**3.3. Effect of branch number, perforated metal layer and overall nanostructure arrangement**

the strong plasmon resonance of Ag, bimetallic AgAu configurations are endowed with a 4 times higher EM field enhancement. The effect of the nanoring on the MB dimer is shown in **Figure 7c** and **d**. The scattered light is reflected back towards the ring centre where the nanostructures are placed, hence increasing the EM field enhancement in the IPS region by a factor 1.65. **Figure 7e** shows the typical Raman spectrum of p-MA (10 μM concentration) on AgAu five-branched dimers in the ring structure with an IPS of 5 nm. The polarization of the incident light was fixed parallel to the IPS axis. Characteristic Raman bands of p-MA can be clearly observed in the acquired spectrum. The impact of metal layer composition on SERS signal enhancement is plotted for the band located at 1077 cm−1 (**Figure 7f**), using 20 nm IPS

**Figure 7.** E-field distribution of five-branched 3D PM nanostructure dimer with Au (a) and AgAu (b) metal layers, in the x–y plane. (c and d) E-field distribution of five-branched 3D PM nanostructure dimer in a ring structure with AgAu metal layers. Reprinted with permission from reference [8]. Copyright 2014 American Chemical Society. (e) Typical SERS spectrum of p-MA molecule (at 10 μM concentration) on five-branched 3D PM nanostructure dimer in the ring structure. The exciting laser source, power and accumulation time were set to 830 nm, 1.4 mW and 15 s, respectively. The incident light polarization was set parallel to IPS axis. (f) SERS signal intensity at 1077 cm−1 as a function of metal layer compositions.

22 Raman Spectroscopy

The 3D nanostructure configuration presented so far allows recycling, long-term stability and reutilization of the SERS substrate, thus reducing fabrication costs. **Figure 8a** shows the regeneration protocol for the present architecture. However, as clearly highlighted by the cleaning steps, the process is fully compatible with any kind of 3D plasmonic configuration. In 3D geometry, the nanostructure morphology is conserved by the underlying Si template, while the plasmon active layer is simply recovered by a mask-less wet etching process, followed by metal redeposition. The chemisorbed analyte molecules used for SERS measurements will be completely removed together with the pre-existent metal layer, thus allowing employment of different molecular species after recycling. In order to test the effectiveness of the regeneration process, we have recycled the AgAu nanostructures up to five times (**Figure 8b**) and investigated their SERS response. The corresponding SEM images are shown in the top panel.

MB dimer structures with 20 nm Au, 20/20 nm Ag/Au and 20 nm Ag metal layers.

So far, we have engineered the five-branched nanostructure dimers to improve the hot-spot intensity as a function of nanostructure height, IPS, incoming light polarization and metal composition. Nevertheless, at very low concentrations (down to the femtomolar scale), the number of molecules available per hot-spot is decreased, and thus it is necessary to endow the nanostructures with high hot-spot densities in order to improve their detection limit. In view of that, single plasmonic nanostructures with multiple branches (4–10) and sharp protrusions (tip radius < 10 nm) were investigated.

The schematic representation of a typical eight-branched bimetallic AgAu nanostructure with 3D PM geometry and 200 nm IPS is shown in **Figure 9a**. Nanostructures with 4–10 branches, 20 nm Ag and 20 nm Au metal layers, 140 nm width, 150 nm high and tip radius of 10 nm were used to investigate the effect of hot-spot density on SERS signal enhancement. Normalincidence SEM image of eight-branched MB nanostructure is shown in **Figure 9b**, where the inset depicts a 54° tilted view. Besides, normal-incidence SEM images of individual MB structures with 4–10 branches are reported in the bottom panel of **Figure 9b**. A typical reflection spectrum of eight-branched 3D PM nanostructures is shown in **Figure 9c**, where multiple plasmonic resonances are clearly observable. To evaluate the effect of branch number on the LSPRs and hot-spot generation, far-field optical spectra and near-field distribution were calculated using RSoft and CST-MW numerical approaches. Distribution of E-field enhancement (E/E<sup>0</sup> ) for 4–10 branched 2D nanostructures at their respective LSPRs is shown in **Figure 10a**–**g**. The corresponding data points are shown in **Figure 10h**, along with their experimental counterparts. The incident polarization is set parallel to the x-axis. In the case of 4 branched nanostructures, two

