Herein we characterize and experimentally demonstrate a new type of a horizontal slot waveguide structure for remarkably enhanced Raman scattering detection in nanometer-scale void channels. As the measurement sensitivity is one of the key limiting factors in nanofluidic detection, it is essential to search advanced solutions for such detection. Combining an all dielectric resonance waveguide grating and a surface enhanced Raman scattering (SERS) substrate in a close proximity it is possible to create high electromagnetic field energy hot zones within an adjustable slot region. This results in a strong enhancement in Raman scattering. We show the theoretical principles and demonstrate, with rhodamine 6G molecules, an approximately 20-fold enhancement compared to a conventional SERS substrate within the corresponding slot arrangement. We foresee potential applications for the proposed approach in the fields of medical, biological and chemical sensing, where the high detection sensitivity is essential due to integration with nanofluidic devices.
© 2013 OSA
With the progress of nanoscale fabrication processes in recent years, various important applications containing nanostructures are developed for chemical sensing and biosensors [1, 2]. Highly attractive approaches related to improving the measurement sensitivity and the resolution of lab-on-chip type devices are featured by pushing the limit of manipulation of nanoparticles or biomolecules from micron scale  to sub-wavelength scale structures [4, 5].
Surface enhanced Raman scattering (SERS) is greatly benefiting from the evolving nanofabrication processes. SERS is an attractive and rapidly expanding field for molecular sensing due to its high sensitivity and fingerprint uniqueness for different molecules . Most commonly SERS spectroscopy is practiced on nanostructured silver and gold surfaces . The principal enhancement mechanisms for SERS are based on the localized electromagnetic field enhancement  and on the charge transfer effect between an adsorbed analyte molecule and photo-excited electrons at metal surface . The local electromagnetic field enhancement follows from the coupling of surface plasmons that are excited by surface modulation. Even higher enhancements, where the nominal enhancement factor exceeds 1013, can be achieved by more complex interactions between several nanoparticles enabling the detection of single molecules . Recently, the ability of focusing plasmons in nanoslits  and the development of nanochannel optofluidic devices [12, 13] have provided potential for through-channel molecular Raman studies.
The resonance waveguide grating (RWG) can be tailored to produce a very high electromagnetic field confinement inside the waveguide region of the grating which can be used to remarkably enhance fluorescence detection [14, 15]. This can also be used for the Raman investigation of materials that are essentially forming the waveguide of an RWG structure . The physical background of an RWG was originally described by Mashev et al.  and it is further widely discussed in literature [18–21]. The phenomenon relies on evanescent leaky modes that interfere together with the propagating mode, resulting in an extraordinary high reflection while the field intensity within the grating and its near field undergoes a significant confinement. The resonance waveguide grating concept is often referred also as a guided-mode resonance filter or a leaky-mode resonance filter and it is used for several applications such as spectral filtering [22, 23], molecular detection , two-photon fluorescence excitation , and second harmonic generation .
Herein, we are taking the advantage of the two field enhancement phenomena by placing a SERS substrate and an RWG at a very close proximity (∼ 100 nm) from each other. This allows a reformation of the electromagnetic field distributions at both surfaces due to the interaction of their evanescent fields. Further, this leads to a very high field confinement in the low index medium between the two tailored surfaces, and hence an interesting analogy to slot waveguides  can be found. The fabrication methods presented here enable a rather straightforward transfer of these promising devices from a laboratory to large scale fabrication.
2. Concept and design
The concept of the device is illustrated in Figs. 1(a) and 1(c). The reduced silver SERS substrate on which the sample molecules are attached is set on the bottom. The RWG, containing a modulated cyclo olefin copolymer (COC) polymer and a layer of TiO2, is brought from the top to a very close proximity of the underlying SERS substrate, leaving a narrow gap hg between the two devices. The modulated TiO2 layer has a higher effective refractive index to its surroundings which fundamentally enables the coupling of the waveguide mode in the RWG. A grating period d = 350 nm has a modulation defined by the fill factor ff = 0.35, the modulation depth hm = 150 nm, and the TiO2 coating thickness ht = 75 nm. The gap distance can be within the range of tens to a few hundred nanometers until the proximity benefit is entirely lost. However, the functionally optimal distance was found to be at the range of sub-hundred nanometers, which is expected as the evanescent fields for both structures fade out exponentially.
