**3. Investigation of optical properties of specific noble metal nanoparticles and nanoarrays by the combination of NSL and multi-hierarchy arrayed micro windows**

The physicochemical properties of nanomaterials significantly depend on their threedimensional (3D) morphologies (sizes, shapes and surface topography), their surrounding media, and their spatial arrangement. Systematically and precisely correlating these parameters with the related physicochemical properties of specific single nanoparticles (NPs) or nanoarrays is a fundamental requirement for the discovery of their novel properties and applications, as well as for advancing the fundamental and practical knowledge required for the design and fabrication of new materials(Song , Zhang et al. 2011). The lack of effective means of fabricating recognizable 3D morphologies controlled NPs and nanoarrays and correlating their structure parameters with their physicochemical properties as observed by different characterization techniques represents an obstacle for studying the 3D morphology-dependent properties of individual NPs and nanoarrays(Song, Zhang et al. 2011). Most current studies investigate the physicochemical properties of the NP ensemble, but not of a single NP(Jin, Cao et al. 2001; Kelly, Coronado et al. 2003; Haes, Zou et al. 2004; Song , Zhang et al. 2011). The ensemble of NPs is typically heterogeneous, because the morphologies of individual NPs prepared by routine chemical synthesis or physical vapor fabrication methods are rarely identical at the nanometre or sub-nanometre scale (Song , Zhang et al. 2011). Effective methods for 3D morphology controlled fabrication of nanomaterials, and to correlate their 3D morphology of single NPs or nanoarrays with their physicochemical properties are also essential to address fundamental and practical questions related to the single NPs (Song , Zhang et al. 2011).

An important research area in nanoscale plasmonic optics is single NP identification and characterization of their 3D morphologies and space-orientation dependent physicochemical

Controlled Fabrication of Noble Metal Nanomaterials

Royal Society of Chemistry.)

via Nanosphere Lithography and Their Optical Properties 515

Fig. 7. The multi-hierarchy arrayed micro windows on the substrate (e.g., glass cover slip). (a) The first tier of the multi-hierarchy arrayed micro window, each local area can be discerned by marking its X and Y number, such as the red-dashed square area of X1–Y2. (b) The second tier of the multi-hierarchy arrayed micro windows, whose scale can be reduced by M or N times, whose local area can also be marked by x and y numbers. If this net area is the sub-tier n the red area of the first tier, it can be labeled as X1–Y2–x3–y3. Step-by-step, the

last tier with several unique-shaped transparent windows can be reached. The open windows can be made with different shapes. (c) The nanoparticles can be fabricated on the micro-pattern by various methods (e.g., nanosphere lithography). In each window, the same nanoparticle can be identified by comparing the images taken by optical microscopy, AFM, or other microscopy methods. Finally, the structural parameters (size, shape, orientation, interparticle spacing, and thickness) can be correlated with their optical responses

(Reprinted from Song Y.; et al., Nanoscale 2011, 3, 31-44, Figure 7, copyright (2011) from the

A typical example to identify NPs and nanoarrays using both AFM and DFOMS is illustrated in Figure 8. Triangular Ag NPs and hexagon-arranged nanoarrays fabricated on the surface of glass cover slips within the nearly circle-shaped micro window can be identified and characterized using AFM (Figure 8A, 8B is the 3D AFM image of the dashsquared area in 8A) and DFOMS equipped with a color camera (Figure 8 C) and chargecoupled device (CCD) camera (Figure 8D). The CCD camera offers higher spatial resolution than the color camera, while the color camera provides the real colors of individual Ag NPs that are generated by LSPR. The center of each individual NP in the optical images recorded by the CCD is located with a single-pixel resolution (each pixel can be 125 nm or 67 nm depending on the CCD resolution and equipment setup) by determining the address of the pixel with the highest intensity of the NP. The positions of individual NPs of interest (e.g. the circled one) within the micro window in the optical images (Figure 8C and D) are then determined with a spatial resolution limited by the optical diffraction limit (~200 nm) and an orientation angle resolution of about 1.0 degree. This approach allows us to correlate AFM images of individual NPs (as the one circled in each image) with the same NP shown in its corresponding optical image and to investigate its 3D morphological-dependent LSPR properties. Clearly, these triangle nanoparticles in this window almost show the same scattering color (Figure 8C) and intensity contrast (Figure 8D). By comparing their scattering color images (Figure 8C) with their AFM images (Figure 8A and B) of these nanoparticles, it

properties(Yang, Matsubara et al. 2007; Song 2009; Song , Zhang et al. 2011). Recently, much attention has been given to the localized surface plasmon resonance (LSPR) of metal NPs because of their promising applications in plasmonic circuits, optoelectronic transducers, optical bioprobes, and surface plasmon resonance interference lithography(Shen, Friend et al. 2000; Prasad 2004; Ozbay 2006; Song 2009; Song , Henry et al. 2009; Song , Jin et al. 2010; Song, Sun et al. 2010; Song , Zhang et al. 2011). Since the plasmonic properties of metal NPs intrinsically rely on their size, shape, surface topography, crystal structure, inter-particle spacing and the dielectric environment around them, methods to correlate their plasmonic properties with the above structural and environmental parameters have become one of the most rapidly developing research directions (Song , Zhang et al. 2011).

