Influence of high-refractive-index oxide cores on ... - OSA Publishing

2 downloads 152 Views 1MB Size Report
nanoshell. One of the benefits of using metal nanoparticles to adjust the optical properties (e.g., in coatings) is that even very small concentrations can cause ...
494

J. Opt. Soc. Am. B / Vol. 31, No. 3 / March 2014

Laaksonen et al.

Influence of high-refractive-index oxide cores on optical properties of metal nanoshells K. Laaksonen,1 S. Suomela,1,* S. R. Puisto,2 N. K. J. Rostedt,2 T. Ala-Nissila,1,3 and R. M. Nieminen1 1

COMP Centre of Excellence, Department of Applied Physics, Aalto University, P.O. Box 11100, FI-00076 Aalto, Finland 2 MatOx Oy, Erottajankatu 19 B, FIN-00130 Helsinki, Finland 3 Department of Physics, Brown University, P.O. Box 1843, Providence, Rhode Island 02912, USA *Corresponding author: [email protected] Received October 3, 2013; revised December 3, 2013; accepted December 29, 2013; posted January 7, 2014 (Doc. ID 198726); published February 13, 2014 We perform computationally a systematic study of the optical properties of nanoscale shell-core metal-oxide particles, where a dielectric core (vacuum, SiO2 , ZrO2 , or TiO2 ) is surrounded by a metal shell (Ag, Au, or Cu). We give a detailed discussion of the observed features in the optical spectra. The calculations are done using Mie theory and the four-flux method. The optical spectra are dominated by the localized surface plasmon resonance (SPR) excitations induced by the metal shell. We find that the symmetric dipole SPR modes can be redshifted up to 1500 nm by decreasing the shell thickness down to 1 nm with a high-refractive-index core. However, this shift comes with a severe loss in the sharpness of the SPR peak as both the decrease of the shell thickness and the high-refractive-index core dampen and broaden the peak. Thus, only shifts up to 500–1000 nm are practical if good selectivity and high extinction are required, as is the case for many near-infrared absorption applications. The choice of core material was found to cause shifts of a few hundred nanometers. © 2014 Optical Society of America OCIS codes: (160.4236) Nanomaterials; (160.4760) Optical properties. http://dx.doi.org/10.1364/JOSAB.31.000494

1. INTRODUCTION When metallic nanoparticles are irradiated with light, absorption of the light occurs at specific frequencies due to a phenomenon known as the localized surface plasmon resonance (SPR). The narrow absorption band caused by the SPR makes metallic nanoparticles promising candidates for applications that require high optical selectivity. For example, metal nanoparticles have found applications in various fields, such as biomedicine (e.g., biomarkers), functional surface coatings, and energy technology (e.g., photovoltaics) [1–4]. The wavelength and intensity of the SPR peak(s) depends on the various properties of the nanoparticles, such as the shape, size, and material. Furthermore, the refractive index of the medium and the concentration of the nanoparticles also influence the SPR. For most metal nanoparticles with a relatively high density of conduction electrons, the SPR is located in the ultraviolet region or at the short wavelengths of the visible spectrum. In order to achieve shifts to longer wavelengths in the visible or even in the near-infrared (NIR) or infrared (IR) regions, it has been shown that a good alternative is to use core-shell nanoparticles that consist of dielectric and metal layers [5]. Both the cases of a metal-dielectric and dielectric-metal core-shell nanoparticles have been under both extensive theoretical research, e.g., [6–11], and experimental research, e.g., Section 3 in [12] and the references therein. In particular, in a recently published paper (Ref. [13]) we studied systematically the effects of surrounding a spherical metal nanoparticle (Ag, Au, and Cu) with an oxide shell (SiO2 , ZrO2 , and TiO2 ). As one of the main results it was shown that redshifts of the SPR mode up to about 1000 nm into the IR and NIR 0740-3224/14/030494-09$15.00/0

regions can be obtained with high-index-of-refraction shell materials. These results raise interesting questions regarding the optical properties of core-shell nanoparticles with an opposite structure, i.e., a dielectric core coated with a metallic shell. In the previous studies of such core-shell particles, e.g., [5–9], the core material has been shown to affect the SPR of a nanoshell. One of the benefits of using metal nanoparticles to adjust the optical properties (e.g., in coatings) is that even very small concentrations can cause significant changes [14]. Such small concentrations also suggest that the nanoparticles do not significantly interact with each other so that computational methods suitable for the independent scattering region [15], such as Mie theory and the four-flux method, can be applied to study their optical properties. The SPR of a thin-metal nanoshell can differ greatly from the corresponding one of a solid metal nanoparticle. For shell thicknesses of few nanometers, the main SPR absorption peak can situate in wavelengths about a thousand nanometers longer than the one of solid spherical nanoparticles. Moreover, for every mode of the SPR there occurs two distinct peaks. The emergence of two different SPR peaks can be well described by the hybridization model [16]. According to the model, the two different SPR peaks are caused by the antisymmetric and symmetric SPR modes of the nanoshell that can be thought as a hybridization of the sphere and cavity modes, as illustrated in Fig. 1 and discussed further in Section 3. In this paper, we systematically study with Mie theory and the four-flux method the tunability of the SPR extinction peak positions without significantly decreasing the intensity of the peak by adding a dielectric core inside a metal nanoshell. The main aim of this work is to follow our previous work [13] and © 2014 Optical Society of America

Laaksonen et al.

