This content has been downloaded from IOPscience. Please scroll down to see the full text.
(http://iopscience.iop.org/0957-4484/26/10/105301)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 138.231.176.8
This content was downloaded on 03/06/2015 at 15:20
E-mail: nlai@lpqm.ens-cachan.fr
Abstract
We investigated the coupling of a single nanoparticle (NP) into a polymer-based photonic
structure (PS). The low one-photon absorption microscopy with a two-step technique allowed us
first to accurately determine the location of a NP and then to embed it as desired into an arbitrary
PS. The coupling of a gold NP and a polymer-based PS was experimentally investigated
showing a six-fold photon collection enhancement as compared to that of a NP in unpatterned
film. The simulation results based on finite-difference time-domain calculation method
confirmed this observation and showed a 2.86-fold enhancement in extraction efficiency thanks
to the NP/PS coupling.
Keywords: direct laser writing, ultralow one-photon absorption, photonic coupling, gold
nanoparticle
(Some figures may appear in colour only in the online journal)
Figure 1. (a) Sketch of the experimental setup of the LOPA-based direct laser writing technique. M: mirror; S: electric shutter; P: polarizer; λ/
2, λ/4: half-wave plate and quarter-wave plate, respectively; PZT: piezoelectric translation stage; DM: dichroic mirror; F: 580 nm long-pass
filter; Ph: pinhole (diameter = 100 μm); L1, L2: lenses; APD: avalanche photodiode. (b) Measured UV-visible absorption spectra of SU8
photoresist (top) and of 50 nm-diameter spherical gold nanoparticles in water (bottom). The vertical line denotes the excitation wavelength of
532 nm.
absorption mechanism [3], which requires an expensive eliminate the excitation light, and detected by an ava-
and complex femtosecond pulsed laser source. The LOPA lanche photodiode.
experimental set-up is sketched in figure 1(a). For PS
fabrication, we used a negative tone SU8 photoresist
(MicroChem Corp.). Gold NPs (50 nm—diameter, in water 3. Fabrication
solution, Sigma-Aldrich) were chosen, as example of
active NPs to be coupled into SU8 PSs. The sample con- The first step of the fabrication process consists in localizing
sisting in a thin monolayer of Au NPs sandwiched between the Au NPs embedded between two SU8 layers. We note that
two SU8 layers was prepared by spin-coating. First, a layer Au NPs emit fluorescence in the visible range when excited
of SU8 2000,5 was spin-coated on a cleaned cover glass. by UV or visible (plasmonic resonance) light (see figure A1
Second, 50 μL of a diluted and perfectly dispersed Au NPs in appendix). Fluorescence images of Au NPs were obtained
by raster scanning the sample through the focusing spot.
solution was spin-coated on the surface of the first layer.
According to the high absorption of Au NPs at the used
Then, the second layer of SU8 2000,5 was spin-coated on
wavelength (figure 1(b)), a very low excitation power (less
top of the Au NPs layer. Note that, after each step, the
than 0.1 mW, or less than 0.0022 W μm−2 at focusing region)
sample was soft-baked on a hot plate at 65 °C (3 min) and
was employed. This power is weak enough not to induce any
95 °C (5 min) to remove the residual solvents. A total film polymerization of the SU8 photoresist, then no structure is
thickness of around 1.0 μm and a smooth surface profile formed during the mapping process. First, a large area
were subsequently confirmed by a profilometer. The (100 × 100 (μm2)) of the sample was scanned and many
sample was mounted on a 3D piezoelectric actuator stage individual Au NPs were found (see figure A2 in appendix).
(PZT). A high numerical aperture oil-immersion objective We then pointed out the position of each NP by scanning a
(Fluar × 100/1.3 NA, Zeiss) placed beneath the glass small area of 2 × 2 (μm2) around the NP, in the (xy)-and (xz)-
coverslip was used to focus the excitation laser beam. A planes, respectively, as shown in figure 2(a). We achieved a
continuous wave laser operating at 532 nm (Oxxius) was lateral resolution of about 243 nm (figure 2(b)), and an axial
used for both mapping and fabrication processes. The resolution of 730 nm (figure 2(c)), which actually corresponds
fluorescence signal emitted by Au NPs was collected by to the diffraction limit of the used objective lens. The curve
the same objective, filtered by a 580 nm long-pass filter to plotted in figure 2(d) is an extract of data corresponding to a
2
Nanotechnology 26 (2015) 105301 M T Do et al
Figure 2. (a) Illustration of a sample area in which a gold NP is sandwiched in between SU8 films. Fluorescence scanning is achieved in the
(xy)-and (xz)-planes, as indicated by the red border rectangles. (b) and (c) fluorescence images obtained by scanning along the (xz)-and (xy)-
planes, respectively. (d) Extract of data corresponds to the dotted line shown in (c) showing the fluorescence intensity as a function of the NP
position and of the precision of its position determination.