In the precise investigation of the relationship between the LSPR properties and their 3D morphologies of specific nanoparticles and nanoarrays, two kinds of methods have been developed recently, or the *in situ* method and the spatial-localization method (Song , Zhang et al. 2011). The *in situ* method combines at least two different instruments together to conduct the structure and property characterization simultaneously: one can be used to characterize the 3D morphology (e.g. AFM or STEM) of NPs and the others will be used to chatacteize the LSPR-related optical properties of the same NPs (e.g. Dark-field microscope and spectroscopy). The spatial-localization method requires using markers to recoganize the same single nanoparticle in different instruments. We have also developed one spatiallocalization method to precisely investigate the 3D morphologies dependent LSPR properties of specific NPs and nanoarrays by the combination of NSL and traditional UV-LIGA, where Ag NPs and nanoarrays can be fabricated by NSL in the pre-formed multihierarchy arrayed transparent micro-windows on the substrates (e.g., glass cover slip) by the UV-LIGA(Song 2009; Song , Zhang et al. 2011). This technique permits easy characterization of the 3D morphologies of single NPs by AFM or SEM and their LSPR spectra using darkfield optical microscopy and spectroscopy (DFOMS). It is also possible to investigate the local morphology dependence of the LSPR spectra of the single NPs and nanoarrays. In this method, multi-hierarchy arrayed micro windows are first fabricated on a glass cover slip using the standard photolithography, whose details are shown in reference 27. Fig. 7A and Fig. 7B show one example of the designed multi-hierarchy arrayed micro windows (3 tiers) and the typical final micro-windows (Fig. 7C) pattern after printing. The multi-hierarchy arrayed micro-windows on the glass cover slip are used to identify the location and orientation of single NPs, whose tiers can be determined by the observed field at desired resolution. For example, in the first tier of the multi-hierarchy arrayed micro windows (Fig. 7A), each local area can be discerned by marking its X and Y number, such as the shaded area X1–Y2. Then, in the second tier of the multi-hierarchy arrayed micro windows (Fig. 7B), the scale can be reduced by M or N times and each local area can also be marked by x and y number. If this area is the sub-tier in the shaded area of the first tier, it can be labeled as X1– Y2–x3–y3. In a similar way, step-by-step, we can reach the last tier with several transparent micro windows available (Fig. 7C), in which the desired nanoparticle can be made by different fabrication methods (e.g., electron beam lithography or nanosphere lithography). Nanoparticles less than 10 nm of different shapes synthesized by a wet-chemical process can be immobilized by a routine diluted deposition process. Consequently, the same nanoparticle in each window can be identified by comparing the images taken by the optical microscope with those characterized by the AFM. Finally, in each window, the same nanoparticle can be characterized by different techniques (e.g., DFOMS and AFM) allowing correlation of its 3D morphology with its optical response(Song 2009; Song , Zhang et al. 2011).

properties(Yang, Matsubara et al. 2007; Song 2009; Song , Zhang et al. 2011). Recently, much attention has been given to the localized surface plasmon resonance (LSPR) of metal NPs because of their promising applications in plasmonic circuits, optoelectronic transducers, optical bioprobes, and surface plasmon resonance interference lithography(Shen, Friend et al. 2000; Prasad 2004; Ozbay 2006; Song 2009; Song , Henry et al. 2009; Song , Jin et al. 2010; Song, Sun et al. 2010; Song , Zhang et al. 2011). Since the plasmonic properties of metal NPs intrinsically rely on their size, shape, surface topography, crystal structure, inter-particle spacing and the dielectric environment around them, methods to correlate their plasmonic properties with the above structural and environmental parameters have become one of the