**Figure 9.** (a) Schematic of bimetallic 3D PM nanostructure for eight-branched nanostructure geometry. (b) Normal incidence SEM image of 3D MB nanostructures with eight-branches arranged in a quadratic array of 200 nm IPS. The inset represents the 54° tilted view of the MBNS with 150nm silicon pillar height. The bottom panel shows SEM images of MBNS with 4–10 branches, where each micrograph shares the same scale bar. (c) A typical reflection spectrum of eight-branched 3D PM nanostructure. (d) Electric field distribution of eight-branched 3D PM nanostructure at its characteristic LSPR position of 685 nm.

hot-spots are observed parallel to the incident light polarization direction. The hot-spots density increases according to the number of branches, inducing a clear blue shift in LSPR spectral position (**Figure 10h**). This behaviour can be explained using a simple dipole theory. An increasing hot-spot density, that goes along with the number of branches, results in a higher total restoring force which blue shifts the nanostructure LSPR. Tailoring of the branch morphology enables precise control over the generation and spatial distribution of the hot-spots on the single nanostructure, thus opening new perspectives in reproducible SERS signal detection from large biomolecules, where the analyte size is many times larger than the individual hot-spot volume.

The influence of 2D, 3D and 3D PM geometries on the optical response and SERS enhancement was experimentally validated, as reported in **Figures 11** and **12**, respectively. Numerically calculated reflectance spectra of the eight-branched nanostructures with 2D, 3D and 3D PM topologies (see schematics in **Figure 11b**–**d**) are shown in **Figure 11a**. The corresponding experimental spectra are summarized in **Figure 11e**. In both cases, the optical spectra are normalized with respect to the flat/unpatterned area of the same sample. The experimental spectra show good correlation with their numerical counterparts. The E-field distribution of the nanostructures at their LSPR position is shown in **Figure 11f**–**h**. In the 2D case, near-field distribution maps (**Figure 11f**) clearly show the E-field confinement at the metal and bulk silicon interface. Due to the overlap of the E-field profiles into the bulk silicon a low-intense E-field enhancement (around 20) is observed. In general, surface plasmon resonances are tightly confined in high refractive index materials (e.g. silicon), which results in low E-field enhancement, low extinction-cross section, large propagation losses, broadening and red shifting of resonances compared to low-index materials [47]. As highlighted in the previous sections, 3D nanostructures decouple the hot-spot confinement from the substrate and enhance its strength by reducing the effective refractive

index of the embedding medium. The reflectance spectrum of the 3D nanostructure (**Figure 11a**) shows a significant blue shift of LSPR (around 485 nm), with a resonance maximum at around 680 nm. **Figure 11g** clearly shows that the generated E-fields (with an enhancement of around 80) are decoupled from the substrate by means of a dielectric nanopedestal. These strong E-field

**Figure 11.** (a) Experimental reflectance spectra of eight-branched nanostructures with 2D, 3D and 3D PM geometries. (b–d) Schematic representations of the structure geometries simulated in (a). (e) Experimentally measured reflection spectra of the nanostructures shown in (a). (f–h). Near-field profiles of E-field distribution in x–z plane for the LSPRs of the nanostructures with 2D, 3D and 3D PM topologies. Reprinted with permission from Ref. [6]. Copyright 2017 John Wiley & Sons.

**Figure 10.** (a–g) E-field distribution of 2D nanostructures with varying branch number (4–10). (h) Corresponding LSPR positions for experimental and calculated data points. Reprinted with permission from Ref. [6]. Copyright 2017 John

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications

http://dx.doi.org/10.5772/intechopen.74066

25

Wiley & Sons.