Vol. 31, No. 3 / March 2014 / J. Opt. Soc. Am. B

εω  nω2 :

495

(1)

The dielectric function of the nanoparticle εω  ε0 ω  iε00 ω can be written as (see [13] for details) εω  εbulk ω 

ω2p ω2p − 2 ; ω  iΓ∞ ω ω  iΓrω 2

(2)

where ω is the angular frequency of the applied electromagnetic wave, εbulk is the bulk dielectric function, and ωp is the bulk plasma frequency [14,20]. Here, the damping constant Γ has been corrected [13] to take into account the influence of the large surface-to-volume ratio to phonon–electron interactions: Γr  Γ∞  A0

Fig. 1. Illustration of the hybridization model for dipole SPR. The SPR modes of the nanoshell can be thought as a hybridization of the sphere and cavity modes. The mode ω corresponds the opposite alignment of the cavity and sphere modes. In the case of ω− , the modes of the cavity and sphere are in parallel. Adapted from [16].

present a systematic and comprehensive study of the influence of a dielectric core to the optical properties of a metallic nanoshell, and to unravel the physical reasons for the observed features in the optical spectra. To this end, we consider a wide range of core radii and shell thicknesses within the nanoscale regime. Since in our case the optically active material is the metallic shell, we shall call these “shell-core” particles in what follows to emphasize the role of the metallic shell as the plasmonic material. The materials considered in this study are the same as in our previous work [13], i.e., the shell metals are silver (Ag), gold (Au), and copper (Cu). These metals were chosen because their SPR peaks are located in the visible wavelength range and they are commonly used in applications. The core materials include vacuum and the oxides SiO2 , ZrO2 and TiO2 , because they cover a large range of the real part values of the refractive index. SiO2 has the lowest refractive index (1.46 at wavelength 700 nm [17]) of the chosen oxide core materials, followed by ZrO2 (2.14 at wavelength 700 nm [18]), and finally TiO2 (2.64 at wavelength 700 nm [17]). For all the materials, we use empirical refractive index data that is wavelength dependent. The outline of this paper is as follows: we start with brief introductions to the optical properties of nanoparticles and the computational methods used in this work, followed by a presentation of the material parameters. Then, the Mie results are presented and analyzed, followed by comparison to the results of our previous work [13] and a short summary of the four-flux results. The four-flux results are presented in more detail in [19]. Finally, conclusions are drawn.

2. METHODS A. Optical Properties of Nanoparticles The dielectric function εω and the complex refractive index nω both describe the optical properties of the system and they are related by

vF ; t

(3)

where vF is the Fermi velocity, t is the thickness of the metal shell, Γ∞ is the damping constant of the bulk metal, and A0 is a constant dependent on the electron-surface interaction and is usually of the order of unity. B. Computational Methods In this paper we use Mie theory based models to investigate the optical properties of the metal-oxide shell-core nanoparticles, i.e., oxide filled metal nanoshells. We have previously used the same methods for metal-oxide core-shell nanoparticles, i.e., metal particles with an oxide coating. As the methodology is exactly the same, we guide the reader to the previous publication [13] for the details of the methods and introduce them below only briefly. Mie theory [21] gives a classical solution to the scattering and absorption of electromagnetic radiation by a sphere embedded in a linear, isotropic, homogeneous medium. Mie theory has been described in detail in the literature [22]. For a coated sphere (core-shell structure) [23] both the particle-coating surface and the coating-medium surface must be taken into account. In this paper, Mie theory was used to study the optical properties of single shell-core nanoparticles. However, as in our previous paper, we also considered briefly how the optical properties of the nanoparticles would be affected by embedding them in a coating layer. The four-flux method [24–26] is a simplification of the multiple-scattering approach and wellsuited for this task. The four-flux results are presented in [19]. C. Material Parameters The calculations were carried out with three different shell materials Ag, Au, and Cu and with dielectric cores of TiO2 , ZrO2 stabilized with Y2 O3 , SiO2 , and vacuum. The bulk complex refractive indices of these metals were obtained from Johnson and Christy [27] and modified according to Eq. (2) with A0  1. In Table 1, the values used for the parameters needed in Eq. (2) are presented. The complex refractive indices of TiO2 and SiO2 were obtained from Palik [17]. The real refractive index of ZrO2 stabilized with 12.0 mol Y2 O3 was formed from the data of Wood and Nassau [18]. We assumed the imaginary part of the refractive index of ZrO2 to be negligible. These oxides were chosen as they span a large range of refractive index values and have small absorption coefficients

J. Opt. Soc. Am. B / Vol. 31, No. 3 / March 2014

Laaksonen et al.

