1. Introduction

In recent years, localized surface plasmons (LSPs) in metallic nanostructures provide an attractive approach to harvest light and enhance the light-matter interaction, because of their remarkable optical properties of confining the optical field into the nanoscale areas beyond the diffraction limit [1,2]. Over the past few years, using their advantages, various applications such as second harmonic generation [3,4], optical sensing [5,6], surface enhanced Raman spectroscopy [7,8], and enhanced two-photon excited fluorescence [9–12] have been demonstrated. However, it has still been challenging to efficiently couple propagating light (PL) into the LSP modes of a single metallic nanostructure, because there is the limitation of the huge scale mismatch among photons, metallic nanostructures, and molecules. In order to realize the efficient excitation of LSPs, several approaches have been reported so far, such as prism coupling [13,14] and objective lens (free-space excitation) [15], but there is still low efficiency coupling of the PL to the LSP modes of the single metallic nanostructure, due to the order-of-magnitude difference between the diffraction-limited spot size of a focused light and a single metallic nanostructure. Thus, improving the PL coupling into LSP modes of individual metallic nanostructure is indispensable to harness the advantages of metallic nanostructures.

In view of the requirements to improve the light coupling to LSP modes of metallic nanostructures, alternative solutions using microcavity structures with higher quality (Q) factor have been recently proposed, such as planar photonic crystal combined with a gold nanorod [16], silver-coated microdisk [17], microring cavity integrated with a plasmonic nanocavity (theoretical) [18], and whispering gallery mode (WGM) microcavity with a metal nanoparticle (theoretical) [19]. In these solutions, the use of microcavity structures makes it possible to efficiently couple light to LSP modes of metallic nanostructures. However, it is difficult to experimentally adjust the coupling conditions between the incident photons and LSPs, because metallic nanostructures were directly placed on the cavity surface and their cavity parameters were manipulated only by the size and shape of the structures. Therefore, the adjustability of the cavity parameters such as photon-cavity and cavity-LSP coupling constants is essential for studying the underlying physics and flexibly attaining the high coupling efficiency.

In our previous work, we have succeeded to realize the efficient coupling of PL to LSP modes using a plasmonic-photonic hybrid system consisting of a gold-coated tip and a fiber-coupled microspherical cavity [20]. The fiber-coupled microspherical cavities own unique advantages of ultrahigh quality (Q) factors (>10⁹), small mode volumes (<10³ μm³), and controllability of coupling conditions [21], which can realize strong PL-LSP coupling owing to sufficiently high intra-cavity enhancement. This hybrid system has the peculiar trait of precise adjustability for the fiber-cavity coupling rate and the cavity-plasmon coupling rate, which is of great important for achieving the critical PL-LSP coupling condition. Adjusting coupling conditions by controlling the separation distances between a tapered fiber and a microspherical cavity and between a microspherical cavity and a gold-coated tip, the PL in an optical fiber can be focused into a gold-coated tip with the effective area of a 58-nm circle, beyond the diffraction limit, with a coupling efficiency of 93% [20]. Although we have experimentally demonstrated the highly efficient coupling of the incident light into the metal tip apex, we only proved the strong dissipation of the incident light, not the efficient LSP excitation, at a gold-coated tip apex via a fiber-coupled microspherical cavity.

In this paper, in order to experimentally prove efficient LSP excitation, two-photon excited fluorescence (TPF) from the pseudoisocyanine (PIC) dye molecules attached on a gold-coated tip via a fiber-coupled microspherical cavity under a weak continuous wave (CW) excitation was investigated. In the experiment, a plasmonic-photonic hybrid system with the coupling efficiency of about 64% to the LSP modes of a PIC-attached gold tip having an effective area of a 79-nm circle and high Q factor of about 4.6 × 10⁵, was initially established. Then, using this plasmonic-photonic hybrid system, the TPF from the PIC-attached gold tip by the use of a weak CW excitation was successfully realized. These results suggest that efficient LSP excitation at the metal tip apex via a fiber-coupled microspherical cavity was achieved.

2. Experimental setup

A fused-silica single-mode optical fiber (Thorlabs, 780HP) was used to fabricate a tapered fiber. The fiber was tapered by means of heating using a ceramic heater at the temperature of about 1400 °C and at the same time pulling at both ends of the fiber [22–24]. Waist diameter of the tapered fiber was about 500 nm in this experiment, which only supported the single-mode propagation. We monitored the transmittance in the tapered fibers at the wavelength of about 780 nm during the fabrication process. The transmittance of the tapered fibers in the experiment were over 0.90. By way of melting the edge of a tapered fiber tip with the irradiation of a carbon dioxide laser (CO₂ laser; wavelength: 10.64 μm), a microspherical cavity with a stem was formed. The microspherical cavity with a diameter of about 50 µm, which can be controlled by the laser intensity and irradiation time, was used in this experiment.

