We report a novel splicing-based pressure-assisted melt-filling technique for creating metallic nanowires in hollow channels in microstructured silica fibers. Wires with diameters as small as 120 nm (typical aspect ration 50:1) could be realized at a filling pressure of 300 bar. As an example we investigate a conventional single-mode step-index fiber with a parallel gold nanowire (wire diameter 510 nm) running next to the core. Optical transmission spectra show dips at wavelengths where guided surface plasmon modes on the nanowire phase match to the glass core mode. By monitoring the side-scattered light at narrow breaks in the nanowire, the loss could be estimated. Values as low as 0.7 dB/mm were measured at resonance, corresponding to those of an ultra-long-range eigenmode of the glass-core/nanowire system. By thermal treatment the hollow channel could be collapsed controllably, permitting creation of a conical gold nanowire, the optical properties of which could be monitored by side-scattering. The reproducibility of the technique and the high optical quality of the wires suggest applications in fields such as nonlinear plasmonics, near-field scanning optical microscope tips, cylindrical polarizers, optical sensing and telecommunications.
© 2011 OSA
Surface plasmon-polaritons (SPPs) have in recent years been extensively studied for their ability to guide and confine light on a subwavelength scale [1,2]. Metallic nanowires can support guided SPP modes, some of which can exhibit unusually low losses – these modes bear some resemblance to long-range surface plasmons guided on symmetrical dielectric-metal-dielectric structures . Such nanowires have recently attracted attention for their potential use in nano-scale integrated optical circuits [4–10], tapered versions having applications in nanofocusing and nonlinear optics [11,12].
Nanowires have recently been integrated into optical fibers using a high-temperature pressure cell approach to pump molten Au and Ag into the hollow channels of solid-core photonic crystal fibers (PCFs) [13,14]. Length-to-diameter (aspect) ratios of 100,000 have been achieved, and the wires are inherently free of impurities, as no chemical precursors are used. Selective filling of individual channels has also been demonstrated . In such metal-filled PCFs, distinct resonances are observed when the axial wavevectors of the guided core mode and the guided SPP modes become phase-matched. Such hybrid PCFs merge the fields of plasmonics and fiber optics, and are finding applications in areas such as optical sensing and subwavelength-scale imaging , [13–16].
In this paper we describe a spliced-fiber pressure-filling technique. Straightforward to implement, this allows selective filling of individual channels with diameters as small as 120 nm. The hollow channels can also be tapered by thermal post-processing, yielding conical nanowires after filling. We report spectral side-scattering and transmission measurements on fibers consisting of a Ge-doped glass core and single parallel gold-filled channel (both uniform and conical).
2. Spliced-fiber pressure-filling technique
A schematic of the spliced-fiber pressure-filling technique is shown in Fig. 1 . First a silica capillary with outer and inner diameters of ~200 μm and ~80 μm was fabricated by standard fiber drawing. A gold wire (diameter ~50 μm, 2 cm long) was inserted into the capillary (Fig. 1a) and pushed further in using a stiff tungsten wire. The end portion of the capillary was cleaved off to provide a fresh and clean endface (Fig. 1b). This capillary end was then spliced to a silica PCF in a Vytran splicing machine (Fig. 1c). Special care was taken to adjust the splicing parameters so that the splice was mechanically strong while keeping the hollow channels open. For fibers with relatively small channel diameters (d < 1 μm), it was sometimes necessary to inflate the channel before splicing by applying appropriate pressure above the softening temperature of silica . In the final step, the fiber was placed in a vertical furnace with the spliced section in the centre and heated up to the melting temperature of gold (~1064°C). Argon gas at a pressure of several hundred bar was then applied through the second end of the capillary, causing molten gold to flow in (Fig. 1d). A prerequisite of this method is that, after melting, the gold plug must completely fill the hollow channels in PCF (Fig. 1e). The technique is suitable for non-wetting materials with melting temperatures significantly lower than the softening temperature of silica (~1400°C), such as Au, Ag, Ga and Ge  and chalcogenide glasses . The splice-filling technique cannot be used with materials such as tellurite glass that strongly wet the silica glass, because this leads to the formation of an annular film (with a central hollow bore) on the inner channel surface. Compared to the pressure-cell filling technique [13,14,20,21], the splice-filling technique is more flexible and safe, requires only small quantities of material, and is easier to adapt for selective channel filling. Similarly to the direct fiber drawing approach , the splicing approach allows fabrication of very thin metallic wires. It is however especially useful for creating large arrays of wires - difficult to realise using fiber drawing since instabilities can occur when there is a large fraction of metal in the cane, leading to structural irregularities.