Table 1. Values Used for the Parameters of the Shell Materials Needed in the Mie Calculations Γ∞ (1013 1∕s) vF (106 m∕s) ωp (eV)

Ag

Au

Cu

3.2 [27] 1.39 [28] 9.15 [27]

11 [27] 1.40 [28] 9.08 [27]

14 [27] 1.57 [28] 8.85 [27]

are more redshifted and thus the symmetric SPR peaks of the shell-cores with Au and Cu cores do not situate anymore in the interband region. Figure 2 presents the Mie calculation results with a core radius R  30 nm and shell thicknesses t  2, 5, and 10 nm for the optical spectra of the nanoparticles with a core material of TiO2 , ZrO2 , SiO2 , and vacuum, and a metal shell of Ag, Au, and Cu. In Fig. 2, the optical spectra is shown only up to the wavelength 1510 nm as for these examples all the relevant optical behavior happens at shorter wavelengths. The extinction (Ext) and absorption (Abs) intensities are expressed as the efficiency Q  σ∕πR  t2 , where R is the core radius, t is the shell thickness, and σ is the corresponding Mie cross section, e.g., the extinction cross section σ Ext in the case of QExt . The absorption efficiency QAbs is obtained from the relation QAbs  QExt − QSca , where QSca is the scattering efficiency. Another often used convention for expressing the intensity is σ ext ∕4∕3πR  t3 . In that case, the intensity would be higher for small particles and lower for larger particles than in the case of using QExt . Thus, if coatings are compared using a constant volume fraction of nanoparticles then the maximum extinction is predicted to be achieved at the core radius and shell thickness where σ ext ∕4∕3πR  t3   3QExt ∕4R  t is highest. The SPR wavelengths of the results

in the relevant wavelength range. For further details we refer to the earlier publication [13].

3. RESULTS AND DISCUSSION The calculations were carried out for different core radii between 5 and 100 nm and shell thicknesses between 1 and 100 nm for wavelengths between 300 and 1970 nm. When the shell thickness decreases to less than few nanometers, the quantum confinement effect starts to become important, which can affect the accuracy of the classical Mie solutions. Two values for the refractive index of the surrounding medium were used, namely 1.00 and 1.50. The results obtained with the refractive index of 1.00 are presented here and the ones with 1.50 are presented in [19]. The most important difference is that in the case of 1.50, the symmetric SPR peaks

Q

(t = 2 nm)

Ext

12

Q

(t = 5 nm)

Q

Ext

(a) Ag−TiO2, R = 30 nm

(b) Ag−ZrO2, R = 30 nm

Efficiency Q

10

Efficiency Q

(t = 10 nm)

8 6 4 2 0

Efficiency Q

5

Q

Ext

(t = 2 nm)

Abs

12

Q

(c) Ag−SiO2, R = 30 nm

10 8 6 4 2 0

(t = 5 nm)

Abs

Efficiency Q

496

12

Q

(t = 10 nm)

Abs

(d) Ag−Vacuum, R = 30 nm

10 8 6 4 2

(e) Au−TiO2, R = 30 nm

(f) Au−ZrO2, R = 30 nm

(g) Au−SiO2, R = 30 nm

(i) Cu−TiO2, R = 30 nm

(j) , R 1000 = 301200 nm 1400 (k) Cu−SiO2, R = 30 nm 400Cu−ZrO 600 800 2 Wavelength (nm)

0

(h) Au−Vacuum, R = 30 nm

4 3 2 1 0

Efficiency Q

4

(l) Cu−Vacuum, R = 30 nm

3

2

1

0

400

600

800 1000 1200 1400

Wavelength (nm)

400

600

800 1000 1200 1400

Wavelength (nm)

400

600

800 1000 1200 1400

Wavelength (nm)

400

600

800 1000 1200 1400

Wavelength (nm)

Fig. 2. Selected examples of the Mie results for the absorption (QAbs ) and extinction (QExt ) efficiencies of shell-core nanoparticles with different shell and core materials. Here, R denotes the core radius and t the shell thickness. The dots denote the symmetric dipole SPR wavelength in the electrostatic limit [Eq. (5)].

Laaksonen et al.

Vol. 31, No. 3 / March 2014 / J. Opt. Soc. Am. B

(according to the data of [27]) [10]. The interband contribution suppresses the intensity of SPR if it occurs at shorter wavelengths than the threshold wavelength. This explains the lower intensities for the shell thickness t  10 nm than for t  5 nm in the Figs. 2(h), 2(k), and 2(l). The sharp peak at 340 nm in Fig. 2(a) for shell thickness 10 nm is mainly caused by the antisymmetric SPR [11]. The SPR modes of shell-core nanoparticles can be understood with the hybridization model [16,29]. According to the model, the SPR modes of a hollow shell-core particle can be thought as a hybridization of the sphere and cavity plasmons (Fig. 1). For a Drude metal with a vacuum core embedded in vacuum, the SPR frequencies in the electrostatic limit are of the form [16,29]:

are in the accordance with the previous studies of nanoshells with a vacuum core [10]. However, there is a difference in the height of the sharp SPR peaks. In Fig. 2, the smooth peaks at wavelengths longer than 400 nm are caused by the symmetric SPRs, mainly by the symmetric dipole SPR. For thin metal shells, the dipole SPR is much more pronounced than the higher SPR modes. Thus, the higher symmetric SPR modes contribute relatively little to the smooth peaks at wavelengths longer than 400 nm in Fig. 2. When the shell thickness is increased to tens of nanometers, the higher SPR modes become gradually more prominent. The interband transitions cause significant absorption at wavelengths shorter than the interband threshold, which is ≈320.4 nm for Ag, ≈496.0 nm for Au, and ≈582.1 nm for Cu R = 5 nm

