Sensitivity Enhancement Study of

Nanostructure-Based Localized Surface Plasmon

Resonance Biosensor

Kyungjae Ma

The Graduate School

Yonsei University

School of Electrical and Electronic Engineering

Sensitivity Enhancement Study of

Nanostructure-Based Localized Surface Plasmon

Resonance Biosensor

A Master’s Thesis

Submitted to the School of Electrical and Electronic

Engineering and the Graduate of Yonsei University

in partial fulfillment of the

requirements for the degree of

Master of Science

Kyungjae Ma

December 2009

감사의 글

졸업을 앞두고 연구실에서 보냈던 2년 여의 시간을 돌아보게 됩니다. 좋은 연구 결과를 얻기 위해 밤을 지샜던 날들은 분명 앞으로 내가 살아가며 이룰 결실들의 밑거름이 될 것이라 믿어 의심치 않습니다. 이제 본 학위 논문이 나오기

까지 많은 가르침과 깨달음을 주신 것에 감사 드리며 여러분들께 감사의 말씀을

드리고자 합니다.

먼저 본 연구를 수행하고 마무리 하는데 있어 아낌없는 지도와 격려를 해

주신 김동현 지도교수님께 감사를 드립니다. 항상 연구를 향한 열정과 늘 새로운

학문에 매진하시는 교수님의 모습을 본받아 졸업 후 어느 곳에 가더라도 이 사회

의 필요한 인재로 성장하도록 앞으로도 노력하겠습니다. 또한 제게 다른 한 분의

지도교수님과도 같으셨던 심은지 교수님께도 감사의 말씀을 드립니다. 특히 저의

학위 논문이 나오기까지 심사위원으로서 여러 가지 조언과 많은 도움을 주신 최

우영 교수님과 이용식 교수님께 감사의 말씀을 드립니다. 또한 2년의 연구 과정

동안 함께 할 수 있었던 물리학과 유경화 교수님, CUHK의 Aaron Ho-Pui Ho 교수

님, 전자과의 이명호 교수님, 의과대학 윤채옥 교수님께도 감사의 말씀을 드립니

다.

본 연구를 진행함에 있어 안 밖으로 많은 도움을 주셨던 우리 연구실 식

구 분들에게도 감사의 말씀을 전합니다. 연구실의 꽃 미남이자 맏형이신 세영 형,

다재 다능하신 신세대 애 아버지 규정 형, 연구실 동기지만 든든하고 배울 점 많

은 동준 형, 연구실 이슈 메이커인 호정 형, 인자한 미소를 가지고 계시지만 요

즘 들어 부쩍 공격력이 상승 중이신 영진 형, 연구실 내 자폭개그의 창시자 종률

군, 그리고 먼저 졸업하셔서 유학길에 올라 고군분투 중이신 순준 형. 이제 연구

실의 새 식구가 될 지유누나, 원주양에게도 감사의 말씀을 전합니다. 또한 즐거

운 시간을 함께한 계산화학 연구실의 희영 누나, 민우 형, 민철에게도 감사의 말

씀을 전합니다. 또한 학부시절 같이 벼락치기하며 날밤을 지샜던 친구들에게 감

사를 전합니다. 특히 결혼을 앞둔 창헌 형, 귀여운 창익 형, 본의 아니게 친구

먹게 되는 바람에 미안했던 재영 형, 춘구 형, 그리고 동갑내기이자 듬직한 혁이

군, 듬지 군, 원석 군에게도 감사를 전합니다.

언제나 날 믿어 주시고, 당신의 아들을 위해 지금껏 눈물로 기도하시며

뒤에서 힘이 되어 주시는 사랑하는 부모님, 그리고 올해 부푼 꿈을 안고 대학에

입학하는 내 동생 경은이에게 감사를 전합니다. 또한 말씀으로 늘 신앙의 버팀목

이 되어 주시는 담임목사님과 사랑하는 교회 식구들, 청년들, 특히 마음이 어렵

고 힘든 시기마다 극복할 수 있도록 기도해 주시고, 밥도 엄청 많이 사주신 (큰

사랑에 보답할 길이 없네요^^) 권성근 전도사님께 감사의 말씀을 드립니다.

마지막으로 오늘날의 내가 있게 해준 사랑하는 나의 조국 대한민국과 궁

극적 감사의 대상이요 앞으로도 나와 함께 하시며 내 꿈도 이루어 주실 지혜의

하나님께 감사와 영광을 돌립니다.

2010년 1월 마경재 올림

Table of contents

Table of contents ································································································ i

List of tables ···································································································· iii

List of figures··································································································· iv

Lists of Abbreviations ····················································································· vii

Abstract ·········································································································· viii

Chapter 1: Introduction ···················································································· 1

Chapter 2: Sensitivity enhancement study on nanostructure-based Localized

surface plasmon resonance (LSPR) sensor ························································ 3

2.1. Theoretical background ···································································· 3

2.1.1. Dispersion relation ································································ 3

2.1.2. Effective medium theory (EMT) ·········································· 4

2.1.3. Numerical analysis ································································ 5

2.2. LSPR detection analysis for phase sensitive configuration ·············· 6

2.3. Sensitivity characteristics of nanostructure-based LSPR sensor ······· 9

2.3.1. Randomly distributed silver islands structure ······················· 9

2.3.2. Periodic gold nanowire structure ········································ 14

Chapter 3: Experimental study on highly sensitive LSPR detection based on

target localization using periodic nanowire structure ····································· 22

3.1. Numerical modeling ······································································ 22

3.1.1. Sensor design ······································································ 22

3.1.2. Sensitivity enhancement factor ·········································· 24

3.2. Materials and methods ···································································· 24

3.2.1. Sensor fabrication······························································· 24

3.2.2. Optical setup ······································································ 25

3.2.3. DNA hybridization process ················································ 26

3.3. Result and discussion ······································································ 27

i

3.3.1. Numerical analysis ····························································· 27

3.3.2. Detection of DNA hybridization ········································ 33

3.3.3. Discussion············································································ 37

Chapter 4: Conclusion ····················································································· 39

References ······································································································· 41

국문요약 ·········································································································· 45

ii

List of tables Table 1. Maximal phase shift (upper row) and FWHM (bottom row) of the

scheme of conventional thin film and periodic nanowire based LSPR structure

regarding film thickness and structure period. (Unit: degree) ····························8

Table 2. SEFUTV for various gold film thicknesses. Theses parameters other

than dTF have been considered from the results presented in Figure .3-4·········30

iii

List of figures

Figure 2-1. Angular based phase shift at target RI change from 1.333 to 1.334

of (a) conventional scheme using gold thin film at thickness of 40 - 52 nm and

(b) LSPR scheme using periodic nanowire periods of 300-400nm. In addition

Its phase variation at the target RI change at maximal resonance condition of

(c) the conventional scheme at gold thin film and (d) the LSPR scheme.)····· ···7

Figure 2-2. SEM image of randomly distributed silver islands structure ·······10

Figure 2-3. The structure of Adenovirus; (1) penton capsomeres, (2) hexon

capsomeres, (3) viral genome (linear dsDNA)······· ··········································11

Figure 2-4. Magnetic field intensity profile at surface of silver islands based

plasmonic sensor································································································12

Figure 2-5. LSPR and conventional SPR detection as change of adenovirus

concentration·····································································································13

Figure 2-6. (a) Schematics with a target binding DNA layer model of LSPR

structure base on periodic nanowire and (b) Conventional thin film based SPR

structure·············································································································15

Figure 2-7. SEFUTV calculated for periodic nanowire structure for DNA

hybridization. The row and column corresponds to dNW = 10, 20 nm and Λ =

300, 400, 500 nm, respectively··········································································17

Figure 2-8. DNA hybridization far-field characteristics of angular based LSPR

using periodic nanowire and conventional SPR structure using thin film ·······19

Figure 2-9. Magnetic field intensity (|Hy|2) distribution at θSP of (a) LSPR and

(b) conventional SPR structure. (c) z axis cross field profiles on (a) and (b)· ·21

Figure 3-1. 3-D Schematic of a periodic nanowire LSPR sensor with design

parameters. The thick arrows represent the evaporation angle of SiO2·············23

iv

Figure 3-2. (a) An SEM image of the periodic nanowire sample after angled

evaporation. (b) A magnified SEM image of the sample confirms the existence

of the dielectric opening by angled evaporation. ··············································25

Figure 3-3. SPR system set-up (PO: polarizer, CH: chopper, M: mirror, PD:

photodiode, and PC: personal computer). ························································26

