Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

11.1 Introduction

In 1966, the Nobel Prize winner C. K. Kao concluded his study on “Dielectric-Fibre Surface Waveguides for optical frequencies” [1] envisaging that “a fibre of glassy material (…) represents a possible practical optical waveguide with important potential as a new form of communication medium” [1, 2]. In the following years, the fiber optics field has undergone a tremendous growth and advancement. Initially conceived as a medium to carry light and images for medical endoscopic applications, optical fibers were later proposed as an adequate information-carrying medium for telecommunication applications. As optical fibers established their position in the telecommunications industry and its commercial market matured, parallel efforts were carried out by a number of different research groups around the world to exploit them also in sensing applications . Initially, fiber sensors were lab curiosities and simple proof-of-concept demonstrations but in a few years their operation and instrumentation have become well understood and developed. Over the years, optical sensors enjoyed an increased acceptance as well as a widespread use for structural sensing and monitoring applications in civil engineering, aerospace, marine, oil & gas, composites, smart structures, bio-medical devices, the electric power industry and many others [3–5].

Although fiber optic sensors have, in many cases, completely replaced and outperformed their counterparts based on conventional technologies, it is an established fact that optical fibers are still mainly conceived as a communication medium. For this reason, the fiber optic industry focused its efforts on a small set of materials and structures able to provide light guidance in the transparency range of silica glass.

A significant technological breakthrough would be the development of components and devices “all in fiber” through the integration of advanced functional materials at micro and nano-scale, exhibiting the more disparate properties, combined with suitable light matter interaction mechanisms. This achievement would be the cornerstone of a new photonics technological revolution that would lead to the definition of a novel generation of micro- and nano- photonic devices “all-in-fiber ”. An additional driving force pushing in this direction is the increasing market demand for highly integrated and multifunctional sensor probes with advanced performances and unrivaled features.

In this context, the growth of the life science field, has driven photonics technology into important interdisciplinary fields such as biophotonics and nanophotonics. In particular the optical biosensor field is experiencing a new surge in activity, seeking to exploit novel optical structures and bio-coating materials capable of detecting and discriminating among large classes of molecules. Thanks to rapid advances in this research field, the highly bio-sensitive/selective optical sensors are becoming a viable alternative to traditional “solution based” assay biosensors for applications in genomics, proteomics and drug discovery, as well as in the food industry, homeland security and environmental monitoring applications.

Optical fibers are well suited to support this revolution also by virtue of their dynamicity and versatility offered by using this unique technology. Ten years ago, first demonstrations of microstructured optical fibers [6, 7] opened up new avenues for the development of advanced light waveguides, adding new fiber functionalities, new insights in light manipulation and control [8, 9]. Microstructured optical fibers, because they have many holes, can be easily filled by fluids and have therefore also provided the basis for novel optofluidic devices and components [8–12].

This impressive driving action combined with an ever increasing market demand for a continuous technological innovation, has determined a strong involvement of the optical fibers research community in looking for new fiber materials to use [13] as well as silica fibers with sub-micro diameters to achieve ultimate light matter interaction levels [14].

All the above considerations are leading an envision of the “Lab-on-Fiber “ concept [15, 16] as a concrete technological solution to actual market demand; this new technology essentially deals with the development of novel and highly functionalized materials, devices and components, completely integrated in a single optical fiber and ready to be incorporated in modern optical systems for communication and sensing applications. This new vision, thus, relies on the development of a technological platform based on a single optical fiber where several materials, macro- and nano-structures are constructed, embedded and connected all together to provide the necessary physical connections and light-matter interactions. The Lab-on-Fiber technology could definitely open the way for the implementation of sophisticated autonomous multifunction sensing and actuating systems “all in fiber”, meaning the existence of adjacent multi operational labs which are able to analyze and exchange sensorial data, providing unique auto diagnostics properties.

