Resonant scattering and local field enhancements by 11-nm-thin gold nanostrip antennas due to constructive interference of counter propagating slow surface plasmon polaritons is investigated. We characterize nanostrips of widths between 50-530 nm using both reflection spectroscopy and nonlinear scanning optical microscopy, in which two-photon-excited photoluminescence (TPL) excited with a strongly focused laser beam at the wavelength 745 nm is detected. We use TPL images to map the local field enhancements from individual nanostrips at a resolution of 0.35µm and compare results with theoretical calculated reflection spectra, enhancement levels and field distributions across the strip.
©2008 Optical Society of America
One of the main research directions in nano-optics is the search for configurations that efficiently interconvert propagating (µm-sized) and strongly localized (nm-sized) optical fields resulting thereby in strongly enhanced local fields, which are indispensable for optical characterization, sensing and manipulation at nanoscale . Various strategies have been suggested and pursued in order to realize a strong and robust field enhancement effect. Resonant interactions in metal nanostructures involving both localized and propagating surface plasmons (SPs) have been intensively and extensively investigated using nanostructures of different shapes and configurations, ranging from individual pointed particles [2,3] to their pairs [4-6] and periodic  and random [8,9] ensembles. While very high intensity enhancements (≥103) have been calculated and deduced from the experiments on surface enhanced Raman scattering (SERS) or two-photon-excited photoluminescence (TPL) the requirements to curvature radii and interparticle gaps involved are rather stringent resulting in large variations in the resonant effects observed for nominally identical structures. Very recently, an alternative route exploiting retardation-based resonances involving (slow) surface plasmon polaritons (SPPs) supported by metal nanostrip antennas has been suggested in a pioneering publication . The nanostrip antennas produced via electron beam lithography (EBL) are highly reproducible and could serve as promising high density substrates for surface enhanced Raman scattering, since the resonances are related to constructive interference of counterpropagating short range (SR) SPPs inside each individual nanostripe.
TPL from metals was previously described [12,13], with spatially resolved TPL studies [14,15] and near-field imaging  being lately used for characterization of local field enhancement. Since TPL has a square dependence on intensity, it is very sensitive to focus adjustment and hence also exhibits large depth selectivity, which we here use for TPL scanning optical microscopy (SOM) of the local field enhancement (TPL resolution ~0.35 µm) from metal nanostrips antennas covered by a layer of refractive index matching oil. We map the local field enhancements using TPL-SOM from individual nanostrips and at the same time even from relatively small structure dimensions (thickness only ~11 nm and widths down to 55 nm) compared to our previous investigations of larger random and periodic ensembles of nanostructures, where in addition the gold thickness was never less than ~50 nm [7,10].
2. Experimental results
Several 100×100 µm2 arrays of 5-µm-long nanostrips of one fixed width w and positioned at 1µm separation along and 10 µm across the strips, were prepared on quartz substrates by EBL for several designed strip widths from 50 to 600 nm at 50 nm increments, gold thicknesses t ~11 nm, and each array being separated from the other by at least 500 µm (Fig. 1). Scanning electron microscopy (SEM) images from the sample show well defined and reproducible gold strip arrays [Fig. 1(b)], although the fabricated actual strip widths were approximately 10-12% less compared to the above-mentioned nominal design parameters and the thickness was t~11.5 nm (±1 nm).