Figure 3-4. SEFUTV calculated for various designs of target-localized periodic

nanowire structure. x- and y-axis indicate evaporation angle θeva in degrees and

SEFUTV, respectively. The row corresponds to (dTF, dNW) = (40 nm, 10 nm), (40

nm, 20 nm), (45 nm, 10 nm), (45 nm, 20 nm), (50 nm, 10 nm), and (50 nm, 20

nm) from top to bottom. The columns represent dm = 1, 3, and 5 nm (from left

to right). Λ in nm and f in %.·············································································29

Figure 3-5. Magnetic field amplitude profiles (|Hy|) calculated for dsDNA: (a)

dTF = 40 nm, (b) 45 nm, and (c) 50 nm. Other parameters are same as those

listed in Table 2. The arrow indicates the incidence direction. (d) Horizontal

and (e) vertical magnetic field distribution show that increased radiation

damping decreases the near field for larger dTF and is associated with less

efficient sensitivity enhancement······································································31

Figure 3-6. Magnetic field amplitude distributions (|Hy|) calculated for the

optimum structure with parameters listed in Table 2 for dTF = 40 nm: from (a)

ssDNA to (b) dsDNA. The arrow indicates the light incidence. (c) Field

difference between the near-fields of (a) and (b), where the arrow indicates the

field difference at localized hotspots formed in the position of target dielectric

opening. ············································································································32

Figure 3-7. Measured reflectance for resonance characteristics of hybridization

process on the periodic nanowire and a thin film control ·································34

Figure 3-8. Real-time reflectance dynamics of DNA hybridization measured by

the periodic nanowire and the control sample, the incident angle was fixed at

v

θin = 60°.······································································································· ····34

Figure 3-9. DNA hybridization in the reverse incidental configuration, the

sample is mounted in the reverse orientation. Calculated near field profile of

(|Hy|): (a) ssDNA and (b) dsDNA, (c) Field difference between the near-fields

of (a) and (b), where the arrow indicates the field difference at localized

hotspots formed in the position of target dielectric opening. (d) Measured

reflectance based resonance characteristics.··················································· ··36

vi

List of Abbreviations

SP Surface Plasmon

SPR Surface Plasmon Resonance

RI Refractive Index

LSP Localized Surface Plasmon

LSPR Localized Surface Plasmon Resonance

EMT Effective Medium Theory

RCWA Rigorous Coupled Wave Analysis

FDTD Finite-Difference Time Domain

RIU Refractive Index Unit

FWHM Full With at Half Maximum

SEM Scanning Electron Microscopy

TIRF Total Internal Reflection Fluorescence

SEF Sensitivity Enhancement Factor

SEFUTV Sensitivity Enhancement Factor per Unit Target Volume

SAM Self-Assemble Monolayer

MCH 6-mercapto-1-hexanol

vii

Abstract

Surface plasmon resonance (SPR) is the optical phenomenon by

propagating electromagnetic field under specific incidental wave condition,

which has been used in molecular detection and imaging as biosensors. This

thesis aims to investigate highly sensitive localized SPR (LSPR) detection

using nanostructure. As a result, several enhancement factors were proved by

theoretical and experimental study including actual chip fabrication process.

In the theoretical analysis, rigorous coupled wave analysis and

finite-difference time-domain methods were employed to evaluate far and near

field characteristics as the LSPR biosensors. By using silver nanoislands based

LSPR, sensitivity improvement was induced by localized near fields. In this

study, actual fabricated islands chip was directly modeled from own SEM

image for the near field calculation. Here, relatively strong and tiny hotspots

were sparsely formed at the sensor surface, this hotspots were effective to

decade-nanoscale molecules detection. Actually, we could confirm

experimentally that the structure gives reasonable reflectance signal for lower

concentration of Adenovirus compared with conventional SPR detection. In

addition, the structure is expected to be used as noble imaging technique by

compensating blurring effect due to condensed fluorescence emission.

In nanowire based LSPR sensing, the scheme has merits of graphical

modeling using various numerical parameters to evaluate as well as to actual

fabricate. Far field analysis shows that the scheme with the optimum

parameters was 3 times more sensitive than conventional detection by thin

film based SPR. This result can be interpreted by broaden resonance angle

shifts accompanied with reflectance damping because of the localized surface

plasmon (LSP) propagation. These LSP modes were formed at both sides of

the nanowire ridge which formed as bimodal hotspots referred to the result of

near field analysis.

viii

Finally, experimental study for investigating highly sensitive LSPR

detection was conducted based on the enhancement issues by hotspots

excitation. Here, angled evaporation for dielectric mask layer was employed to

be exclusively raised the interaction event only at intended localized-areas

which overlap localized hotspots in the structure. By DNA hybridization

detection, sensitivity enhancement by resonance angular shift per unit target

volume was observed to be 64 times compared with conventional thin film

based SPR detection. In conclusion, the nanostructure-based LSPR detection

was excellent to characterize for sensitivity improvement of SPR biosensing,

and is expected to be used more enhanced detection method by additional

improvement using appropriate mask layer for target localization.

Keywords: Nanoplasmonics, Biosensor, Sensitivity enhancement,

Nanostructure, Target localization

ix

Chapter 1 Introduction

Surface plasmon (SP) is a longitudinal electromagnetic wave formed

between a dielectric and metal interface when TM wave is incident. Condition

of SP resonance (SPR) varies in response to molecular interactions which

occur on the metal surface. Due to the capability to detect by label-free as

refractive index (RI) changes, SPR has been used in well-known biosensing

technique. Basically, the dispersion relation between SP momentum and

energy can be postulated from Maxwell’s equations, which is given by

SPS0dm

dmSP sinθnk

εεεε

cωk =

+= , (1)

where θSP, εm, εm and ns are the incident angle at resonance condition, metal

permittivity, dielectric permittivity of target and RI of a dielectric prism

substrate. In Eq. (1), kSP is the SP wave vector and ω is angular frequency. c

and k0 are the speed of light and wave vector in free space. kSP and θSP in Eq.

(1) are required to be complex values as εm is complex. For gold or silver that

is typical metal used to excite SPs, |Re(εm)| >> |Im(εm)|, so that sinθ values of a

measured SPR condition can be approximated as [εmεd/(εm + εm)]1/2/ ns from Eq.

(1). It is the change of εd related in molecular interactions that one measures

using SPR.

Recently, intensive research have been made to enhance the sensitivity

limit of SPR: for example, by signal amplification using functionalized

nano-particles [1], magneto-optic methods [2], localized SPR (LSPR) based on

periodic various nanostructure [3,4,5,6,9,10,11,12,13,20,29,30] and detection

schemes of phase-sensitive SPR [7,8]. Especially, nanostructure-based LSPR,

typically, which is characterized by intensity measurement of periodic groove

patterns on the metallic film, improves the sensitivity by means of generating

１

partially distributed hotspots. It was also found that the sensitivity

enhancement relies on the molecular interaction [9] and can be acquired

additionally by localizing the target molecules in the hotspots [10].

This research contains theoretical background which describes

numerical analysis on LSPR detection based on metallic nanowire structure.

Then, we will consider sensitivity enhancement effect associated with

localized SP (LSP) based on randomly distributed silver nanoislands and

metallic nanowire structure.

The thesis’s goal is ultimately to investigate highly sensitive LSPR

detection system using numerical and experimental methods. For this reason,

it is important to consider the spatial distribution of localized near-fields on a

sensor surface and the target localization effect on the distributed field. This

research thus contains theoretical background which describes numerical

analysis on LSPR detection based on metallic nanowire structure. Then, we

will consider sensitivity enhancement effect associated with LSP based on

silver nanoislands and gold nanowire-structure. Then, for the experimental

study using periodic nanowire based LSPR detection, note that angled

evaporation process is additionally used in chip fabrication for target

localization. This research can be potentially important to taking advantage of

nanophotonic fields for biomolecular sensing and imaging.

２

Chapter 2 Sensitivity enhancement study on nanostructure based Localized surface plasmon resonance (LSPR) sensor 2.1. Theoretical background 2.1.1. Dispersion relation An analytic solution does not exist in a closed form to describe

plasmonic phenomenon of electromagnetic fields in nanostructures.