11.2 Lab on Fiber Technologies

In recent years, remarkable theoretical, numerical and experimental studies on novel optical fibers and novel optical fiber components have been conducted with the aim of concurring to launch the “Lab on Fiber” technologies [15–17]. From this perspective, we refer to “Lab on Fiber” technology as the basis for the development of a technological world completely contained in a single optical fiber where several structures and materials at sub or wavelength scale are constructed, embedded and connected all together to provide the necessary physical connections and light-matter interactions useful in providing a wide range of functionalities and unparalleled performances. Optical fiber technology combined with the new concept of “Lab on Fiber” has the potential to enable the implementation of sophisticated autonomous multifunctioning sensing and actuating systems all of them integrated in optical fibers with unique advantages in terms of miniaturization, light weight, cost effectiveness, robustness and power consumption.

The most obvious technological strategy to produce multi-material optical fiber with new functionalities is to create a new fiber fabrication process . The idea relies on the preparation of a multi-material pre-form suitable for the fiber drawing technique . This technological approach was successfully adopted to produce novel optical fibers acting as photodetectors to visible and infrared light [18] or optical fibers with chemiluminescent features for sensing applications [19] as well as piezoelectric fibers for acoustic transduction [20].

Alternative approaches recently proposed are focused on the integration of photonic crystal structures and/or plasmonic nanostructures on the end facet of optical fibers. In this scenario, the creation of periodic micro and nanostructures on the end facet of optical fibers is of great interest because it may yield versatile optical devices well-suited to serve as miniaturized probes for remote sensing applications. The reduction of feature sizes in micrometer and nanometer scales permits the possibility of obtaining intriguing spectral features, even by lighting the structured pattern out of the plane of periodicity [16, 21–23]. Several approaches have been recently introduced to fabricate metallic and dielectric structures on the optical fiber end facet [2]. Some approaches rely on technique of transferring planar nano-scale structures, fabricated on a planar wafer by means of standard lithographic techniques, onto the optical fiber end facet. These methods exploit well-assessed fabrication processes developed for planar substrates, but they are limited by the final transferring step that plays a fundamental role in determining both the fabrication yield and the performance of the final device. For example it is worth focusing attention on [21–25] where a transfer technique was demonstrated with regards to various metallic patterns. They were first defined by electron-beam lithography (EBL) and then moved to the optical fiber facets by using a sacrificial polymeric film to strip and transfer the metallic patterns. Successively, in order to increase the production throughput, the same authors developed an alternative procedure to realize metallic nanostructures by embedding and sectioning a metallized array of epoxy nanoposts using an ultramicrotome [26]. After the sectioning stage, the thin epoxy slabs are manually transferred to the optical fibers tips and then the epoxy matrix is etched, leaving the metallic pattern on the optical fiber.

Following this basic approach, featured by the two distinct phases of generation and transfer of the nanostructures, the fabrication of a silicon photonic crystal (PC) on the fiber tip was also recently proposed [27]. Similarly, Shambat et al. [28] realized a photonic crystal exhibiting resonant cavities via EBL, transferred it onto the fiber by micromanipulator and bonded the PC on the fiber by depositing small amounts of epoxy, acting as adhesive, on the cladding surface. Scheerlinck et al. [29] developed a procedure in which the realization and transferring of metallic gratings on the optical fiber occurred in a single step by means of UV-based nano imprint and transfer lithography. By following these approaches, multifunctional optical fiber nanoprobes composed of a variety of gold and silver patterns on the facet of an optical fiber were proposed and experimentally demonstrated by highlighting the potential impact of fiber tip probes for sensing applications.

Alternative approaches are based on direct-write patterning of the fiber tip. These methods, based on conventional lithographic techniques adapted to operate on unconventional substrates such as the optical fiber tip, are able to efficiently provide nanostructured devices on the optical fiber, but they require complex and expensive fabrication procedures with a relatively low throughput.