We have previously used reflection spectroscopy to experimentally obtain a quick overview of resonances within fabricated arrays of nanostructures  and recently we have demonstrated successful mapping of scattering resonances from strip resonators . Our experimental setup for reflection spectroscopy (Fig. 2) is based on a conventional optical microscope equipped with a Mitutoyo infinity-corrected long working distance objective (×100, numerical aperture, N.A.=0.70) and a possibility to access the light reflected from the sample with a fiber spectrometer recording between 540 and 1150 nm. The illumination is a standard 50 W halogen bulb producing sufficient light up to ~850 nm, and a polarizer is inserted both before the sample and in front of the spectrometer entrance to fully control the polarization. The nanostrips are placed in the focal plane of the illumination where the wave front is flat and the beam waist is furthermore much larger than a single strip since the whole array of strips is illuminated. The sample is covered with oil matching the refractive index (n=1.452) of the quartz substrate and positioned on top of a 1-cm-high glass cube, ensuring that only the reflected part of the light focused onto the gold strips and a small defocused part reflected from the top oil-air interface are collected by the spectrometer. For each spectrum measured from the strips we immediately record individual references at positions ~250 µm away from the concerned gold strip array in order to minimize the influence of any fluctuations in radiation spectra from the lamp.
Using the experimental data to compute the ratio [(Rstrip-Rref)/Rref], where Rstrip is the reflection measured from the strip array and Rref is the corresponding reference, the experimental reflection spectra can be directly compared with scattering cross sections estimated theoretically for the same gold strip parameters via the Green function surface integral equation method for the magnetic field described previously  (Fig. 3). It should be noted that the experimental reflection spectra were obtained with p-polarized illumination (electric field across the strip) and that the spectra obtained with s-polarized illumination were essentially flat except for a general very small increase in the reflection (proportional to the gold covered area) compared to that of glass (as expected).
The theoretical reflection spectra feature peaks for 55- and 90-nm-wide strips and both of these peaks are well reproduced in the experiment. For 135- and 175-nm-wide strips the reflection peaks are outside the experimentally accessible wavelength range from 550–825 nm and in the experimental spectral for 135- and 175-nm-wide strips we only observe a small difference between minimum and maximum reflection. For the 135 and 175 nm widths the increase in reflection from the shortest to the longest wavelengths resemble the smooth increase in the theoretical spectra for these widths, whereas the curvature on top of these reflection spectra at ~725-750 nm would also be present for s-polarized illumination, due to the additional reflection from gold covered areas. Small offsets between the curves can be ascribed to variations in the lamp intensity between measurements of strip reflection and glass reference, which becomes important for thin and narrow structures with weaker reflections. Although the reflection spectra are offset in level with respect to each other the method still provides a quick characterization of resonances.
As the next step, we then used TPL microscopy to evaluate the actual intensity enhancement inside the gold nanostrips. Our experimental setup for TPL-SOM is essentially the same as used in previous experiments , but adjustments have now been significantly improved [Fig. 4(a)]. It consists of a scanning optical microscope in reflection geometry built on the base of a commercial microscope and a computer-controlled translation stage (steps down to 50 nm, accuracy ~4 nm). The linearly polarized light beam from a mode-locked pulsed (pulse duration ~200 fs, repetition rate ~80 MHz) Ti-Sapphire laser is used as a source of sample illumination at the fundamental harmonic (FH) wavelength (λ=745nm, δλ~10 nm). After passing an optical isolator (to avoid back-reflection), half-wave plate, polarizer, red colour filter and wavelength selective beam splitter, the laser beam is focused on the sample surface at normal incidence with a Mitutoyo infinity-corrected objective (×100, N.A.=0.70). The half-wave plate and polarizer allow accurate adjustment of the incident power, which was selected in the range 0.1–0.25 mW depending on the obtained TPL signals. TPL radiation generated in reflection and the reflected FH beam are collected simultaneously with the same objective, separated by the wavelength selective beam splitter, directed through the appropriate filters and detected with two photomultiplier tubes (PMTs), the tube for TPL photons being connected with a photon counter giving typically only ~20 dark counts per second (cps).