The device parameters were designed by the electromagnetically fully rigorous Fourier-Modal Method (FMM) and an S-matrix algorithm . The handling of the conical illumination is based on the approach successfully applied in the reference . The basis of the structure design is to capture the excitation light inside the horizontal slot between the RWG and the SERS substrate. The optimal design was defined by searching the parameters that provided the condition where minimum reflection occurs as is seen in Fig. 1(b). This optimization was carried out by using the Nelder-Mead optimization algorithm. This was taken as a starting point for the further optimization process where the maximum electromagnetic field intensity inside the slot area was the target. Figure 1(b) shows the reflection efficiency for a conical illumination where θ and φ are the conical angles. The circle represents an angular view through a 0.45 NA objective.
As the rigorous modeling of an electromagnetic field behavior for a reduced silver substrate is rather complex due to variations in shape and size of the functional particles, a simpler approach was chosen to model the functionality. At first, only a planar 50 nm thick Ag film on an SiO2 substrate was assumed to be the underlying layer. The simulations were carried out for the excitation wavelength of 514 nm with the TM polarization and the excitation was assumed to be in the plane wave form. Figure 2 shows the calculated time average electric energy densities inside the system. The optimum incident angle θ is around nine degrees (φ = 0), the optimized gap distance is 50 nm, and (a) only the RWG is set above the SiO2 substrate, (b) only the planar Ag-film appears on the SiO2 substrate, and (c) the complete slot channel is formed by the RWG and the planar Ag-film. The borders of the structures are highlighted with white solid lines. The electric energy densities along the dashed white lines in Figs. 2(a)–2(c) are plotted in Fig. 2(d), which clearly shows the remarkably increased intensity within the gap region for the stack configuration. The scaling ratio ω/ω0 gives the electric energy density ω relative to the maximum of the incident (z ¡ 0) field ω0.
The simulations shown in Fig. 2 clearly predict the significant field enhancement in the gap region when the RWG-Ag-stack is formed. However, surface plasmons do not play a significant role as the silver film has no features. For the next simulation step, the surface roughness was added for the silver film. A periodic, 10 nm high, modulation was added while the same structure parameters for the RWG were used to examine the field behavior. Only the gap distance and the optimum incident angle were varied. Figure 3 shows, respectively to Fig. 2, the time average electric densities for (a) the featured Ag-film and (b) the RWG-Ag-slot channel. The mode shapes along the z-axis are plotted for x equals 130 nm (c) and x equals 177 nm (d). Now, the modulation on the silver film excites surface plasmons but without the RWG the field maximums are only moderately enhanced. When the RWG is added, roughly a 300-fold increase in the field intensity is achieved in comparison with the incident field.
The two main benefits of the proposed structure are as follows. Firstly, the plasmon field is enhanced at the spots where it appears also without the RWG and, secondly, the field enhancement takes place over the gap volume. The latter will drastically enhance the excitation of molecules that are not critically located in the plasmon mode hotspots. Even though the enhancement is not as high as it is in the immediate presence of Ag modulation, still up to about 60-fold, the enhancement mode volume is remarkably larger than in the case when only the silver SERS substrate is used.
3. Experimental results and discussion
The RWG fabrication was carried out by the well-established lithographic fabrication methods. An imprint mold was prepared on a silicon substrate using electron beam lithography (Vistek EBPG5000+ES HR) on a spin-coated hydrogen silsesquioxane (HSQ) resist. The resist was developed with the M 361 developer and was further hard baked in a convection oven and silane treated to improve its performance in the replication process. The imprint mold fabrication process is described in full details in the reference . The mold replication into the COC (n = 1.53 at 514 nm) film was made by hot embossing (Obducat, NIL Eitre 3). The high refractive index layer of TiO2 (n = 2.45 at 514 nm) was coated by atomic layer deposition (Beneq, TFS 200).