As a metal nanostructure, a commercial silicon atomic force microscope (AFM) probe tip (Olympus, OMCL-AC160TS-C3, curvature diameter 14 nm) covered by a gold thin film using a helicon sputter (ULVAC, MPS-4000C1/HC1) was used. The gold thickness was designed to be 50 nm and the curvature diameter of the gold-coated tip was smaller than 100 nm. Figure 1(a) shows the image of a gold-coated tip measured by a field emission-scanning electron microscope (JEOL, FE-SEM, JSM-6700FT). In order to confirm the LSP resonance property of the gold-coated tip, the scattering spectrum from the apex of a gold-coated tip was measured, when a white light was launched into a tapered fiber and the gold-coated tip was contacted to the surface of the tapered fiber. The surface plasmon resonance scattering spectrum of the gold-coated tip is shown in Fig. 1(b), which was obtained from the scattered intensity spectrum at the apex of the gold-coated tip collected by the objective lens divided by the transmitted intensity spectrum from the end of the tapered fiber without the gold-coated tip. The scattering spectrum of this metal tip exhibited a resonant peak at the wavelength of about 730 nm, and the full width at half maximum of the peak was about 200 nm. From the result, since the wavelength of the excitation light (~780 nm) was near to the LSP resonance peak wavelength, the efficient excitation of LSPs at the tip apex could be expected.

 

Fig. 1 (a) FE-SEM image of a gold-coated tip and (b) surface plasmon resonance scattering spectrum of a gold-coated tip.

Download Full Size | PDF

PIC molecules, as dye molecules, have a large two-photon absorption cross section at the excitation wavelength 780 nm (~200 GM) [25–27]. For the purpose of the PIC-attached gold tip preparation, a self-assembly technique [28] was used to attach PIC molecules on the surface of a gold-coated tip. Firstly, the gold-coated tip was dipped into 1 mM ethanol solution of 3-mercaptopropionic acid (MPA) for 20 min at room temperature; excess thiol MPA molecules were removed by rinsing the tip with an ethanol solution. The thiol function groups in MPA facilitate the attachment of the gold-coated tip surface; therefore the MPA was chemically bonded to the surface of the gold-coated tip. Secondly, in order to activate the reactive chemical group of the MPA, which are able to bind to PIC dye molecules, the gold-coated tip was dipped into 0.1 mM aqueous solution of AgNO₃ for 20 min at room temperature. Similarly, for removing excess AgNO₃, the tip was rinsed by pure water. Lastly, the gold-coated tip was dipped into 0.01 mM ethanol solution of PIC dye molecules for 1 h at room temperature. After rinsing the tip by pure water to remove unbound PIC dye molecules, the PIC-attached gold tip was obtained.

Figure 2(a) shows the experimental setup. A tunable external cavity diode laser with a linewidth of ~300 kHz (New Focus, TLB-6312, Wavelength ~780 nm) was launched into a tapered fiber as a probe and excitation light. The laser frequency was swept by applying a voltage from a function generator (FG: NF, DF1906) to measure the transmission spectra. The pair components of a half wave plate (HWP1) and a polarization beam splitter (PBS) controlled the power of the incident light coupled into the tapered fiber. In order to efficiently induce TPF from the PIC-attached gold tip, two quarter wave plates (QWP1, 2) and a half wave plate (HWP2) were used to adjust the polarization states of the excitation light to selectively excite transverse magnetic modes (TM) parallel to the tip axis. The power of the incident light was monitored using an optical power meter (Newport, 2935-C) connected to the output of a fused fiber coupler (95:5). To control the coupling conditions, the separation distances between the tapered fiber and microspherical cavity and also between the PIC-attached gold tip and microspherical cavity were controlled by three-dimensional piezo manipulators (PI-Polytec, P-620.ZCL, P-621.1CD, P-621.ZCD). The origin of the distance between the tapered fiber and the microspherical cavity and the origin of the distance between the microspherical cavity and the PIC-attached gold tip were determined as the position where the tapered fiber was contacted to the surface of the microspherical cavity and the PIC-attached gold tip was contacted to the surface of the microspherical cavity, respectively. The tapered fiber and the PIC-attached gold tip were positioned at the opposite sides of the equator of the microspherical cavity. The axis of the tip was perpendicular to both microsphere surface and the tapered fiber. In order to satisfy the condition, the tip was oriented by three-dimensional piezo manipulators, which was monitored using a highly sensitive charge-coupled device camera (CCD) equipped by a microscope. These components and samples were placed in a plastic box, and filled with dry air to keep at a stable condition and reduce the humidity that may cause adhesion force at the surface of microspherical cavity decreasing the Q factor [23,24]. A high sensitive photodiode (Thorlabs, DET36A) and a digital oscilloscope (Tektronix, TDS5034) were used to measure the transmitted light from the end of the tapered fiber. The TPF intensity emitted from the PIC-attached gold tip was measured by a microscopy system (Olympus, BX-51) set on the top of the samples. The emission signal was collected by an objective lens with 0.42 NA, and then passed through a dichroic mirror and short-pass filters to eliminate the excitation light, and was incident into the multimode fiber bundle connected to a spectrometer (JASCO Corporation; iDus; Andor).