By following direct-writing approaches, fiber based probes composed of a gold layer with sub-wavelength apertures [30] and periodic gold dots arrayed [31] on the facet of an optical fiber were already proposed and experimentally demonstrated by highlighting the potential impact of fiber tip nanoprobes for chemical and biological sensing applications. Such prototypes were possible via Focused Ion Beam (FIB) milling and Electron Beam Lithography (EBL). On this line, recently, spin coating combined with EBL nanotechnology, has been proposed for creating hybrid metallo-dielectric nanostructures, giving rise to resonant effects exploitable for label-free chemical and biological sensing [16].

The direct writing techniques have the merit to be able to sculpt the desired micro and nanostructures with high definition on the optical fiber. So arbitrary sizes and exotic shapes can be defined, inspiring novel configurations for completely new functionalities. The choice of materials is also increased insofar as diversified deposition techniques are developed and assessed on the optical fiber substrates. Nevertheless, all these methods are evidently serial because each device must be written individually by using uneconomical technological equipment. Overall the direct writing approaches, while they are able to efficiently provide nanostructured devices on the optical fiber, are still very challenging by requiring complex and expensive fabrication procedures with a relatively low throughput.

In order to create patterned structures onto the optical fiber tip usinga low cost fabrication approach, we propose the exploitation of self-assembly processes, to be specialized for operating on the optical fiber substrates [32–34]. The autonomous organization of micro and nano structures in regular ensembles [35–37] on the optical fiber end facet is the key tool in obtaining ordered patterns on a high number of optical fibers simultaneously by using easy and low cost fabrication strategy, suitably employed in the mass production of technologically advanced devices.

In the research effort of developing a direct and low-cost method that enables the fabrication of self-assembled periodic structures , the Breath Figures (BFs) are a valid option [34]. In fact, by exploiting this matter auto-organization phenomenon , thin polymeric films having a distribution of micrometric cavities arranged in hexagonal fashion are readily obtainable: (i) through a single step approach, (ii) with no need for lithographic processing, (iii) in a few seconds, and (iv) by means of very simple and accessible laboratory equipment. Indeed, the BF process is a templating method in which the template consists of an ordered array of water droplets that can be removed by simple evaporation. This is an undisputable advantage with respect to most of the other known templating approaches, where the templates need to be removed after the fabrication of the porous films and in most of the cases they are not easily prepared or eliminated. For these reasons, BF is one of the most widely employed methods for the fabrication of porous polymer films [35–37].

On this line, in this chapter, we present recent results on the research activities aimed to use self-assembly to create ordered metallo-dielectric structures on the optical fiber tip. In particular, we report details on the fabrication process based on the BF technique, supported by a wide morphological analysis, and discuss the related principle of operation through numerical and experimental analysis. Finally, the functionality of the realized probes for sensing applications is also demonstrated.

11.3 Breath Figure Structures onto the Optical Fiber Tip

The strategy for the fabrication of periodic structures onto the optical fibers initially involves the coating of the fiber facet by a microporous polymer film, realized via the BF technique. This self-assembly approach, in fact, allows for the preparation of honeycomb-like arrays at micro scale with a high degree of order despite the simplicity of the process. The mechanism of the BF formation has been described in detail by different authors [38–40], and will be, here, briefly recalled. The whole process relies on the precipitation of a polymer around condensed water droplets, initiated by rapid evaporation of a polymer solution in a humid environment. Its salient stages are depicted in Fig. 11.1. In the initial stage, the endothermic evaporation of the solvent results in a decrease of the system temperature, thus triggering the water condensation. Once a droplet of water has nucleated on the polymer solution, it grows at the expense of the vapor of the surrounding atmosphere. During this stage, the growing droplets auto-organize at the polymer solution/air interface into a close hexagonal arrangement. Once the film returns to ambient temperature, the condensed water and residual solvent evaporate, leaving behind a honeycomb structure. When the process is adequately controlled and the key parameters (polymer concentration, kind of solvent, evaporation rate and relative humidity) accurately adjusted, the manipulation of some morphology features in the final film, i.e. degree of order, distribution and size of the cavities, is easily feasible.