In the precise investigation of the relationship between the LSPR properties and their 3D morphologies of specific nanoparticles and nanoarrays, two kinds of methods have been developed recently, or the *in situ* method and the spatial-localization method (Song , Zhang et al. 2011). The *in situ* method combines at least two different instruments together to conduct the structure and property characterization simultaneously: one can be used to characterize the 3D morphology (e.g. AFM or STEM) of NPs and the others will be used to chatacteize the LSPR-related optical properties of the same NPs (e.g. Dark-field microscope and spectroscopy). The spatial-localization method requires using markers to recoganize the same single nanoparticle in different instruments. We have also developed one spatiallocalization method to precisely investigate the 3D morphologies dependent LSPR properties of specific NPs and nanoarrays by the combination of NSL and traditional UV-LIGA, where Ag NPs and nanoarrays can be fabricated by NSL in the pre-formed multihierarchy arrayed transparent micro-windows on the substrates (e.g., glass cover slip) by the UV-LIGA(Song 2009; Song , Zhang et al. 2011). This technique permits easy characterization of the 3D morphologies of single NPs by AFM or SEM and their LSPR spectra using darkfield optical microscopy and spectroscopy (DFOMS). It is also possible to investigate the local morphology dependence of the LSPR spectra of the single NPs and nanoarrays. In this method, multi-hierarchy arrayed micro windows are first fabricated on a glass cover slip using the standard photolithography, whose details are shown in reference 27. Fig. 7A and Fig. 7B show one example of the designed multi-hierarchy arrayed micro windows (3 tiers) and the typical final micro-windows (Fig. 7C) pattern after printing. The multi-hierarchy arrayed micro-windows on the glass cover slip are used to identify the location and orientation of single NPs, whose tiers can be determined by the observed field at desired resolution. For example, in the first tier of the multi-hierarchy arrayed micro windows (Fig. 7A), each local area can be discerned by marking its X and Y number, such as the shaded area X1–Y2. Then, in the second tier of the multi-hierarchy arrayed micro windows (Fig. 7B), the scale can be reduced by M or N times and each local area can also be marked by x and y number. If this area is the sub-tier in the shaded area of the first tier, it can be labeled as X1– Y2–x3–y3. In a similar way, step-by-step, we can reach the last tier with several transparent micro windows available (Fig. 7C), in which the desired nanoparticle can be made by different fabrication methods (e.g., electron beam lithography or nanosphere lithography). Nanoparticles less than 10 nm of different shapes synthesized by a wet-chemical process can be immobilized by a routine diluted deposition process. Consequently, the same nanoparticle in each window can be identified by comparing the images taken by the optical microscope with those characterized by the AFM. Finally, in each window, the same nanoparticle can be characterized by different techniques (e.g., DFOMS and AFM) allowing correlation of its 3D morphology with its optical response(Song 2009; Song , Zhang et al.

most rapidly developing research directions (Song , Zhang et al. 2011).

2011).

Fig. 7. The multi-hierarchy arrayed micro windows on the substrate (e.g., glass cover slip). (a) The first tier of the multi-hierarchy arrayed micro window, each local area can be discerned by marking its X and Y number, such as the red-dashed square area of X1–Y2. (b) The second tier of the multi-hierarchy arrayed micro windows, whose scale can be reduced by M or N times, whose local area can also be marked by x and y numbers. If this net area is the sub-tier n the red area of the first tier, it can be labeled as X1–Y2–x3–y3. Step-by-step, the last tier with several unique-shaped transparent windows can be reached. The open windows can be made with different shapes. (c) The nanoparticles can be fabricated on the micro-pattern by various methods (e.g., nanosphere lithography). In each window, the same nanoparticle can be identified by comparing the images taken by optical microscopy, AFM, or other microscopy methods. Finally, the structural parameters (size, shape, orientation, interparticle spacing, and thickness) can be correlated with their optical responses (Reprinted from Song Y.; et al., Nanoscale 2011, 3, 31-44, Figure 7, copyright (2011) from the Royal Society of Chemistry.)

A typical example to identify NPs and nanoarrays using both AFM and DFOMS is illustrated in Figure 8. Triangular Ag NPs and hexagon-arranged nanoarrays fabricated on the surface of glass cover slips within the nearly circle-shaped micro window can be identified and characterized using AFM (Figure 8A, 8B is the 3D AFM image of the dashsquared area in 8A) and DFOMS equipped with a color camera (Figure 8 C) and chargecoupled device (CCD) camera (Figure 8D). The CCD camera offers higher spatial resolution than the color camera, while the color camera provides the real colors of individual Ag NPs that are generated by LSPR. The center of each individual NP in the optical images recorded by the CCD is located with a single-pixel resolution (each pixel can be 125 nm or 67 nm depending on the CCD resolution and equipment setup) by determining the address of the pixel with the highest intensity of the NP. The positions of individual NPs of interest (e.g. the circled one) within the micro window in the optical images (Figure 8C and D) are then determined with a spatial resolution limited by the optical diffraction limit (~200 nm) and an orientation angle resolution of about 1.0 degree. This approach allows us to correlate AFM images of individual NPs (as the one circled in each image) with the same NP shown in its corresponding optical image and to investigate its 3D morphological-dependent LSPR properties. Clearly, these triangle nanoparticles in this window almost show the same scattering color (Figure 8C) and intensity contrast (Figure 8D). By comparing their scattering color images (Figure 8C) with their AFM images (Figure 8A and B) of these nanoparticles, it