 

Fig. 2 (a) Schematic of experimental setup. FG: Function generator; LD: Laser diode; HWP: Half wave plate; PBS: Polarization beam splitter; QWP: Quarter wave plate; FC: Fiber collimator; FFC: Fused fiber coupler (95:5); PD: Photodiode; DM: Dichroic mirror; Inset: SEM images of a microspherical cavity and a tapered fiber. (b) Schematic of optical coupling into a metal nanostructure using a fiber-coupled microspherical cavity.

Download Full Size | PDF

According to [20,21,29], the coupling conditions among a tapered fiber, a microspherical cavity, and a PIC-attached gold tip, are basically described by three parameters, the fiber-cavity coupling rate (κ_ex), the intrinsic cavity loss rate (κ₀), and the tip-cavity coupling rate (κ_tip), as shown in Fig. 2(b). In the steady state, the transmittance (T) of the fiber output under the resonant condition can be described by,

3. Results and discussion

For the experimental examination of the LSP property of a PIC-attached gold tip, we investigated the dependence of the TPF intensity from a PIC-attached gold tip on the polarization of the excitation light. Pulses from a femtosecond Ti:sapphire laser (Spectra Physics, wavelength 780 nm, duration time 100 fs, repetition rate 80 MHz) focused onto the tip were used for the TPF excitation. The TPF from the tip was collected by an objective lens, filtered by the combination of a dichroic mirror and short-pass filters (Opto-Line, FF01-720/SP25; 720-1100nm: OD>6, 350-690nm: Transmittance > 90%), and detected by a single photon counting modules (SPCM; EG&G, SPCM-AQR-14). Figure 3 shows the excitation polarization dependence of the TPF intensity (I_em) from the tip when the angle (θ) between the polarization of the excitation light and the tip axis was changed. When the polarization was paralleling to the tip axis (θ = 0 deg.), the TPF intensity approached maximum. On the contrary, when the polarization was perpendicular to the axis of the tip (θ = 90 deg.), the TPF intensity decreased to minimum. Note that the minimums shown in Fig. 3 are about 100 counts/100ms were due to the imperfection of linear polarization state of the excitation light and the mismatch at the high angles was due to the degradation of PIC dyes, because we used intense pulsed laser excitation for the observation of TPF in the experiment. The reports [9–12] demonstrated that the corresponding TPF intensity (I_em) was proportional to (I_sca)² ~cos⁴(θ), since the scattering intensity of the excitation light (I_sca) was proportional to cos²(θ) owing to the existence of surface plasmon from the metallic nanostructures. As expected, the TPF intensity in this experiment followed cos⁴(θ) dependence on the excitation polarization (solid line in Fig. 3). These results suggest that the TPF intensity was sensitive to the incident polarization light, and the incident polarized light paralleling to the tip axis could efficiently excite the TPF from the tip, as a result of the LSP excitation at the tip.

 

Fig. 3 Excitation polarization dependence of TPF from the apex of a PIC-attached gold tip under a femtosecond pulsed laser excitation (excitation peak power density: ~60 MW/cm²) in free space. The fitting solid curve shows a cos⁴(θ) function. Inset: TPF images of the tip with parallel and perpendicular excitation polarizations, respectively. Image size is 20 × 20 μm². In the images, circles and double-arrows indicate the tip apex region and the polarization of the excitation light.