R = 10 nm

R = 20 nm 2000

R = 40 nm

(b) Ag−ZrO

2000

R = 60 nm

R = 80 nm

2000 1800

1500

1500

1500

1500

1000

1000

1000

1000

2

Wavelength (nm)

R = 30 nm

2000

(a) Ag−TiO

1600

497

2

2000

(c) Ag−SiO

2

R = 100 nm

(d) Ag−Vacuum

1400 500

1200

500 1

2

3

4

5

500 1

2

3

4

5

500 1

2

3

4

5

1

2

3

4

5

1000 800 600 400

(f) Ag−ZrO2

(e) Ag−TiO2

12

(g) Ag−SiO2

(h) Ag−Vacuum

10

Ext

QMax

8 6 4

Wavelength (nm)

2 0 2000

2000

1800

1500

1500

1500

1500

1000

1000

1000

1000

1600

2000

(i) Au−TiO2

2000

(j) Au−ZrO2

2000

(k) Au−SiO2

(l) Au−Vacuum

1400 500 1200

500 1

2

3

4

5

500 1

2

3

4

5

500 1

2

3

4

5

1

2

3

4

5

1000 800 600 400 8

(n) Au−ZrO2

(m) Au−TiO2

(o) Au−SiO2

(p) Au−Vacuum

7

QExt

Max

6 5 4 3 2 1 0

0

20

40

60

80

Shell thickness (nm) Fig. 3.

100 0

20

40

60

80

Shell thickness (nm)

100 0

20

40

60

80

Shell thickness (nm)

100 0

20

40

60

80

100

Shell thickness (nm)

Positions and intensities (QMax Ext ) of the dipole SPR peaks (symmetric mode) for the shell-core particles with an Au and Ag shell.

498

J. Opt. Soc. Am. B / Vol. 31, No. 3 / March 2014

ω2l

Laaksonen et al.

2 3 s   ω2p 4 1 R 2l1 5 1  4l  1  1 ; 2l  1 tR 2

where f s  1 − R3 ∕R  t3 and εc , εs , εm denote the dielectric functions of the core, the shell, and the medium, respectively. In Eq. (5), the term with  is for the symmetric dipole SPR and the one with − for the antisymmetric dipole SPR. In the elecrostatic limit only the shell/core ratio, not the actual size of the nanoparticle, affects the SPR wavelength. In Fig. 2, the small dots are the symmetric dipole SPR wavelengths given by Eq. (5). From Fig. 2, we can see that the SPR wavelength of the elecrostatic limit differs tens or even hundreds of nanometers from the Mie solutions. This shows that the size-dependent effects, i.e., retardation effects, are already significant with the core radius 30 nm. The effects of core radius and shell thickness on the intensity and position of the symmetric dipole SPR peak are presented in more detail in Figs. 3 and 4. It should be noted that the values in Figs. 3 and 4 are not for pure symmetric dipole SPR peaks. Overlap with other contributions, such as interband absorption, scattering from the core, and the antisymmetric and other symmetric modes have some influence on the values, especially near the interband threshold. As can be seen from Figs. 2–4, the highest intensities for the symmetric extinction peak were achieved when the core was coated with an Ag shell and the lowest ones when coated with Cu, as expected from the earlier results of corresponding metal-dielectric core-shell particles [13]. In the case of a 1 nm shell thickness, the SPR peak is very broad and the maximum QExt is relatively low. When the shell thickness increases from 1 to 20 nm, the SPR peak blueshifts very rapidly and QMax Ext increases due to the increasing absorption and scattering of the metal shell. Moreover, the antisymmetric and symmetric modes start to overlap, which also causes an increase in the maximum intensity and explains the rapid growth of QExt in the case of Ag-TiO2 with a core radius of 5 nm [Fig. 3(e)]. As an