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications http://dx.doi.org/10.5772/intechopen.74066 25

**Figure 10.** (a–g) E-field distribution of 2D nanostructures with varying branch number (4–10). (h) Corresponding LSPR positions for experimental and calculated data points. Reprinted with permission from Ref. [6]. Copyright 2017 John Wiley & Sons.

hot-spots are observed parallel to the incident light polarization direction. The hot-spots density increases according to the number of branches, inducing a clear blue shift in LSPR spectral position (**Figure 10h**). This behaviour can be explained using a simple dipole theory. An increasing hot-spot density, that goes along with the number of branches, results in a higher total restoring force which blue shifts the nanostructure LSPR. Tailoring of the branch morphology enables precise control over the generation and spatial distribution of the hot-spots on the single nanostructure, thus opening new perspectives in reproducible SERS signal detection from large biomolecules, where the analyte size is many times larger than the individual hot-spot volume.

Electric field distribution of eight-branched 3D PM nanostructure at its characteristic LSPR position of 685 nm.

24 Raman Spectroscopy

**Figure 9.** (a) Schematic of bimetallic 3D PM nanostructure for eight-branched nanostructure geometry. (b) Normal incidence SEM image of 3D MB nanostructures with eight-branches arranged in a quadratic array of 200 nm IPS. The inset represents the 54° tilted view of the MBNS with 150nm silicon pillar height. The bottom panel shows SEM images of MBNS with 4–10 branches, where each micrograph shares the same scale bar. (c) A typical reflection spectrum of eight-branched 3D PM nanostructure. (d)

The influence of 2D, 3D and 3D PM geometries on the optical response and SERS enhancement was experimentally validated, as reported in **Figures 11** and **12**, respectively. Numerically calculated reflectance spectra of the eight-branched nanostructures with 2D, 3D and 3D PM topologies (see schematics in **Figure 11b**–**d**) are shown in **Figure 11a**. The corresponding experimental spectra are summarized in **Figure 11e**. In both cases, the optical spectra are normalized with respect to the flat/unpatterned area of the same sample. The experimental spectra show good correlation with their numerical counterparts. The E-field distribution of the nanostructures at their LSPR position is shown in **Figure 11f**–**h**. In the 2D case, near-field distribution maps (**Figure 11f**) clearly show the E-field confinement at the metal and bulk silicon interface. Due to the overlap of the E-field profiles into the bulk silicon a low-intense E-field enhancement (around 20) is observed. In general, surface plasmon resonances are tightly confined in high refractive index materials (e.g. silicon), which results in low E-field enhancement, low extinction-cross section, large propagation losses, broadening and red shifting of resonances compared to low-index materials [47]. As highlighted in the previous sections, 3D nanostructures decouple the hot-spot confinement from the substrate and enhance its strength by reducing the effective refractive

**Figure 11.** (a) Experimental reflectance spectra of eight-branched nanostructures with 2D, 3D and 3D PM geometries. (b–d) Schematic representations of the structure geometries simulated in (a). (e) Experimentally measured reflection spectra of the nanostructures shown in (a). (f–h). Near-field profiles of E-field distribution in x–z plane for the LSPRs of the nanostructures with 2D, 3D and 3D PM topologies. Reprinted with permission from Ref. [6]. Copyright 2017 John Wiley & Sons.

index of the embedding medium. The reflectance spectrum of the 3D nanostructure (**Figure 11a**) shows a significant blue shift of LSPR (around 485 nm), with a resonance maximum at around 680 nm. **Figure 11g** clearly shows that the generated E-fields (with an enhancement of around 80) are decoupled from the substrate by means of a dielectric nanopedestal. These strong E-field regions can be easily accessible to the analytes in Raman measurements if compared to the planar configuration. In the case of 3D PM geometry, the numerical spectrum (**Figure 11a**) shows various modes corresponding to the LSPR of the nanostructure and the underlying nanohole, and surface plasmon polaritons (SPP) at metal/air and metal/substrate interfaces. The LSPR position of the 3D PM structure (650 nm) is slightly shifted (a blue shift of 30 nm) from the 3D geometry, and a high E-field enhancement of 120 is observed. A deeper analysis of the perforated metal contribution has been discussed elsewhere [6]. The spectral features corresponding to the LSPRs of the 3D PM nanostructure and star-shaped hole cavity are located at about 620 and 1350 nm, while the 543 and 1050 nm resonances can be associated to the SPPs of the PM layer.