Download Full Size | PDF

In order to obtain TPF from the PIC-attached gold tip via a fiber-coupled microspherical cavity using a weak CW excitation, building a system with high coupling efficiency between the incident light and the tip was necessary. The system must satisfy the critical PL-LSP coupling condition (κ_ex = κ_tip >> κ₀) as mentioned in section 2. First of all, the resonance property of the fiber-coupled microspherical cavity without the PIC-attached gold tip was analyzed and the intrinsic cavity loss rate κ₀ of a microspherical cavity was measured. Before the measurements, the polarization controller was adjusted to obtain a linearly polarized light at the tapered region of the fiber by same method proposed in [30]. The polarization parallel to the tip axis (TM modes of a microspherical cavity) has been confirmed to efficiently induce TPF from the PIC-attached gold tip. During the measurements, the transmittance spectra were observed by scanning the laser frequency.

The transmittance spectra were measured by changing the separation distance between the microspherical cavity and the tapered fiber (diameter ~500 nm) at 20-nm interval step. Figure 4(a) represents the transmittance spectra with five different separation distances between the microspherical cavity and the tapered fiber. In Figs. 4(b) and 4(c), the transmittance and linewidth of the resonant dip are plotted as a function of the separation distances. When the tapered fiber was moved to the surface of the microspherical cavity, and was far away from the surface of the microspherical cavity (>920 nm), no dip was observed and the transmittance was almost equal to 1. At the distance of 860 nm, a small resonant dip in the transmittance spectrum appeared, whose linewidth (∆υ₀) of the dip was about 19.8 MHz. The fiber-cavity coupling rate κ_ex was very small, because the tapered fiber was enough far away from the surface of the microspherical cavity. Therefore, the linewidth of the small dip was considered to be due to the intrinsic cavity loss rate (κ₀ ~2π × 19.8 × 10⁶ s⁻¹), corresponding to the Q factor of around 2.0 × 10⁷. When the distance decreased to 380 nm, the dip transmittance became zero and its linewidth was about 41.2 MHz (κ₀ + κ_ex ~2π × 41.2 × 10⁶ s⁻¹). As the linewidth became almost twice of κ₀ (κ_ex ~κ₀), the incident light was coupled to the microspherical cavity with the 100% coupling efficiency. Further decreasing the distance, the dip became wider and shallower. When the tapered fiber was in contact with the surface of the microspherical cavity by the use of a three-dimensional piezo manipulator, the transmittance was nearly flat but small dip still remained. At this time, the transmittance (T₀) and linewidth (∆υ₀ + ∆υ_ex) were measured to be about 0.67 and 424.5 MHz, and the fiber-cavity coupling rate κ_ex was estimated to be about 2π × 404.7 × 10⁶ s⁻¹. Therefore, the results suggest that different coupling regimes were clearly identified and the over coupling condition of κ_ex >> κ₀ was achieved, which was necessary for achieving the efficient light coupling into the gold-coated tip.

 

Fig. 4 (a) Transmittance spectra from the end of a tapered fiber at five different gap distances between the microspherical cavity and the tapered fiber. d in (a) shows the distance between the tapered fiber and microspherical cavity, and the zero point corresponds to the point where the tapered fiber was in contact with the surface of microspherical cavity. (b) Transmittance and (c) linewidth on resonance as a function of the distance between the tapered fiber and microspherical cavity.

Download Full Size | PDF

Then, when the PIC-attached gold tip from the opposite side of the tapered fiber was moved to the surface of the microspherical cavity, the critical PL-LSP coupling condition between the tip and the fiber-coupled microspherical cavity was realized. Simultaneously, the dependence of emission intensity from the tip on the distance (d) between the fiber-coupled microspherical cavity and the tip was measured. The tip was moved toward the microspherical cavity surface at 20-nm intervals, while the tapered fiber was in contact with the surface of the microspherical cavity. With the purpose of analyzing the coupling condition between the fiber-coupled microspherical cavity and the tip, the frequency of the incident light was continuously scanned and the transmittance spectra were measured during the measurements. In Figs. 5(a) and 5(b), the transmittance spectra from the end of the tapered fiber and emission spectra from the apex of the tip at five different distances between the tip and microspherical cavity are presented. When the tip was moved to the microspherical cavity surface, the depth and width of the resonant dip became deeper and broader and the emission intensity consequently increased. It is noted that due to the short-pass filters that eliminated the incident light, the emission spectral shapes at the longer wavelength regions (> 700 nm, in Fig. 5(b)) were strongly modified. In the experiment, we also performed the measurements using an Au-coated tip without PIC dye molecules. However, we could not observe any emission spectrum, and only after the PIC dye molecules were attached on the surface of the Au-coated tip, we could observe the emission. Thus, we concluded that the observed emission spectrum was from PIC dye molecules, not from the Au-coating.