Controlled Fabrication of Noble Metal Nanomaterials

**200 nm 200 nm** 

**B A** 

via Nanosphere Lithography and Their Optical Properties 517

**C**

**0 0.5 1 1.5 2**

**N.M.Intensity (a.u.)**

**400 600 800 1000 1200 Wavelength (nm)**

**759.6**

**1020**

**d** 

**c b** 

**a** 

**562.3 659.0 540 662**

**D**

Fig. 9. (A): The AFM image for one single triangular shaped Ag nanoparticle with edge length of 375-420 nm; (B) the height mapping of the triangle shape Ag nanoparticle along the direction of the arrow in Fig. 9(A), showing the out-of-plane height of this nanoparticle of 16.1 nm; (C) the real scattering color image of this triangular shaped Ag nanoparticle; (Da): the LSPR absorption spectrum of this nanoparticle by DDA; (D-b) the LSPR scattering by experiment; (D-c) the LSPR scattering by DDA; (D-d): the LSPR extinction by DDA. In order to identify the location and the orientation of these positions around the NPs, the AFM image and color image were netted by dashed lines with each square unit of 125 nm ×125 nm after their distances and orientations were corrected. (Adapted from references: Y. Song, China Patent, Appl. No. CN200910085973.9; Song Y.; et al., Nanoscale 2011, 3, 31-44, Figure 13, copyright (2011) from the Royal Society of Chemistry. Adapted with permission.).

<sup>i</sup> <sup>i</sup><sup>i</sup> i

This combined method based on the NSL and the multi-hierarchy arrayed micro windows also allows us to investigate the 3D morphology dependent tip-tip LSPR coupling of triangular nanoparticle pairs. The zoom-in AFM image for the detailed 3D morphology of one typical Ag nanoprism pair is shown in Figure 10A. The nanoprisms have almost the same edge size ~ 375 nm and maximum out-of-plane height ~ 17.1 nm shown in Figure 10B by the typical height map along the arrowed tip-tip direction in Figure 10A. The real scattered color for the nanoprism pair, taken from dark-field microscopy, is shown in Figure 10C. Both of the nanoprisms in the pair give red color with different brightness, which might be due to variation in their surface roughness, slightly difference in the underlying surrounding dielectrics, and the focusing distance during image recording. The middle area between the two nanoprisms clearly shows more reddish color than the optical centers of the two nanoprisms. The LSPR spectrum for the middle area of the two optical centers (representing the tip-tip-coupling) is recorded in Figure 10D using their CCD image (not shown here) for the location identification, together with that obtained by the discrete dipole approximation (DDA) calculation of the nanoprism pair. According to its 3D morphology of the nanoparticle pair, the two nanoprisms can be treated as regular triangular nanoprisms with the bottom edge length of 375 nm, the top edge length of 125 nm and out-of-plane height of 17.1 nm for conducting the DDA calculation of the nanoprism pair. The recorded LSPR spectrum (Figure 10D: a) at the middle optical center of the two nanoprisms shows three distinct peaks, one strongest peak at 605 nm, one shoulder at 536 nm, and one secondary strong peak at 754 nm. By comparing the experimental result for the tip-tip coupling of the nanoprism pair with the DDA calculation (Figure 10D: b), it can be deduced that the peak at 605 nm represents the in-plane quadrupole resonances originated from the two source nanoprisms and the peak at 536 nm is from the out-of-plane quadrapole resonances of the two source nanoprisms. Although the DDA simulation does not show one distinct peak at 754 nm, our experiment result suggest one strong peak at this wavelength, which is probably from the strong tip-tip coupling. In order to reveal whether the peak at 754 nm is mainly from the tip-tip coupling or not, the LSPR spectra from the optical centers

is once again showing that NSL is powerful method in the fabrication of uniform triangular nanoparticles and nanoarrays.