Fig. 11.1
figure 1

Schematic overview of the fabrication process. The detail of the microporous film is a real view taken by a confocal microscope, in the central area of the fiber face

Full size image

Even though the applicability of the BF approach to the fabrication of both metallic disk arrays and metallic meshes has already been established [41, 42], this is the first study, to the best of our knowledge, that makes use of BFs as a tool for realizing self-assembled metallo-dielectric periodic structures on the end facet of a standard single mode optical fiber.

To reach this goal, the standard setup, which is normally utilized for making honeycomb films onto glass or silicon substrates, must be adapted to correctly work onto non-conventional substrates as is the case with optical fibers. The main issue to be addressed deals with the fiber tip size: 125 µm of diameter for a standard single mode fiber. Typically, in fact, highly regular BF arrays are obtained by spreading a few droplets of the polymer solution over a flat surface of 1–5 cm2, hence the fiber facet is definitely too small for such an approach. To overcome this issue, the optical fiber was embedded in a ceramic ferule with a diameter of 2.5 mm (similar to the ferule used in standard fiber connectors FC/PC) and then accurately polished. This modified fiber tip offers a larger surface as compared to the bare fiber, therefore resulting in easier handling during both the fabrication process and the operative stage of the final device.

Even though we have experimented with the possibility of preparing BF films directly onto the fiber tip, owing to the little casting volume (4 µL) and consequently quicker evaporation of the polymer solution, we found the reproducibility of this approach rather poor. Therefore, the use of a holder that enlarges the deposition surface, is recommendable. To this purpose, an Al holder of 20 × 20 × 8 mm was shaped in order to exactly fit the ceramic ferule external diameter, thus preventing any solvent from leaking, so that the polymer solution can be easily drop cast over an area of about 1 cm2 (see initial sketch in Fig. 11.1). Since the fiber is placed exactly in the center of the holder, only the central portion of the film is deposited on it, which overcomes also the problem of poor pattern homogeneity that is usually encountered on the edges of BF films [43]. Moreover, this approach is in principle suitable for patterning at the same time more than one fiber per time, simply by arranging multi-slots in the Al holder and by casting a wider film. The Al holder was pre-cooled by placing it on a cold stage which was maintained at a temperature of +10 °C, and removed from there just before starting the film deposition. In fact, given the high thermal conductibility of the metal substrate, as compared to glass, the evaporation-induced cooling of the polymer solution is mainly dispersed through the Al substrate, resulting in poor water condensation. For this reason, such a pre-cooling step is required to promote an efficient water condensation on the evaporating film and, consequently, the BF formation.

Films were prepared by drop casting on the fiber/holder assembly a 4 mg/mL CS2 solution of a fluorinated fluorescent dye-terminated linear polystyrene, which we had already described as a good candidate for producing highly ordered BF structures [44]. To speed up the solvent evaporation and promote water condensation, a flux of moist nitrogen (75 % R.H. at 25 °C) was directed towards the fiber facet. The solvent completely evaporated within 20 s, leaving on the holder and the fiber surface an opaque film, which shone in bright iridescent colors indicating a periodic variation in the refractive index (air holes/polymer walls) throughout the surface of the film. Thanks to the marked fluorescent emissions of the polymer, we were able to rapidly check the quality of patterns formed on the fiber by fluorescence microscopy. Once the BF process was complete, the fiber was carefully detached from the holder, retaining on its tip a portion of the honeycomb film.

Since all other external parameters (temperature and air humidity) as well as those related with the preparation procedure (solvent, polymer concentration and casting volume) were maintained constant, the control on the pore size was achieved by varying the flow of humid nitrogen directed at the fiber tip. Even though a flow rate /size dependence was not accurately determined at this stage of the study, by varying the air flow between 100 and 300 L/h it was possible to obtain honeycomb arrays with cavities ranging from 2.5 to 1.0 µm (external diameter).

