Detailed Observation of Multiphoton Emission Enhancement from a

Aug 8, 2016 - The enhancement of multiphoton emission from a single colloidal nanocrystal quantum dot (NQD) interacting with a plasmonic nanostructure...
0 downloads 0 Views 3MB Size
Letter pubs.acs.org/NanoLett

Detailed Observation of Multiphoton Emission Enhancement from a Single Colloidal Quantum Dot Using a Silver-Coated AFM Tip Hiroki Takata,† Hiroyuki Naiki,‡ Li Wang,† Hideki Fujiwara,§ Keiji Sasaki,§ Naoto Tamai,† and Sadahiro Masuo*,‡ †

Department of Chemistry, Kwansei Gakuin University, 2-1 Gakuen, Sanda, Hyogo 669-1337, Japan Department of Applied Chemistry for Environment, Kwansei Gakuin University, 2-1 Gakuen, Sanda, Hyogo 669-1337, Japan § Research Institute for Electronic Science, Hokkaido University, Sapporo, 001-0020 Japan ‡

S Supporting Information *

ABSTRACT: The enhancement of multiphoton emission from a single colloidal nanocrystal quantum dot (NQD) interacting with a plasmonic nanostructure was investigated using a silver-coated atomic force microscopy tip (AgTip) as the plasmonic nanostructure. Using the AgTip, which exhibited a well-defined localized surface plasmon (LSP) resonance band, we controlled the spectral overlap and the distance between the single NQD and the AgTip. The emission behavior of the single NQD when approaching the AgTip at the nanometer scale was measured using off-resonance (405 nm) and resonance (465 nm) excitation of the LSP. We directly observed the conversion of the single-photon emission from a single NQD to multiphoton emission with reduction of the emission lifetime at both excitation wavelengths as the NQD-AgTip distance decreased, whereas a decrease and increase in the emission intensity were observed at 405 and 465 nm excitation, respectively. By combining theoretical analysis and the numerical simulation of the AgTip, we deduced that the enhancement of the multiphoton emission was caused by the quenching of the single-exciton state due to the energy transfer from the NQD to the AgTip and that the emission intensity was increased by enhancement of the excitation rate due to the electric field of the LSP on the AgTip. These results provide evidence that the photon statistics and the photon flux from the single NQD can be manipulated by the plasmonic nanostructure through control of the spectral overlap and the distance. KEYWORDS: Quantum dot, single photon, multiphoton, multiexciton, plasmon, tip enhancement

M

can be emitted by the cascade emission from the BX or TX to the ground state (GS) via a single exciton state (SX).31 The emitted multiple photons can behave as an entangled photon pair, which is important for quantum information technologies.32 However, when multiple excitons are produced in a single NQD, the excitons decay from the MX to the SX by nonradiative Auger recombination.33 Subsequently, the remaining single exciton decays from the SX to the GS by emitting a single photon. Therefore, a single NQD can exhibit singlephoton emission through Auger recombination.34−37 Singlephoton emission is also important for quantum information and communication technologies. Generally, the emission

odification of the emission photon statistics, i.e., singlephoton emission/multiphoton emission, from a single colloidal nanocrystal quantum dot (NQD) using plasmonic nanostructures has attracted considerable attention from the viewpoint of fundamental science as a new phenomenon induced by a plasmonic nanostructure and from the viewpoint of applications for quantum information technologies.1−17 NQDs are a unique class of tunable, dispersible fluorophores that have drawn intensive interest for their potential applications in a wide range of optoelectronic devices18−27 and biological detection systems.28−30 One of the particularly remarkable characteristics of the NQD is single-photon and multiphoton emission from a single NQD. When multiple excitons are produced in a single NQD, they form a multiple exciton state (MX), such as a triexciton state (TX) or biexciton state (BX), through the Coulomb interaction between excitons. In the emission process, multiple photons © 2016 American Chemical Society

Received: June 17, 2016 Revised: August 3, 2016 Published: August 8, 2016 5770

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters

silver-coated AFM cantilever (AgTip) as an MNS. Distance control with nanometer-scale precision can be achieved through the use of an AFM system with an AgTip, as demonstrated previously.41−43 In addition, using the AgTip, which exhibits a well-defined LSPR band, enabled control of both the distance and the spectral overlap. By observing the dependence of the emission behavior of single NQDs on the excitation wavelength and the NQD-AgTip distance, we elucidated in detail the mechanism of the control of NQD photon statistics by MNS. Figure 1 shows the scanning electron microscopy (SEM) images of an Si cantilever before sputtering (a) and after