The filling dynamics are dominated by the capillary effect, and are described by the Washburn equation . Neglecting gravitational forces (an excellent approximation at the high pressures used), the filling length L is given by [19,23]:24,25]), a minimum pressure of p min = −(2γ/R)cosθ is required to initiate the filling process. In our set-up the maximum available pressure is 300 bar, yielding a minimum fillable capillary diameter (critical diameter) of 27 nm (Fig. 2c ). The calculated pressure needed to fill 10 cm long silica nano-channels within one hour is plotted versus channel diameter in Fig. 2a for molten glass, molten gold and water (parameters given in the Appendix—see Table 1 ). The smallest channel diameter so far filled with gold is 120 nm (filling length 3 cm after 20 min); a scanning-electron-micrograph (SEM) of the resulting structure is shown in the Fig. 2b (the endface is polished using focused ion-beam milling).
Low-melting-point compound glasses can be used if their softening temperatures are below ~1200°C (the softening point of fused silica). Figure 2a shows that channels 10 cm long with diameters of ~1 μm can be filled within 1 hour at 300 bar.
Figure 3 shows optical micrographs of several splices and SEMs of different Au-filled fibers. First a solid-core PCF with cladding hole diameter 1 μm and pitch 3 μm was completely filled by applying a pressure of ~50 bar (Fig. 3a). Selective filling was achieved by splicing an intermediate capillary between the PCF and gold capillary (Fig. 3b). This additional capillary had a eccentric channel (diameter 3 μm, center offset 6 µm) and acted as a mask, allowing selective channel filling. An example of what can be achieved with this masking technique is given in Fig. 3b right, where two holes in the second and third cladding rings are filled with Au. The second type of fiber, used for the optical experiments in this paper, was a modified step index fiber (MSIF) consisting of a 510 nm hollow channel running parallel to a Ge-doped silica core (core diameter ~1100 nm, dopant concentration 20 mol%) with a centre-centre spacing of 3.5 μm . The hollow channel was inflated before splicing and then filled with gold by applying ~100 bar of Ar pressure at 1100°C (Fig. 3c). The entire sample had contiguous filled and unfilled sections 10 cm and 50 cm long.
To investigate the structural properties of the nanowires, several fiber capillaries with different IDs (120 nm, 400 nm, 500 nm, 700 nm, 900 nm, 1.2 µm, 1.5 µm, 2 µm) were filled with gold at 1150°C in a vertical tube furnace with a uniform hot-zone of length ~10 cm. For capillaries with IDs larger than 1.2 µm, continuous wires several cm in length were obtained (confirmed by electrical conductivity measurements). For smaller IDs, however, the wires showed nm long breaks at ~100 µm intervals. For example, the longest continuous wire at an ID of 500 nm was 122 µm.
3. Scattering of guided SPP modes on Au-nanowire
As a next step, we investigated the optical properties of the gold-filled MSIF (wire diameter 510 nm) introduced in the previous section (Fig. 3c). When broadband (450 to 1700 nm) white light was launched into the glass core, bright red-coloured scattering points appeared at the end of each continuous nanowire section (Fig. 4c , light launched from right). The brightness of these spots was almost independent of launched polarization state. To investigate the polarization of the scattered light, a linear polarizer was inserted between the fiber and the microscope. The scattered light was strongest when the polarizer was oriented perpendicular to the nanowire and lowest when the polarizer is parallel to it (Fig. 5b ).
The spectrum of the side-scattered light was measured by scanning a multimode fiber along the filled MSIF and delivering the collected light to an optical spectrum analyzer. To enhance the signal, a droplet of index matching fluid was placed between the multimode fiber and the Au-filled MSIF. Figure 5a shows the scattering spectra at 1.5 mm intervals along the fiber. A clear peak is observed at ~700 nm at the starting point of the nanowire section (z = 0). Over the wavelength band investigated the scattered intensity dropped exponentially along the MSIF, the strongest decay rate being observed at 700 nm. Fitting the data at each wavelength to an exponential function resulted in the attenuation spectrum represented by the black curve in Fig. 6b .
4. Theory and analysis
The coupled system of dielectric core mode (C) and guided SPP mode (P) has two eigenmodes with propagation constants and modal shapes given approximately by :Equation (2) is valid in the vicinity of an anti-crossing, at the centre of which ϑ = 0. It shows that one “plasmon-like” eigenmode has very high attenuation ( + sign), the fields being concentrated on the nanowire, whereas the other “dielectric-like” mode has very low attenuation (− sign), the fields being predominantly in the glass core. When light guided in the glass core reaches the point where the nanowire starts, boundary conditions dictate that the dielectric-like mode will dominate, the high loss mode being more weakly excited and in any case being rapidly absorbed in the wire. When the surviving low-loss mode reaches a break in the nanowire, light is strongly scattered, the strength of the scattered signal being determined by the proportion of optical power travelling on the nanowire, which in turn depends on the balance between loss and coupling.
