**2. Numerical investigation of plasmonic nanostructures**

Since the discovery of the plasmonic effect, many text books have been produced [2, 13–16] and the theory of this interesting effect can be found easily. We present here only a numerical method to model the analytical theories, especially for some particular plasmonic structures. Actually, there exist different simulation methods which allow to create models of an arbitrary plasmonic structure and to obtain its plasmo-optical properties. Each method presents its own advantages and disadvantages, such as required time processing, computer memory, and simulation precision. We have adopted a very well-known simulation method, which allowed us to achieve rapidly and precisely the plasmo-optical properties of any metallic NSs.

### **2.1. Simulation methodology**

on [2–4]. The physical phenomenon of SPR is arisen by light-matter interaction at the interface of metallic and dielectric materials, especially in noble metals, for example, silver and gold [2–4]. The physical mechanism of the SPR can be briefly explained as following. Under irradiation of an electromagnetic (EM) wave, the free electrons in metallic materials are driven to oscillate at the frequency of the external EM field. This electron oscillation at the metallic surface causes a charge separation with respect to the ionic lattice, forming a dipole oscillation along the direction of the electric field of the light. The amplitude of the oscillation reaches maximum at a specific frequency, called SPR frequency, when the incident EM frequency matches the eigenfrequency that relates to the restoring force stemming from the lattice of positive nuclei. For a metallic object with finite dimension, only the electrons on its surface are the most significant since the EM wave can only penetrate a limited depth in metal. Therefore,

The SPR band intensity and spectra depend on several factors affecting the electron charge density on the metallic surface, such as the metal type, the dielectric constant of the surrounding medium, particle size, shape, structure, and composition [5]. The plasmonic structures are therefore distinguished into three categories: (i) surface plasmon polariton; (ii) localized

Surface plasmon polaritons (SPPs) are EM excitations propagating along the interface between a dielectric and a metal. These EM surface waves arise via the coupling of the EM field to oscillations of the conductor's electron plasma. As a SPP loses energy to the metal due to absorption and scattering, it can only propagate for a finite distance along the interface. Likewise, perpendicularly to the interface, electric field falls off evanescently and can only penetrate into the metal a certain tiny "skin depth" [2], while such an evanescent electric field extends more in the dielectric. Therefore, SPPs are very sensitive to any perturbation within the skin depth and are often used as a very sensitive chemical, biological, or gas sensor [6].

In contrast to SPP, localized surface plasmon resonance (LSPR) occurs when collective oscillations of free electrons are confined to a finite volume, such as metal nanoparticles (NPs), as EM standing waves. The LSPR is excited when the frequency of the incident photons matches the resonance frequency of the NPs. Generally, the LSPR in visible range is obtained with noble NPs with dimensions below 100 nm. The plasmonic properties of metallic NPs vary with their shape and size and are also affected by the refractive index of the surrounding medium. This results in various applications of metallic NPs in different domains, from fun-

Clearly, the LSPR can strongly and locally amplify the EM field near the metallic surface. It is recently demonstrated that this effect becomes much stronger when two or multiple metallic nano-objects are arranged very close to form the so-called plasmonic nanostructure (PNS) [8]. Two NPs, for example, under an EM illumination can be described as two point dipoles. When (i) NPs are closely spaced (separation ≪ light wavelength) and (ii) light polarization is appropriated (longitudinally), these two point dipoles can interact via their near-fields: the restoring force acting on the oscillation electrons of each NP is increased by the charge distribution of their neighboring NP. Therefore, EM field in between this "dimer" can be locally and intensely enhanced, resulting in a hotspot [9]. Following this idea, the plasmonic

the collective oscillations of such electrons are called SPR [2].

66 Plasmonics

surface plasmon resonance; and (iii) plasmonic nano-structures.

damental science to practical applications [7].