photon statistics of the NQD are governed by the multiexciton dynamics based on the quantum confinement depending on the size, shape, and atomic composition of NQDs themselves. Researchers have recently demonstrated that the probability of multiphoton emission, i.e., BX emission, through interactions with plasmonic nanostructures can be increased.1−17 Plasmonic nanostructures, i.e., metallic nanostructures (MNSs), can enhance the excitation and the relaxation processes of nearby fluorophores. In the case of excitation enhancement, the excitation rate of the fluorophore can be enhanced by the electric field of the localized surface plasmon (LSP) generated on the MNS by incident light. For this enhancement, overlap of the absorption spectrum of the fluorophore with the localized surface plasmon resonance (LSPR) band is required. In the case of relaxation enhancement, both the radiative and the nonradiative rates of the fluorophore can be enhanced. For this enhancement, overlap of the photoluminescence (PL) spectrum of the fluorophore with the LSPR band is required. The LSP can be generated on the MNS by resonance energy transfer, and the LSP then decays radiatively or nonradiatively. The enhancement of the radiative and nonradiative rates, i.e., whether the LSP decays radiatively or nonradiatively, depends on the scattering properties of the MNS itself. The enhancement of the nonradiative process is referred to as quenching of the excitons. Increasing the probability of BX emission from a single NQD through interactions with MNSs requires consideration of the factors that enhance the relaxation process. Currently, two possibilities have been proposed. One mechanism involves enhancement of the BX emission rate by enhancing the radiative rate. With such an enhancement, the single NQDMNS can emit multiphotons before the excitons are annihilated by Auger recombination.1,3−6,8,11−15,17 The other mechanism is the quenching of SX by the MNS, i.e., a decrease in the PL quantum yield (ΦPL) of SX emission (ΦSX) rather than an actual increase in the ΦPL of the BX emission (ΦBX). When the excitons are quenched by MNS via energy or electron transfer,9 the quenching of SX is more efficient compared with the quenching of BX because of the longer lifetime of SX. As ΦSX decreases, the contribution of BX emission increases.9,38−40 Although these two mechanisms have been proposed, the process has not yet been fully elucidated. To elucidate the mechanism, a single NQD-MNS system in which both the spectral overlap and the distance are fully controlled would be used because the fluorophore−MNS interaction strongly depends on the spectral overlap and the distance between the fluorophore and MNS. However, no investigation using such systems has been reported. In this context, using atomic force microscopy (AFM) manipulation of a single Au nanocube (AuCube) with a well-defined shape and size to directly observe the change of the emission behavior of the single NQD, we previously reported that the photon statistics of the NQD were converted from single-photon emission to multiphoton emission because of the enhancement of the BX emission rate when approaching the AuCube.17 This previous work clearly demonstrates that enhancement of the BX emission rate could be induced by the spectral overlap between the PL of the NQD and the LSPR of MNS. However, in the case of AFM manipulation, control of the distance between the single NQD and the AuCube with nanometerscale precision was difficult. In this work, to elucidate the control of the MX dynamics and subsequent photon statistics using an MNS, we used a

Figure 1. Scanning electron microscopy (SEM) images of the Si cantilever before Ag sputtering (a) and after Ag sputtering (b). The scale bars in the images represent 100 nm. (c) A scattering spectrum of the AgTip (black line) and the absorption (green line) and PL (red line) spectra of the NQD dispersed in toluene. The vertical purple and blue lines indicate the 405 and 465 nm excitation wavelengths.

sputtering (b) with Ag. A comparison of these images confirms the coating of the cantilever by Ag. The tip radius of the AgTip, a, was estimated to be 20 ≤ a ≤ 30 nm from the SEM images. The scattering spectrum of the AgTip is shown in Figure 1c, along with the absorption and PL spectra of the NQD. The AgTip showed an LSPR band with a peak wavelength at 450 nm. The LSPR fully overlaps the absorption spectrum of the NQD and slightly overlaps the PL spectrum. In this work, 405 and 465 nm lasers were used as the excitation light source. On the basis of the spectral relationship, we expected that the excitation rate of the NQD could be enhanced by the electric field of the LSP generated on the AgTip at 465 nm excitation. The relaxation rate of the NQD could also be modified by the AgTip independent of the excitation wavelength because of the spectral overlap between the PL and LSPR. The PL and AFM images of the single NQDs dispersed on a coverslip are summarized in Figure 2. Figure 2a shows a typical PL image of the individual NQDs measured at 405 nm excitation without advancing the AgTip. The individual NQDs exhibit a double-lobed PL intensity pattern, which is characteristic of radially polarized beam excitation.41,42,44 From the cross-section of a single PL spot (Figure 2b), the double-lobed pattern was confirmed, and the full-width at half-maximum (fwhm) of the center lobe was estimated to be 220 nm. The PL and AFM images of the same area as that shown in image (a) 5771