Fig. 8. One example for the identification of the specific nanoparticles and nanoarrays in different instruments via multi-hierarchy arrayed micro windows based on Ag triangle nanoparticles and nanoarrays fabricated by NSL in one nearly circle-shaped window. (A) The plan view of the hexagon-arrayed triangle Ag NPs in one circle micro-window scanned by atomic force microscope (AFM); (B) the 3D view of the hexagon-arrayed triangle Ag NPs marked in the large pink dash-square in (A); (C) the real scattering color of these hexagonarrayed triangle Ag NPs observed under a dark-field microscope; (D) the CCD images of the scattering light of these hexagon-arrayed triangle Ag NPs recorded by a CCD camera equipped in the dark-field microscope. The dashed circles in each image refer to the same specific particle and the dashed squares in each image refer to the same specific nanoparticle pair. (Adapted from reference Y. Song, China Patent, Appl. No. CN200910085973.9).

We have used it to investigate size- and shape-dependent LSPR spectra of single Ag NPs by the analysis of the experimental results with the theoretical calculation (i.e. DDA simulation)(Song , Zhang et al. 2011). Figure 9 gives the AFM images of one specific triangle-shaped Ag NPs characterized by multi-hierarchy arrayed micro windows. The AFM image of the triangular silver NP shows that it has the edge length of 375-420 nm (Figure 9A) and the out-of-plane height of about 16.1 nm (Figure 9B). This NP shows multi LSPR scattering colors (Figure 9C), as further evidenced by its multi-mode LSPR peaks at 562.3 nm, 659.9 nm and 759.6 nm (Figure 9D-b). The peak wavelengths, peak ratios, and line widths (FWHM) at 562.3 nm and 659.9 nm from experiment are in good agreement with DDA simulation for its LSPR scattering (Figure9D-c), as have been summarized together with other shaped nanoparticles fabricated by the modified NSL in reference 27. In general, the DDA simulation shows best agreement with the experimental spectra for NPs, hence their shapes can be accurately modeled. However, it can also be seen that for wavelengths longer than 650 nm for the investigated NPs, the experimental result has a lower intensity than the simulation(Song , Zhang et al. 2011). By analysis the instrument errors and the wavelength dependent CCD quantum efficiency, these deviations are deduced by the precision in the shape construction during DDA simulations. From these results, it was also found that when the shapes and 3D morphologies of the NPs became more complicated, the deviation between the DDA simulation and the experimental result increased (Song , Zhang et al. 2011). This is due to the geometrical deviation between the real NPs and the regular species used in the calculations. If these two instrumental factors and the geometrical deviation of NPs are considered, the corrected experimental results will match with the DDA simulation very well. This result also confirms that our experimental method (DFOMS), based on the far field detection, preserves the ability to detect the near-field LSPR signal.

is once again showing that NSL is powerful method in the fabrication of uniform triangular

**D** 

Fig. 8. One example for the identification of the specific nanoparticles and nanoarrays in different instruments via multi-hierarchy arrayed micro windows based on Ag triangle nanoparticles and nanoarrays fabricated by NSL in one nearly circle-shaped window. (A) The plan view of the hexagon-arrayed triangle Ag NPs in one circle micro-window scanned by atomic force microscope (AFM); (B) the 3D view of the hexagon-arrayed triangle Ag NPs marked in the large pink dash-square in (A); (C) the real scattering color of these hexagonarrayed triangle Ag NPs observed under a dark-field microscope; (D) the CCD images of the scattering light of these hexagon-arrayed triangle Ag NPs recorded by a CCD camera equipped in the dark-field microscope. The dashed circles in each image refer to the same specific particle and the dashed squares in each image refer to the same specific nanoparticle

pair. (Adapted from reference Y. Song, China Patent, Appl. No. CN200910085973.9).

We have used it to investigate size- and shape-dependent LSPR spectra of single Ag NPs by the analysis of the experimental results with the theoretical calculation (i.e. DDA simulation)(Song , Zhang et al. 2011). Figure 9 gives the AFM images of one specific triangle-shaped Ag NPs characterized by multi-hierarchy arrayed micro windows. The AFM image of the triangular silver NP shows that it has the edge length of 375-420 nm (Figure 9A) and the out-of-plane height of about 16.1 nm (Figure 9B). This NP shows multi LSPR scattering colors (Figure 9C), as further evidenced by its multi-mode LSPR peaks at 562.3 nm, 659.9 nm and 759.6 nm (Figure 9D-b). The peak wavelengths, peak ratios, and line widths (FWHM) at 562.3 nm and 659.9 nm from experiment are in good agreement with DDA simulation for its LSPR scattering (Figure9D-c), as have been summarized together with other shaped nanoparticles fabricated by the modified NSL in reference 27. In general, the DDA simulation shows best agreement with the experimental spectra for NPs, hence their shapes can be accurately modeled. However, it can also be seen that for wavelengths longer than 650 nm for the investigated NPs, the experimental result has a lower intensity than the simulation(Song , Zhang et al. 2011). By analysis the instrument errors and the wavelength dependent CCD quantum efficiency, these deviations are deduced by the precision in the shape construction during DDA simulations. From these results, it was also found that when the shapes and 3D morphologies of the NPs became more complicated, the deviation between the DDA simulation and the experimental result increased (Song , Zhang et al. 2011). This is due to the geometrical deviation between the real NPs and the regular species used in the calculations. If these two instrumental factors and the geometrical deviation of NPs are considered, the corrected experimental results will match with the DDA simulation very well. This result also confirms that our experimental method (DFOMS), based on the far field detection, preserves the ability to detect the near-field LSPR