Nevertheless, an alternative approach can be offered use of effective media

approximately. Dispersion relation for a surface mediated nanostructure can be

revised from Eq. (1) as [11,12,13]

SPS0effm

effmSP sinθnk

εεεε

cωk =

+= , (2)

In Eq. (2), εeff indicates the effective permittivity of the nanostructures.

kSP is significantly varied by the condition of a nanostructure, as it modulates

εeff in Eq. (2). It tends to take stronger momentum to establish plasmonic

oscillations in a surface mediated nanostructure. This suggests the SP

generation at shorter plasmon wavelength (λSP = 2π/kSP) and implies the

increased degree of field localization. In other words, an increased momentum

may functionalize as a sensitivity enhancement indicator. Especially, If

periodic nanowire at a smaller pitch and an extreme fill factor are used to

induce SPs, back-bending damping of plasmon momentum may be

significantly appeared [14]. This induces the decreased momentum at

resonance, although the configuration nature is complicated by the diverse

design parameters. Such a momentum variation is often accompanied by the

significant enhancement of plasmon fields-formation and sensitivity [4].

３

2.1.2. Effective medium theory (EMT) In this section, we describe periodic nanowire as an effective medium

related in the Rytov effective medium theory (EMT) for qualitative analysis

[15,16]. If the ambient medium including target area on the periodic nanowire

is approximated by εd, an overall effective medium that considers both sensing

target and nanowire may be given by

dmTEeff ff εεε )1()0(, −+= , (3)

dmTMeff

ffεεε−

+=11

)0(,

, (4)

22)0(,

3)0(,

222

)0(,

11)()1(3

⎟⎠⎞

⎜⎝⎛ Λ

⎟⎟⎠

⎞⎜⎜⎝

⎛−−+=

λεεεεπεε

dmTEeffTMeffTMeffeff ff (5)

with ε(0)eff,i and λ as the dielectric constant of effective medium at first order in

TE or TM polarization and wavelength of incident light, respectively.

While this approximation has been applied only when obtained

phenomenally and used in the Fresnel analysis [17,18], the EMT introduced in

Eqs. (3)-(5) refers to only qualitative viewpoint and has limitation of

completive analysis of the SPs excitation process in periodic nanowire,

because of the following reasons in [19]. First, the second-order EMT shown

in Eqs. (3)-(5) normally does not provide valid results which converge for

metallic nanowire. Second, the periodic nanowire and a metallic film layer are

so thin that entering wave can be directly coupled to outgoing wave. Moreover,

what an EMT does is only limited to homogenize the distributed field

produced by a periodic nanostructure to approximate the far field, in contrary

to the truth that evanescence waves are localized in the near-field. Therefore,

we have mainly used in numerical analysis by following analysis in next

section.

４

2.1.3. Numerical analysis The analysis was numerically calculated to investigate enhanced

sensitivity using rigorous coupled wave analysis (RCWA) and the

finite-difference time domain (FDTD). RCWA has been successfully

employed to presenting optical characteristics of various nanostructures. If the

ambient media are assumed to be linear in terms of the electromagnetic fields,

isotropic, time-invariant, homogeneous and source-free, Maxwell’s equations

are derived to a generalized by Helmholthz equation

0z)(x,Hnkz)(x,H y22

y2 =+∇ (6)

for TM polarized light. RI parameter, n = [ε(x,z) μ(x,z)]1/2. For a boundary

condition as pseudo-periodic, the Floquet-Bloch theorem is derived as follows: )sin(),(),( θi

yy ezxHzxH Λ+= (7)

for 0 < x < Λ and -∞ < z < ∞. Also, θ can be denoted as the incident angle. In

this event, the Rayleigh expansion holds as a solution of the near field:

)(),( zikxik

mmy

zmxmeAzxH +∞

−∞=∑= (8)

where Am may be the reflection or transmission coefficient in place. The z

dependence of the complex permittivity can be neglected for an infinitely long,

one-dimensional nanowire, thus, it is possible to have solution inside the

grooves as a Fourier expansion. The unknown values such as Am can be

determined based on the tangential-components continuity rule of the

electromagnetic fields at the boundaries between two media. For the

simulation in this study, spatial harmonics have been employed by 15-30.

５

2.2. LSPR detection analysis for phase sensitive configuration

Phase sensitive SPR detection is typical method for sensitivity

enhancement using a modified optical interferometric scheme. The system is

basis on detection to reflective differential phase-angles between TM and TE.

Here, TE is only used in a reference signal with respect to TM variation

because TE is not strongly varied for SPR condition. In the sensitivity aspects,

RI unit (RIU) by the order of 10-8 was reported [8], which refers a result by 10

times enhancement than conventional reflectance sensitive based SPR

detection. Note that RIU can be estimated by maximal RI detection

performance multiplied by sensing resolution.

Here, I want to brief phase sensitive effect of LSPR structure such as

periodic nanowire for a moment. We can carry out two sensitivity criterions to

compare sensing performance between structure of conventional SPR and

LSPR. The first is ∂φ/∂n that denoted differential phase shift for target RI

change, and the other is full with at half maximum (FWHM) of angular profile.

Former indicates that how phase shift are generated for target RI change, later

implies that how much angular margins that lead to reasonable sensing results

at least half phase shift.

６

Figure 2-1. Angular based phase shift at target RI change from 1.333 to 1.334 of (a) conventional scheme using gold thin film at thickness of 40 - 52 nm and (b) LSPR scheme using periodic nanowire periods of 300-400nm. In addition, Its phase variation at the target RI change at maximal resonance condition of (c) the conventional scheme at gold thin film and (d) the LSPR scheme.

７

Table 1. Maximal phase shift (upper row) and FWHM (bottom row) of the scheme of conventional thin film and periodic nanowire based LSPR structure regarding film thickness and structure period. (Unit: degree)

Conventional Period = 300nm Period = 400nm Period = 500nm

40 nm 6.2538 28.057 144.8507 63.7226

2.52 0.34 0.095 0.16

44 nm 12.7572 6.594 9.541 8.995

0.99 1.13 0.84 0.86

48 nm 162.0628 3.542 4.634 4.661

0.085 1.58 1.3 1.27

52 nm 13.12 2.283 2.896 2.998

0.56 1.85 1.58 1.53

Fig. 2-1 and Table 1 show numerical evaluation of phase sensitive

detection results. RCWA method was used in this simulation. Fig. 2-1 (a)

shows dramatic thickness dependence of a phase shift in the case of

conventional SPR. This result implies that each maximal phase shifts are

strongly dependant on the thin film thickness. In particular, appropriate

thickness was 48 nm. Fig. 2-1 (b) indicates each periodic nanowire case when

film thickness is 40 nm. Other cases of thickness are listed in Table 1. In this

result, 400 nm periodic nanowire shows is superior in the listed result, yet the

effect of that can’t exceed conventional performance. Figs. 2-1 (c) and (d)

show phase variation at target RI changes upon ambient condition using

conventional and periodic nanowire, respectively. Those are in the fixed

incident angle at resonance condition that indicates the maximal phase shift in

Figs. 2-1 (a) and (b). Here we can confirm a nonlinear variation tendency of

phase at fixed incident angle and ∂φ/∂n of each case is the largest at center of

RI change. In addition, Table 1 implies that FWHM is in inverse proportion to

the maximal phase shift, and an issue is that there is case that FWHM is less

８

than 0.1˚. Therefore, it is possible to causing a serious damage to the sensing

performance when the phase detection system employs a scheme which is

vulnerable to mechanical vibration or has sensing fluctuation range above 0.1˚.

Eventually, the phase sensing based LSPR structure could enhance

only in case the thin film thickness is equal to 40 nm. Therefore, use of

periodic nanowire structure is only rational to be applied when structure

conditions are specific and limited [20].

2.3. Sensitivity characteristics of nanostructure-based LSPR sensor In this section, sensitivity characteristics of LSPR sensors are

evaluated numerically using RCWA and FDTD methods. Formal is silver

islands structure that is used to randomly generate LSP mode. However, this

configuration had a limit of analysis using above evaluation methods since the

structure is formed randomly. In contrast, later, periodic gold nanowire

structure was properly adopted in this numerical analysis. For this reason,

feasible results were only presented about silver islands structure. Adenovirus

detection and DNA hybridization were assumed to biomolecular event for

characteristics analysis of silver islands and gold nanowire based LSPR

detection respectively.

2.3.1. Randomly distributed silver islands structure Randomly distributed silver islands structure generates stronger field

localization between each island than that of normally generated SP modes

from silver thin film. However, it is actually impossible that not only

numerical evaluation but also specific modeling due to random distribution of

islands. Therefore, surface near field profile was calculated using an actual

produced silver islands pattern.

９

The islands structure was fabricated simply by annealing process for a

little bit minutes. In detail, silver thin film upon SF10 slide glass was coated

by thermal evaporation. Then, the metal film heated using a hot plate at 175˚C

for 5min and cooled at room temperature. Fig. 2-2 shows scanning electron

microscopy (SEM) image of islands pattern by treating above process. A

shape of each island was round and embossing approximately, some got

tangled like chunks or preserved their shape. The average size and separation

was 100nm and 50nm, respectively.