scan along the dotted line passing by the center of the NP linking process, followed by the development step. The pat-
(figure 2(c)), showing the fluorescence intensity as a function terned sample was then placed again on the PZT stage, at the
of position. The sharp peak reveals a precision <20 nm for the same position as in the fabrication step, in order to perform
position determination. optical characterization and comparison. The morphology and
Once the position of a single NP is determined, the surface topography of each structure was subsequently
fabrication of PSs containing this NP will be realized. For the examined by optical microscopy and scanning electron
fabrication step, the excitation power was increased due to the microscopy (SEM). Figure 3(a) shows a SEM image of 10
ultralow absorption of SU8 photoresist at 532 nm excitation hexagonal structures, each of them containing a single Au NP
wavelength [7]. In this work, we have chosen a power of at its center. We expected to obtain a micropillar containing a
3.8mW (0.084 W μm−2 at focusing region) for all fabrica- single Au NP. Surprisingly, a microsphere was formed at the
tions. The target structure is made of micropillars arranged in position of the Au NP instead of a micropillar. In this image,
a hexagonal 2D photonic crystal. The NP was chosen to be all micropillars patterns have the same parameters (center-to-
located at the central pillar of the hexagonal structure. The center distance of 1.50 μm, diameter of each pillar of about
fabrication of each pillar was realized by scanning the 0.32 μm and the height of a pillar = 1.00 μm), the micro-
focusing spot (scanning speed = 1 μm s−1) along the vertical spheres at the center display different diameters varying from
direction (z-axis) and through the total SU8 film thickness 0.76 to 1.40 μm. Figures 3(b) and (c) show SEM images of a
(total moving distance of 2 μm). In this work, we did not aim hexagonal structure, in which a quasi-sphere with a diameter
at investigating the effect of the periodic microstructure, so of 1.12 μm and containing a single Au NP was obtained at the
the geometric parameters of the micropillars pattern were structure center.
chosen to keep the photonic bandgap (if any) of the micro- The formation of the microsphere can be explained by a
pillars hexagonal structure remote from the emission band of thermodynamic model shown in figure 3(d). During the
the Au NP embedded inside the structure. The micropillar continuous irradiation by the focused laser beam, Au NP
pattern was made here only to ensure the accurate localization strongly absorbs green light (see figure 1(b)) due to the
of NPs by identifying the pillar containing the single Au NP. strongly localized surface plasmon resonance. The absorp-
After the exposure step, the sample was post-baked on a hot tion energy decays mainly non-radiatively via electron–
plate at 65 °C (3 min) and 95 °C (5 min) to finalize the cross- phonon collisions followed by phonon–phonon relaxation.
3
Nanotechnology 26 (2015) 105301 M T Do et al
Figure 3. (a) SEM image of different fabricated structures, each of them containing a single gold NP at the center. (b) SEM image of an
individual structure corresponding to number 5 in (a). (c) Zoom in of the center of the structure. (d) Thermodynamic model used to explain
the formation of the spherical structure containing a gold NP. The two-ends arrows passing by the NP denote the heat diffusion from NP,
which originates the thermopolymerization effect around the NP.
As a result, the NP is continuously heated and radially dif- controllable fabrication processes. With this given experi-
fuses heat beyond its surface to the surrounding medium mental setup, there are two possible ways for controlling the
[9, 10]. Meanwhile, SU8 photoinitiators within the focal sphere size: either by excitation intensity or by exposure
spot absorb one-photon energies from the incident light to time. It is worth to mention that the formation could be out
generate a certain number of strong acids (Lewis acid). The of control if light intensity is too high, because the tem-
AuNP at that time plays the role as a ‘hot spot’: in the perature at the Au NP position can exceed either its defor-
vicinity of the Au NP, the temperature is noticeably mation point (such as melting point, fragmentation
increased, then accelerate the generation rate of acids. Since threshold) [10] or the boiling point of the host SU8 material.