Finite-difference time-domain (FDTD) is a numerical analysis method allowing to obtain approximate solutions of electrodynamics problems. FDTD workground was first set in 1928 [17]. However, it only became popular since the1980s thanks to the revolution of information technology [18]. Recently, with the assistance of modern computers, FDTD has become a powerful technique for researchers to predict the EM response of any structure, in particular PNSs. As other finite-difference modeling methods, FDTD is grid-based. It means that computational domains will be meshed into minor units, mostly in a cube shape, with associated vector components of the electric (**E**) field and magnetic (**H**) field, which are determined by Maxwell's equations. In simulation, the **E** and **H** fields are calculated at every point in space, forward in time. Additionally, FDTD is a time-domain method, so it can cover a wide range of frequencies with a single simulation run [19]. Thanks to these advantages, FDTD provides us a natural way to treat any electrodynamics problem, especially periodic structures and broadband sources in PNSs complications.

**2.2. Localized surface plasmon resonance**

y-, and z-axis. The mesh size of 1 × 1 × 1 nm3

blue shift as the aspect ratio R increases.

Inset: design of Au nano-prism.

is a 1 × 1 × 1 μm3

Au-medium interface.

Au NPs possessing particular shapes, such as nano-sphere, nano-rod, and nano-prism, are often fabricated by standard chemical methods, for example, the Turkevich [22], Brust [23] seeded growth [24] or other miscellaneous methods [25]. Those NPs are usually randomly distributed and initially suspended in a solvent such as water. Therefore, they are totally independent, that is, the plasmonic resonance is purely localized and uncoupled. To study the plasmon effect of Au NPs, a simple model consisting of a single NP is demonstrated. The simulation region

The plasmon resonance frequency is highly sensitive to the refractive index (n) of the surrounding environment. Hence, a change of n results in a shift in the resonant frequency. We have considered a spherical Au NP immersed in various media, such as air, water, or glass. As n near the NP surface increases, the absorption coefficients increase and the absorption spectrum shifts to longer wavelengths as shown in **Figure 2(a)**. Theoretically, this means that the LSPR peak location will be red-shifted if the surrounding medium changes from air to water and to oil.

Optical properties of Au NPs can also be tuned by varying their sizes and shapes. **Figure 2(b)** shows the calculated absorption spectra of Au nano-rods with different aspect ratios (the diameter a was fixed at 15 nm, and the ratio R = 1, 2, 2.5, and 3). We can see that the longitudinal mode, which is commonly referred to the plasmonic band induced by exciting along the long axis of nano-rods, is significantly red-shifted, while the transverse mode, which is commonly referred to the plasmonic band when exciting along the short axis, exhibits a slight

Consider now another kind of NP, namely Au nano-prism. The size and shape dependence of plasmonic band can also be seen clearly in this case. The LSPR position is highly sensitive to the edge length and the aspect ratio (defined as the ratio d/t between the edge length, d, and

**Figure 2.** (a) Numerical calculation of absorption spectra of Au NPs (diameter d = 50 nm) in different media (air, water, and glass, indicated by its refractive index, n = 1, 1.33, and 1.5, respectively). Inset: Illustration of a single Au NP in a medium with refractive index n. (b) Calculated absorption spectra of Au nano-rods in water with different aspect ratios, R. The diameter of the Au nano-rod is fixed at a = 15 nm. Inset: design of Au nano-rod. (c) Size-dependent absorption spectra of Au prism in water, calculated with different edge-lengths, d. The thickness of the Au prism is fixed at t = 25 nm.