Figure 2-2. SEM image of randomly distributed silver islands structure

１０

We calculated surface near field distribution of the islands pattern

using FDTD methods. In actual, it is proper to exam characteristics of periodic

structure rather than the islands structure. Therefore, the calculation period

was assumed to enough large values to avoid useless coupling effect as

comparing a size of detection target, adenovirus.

Adenoviruses are composed of a nucleocapsid and a double-stranded

DNA genome, which are responsible for 5–10% of upper respiratory

infections, and many infections in adults as well as children. Typical size of

adenovirus is range of 80-90 nm, a basic structure is Fig. 2-3. The image was

obtained from [21].

Figure 2-3. The structure of Adenovirus; (1) penton capsomeres, (2) hexon capsomeres, (3) viral genome (linear dsDNA)

Fig. 2-4 shows the exited surface near field profile at 60 degree

incidental TM wave. A number of local hotspots are confirmed within range of

1.5um ×1.5um plane of pattern. Diameters of hotspots are less than 100nm,

approximately. These are formed by field localization from LSP coupling. In

particular, overall size of each hotspot was well matched to average size of

one adenovirus. It is implies that the hotspots are appropriate to identify

１１

characteristics of individual adenovirus. In other word, this technique might be

possible to use toward to analysis the single molecule like adenovirus.

Especially, it enables us to improve resolution issue of total internal reflection

fluorescence (TIRF), etc. For example, TIRF microscopy adopted the islands

pattern can distinguish unit pixel from an image sensor more accurately and

brightly at localized hotspots areas to be compared with conventional TIRF

system since it generates only uniformly monotonous near field distribution.

Especially, labeled fluorescence dyes are significantly exited at localized

hotspot area, which can be effective to obtain enhanced resolution image.

However, the resolution enhancement is far from our overall subject related in

study on sensitivity characteristics of LSPR.

Figure 2-4. Magnetic field intensity profile at surface of silver islands based plasmonic sensor

１２

Experimental result proved that the islands structure contributes to

sensitivity enhancement of LSPR detection. Fig. 2-5 shows the reflectance

variation as change adenovirus concentration at fixed incident angle. In this

figure, an error bar means detected reflectance fluctuation when incident angle

is varied in range of ± 1˚.

Figure 2-5. LSPR and conventional SPR detection as change of adenovirus concentration

The result implicates intuitively sensitivity enhancement effect and

detailed measurement method will be introduced in next chapter. In this figure,

two configurations were not able to detect low concentration less than 106

units/μl. Especially, in conventional, reflectance varied unclearly as increased

adenovirus concentration until 1010 units/μl. These values were fluctuated

within the scope of error. Then, extremely rapid reflectance jump was found to

１３

concentration of 1011 units/μl. In contrast, in silver islands structure,

reflectance was increased constantly as increased adenovirus concentration

from 1011 units/μl. In addition, slope of conventional and islands pattern were

approximately 0.012 and 0.008, respectively, in the concentration range from

1010 to1011 units/μl. Thus, in the high concentration detection, conventional

would be more effective sensing tool because this reflectance difference was

larger as increased the virus concentration. However, island structure

guarantees low concentration range from 107 to1010 units/μl, it enhances by

thousand-fold compared with conventional structure. In this result, we can

induce that sensitivity enhancement effect was caused by interaction between

hotspot and adenovirus and there were equivalent size approximately.

Silver island structure based LSPR detection is beneficial to mass

production and industrialization since it can be made by relatively easy

procedure, annealing. However, this process is vulnerable of reproducibility.

In other word, this technique is not appropriate for LSPR detection due to a

significant issue with random condition.

2.3.2. Periodic gold nanowire structure We have confirmed briefly sensitivity characteristics using random

distributed silver islands based SPR detection. As a result, hotspots were

important factor related in reflectance based sensitivity enhancement. In this

section, we will deal with LSPR detection using periodic gold nanowire

structure since various enhancement factors as well hotspots issues can be

modeled and calculated by numerical methods mentioned above. For this

purpose, LSPR detection to measure DNA hybridization was modeled, which

was compared with conventional thin film based SPR detection. DNA

hybridization generally refers to a biology technique that measures the

１４

polymerization between each complementary nucleotide. This process is help

to identify genetic characterization of bacteria, virus, and etc.

In the event, the hybridization process was assumed as changing

ssDNA to dsDNA abbreviated single and double strand of DNA, respectively.

Optical constants of ssDNA and dsDNA were referred and obtained from [22]

as n = 1.45 and 1.52. Ambient condition was assumed distillated water nature.

The thickness of DNA layer was assumed to be 8nm and uniformly conformed

on the sensor surface although the thickness might be actually fluctuated due

to hybridization process in liquid.

Figure 2-6. (a) Schematics with a target binding DNA layer model of LSPR structure base on periodic nanowire and (b) Conventional thin film based SPR structure

Figs. 2-6 (a) and (b) present a schematic configuration of a LSPR

using gold nanowire structure and conventional thin film based SPR structure.

The thickness of the gold (dTF) and gold nanowire (dNW) are assumed as 40-50

nm and 10-20 nm, respectively. Chromium layer was considered as addition

layer for adhesion between metal and prism. Optical constants of both metal

and prism are obtained from [23]. In addition, period (Λ) and fill factor (f) of

１５

nanowire have also been considered as design parameters. In this scheme,

conjugated DNA layer was added on the sensor surface.

Here, the plasmonic sensitivity enhancement has been quantitatively

carried out by sensitivity enhancement factor (SEF) that is ratio of resonance

angle shift in terms of target binding on a LSPR sensing to that of a thin film

based conventional SPR sensing. SEF is defined as

[ ][ ]TF

NW

TF

NW

dsDNAssDNAdsDNAssDNA

SEF)()()()(

θθθθ

θθ

Δ−ΔΔ−Δ

=ΔΔ

= (9)

Δθ represents the resonance angle shift as a result of RI change from

ssDNA to dsDNA of DNA layer. This expression presents enhancement effect

for DNA hybridization process when we use the detection method of using

nanowire based LSPR compared with that of using thin film based SPR. The

notation is well known expression as reported research using periodic

nanowire structure [3,4,5,9,10,11,12,13,29,30]. SEF can be lower than 1,

which refers to negative effect of nanowire structure in terms of some special

cases.

In Figs. 2-6 (a) and (b), however, DNA layers are not identical

because of difference of sensor surface shape. This difference actually implies

that nanowire scheme can include much more DNA oligomers than

conventional SPR. Consequently, it is necessary to induce precisely a way to

compensate the difference of the sensing target volume for performance

analysis. Hence, we have newly employed an SEF per unit target volume

(SEFUTV), which is ratio of resonance angle shift per unit target of

cross-section of nanowire as the target volume. The SEFUTV is derived as

Λ+Λ

=ΛΔ

Λ+Δ=

NWTF

NWNWUTV d

SEFdSEF2/

)2/(θ

θ (10)

The SEFUTV maybe imply more reasonable analysis result than the SEF when

we consider near field effect which was dealt with previous section. Because

１６

this notation is appropriate to quantify the sensitivity enhancement of LSPR

sensing associated with all of target tiny molecular sensing as well as DNA

binding event. In particular, target localization effect related in The SEFUTV

for various sensing position was reported using self-assembled monolayer

(SAM) [10]. The SEFUTV will be revised at next section based on newly

considered fabrication.

The nanowire structure was selected for simulation by inducing SEF

and SEFUTV with few design parameter (dTF, dNW, Λ and f). The parameters

were varied in range of dTF = 40-50 nm, dNW = 10-20 nm, Λ = 300-500 nm and

f = 50-70%. These parameters were referred with respect to actual fabrication

process.

Figure 2-7. SEFUTV calculated for periodic nanowire structure for DNA hybridization. The row and column corresponds to dNW = 10, 20 nm and Λ = 300, 400, 500 nm, respectively.

１７

Fig. 2-7 shows the simulation results of various nanowire structures

for the various design parameters. First of all, we could confirm overall trends

of decrease SEFUTV with lager fill factor in terms of different dNW. In case of

dNW = 10nm, SEFUTV has a constant trend as increasing the fill factor. Instead,

SEFUTV tends to decrease slightly in near of 1.5 with lager Λ. However, thicker

dNW leads to dramatic SEFUTV variation in terms of fill factors compared with

thinner case. In particular, the structures with condition of Figs. 2-7 (d) and (e)

have not only maximum values at f = 50% among total result but also negative

values at f = 70%, negative effect corresponds to sensitivity degradation or

back bending effect. The optimum SEFUTV that was numerically described was

3.12 with design parameters of dTF = 45 nm, dNW = 20 nm, Λ = 400 nm and f =

50 %.