heat is diffused according to a spherical symmetry, the In such a case, heat storage may result in a micro-explosion.
polymerization rate of SU8 is much higher in a sphere In this work, fabrication parameters were the same for all
centered on Au NP. As long as the temperature is high structures and weaker than those of damage threshold, but
enough (>75) °C, the SU8 polymer chains are continuously the microsphere diameter varied from 0.76 to 1.40 μm due to
formed [11]. As a result, a spherical structure of a poly- the variation of the Au NPs size. It is important to mention
merized photoresist was obtained during the fabrication that the microsphere is not the PS that we intended to fab-
process. The amount of heat generated by the plasmonic NP ricate. However, as it will be shown later, such a photonic
can be controlled by the particle size and shape, as well as microsphere could be a valuable PS for the enhancement of
illumination strength, wavelength and irradiation duration. NP optical properties. In general case, when working with
Hence the interplay of the AuNP with the excitation light NPs without thermal effects, such as diamond NP, nonlinear
allows the spatial and temporal heat management leading to NP, etc, this very simple LOPA-based two-step technique
4
Nanotechnology 26 (2015) 105301 M T Do et al
Figure 4. Fluorescence images of fabricated structures obtained in (xy)-plane (a), (c) and in (xz)-plane (b). (c) Image of the microsphere
containing a single Au NP. (d) Comparison of fluorescence signals in two cases, Au NP in a microsphere (blue curve) and in an unpatterned
SU8 photoresist film (red curve).
should allow to fabricate structures as desired, including dominant role, since the optical properties of SU8 do not
microspheres. change significantly between the two states. Actually, a
similar structure, called solid immersion lens [12], allowing to
enhance the fluorescence emission of a single emitter has
been also reported.
4. Characterization: fluorescence measurement
5
Nanotechnology 26 (2015) 105301 M T Do et al
Figure 5. Calculation of fluorescence extraction efficiency in two situations: a single Au NP in a photoresist microsphere (red curves) and a
single Au NP in unpatterned SU8 photoresist film (blue curves). Simulations are based on FDTD method, with (see figure 3(d)): d = 503 nm,
r = 553 nm, emitted wavelength λ = 650 nm. We also assumed that a single NP corresponds to an absorption/emission dipole oriented parallel
to the air/glass interface. (a) Extraction efficiency versus half θ-angle. (b) Polar diagrams of normalized electric field intensity in (xz)-plane
shows the spatial distribution of the emitted light (radiation pattern).
coupling system, but we focused only on the ‘extraction belongs to the collection cone of the microscope objective,
efficiency’, which is defined as the ratio of the collected and could be detected. On the contrary, the radiation pattern
power to the total power radiated by the electric dipole [15]. of the F configuration (blue) is oriented at a larger angle and
We note that the microscope objective is characterized by its the radiation intensity is three times weaker, leading to a loss
numerical aperture NA = n sin θ , where n is the refractive of photons propagating out of the collection cone of the
index of the immersion medium and θ is the half aperture microscope objective. Certainly, we cannot directly compare
angle. the experimental results with the numerical calculations
Figure 5(a) shows simulation results of the extraction because we have simplified the coupling by considering a Au
efficiency as a function of the θ-angle for both S (red curve) NP (50 nm) as a single electric dipole. A complete model and
and F (blue curve) configurations. It can be seen that, with full mathematical calculation may be necessary for future
NA = 1.3, while only ≃22% of the radiated power was complete investigation of such NP/PS coupling.
obtained for the case of NP in a SU8 film, ≃63% of the Beyond the plasmonic NPs used in this work, our tech-
radiated power can be collected by the NP/microsphere cou- nique can be applied for a wide range of single solid-state
pling, resulting in a 2.86-fold enhancement in the extraction NPs, such as diamond NP containing a single emitter, non-
efficiency. To explain these simulation results, we note that linear NPs, etc. The LOPA-based direct laser writing could
for small particles behaving like dipoles close to a dielectric allow to fabricate other 2D and 3D PSs, which are very
interface, the radiated power is principally emitted towards promising for controlling and manipulation of emission
the denser medium at the critical angle [16]. Because a SU8 photons.