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters

Figure 2e and f, respectively. In this case, the center lobe of the individual PL spots appeared in the PL image. From the crosssection of the single PL spot (Figure 2h), the fwhm of the center lobe was smaller (97 nm) than the cross-section (b) obtained at 405 nm excitation without advancing the AgTip. These results indicate that the PL was not quenched and that the spatial resolution of the PL image was increased41,42,45 by the approach of the AgTip at 465 nm excitation. The significance of these results is that the PL was quenched and unquenched by 405 and 465 nm excitations, respectively. Below, the details of the emission behavior of the individual NQDs are discussed. The representative emission behavior of a single NQD depending on the NQD-AgTip distance (z-distance) is summarized in Figure 3. The z-distance was defined as the distance from the top of the NQD to the top of the AgTip. These results were obtained from the same single NQD at 405 nm excitation. In the time traces of the PL intensity (Figure 3a−e), the PL intensity decreased with decreasing z-distance: 80 cts/ms (before approaching the AgTip; a); 60 cts/ms (z = 10 nm; b); 30 cts/ms (z = 6 nm; c); and 20 cts/ms (z = 2 nm; d). When the AgTip was retracted, the intensity returned to the original intensity before the approach of the AgTip (80 cts/ms; e). In the PL decay curves (Figure 3k−o), the curves show faster decay with decreasing z-distance. The lifetime (τ) and the normalized amplitude of the lifetime (α) were estimated as follows by fitting with two- or three-exponential functions: τ1 = 1.2 ns (30.0%) and τ2 = 25.0 ns (70.0%) before the approach of the AgTip; 1.0 ns (56.8%), 5.6 (40.6%), and 17.8 ns (2.6%) for z = 10 nm; 0.3 ns (99.2%) and 3.9 ns (0.8%) for z = 6 nm; and 0.3 ns (99.9%) and 5.1 (0.1%) for z = 2 nm. Because the shortest lifetime of 0.3 ns was the same as the instrument response function (IRF), the actual lifetime was probably much shorter than 0.3 ns. After the AgTip was retracted, the lifetime returned to almost the original values before the approach of the AgTip: 1.5 (28.3%) and 23.2 ns (71.7%). The decrease in the PL intensity with decreasing lifetime caused by the approach of the AgTip clearly indicates that the PL was quenched; i.e., the nonradiative decay rate of the NQD was enhanced by the AgTip as a result of the resonance energy transfer from the NQD to the AgTip (vide infra). In the photon correlation histogram (Figure 3f−j), the second-order correlation function g(2)(0), which is defined as the ratio of the number of detection events at a delay time of 0 ns to the average number of detection events at other delay times, provides information about the photon statistics in the PL emission; specifically, the probability of single-photon emission increases when g(2)(0) approaches zero. In addition, the g(2)(0) value corresponds to the efficiency of BX emission, ΦBX/ΦSX, at low excitation power, such as under our excitation conditions (the average NQD exciton occupancy ⟨N⟩ ≪ 1).9,14 The g(2)(0) value increased with the approach of the AgTip: 0.09 before the approach of the AgTip, 0.17 for z = 10 nm, and 0.85 for z = 6 nm. In the case of z = 2 nm, the PL intensity was too low for a photon correlation histogram to be constructed. After retracting the AgTip, the g(2)(0) value returned to the original value of 0.08. These results indicate that the emission photon statistics changed from single-photon emission to multiphoton emission with the approach of the AgTip. In the PL spectra (Figure 3p), no clear change of the spectra was observed. Because of the small red-shift of the BX emission spectrum compared with the SX emission spectrum,31,46 distinguishing the BX emission spectrum from the SX emission

Figure 2. (a) A PL image of individual NQDs obtained at 405 nm excitation without the approach of the AgTip. (b) A cross-section of the single PL spot indicated by the dotted line in image a. (c, d) PL and AFM images of the individual NQDs obtained from the same area as image (a) at 405 nm excitation with an approaching AgTip. (e, f) PL and AFM images of the individual NQDs obtained at 465 nm excitation with an approaching AgTip. (g, h) Cross sections of the single PL spots indicated by the dotted lines in images c and e, respectively. The size of all images is 5 μm × 5 μm. The scale-bar in image (a) represents 1 μm.

were then measured again by 405 nm excitation with advancement of the AgTip (Figure 2c,d). The center lobe of each PL spot disappeared with the approach of the AgTip (c). From the cross-section of a single PL spot (Figure 2g), the disappearance of the center lobe was clearly confirmed. This result indicates that the PL from the single NQDs was quenched; i.e., the nonradiative decay rate was enhanced by the approach of the AgTip at 405 nm excitation. The PL and AFM images obtained from another area at 465 nm excitation with the approach of the AgTip are shown in 5772

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters

Figure 3. Time traces of the PL intensity (a−e), photon correlation histograms (f−j), PL decay curves (k−o), and the PL spectra (p) detected from the same single NQD depending on the z-distance at 405 nm excitation: (a, f, k, and black line in p) before the approach of the AgTip; (b, g, l, and blue line in p) z = 10 nm; (c, h, m, and red line in p) z = 6 nm; (d, i, n, and green line in p) z = 2 nm; and (e, j, o, and orange line in p) after the AgTip was retracted.