This explains qualitatively the origin of the bright red scattering points seen in the experiment (Fig. 4c). For a more quantitative understanding, finite-element (FE) simulations were carried out on a structure in which a small air gap was introduced between the wire and the glass – expected because the gold has a much larger coefficient of thermal expansion (15 × 10−6 K−1) than silica (0.5 × 10−6 K−1). For the parameters in the experiment we estimate this gap to be 3.5 nm wide, and in the calculations we assume this to be constant around the circumference of the wire. The dispersion of the gold and silica was taken from  and .
The FE modelling results (red curve in Fig. 6b) show that the dielectric-like mode has a strong attenuation peak at 700 nm, in excellent agreement with the experiments. After fitting the coupled mode expressions in Eq. (2) to the data, making use of the exact dispersion of the isolated core and guided SPP modes (Fig. 6a) and adjusting the coupling constant so that the calculated peak attenuation matches the experimental value (this yields κ = 1.8 mm−1), the blue curve in Fig. 6b is obtained. The attenuation of the plasmon-like mode turns out to be 340 dB/mm, compared to 0.695 dB/mm for the dielectric-like mode.
The Poynting vector distributions of the dielectric- and plasmon-like modes, calculated at the resonance wavelength using the FE code (Fig. 6c & d), show that the plasmon-like mode has 0.96% of its power on the wire, whereas the dielectric-like mode has only 0.0026%. Since α/κ ≈43 >> 1, the attenuation of the plasmon-like mode at resonance is almost identical to that of the guided SPP on the isolated wire (340.5 dB/mm compared to 341.2 dB/mm) - inset of Fig. 6b (the attenuation of a surface plasmon at a planar interface is ~1000 dB/cm at 700 nm ). The dielectric-like mode, in contrast, has an on-resonance attenuation of only 0.7 dB/mm, which is remarkably low for a plasmonic device and can be adjusted over a wide range by changing the core-wire spacing. The measured on-resonance loss value is comparable to that of recently reported devices (e.g., 1 dB/cm at 1.31 µm ).
5. Conically-tapered nanowire
In a separate experiment, thermal post-processing was used to taper the hollow channel while keeping the diameter of the glass core constant . The tapered channel was then filled with gold (Fig. 7a ). Similarly to the untapered fiber, continuous wires ~100 µm long were separated by μm-wide gaps. The initial channel diameter was ~740 nm, the core diameter 1.1 μm and the center-center core-wire spacing was 3.5 μm. The conical hole diameter causes the resonant wavelength to vary along the structure, with the result that the side-scattered peak shifts toward longer wavelength over a distance of ~3.5 mm. Using the scattering data the local diameter of the nanowire could be calculated by matching the wavelength of the measured peak to the data from a series of previous experiments (inset of Fig. 7c) which showed that a 1 nm change in wire diameter causes a 0.44 nm shift in the resonance wavelength. The resulting diameter profile of the conical wire is plotted in Fig. 7c.
We have described a novel and easy-to-implement fiber-splicing technique for pumping molten metals into nano-scale hollow channels in silica glass fibers. The technique is used to fill a single hollow channel placed adjacent to a conventional step-index core in a MSIF. Guided SPP waves are excited at specific resonant wavelengths where the glass core mode is phase-matched to one of the guided SPP modes. Breaks in the nanowires act as efficient antennae, causing strong scattering of resonant light (at 700 nm) out of the fiber. The absence of scattering along continuous nanowire sections (aspect ratios of 50:1) indicates that the nanowires are of excellent quality with low surface roughness. Coupled mode theory can be used to model the behaviour of the glass-core/nanowire system, showing that two eigenmodes exist, one (a plasmon-like mode) with very high attenuation and the other (a dielectric-like mode) with attenuation as low as 7 dB/cm at resonance. The directional coupler geometry allows gentle energy transfer on to the wire at a rate that can be adjusted precisely by varying the core-nanowire spacing. The nanowire MSIF represents a merging of the fields of plasmonics and fiber optics and may find applications in various fields such as photonic-plasmonic integration, nonlinear plasmonics, near-field scanning optical microscope tips, optical sensing and telecommunications.