box with perfectly matched layer boundary conditions (PML BCs) for all x-,

Arbitrary Form Plasmonic Structures: Optical Realization, Numerical Analysis and Demonstration…

is set in the region encompassing the Au NP and

http://dx.doi.org/10.5772/intechopen.79236

69

For calculation of plasmonic properties of all gold (Au) nanostructures (NSs) presented in this chapter, we have performed simulations by using the FDTD method (Lumerical software). **Figure 1** shows the FDTD simulation model. The simulation area was bounded in x- and y-directions by parallel planes in which periodical boundary conditions are defined, while perfectly matched layer metal boundary conditions were applied in top and bottom boundaries to prevent any reflections. The absorbance spectra were calculated from Fourier transform time-dependent transmission monitor. These metallic NSs could be computationally created or imported from real structures, which are reconstructed by using the scanning electron microscope (SEM) and atomic force microscope (AFM) data. Indeed, by using Lumerical software, it is possible to construct models for single or multiple NPs, with changeable particle parameters (size, shape, materials, compositions, etc.). It is also able to simulate NPs organized in order, such as dimer and arrays, with uniform distribution or random distribution. This software also allows to import real SEM and AFM images to simulate with true fabricated structures. Other FDTD model parameters are set as close to characterization conditions: the optical properties of materials were taken from [20] for SiO2 substrate and from [21] for Au thin films.

**Figure 1.** FDTD model used to simulate the plasmonic properties of Au NSs, including Au NPs of different shapes and Au NHAs. These NSs could be computationally created or imported from real structures, which are reconstructed by using SEM and AFM data.

### **2.2. Localized surface plasmon resonance**

technology [18]. Recently, with the assistance of modern computers, FDTD has become a powerful technique for researchers to predict the EM response of any structure, in particular PNSs. As other finite-difference modeling methods, FDTD is grid-based. It means that computational domains will be meshed into minor units, mostly in a cube shape, with associated vector components of the electric (**E**) field and magnetic (**H**) field, which are determined by Maxwell's equations. In simulation, the **E** and **H** fields are calculated at every point in space, forward in time. Additionally, FDTD is a time-domain method, so it can cover a wide range of frequencies with a single simulation run [19]. Thanks to these advantages, FDTD provides us a natural way to treat any electrodynamics problem, especially periodic structures and

For calculation of plasmonic properties of all gold (Au) nanostructures (NSs) presented in this chapter, we have performed simulations by using the FDTD method (Lumerical software). **Figure 1** shows the FDTD simulation model. The simulation area was bounded in x- and y-directions by parallel planes in which periodical boundary conditions are defined, while perfectly matched layer metal boundary conditions were applied in top and bottom boundaries to prevent any reflections. The absorbance spectra were calculated from Fourier transform time-dependent transmission monitor. These metallic NSs could be computationally created or imported from real structures, which are reconstructed by using the scanning electron microscope (SEM) and atomic force microscope (AFM) data. Indeed, by using Lumerical software, it is possible to construct models for single or multiple NPs, with changeable particle parameters (size, shape, materials, compositions, etc.). It is also able to simulate NPs organized in order, such as dimer and arrays, with uniform distribution or random distribution. This software also allows to import real SEM and AFM images to simulate with true fabricated structures. Other FDTD model parameters are set as close to characterization conditions: the optical properties

**Figure 1.** FDTD model used to simulate the plasmonic properties of Au NSs, including Au NPs of different shapes and Au NHAs. These NSs could be computationally created or imported from real structures, which are reconstructed by

substrate and from [21] for Au thin films.

broadband sources in PNSs complications.

68 Plasmonics

of materials were taken from [20] for SiO2

using SEM and AFM data.

Au NPs possessing particular shapes, such as nano-sphere, nano-rod, and nano-prism, are often fabricated by standard chemical methods, for example, the Turkevich [22], Brust [23] seeded growth [24] or other miscellaneous methods [25]. Those NPs are usually randomly distributed and initially suspended in a solvent such as water. Therefore, they are totally independent, that is, the plasmonic resonance is purely localized and uncoupled. To study the plasmon effect of Au NPs, a simple model consisting of a single NP is demonstrated. The simulation region is a 1 × 1 × 1 μm3 box with perfectly matched layer boundary conditions (PML BCs) for all x-, y-, and z-axis. The mesh size of 1 × 1 × 1 nm3 is set in the region encompassing the Au NP and Au-medium interface.