nanoparticles and nanoarrays.

**1μm** 

signal.

Fig. 9. (A): The AFM image for one single triangular shaped Ag nanoparticle with edge length of 375-420 nm; (B) the height mapping of the triangle shape Ag nanoparticle along the direction of the arrow in Fig. 9(A), showing the out-of-plane height of this nanoparticle of 16.1 nm; (C) the real scattering color image of this triangular shaped Ag nanoparticle; (Da): the LSPR absorption spectrum of this nanoparticle by DDA; (D-b) the LSPR scattering by experiment; (D-c) the LSPR scattering by DDA; (D-d): the LSPR extinction by DDA. In order to identify the location and the orientation of these positions around the NPs, the AFM image and color image were netted by dashed lines with each square unit of 125 nm ×125 nm after their distances and orientations were corrected. (Adapted from references: Y. Song, China Patent, Appl. No. CN200910085973.9; Song Y.; et al., Nanoscale 2011, 3, 31-44, Figure 13, copyright (2011) from the Royal Society of Chemistry. Adapted with permission.).

This combined method based on the NSL and the multi-hierarchy arrayed micro windows also allows us to investigate the 3D morphology dependent tip-tip LSPR coupling of triangular nanoparticle pairs. The zoom-in AFM image for the detailed 3D morphology of one typical Ag nanoprism pair is shown in Figure 10A. The nanoprisms have almost the same edge size ~ 375 nm and maximum out-of-plane height ~ 17.1 nm shown in Figure 10B by the typical height map along the arrowed tip-tip direction in Figure 10A. The real scattered color for the nanoprism pair, taken from dark-field microscopy, is shown in Figure 10C. Both of the nanoprisms in the pair give red color with different brightness, which might be due to variation in their surface roughness, slightly difference in the underlying surrounding dielectrics, and the focusing distance during image recording. The middle area between the two nanoprisms clearly shows more reddish color than the optical centers of the two nanoprisms. The LSPR spectrum for the middle area of the two optical centers (representing the tip-tip-coupling) is recorded in Figure 10D using their CCD image (not shown here) for the location identification, together with that obtained by the discrete dipole approximation (DDA) calculation of the nanoprism pair. According to its 3D morphology of the nanoparticle pair, the two nanoprisms can be treated as regular triangular nanoprisms with the bottom edge length of 375 nm, the top edge length of 125 nm and out-of-plane height of 17.1 nm for conducting the DDA calculation of the nanoprism pair. The recorded LSPR spectrum (Figure 10D: a) at the middle optical center of the two nanoprisms shows three distinct peaks, one strongest peak at 605 nm, one shoulder at 536 nm, and one secondary strong peak at 754 nm. By comparing the experimental result for the tip-tip coupling of the nanoprism pair with the DDA calculation (Figure 10D: b), it can be deduced that the peak at 605 nm represents the in-plane quadrupole resonances originated from the two source nanoprisms and the peak at 536 nm is from the out-of-plane quadrapole resonances of the two source nanoprisms. Although the DDA simulation does not show one distinct peak at 754 nm, our experiment result suggest one strong peak at this wavelength, which is probably from the strong tip-tip coupling. In order to reveal whether the peak at 754 nm is mainly from the tip-tip coupling or not, the LSPR spectra from the optical centers

Controlled Fabrication of Noble Metal Nanomaterials

nanoarrays can be precisely accounted for in the DDA model.

low.