For these experiments we decided to keep the excitation wavelength fixed at 745 nm, since the resonance distributions were rather broad (Fig. 3) compared to the available Ti:Sapphire wavelength range and 745 nm seem to be the central resonance position for 90-nm-wide nanostrips. During all TPL-SOM investigations we keep the sample covered with index matching oil (n=1.452) to maintain a symmetric environment around the strips supporting the SR-SPPs resonances. We decreased the scanning steps across the strips down to 50 nm with TPL integration times at each point of 100 ms in order to improve signal accuracy and improve the TPL image quality. Even at the low incident powers, some absorbing samples can be quite sensitive to damage by heating from the strongly focused laser beam and the local field enhancement. Here, we illuminate approximately the same place for 1-2 seconds and to minimize damage we had to use relatively low incident powers (only 0.1–0.25 mW) and at the same time increase the TPL signal efficiency by removing polarizers on the TPL detection side. Furthermore, to improve the image resolution compared to our previous investigations , we have carefully adjusted the incident beam to fill the entire numerical aperture of the microscope.
FH images obtained from the nanostrips show rather low contrast [Fig. 4(b)] due to the diffuse reflection from the air/oil interface. On the other hand, the TPL images show only signals from the in-depth focus plane at the nanostrips, even though this is below the top oil surface [Fig. 4(c) and (d)]. This can be ascribed to the fact that the two-photon absorption has a square dependence on intensity and the TPL signals are collected in a nearly confocal configuration. In addition, it is a very important fact that we basically don’t observe any TPL signal from the index matching oil, which would have made these measurements impossible. The TPL images [Fig. 4(c) and (d)] show very high contrast with signals varying two orders of magnitude from nearly only dark counts ~20-30 cps at the glass substrate and up to ~2-3 kcps from the nanostrips.
We investigated the influence of both p- and s-polarized illumination as indicated by arrows on the images [Fig. 4(b)-(d)]. The FH image is here only shown for p-polarization and appears similar for s-polarization, whereas there is remarkable polarization dependence for the TPL images [Fig. 4 (c) and (d)]. For wide strips and p-polarized excitation the TPL signal splits into two peaks along the strip, whereas for s-polarization the maximum signal is at the ends of the strip. The TPL image for p-polarization can be explained by the fact that for sufficiently wide strips (here 530 nm) and strong focus (TPL resolution ~0.35 µm) the SRSPPs are mainly excited when the beam is at the edges of the nanostrip and for an ideally focused beam there is no SR-SPP excitation when the beam is in the center of the nanostrip. Even when the excitation wavelength is off-resonance with the strip width, we still expect increased TPL from the sides due to edge effects and breaking of symmetry, similar to what is observed for second harmonic generation . For s-polarization both these effects are practically absent and we only observe maximum signals at the ends of the strip. However, some TPL will be obtained everywhere on any gold layer due to the strong focusing and surface roughness . For the off-resonant combination of strip width (~530 nm) and excitation wavelength (745 nm) shown here (Fig. 4), the obtained maximum TPL signals (relative to the used excitation powers) were rather similar for p- and s-polarization, but for resonant cases the TPL was at least one order of magnitude larger for p-polarized excitation.
To further elucidate the TPL dependence on strip width, we have carefully obtained detailed TPL images from individual strips for each of the fabricated actual strip widths (~55, 90, 135, 175, 215, 255, 295, 340, 390, 435, 480, and 530 nm) at a fixed excitation wavelength of 745 nm [Fig. 5]. The evolution from one into two TPL maxima across the strip can be followed in the images [Fig. 5(a)-(l)] and from cross sections taken across the strips [Fig. 5(m)] the splitting into two peaks from around 390-nm-wide strips is clearly seen. Each cross-section is averaged over 5 adjacent lines and taken at positions away from strong TPL spots, which are most likely coursed by increased roughness of the very thin (11.5 nm) gold films.
Both FH and TPL resolutions have been significantly improved by filling the entire numerical aperture of the microscope objective (N.A.=0.7), and from the TPL image cross sections [Fig. 5(m)], the TPL resolution at full width half maximum is found to ~0.35 µm, whereas the FH resolution was evaluated to ~0.75 µm.