The plasmon resonance frequency is highly sensitive to the refractive index (n) of the surrounding environment. Hence, a change of n results in a shift in the resonant frequency. We have considered a spherical Au NP immersed in various media, such as air, water, or glass. As n near the NP surface increases, the absorption coefficients increase and the absorption spectrum shifts to longer wavelengths as shown in **Figure 2(a)**. Theoretically, this means that the LSPR peak location will be red-shifted if the surrounding medium changes from air to water and to oil.

Optical properties of Au NPs can also be tuned by varying their sizes and shapes. **Figure 2(b)** shows the calculated absorption spectra of Au nano-rods with different aspect ratios (the diameter a was fixed at 15 nm, and the ratio R = 1, 2, 2.5, and 3). We can see that the longitudinal mode, which is commonly referred to the plasmonic band induced by exciting along the long axis of nano-rods, is significantly red-shifted, while the transverse mode, which is commonly referred to the plasmonic band when exciting along the short axis, exhibits a slight blue shift as the aspect ratio R increases.

Consider now another kind of NP, namely Au nano-prism. The size and shape dependence of plasmonic band can also be seen clearly in this case. The LSPR position is highly sensitive to the edge length and the aspect ratio (defined as the ratio d/t between the edge length, d, and

**Figure 2.** (a) Numerical calculation of absorption spectra of Au NPs (diameter d = 50 nm) in different media (air, water, and glass, indicated by its refractive index, n = 1, 1.33, and 1.5, respectively). Inset: Illustration of a single Au NP in a medium with refractive index n. (b) Calculated absorption spectra of Au nano-rods in water with different aspect ratios, R. The diameter of the Au nano-rod is fixed at a = 15 nm. Inset: design of Au nano-rod. (c) Size-dependent absorption spectra of Au prism in water, calculated with different edge-lengths, d. The thickness of the Au prism is fixed at t = 25 nm. Inset: design of Au nano-prism.

thickness, t, of nano-prisms). The larger edge lengths and higher aspect ratio generally result in red shifted resonances. The red shift of the position of the peak maxima (from 642 to 684 nm) corresponds to the edge length of Au nano-prism increases (from 56 to 112 nm) as illustrated in **Figure 2(c)**. This result is quite interesting as compared to that obtained by a spherical NP having a similar size. That allows to easily create different colors by using metallic NPs with different shapes.

From three examples of nano-sphere, nano-rod, and nano-prism, there exists thus a strong dependence of the LSPR peaks (number of modes, location, and intensity) on metallic NPs shape, size, and surrounding medium. Generally, a red shift of absorption spectrum is noticed when NPs size increases. In addition, increment of surrounding refractive index can increase absorption coefficient and enlarge red-shift absorption spectrum. This general observation gives us an insight to construct a numerical model for multiple-sizes/shape nano-islands (NIs) as well as to qualitatively explain their use as plasmonic sensors and/or data storage and color nanoprinter.

### **2.3. Surface plasmon resonance of randomly distributed Au nanoparticles**

Another method, called thermal annealing dewetting technique [19, 26, 27], has recently been demonstrated as an excellent way to produce monolayers of randomly distributed Au NPs on a substrate. In this technique, discontinuous thin metallic films are first deposited on a substrate, such as a glass, and then annealed at a temperature of about 500°C. The discontinuous metallic films are thermally melted resulting in isolated metallic NIs. In this case, it is not correct if one calculates the plasmonic properties of only a single NP but it should be done for an ensemble of NPs of various sizes. There exist a few of works in literature, which attempted to model the optical properties of a monolayer of Au NIs [28, 29]. However, most of the works were based on statistical methods to estimate the median parameters of NIs, which required a lot of raw data and statistical analysis efforts. Moreover, in order to simplify the models, an assumption of semispherical NIs was generally used, which might greatly alter the final calculation results since LSPR is highly sensitive to size and shape of NIs.