In order to assess the impact of hot-spot density on SERS enhancement, Raman measurements were performed on 4–10 branched nanostructures with p-MA molecules chemisorbed at 1 μM concentration. Typical SERS spectra of 8 branched nanostructures with 2D, 3D and 3D PM geometries are shown above. The incident laser wavelength, acquisition time and power were set to 785 nm, 30 s and 1 mW, respectively, and the incident light polarization was fixed along the x-axis. Characteristic Raman modes of p-MA are clearly visible in the spectra acquired on 3D PM MB nanostructures. In the planar case, the peaks centered around 1077 and 1590 cm−1 are experimentally observable, but the other bands are buried into the background noise. When the nanostructures are decoupled from the substrate via a dielectric nanopedestal (3D case), a significant rise in SERS signal intensity is observed along with the presence of all characteristic p-MA Raman modes. Furthermore, an additional enhancement in SERS signal intensity is observed for the 3D PM nanostructure geometry due to the coupling of the MBNS with the reflected light coming from the perforated metal layer underneath. The 3D PM MB nanostructure shows an absolute SERS enhancement in the order of 1011 obtained from the evaluation of the peak intensity at 1077 cm−1, with reference to Raman spectra of p-MA molecules on a planar SERS active gold film. Details on the corresponding calculations can be found in [6].

In order to understand the wavelength-dependent hot-spot generation and SERS enhancement, 4–10 branched 3D PM nanostructures were excited with four different laser sources (**Figure 12b**). The SERS signal intensity of the 1077 cm−1 band is plotted with respect to the number of branches and excitation lasers (532, 633, 785 and 830 nm). Due to different spectral power densities of the excited lasers, the SERS intensities are normalized independently with respect to the highest peak intensity obtained in the series. A monotonic increment of the SERS signal is observed by raising the number of branches for both 785 and 830 nm laser excitations, which are off-resonance with respect to the nanostructure LSPRs (see, **Figure 11i**). In this scenario, the SERS signal intensity can be associated to the increment of the number of hot-spots with multiple branches. For 633 nm laser excitation, the eight-branched nanostructure shows highest SERS signal intensity owing to the overlap between the LSPR and the laser source. A similar trend has been observed for 532 nm excitation source.

surface. For 1 fM concentration, only the Raman bands at 1077 and 1590 cm−1 are clearly visible, while the other features lie in the background noise. Despite of this, analyte detection at 1 fM showed the ultra-sensing capability of the substrate towards single/few molecules detection. In order to assess the effect of branch number on the detection limit, SERS measurements were performed on 4–10 branched 3D PM nanostructures with p-MA chemisorbed from a 1 fM concentrated solution (**Figure 12d**). An increment in SERS signal intensity at 1077 cm−1 is observed

**Figure 12.** (a) SERS spectra of p-MA molecules chemisorbed at 1 μM concentration on eight-branched nanostructures with 2D, 3D and 3D PM geometries. (b) SERS signal intensity at 1077 cm−1 as a function of a number of branches and different exciting wavelengths (532, 633, 785 and 830 nm). (c) SERS spectra of p-MA molecules at 1 μM, 1 nM, 1 pM and 1 fM concentrations taken on eight-branched 3D PM nanostructures for 785 nm laser excitation. (d) SERS signal intensity at 1077 cm−1 versus number of branches at ultra-low concentration, 1 fM. Reprinted with permission from Ref. [6].

