SERS物理增强机理(精)

Simulation of Raman Enhancement in SERS-Active Substrates with Au Layer

Considering Different Geometry of Nanoparticles

Hui-Wen Cheng1 and Yiming Li1,2,3,*

Institute of Communications Engineering, National Chiao Tung University, 1001 Ta-

Hsueh Road, Hsinchu 300, Taiwan 2

Department of Electrical Engineering, National Chiao Tung University, 1001 Ta-Hsueh

Road, Hsinchu 300, Taiwan

3

National Nano Device Laboratories, Hsinchu 300, Taiwan

*

Tel: +886 3 5712121 ext 52974; Fax: +886 3 5726639; E-mail:

ymli@

Abstract-In this work, we study surface enhanced Raman identification of Rhdamine

6G (R6G) are examined. This paper spectroscopy (SERS) active substrates for the

detection of is organized as follows. In Sec. II, we introduce the fabrication Rhodamine

6G. To examine the electromagnetic enhancement, process and computational

technique for the SERS- active with different shape of nanoparticle, we apply the finite-

substrates. In Sec. III, the local field enhancements of

difference timedomain (FDTD) algorithm to analyze the nanoparticle with different

shapes are calculated by three-

structures by solving a set of coupled Maxwell’s

equations in dimensional (3D) finite-difference time-domain (FDTD) differential form.

The field enhancements are investigated in the numerical simulation. Finally, we draw

the conclusions and visible regime with the wavelength of 633 nm. In the suggest the

future work.

experimental measurement, the surface enhanced Raman scattering signals from the

surface of substrates with 12-hour

II. FABRICATION AND COMPUTATIONAL TECHNIQUE hydrothermal treatment

and without treatment are performed

and compared. Through the three-dimensional (3D) FDTD For the flow of fabrication,

as shown in Fig. 1, first, calculation, the enhancements with different shape of buffered

oxide etchant (BOE) and standard RCA cleaning are

nanoparticle are tested and obtained which are nanoparticle, carried out to prepare clean

silicon substrates (Boron-doped

gold nanocage and gold/silver alloy for spherical, cubic and pyramidical shapes. The

results show that the enhancement of

(i)spherical and cubic shapes can be much improved by nanocage

and gold/siliver alloy structures.

Keywords- Surface-Enhanced Raman spectroscopy (SERS), electromagnetic

enhancement, nanoparticle, gold nanocage, gold/silver alloy, finite-difference time-

domain, hydrothermally treated substrate.

(ii)

I. INTRODUCTION

Surface-enhanced Raman Scattering (SERS) is one of the characterization techniques,

which is sensitive to the enhanced electromagnetic fields [1-6]. SERS-active substrates

have recently attracted a great deal of attention for rapid identification of chemical and

bacterial samples [5-7]. The fabricated nanostructures for both bottom-up and top-down

approaches have been reported. And, the degree of Raman enhancement is strongly

dependent on the morphology of formulated nanostructures [8]. Recently, a top-down

approach for the fabrication of SERS-active substrate was proposed [9-12]. However,

the expensive substrate, equipments and complicated process are needed. Therefore, a

low cost, environment friendly and simple fabrication for SERS-active substrates will

be of great interest for basic and clinical researchers as well as for biotechnologies. In

this study, we experimentally and computationally study the local field enhancements

of nanoparticles on hydrothermally roughened SERS-active substrates, where the

effects of shape and size of Au particles and application of the fabricated samples in

(iii)

(iv)

Figure 1. Schematic representation for the fabrication of SERS-active substrate. First,

silicon wafers were cleaned by BOE and standard RCA cleaning procedures. Then, Ti

films were deposited on the pre-cleaned silicon wafers using reactive DC magnetron

sputtering system. The asdeposited samples were cleaved and treated under

hydrothermal conditions for various durations. Subsequently, Au was thermal

evaporated onto the hydrothermally roughened substrates for sensing. .

Figure. 2 (a) The AFM image of titanium thin films treated under hydrothermal

condition for 12 hours treatment duration. (b) The plot of simulated substrate which is

part of real substrate, where the matrix of nanoparticles is 3 x 5 due to periodical

property of the simulated structure.

p<100>). Then, 100-nm-thick titanium films are deposited on the pre-cleaned silicon

wafers using reactive DC magnetron sputtering system. The as- deposited sample is

cleaved into 0.5 cm x 1 cm squares and rinsed with ethanol, and de- ionized water.

Subsequently, the sample is put into a 23 mL Teflon-lined stainless steel autoclave

filled with 20 mL distilled water, which is sealed, and heated at 200oC for 2, 4, 6, 8, 10,

and 12 hours, respectively. Then the treated sample is cooled to room temperature

naturally, washed with distilled water for several times, and dried with a stream of

cylinder air. For example, the image of Fig. 2(a) shows the AFM images represent

titanium thin films treated under hydrothermal conditions for 12 hours treatment

duration.

The image of Fig. 2(b) shows the plane view of the gold-coated nanoparticular structure,

where the matrix of nanoparticles is 3 x 5 due to periodical property of the simulated

structure. Numerical simulation using a 3D FDTD method is conducted to investigate

the local field enhancement of substrate [13-

15]. The Maxwell’s curl equations in linear,

isotropic, nondispersive, lossy materials are

?B

K

KK?=?

?

×E, (1) ?EKKt?t=?J1KK

ε+με

?

×

B, (2)

?

?

BK

=0, (3)

F

igure 3. The simulation procedure of solving the Maxwell’s equations.

?

K

?

EK=ρ

ε

, (4)

where EK and BK

are the vectors of electric and magnetic fields, respectively,

?

and μ are permeability and

permittivity and JK and ρ are the current density vector and c

harge density. For a

globally defined curvilinear space, Maxwell’s equations are easily implemented in their

differential form, where Faraday’s law is Eq. (1) and Ampere’s law is Eq. (2).

The FDTD method solves Maxwell’s equations by first discretizing all

equations via

central differences in time and space. Then, based upon a 3D Yee’s mesh and

components of the electric and magnetic fields at points, the discretized spacing in the x,

y, and z directions adopted in our simulation are |x| = 0.01 um, |y| = 0.01 um and |z| =

0.01 um, where the time step Δt is 0.0004 and the time duration T is 3 in units of

femtoseconds. The discretized equations are iteratively solved in a leapfrog manner,

alternating between computing the E and

H fields at subsequent Δt/2 inte

rvals, as shown in Fig. 3. Notably, we employ the

perfectly matched layer as the simulation domain boundaries in which both electric and

magnetic conductivities are introduced in such a way that wave impedance remains

constant, absorbing the energy without

inducing reflections. III. RESULTS AND DISCUSSION

In order to have less light absorption, the larger scattering of substrate is better to

achieve larger field enhancement. For chemical sensing, the hydrothermally roughened

substrates are treated with aqueous solutions of 10-4 M R6G. The The chemical

structure of R6G is shown in Fig. 4(a). Fig. 4(b) shows that the characteristic Raman

vibrational modes of R6G immobilized on the substrate with or without hydrothermal

treatment. The substrate with hydrothermal treatment shows