spectrum is difficult at room temperature.1,17 Similar changes in the emission behavior were observed from other single NQDs reproducibly, as shown in the Supporting Information (SI). The observed emission behavior can be interpreted as follows. In the case of 405 nm excitation, no excitation enhancement occurs because the LSP is not generated on the AgTip. By contrast, enhancement of the relaxation process, mainly enhancement of the nonradiative rate (quenching), occurs through the spectral overlap between the PL and LSPR band. Because quenching was observed at z ≈ 10 nm, the mechanism of the quenching is likely resonance energy transfer from the NQD to the AgTip rather than electron transfer. Thus, we concluded that the efficient quenching of the SX of the NQD caused the increase in the probability of the BX emission, as reported previously.9,38−40 In the case of 465 nm excitation, enhancement of the excitation rate is expected. In addition, the enhancement of the relaxation process also occurs, as in the case of the 405 nm excitation, because the spectral overlap between the PL and the LSPR is unchanged. The representative emission behavior of the single NQD observed at 465 nm excitation with the approach of the AgTip is shown in Figure 4. Unlike the 405 nm

excitation, the PL intensity from the single NQD increased with the approach of the AgTip: 60 cts/ms before the approach of the AgTip (a), increasing to 140 cts/ms for z = 10 nm (b), 160 cts/ms for z = 8 nm (c), and 140 cts/ms for z = 6 nm (d). When the AgTip further approached the single NQD, i.e., at z = 2 nm, the intensity decreased to 40 cts/ms. The intensity then returned to the same value as before the approach (60 cts/ ms) after the AgTip was retracted. In the photon correlation histograms (Figures 4g−l), the g(2)(0) value increased with the approach of the AgTip, as in the case of the 405 nm excitation; the g(2)(0) value of 0.09 before the approach of the AgTip (g) increased to 0.29, 0.56, 0.88, and 0.92 for z = 10 nm (h), 8 nm (i), 6 nm (j), and 2 nm (k), respectively. Thus, the probability of BX emission increased with decreasing z-distance. After the AgTip was retracted, the g(2)(0) value returned to 0.09. The decay curves (Figures 4m−r) were also shortened with the approach of the AgTip: the 26.7 ns (100%) lifetime before advancing the AgTip was modified to 0.8 ns (17.9%), 2.3 ns (81.9%), and 21.2 ns (0.2%) for z = 10 nm, to 0.7 ns (65.9%), 1.3 ns (34.0%), and 11.4 ns (0.1%) for z = 8 nm, to 0.4 ns (99.5%) and 2.8 ns (0.5%) for z = 6 nm, and to 0.3 ns (99.4%) and 2.7 ns (0.6%) for z = 2 nm. After the AgTip was retracted, 5773

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters

Figure 4. Time traces of the PL intensity (a−f), photon correlation histograms (g−l), PL decay curves (m−r), and the PL spectra (s) detected from the same single NQD depending on the z-distance at 465 nm excitation: (a, g, m, and black line in s) before the approach of the AgTip; (b, h, n, and blue line in s) z = 10 nm; (c, i, o, and right blue line in s) z = 8 nm; (d, j, p, and red line in p) z = 6 nm; (e, k, q, and green line in s) z = 2 nm; and (f, l, r, and orange line in s) after the AgTip was retracted.

the lifetime increased to 2.8 ns (15.4%) and 17.3 ns (84.6%) and did not return to the original value before the AgTip was advanced (26.7 ns). This shortening of the lifetime was caused by damage to the NQD during the measurement. In the PL spectra (Figure 4s), no change in the spectrum was observed, as in the case of 405 nm excitation. Interestingly, the PL intensity increased with the low probability of BX emission at z = 10 nm, whereas in the cases of z = 8 and 6 nm, the probability of BX emission increased with increasing intensity. Similar changes in the emission behavior with the approach of the AgTip were observed from other single NQDs, as shown in the SI. These results demonstrate that control of single-photon and multiphoton emission with increased intensity is possible through nanometer-scale control of the z-distance. The emission behavior observed at 465 nm excitation can be interpreted as follows. With the enhancement of the excitation rate, the PL intensity increased because the LSP could be generated on the AgTip at 465 nm excitation. In addition, the

quenching of SX also occurred, which caused an increase in BX emission with decreased PL lifetime, as in the case of the 405 nm excitation. The details of the mechanism of the emission behavior are discussed below. Figure 5 shows the enhancements of the PL intensity and g(2)(0) value as a function of the z-distance built as average values of 10 single NQD measurements at 465 nm excitation. In Figure 5a, the PL intensity increased with decreasing zdistance, and the intensity reached the maximum value, i.e., enhancement of 2.3 times on average, near z ≈ 10 nm. The intensity then decreased as the z-distance decreased further. In Figure 5b, the g(2)(0) value increased with decreasing z-distance below 10 nm. The maximum value of the enhancement was 10.5 times, on average, at z ≈ 2 nm. We discuss the mechanism of the change in the emission behavior induced by the approach of the AgTip using the enhancement of the PL intensity as a function of the z-distance (Figure 5a). Previously, Novotny et al. investigated the change of the PL intensity from a single 5774