Once the optical fiber facet was covered with the polymeric honeycomb structure, the quality of the pattern in the center of the fiber was checked by fluorescence microscopy, so that well-ordered and defect-free arrays with the desired dimensions were selected for the following step. The second fabrication step consisted in the vacuum evaporation of a thin layer of Au over it. To this purpose, the fiber was placed in a vacuum evaporation chamber and 30–40 nm of Au were deposited on top. By this two-step procedure, prototypes consisting of a single mode optical fiber end-coated with ordered metal-dielectric crystals were successfully produced (Fig. 11.1, final sketch).

11.4 Morphological Characterization

A complete morphological characterization of a representative sample achieved by means of a flow of humid nitrogen of 300 L/h is, here, reported and discussed.

Figure 11.2a shows a scanning electron microscope (SEM) top view image of the sample in which the ceramic ferule (diameter 2.5 mm) with smoothed edge can be seen while the pattern is poorly visible. In Figs. 11.2b, c, we show a magnified SEM image and atomic force microscope (AFM) image of the structure, respectively, allowing to accurately study the topography of the patterned region. By analyzing the images, the key parameters of the periodic pattern of holes have been measured revealing a lattice period and hole diameter of 2.67 µm and 0.95 µm, respectively. Relative standard deviation (%RSD) of both geometrical parameters has been estimated to be 1.3 and 3.2%, respectively. Additionally, the AFM profile, shown in Fig. 11.2d, made it possible to measure the mean pores depth, the result being 1.78 µm (RSD of 3.5 %). Finally, in order to investigate the overall structure’s thickness and uniformity, we firstly removed the polystyrene on half the area of the ceramic ferule by using an excimer laser (operating at the wavelength of 248 nm), and then measured the structure’s height with AFM analysis. Laser fluence and repetition rate were opportunely selected in order to ablate the polystyrene layer without damage to the ceramic surface. A SEM top view image of the carved sample is shown in Fig. 11.3a where the treated region is clearly evident. In Fig. 11.3b, we report an AFM profile of the treated region edge, revealing a polystyrene height of 2.5 µm with a relative standard deviation of 2.3 %. The polystyrene thickness is evidently higher than the holes height, revealing the presence of a polystyrene uniform undercoating estimated to be 0.72 µm thick.

Fig. 11.2
figure 2

Morphological characterization via SEM and AFM images: a SEM top view image; b magnified SEM image; c AFM image of patterned region; and d AFM profile

Full size image
Fig. 11.3
figure 3

Morphological characterization via SEM and AFM images after laser treatment: a SEM top view image; b AFM profile of the pattern edge

Full size image

From these results it is evident that the breath figure technique enables us to realize metallo-dielectric structures directly self-assembled on a fiber optic tip with a highly regular pattern.

11.5 Spectral Reflectance via Numerical and Experimental Analysis

Spectral reflectance measurements were carried out by means of an easy measurement setup involving a 3 dB 1 × 2 coupler connected to a broadband light source (covering the wavelength range 1,250–1,650 nm), the fiber probe and the optical spectrum analyser. In addition, to compensate for intensity fluctuations of the source vs wavelength, the sample reflectance was normalized by using a fiber-optic reference mirror, fabricated by depositing a 160 nm-thick gold film on the tip of a standard single-mode fiber. The schematic of the characterization setup is reported in the Fig. 11.4a. In Fig. 11.4b, we show the experimental reflectance spectrum (blue line) of the selected sample, revealing a broadband peak, centred around 1400 nm.

Fig. 11.4
figure 4

a Schematic optoelectronic setup; and b Numerical and experimental reflectance spectra

Full size image

In order to investigate the physics underlying the observed spectra, we performed a numerical analysis on the basis of the results obtained from the morphological characterization reported in the previous section. The structure realized is numerically reconstructed in the form of a thin polystyrene uniform layer supporting the BF pattern as schematized in Fig. 11.5a. The whole structure is thus assumed to be covered by a conformal thin film of gold.