In the hybridization process, angular based far-filed characteristics of

LSPR applied the optimum parameters are compared with that of conventional

SPR as shown Fig. 2-8. In this figure, each change from solid to dashed line

implies reflectance variation due to the hybridization form ssDNA to dsDNA.

In Fig. 2-8, LSPR structure leads to broadening and increased reflectance

variation at resonance angle (θSP) compared with conventional SPR structure.

θSP of LSPR structure was larger than that of conventional SPR structure. This

far-field analysis is obvious to investigate shift of ΔθSP which corresponds to

sensing signals.

１８

Figure 2-8. DNA hybridization far-field characteristics of angular based LSPR using periodic nanowire and conventional SPR structure using thin film

In addition to far-field analysis, we can concentrate on near-field

characterization. Fig. 2-9 presents the distribution of magnetic field intensity

(|Hy|2) of the near-field in the course of the DNA hybridization. Figs. 2-9 (a)

and (b) show the field formed by the optimum structure at θSP of incident (θSP

= 68.45˚ at nanowire and θSP = 60.58˚ at conventional structure) in medium.

Contrary to uniformly distributed in terms of conventional of Fig. 2-9 (b), the

fields were localized at both side of nanowire in Fig. 2-9 (a). Especially,

intensity of primary peak is larger than that of secondary peak; this position

depends on the incident direction. In addition, conventional SPR structure

leads to matching completely DNA layer to localized hotspots; however LSPR

structure enables only small hotspots to coincident side target DNA layer.

Despite this unfair condition, LSPR detection induced larger ΔθSP as change

１９

from the hybridization process than conventional SPR detection. Here, larger

ΔθSP can be explained by target volume difference with LSPR and

conventional structure since larger volume makes more broaden ΔθSP.

However, it is not easy to explain the large SEFUTV of LSPR using only this

volume difference. Fig. 2-9 (c) proves that extraordinary amplified localized

fields contribute sensitivity enhancement of the far-field analysis with respect

to ΔθSP. As shown in Fig. 2-9 (c), z-cross profile at hotspot of LSPR structure

and normal penetration of conventional SPR are plotted. Both maximum fields

are located at near dNW = 45 nm. This obviously corresponds with sensitivity

enhancement related in randomly distributed amplified hotspot that was dealt

with chapter 2.3.1. However, in LSPR structure; field intensity is rapidly

decayed as increase z-axis compared to the conventional structure. In detail,

field intensity was decayed and halved at approximately 10 nm and 50nm

from the maximum point in case of LSPR and conventional SPR, respectively.

In other word, these results suggest that available target size for LSPR is

limited by rapid field decaying. However, this truth suggests that a LSPR

method might be more sensitive to detect tiny molecules less than the

limitation more than a conventional SPR scheme.

２０

Figure 2-9. Magnetic field intensity (|Hy|2) distribution at θSP of (a) LSPR and (b) conventional SPR structure (c) z axis cross field profiles on (a) and (b)

２１

Chapter 3 Experimental study on highly sensitive LSPR detection based on target localization using periodic nanowire structure It is necessary to match target localization area to hotspots for

investigating highly sensitive LSPR detection when we judged many results in

previous chapter. For this purpose, we will deal with DNA hybridization

detection using periodic nanowire structure with newly additional fabrication

for target localization. In addition, numerical enhancement results as various

sensor design parameter and experimental evaluations for optimum structure

are presented as following.

3.1. Numerical modeling 3.1.1. Sensor design We examine the feasible study for this approach experimentally, where

target molecular interactions was designed to occur only in the strongly

localized evanescent near field, i.e. hotspots. In particular, we modeled

localized target on a nanowire which has dealt with chapter 2.3 using angled

evaporation to add a SiO2 dielectric mask to obtain a small opening. Hence,

we attempted to make highly sensitive LSPR detection by coinciding localized

DNA area and hotspots excited by LSP. Note that adopting periodic nanowire

of matalodielectric double layers consisting of a dielectric opening that

separates each nanowire and a metallic thin film would enhance the sensitivity

of detection by manipulating an additional control over the coupling effect

between LSP modes and propagating SPs.

２２

Figure 3-1. 3-D Schematic of a periodic nanowire LSPR sensor with design parameters, the thick arrows represent the evaporation angle of SiO2.

Fig. 3-1 represents a schematic diagram of the periodic nanowire

fabricated by angled evaporation. θeva indicates the evaporation angle.

Evaporated dielectric thickness is varies at side and top of nanowire in terms

of shot thickness of mask layer (dm) and θeva. It prevents non-specific binding

between thiolated probe ssDNA bare gold surface that is not modified by SiO2

[23]. Only the openings beside side of nanowire are opened to DNA

immobilization. Fig. 3-1 indicates the helixes in which immobilized and

hybridized DNAs. Other design parameters are same as chapter 2.

２３

3.1.2. Sensitivity enhancement factor As mentioned, we have newly employed an SEFUTV. Because an SEF

can’t represent the amount of target molecules that is needed to achieve a

resonance shift, SEFUTV can be inevitably more appropriate to molecular

sensing. Note that for newly modified periodic gold nanowire structure,

SEFUTV may be approximated as an SEFUTV, i.e.,

evaTF

evaNWNWNWUTV SEF

ddSEF

θθθθ

tan1/)tan/(

+Λ

=ΛΔ

+Δ≈ ,

(11),

which explains the dependence of SEFUTV on various fabrication parameters.

3.2. Materials and methods 3.2.1. Sensor fabrication For fabrication of designed periodic nanowire structure, a gold film

(thickness: dTF = 40 nm) was basically evaporated on an SF10 substrate. Gold

was considered only as a metal for nano-plasmonic sensor due to the superior

biocompatibility. Periodic nanowire was then patterned by electron beam

lithography method with the optimum fabrication parameter. Then, SiO2 was

evaporated on the sample by angled-mounting by θeva = 30°. An SEM image

of the fabricated nanowire is shown in Fig. 3-2 (a). Approximately 11-nm

wide openings were created at left side of nanowire by the angled evaporation.

A magnified SEM image of the periodic nanowire sample in Fig. 3-2 (b)

confirms the opening to expose DNA hybridizations.

２４

Figure 3-2. (a) An SEM image of the periodic nanowire sample after angled evaporation. (b) A magnified SEM image of the sample confirms the existence of the dielectric opening by angled evaporation.

3.2.2. Optical setup For experiment, a SPR instrument was set up, as shown in Fig. 3-3,

with dual motorized rotation stages (URS75PP, Newport, Irvine, CA). Light

source from a He-Ne laser (36 mW, λ = 632.8 nm, Melles-Griot, Carlsbad, CA)

is TM polarized coherent beam and incident on the chip from a SF10 prism.

Incident angle (θin) was scanned by a rotation stage, while a connected

photodiode (818-UV, Newport, Irvine, CA) was rotated by same θin of

opposite direction to measure the reflected light. The angular resolution of the

rotation stage was 0.0002°. The measured signal was taken by a lock-in

amplifier (Model SR830, Stanford Research Systems, Sunnyvale, CA) to

compensate noise. The measurement procedure is controlled by LabView™

program (National Instruments, Austin, TX). Each measurement was taken 5

times and averaged. The resolution of the SPR sensor set-up is estimated to be

Δn ~ 2 x 10-6 in RIU.

２５

Figure 3-3. SPR system set-up (PO: polarizer, CH: chopper, M: mirror, PD: photodiode, and PC: personal computer).

3.2.3. DNA hybridization process Complementary 24-mer length of sequence are following:

5΄thiolated-TTT TTT CGG TAT GCA TGC CAT GGC-3΄ and

complementary 5΄- GCC ATG GCA TGC ATA CCG AAA AAA-3΄ (Bioneer

Co., Daejeon, Korea), abbreviated as HS-ssDNA and C-ssDNA,

respectively.

DNA hybridization protocol are referred to [25,26] in detail. A

conventional thin film based chip and a periodic nanowire chip were used in

this experiment. Both samples were cleaned by sonication in acetone for 10

min and in 70% ethanol for another 10 min. Then, they were surface-treated

for attaching the gold surface to thiol compound at HS-ssDNA end using

plasma cleaner (Harrick Scientific Products, USA). In immobilization of

HS-ssDNA, the processed chips were dipped in 1.0-M KH2PO4 buffer mixed 2

μM HS-ssDNA at 4ºC for 5 hours. From this process, DNA immobilized chips

were treated in 1.0 mM 6-mercapto-1-hexanol (MCH) (Sigma-Aldrich) in

２６

ethanol for 1 hour. Then, the chips were incubated in TE buffer (10 mM Tris

buffer (pH 7.6) and 1 mM EDTA) (Bioneer Co., Daejeon, Korea) at 65ºC.