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters ΦPL =

γr /γ 0 r γr /γ 0 r + γq /γ 0 r + (1 − Φ0PL)/Φ0PL

(3)

where γr refers to the radiative rate and γq indicates the additional nonradiative rate (quenching) of the single NQD by the energy transfer rate from the single NQD to the AgTip. Parameter γr/γ0r represents the enhancement of the radiative rate. In this work, we assumed γr/γ0r = 1 to exclude the enhancement of the radiative rate. In eq 3, the ratio γq/γ0r can be expressed as γq γ 0r

organic molecule by the approach of a gold nanoparticle (AuNP) attached to the end of a pointed optical fiber, and they reported an increase and a decrease in the PL intensity depending on the distance between the single molecule and the AuNP.41,43 Using theoretical analysis, they clearly demonstrated that the change of the PL intensity as a function of the distance could be interpreted by the combination of the enhancement of the excitation rate by the LSP of the AuNP and the quenching of the PL by resonance energy transfer from the single molecule to the AuNP.43 To elucidate the mechanism, we applied this theoretical analysis to the enhancement of the PL intensity (Figure 5a). The enhancement of the PL intensity can be expressed as γ Φ I = 0exc PL 0 I γ exc Φ0PL (1) where γexc and ΦPL represent the excitation rate and the PL quantum yield, respectively, and the superscript “0” refers to the value without the approach of the AgTip. The enhancement of the PL intensity (Figure 5a) was fitted using eq 1. In eq 1, γ the enhancement of the excitation rate 0exc by the electric field exc

of the LSP can be expressed as γexc γ 0 exc

= 1+2

ε(ω1) − 1 a3 3 (a + z) ε(ω1) + 2

2 2 2 ε(ω2) − 1 1 (Px + Py + 2Pz ) 3 Im ε(ω2) + 1 k 2 3z 3 16 |P|2

(4)

where ε(ω2) and κ2 refer to the dielectric constant of Ag at the frequency ω2 of the PL wavelength (610 nm) and the wavevector at ω2, respectively. P = [Px, Py, Pz] is the dipole moment. In these equations, a and Φ0PL were treated as adjustable parameters, whereas other values were treated as fixed parameters.47 The red and blue lines in Figure 5a show the theoretical curves obtained by assuming Φ0PL = 30% and 40%, respectively, with a = 20 nm. The theoretical curves are clearly in good agreement with the experimental results with reasonable Φ0PL and a values. This result indicates that the observed change of the emission behavior can be interpreted by the combination of the excitation rate enhancement and the quenching by the resonance energy transfer from the NQD to the AgTip. Therefore, the increase in the probability of BX emission can be explained by the quenching of the SX due to energy transfer. Because the quenching rate increases with decreasing zdistance, the g(2)(0) value increased with a decrease in the zdistance (Figure 5b). This quenching occurred independent of the excitation wavelength. Thus, the increase in BX emission was observed for both 405 and 465 nm excitation wavelengths. In the case of 465 nm excitation, enhancement of the excitation rate also occurred. Hence, the PL intensity increased with decreasing z-distance, and as the distance further decreased, the PL intensity decreased because of the contribution of energy transfer. In the case of 405 nm excitation, the PL intensity decreased with decreasing z-distance because no enhancement of the excitation rate occurred. To confirm the validity of the aforementioned interpretation, i.e., enhancement of the excitation rate by LSP on the AgTip, we conducted numerical simulations of the AgTip. The primary difficulty in the simulations was the determination of the shape and effective length of the AgTip. Thus, the tip radius of the AgTip (a) was fixed as a = 20 and 30 nm, which was estimated from the SEM image of the AgTip, and the shape and the effective length were varied to reproduce the scattering spectrum of the AgTip (details in SI). We achieved good agreement between the simulated extinction spectrum and the observed scattering spectrum by assuming a cone-shaped AgTip with an effective length (h) = 60 nm and a bottom radius (ab) = 44 nm. Figure 6a shows the distribution of the electric field enhancement of the cone-shaped AgTip obtained by the simulation for illumination at a wavelength similar to that of the excitation laser (465 nm) with z-direction of the electric field. The simulated extinction spectra of the AgTip are shown in Figure 6b, along with the scattering spectrum of the AgTip. The red and blue lines are the extinction spectra simulated under the assumptions of a = 20 and 30 nm, respectively. The peak wavelengths of the extinction spectra

Figure 5. Enhancements of the PL intensity, I/I0 (a), and g(2)(0) value, g(2)(0)/g(2)(0)0 (b), as a function of the z-distance. The superscript 0 indicates the values before the approach of the AgTip. These figures were built from the average results of 10 single NQD measurements at 465 nm excitation. The red and blue lines in panel a show the theoretical curves obtained by assuming Φ0PL = 30% and 40%, respectively, with a = 20 nm.