Fig. 11.5
figure 5

a Schematic side view of the hybrid metallo-dielectric structures; and b Computational domain

Full size image

The simulations were carried out by the finite-element method (FEM) using the commercial modeling tool COMSOL Multiphysics (RF Module). In order to numerically retrieve the reflectance of photonic crystals by FEM, a common approach requires the calculation of the scattering parameters of a slab composed of the unit cell under periodic boundary conditions. Rather than considering a (honeycomb) primitive unit cell replicated by the Blok periodic boundary conditions, we selected a non-primitive rectangular cell of the hexagonal tiling and, to reduce the computation burden, we restricted the computational domain to one quarter of this rectangular cell (see Fig. 11.5b). At the bottom side of the slab, we assumed an homogenous glass substrate (having the role of an optical fiber), while at the top side we posed an air thickness (having the role of the surrounding medium). By exploiting the crystal symmetry, the quarter of the cell is transversely terminated with two horizontal perfectly-electric-conducting and two vertical perfectly-magnetic-conducting walls, so as to simulate a normal incident plane-wave with a vertically-polarized electric field [21, 45, 46]. The input power is set to 1 W. The structural parameters used for the simulations are retrieved from the morphological characterizations. Refractive index of glass and polystyrene are supposed 1.45 and 1.58 respectively. Both losses and dispersion of gold have been taken into account [16].

The dimensional parameters used for the simulation are: holes diameter of 0.95 µm, pitch of 2.67 µm, depth of 1.78 µm, basement height of 0.72 µm whereas the gold thickness is assumed to be 33 nm.

In Fig. 11.4b, the numerical reflectance spectrum (green line) is reported and compared with the experimental results (blue line) obtained in the wavelength range 1,250–1,650 nm.

From the reported data, it can be appreciated that numerical and experimental data agree with each other, even if two main differences can be easily observed. The first is related to the reflectance values. Maximum reflectance of 25 and 50 % have been found for the experimental and numerical data, respectively. The observed mismatch can be explained taking into account the unavoidable fabrication defects in respect to the ideal structure considered in the numerical analysis. The second difference relies on the presence in the numerical spectra of narrow peaks and dips superimposed on the broadband reflectance peak, such as those located at 1,260, 1,330 and 1392 nm. By means of a careful analysis of the electric field distributions, it can be easily recognized that these spectral features are associated to the excitation of surface plasmonic resonances .

Obviously the presence of sharp spectral features, erroneously foreseen by the numerical analysis but not observable in the experimental spectra, limits the prediction capability of the unit cell based model to study and design advanced structures conceived on the optical fiber tip.

In order to improve the numerical prediction capability, we developed a numerical model offering a more faithful representation of the realty with respect to the mostly used unit cell model. The major discrepancies between the experimental scenario and the numerical simulations rely on the waveform impinging on the structure under investigation and on the finiteness of the periodic crystal. In the model, using the unit cell of the periodic structure and its symmetries, the light source is a plane wave with the direction of propagation orthogonal to the plane of periodicity and the periodic structure is assumed infinite in the plane of periodicity in virtue of the imposed boundary conditions. The developed numerical model, instead, admits a Gaussian-like mode as input light source and it is characterized by a space-limited periodic pattern. The corresponding computational domain, shown in Fig. 11.6a, involves a thin slice of optical fiber, composed of a core with a diameter of 9 µm and a reduced cladding with diameter of 16 µm terminated by a perfect electric conductor (PEC) condition. This optical fiber supports a single mode clearly resembling the fundamental mode of the standard single mode fibers . Consequently, as evident in Fig. 11.6b, showing the input mode field distribution superimposed on the computational domain, in the numerical simulations, most light content impinges on a restricted area close to the optical fiber core which is the only one effectively illuminated in the experimental scenario.