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications

http://dx.doi.org/10.5772/intechopen.74066

27

by raising the number of branches, owing to the larger number of generated hot-spots.

Copyright 2017 John Wiley & Sons.

Reproducible identification and detection of biological samples/chemicals at ultra-low concentrations remains a huge challenge due to lack of high hot-spot density substrates. At ultralow concentrations, single/few molecules adsorbed in the vicinity of the hot-spot sites provide the majority of the SERS signal intensity. In this situation, if the molecules are not absorbed in the proximity of the hot-spot, the molecular fingerprint of the analyte cannot be identified. Thus, development of plasmonic nanostructures with a high hot-spot density that enables reproducible detection at ultra-low concentrations is of paramount importance in the field of molecular sensing. Multi-branched nanostructure designs hold concrete promises in this

Ultra-sensitive detection of analyte molecules was probed on eight-branched 3D PM nanostructures with p-MA molecules at concentrations ranging from 1 μM to 1 fM (**Figure 12c**). The incident wavelength, power and acquisition time were set to 785 nm, 1 mW and 3 s, respectively. At 1 μM concentration, the SERS spectra show the characteristic Raman bands of p-MA with good SNR. A decrease in SERS signal intensity is observed upon reduction of the molecular concentration from 1 μM to 1 fM, due to a lower number of adsorbed molecules on the nanostructure

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications http://dx.doi.org/10.5772/intechopen.74066 27

regions can be easily accessible to the analytes in Raman measurements if compared to the planar configuration. In the case of 3D PM geometry, the numerical spectrum (**Figure 11a**) shows various modes corresponding to the LSPR of the nanostructure and the underlying nanohole, and surface plasmon polaritons (SPP) at metal/air and metal/substrate interfaces. The LSPR position of the 3D PM structure (650 nm) is slightly shifted (a blue shift of 30 nm) from the 3D geometry, and a high E-field enhancement of 120 is observed. A deeper analysis of the perforated metal contribution has been discussed elsewhere [6]. The spectral features corresponding to the LSPRs of the 3D PM nanostructure and star-shaped hole cavity are located at about 620 and 1350 nm,

In order to assess the impact of hot-spot density on SERS enhancement, Raman measurements were performed on 4–10 branched nanostructures with p-MA molecules chemisorbed at 1 μM concentration. Typical SERS spectra of 8 branched nanostructures with 2D, 3D and 3D PM geometries are shown above. The incident laser wavelength, acquisition time and power were set to 785 nm, 30 s and 1 mW, respectively, and the incident light polarization was fixed along the x-axis. Characteristic Raman modes of p-MA are clearly visible in the spectra acquired on 3D PM MB nanostructures. In the planar case, the peaks centered around 1077 and 1590 cm−1 are experimentally observable, but the other bands are buried into the background noise. When the nanostructures are decoupled from the substrate via a dielectric nanopedestal (3D case), a significant rise in SERS signal intensity is observed along with the presence of all characteristic p-MA Raman modes. Furthermore, an additional enhancement in SERS signal intensity is observed for the 3D PM nanostructure geometry due to the coupling of the MBNS with the reflected light coming from the perforated metal layer underneath. The 3D PM MB nanostructure shows an absolute SERS enhancement in the order of 1011 obtained from the evaluation of the peak intensity at 1077 cm−1, with reference to Raman spectra of p-MA molecules on a planar SERS active gold film. Details on the corresponding calculations can be found in [6]. In order to understand the wavelength-dependent hot-spot generation and SERS enhancement, 4–10 branched 3D PM nanostructures were excited with four different laser sources (**Figure 12b**). The SERS signal intensity of the 1077 cm−1 band is plotted with respect to the number of branches and excitation lasers (532, 633, 785 and 830 nm). Due to different spectral power densities of the excited lasers, the SERS intensities are normalized independently with respect to the highest peak intensity obtained in the series. A monotonic increment of the SERS signal is observed by raising the number of branches for both 785 and 830 nm laser excitations, which are off-resonance with respect to the nanostructure LSPRs (see, **Figure 11i**). In this scenario, the SERS signal intensity can be associated to the increment of the number of hot-spots with multiple branches. For 633 nm laser excitation, the eight-branched nanostructure shows highest SERS signal intensity owing to the overlap between the LSPR and the laser

while the 543 and 1050 nm resonances can be associated to the SPPs of the PM layer.