 

Fig. 5 (a) The transmittance spectra from the end of the tapered fiber and (b) emission spectra from the apex of the PIC-attached gold tip with five different distances between the tip and the surface of microspherical cavity. d in (a) and (b) indicates the distance between the tip and microspherical cavity, and the zero point corresponds to the point in which the tip was in contact with the surface of microspherical cavity. (c) Transmittance, (d) linewidth of the resonant dip, and (e) emission intensity at the apex of the PIC-attached gold tip as a function of the distance between the PIC-attached gold tip and the microspherical cavity surface. The tapered fiber was in contacted with the surface of microspherical cavity. (f) Numerical result of the cross-section of WGM.

Download Full Size | PDF

From the results in Figs. 5(a) and 5(b), the transmittance, linewidth of the resonant dip, and emission intensity at the PIC-attached gold tip are plotted as a function of the distance between the tip and the microspherical cavity surface, as shown in Figs. 5(c)-5(e). When the distance between the tip and the surface of microspherical cavity was enough far away, no change from the spectrum shown in the bottom of Fig. 4(a) was observed. The dip transmittance (T₀) and linewidth (∆υ₀ + ∆υ_ex) were measured to be about 0.67 and 424.5 MHz from the Figs. 5(c) and 5(d), and no emission light form the apex of the tip was detected (Fig. 5(e)). When the tip was continued to move toward the microspherical cavity surface, the resonant dip became deeper and broader and the emission intensity from the tip increased. When the tip was contacted to the microsphere surface (d = 0 nm), the transmittance of the resonant dip became its minimum (T_min ~0.03), whose linewidth was almost twice (~840.7 MHz) of the initial value (~424.5 MHz), and the emission intensity reached the maximum. According to the obtained values of κ_ex and κ₀, we estimated the tip-cavity coupling rate κ_tip to be ~2π × 416.2 × 10⁶ s⁻¹. At this distance, the critical PL-LSP coupling condition of κ_ex ≈κ_tip >> κ₀ was achieved, since κ_tip (~2π × 416.2 × 10⁶ s⁻¹) almost equivalent to κ_ex (~2π × 404.7 × 10⁶ s⁻¹) was much larger than κ₀ (~2π × 19.8 × 10⁶ s⁻¹). Moreover, even though the PIC-attached gold tip was contacted to the microsphere surface, the total Q factor of this system was still as high as about 4.6 × 10⁵. Based on the obtained values and Eq. (1), the transmittance under the critical PL-LSP coupling condition was calculated to be 0.001, and this value nearly agreed with the experimental data (T_min). Note that a slight difference in the transmittance would result from the imperfections of polarization state of the incident light at the tapered region of the fiber and the tip-cavity position. Additionally, according to the κ_tip and Eq. (3), the effective extinction cross-section σ of the tip was estimated to be about 4.9 × 10³ nm², corresponding to the circular area with the 79-nm diameter, when the cross-section of WGM A and effective cavity length l_c were evaluated to be 2.5 µm² and 227.6 µm using the finite-element method (See Fig. 5(f)), respectively. Comparing the effective extinction cross-section of the tip σ and the cavity mode cross section A, the PL-LSP coupling efficiency would be roughly enhanced by a factor of ~5 × 10². Furthermore, according to the amount of the transmittance change (ΔT = T₀ - T_min) in Fig. 5(c), the incident light coupled to the PIC-attached gold tip with the effective area of the 79-nm circular via the fiber-coupled microspherical cavity was evaluated to be about 64%. Thus, we confirmed that the high coupling efficiency and strong light focusing on the apex of the tip could result in maximizing the two-photon excited emission intensity as shown in Fig. 5(e).