(4)

where l is the mode of the SPR, i.e., l  1 is the dipole mode, l  2 the quadrupole mode, etc., and ωl and ωl− correspond to the antisymmetric and the symmetric coupling of the sphere and cavity, respectively. The dipole moments of the sphere plasmon and cavity plasmon are in opposite alignment in the antisymmetric mode and in parallel alignment in the symmetric mode. Thus, the symmetric mode has a larger net dipole moment and a stronger coupling with light. As a consequence, the SPR extinction caused by the symmetric mode is usually more pronounced than that of the antisymmetric mode. This is seen in Fig. 2, where the symmetric mode can be more clearly distinguished than the antisymmetric for all shell-core combinations. Moreover, in the present cases the antisymmetric mode occurs near the ultraviolet wavelengths, where the extinction is mainly caused by the interband transitions of the metal shell and scattering from the core. Therefore, in this work we will focus on analyzing the behavior of the symmetric mode. Equation (4) gives only the SPR frequency in the electrostatic limit for a Drude metal shell. More generally, the dipole SPR wavelength in the electrostatic limit is given by the resonance condition [30]: Refεs g  −Refεm p1  p21 − εc ∕εm 1∕2 g;

p1 

  εc 3 1 3 1  − − ; 2f s 2 εm 4f s 2

R = 5 nm

R = 10 nm

(6)

R = 20 nm 2000

R = 30 nm

R = 40 nm

R = 80 nm

2000

1800

1500

1500

1500

1500

1000

1000

1000

1000

(a) Cu−TiO

1600

2000

R = 60 nm

2000

2

Wavelength (nm)

(5)

(b) Cu−ZrO

2

2000

(c) Cu−SiO

2

R = 100 nm

(d) Cu−Vacuum

1400 500 1200

500 1

2

3

4

5

500 1

2

3

4

5

500 1

2

3

4

5

1

2

3

4

5

1000 800 600 400 7

(f) Cu−ZrO2

(e) Cu−TiO2

(g) Cu−SiO2

(h) Cu−Vacuum

6

Ext

QMax

5 4 3 2 1 0

0

20

40

60

80

Shell thickness (nm)

100 0

20

40

60

80

Shell thickness (nm)

100 0

20

40

60

80

Shell thickness (nm)

100 0

20

40

60

80

100

Shell thickness (nm)

Fig. 4. Positions and intensities (QMax Ext ) of the dipole SPR peaks (symmetric mode) for the shell-core particles with a Cu shell.

Laaksonen et al. 8

Vol. 31, No. 3 / March 2014 / J. Opt. Soc. Am. B

Ag−TiO , t+R = 80 nm 2

7

(R = 10 nm)

Q

(R = 0 nm)

Ext Ext

6

Efficiency Q

is negligible as the optical response primarily happens near the surface of the particle. As was expected from the previous studies (e.g., [6]), for all the shell materials, a high-refractive-index oxide core causes the symmetric SPR to occur at longer wavelengths than the symmetric SPR of the corresponding shell-core particle with a core of a lower refractive index. The thinner the shell, the stronger the influence of the core material. For example, the symmetric SPR peak occurs at 852.1 nm for an Au-TiO2 shellcore particle with a core radius of 30 nm and shell thickness of 5 nm. For the corresponding Au-SiO2 and Au-vacuum particles, the SPR occurs at 670.1 and 625.1 nm, respectively. Thus, the position of the symmetric SPR peak can be adjusted by changing the core material: the larger the refractive index of the core, the more the SPR peak shifts. However, the refractive index of the core also influences the maximum intensity of the SPR peak. The SPR peak quickly broadens and the maximum QExt decreases as the refractive index of the core increases, as seen clearly in Figs. 2(a)–2(d). Furthermore, a high-refractive-index core increases scattering from the core. When the core radius is increased, the scattering contribution, and the possible absorption contribution depending on the core material, strengthens and redshifts, as seen in Fig. 6. Thus, for example, for NIR absorption coatings, the refractive index of the core should be small enough to prevent the scattering contribution of the core to be redshifted to the region of visible light. As discussed in [11], the refractive index of the core also influences the magnitude of the antisymmetric modes. For shell-core particles with a SiO2 or vacuum core, the antisymmetric peak cannot be distinguished in the optical spectra. However, for TiO2 and ZrO2 , there is a sharp peak at ≈350 nm in the case of Ag shells. The peak, which is mainly caused by the antisymmetric SPR, can be most clearly seen when the core radius and shell thickness are of the same order. The effect of changing the R∕t ratio is illustrated in Fig. 7. In Fig. 7, the peaks at wavelengths shorter than 400 nm are mainly caused by the antisymmetric modes. When the shell thickness is increased or the core radius is decreased, the antisymmetric mode peak redshifts as predicted by the hybridization model.

QExt(R = 20 nm) Q

QAbs(R = 20 nm)

5

QAbs(R = 10 nm)

4

QAbs(R = 0 nm)

3 2 1 0 300

400

500

600

700

Wavelength (nm) Fig. 5. Extinction efficiency (QExt ) and the absorption efficiency (QAbs ) for Ag-TiO2 shell-core nanoparticles with a total radius (t  R) of 80 nm and a core radii of 0,10, and 20 nm.