Also, TE buffer mixed with 2 μM C-ssDNA and1 M NaCl solution was

conserved in water at 65ºC for hybridization experiments. In hybridization

process, the C-ssDNA solution was injected onto the DNA chip by a fluidic

system for 200 min using the syringe pump (KD Scientific Inc., Holliston,

MA). Pumping rate was at 10 μl/min. After hybridization, TE buffer was

injected to remove the DNA chip from the residue C-ssDNA particle in

medium. Because thiolated DNA has a tendency to immobilize onto gold

rather than to SiO2 and residue HS-ssDNA is washed out in TE buffer

injection process after immobilization, non-specific binding event of DNA to

SiO2 was neglected.

3.3. Result and discussion 3.3.1. Numerical analysis The periodic nanowire was selected for fabrication by calculating SEF

and SEFUTV with the design parameters (dTF, dNW, dm, Λ, and f). Especially,

parameters were varied in the range of dm = 1 to 5 nm, θeva = 30° to 60° and

the others were varied same range of chapter 2.

Fig. 3-4 represents a few inductions in a worth result. Moreover,

overall increased trends of SEFUTV with smaller θeva were critical. Target DNA

molecules can be better localized with smaller values of θeva, which affects

SEFUTV to increase. When effect of damping affects the SPs characteristics,

high f was predominant issues, for instance, in the case of dTF = 50 nm and

dNW = 20 nm, the overall trends were reversed. In many cases, damping gives

to broaden effect of the reflectance for θin so much, so that the structure can’t

be serving as a good SPR biosensor any more.

２７

In Fig. 3-4, thicker dNW and dm were observed to increase SEFUTV,

because of amplified localized field, although the effect strongly depends on

the specific position of a nanowire. Also, a thinner gold film based nanowire

structure makes higher SEFUTV effect, as listed in Table 2. The preference of a

thinner gold film for higher SEFUTV is related to an increase of damping effect,

as the sensitivity enhancement is often correlated with strongly excited

damping effect. The increase factor of SEFUTV by a thin gold film only appears

to be independent of any other parameters, because the parameters other than

dTF do not affect the optimum film thickness. The optimum design parameter

and SEFUTV that was determined was 12.02 for dTF = 40 nm.by numerical

methods, which were listed in Table 2. Ten times of enhancement is very

important, because it is quite difficult challenges to attain significance

improvement in liquid medium due to reduced index contrast [9].

２８

Figure 3-4. SEFUTV calculated for various designs of target-localized periodic nanowire structure. x- and y-axis indicate evaporation angle θeva in degrees and SEFUTV, respectively. The row corresponds to (dTF, dNW) = (40 nm, 10 nm), (40 nm, 20 nm), (45 nm, 10 nm), (45 nm, 20 nm), (50 nm, 10 nm), and (50 nm, 20 nm) from top to bottom. The columns represent dm = 1, 3, and 5 nm (from left to right). Λ in nm and f in %.

２９

Table 2. SEFUTV for various gold film thicknesses. Theses parameters other than dTF have been considered from the results presented in Fig. 3-4

dTF dNW dm Λ f θeva SEFUTV

40 nm

20 nm 5 nm 400 nm 50% 30°

12.02

45 nm 11.55

50 nm 11.32

Figs. 3-5 (a)-(c) show the magnetic field distributions (|Hy|) for dTF =

40, 45, and 50 nm, whereas other parameters are same as listed in Table 2.

Figs. 3-5(d)-(e) provide horizontal and vertical field profile (|Hy|) which point

maximum values. In Fig. 3-5(d), the distribution shows typically a couple of

localized hotspots: a primary peak formed at the left-hand side which exists

opening and a secondary peak formed at the other side. The intensity ratio of

the primary to secondary peak was 2.27, 2.08, and 1.91 for dTF = 40, 45, and

50 nm, respectively. The intensity of primary peak was approximately 2 times

larger than that of secondary peak. In addition, the values were strongly

dependant on dTF variation. Therefore, the increased radiation damping related

in SP modes makes an negative effect of the near field amplification for larger

dTF, and is also associated with less efficient sensitivity enhancement.

In addition, note that the target openings induce effect of additional

field localization compared to a structure without the openings by complete

evaporation of SiO2. In terms of the primary peak field comparison, the

additional enhancement by the opening to be 8.5% compared to using a

complete SiO2 covered chip.

３０

Figure 3-5. Magnetic field amplitude profiles (|Hy|) calculated for dsDNA: (a) dTF = 40 nm, (b) 45 nm, and (c) 50 nm. Other parameters are same as those listed in Table 2. The arrow indicates the incidence direction. (d) Horizontal and (e) vertical magnetic field distribution show that increased radiation damping decreases the near field for larger dTF and is associated with less efficient sensitivity enhancement.

Fig. 3-6 shows the distribution of magnetic field (|Hy|) in the near-field

in response to DNA hybridization process from ssDNA to dsDNA. Figs. 3-6 (a)

and (b) represent the fields distribution formed by the optimum structure (dTF

= 40 nm) at resonance angle of incidence (θin = 66.29° for ssDNA and 66.60°

for dsDNA) in TE buffer. Fields are localized in a small area of hotspots near

side of nanowire. The near-field variation induced by the DNA hybridization

can’t be appeared significant if Figs. 3-6(a) and (b) are feasibly compared.

Once ssDNA is hybridized to dsDNA, both the two peaks of field near

nanowire ridge slightly increase, whereas the hotspots are more localized in

the primary peak. In Fig. 3-6(c), it provides the numerical differences between

the near fields visualized in Figs. 3-6(a) and (b). The minute difference in the

fields can be induced as a significant change related in a molecular interaction,

３１

since the interaction event occurs in the opening near primary peak that

roughly overlaps the hotspots area of large difference of magnetic field. This

implies that the lager field difference with smaller dTF, corresponds tendency

of SEFUTV presented in Fig. 3-4.

x [nm]

(a)

x [nm]

z [n

m]

(c)

z [n

m]

x [nm]

(b)

z [n

m]

Figure 3-6. Magnetic field amplitude distributions (|Hy|) calculated for the optimum structure with parameters listed in Table 2 for dTF = 40 nm: from (a) ssDNA to (b) dsDNA. The arrow indicates the light incidence. (c) Field difference between the near-fields of (a) and (b), where the arrow indicates the field difference at localized hotspots formed in the position of target dielectric opening.

３２

3.3.2. Detection of DNA hybridization Fig. 3-7 shows the reflectance based resonance characteristics

measured by the fabricated periodic nanowire sample and a conventional thin

film control (dTF = 40 nm). Measured angle shifts were ΔθNW = 0.17° ± 0.05°

and ΔθTF = 0.05° ± 0.01°, i.e., SEF = 3.4 ± 2.1. In other words, angular

sensitivity has 3.4 times of enhancement factor. This implies to SEFUTV = 65 ±

40. This result confirms that the significant enhancement is induced by

localizing of target interactions at hotspots using periodic nanowire structure.

Note that the measured SEFUTV was larger than predicted SEFUTV mainly since

a larger ΔθNW and a smaller ΔθTF give a result of larger SEF than a numerical

result.

A smaller ΔθTF obtained experimentally might be related to the some

limited factors of the numerical modeling, for example, the model did not

consider the concentration of DNA layer and it dealt with DNA hybridization

only as the RIs variation. In addition, a larger ΔθNW can be interpreted partly a

cause of the angled evaporation process. The dielectric opening size might

have been narrower than designed due to angled evaporation nature. The

narrower opening may have an amplifying the hotspots even more expectation.

In addition, denser ssDNA molecules might have attached at smaller opening

for immobilization.

３３

Figure 3-7. Measured reflectance for resonance characteristics of hybridization process on the periodic nanowire and a thin film control

Figure 3-8. Real-time reflectance dynamics of DNA hybridization measured by the periodic nanowire and the control sample, the incident angle was fixed at θin = 60°.

３４

Real-time reflectance dynamics of DNA hybridization were detected

by measuring reflected light from the sample at θin = 60°, as shown in Fig. 3-8.

