of the spheres [the step shown in Fig 1.10(g)] A 3D view of the silica lattice of the sample

is shown in Fig 2.13(a) Nearly perfect structures of the silica hexagonal lattice with tip structures located at the interstitial position of the honeycomb lattice can be

clearly seen in the AFM image, which shows P 6mm symmetry Fig 2.13(b) and (c)

show the cross-section images of the sample along the mm’ and nn’ directions

indicated in Fig 2.13(d) As expected, the mean period [as shown in Fig 2.13(b)] of the circular openings is about 312 nm which agrees well with that measured by SEM The height of the sidewall between two pores of the honeycomb structures is about 70

nm The height of the protruding tips located at the interstitial position of the honeycomb lattice, i.e at the corner between 3 adjunct holes, is about 122 nm, which

is about 50 nm higher than the sidewall of the silica nanostructures The full width at half maximum of the protruding tips is around 83 nm The nearest distance between the tips is 180 nm which is almost the same as 180.1 nm, calculated from simple geometry for the closely packed spheres with a diameter of 312 nm In addition, the full width at half maximum of the walls of the pore is measured to be ~74 nm Obviously, the structural features of these nanostructure arrays can be easily adjusted

by changing the RIE etching durations if the sizes of the original nanospheres are kept the same In principle, smaller pores, protruding tips, and walls between the pores can be obtained by using smaller nanospheres

Based on the experimental results, we propose a model for evolutions in the formation of nanostructures with RIE dry etching Fig 2.13(e) shows the cross-

sections [along the mm’, and nn’ direction of the nanostructures shown in Fig 2.13(d)]

of the possible nanostructures obtained at different etching stages At the earlier stage, the RIE homogenously etches the silica film until the etching front reaches the PS spheres Further etching results in the circular openings of the spherical cavities of the

Fig 2.13 (a) Perspective view of the sample, where the silica was etched below the horizontal diameter plane of the sphere at step g shown in Fig 10 (b) and (c) show

line scan taken along the directions of mm’ and nn’ as illustrated in (d) (d) Top view

of AFM image for the same sample (e) Schematic illustrations of the evolution

[cross-section views along mm’ and nn’ direction, respectively, as shown in (d)] of

the hexagonal close-packed silica nanostructure arrays at different stages of etching

silica nanostructures (the case as seen in Fig 2.12) Plasma generated during RIE has

a strong momentum, resulting in a higher etch rate for the edges of the openings Therefore, the diameters of the circular openings are quickly increased because the etch rate at the edges of the opening is faster than other portions of the nanostructures Dry etching conditions should be carefully controlled at this stage if we want to obtain spherical cavities with different diameters of openings The slowest etching rate of the spherical cavity arrays during the RIE should be at the interstitial position between three spheres Hence, further etching results in protruding tip formation

located at an interstitial position of the hexagonal structure, which shows P6 mm

symmetry [the case as seen in Fig 2.13(a)]

2.3.3 Creating nanostructures with multi-features in one step

In practice, it is useful to have the location of the nanostructures formed at the desired places Furthermore, the integration of different nanostructures in one chip also benefits applications in nanophotonics, biosensors, and displays However, so far only EBL and nanoimprinting techniques can be used to fabricate nanostructures with different features in one step Even nanoimprinting, the mold is also made by EBL Here, a simple, inexpensive and high throughput approach to create surface nanostructures in one step by using self-assembled nanospheres combined with photolithography is presented

Principle and experiments

Figure 2.14 displays a schematic diagram of the principle to fabricate an ordered array of nanoholes with multi-size features A SiO2 film is firstly deposited

on a flat substrate, such as Si or GaAs, which is then etched to have different regions

with different thickness h 1 and h 2 , h 3 , h 4 … The hcp monolayer or bilayer

arrangement of polymer nanospheres can be self-assembled uniformly on the SiO2

film, as described in section 2.2.2.3, if the thickness difference (for example h 1 and h 2)

is not larger than a half of the diameter of the spheres We can select proper etchants such that the polymer spheres can be etched during dry etching of SiO2, resulting in the tilted wall formation for the SiO2 nanostructures as depicted in Fig 2.14 This gives us an opportunity to fabricate the multi-sized nanoholes in a single step process

At the earlier stage, the dry etching of SiO2 results in an etch front with the diameter

of a (assuming it is circles) in the SiO2 film through the interstitial voids of the spheres With continuous etching, the etch front goes deep into the SiO2 film in both

thickness regions, while the diameter of the spheres is reduced at the same time Therefore the top portion of the nanoholes in the SiO2 film can

while the etching front is still inside the SiO2 film for thicker SiO2 film regions (h 1)

The size of the bottom of the hole openings a’ can be further enlarged simply by increasing the etching duration t When the hole has reached the substrate its size

increases with etching time, given by 2tR p When the etch front in the thicker SiO2

region reaches the substrate, the hole size is still a, but the hole size in the thinner regions becomes d Thus it is seen that in a single etch step, two different sized holes are formed at the substrate The smaller hole is formed in the thicker silica film (h 1), while the larger hole is formed in the thinner SiO2 film regions (h 2) The size of the opening at the top of the SiO2 for both thicknesses are however the same

Normally, the ratio of the dual-sized silicon dioxide nanoopenings d/a is

determined by the thickness of the SiO2 film h1 and h 2 and the respective lateral and

vertical etching rates (R p and R s) for the polymer nanospheres and the SiO2 films By controlling these parameters, it is easier to modulate the structural features of the ordered silicon dioxide templates to make various semiconductor nanostructures (d a2R p(h1h2)/R s) When three or more thicknesses of SiO2 thin films are used, multi-sizes of ordered nanohole arrays can be further fabricated on the micropatterned SiO2 thin films It is assumed that the micro-patterned substrate has n regions with decreased thickness of h 1 , h 2 , h 3 , … … h n, respectively It is also required

that h n -h 1 < half the diameter of the nanospheres After etching, the bottom sizes of

the openings are l 1 , l 2 , l 3 ,… l n increase correspondingly The size for the nanoopenings

located at the nth region can be expressed by

s n p

l  12 ( 1 )/ (4)

Although hcp monolayer of the spheres can be used to create the two-sized

nanoholes in SiO2 film, the quality of the holes is poor Therefore, a bilayer array of the PS spheres was used in the experiments The SiO2 film of ~100 nm in thickness was deposited by PECVD followed by photolithographic patterning on SiO2 film (8 

8 µm squares separated by 5 µm) and further etching of SiO2 inside the square down

to ~50 nm by RIE After removing the photoresist, an array of SiO2 square well is remained and a bilayer of nanospheres is subsequently self-assembled on the whole