plasmonic-photonic crystal cavity [32], grating coupler with waveguide resonant grating [33], and so on. In this section, Au NHAs are demonstrated theoretically as an excellent optical bandpass filter via the well-known extraordinary optical transmission (EOT) phenomenon [34]. For complete simulations of NHAs, the parameters were selected and cat-

**Figure 3.** Simulation result of absorbance spectra of Au NPs randomly distributed on glass substrates. Different curves correspond to samples having different average NPs sizes. Insets show SEM images of a thin film Au sample obtained before annealing, corresponding to the absorption spectrum of gray color, and of an Au NPs sample obtained after

Arbitrary Form Plasmonic Structures: Optical Realization, Numerical Analysis and Demonstration…

http://dx.doi.org/10.5772/intechopen.79236

71

thermal annealing, corresponding to one of the plasmonic resonance absorption spectra.

**Fixed parameters**: NHA structures with round holes were chosen because it is the most similar to the structure fabricated by direct laser writing lithography, which will be shown in the fabrication section. The optical constants of silica, Au, and Cr were taken from [20, 21], respectively. Note that Cr is usually used for enhancing the adhesion of Au material with substrates, and it should be thin to avoid the influence of plasmonic properties of Au structures. The periodicity of NHAs was fixed at 1000 nm. The net transmission light was calculated using the arithmetic average of the simulation results of individual and orthogonal polarizations.

**Swept parameters**: the thickness of Au layer and Cr layer were swept from tAu = 10 to 90 nm and tCr = 0 to 20 nm, respectively. Diameter of the hole was also varied from dhole = 300 to 900 nm. In addition, monitors were set within computation domain to analyze transmission spectra along with EM field distributions. Results were normalized to the transmission spectra obtained from a glass substrate (no NHA) and evaluated right after each simulation to

**Figure 4** shows calculated transmission spectra of Au NHA as a function tAu, tCr, and dhole. As can be seen in **Figure 4(a)**, the EOT peak position and width depend strongly on the thickness of Au NHA. In detail, this peak was blue-shifted and became shaper when the Au thickness increased. In contrast, an increase of Cr layer thickness suppresses the transmission peaks dramatically (**Figure 4(c)**). Another parameter, which influences greatly the transmission spectrum, is hole diameter, dhole. Bigger holes allow higher transmission coefficients; however,

confine the parameters, with a purpose of the fastest convergence.

egorized into two groups as described below:

The calculation method shown in **Figure 1** allowed thus solving completely the problem. This process required only a SEM image and an AFM image. SEM image is utilized to extract the top-view sizes/shapes and (x,y) position coordinates of NIs. This two-dimensional (2D) map is then extruded to three-dimensional (3D) structures with the estimated height from AFM data. **Figure 3** shows the plasmonic resonant spectra of monolayer films of Au NIs with different average NIs sizes. The plasmonic peak located around 550 nm is clearly evidenced. When the particles size increases, the corresponding resonant peaks of the absorption spectra are red-shifted. Theoretically, the resonance peak shifts by a quantity of about 48.5 nm when the average size of NIs changes from 20 to 100 nm. This shift suggests that the structuration of Au NPs by the dewetting method can be a good idea for applications such as tunable absorbers and color nanoprinter.

### **2.4. Surface plasmon resonance in periodic Au nano-holes array**

Recently, several strategies have been employed to realize a metallic optical filter, including plasmonic resonators [30], plasmonic metasurfaces for perfect light absorption [31], Arbitrary Form Plasmonic Structures: Optical Realization, Numerical Analysis and Demonstration… http://dx.doi.org/10.5772/intechopen.79236 71

thickness, t, of nano-prisms). The larger edge lengths and higher aspect ratio generally result in red shifted resonances. The red shift of the position of the peak maxima (from 642 to 684 nm) corresponds to the edge length of Au nano-prism increases (from 56 to 112 nm) as illustrated in **Figure 2(c)**. This result is quite interesting as compared to that obtained by a spherical NP having a similar size. That allows to easily create different colors by using metallic NPs with