Metallic silver nodules on a glass substrate were achieved by a modified Tollen’s method . The method is based on a chemical reduction of the complex cation [Ag(NH3)2]+ by d-glucose. 10 mM Tollen’s reagent was made by adding 25% NH3 to a 25 mM AgNO3 solution in a stepwise manner. At first, a precipitate of AgO was formed which was then solubilized after adding more ammonia. At the end of the reaction the final concentration of the formed [Ag(NH3)2]+ was adjusted to 10 mM by adding ddH2O. pH of the reagent was measured to be 11. 40 μl of freshly made Tollen’s reagent was applied on a glass microscope slide and 5 μl of 1.0% d-glucose was added and mixed in a drop. The glass slide was then heated to 60°C by placing it onto a thermal block. After 5 min. the slide was thoroughly rinsed in ddH2O and dried in an air stream. Then 1.0 μM rhodamine 6G was applied onto the surface and incubated for 20 min. at room temperature. The excess of rhodamine 6G was washed out with ddH2O and the slide was air dried prior to measurements.
The scanning electron microscope (SEM LEO 1550 Gemini) imaging was used to ensure the quality of the SERS and RWG structures. Figure 4 shows (a) the SEM image of the SERS substrate and (b) a cross-sectional SEM image of the RWG structure. It is visible that there is some room to improve the surface uniformity of the SERS substrate but the functionality of the SERS substrate was not the main goal of this work. For the SEM cross-sectional imaging of the RWG the grating was cut in half by cooling it down in liquid nitrogen and bending after that. Therefore, polymer lines have been torn and there appears a visible fraction between TiO2 and polymer layers. Also, due to fabrication the features in polymer have been rounded, which leads to less optimal features in the final structure resulting in some losses and reduction in the device performance.
The performance of our reduced silver SERS substrates was proved by comparing them to silver nanoparticle SERS substrates developed by ion-exchange that were reported earlier in . The Raman spectroscopy measurements were carried out by a Renishaw in Via Raman microscope with the excitation wavelength of 514 nm, and a 50X, 0.45 NA objective. Figure 5 justifies that, by using the same immersion technique for Rh6G molecules on our reduced silver substrate (top red) and the ion-exchanged silver nanoparticles (middle blue), the performance was exceptionally good. For the comparison, the black line on the bottom shows the Raman signal of Rh6G from a planar SiO2 surface that was used as a substrate for both SERS methods. A ten-second integration time and 56 mW laser power were used for all the measurements.
As the main result of this work, we have demonstrated the Raman signal enhancement in the RWG-SERS-formed nanoslot channel compared to a conventional SERS substrate in an equal slot configuration. The slot configuration was formed by compressing and trapping the two functional surfaces together within vacuum in the clean room conditions. In Fig. 6, the purple line on the top shows the measured Raman signal of Rh6G from the RWG-Ag-slot channel and the red line in the middle shows the Raman intensity from the slot channel formed by the reduced silver SERS substrate and the unmodulated polymer and the TiO2 films. The latter signal is given as an average over nine measured values and the RWG signal is the peak value of the measurements. The gap width highly influences the performance of the RWG-Ag-system and the strict control of it was not possible in the measurements. However, the enhancement was demonstrated over several samples and the significant enhancement was only found when the tight compression was applied in the clean room conditions. Black line at the bottom shows the measured Raman signal for the RWG-glass-channel without the functional SERS layer. Again, a ten-second integration time and 56 mW laser power were used.