From the transmittance dependent on the distance (d) between the PIC-attached gold tip and the microspherical cavity surface as shown in Fig. 5(c), the transmittance follows T(d) ~T₀(1-exp(-αd)), where T₀ and α are the initial transmittance and coefficient of exponential function. This is because the dissipation by the PIC-attached gold tip was dependent on the spatial profile of the evanescent field around the microspherical cavity. When the obtained experimental data (T₀ ~0.67) was used and the transmittance was fitted by the exponential function (blue solid line in Fig. 5(c)), the coefficient α was estimated to be about 0.0075 nm⁻¹. The TPF intensity was examined as I_em ~(I_in)² ~(η)² ~exp(−2αd), since the intensity of the incident light (I_in) coupled into the tip was proportional to the coupling efficiency η = T₀ – T(d). After fitting the data (blue solid line of Fig. 5(e)), the coefficient of exponential function was about 0.015 nm⁻¹. It was found that the obtained value of about 0.015 nm⁻¹ was just twice of α ~0.0075 nm⁻¹, implying the quadratic dependence on the incident power. Note that the deviations from the fitting curves in Figs. 5(c) and 5(e) would be due to instability of the detection system and position fluctuation of the metal tip and the microsphere. Therefore, the efficient interaction between light and the PIC-attached gold tip owing to the synergetic effect of strong optical confinement effect of the fiber-coupled microspherical cavity as well as optical antenna effect of the gold-coated tip resulted in the enhancement of two-photon excitation process and the TPF from the PIC-attached gold tip could be observed.

In addition, after the efficient light coupling system was constructed, the TPF from the PIC-attached gold tip was also confirmed by measuring the incident power dependence of the TPF intensity. In order to keep the positions of the fiber-coupled microspherical cavity and tip, we monitored the transmission spectra from the end of the tapered fiber which were sensitive to the position change. Keeping the positions of the fiber-coupled microspherical cavity and tip, we gradually increased the power of the incident CW light, and the emission spectra from the tip were measured by a spectrometer. The dependence of the TPF intensity on the power of the incident light is shown in Fig. 6 and a slope value was evaluated to be about 2.2, indicating that the two-photon excitation process from the PIC dye molecules attached on the gold-coated tip was occurred. By taking into account the influence of the frequency scan, we estimated the effective power density at the tapered region in order to compare with the previous studies [9–12], in which TPF has typically been induced by a high-power pulsed laser excitation. Because only the incident light coupled into WGM modes would induce the TPF at the tip, the effective incident power was compensated by the ratio of the linewidth of resonant dip to the total frequency scanning range. The effective power density at the tapered region was estimated to be several kW/cm², which was much smaller than those excited by a high-power pulsed laser (typical peak power density: several MW/cm²). Especially, we note that Sanchez et al. [11] have demonstrated that the TPF from PIC dye molecules dispersed on a substrate was enhanced by a gold tip, where the incident pulsed laser light was focused on the tip. Although in their paper, the plasmonic enhancement at the tip apex also played an important role for improving the excitation and emission efficiencies for observing the TPF, the coupling of incident light into the gold tip would be low due to the scale mismatch of the focused spot and the apex of the tip, resulting in the use of a high intense pulsed laser excitation (peak power density: several MW/cm²). Meanwhile, we also previously reported the TPF from PIC dye molecules attached on the gold-coated tip via a thin tapered fiber under a CW excitation condition, and the effective excitation intensity for TPF decreased to be about several tens kW/cm² due to the optical antenna effect of the gold-coated tip and light focusing effect of the thin tapered fiber [22]. Comparing with these results, the excitation intensity of the TPF from the PIC-attached gold tip using our plasmonic-photonic hybrid system via the fiber-coupled microspherical cavity further decreased. This enhancement was attributed to the strong optical confinement effect of the fiber-coupled microspherical cavity as well as the optical antenna effect of the gold-coated tip. Thus, these results suggest that using the hybrid system via the fiber-coupled microspherical cavity, the highly efficient LSP excitation can be achieved, which lead to the strong light-matter interaction at the gold-coated tip.

 

Fig. 6 Log-log plot of the incident power dependence of the TPF intensity. The slope value is about 2.2 (blue solid line).

Download Full Size | PDF

4. Conclusions

Using the plasmonic-photonic hybrid system, in which a fiber-coupled microspherical cavity was employed to efficiently couple light into a PIC-attached gold tip, we achieved the 64% of light localized into a nanoscale domain of the tip apex with the effective area of a 79-nm circle and high Q-factor of about 4.6 × 10⁵. To experimentally prove efficient LSP excitation at the gold-coated tip, we successfully observed the TPF from a PIC-attached gold-coated tip under a weak CW excitation condition (effective power density: several kW/cm²) using this plasmonic-photonic hybrid system. These results suggested the possibility that, using our proposed plasmonic-photonic hybrid system, the synergetic effect of strong optical confinement effect of the fiber-coupled microspherical cavity and optical antenna effect of the gold-coated tip tremendously leads to the strong light-matter interaction at the PIC-attached gold tip.