26 Raman Spectroscopy

source. A similar trend has been observed for 532 nm excitation source.

Ultra-sensitive detection of analyte molecules was probed on eight-branched 3D PM nanostructures with p-MA molecules at concentrations ranging from 1 μM to 1 fM (**Figure 12c**). The incident wavelength, power and acquisition time were set to 785 nm, 1 mW and 3 s, respectively. At 1 μM concentration, the SERS spectra show the characteristic Raman bands of p-MA with good SNR. A decrease in SERS signal intensity is observed upon reduction of the molecular concentration from 1 μM to 1 fM, due to a lower number of adsorbed molecules on the nanostructure

**Figure 12.** (a) SERS spectra of p-MA molecules chemisorbed at 1 μM concentration on eight-branched nanostructures with 2D, 3D and 3D PM geometries. (b) SERS signal intensity at 1077 cm−1 as a function of a number of branches and different exciting wavelengths (532, 633, 785 and 830 nm). (c) SERS spectra of p-MA molecules at 1 μM, 1 nM, 1 pM and 1 fM concentrations taken on eight-branched 3D PM nanostructures for 785 nm laser excitation. (d) SERS signal intensity at 1077 cm−1 versus number of branches at ultra-low concentration, 1 fM. Reprinted with permission from Ref. [6]. Copyright 2017 John Wiley & Sons.

surface. For 1 fM concentration, only the Raman bands at 1077 and 1590 cm−1 are clearly visible, while the other features lie in the background noise. Despite of this, analyte detection at 1 fM showed the ultra-sensing capability of the substrate towards single/few molecules detection. In order to assess the effect of branch number on the detection limit, SERS measurements were performed on 4–10 branched 3D PM nanostructures with p-MA chemisorbed from a 1 fM concentrated solution (**Figure 12d**). An increment in SERS signal intensity at 1077 cm−1 is observed by raising the number of branches, owing to the larger number of generated hot-spots.

Reproducible identification and detection of biological samples/chemicals at ultra-low concentrations remains a huge challenge due to lack of high hot-spot density substrates. At ultralow concentrations, single/few molecules adsorbed in the vicinity of the hot-spot sites provide the majority of the SERS signal intensity. In this situation, if the molecules are not absorbed in the proximity of the hot-spot, the molecular fingerprint of the analyte cannot be identified. Thus, development of plasmonic nanostructures with a high hot-spot density that enables reproducible detection at ultra-low concentrations is of paramount importance in the field of molecular sensing. Multi-branched nanostructure designs hold concrete promises in this direction. In view of that, eight-branched 3D PM nanostructures arranged in the form of single, dimer, 3 × 3 array of clusters and chain of nanostructures, as shown in **Figure 13**, were investigated by keeping the interparticle distance fixed at 200 nm.

Near-field distributions of 2D single, dimer, 3 × 3 periodic array and chain of eight-branched nanostructures at their characteristic LSPRs, are shown in the **Figure 14a**–**d**, respectively. In order to reduce the computational time, 2D structures were used to compare the hot-spot density and E-field enhancement. The corresponding LSPR positions (experimental and calculated) are shown in **Figure 15**. A red-shift in the plasmon resonance has been observed with different layouts, as a consequence of the interaction between adjacent nanostructures. It is clearly visible that the hot-spot density is increased with respect to the arrangement schemes and, the highest E-field enhancement is observed for nanostructures arranged in the form of a chain.