example of the shift magnitude, for an Au-TiO2 particle with a core radius of 20 nm, the SPR peak blueshifts from 1293.9 to 628.6 nm as the shell thickness increases from 1 to 10 nm [Fig. 3(i)]. The same behavior of the blueshift is predicted by the hybridization model [Eq. (4)]. As the shell thickness is further increased, the behavior starts to resemble the one corresponding to a noncoated metal sphere, i.e., the SPR peak redshifts and broadens. As shown in Figs. 3 and 4, when the shell thickness is increased from 60 nm up, the SPR peak redshifts almost linearly and resembles the radius dependence of a homogeneous metal sphere. This is a consequence of the fact that when R∕t → 0, the Mie coefficients of a shell-core particle reduces to those of a homogeneous sphere of the shell material with radius t [22]. In fact, the whole spectrum depends almost solely on the total radius RTot  R  t when the shell thickness is relatively large (R∕t ≲ 0.5), as illustrated in Fig. 5 and [11]. That is, for large shell thicknesses, the effect of the core

6

(a) Ag−TiO2, R = 50 nm

Q

(t = 0 nm)

Q

(t = 5 nm)

Ext Ext

Q

Sca

4

1000 1200 1400 1600 1800

Sca

(t = 5 nm)

3

1

Wavelength (nm)

QSca (t = 0 nm) Q

1

800

(t = 5 nm)

4

2

600

Q 5

2

400

(t = 0 nm)

Ext

(t = 5 nm)

3

0

(b) Ag−TiO2, R = 100 nm Q

Ext

Efficiency Q

Efficiency Q

6

QSca (t = 0 nm)

5

499

0

400

600

800

1000 1200 1400 1600 1800

Wavelength (nm)

Fig. 6. Extinction efficiency (QExt ) and the scattering efficiency (QSca ) for Ag-TiO2 shell-core nanoparticles with a core radius of (a) 50 nm and (b) 100 nm, and a shell thickness of 0 and 5 nm.

500

J. Opt. Soc. Am. B / Vol. 31, No. 3 / March 2014 8

Ag−TiO2, R = 20 nm

7

Laaksonen et al.

In Fig. 8(b), the maximum intensity of the SPR peak is given in the maximum extinction cross section of the peak σ Max Ext per particle volume V  4∕3πR3Tot , i.e., σ Max Ext is scaled by the amount of core and shell material. Thus, the maximum of a curve with constant t∕RTot gives the total radius by which the extinctance of a coating is estimated to be highest if the volume or mass fraction of particles is kept constant. For t∕RTot  0.1 and 0.2 the maximum of σ Max Ext ∕V is found with shell thicknesses 4–12 nm. For very thin shells, the damping constant increases rapidly due to the surface-to-volume correction term A0 vF ∕t in Eq. (3). This causes the intensity of the SPR peak to drop significantly. The choice of A0 affects the exact shell thickness by which the maximum of σ Max Ext ∕V is achieved. In our simulations A0 was assumed to be unity. In [13], we studied metal nanospheres coated with an oxide shell and considered their suitability for applications where NIR absorption is required. The same materials were used as in this work. It was shown that with a high-refractive-index oxide coating and increasing the size of the metal sphere, the dipole SPR can be redshifted even close to 1000 nanometers. However, for metal nanospheres, the other SPR modes redshift at a slower rate. As a consequence, when a large redshift of the dipole SPR peak is obtained, there occur multiple SPR peaks in the optical spectrum at shorter wavelengths than the dipole SPR wavelength. Moreover, the large redshifts of the dipole SPR peak come with a severe dampening and broadening of the peak. Thus, the investigated metal nanospheres (Ag, Au, and Cu) in [13] are not best suited for NIR absorption applications. With metal nanoshells, large redshifts of the symmetric SPR peak can be achieved without splitting up to multiple peaks corresponding to the different modes, such as the dipole SPR and the quadrupole SPR. Furthermore, the antisymmetric SPR situates in the ultraviolet region for thin shells. Therefore, the metal-dielectric shell-core particles are a suitable candidate for NIR absorption applications as they absorb light in the NIR region due to the symmetric SPR without blocking the visible light. Due to the properties mentioned in the paragraphs above, according to our data the metaldielectric shell-core particles with small enough shell

QExt(t = 5 nm) Q

(t = 15 nm)

Q

(t = 30 nm)

Q

(t = 5 nm)

5

Q

(t = 15 nm)

4

QAbs(t = 30 nm)

Ext Ext

6

Efficiency Q

Abs Abs

3 2 1 0 300

400

500

600

700

Wavelength (nm) Fig. 7. Extinction efficiency (QExt ) and the absorption efficiency (QAbs ) for Ag-TiO2 shell-core nanoparticles with a core radius of 20 nm and a shell thickness of 5,15, and 30 nm.