γ

=

2

(2)

where a, z, and ε(ω1) refer to the tip radius of the AgTip, the distance between the single NQD and the AgTip, and the dielectric constant of Ag at the frequency ω1 of the excitation wavelength (465 nm), respectively. In eq 1, the ΦPL of the single NQD interacting with the AgTip can be expressed as 5775

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters

multiphoton emission increased with decreasing z-distance, similar to the behavior observed at 405 nm excitation. By comparing the experimental data with the results of theoretical calculations, we elucidated that the increase in the probability of the multiphoton emission was induced by the quenching of the SX by the resonance energy transfer from the NQD to the AgTip and that the enhancement of the PL intensity was induced by the enhancement of the excitation rate by the LSP on the AgTip. This mechanism differs from the previously reported enhancement of the BX emission rate by a plasmonic nanostructure.1,3−6,8,11−15,17 Although the increase in the probability of the BX emission due to the quenching of SX has been previously reported, the PL intensity was decreased in the case of this enhancement.9,38−40 However, in our case, enhanced multiphoton emission with enhanced PL intensity was observed by the combination of quenching of the SX and enhancement of the excitation rate. Furthermore, the photon statistics and the intensity were modified through control of the distance. These results clearly demonstrated that the emission photon statistics and the emission intensity could be modified by the distance, the spectral overlap, and the excitation wavelength, i.e., by combining the enhancement of the excitation rate, the radiative rate, and the nonradiative rates using the plasmonic nanostructure. An increase in the probability of multiphoton emission can be achieved via enhancement of the radiative and nonradiative rates, the intensity can be increased via enhancement of the excitation rate and the radiative rate, and the enhancement factor of these enhancements can be controlled by the distance. Methods. Commercially available colloidal CdSe/ZnS core/ shell NQDs (average core radius: 2.6 nm; maximum PL wavelength: 610 nm) were purchased from Invitrogen. The AgTip was prepared by sputtering Ag onto a silicon AFM cantilever with a 7 nm tip radius (Olympus, OMCL-AC160TSR3). The scattering spectrum of the AgTip was measured by contacting the AgTip with a taper fiber coupled with a white light source,48 and numerical simulations of the AgTip were conducted using finite element analysis. Details are provided in the SI. The sample was prepared by spin-coating a toluene solution of colloidal CdSe/ZnS NQDs onto a clean glass coverslip. The PL emission behavior of the single NQD as a function of the NQD-AgTip distance (z-distance) was measured using an inverted confocal microscope (Olympus, IX-71) combining with an AFM system (JPK Instruments, NanoWizard II).17 In addition to the three closed-loop piezo-driven axes of the AFM head, a two-axis, closed-loop piezo-driven sample stage was used. To produce the z-polarized excitation beam, the linear polarization of pulsed laser beams (405 and 465 nm, 10.0 MHz, 90 ps fwhm, PicoQuant) was converted to radial polarization by a converter (ARCoptix); the beam was then focused to a diffraction-limited spot on the sample using an objective lens (Olympus, NA 1.4). The excitation intensity was adjusted to produce 0.17 excitons/pulse in a single NQD by taking into account the excitation intensity and the absorption crosssection of the NQD. The photons emitted from the NQD were collected by the same objective lens and passed through a confocal pinhole and long-pass and a short-pass filters to remove the excitation laser and AFM laser. Subsequently, half of the photons were detected by a spectrograph (Acton Research Corporation, SpectraPro2358) equipped with a cooled CCD camera (Princeton Instruments, PIXIS400B). The remaining half of the photons were passed through a band-

Figure 6. (a) Distribution of the electric field enhancement simulated by assuming a cone-shaped AgTip with a = 30 nm, ab = 44 nm, and h = 60 nm. (b) The extinction spectra simulated under the assumptions a = 20 (red line) and 30 nm (green line) with the same ab and h as panel a. The spectrum indicated by a black line shows the scattering spectrum of the AgTip. (c) The electric field enhancement as a function of the z-distance obtained from panel a. In this case, zdistance indicates the distance from the top of the AgTip.

were approximately the same, whereas the spectral widths were strongly dependent on a. The simulated spectrum corresponding to a = 30 nm was in better agreement with the experimental scattering spectrum. Figure 6c shows the electric field enhancement as a function of the z-distance obtained from Figure 6a. From this result, we observed that the field enhancement increased below 50 nm. This result is in good agreement with the z-dependent PL intensity enhancement (Figure 5a), which is evidence that the PL intensity was increased by the electric field of LSP on the AgTip. In this work, we observed the dependence of the emission behavior of single NQDs on the z-distance with the approach of an AgTip to elucidate the mechanism of the modification of the emission photon statistics. In the case of 405 nm excitation, the PL intensity decreased with decreasing lifetime and the probability of multiphoton emission increased with decreasing z-distance. In the case of 465 nm excitation, the PL intensity increased and then decreased with decreasing z-distance, whereas the lifetime was shortened and the probability of the 5776

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Nano Letters



pass filter and were detected by two avalanche single-photon counting modules (PerkinElmer, SPCM-AQR-14) for Hanbury-Brown and Twiss-type photon correlation and lifetime measurements. The time-resolution of the lifetime measurement IRF was approximately 0.3 ns. The details are provided in the SI. The measurement procedure was as follows. Initially, the AgTip was coupled to the center of a focused laser by the piezo of the AFM head. AFM topography and PL images of the sample were collected simultaneously by scanning the sample stage. By choosing a PL spot corresponding to the single NQD in the PL image, we measured the emission behavior before advancing the AgTip. Subsequently, the same single NQD was selected in the AFM image, and the AgTip was advanced toward the single NQD to measure the dependence of the emission behavior on the z-distance. The z-distance was defined as the distance from the top of the NQD to the top of the AgTip. The distance was controlled by a closed-loop feedback system. The real-time displacement of the z-distance was monitored by the software of the AFM system. The AFM topography measurement and the approach of the AgTip were performed in tapping and contact modes, respectively. All measurements were performed at room temperature under ambient conditions.