SERS spectra were acquired in order to address the effect of different geometrical configurations on the measured signal. **Figure 16a** shows a typical SERS spectrum of p-MA (at 1 μM concentration) on eight-branched 3D PM structures arranged in the form of a chain. The exciting laser, acquisition time and power were set to 785 nm, 10 s and 1 mW, respectively, while the impinging light polarization was fixed along the x-axis. **Figure 16b** shows the SERS signal intensity for the band at 1077 cm−1 with respect to the nanostructure arrangement. Dimer configuration presents higher field enhancement in comparison to the isolated geometry, as

**Figure 14.** E-field distribution of single, dimer, 3 × 3 periodic array and chain of MBNS (a–d) at their characteristic LSPRs.

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications

http://dx.doi.org/10.5772/intechopen.74066

29

**Figure 15.** LSPR positions (theoretical and experimental) of 2D eight-branched nanostructures arranged in the form of

**Figure 16.** (a) The SERS spectrum of p-MA at 1 μM concentration taken on eight-branched 3D PM nanostructures arranged in the form of a chain. (b) Normalized SERS signal intensity variation at 1077 cm−1 with respect to different

single, dimer, 3 × 3 periodic array of clusters and chain of nanostructures.

arrangements of the nanostructures at 1 μM p-MA concentration.

The incident light is polarized along the x-axis.

**Figure 13.** (a–d). Normal incidence SEM images of the eight-branched 3D PM MB nanostructures in the form of single, dimer, 3 × 3 periodic array of clusters and chain of nanostructures, respectively. The inset shows the magnified view of the nanostructures.

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications http://dx.doi.org/10.5772/intechopen.74066 29

direction. In view of that, eight-branched 3D PM nanostructures arranged in the form of single, dimer, 3 × 3 array of clusters and chain of nanostructures, as shown in **Figure 13**, were

Near-field distributions of 2D single, dimer, 3 × 3 periodic array and chain of eight-branched nanostructures at their characteristic LSPRs, are shown in the **Figure 14a**–**d**, respectively. In order to reduce the computational time, 2D structures were used to compare the hot-spot density and E-field enhancement. The corresponding LSPR positions (experimental and calculated) are shown in **Figure 15**. A red-shift in the plasmon resonance has been observed with different layouts, as a consequence of the interaction between adjacent nanostructures. It is clearly visible that the hot-spot density is increased with respect to the arrangement schemes and, the highest E-field enhancement is observed for nanostructures arranged in the form of a chain.

SERS spectra were acquired in order to address the effect of different geometrical configurations on the measured signal. **Figure 16a** shows a typical SERS spectrum of p-MA (at 1 μM concentration) on eight-branched 3D PM structures arranged in the form of a chain. The exciting laser, acquisition time and power were set to 785 nm, 10 s and 1 mW, respectively, while the impinging light polarization was fixed along the x-axis. **Figure 16b** shows the SERS signal intensity for the band at 1077 cm−1 with respect to the nanostructure arrangement. Dimer configuration presents higher field enhancement in comparison to the isolated geometry, as

**Figure 13.** (a–d). Normal incidence SEM images of the eight-branched 3D PM MB nanostructures in the form of single, dimer, 3 × 3 periodic array of clusters and chain of nanostructures, respectively. The inset shows the magnified view of

the nanostructures.

investigated by keeping the interparticle distance fixed at 200 nm.

28 Raman Spectroscopy

**Figure 14.** E-field distribution of single, dimer, 3 × 3 periodic array and chain of MBNS (a–d) at their characteristic LSPRs. The incident light is polarized along the x-axis.

**Figure 15.** LSPR positions (theoretical and experimental) of 2D eight-branched nanostructures arranged in the form of single, dimer, 3 × 3 periodic array of clusters and chain of nanostructures.

**Figure 16.** (a) The SERS spectrum of p-MA at 1 μM concentration taken on eight-branched 3D PM nanostructures arranged in the form of a chain. (b) Normalized SERS signal intensity variation at 1077 cm−1 with respect to different arrangements of the nanostructures at 1 μM p-MA concentration.