By adding the RWG, a significant 20-fold increase in the Raman intensity for 1575 cm−1 line was observed in comparison with the plain SERS substrate. As a part of the Raman signal is also wavelength dependently coupled in the RWG, the enhancement slightly differs depending on the Raman band being in the range of 15–20-fold for the bands shown in Fig. 6. Some of the SERS substrate performance is lost due to additional absorption and scattering of the COC-TiO2-film covering the SERS substrate. In addition, the compression of the structures quite certainly causes degradation in the SERS substrate reducing the performance. These results are highly promising, in particular for applications that require the encapsulation or covering of the sample, with the lab-on-chip with integrated nanofluidistics being a good example.
It can be seen from the simulations that even greater enhancement is predicted if all the excitation light came from a single angle. Here however, the results were achieved simply by an unmodified Raman microscope assembly with the standard through lens configuration where the excitation angle is conical, which means that only a relatively small portion of the excitation light was used beneficially. Slight modifications to the measurement configuration should provide a higher enhancement. The main benefit of applying the RWG is to control the reflected light from the SERS substrate which leads to a more efficient use of excitation energy. Due to the randomness of the SERS particles a part of excitation is, against the simulation presumptions, scattered to undesired directions degrading the performance of the device. Also, a SERS substrate with more controlled features should provide a better performance. Considering the mentioned drawbacks, the experiments clearly demonstrate that the concept is capable of enhancing the Raman signal in the given slot configuration. From the point of view of the current capabilities of nanofabrication techniques, it is fully possible to employ the proposed structure in integrated lab-on-chip type devices.
We have demonstrated a Raman enhancement technique that combines the benefits of the surface enhanced Raman spectroscopy substrate and an all dielectric resonance waveguide grating in a horizontal slot channel. A 20-fold increase in the enhancement was measured compared to the sole SERS substrate performance. The concept shown here can be easily applied to any SERS substrate that allows bringing the RWG structure close enough and it can be tailored for different needs in terms of the channel width. The benefits compared with the conventional SERS detection in a slot volume are the ability to enhance the SERS substrate performance and the remarkably stronger enhancement in a volume that is not directly at the SERS hotspots. The wider enhancement volume makes the approach potentially valuable for online measurement devices. The possible applications are in the fields of medical, biological, and chemical sensing, and due to the optimum performance in the channels up to hundred nanometers in size, it is a highly promising sensing technique for nanofluidic channels. In addition, the possibility to use inexpensive materials and fabrication methods provides a great potential for applications in which low-cost is of crucial importance.
References and links
2. M. M. Bellah and S. M. C. S. M. Iqbal, “Nanostrutures for medical diagnostics.” J. Nanomater 2012, 486301 (2012) [CrossRef] .
3. D. Psaltis, S. R. Quake, and C. Yang, “Developing optofluidic technology through the fusion of microfluidics and optics.” Nature 33, 381–386 (2006) [CrossRef] .
4. A. H. J. Yang, S. D. Moore, B. S. Schmidt, M. Klug, M. Lipson, and D. Erickson, “Optical manipulation of nanoparticles and biomolecules in sub-wavelength slot waveguides.” Nature 457, 71–75 (2009) [CrossRef] .
6. Z. Q. Tian, “Surface-enhanced raman spectroscopy: advancements and applications.” J. Raman Spectrosc . 36, 466–470 (2005) [CrossRef] .
7. M. Moskovits, “Surface-enhanced raman spectroscopy: a brief retrospective.” J. Raman Spectrosc . 36, 485–496 (2005) [CrossRef] .
8. M. Moskovits, “Surface-enhanced spectroscopy.” Rev. Mod. Phys. 57, 783–826 (1985) [CrossRef] .
9. A. Otto, “The ‘chemical’ (electronic) contribution to surface-enhanced raman scattering.” J. Raman Spectrosc. 36, 497–509 (2005) [CrossRef] .
11. C. Chen, J. A. Hutchison, P. Van Dorpe, R. Kox, I. De Vlaminck, H. Uji-i, J. Hofkens, L. Lagae, G. Maes, and G. Borghs, “Focusing plasmons in nanoslits for surface-enhanced raman scattering.” Small 5, 2876–2882 (2009) [CrossRef] [PubMed] .