3. Numerical results
Theoretical distributions of the electric field across gold nanostrips on quartz substrates covered by refractive index matching oil can be calculated using the Green function surface integral equation method  for similar gold strip parameters (thickness t~11 nm, widths w~100, 250, 350, and 500 nm) and fixed wavelength (λ=745 nm) as in the experiment. We calculate the field enhancement with respect to a fixed plane wave incident field and use the same maximum scale to visualize the dependence on strip width (Fig. 6).
We obtain increased fields for the strip widths 100 and 350 nm, where standing wave patterns of SR-SPPs reflected at the strip edges are appearing inside the nanostrip . In addition, the maximum field enhancements are obtained just outside the nanostrip due to the boundary conditions for the normal component of the electric field across the strip edges . We can compare theoretically and experimentally obtained electric field enhancements by instead analyzing the TPL signal from the nanostrips. Experimentally, the level of TPL enhancement can be objectively evaluated by taking into account the area and incident power producing the TPL signal . We have previously estimated the degree of TPL enhancement by comparing TPL signals from gold dots to those from smooth gold films . Following this approach the intensity enhancement factor α can be found using the relation:
where TPL is the obtained TPL signal, <P> is the used average incident power, and A is the gold area within the FH focus spot (diameter ~0.75 µm) producing the enhancement. To compare the experimentally obtained enhancement values with the theory we average the theoretical field distributions by integrating the intensity enhancement from a cross section across the strip width and then dividing this integral by the strip width:
where w is the strip width, E(x,y=0) is the calculated enhanced electric field depending on x across the strip, and E 0(x,y=0) is the incident electric field across the strip.
For strip widths up to 300 nm there is good agreement between theoretically and experimentally obtained intensity enhancement values, where especially the peak around 100-nm-wide strips is very well reproduced [Fig. 7]. Comparing theory and experiment for larger strip widths above 300 nm, we observe only a smaller increase of the intensity enhancement for strip widths around 350 nm [Fig. 7]. This discrepancy could probably be ascribed to the size of the TPL spot relative to the strip width, since we obtain a resolution of ~0.35 µm in the TPL images and therefore only collect signals from a smaller part of the stripe at each scan position instead of integrating over the entire strip as in Eq. (2). We have tried to deal with this problem using a different objective (×50, N.A.=0.55), where the TPL spot size was around 0.60 µm, but the obtained intensity enhancements were approximately the same although more uncertain due to the lower TPL efficiency and growing risk of damaging the sample through heating by increased incident laser powers.
In conclusion, we have investigated resonant scattering and local field enhancements by 11-nm-thin gold nanostrip antennas of widths between 55-530 nm using both reflection spectroscopy and nonlinear scanning optical microscopy, in which TPL excited with a strongly focused laser beam at the wavelength 745 nm is detected. Even from the smallest (55-nm-wide and 11-nm-thin) structures TPL could still be easily detected through the layer of refractive index matching oil. We used the TPL images to map the local field enhancements from individual nanostrips at a resolution of 0.35 µm and obtain good agreement with theoretical calculated reflection spectra, enhancement levels and field distributions across the strip. The theoretical field distributions clearly show the standing wave patterns of SR-SPPs inside the nanostrips of different widths and both experimental and theoretical results show increased field enhancements at the strip edges along the nanostrip.
The authors gratefully acknowledge financial support (TS) from the Danish Research Council for Technology and Production, (JB and SB) from the European Network of Excellence, PLASMONANO-DEVICES (FP6-2002-IST-1-5078789), and (TS, SN and SB) from the NABIIT project financed by the Danish Research Agency (contract No. 2106-05-033).