3D structures were investigated with five-branched nanostar dimers in the ring structures. The effect of the number of branches, varying from 4 to 10, on the hot-spot generation and SERS enhancement was evaluated by using individual nanostructures (separated by 200 nm IPS). Moreover, the arrangement of MB nanostructures in various configurations (single, dimer, 3 × 3 array of clusters and chain of nanostructures) was evaluated to improve the device detection limit towards the single-molecule regime. 3D multi-branched nanostructures exhibit enhancement factors in the order of 1011 with an extremely high sensing capability (down to 1fM concentration). In this view, engineering the aforementioned architectures for high hot-spot density paves the way towards commercial biosensing applications, which require single/few-molecule detection sensitivity, with scalable manufacturing methods and cost-effective approaches. The proposed devices do not require specific labeling of the investigated analytes and multiple testing can be evaluated on the same platform. A new class of biological experiments will be therefore feasible, including monitoring growth factors that are produced from cultured cells. Moreover, our plasmonic nanostructures can be employed for direct detection of proteins within biological samples and real-time monitoring of chemical reactions. When applied to biomedicine, the present results, combined with already available purification methods, suggest the possibility of improving the early detection of several diseases, including cancer, where the number of clinically significant molecules at the onset

Engineering 3D Multi-Branched Nanostructures for Ultra-Sensing Applications

http://dx.doi.org/10.5772/intechopen.74066

31

\*, Alexander S. Roberts3

, Sergey I. Bozhevolnyi3

, Remo Proietti Zaccaria4,5, Peter Kjær Kristensen<sup>2</sup>

, Kjeld Pedersen<sup>2</sup>

, Andrea Cerea<sup>4</sup>

and

,

,

of the pathology is very small and often generated by a single cell.

\*Address all correspondence to: mch@nano.aau.dk and andrea.toma@iit.it

1 Interdisciplinary Nanoscience Center (iNANO), Århus University, Aarhus C, Denmark

2 Department of Physics and Nanotechnology, Aalborg University, Aalborg Øst, Denmark

5 Cixi Institute of Biomedical Engineering, Ningbo Institute of Materials Technology and

[1] Cao Y, Zhang J, Yang Y, Huang Z, Long NV, Fu C. Engineering of SERS substrates based on noble metal nanomaterials for chemical and biomedical applications. Applied Spec-

3 Centre for Nano Optics, University of Southern Denmark, Odense M, Denmark

troscopy Reviews. 2015;**50**:499. DOI: 10.1080/05704928.2014.923901

Anisha Chirumamilla1,2, Manohar Chirumamilla<sup>2</sup>

4 Istituto Italiano di Tecnologia, Genova, Italy

Engineering, Chinese Academy of Sciences, Ningbo, China

, Francesco De Angelis<sup>4</sup>

, Duncan S. Sutherland<sup>1</sup>

**Author details**

Esben Skovsen<sup>2</sup>

Roman Krahne<sup>4</sup>

\*

Andrea Toma<sup>4</sup>

**References**

**Figure 17.** Normalized SERS signal intensity variation at 1077 cm−1 with respect to different arrangements of the eightbranched 3D PM nanostructures at 1 fM p-MA concentration.

a consequence of the strong hot-spot confined in the IPS region. Periodic 3 × 3 arrays of nanostructures show higher E-field enhancement compared to single nanostructures, and a lower E-field enhancement with respect to the dimer layout, which is in good agreement with the E-field enhancements observed in **Figure 14**. Nanostructures arranged in the form of chain show the highest E-field enhancement (a factor of 62) in comparison to the other geometries. **Figure 17** shows the SERS signal intensity at 1077 cm−1 as a function of the arrangements of nanostructures at 1 fM p-MA concentration. As in the previous situation, the highest SERS intensities were observed for chains, as a result of the higher hot-spot density.