13. M. Huang, A. Yanik, T.-Y. Chang, and H. Altug, “Sub-wavelength nanofluidics in photonic crystal sensors.” Opt. Express 17, 24224–24233 (2009) [CrossRef] .
14. P. Karvinen, T. Nuutinen, J. Rahomäki, O. Hyvärinen, and P. Vahimaa, “Strong fluorescence-signal gain with single-excitation-enhancing and emission-directing nanostructured diffraction grating.” Opt. Lett. 34, 3208–3210 (2009) [CrossRef] [PubMed] .
15. T. Nuutinen, P. Karvinen, J. Rahomäki, and P. Vahimaa, “Resonant waveguide grating (rwg): Overcoming the problem of angular sensitivity by conical, broad-band illumination for fluorescent measurements.” Anal. Method . 5, 281–284 (2013) [CrossRef] .
16. P. Päivänranta, Nanostructured surfaces for photonic and biophotonic applications (Joensuu University, 2009).
17. L. Mashev and E. Popov, “Diffraction efficiency anomalies of multicoated dielectric gratings,” Opt. Commun. 51, 131–136 (1984) [CrossRef] .
18. S. S. Wang, R. Magnusson, J. S. Bagby, and M. G. Moharam, “Guided-mode resonances in planar dielectric-layer diffraction gratings.” J. Opt. Soc. Am. A 7, 1470–1474 (1990) [CrossRef] .
19. R. Magnusson and S. S. Wang, “New principles for optical filters.” Appl. Phys. Lett . 61, 1022–1024 (1992) [CrossRef] .
20. D. Rosenblatt, A. Sharon, and A. A. Friesem, “ Resonant grating waveguide structures.” IEEE J. of Quantum Elect . 33, 2038–2059 (1997) [CrossRef] .
21. A. Sato, N. Iwai, and M. Sato, “Large incident angle tolerance of guided-mode resonant gratings by light coupling via waveguide end face.” J. Opt. Soc. Am. A 27, 1671–1678 (2010) [CrossRef] .
23. Z. S. Liu, S. Tibuleac, D. Shin, P. P. Young, and R. Magnusson, “High-efficiency guided-mode resonance filter.” Opt. Lett . 23, 1556–1558 (1998) [CrossRef] .
25. S. Soria, T. Katchalski, E. Teitelbaum, A. A. Friesem, and G. Marowsky, “Enhanced two-photon fluorescence excitation by resonant grating waveguidestructures.” Opt. Lett . 29, 1989–1991 (2004) [CrossRef] [PubMed] .
26. A. Saari, G. Genty, M. S. P. Karvinen, P. Vahimaa, M. Kuittinen, and M. Kauranen, “Giant enhancement of second-harmonic generation in multiple diffraction orders from sub-wavelength resonant waveguide grating.” Opt. Express 18, 122298–12303 (2010) [CrossRef] .
28. H. Kim, J. Park, and B. Lee, Fourier Modal Method and Its Applications in Computational Nanophotonics (Tailor & Francis Group, 2012).
29. M. R. Saleem, P. Stenberg, T. Alasaarela, P. Silfsten, M. B. Khan, S. Honkanen, and J. Turunen, “Towards athermal organic-inorganic guided mode resonance filters.” Opt. Express 19, 24241–24251 (2011) [CrossRef] [PubMed] .
30. Y. Yin, Z.-Y. Li, Z. Zhong, B. Gates, Y. Xia, and S. Venkateswaran, “Synthesis and characterization of stable aqueous dispersions of silver nanoparticles through the tollens process.” J. Mater. Chem . 12, 522–527 (2002) [CrossRef] .
31. Y. Chen, L. Karvonen, A. Säynätjoki, C. Ye, A. Tervonen, and S. Honkanen, “Ag nanoparticles embedded in glass by two-step ion exchange and their sers application.” Opt. Mater. Express 1, 164–172 (2011) [CrossRef] .