REFERENCES

(1) Masuo, S.; Naiki, H.; Machida, S.; Itaya, A. Appl. Phys. Lett. 2009, 95, 193106. (2) Yuan, C. T.; Yu, P.; Ko, H. C.; Huang, J.; Tang, J. ACS Nano 2009, 3, 3051−3056. (3) Mallek-Zouari, I.; Buil, S.; Quelin, X.; Mahler, B.; Dubertret, B.; Hermier, J. P. Appl. Phys. Lett. 2010, 97, 053109. (4) Vion, C.; Spinicelli, P.; Coolen, L.; Schwob, C.; Frigerio, J. M.; Hermier, J. P.; Maitre, A. E. Opt. Express 2010, 18, 7440−7455. (5) Canneson, D.; Mallek-Zouari, I.; Buil, S.; Quelin, X.; Javaux, C.; Mahler, B.; Dubertret, B.; Hermier, J. P. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 84, 245423. (6) Naiki, H.; Masuo, S.; Machida, S.; Itaya, A. J. Phys. Chem. C 2011, 115, 23299−23304. (7) Masuo, S.; Tanaka, T.; Machida, S.; Itaya, A. J. Photochem. Photobiol., A 2012, 237, 24−30. (8) Leblanc, S. J.; McClanahan, M. R.; Jones, M.; Moyer, P. J. Nano Lett. 2013, 13, 1662−1669. (9) Park, Y.-S.; Ghosh, Y.; Chen, Y.; Piryatinski, A.; Xu, P.; Mack, N. H.; Wang, H.-L.; Klimov, V. I.; Hollingsworth, J. A.; Htoon, H. Phys. Rev. Lett. 2013, 110, 117401. (10) Park, Y.-S.; Ghosh, Y.; Xu, P.; Mack, N. H.; Wang, H.-L.; Hollingsworth, J. A.; Htoon, H. J. Phys. Chem. Lett. 2013, 4, 1465− 1470. (11) Yuan, C. T.; Wang, Y. C.; Cheng, H. W.; Wang, H. S.; Kuo, M. Y.; Shih, M. H.; Tang, J. J. Phys. Chem. C 2013, 117, 12762−12768. (12) Dey, S.; Zhou, Y. D.; Tian, X. D.; Jenkins, J. A.; Chen, O.; Zou, S. L.; Zhao, J. Nanoscale 2015, 7, 6851−6858. (13) Wang, F.; Karan, N. S.; Nguyen, H. M.; Ghosh, Y.; Hollingsworth, J. A.; Htoon, H. Sci. Rep. 2015, 5, 14313. (14) Wang, F.; Karan, N. S.; Nguyen, H. M.; Ghosh, Y.; Sheehan, C. J.; Hollingsworth, J. A.; Htoon, H. Nanoscale 2015, 7, 9387−9393. (15) Wang, F.; Karan, N. S.; Nguyen, H. M.; Mangum, B. D.; Ghosh, Y.; Sheehan, C. J.; Hollingsworth, J. A.; Htoon, H. Small 2015, 11, 5028−5034. (16) Hoang, T. B.; Akselrod, G. M.; Mikkelsen, M. H. Nano Lett. 2016, 16, 270−275. (17) Masuo, S.; Kanetaka, K.; Sato, R.; Teranishi, T. ACS Photonics 2016, 3, 109−116. (18) Klein, D. L.; Roth, R.; Lim, A. K. L.; Alivisatos, A. P.; McEuen, P. L. Nature 1997, 389, 699−701. (19) Klimov, V. I.; Mikhailovsky, A. A.; Xu, S.; Malko, A.; Hollingsworth, J. A.; Leatherdale, C. A.; Eisler, H.; Bawendi, M. G. Science 2000, 290, 314−317. (20) Coe, S.; Woo, W. K.; Bawendi, M.; Bulovic, V. Nature 2002, 420, 800−803. (21) Nozik, A. J. Phys. E 2002, 14, 115−120. (22) Klimov, V. I. J. Phys. Chem. B 2006, 110, 16827−16845. (23) Robel, I.; Subramanian, V.; Kuno, M.; Kamat, P. V. J. Am. Chem. Soc. 2006, 128, 2385−2393. (24) Klimov, V. I.; Ivanov, S. A.; Nanda, J.; Achermann, M.; Bezel, I.; McGuire, J. A.; Piryatinski, A. Nature 2007, 447, 441−446. (25) Kamat, P. V. J. Phys. Chem. C 2008, 112, 18737−18753. (26) Qian, L.; Zheng, Y.; Xue, J. G.; Holloway, P. H. Nat. Photonics 2011, 5, 543−548. (27) Kramer, I. J.; Sargent, E. H. Chem. Rev. 2014, 114, 863−882. (28) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013−2016. (29) Chan, W. C.; Nie, S. Science 1998, 281, 2016−2018. (30) Alivisatos, P. Nat. Biotechnol. 2004, 22, 47−52. (31) Fisher, B.; Caruge, J. M.; Zehnder, D.; Bawendi, M. Phys. Rev. Lett. 2005, 94, 087403. (32) Benson, O.; Santori, C.; Pelton, M.; Yamamoto, Y. Phys. Rev. Lett. 2000, 84, 2513. (33) Klimov, V. V.; Mikhailovsky, A. A.; McBranch, D. W.; Leatherdale, C. A.; Bawendi, M. G. Science 2000, 287, 1011−1013. (34) Lounis, B.; Bechtel, H. A.; Gerion, D.; Alivisatos, P.; Moerner, W. E. Chem. Phys. Lett. 2000, 329, 399−404.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b02479. Instrument setup, scattering spectrum measurement of the AgTip, numerical simulation of the AgTip, and emission behavior of the single NQDs with an approach of the AgTip (PDF)