From Figs. 3 and 4, it can be seen that the position of the symmetric SPR extinction peak strongly depends on the ratio of the shell thickness to the core radius for small shell thicknesses rather than just the shell thickness itself. In the electrostatic limit, the wavelength of the SPR depends only on the ratio of the shell thickness to the core radius [Eq. (5)]. However, outside the electrostatic limit, the SPR peak redshifts if the total radius RTot is increased and the ratio of t∕RTot is kept constant. Figure 8 presents examples of the redshift. The SPR wavelength was determined for all the shell-core combinations with ratios t∕RTot  0.1; 0.2; 0.3; 0.4, and 0.5, while the total radius was changed from 10 to 100 nm (results presented in [19]). It was found out that the magnitude of the redshift increases when the refractive index of the core is increased or the ratio t∕RTot is decreased. For t∕RTot  0.1 and TiO2 core, the SPR peak shifted from 887.1 to 1028.3 nm in the case of Ag shell [Fig. 8(a)], from 940.0 to 1092.2 nm in the case of Au shell, and from 932.2 to 1086.7 nm in the case of Cu shell. 0.1

(a)

1000

0.09

950

0.08

900

0.07 Max −1 σExt /V (nm)

Wavelength (nm)

1050

850 800 750

0.02

600

0.01 60

Total radius (nm)

80

100

Au−SiO , 2 t/RTot = 0.2

0.04

650

40

Au−SiO2, t/RTot = 0.1

0.05

0.03

20

Ag−TiO , 2 t/RTot = 0.2

0.06

700

550

Ag−TiO , 2 t/RTot = 0.1

(b)

0

Cu−Vacuum, t/R = 0.1 Tot

Cu−Vacuum, t/RTot = 0.2 20

40

60

80

100

Total radius (nm)

Fig. 8. Examples of (a) the wavelength and (b) the intensity of the symmetric dipole SPR peak as a function of the total radius (RTot  R  t) when 3 the ratio t∕RTot is kept constant. Here, σ Max Ext denotes the maximum extinction cross section of the peak and V  4∕3πRTot denotes the particle volume.

Laaksonen et al.

thicknesses to induce the symmetric SPR peak at longer wavelengths than 750 nm, are the most suitable candidates for NIR absorption applications. Finally, we also used the four-flux method in the same way as in our previous work [13] to study the optical properties of the shell-core nanoparticles embedded in a dielectric coating layer. In the four-flux method, the multiple scattering is taken into account, e.g., the light scattered by one particle can be absorbed by another particle. With the four-flux method the reflectance versus absorptance ratio can be estimated better than directly from the scattering and absorption coefficients obtained from Mie theory. The results of our metal-oxide shellcore nanoparticles (results shown in [19]) demonstrate, as in the previous case of metal-oxide core-shell particles, that in a coating the absorption contribution of the extinction becomes emphasized and only a very small concentration of nanoparticles is required for an observable change in the optical behavior.

4. SUMMARY AND CONCLUSIONS In the present work we have systematically studied the influence of an oxide core on the SPR of metal nanoshells by using Mie-based computational methods for all possible combinations of the metals Ag, Au and Cu, and oxides SiO2 , ZrO2 , TiO2 , and the vacuum. Our work demonstrates quantitatively the dramatic influence of the thickness of the metal shell. For all the metal-oxide shell-core combinations, it was found that the symmetric SPR absorption peak can be redshifted to IR wavelengths by decreasing the thickness of the shell to less than 10 nm. By changing the shell thickness, we observed even shifts over 1500 nm. However, for the thinnest shells of just a few nanometers in thickness, the symmetric SPR peak becomes very broad and highly dampened, limiting the practical applicability of the thinnest shells. Thus, practically obtainable shifts are of the order of 500–1000 nm. The core material was found to have a smaller influence, only few hundred nanometers in general. The higher the refractive index of the core, the more redshifted is the symmetric SPR peak. For example, for an Ag nanoshell with a core radius of 30 nm and a shell thickness of 5 nm, there was a difference of 251 nm between the SPR peaks of Agvacuum and Ag-TiO2 particles. Even larger shifts can be achieved with thinner shells but they are accompanied by a severe dampening of the symmetric SPR peak. Moreover, the applicability of very high refractive-index cores is decreased by the scattering from the core which becomes prominent especially in the case of large cores. Since many of the common oxide-core materials have refractive indices in the mid-range, the choice of the core material has only a relatively minor effect on the peak positions. Consequently, in experiments the core material can be chosen among such materials relatively freely without significantly influencing the optical properties of the shell-core nanoparticles. Compared to our previous results [13] of metal nanospheres coated with an oxide shell for the same materials, the metal nanoshells with an oxide core were found to be better suited for NIR absorption applications, as the symmetric SPR can cause absorption of light in the NIR region without blocking the visible light.

Vol. 31, No. 3 / March 2014 / J. Opt. Soc. Am. B

501

ACKNOWLEDGMENTS Authors from Aalto University acknowledge support by the Academy of Finland through its COMP Center of Excellence grant (project no. 251748). MatOx Oy acknowledges support by the Finnish Funding Agency for Technology and Innovation (TEKES) and by Chris Lowe, Bengt Ingman, and James Maxted of Becker Industrial Coatings Ltd. We are also grateful to G. A. Niklasson from Uppsala University and W. E. Vargas from University of Costa Rica for allowing us to use their implementation of the four-flux method.