From three examples of nano-sphere, nano-rod, and nano-prism, there exists thus a strong dependence of the LSPR peaks (number of modes, location, and intensity) on metallic NPs shape, size, and surrounding medium. Generally, a red shift of absorption spectrum is noticed when NPs size increases. In addition, increment of surrounding refractive index can increase absorption coefficient and enlarge red-shift absorption spectrum. This general observation gives us an insight to construct a numerical model for multiple-sizes/shape nano-islands (NIs) as well as to qualitatively explain their use as plasmonic sensors and/or data storage and color

Another method, called thermal annealing dewetting technique [19, 26, 27], has recently been demonstrated as an excellent way to produce monolayers of randomly distributed Au NPs on a substrate. In this technique, discontinuous thin metallic films are first deposited on a substrate, such as a glass, and then annealed at a temperature of about 500°C. The discontinuous metallic films are thermally melted resulting in isolated metallic NIs. In this case, it is not correct if one calculates the plasmonic properties of only a single NP but it should be done for an ensemble of NPs of various sizes. There exist a few of works in literature, which attempted to model the optical properties of a monolayer of Au NIs [28, 29]. However, most of the works were based on statistical methods to estimate the median parameters of NIs, which required a lot of raw data and statistical analysis efforts. Moreover, in order to simplify the models, an assumption of semispherical NIs was generally used, which might greatly alter the final

The calculation method shown in **Figure 1** allowed thus solving completely the problem. This process required only a SEM image and an AFM image. SEM image is utilized to extract the top-view sizes/shapes and (x,y) position coordinates of NIs. This two-dimensional (2D) map is then extruded to three-dimensional (3D) structures with the estimated height from AFM data. **Figure 3** shows the plasmonic resonant spectra of monolayer films of Au NIs with different average NIs sizes. The plasmonic peak located around 550 nm is clearly evidenced. When the particles size increases, the corresponding resonant peaks of the absorption spectra are red-shifted. Theoretically, the resonance peak shifts by a quantity of about 48.5 nm when the average size of NIs changes from 20 to 100 nm. This shift suggests that the structuration of Au NPs by the dewetting method can be a good idea for applications such as tunable absorbers

Recently, several strategies have been employed to realize a metallic optical filter, including plasmonic resonators [30], plasmonic metasurfaces for perfect light absorption [31],

**2.3. Surface plasmon resonance of randomly distributed Au nanoparticles**

calculation results since LSPR is highly sensitive to size and shape of NIs.

**2.4. Surface plasmon resonance in periodic Au nano-holes array**

different shapes.

70 Plasmonics

nanoprinter.

and color nanoprinter.

**Figure 3.** Simulation result of absorbance spectra of Au NPs randomly distributed on glass substrates. Different curves correspond to samples having different average NPs sizes. Insets show SEM images of a thin film Au sample obtained before annealing, corresponding to the absorption spectrum of gray color, and of an Au NPs sample obtained after thermal annealing, corresponding to one of the plasmonic resonance absorption spectra.

plasmonic-photonic crystal cavity [32], grating coupler with waveguide resonant grating [33], and so on. In this section, Au NHAs are demonstrated theoretically as an excellent optical bandpass filter via the well-known extraordinary optical transmission (EOT) phenomenon [34]. For complete simulations of NHAs, the parameters were selected and categorized into two groups as described below:

**Fixed parameters**: NHA structures with round holes were chosen because it is the most similar to the structure fabricated by direct laser writing lithography, which will be shown in the fabrication section. The optical constants of silica, Au, and Cr were taken from [20, 21], respectively. Note that Cr is usually used for enhancing the adhesion of Au material with substrates, and it should be thin to avoid the influence of plasmonic properties of Au structures. The periodicity of NHAs was fixed at 1000 nm. The net transmission light was calculated using the arithmetic average of the simulation results of individual and orthogonal polarizations.