**fabrication** 

(Figure 11: a and b).

signal magnification.

via Nanosphere Lithography and Their Optical Properties 519

tip-tip distance is more than about 400 nm due to the coupling intensity becomes extremely

Generally, the combination of NSL and the multi-hierarchy arrayed micro windows fabricated by the routine UV-LIGA shows a powerful ability not only in the identification of nanoparticles and nanoarrays but also in the precise investigation of the fundamental theory related to the 3D morphology dependent LSPR and LSPR coupling. In our study, the detector is far-field while the DDA calculation is based on the near-field. Thereby, the results indicate that the near-field LSPR of single NPs and the coupling signals of nanoarrays can be detected by the far-field detector if the 3D morphologies of NPs or

**4. Microfluidic biosensing system based on NSL and microfluidic reactor** 

Recently, Song has developed a high-throughput single Ag NPs biosensing device by coupling a variety of functionalized Ag NPs fabricated by NSL into a series of microfluidic channels (Song 2009). The designed microfluidic biosensing system based on Ag single nanoparticles and nanoparticle arrays is illustrated in Figure 11. Samples were fabricated by the combination of NSL and the traditional UV-LIGA process for the microfluidic reactor fabrication (Song 2009; Song 2010; Song and Elsayed-Ali 2010). In this biosensing system, the corresponding microfluidic channels are fabricated on the designed patterns where series of single Ag NPs or arrays (Figure 11: a) have been fabricated by careful alignment. The glass cover is then connected with glass optical fiber binding on the top of the microfluidic channels after careful alignment with the desired single Ag NPs or nanoarrays. In order to alleviate the non-specific absorption in the microfluidic channels, the channels are modified by polyvinylacholol (PVA) or polyethylene glycol (PEG) solution. After that, the single Ag NPs will be surface modified by a mixture of at least two thiol compounds with one having carboxyl group or amine group as the conjugating compound (e.g. 11-mercaptoundecanoic acid: MUA), and another thiol compound without carboxyl group or amine group as spacer (e.g. 6 mercapto-1-hexanol: 6-MCH, 1-octanethiol: 1-OT). The modification reaction is shown in equation (1). The Ag NPs can then be functionalized with biomolecules, as reporter (e.g. IgG), by a conventional 1-ethyl-3-(3- dimethylaminopropyl)-carbodiimide (EDC) coupling process to form the f, as shown in equation (2) and (3) for the functionalization of Ag NPs

The number per Ag NPs can be controlled by the ratio of the conjugation compounds and spacers, which can be used to calculate the number of the responding biomolecules (e.g. Protein A) that can bind with the reporters, which can be directly sensed by the LSPR peak shift. As shown in Figure 11, the solution having a specific concentration of the corresponding detected biomolecules can be delivered into the microfluidic channels (Figure 11: g). The channel widths are designed from several hundreds micro meter to ten micrometers that will play a role like a dark-field condenser for incident white light. The scattering color changes and the LSPR spectrum variations of Ag NPs (a) caused by the binding of the detected biomolecules on the reporters, (b) will be collected in the opened windows, (d) and transported into the detector and analyzer, (f) by the glass fiber, (e) after

of the source nanoprisms, are recorded (not shown here), showing one strong peak at the same position. Generally, one can see that the peak positions and shape resonances for the two nanoprisms are almost the same, suggesting that the nanosphere lithography process is very powerful in the fabrication of the nanoprisms with almost identical 3D morphologies and surroundings. Both of the two triangle nanoprisms do not give the peak at 754 nm as strong as the pair, confirming that the additional peak at 754 nm indeed is from the tip-tip coupling. However, previous investigations did not show additional strong peak due to tiptip coupling (Su, Wei et al. 2003; Zhao, Kelly et al. 2003). The reason for this significant coupling between the nanopair may be caused by the unique size of our particles that is just lying in the range of half wavelength of visible light, which can cause a strong long-range electrodynamic interaction among light and the collective electrons on the particle surfaces.

Fig. 10. Tip-tip coupling of one typical pair of triangular Ag nanoprisms in the arrays with interspacing of 103 nm is characterized using (A) AFM; (B) the scan of out-of-plane height of the two nanoprisms along the arrow direction in (A); (C) the color image taken from darkfield microscopy; (D: a) the LSPR scattering spectrum at the central locations (3 pixels) of the source nanoprisms and (D:b) the LSPR scattering spectrum of the pair calculated by DDA.