REFERENCES 1. M. Garcia, “Surface plasmons in metallic nanoparticles: fundamentals and applications,” J. Phys. D 44, 283001 (2011). 2. D. A. Giljohann, D. S. Seferos, W. L. Daniel, M. D. Massich, P. C. Patel, and C. A. Mirkin, “Gold nanoparticles for biology and medicine,” Angew. Chem., Int. Ed. 49, 3280–3294 (2010). 3. J. R. Lakowicz, Principles of Fluorescence Spectroscopy (Springer, 2006). 4. R. G. Chaudhuri and S. Paria, “Core/shell nanoparticles: classes, properties, synthesis mechanisms, characterization, and applications,” Chem. Rev. 112, 2373–2433 (2012). 5. N. J. Halas, S. Lal, W.-S. Chang, S. Link, and P. Nordlander, “Plasmons in strongly coupled metallic nanostructures,” Chem. Rev. 111, 3913–3961 (2011). 6. E. Prodan, A. Lee, and P. Nordlander, “The effect of a dielectric core and embedding medium on the polarizability of metallic nanoshells,” Chem. Phys. Lett. 360, 325–332 (2002). 7. D. Wu, X. Xu, and X. Liu, “Influence of dielectric core, embedding medium and size on the optical properties of gold nanoshells,” Solid State Commun. 146, 7–11 (2008). 8. R. Bardhan, N. K. Grady, T. Ali, and N. J. Halas, “Metallic nanoshells with semiconductor cores: optical characteristics modified by core medium properties,” ACS Nano 4, 6169–6179 (2010). 9. J. Butet, I. Russier-Antoine, C. Jonin, N. Lascoux, E. Benichou, and P.-F. Brevet, “Effect of the dielectric core and embedding medium on the second harmonic generation from plasmonic nanoshells: Tunability and sensing,” J. Phys. Chem. C 117, 1172–1177 (2013). 10. O. Peña, U. Pal, L. Rodríguez-Fernández, and A. Crespo-Sosa, “Linear optical response of metallic nanoshells in different dielectric media,” J. Opt. Soc. Am. B 25, 1371–1379 (2008). 11. Q. Li and Z. Zhang, “Broadband tunable and double dipole surface plasmon resonance by tio 2 core/ag shell nanoparticles,” Plasmonics 6, 779–784 (2011). 12. M. R. Jones, K. D. Osberg, R. J. Macfarlane, M. R. Langille, and C. A. Mirkin, “Templated techniques for the synthesis and assembly of plasmonic nanostructures,” Chem. Rev. 111, 3736–3827 (2011). 13. K. Laaksonen, S. Suomela, S. R. Puisto, N. K. Rostedt, T. Ala-Nissila, and R. M. Nieminen, “Influence of high-refractiveindex oxide coating on optical properties of metal nanoparticles,” J. Opt. Soc. Am. B 30, 338–348 (2013). 14. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters (Springer, 1995). 15. G. Niklasson, “Modeling the optical properties of nanoparticles,” SPIE Newsroom, doi: 10.1117/2.1200603.0182 (2006). 16. E. Prodan, C. Radloff, N. Halas, and P. Nordlander, “A hybridization model for the plasmon response of complex nanostructures,” Science 302, 419–422 (2003). 17. E. Palik and G. Ghosh, Handbook of Optical Constants of Solids (Academic, 1998). 18. D. Wood and K. Nassau, “Refractive index of cubic zirconia stabilized with yttria,” Appl. Opt. 21, 2978–2981 (1982). 19. S. Suomela, “Nanoparticle plasmonics,” http://physics.aalto.fi/ groups/comp/msp/research/plasmonics/. 20. J. Zhang, Optical Properties and Spectroscopy of Nanomaterials (World Scientific, 2009). 21. G. Mie, “Beiträge zur optik trüber medien, speziell kolloidaler metallösungen,” Ann. Phys. 330, 377–445 (1908).

502

J. Opt. Soc. Am. B / Vol. 31, No. 3 / March 2014

22. D. Huffman and C. Bohren, Absorption and Scattering of Light by Small Particles (Wiley Interscience, 1983). 23. A. Aden and M. Kerker, “Scattering of electromagnetic waves from two concentric spheres,” J. Appl. Phys. 22, 1242–1246 (1951). 24. B. Maheu, J. Letoulouzan, and G. Gouesbet, “Four-flux models to solve the scattering transfer equation in terms of Lorenz-Mie parameters,” Appl. Opt. 23, 3353–3362 (1984). 25. B. Maheu and G. Gouesbet, “Four-flux models to solve the scattering transfer equation: special cases,” Appl. Opt. 25, 1122–1128 (1986).

Laaksonen et al. 26. W. Vargas and G. Niklasson, “Pigment mass density and refractive index determination from optical measurements,” J. Phys. 9, 1661–1670 (1997). 27. P. Johnson and R. Christy, “Optical constants of the noble metals,” Phys. Rev. B 6, 4370–4379 (1972). 28. N. Ashcroft and N. Mermin, Solid State Physics, Thomson Learning (Saunders College, 1976). 29. E. Prodan and P. Nordlander, “Plasmon hybridization in spherical nanoparticles,” J. Chem. Phys. 120, 5444–5454 (2004). 30. N. G. Khlebtsov, “T-matrix method in plasmonics: an overview,” J. Quant. Spectrosc. Radiat. Transfer 123, 184–217 (2013).