Letter

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Prof. Tadaaki Kaneko, Mr. Koji Ashida, Mr. Daichi Dojima, and Mr. Satoshi Ito at Kwansei Gakuin University for assistance with SEM observations. This work was partly supported by JSPS KAKENHI Grant Number JP26390023 and JP26107005 in Scientific Research on Innovation Areas “Photosynergetics”.



ABBREVIATIONS NQD, colloidal nanocrystal quantum dot; AgTip, silver-coated atomic force microscopy tip; AFM, atomic force microscopy; LSP, localized surface plasmon; LSPR, localized surface plasmon resonance; MX, multiexciton; TX, triexciton; BX, biexciton; SX, single exciton; GS, ground state; MNS, metallic nanostructure; PL, photoluminescence; SEM, scanning electron microscopy; AuCube, cubic gold nanoparticle; AuNP, gold nanoparticle; fwhm, full-width at half-maximum; IRF, instrument response function 5777

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778

Letter

Nano Letters (35) Michler, P.; Imamoglu, A.; Mason, M. D.; Carson, P. J.; Strouse, G. F.; Buratto, S. K. Nature 2000, 406, 968−970. (36) Messin, G.; Hermier, J. P.; Giacobino, E.; Desbiolles, P.; Dahan, M. Opt. Lett. 2001, 26, 1891−1893. (37) Brokmann, X.; Giacobino, E.; Dahan, M.; Hermier, J. P. Appl. Phys. Lett. 2004, 85, 712−714. (38) Cheng, H. W.; Yuan, C. T.; Wang, J. S.; Lin, T. N.; Shen, J. L.; Hung, Y. J.; Tang, J.; Tseng, F. G. J. Phys. Chem. C 2014, 118, 18126− 18132. (39) Gao, Y.; Roslyak, O.; Dervishi, E.; Karan, N. S.; Ghosh, Y.; Sheehan, C. J.; Wang, F.; Gupta, G.; Mohite, A.; Dattelbaum, A. M.; Doorn, S. K.; Hollingsworth, J. A.; Piryatinski, A.; Htoon, H. Adv. Opt. Mater. 2015, 3, 39−43. (40) Liu, J.; Kumar, P.; Hu, Y. W.; Cheng, G. J.; Irudayaraj, J. J. Phys. Chem. C 2015, 119, 6331−6336. (41) Anger, P.; Bharadwaj, P.; Novotny, L. Phys. Rev. Lett. 2006, 96, 113002. (42) Kuhn, S.; Hakanson, U.; Rogobete, L.; Sandoghdar, V. Phys. Rev. Lett. 2006, 97, 017402. (43) Bharadwaj, P.; Novotny, L. Opt. Express 2007, 15, 14266− 14274. (44) Novotny, L.; Beversluis, M. R.; Youngworth, K. S.; Brown, T. G. Phys. Rev. Lett. 2001, 86, 5251−5254. (45) Gerton, J. M.; Wade, L. A.; Lessard, G. A.; Ma, Z.; Quake, S. R. Phys. Rev. Lett. 2004, 93, 180801. (46) Achermann, M.; Hollingsworth, J. A.; Klimov, V. I. Phys. Rev. B: Condens. Matter Mater. Phys. 2003, 68, 245302. (47) Johnson, P. B.; Christy, R. W. Phys. Rev. B 1972, 6, 4370−4379. (48) Ren, F.; Takashima, H.; Tanaka, Y.; Fujiwara, H.; Sasaki, K. Opt. Express 2013, 21, 27759−27769.

5778

DOI: 10.1021/acs.nanolett.6b02479 Nano Lett. 2016, 16, 5770−5778