In additon, our experimental observations show that nanoprism coupling does not affect the quadrupole mode in LSPR significantly, resulting in little shifts in the highest peak at 598- 605 nm (the in-plane quadrupole mode). However, one additional peak (i.e. 754 nm) as compared with the in-plane quadrupole mode can be observed. This peak resulted from LSPR coupling is in good agreement with the prediction by the semianalytical model by Schatz et al. (Zhao, Kelly et al. 2003) In the present study, the edge lengths of the triangular nanoprisms are more than λ/2π (64-128 nm), which is more than the critical scale in the semianalytical model in the DDA.(Zhao, Kelly et al. 2003) Therefore, the long-range electrodynamic interaction, not electrostatic effects, will be dominant in the LSPR of the two nanoparsms. The center-to-center interspacing of the two nanoprisms is ~ 532 nm, more than the critical interspacing. As a consequence, the coupling will be mainly dertermined by the long-range radiative dipolar interactions (or radiative damping effects),(Zhao, Kelly et al. 2003) and phase retardance effects,(Su, Wei et al. 2003) resulting in one new peak with wavelength more than the highest peak for the two nanoprisms.

Based on this combined method, we have investigated the distance dependent tip-tip coupling between triangular Ag nanoprism pairs with dimensions at the range of half wavelength of visible light and distance ranging from 100 nm to 400 nm. It has been found that the coupling peak wavelength increases and the coupling intensity decreases with the increased tip-tip distance, and finally the coupling disappears (no coupling peak) when the

of the source nanoprisms, are recorded (not shown here), showing one strong peak at the same position. Generally, one can see that the peak positions and shape resonances for the two nanoprisms are almost the same, suggesting that the nanosphere lithography process is very powerful in the fabrication of the nanoprisms with almost identical 3D morphologies and surroundings. Both of the two triangle nanoprisms do not give the peak at 754 nm as strong as the pair, confirming that the additional peak at 754 nm indeed is from the tip-tip coupling. However, previous investigations did not show additional strong peak due to tiptip coupling (Su, Wei et al. 2003; Zhao, Kelly et al. 2003). The reason for this significant coupling between the nanopair may be caused by the unique size of our particles that is just lying in the range of half wavelength of visible light, which can cause a strong long-range electrodynamic interaction among light and the collective electrons on the particle surfaces.

Fig. 10. Tip-tip coupling of one typical pair of triangular Ag nanoprisms in the arrays with interspacing of 103 nm is characterized using (A) AFM; (B) the scan of out-of-plane height of the two nanoprisms along the arrow direction in (A); (C) the color image taken from darkfield microscopy; (D: a) the LSPR scattering spectrum at the central locations (3 pixels) of the source nanoprisms and (D:b) the LSPR scattering spectrum of the pair calculated by DDA. In additon, our experimental observations show that nanoprism coupling does not affect the quadrupole mode in LSPR significantly, resulting in little shifts in the highest peak at 598- 605 nm (the in-plane quadrupole mode). However, one additional peak (i.e. 754 nm) as compared with the in-plane quadrupole mode can be observed. This peak resulted from LSPR coupling is in good agreement with the prediction by the semianalytical model by Schatz et al. (Zhao, Kelly et al. 2003) In the present study, the edge lengths of the triangular nanoprisms are more than λ/2π (64-128 nm), which is more than the critical scale in the semianalytical model in the DDA.(Zhao, Kelly et al. 2003) Therefore, the long-range electrodynamic interaction, not electrostatic effects, will be dominant in the LSPR of the two nanoparsms. The center-to-center interspacing of the two nanoprisms is ~ 532 nm, more than the critical interspacing. As a consequence, the coupling will be mainly dertermined by the long-range radiative dipolar interactions (or radiative damping effects),(Zhao, Kelly et al. 2003) and phase retardance effects,(Su, Wei et al. 2003) resulting in one new peak with

**μm** 

**500 nm** 

**C D** 

Based on this combined method, we have investigated the distance dependent tip-tip coupling between triangular Ag nanoprism pairs with dimensions at the range of half wavelength of visible light and distance ranging from 100 nm to 400 nm. It has been found that the coupling peak wavelength increases and the coupling intensity decreases with the increased tip-tip distance, and finally the coupling disappears (no coupling peak) when the

wavelength more than the highest peak for the two nanoprisms.

**500 nm** 

**B nm** 

**A** 

tip-tip distance is more than about 400 nm due to the coupling intensity becomes extremely low.

Generally, the combination of NSL and the multi-hierarchy arrayed micro windows fabricated by the routine UV-LIGA shows a powerful ability not only in the identification of nanoparticles and nanoarrays but also in the precise investigation of the fundamental theory related to the 3D morphology dependent LSPR and LSPR coupling. In our study, the detector is far-field while the DDA calculation is based on the near-field. Thereby, the results indicate that the near-field LSPR of single NPs and the coupling signals of nanoarrays can be detected by the far-field detector if the 3D morphologies of NPs or nanoarrays can be precisely accounted for in the DDA model.