References and links
1. Optics of Nanostructured MaterialsV. M. Markel and T. F. George, eds. (John Wiley and Sons, New York, NY, 2001).
2. G. T. Boyd, Th. Rasing, J. R. R. Leite, and Y. R. Shen, “Local-field enhancement on rough surfaces of metals, semimetals, and semiconductors with the use of optical second-harmonic generation,” Phys. Rev. B 30, 519–526 (1984), and references therein. [CrossRef]
3. E. J. Sánchez, L. Novotny, and X. S. Xie, “Near-Field fluorescence microscopy based on two-photon excitation with metal tips,” Phys. Rev. Lett. 82, 4014–4017 (1999). [CrossRef]
4. K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari, and M. S. Feld, “Surface-enhanced Raman scattering and biophysics,” J. Phys. Condens. Matter 14, R597–R624 (2002). [CrossRef]
5. P. J. Schuck, D. P. Fromm, A. Sundaramurthy, G. S. Kino, and W. E. Moerner, “Improving the mismatch between light and nanoscale objects with gold bowtie nanoantennas,” Phys. Rev. Lett. 94, 017402 (2005). [CrossRef] [PubMed]
7. A. Hohenau, J. R. Krenn, S. G. Rodrigo, L. Martin-Moreno, F. Garcia-Vidal, J. Beermann, and S. I. Bozhevolnyi, “Spectroscopy and nonlinear microscopy of gold nanoparticle arrays on gold films,” Phys. Rev. B 75, 085104 (2007). [CrossRef]
8. A. K. Sarychev, V. M. Shalaev, and M. I. Stockman, “Local fields’ localization and chaos and nonlinear-optical enhancement in clusters and composites,” in Optics of Nanostructured Materials, Ref. 1, p. 313, and references therein.
10. J. Beermann and S. I. Bozhevolnyi, “Microscopy of localized second-harmonic enhancement in random metal nanostructures,” Phys. Rev. B 69, 155429 (2004). [CrossRef]
11. T. Søndergaard and S. I. Bozhevolnyi, “Slow-plasmon resonant nanostructures: Scattering and field enhancements,” Phys. Rev. B. 75, 073402 (2007). [CrossRef]
12. A. Mooradian, “Photoluminescence of metals,” Phys. Rev. Lett. 22, 185–187 (1969). [CrossRef]
13. G. T. Boyd, Z. H. Yu, and Y. R. Shen, “Photoinduced luminescence from the noble metals and its enhancement on roughened surfaces,” Phys. Rev. B 33, 7923–7936 (1986). [CrossRef]
14. M. R. Beversluis, A. Bouhelier, and L. Novotny, “Continuum generation from single gold nanostructures through near-field mediated intraband transitions,” Phys. Rev. B 68, 115433 (2003). [CrossRef]
15. P. Ghenuche, S. Cherukulappurath, T. H. Taminiau, N. F. van Hulst, and R. Quidant, “Spectroscopic Mode Mapping of Resonant Plasmon Nanoantennas,” Phys. Rev. Lett. 101, 116805 (2008). [CrossRef] [PubMed]
16. A. Bouhelier, M. R. Beversluis, and L. Novotny, “Characterization of nanoplasmonic structures by locally excited photoluminescence,” Appl. Phys. Lett. 83, 5041–5043 (2003). [CrossRef]
17. T. Søndergaard, J. Beermann, A. Boltasseva, and S. I. Bozhevolnyi, “Slow-plasmon resonant-nanostrip antennas: Analysis and demonstration,” Phys. Rev. B. 77, 115420 (2008). [CrossRef]
18. T. Søndergaard, “Modeling of plasmonic nanostructures: Green’s function integral equation methods,” Phys. Status Solidi B 244, 3448–3462 (2007). [CrossRef]
19. J. Beermann, I. P. Radko, A. Boltasseva, and S. I. Bozhevolnyi, “Localized field enhancements in fractal shaped periodic metal nanostructures,” Opt. Express15, 15234–15241 (2007), http://www.opticsinfobase.org/abstract.cfm?URI=oe-15-23-15234. [CrossRef] [PubMed]
20. R. W. Boyd, Nonlinear Optics (Academic Press, London, 1992).
21. T. Søndergaard and S. I. Bozhevolnyi, “Metal nano-strip optical resonators,” Opt. Express15, 4198–4204 (2007), http://www.opticsinfobase.org/abstract.cfm?URI=oe-15-7-4198. [CrossRef] [PubMed]