**Swept parameters**: the thickness of Au layer and Cr layer were swept from tAu = 10 to 90 nm and tCr = 0 to 20 nm, respectively. Diameter of the hole was also varied from dhole = 300 to 900 nm. In addition, monitors were set within computation domain to analyze transmission spectra along with EM field distributions. Results were normalized to the transmission spectra obtained from a glass substrate (no NHA) and evaluated right after each simulation to confine the parameters, with a purpose of the fastest convergence.

**Figure 4** shows calculated transmission spectra of Au NHA as a function tAu, tCr, and dhole. As can be seen in **Figure 4(a)**, the EOT peak position and width depend strongly on the thickness of Au NHA. In detail, this peak was blue-shifted and became shaper when the Au thickness increased. In contrast, an increase of Cr layer thickness suppresses the transmission peaks dramatically (**Figure 4(c)**). Another parameter, which influences greatly the transmission spectrum, is hole diameter, dhole. Bigger holes allow higher transmission coefficients; however,

**Figure 4.** Calculated transmission spectra of Au NHAs as a function of Au layer thickness, tAu (a); of nano-hole diameter, dhole (b); and of Cr layer thickness, tCr, (c), respectively.

band selection is poor. Conversely, smaller holes allow fewer transmission resonance modes, which make transmission peak sharper but also decrease the transmission coefficients. Based on this insight along with advantages and disadvantages of the direct laser writing method, an optimum NHA should have the following parameters: tAu = 50 nm, tCr = 3 nm, Λ = 1000 nm, and dhole = 400 nm, respectively.

spectrum of both S1805 and Au materials, is tightly focused into samples by a high numerical aperture (NA) objective lens (OL). Since the DLW operates with an OPA mechanism, the required laser power is very modest, in the range of few microwatts for S1805 photoresist and a few dozens of milliwatts for Au films. Thanks to the use of a high NA OL, the light intensity at the focusing region is, however, very high, which is enough for depolymerizing the S1805 photoresist and thermally dewetting the Au films. 3D piezoelectric translator (PZT) connected to a computer control allows the focusing spot to move through the sample following a desired trajectory. By controlling the laser power and the exposure time, the exposure doses are adjusted resulting in structures with desired sizes and forms, as illustrated in **Figure 5(b)**. A detection system consisting of a lenses ensemble, a pinhole, and an avalanche photodiode is used to determine the focusing position, which should be practically located on substrate surface. It should be noted that this DLW is time consuming, like e-beam lithography. Also, in order to keep high resolution, the total area of fabricated structures is limited, generally

**Figure 5.** (a) Illustration of the DLW technique used to realize arbitrary 2D structures on photoresist and Au film. PZT: piezoelectric translator; DM: dichroic mirror; OL: objective lens. (b) Control of filling factor of structures fabricated on a positive photoresist by adjusting the exposure dose. A 2D square structure is obtained by scanning continuously the focal spot in x- and y-directions. Top: theoretical light pattern; bottom: experimental demonstration. The separation between lines, that is, the period of structure, is Λ = 0.8 μm, and the structures change from negative (air-holes) to

Arbitrary Form Plasmonic Structures: Optical Realization, Numerical Analysis and Demonstration…

http://dx.doi.org/10.5772/intechopen.79236

73

necessary, it could be enlarged by using a PZT with a larger scanning range, together with an

The indirect fabrication of PNSs consists of two steps: (i) fabrication of photoresist templates by DLW method and (ii) transferring templates to metallic structures by evaporation of method and template lift-off. **Figure 6(a)** illustrates the fabrication process of Au NSs by this indirect method. This process is very similar to the fabrication of PNSs by e-beam lithography [35–37]. To fabricate desired structures, the positive photoresist was first coated on cleaned glass substrates and exposed by the DLW system. The samples were then developed, removing all exposed parts and leaving unexposed parts as desired structures. It is demonstrated

**3.2. Realization of plasmonic structure by an indirect method**

. This surface should be enough for various applications, and in case

of about 100x100 μm2

increase of fabrication time.

positive (polymeric cylinders) forms.