Hybridization kinetics was continually tracked to fulfill stabilized interactions

between each complementary ssDNA stands. A drop at 800 sec in the

reflectance was result of the injection of C-ssDNA solution of TE buffer

mixed NaCl. As the C-ssDNA injection was stopped at around 12,000 second

later, the reflectance jump was observed as normal TE buffer. The slope (S =

ΔR/Δt) of reflectance during DNA hybridization process was constant for both

control and periodic nanowire indicating a fixed state of an interaction and

provides information on the hybridization kinetics between C-ssDNA and

HS-ssDNA. Measured slope of reflectance increase is S = -5 pW/sec and -10

pW/sec for the periodic nanowire and control sample, respectively. In this

evaluation, S can’t be necessarily representative detection of the enhanced

sensitivity. S is actually smaller for the periodic nanowire sample, because its

reflectance based resonance characteristics are broader due to strong damping

effect and thus a smaller reflectance change is induced by an interaction in

spite of a larger θSP shift. Also, the measured reflectance has a higher than that

of a conventional thin film based control, which was detected higher

reflectance for a periodic nanowire sample as shown in Fig. 3-8.

３５

Figure 3-9. DNA hybridization in the reverse incidental configuration, the sample is mounted in the reverse orientation. Calculated near field profile of (|Hy|): (a) ssDNA and (b) dsDNA, (c) Field difference between the near-fields of (a) and (b), where the arrow indicates the field difference at localized hotspots formed in the position of target dielectric opening. (d) Measured reflectance based resonance characteristics.

３６

Other experiments were additionally conducted to evaluate reverse

configuration. In this event, the beam was incident from the reverse direction,

i.e., the incident angle was negative so that the secondary hotspot coincides

with DNA hybridization as shown in Fig. 3-9. Reflectance based resonance

characteristics were measured as an angular shift by Δθsn = 0.06° ± 0.05°. Note

that the data of control should be same as that of previous experiment for thin

film. Thus, a smaller SEF was obtained as SEF = 1.2 and SEFUTV should be 23

under numerical assumption, which is still significant because of the

secondary hotspot that exists where target is localized.

3.3.3. Discussion In this section, I want to concentrate on the implication of SEF in detail.

The LPSR detection has been related in the increase of detection volume and

the enhanced surface sensitivity due to the excitation of localized hotspots. In

this study, we conducted angled evaporation to make dramatically the

localized sensing space to overlap hot spots excitation area. The SEF was

numerically expectation value that presumes an evenly distributed molecular

interaction. In this respect, when calculating SEF, it was an indicator of the

amount of reaction molecules that participate in a binding to cause a unit

resonance angular shift and thus SEFUTV, i.e., the efficiency of angular shifts

per unit volume compared with conventional thin film based sensor

The detection criterion that expresses the sensitivity limitation of 1

ng/mm2 as 0.1º shifts eventually of the resonance angle from molecular

interaction on the sensing surface [27,28]. Based on the criterion, the number

of DNA molecules in case of the control experiment on the conventional SPR

detection can be estimated as 3.91 x 108 units assuming a beam for detection

area was collimated by forms 100 μm × 100 μm. Because the surface where

probe DNA helixes are immobilized on the opening of the ridge of nanowire

３７

became smaller approximately by 7% compared to that on the conventional

thin film, it can be induced that 0.02 ng/mm2 results in 0.1º resonance angle

shift on the sensing surface matched localized hotspots of the periodic

nanowire.

Instead, assuming SPR detection system with performance of 1 pg/mm2,

the nanowire structure suggests roughly 20 fg/mm2 sensing efficiency, if

localized molecular interaction in the dielectric opening for subsequent SPR

measurement is presumed.

The cause for significant sensitivity enhancement reported in this thesis

can be mainly the hotspots from localized field. Due to the limitation of target

area, fewer binding event can be detected by an appreciable signal. Note that

practical applications without the localization of target have been improved by

two fold of sensitivity enhancement in numerical and experimental way

[9,29,30]. Another issue is that denser DNA may immobilized on the opening

compared with area of conventional thin film, since a relatively limited

adsorption at target opening results in staying a lasting number of DNA

molecules for binding event in liquefied nature throughout the process.

This result suggests that the detection by LSPR with target localization

doesn’t merit of revealing individual or interactions of molecules effectively

so far. Nevertheless, there is still a plenty of room by some improvement.

Assuming that target molecules of 0.1 pg/mm2 can be detectable and that this

study realize sensitivity enhancement by an order-of-magnitude increasing

using periodic nanowire including target localization, a very cautious estimate

would make the sensitivity limit achievable toward to 1 fg/mm2 in the near

future. In appropriate optical system with well-maintained related in issues

such as temperature and humidity, potentially 800 DNA molecules can be

detectable.

３８

Chapter 4 Conclusion

The thesis aims to analysis the fundamental and sensitivity

characteristics of LSPR detection and to investigate high sensitive LSPR

scheme using nanostructure such as randomly distributed nanoislands and

periodic nanowire.

Silver nanoislands structure generates localized hotspots distributed

randomly on the structure since each island is randomly distributed on the film.

Nevertheless, this was effective tool as LSPR detection to detect tiny

molecules like adenovirus. However, the islands structure can’t be appropriate

LSPR detection methods due to the issues related in weak of reproducibility

and numerical modeling. Instead, the structure would be used to overcome

resolution failure including the blurring effect in fluorescence-based imaging

sensor because randomly localized-hotspots in islands could significantly

enhance fluorescence excitation.

In contrast, it is useful LSPR detection using periodic nanowire based

on practical fabrication. By using RCWA and FDTD method, optimum design

parameters were obtained as the modeling of DNA hybridization process. In

this event, we have confirmed strong relation with near-field and far-field to

induce amplified SPR detect signal. In other word, broader plasmon angular

shift could be detectable compared with conventional thin film based SPR

detection.

Based on above analysis, we have investigated the high sensitive LSPR

detection using periodic nanowire structures. Using numerical methods, we

modeled a nanowire structure in which event related in molecular interactions

are localized in hotspots by only LSPs modes. These results to optimize

nanowire structures proved that sensitivity enhancement by an order of

magnitude and it may be feasible on a per-unit-volume basis. Then, we used

３９

angled evaporation to make target-localized periodic nanowire samples for

experimental way and employed DNA hybridization process as the target

interaction model to confirm the sensitivity enhancement. Eventually, the

measured data gave a proof that the LSPR detection scheme shows feasibility

of individual molecular detection for sensitivity enhancement.

４０

References [1] L. He, M. D. Musick, S. R. Nicewarner, F. G. Salinas, S. J. Benkovic, M.

J. Natan, and C. D. Keating, “Colloidal Au-enhanced surface plasmon

resonance for ultrasensitive detection of DNA hybridization,” J. Am. Chem.

Soc., vol. 122, no. 38, pp. 9071-9077, Sep. 2000.

[2] B. Sepúlveda, A. Calle, L. Lechuga, and G. Armelles, “Highly sensitive

detection of biomolecules with the magneto-optic surface-plasmon-resonance

sensor,” Opt. Lett., vol. 31, no. 8, pp. 1085-1087, Apr. 2006.

[3] K. Byun, S. Kim, and D. Kim, “Design study of highly sensitive

nanowire-enhanced surface plasmon resonance biosensors using rigorous

coupled wave analysis,” Opt. Express, vol. 13, no. 10, pp. 3737-3742, May

2005.

[4] K. Kim, S. J. Yoon, and D. Kim, “Nanowire-based enhancement of

localized surface plasmon resonance for highly sensitive detection: a

theoretical study,” Opt. Express, vol. 14, no. 25, pp. 12419-12431, Dec. 2006.

[5] K. M. Byun, S. J. Yoon, D. Kim, and S. J. Kim, “Experimental study of

sensitivity enhancement in surface plasmon resonance biosensors by use of

periodic metallic nanowires,” Opt. Lett., vol. 32, no. 13, pp. 1902-1904, Jul.

2007.

[6] L. Malic, B. Cui, T. Veres, and M. Tabrizian, “Enhanced surface plasmon

resonance imaging detection of DNA hybridization an periodic gold

nanoposts,” Opt. Lett., vol. 32, no. 21, pp. 3092-3094, Nov. 2007.

[7] H. P. Ho, W. C. Law, S. Y. Wu, C. Lin, and S. K. Kong, “Real-time optical

biosensor based on differential phase measurement of surface plasmon

resonance,” Biosens. Bioelectron., vol. 20, no. 10, pp. 2177-2180, Apr. 2005.

[8] S. Y. Wu, H. P. Ho, W. C. Law, C. Lin, and S. K. Kong, “Highly sensitive

differential phase-sensitive surface plasmon resonance biosensor based on the

４１

Mach-Zehnder configuration,” Opt. Lett., vol. 29, no. 20, pp. 2378-2380, Oct.

2004.