References and links
2. S. A. Maier, “Plasmonics: The promise of highly integrated optical devices,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1671–1677 (2006). [CrossRef]
4. A. V. Akimov, A. Mukherjee, C. L. Yu, D. E. Chang, A. S. Zibrov, P. R. Hemmer, H. Park, and M. D. Lukin, “Generation of single optical plasmons in metallic nanowires coupled to quantum dots,” Nature 450(7168), 402–406 (2007). [CrossRef]
6. H. Ditlbacher, A. Hohenau, D. Wagner, U. Kreibig, M. Rogers, F. Hofer, F. R. Aussenegg, and J. R. Krenn, “Silver nanowires as surface plasmon resonators,” Phys. Rev. Lett. 95(25), 257403 (2005). [CrossRef]
7. J. Tian, Z. Ma, Q. A. Li, Y. Song, Z. H. Liu, Q. Yang, C. L. Zha, J. Akerman, L. M. Tong, and M. Qiu, “Nanowaveguides and couplers based on hybrid plasmonic modes,” Appl. Phys. Lett. 97(23), 231121 (2010). [CrossRef]
10. X. Guo, M. Qiu, J. M. Bao, B. J. Wiley, Q. Yang, X. N. Zhang, Y. G. Ma, H. K. Yu, and L. M. Tong, “Direct coupling of plasmonic and photonic nanowires for hybrid nanophotonic components and circuits,” Nano Lett. 9(12), 4515–4519 (2009). [CrossRef]
12. A. R. Davoyan, I. V. Shadrivov, A. A. Zharov, D. K. Gramotnev, and Y. S. Kivshar, “Nonlinear nanofocusing in tapered plasmonic waveguides,” Phys. Rev. Lett. 105(11), 116804 (2010). [CrossRef]
13. H. W. Lee, M. A. Schmidt, H. K. Tyagi, L. P. Sempere, and P. St. J. Russell, “Polarization-dependent coupling to plasmon modes on submicron gold wire in photonic crystal fiber,” Appl. Phys. Lett. 93(11), 111102 (2008). [CrossRef]
14. M. Schmidt, L. Prill Sempere, H. Tyagi, C. Poulton, and P. Russell, “Waveguiding and plasmon resonances in two-dimensional photonic lattices of gold and silver nanowires,” Phys. Rev. B 77(3), 033417 (2008). [CrossRef]
15. H. K. Tyagi, H. W. Lee, P. Uebel, M. A. Schmidt, N. Joly, M. Scharrer, and P. St. J. Russell, “Plasmon resonances on gold nanowires directly drawn in a step-index fiber,” Opt. Lett. 35(15), 2573–2575 (2010). [CrossRef]
18. M. J. Weber, Handbook Of Optical Materials (CRC Press, 2002).
19. N. Da, L. Wondraczek, M. A. Schmidt, N. Granzow, and P. St. J. Russell, “High index-contrast all-solid photonic crystal fibers by pressure-assisted melt infiltration of silica matrices,” J. Non-Cryst. Solids 356(35-36), 1829–1836 (2010). [CrossRef]
20. M. A. Schmidt, N. Granzow, N. Da, M. Y. Peng, L. Wondraczek, and P. St. J. Russell, “All-solid bandgap guiding in tellurite-filled silica photonic crystal fibers,” Opt. Lett. 34(13), 1946–1948 (2009). [CrossRef]
21. H. K. Tyagi, M. A. Schmidt, L. Prill Sempere, and P. S. Russell, “Optical properties of photonic crystal fiber with integral micron-sized Ge wire,” Opt. Express 16(22), 17227–17236 (2008). [CrossRef]
22. E. W. Washburn, “The dynamics of capillary flow,” Phys. Rev. 17(3), 273–283 (1921). [CrossRef]
23. N. Da, A. A. Enany, N. Granzow, M. A. Schmidt, P. St. J. Russell, and L. Wondraczek, “Interfacial reactions between tellurite melts and silica during the production of microstructured optical devices,” J. Non-Cryst. Solids 357(6), 1558–1563 (2011). [CrossRef]
24. S. Z. Beer, Liquid Metals (Marcel Dekker, Inc, 1972).
25. T. Iida, and R. Guthrie, The Physical Properties of Liquid Metals (Clarendon Press, 1988).
26. A. W. Snyder, and J. Love, Optical Waveguide Theory , 1st ed. (Springer, 1983).
27. P. G. Etchegoin, E. C. Le Ru, and M. Meyer, “An analytic model for the optical properties of gold,” J. Chem. Phys. 127(18), 164705 (2007). [CrossRef]