film possesses a high refractive index n SU8 ≃ 1.6 with respect
to that of glass substrate (n glass = 1.518), the emission from
the AuNP suffered a total internal reflection (TIR) effect, so 6. Conclusions
that all the emitted light at angles larger than the critical angle
are completely reflected. In contrast, in the case of a Au NP In conclusion, we have demonstrated a simple and low-cost
embedded in a SU8 microsphere, the Au NP is bounded by a technique to precisely position and embed a single Au NP
small SU8 volume, surrounded by air, resulting in a low into a photonic microstructure. The coupled NP/photonic
effective refractive index, as compared to that of glass sub- microstructure was realized by a confocal system working in
strate. Therefore there is no limitation caused by the TIR a very low linear absorption regime, and with a double-step
effect, then the radiated light is transmitted into the glass process: ultralow excitation power for determination of NP
substrate and most of it is collected by the objective. This is position, and low excitation power for the fabrication of
confirmed by the radiation pattern, i.e., the electric field desired microstructures. The coupling of NP to the micro-
intensity distribution in the (xz)-plane, shown in figure 5(b). A sphere allowed a great enhancement, up to six-fold, of the
single dipole, oriented parallely to the dielectric interface, extraction efficiency of its fluorescence, as compared to the
emits light mainly in the direction close to the critical angle, case without coupling. Simulations based on FDTD method
represented by two lobes. We can see that in the case of a S confirmed the experimental observation, showing a 2.86-fold
configuration (red), a significant portion of the emitted light is enhancement. This LOPA-based direct laser writing with a
located in the vicinity of θc ≃ arcsin(1 n glass ) = 41.2°, which double-step process is very simple but powerful as compared
6
Nanotechnology 26 (2015) 105301 M T Do et al
Appendix
Acknowledgments
[1] Dousse A, Suffczynski J, Braive R, Miard A, Lemaitre A,
Sagnes I, Lanco L, Bloch J, Voisin P and Senellard P 2009
The authors acknowledge the financial support from the Appl. Phys. Lett. 94 121102
Labex NanoSaclay in the frame of ONE-FAB-3D and [2] Henderson M R et al 2011 Adv. Mater. 23 2806
7
Nanotechnology 26 (2015) 105301 M T Do et al
[3] Schell A W, Kaschke J, Fischer J, Henze R, Wolters J, [10] Hashimoto S, Werner D and Uwadab T 2012 J. Photochem.
Wegener M and Benson 2013 Sci. Rep. 3 1577 Photobiol. C: Photochem. Rev. 13 28
[4] Faraon A, Santori C, Huang Z, Fu K C, Acosta V M, [11] Mack C 2008 Fundamental Principles of Optical Lithography
Fattal D and Beausoleil R G 2013 New J. Phys. 15 025010 (New York: Wiley)
[5] Barth M, Schietinger S, Fischer S, Becker J, Nusse N, [12] Hadden J P, Harrison J P, Stanley-Clarke A C, Marseglia L,
Aichele T, Lochel B, Sonnichsen C and Benson O 2010 Ho Y-L D, Patton B R, OBrien J L and Rarity J G 2010
Nano Lett. 10 891 Appl. Phys. Lett. 97 241901
[6] Li Q, Do M T, Ledoux-Rak I and Lai N D 2013 Opt. Lett. [13] Kelly K L, Coronado E, Zhao L L and Schatz G C 2003
38 4640 J. Phys. Chem. B 107 668
[7] Do M T, Nguyen T T N, Li Q, Benisty H, Ledoux-Rak I and [14] Huang C, Bouhelier A, Colas des Francs G, Bruyant A,
Lai N D 2013 Opt. Express 21 20964 Guenot A, Finot E, Weeber J C and Dereux A 2008 Phys.
[8] Lourtioz J M, Benisty H, Berger V, Gerard J M, Maystre D and Rev. B 78 155407
Tchelnokov A 2008 Photonic Crystals Towards Nanoscale [15] Schell A W, Neumer T and Benson O 2014 Opt. Lett. 39 4639
Photonic Devices (Berlin: Springer) [16] Lukosz W and Kunz R E 1977 J. Opt. Soc. Am. 67 1607–15
[9] Merabia S, Shenogin S, Joly L, Keblinski P and Barrat J 2009 Lukosz W and Kunz R E 1977 J. Opt. Soc. Am. 67 1615–9
Proc. Natl Acad. Sci. 106 15113 Lukosz W and Kunz R E 1979 J. Opt. Soc. Am. 69 1495–503