[9] S. J. Yoon and D. Kim, “Target dependence of the sensitivity in periodic

nanowire-based localized surface plasmon resonance biosensors,” J. Opt. Soc.

Am. A, vol. 25, no. 3, pp. 725-735, Mar. 2008.

[10] K. M. Byun, S. M. Jang, S. J. Kim, and D. Kim, “Effect of target

localization on the sensitivity of a localized surface plasmon resonance

biosensor based on subwavelength metallic nanostructures,” J. Opt. Soc. Am.

A, vol. 26, no. 4, pp. 1027-1034, Apr. 2009.

[11] D. Kim and S. J. Yoon, “Effective medium-based analysis of

nanowire-mediated localized surface plasmon resonance,” Appl. Opt., vol. 46,

no. 6, pp. 872-880, Feb. 2007.

[12] D. Kim, “Effect of resonant localized plasmon coupling on the sensitivity

enhancement of nanowire-based surface plasmon resonance biosensors,” J.

Opt. Soc. Am. A, vol. 23, no. 9, pp. 2307-2314, Sep. 2006.

[13] D. Kim, “Nanostructure-based localized surface plasmon resonance

biosensors,” in Optical Guided-Wave Chemical and Biosensors, Springer

Series on Chemical Sensors and Biosensors, vol. 7, M. Zourob and A.

Lakhtakia, Eds. New York: Springer, 2009, in press.

[14] J. Cesario, R. Quidant, G. Badenes, and S. Enoch, “Electromagnetic

coupling between a metal nanoparticle grating and a metallic surface,” Opt.

Lett., vol. 30, no. 24, pp. 3404-3406, Dec. 2005.

[15] S. M. Rytov, “Electromagnetic properties of a finely stratified medium,”

Sov. Phys. JETP, vol. 2, no. 3, pp. 466-475, May 1956.

[16] P. Lalanne and D. Lemercier-Lalanne, “On the effective medium theory of

subwavelength periodic structures,” J. Modern Opt., vol. 43, no. 10, pp.

2063-2085, Oct. 1996.

４２

[17] G. T. Hermanson, Bioconjugate Techniques. London: Academic Press,

1996.

[18] S. Moon and D. Kim, “Fitting-based determination of an effective

medium of a metallic periodic structure and application to photonic crystals,” J.

Opt. Soc. Am. A, vol. 23, no. 1, pp. 199-207, Jan. 2006.

[19] P. Lalanne and M. Hutley, “The optical properties of artificial media

structured at a subwavelength scale,” in Encyclopedia of Optical Engineering,

R. Driggers, Ed. Marcel Dekker: New York, 2003, pp. 62-71.

[20] K. Ma, D. J. Kim, and D. Kim, "Numerical evaluation of periodic

nanowire-based phase sensitive surface plasmon resonance detection," Current

Applied Physics, 9(4), Suppl. 1, e239-e242, July, 2009 (Nanobio-Seoul 2008

Special Issue).

[21] http://en.wikipedia.org/wiki/Adenovirus

[22] S. Elhadj, G. Singh, and R. F. Saraf, “Optical properties of an

immobilized DNA monolayer from 255 to 700 nm,” Langmuir, vol. 20, no. 13,

pp. 5539-5543, Jun. 2004.

[23] E. D. Palik and G. Ghosh, Handbook of optical constants of solids.

Orlando; London: Academic Press, 1985.

[24] G. T. Hermanson, Bioconjugate Techniques. London: Academic Press,

1996.

[25] T. M. Herne and M. J. Tarlov, “Characterization of DNA probes

immobilized on gold surfaces,” J. Am. Chem. Soc., vol. 119, no. 38, pp.

8916-8920, Sep. 1997.

[26] A. W. Peterson, L. K. Wolf, and R. M. Georgiadis, “Hybridization of

mismatched or partially matched DNA at surfaces,” J. Am. Chem. Soc., vol.

124, no. 49, pp. 14601-14607, Dec. 2002.

[27] L. S. Jung, C. T. Campbell, T. M. Chinowsky, M. N. Mar, and S. S. Yee,

“Quantitative interpretation of the response of surface plasmon resonance

４３

sensors to adsorbed films,” Langmuir, vol. 14, no. 19, pp. 5636-5648, Sep.

1998.

[28] D. K. Roper, “Determining surface plasmon resonance response factors

for deposition onto three-dimensional surfaces,” Chem. Eng. Sci., vol. 62, no.

7, pp. 1988-1996, Apr. 2007.

[29] K. M. Byun, D. Kim, and S. J. Kim, “Investigation of the profile effect on

the sensitivity enhancement of nanowire-mediated localized surface plasmon

resonance biosensors,” Sens. Actuators B, vol. 117, no. 2, pp. 401-407, Oct.

2006.

[30] K. M. Byun, S. J. Kim, and D. Kim, “Profile effect on the feasibility of

extinction-based localized surface plasmon resonance biosensors with metallic

nanowires,” Appl. Opt., vol. 45, no. 14, pp. 3382-3389, May 2006.

４４

국문요약

국소 나노플라즈모닉 바이오센서의 광학적 특성 및

측정 감도 개선 연구

표면 플라즈몬 공명은 일정 광 입사 조건에서 금속 층에 전자적

표면진동파가 형성되는 현상으로, 일반적으로 극 미량 물질의 검출 및 이

미징을 위한 센서로서 이용되고 있다. 본 논문은 표면 플라즈몬 공명현상

을 이용한 국소나노플라즈모닉 센서의 측정 감도를 높이기 위해 진행되었

다. 이를 위해 이론적 접근을 비롯하여 실제 센서 제작 공정 및 실험을 통

해 그 효과를 입증하였다.

먼저 다음의 목적을 이론적으로 접근하기 위해 맥스웰 방정식에

기초한 엄 결합파동분석법 (Rigorous coupled wave analysis) 과 유한 요소

법 (Finite-difference time-domain)을 이용한 원거리 및 근거리 장의 반사광

특성 분석을 하였다. 이를 이용하여, 랜덤하게 분포하는 은 나노섬

(Nanoislands)구조의 경우 기판에 형성되는 국소 근거리장의 분포가 센서의

민감도 개선에 이용될 수 있음을 확인 할 수 있었다. 무질서 하게 분포하

는 패턴은 정해진 구조적 파라미터가 없기 때문에 반도체 공정으로 얻은

구조의 SEM 이미지를 토대로 직접 시각화 모델링 하였다. 센서의 근거리

장 분포를 분석한 결과, 센서의 표면 곳곳에 일부 강한 세기의 여기장 성

분 (Hotspots) 들이 형성 됨을 확인 할 수 있었다. 이러한 요인들은 기존의

단순박막구조와 비교해 볼 때 보다 작은 물질을 감지하는데 효과적이었으

며, 특히 실제 아데노바이러스 (Adenovirus) 측정을 통해 보다 저 농도의

물질을 감지할 수 있음이 확인 되었다. 또한 이미징 개선 측면에서 볼 때

에 이미지의 퍼짐현상 (Blurring)을 개선하기 때문에, 해상도 개선 효과를

기대 할 수 있다.

한편 나노선 구조의 경우 구조 자체의 주기적 성질와 제작의 용이

４５

함으로 인해 다양한 공정 파라미터에 대한 민감도의 계산이 가능했다. 원

거리장 분석 결과 최대 3배 이상의 센서의 민감도 개선을 확인 할 수 있었

으며, 이것은 강한 국소 표면 플라즈몬 공명으로 인한 반사광의 감쇄 현상

과 공명각 변화 차이에 기인한다. 특히 근거리장 분석을 통해 나노선 양단

의 접합면 부근에 강한 국소 여기장이 형성되는 것을 확인할 수 있었다.

마지막으로 국소화된 여기장은 센서의 감도 향상에 영향을 미친다

는 이론적 결과를 토대로 고감도 측정을 위한 국소표면 플라즈몬 센서를

제작하였다. 특히, 강한 여기장이 발생하는 부분에만 선택적인 감지가 일어

나도록 금속 나노선에 기울임 증착을 통한 유전체 마스크층을 형성하였다.

이렇게 제작된 센서는 DNA 중합 반응 실험 결과 기존 대비 약 64배 이상

의 단위 접합면 당 공명 각 변화 감도 개선효과를 가져 왔다. 결과적으로

나노선 구조는 특성화 분석, 모델링 및 감도 개선 정량화에 유리하며 마스

크층을 이용한 추가적 감도 개선효과를 얻을 수 있음을 확인 할 수 있었

다.

핵심되는 말: 나노플라즈모닉, 바이오센서, 민감도 개선, 나노구조, 국소단

위측정

４６