Fig. 11.6
figure 6

a Computational domain of the optical fiber based model; b Normalized electric field distribution of the input light source superimposed to the periodic pattern onto the optical fiber; c Experimental reflectance spectrum in comparison with the numerical reflectance retrieved by the optical fiber based model

Full size image

In Fig. 11.6c, we show the numerical reflectance spectrum (green line) retrieved using the “optical fiber based model ” in comparison with the experimental one (blue line) demonstrating a better agreement when compared with the unit cell model.

Having improved the numerical prediction capability , in Fig. 11.7 we show the electric field distribution at different wavelengths, representative of the investigated spectral range. At a first glance, the normalized electric field appears quite similar to that observed in a simple interferometric structure, featured by sinusoidal behavior along the propagation direction. However, remarkable differences are evident in correspondence to the metal-dielectric interfaces and in correspondence to the holes pattern, where a significant field localization can be clearly appreciated. In this regard, the Fig. 11.7 also shows the top view of a slice positioned in the middle of the patterned region, where field localization can be easily noticed especially close to the gold disks and at the interfaces between the thin layer of gold and the surrounding dielectric.

Fig. 11.7
figure 7

Normalized electric field distributions at the wavelengths 1,250, 1,400 and 1,550 nm and top view of the normalized electric field distributions in a middle slice at the same wavelengths

Full size image

These considerations lead us to envision the interplay taking place between the classical interference mechanism and the plasmonic excitation occurring at the metal interfaces.

11.6 Sensing Characteristics

All the results so far reported open up very intriguing scenarios for the development of a novel generation of miniaturized and cost-effective fiber optic nanoprobes useful in many applications including physical, chemical, and biological sensing. In this section we report on experimental results aimed at investigating the refractive index sensitivity of the realized sample. Since the physical mechanism at the basis of the realized probes relies on the interplay between interferometric and plasmonic effects , a significant refractive index sensitivity is expected especially due to the local field enhancement occurring at the metal interface.

In order to assess the sensitivity of the fabricated samples to the Surrounding Refractive Index (SRI), the optical fiber probes were immersed in different liquids with known refractive indices and the corresponding reflectance spectra were thus measured. To this aim, water (n = 1.333), ethanol (n = 1.362), isopropyl alcohol (n = 1.378), and some solutions with different concentration of water and ethanol were used as external medium. In Fig. 11.8a, the spectra of the investigated sample for different SRIs are shown. It can be noted that the maximum reflectance value keeps quasi unchanged, when the external medium changes from air to water, while it monotonically increases for higher SRIs. A red-shift of the spectral curves with increasing values of the SRI can be clearly appreciated. In Fig. 11.8b we plot the wavelength of the reflection minimum as a function of the SRI, revealing an impressive sensitivity of 2,300 nm/RIU in the refractive index range of 1.33–1.378.

Fig. 11.8
figure 8

SRI analysis: a Experimental reflectance spectra as function of the surrounding refractive index; b wavelength of the reflection dip versus SRI

Full size image

Despite the enhanced sensitivity for SRI slightly higher than 1.33, the SRI sensitivity results are quite low in correspondence to the air as a surrounding medium. This behavior resembles the typical sensitivity curves exhibited by most of the evanescent wave sensors and thus is consistent with the plasmonic nature of the field enhancement occurring at the metal interface, while the interplay between interferometric and plasmonic effects can explain the exceptional refractive index sensitivity here observed.

It is worth highlighting that the reported sensitivity is significantly higher than that demonstrated in the past concerning Local Surface Plasmon Resonance (LSPR)-based fiber tip probes (typically hundreds of nm/RIU) [16, 27, 30, 31]. Moreover, the results obtained are significant higher than those reported in an excellent review by Van Duyne et al. [47] concerning LSPR sensors implemented in “planar” configurations (reporting sensitivities ranging from 200 nm/RIU to 538 nm/RIU).

11.7 Conclusions

We have demonstrated a fabrication process which enables the integration of metallo-dielectric crystals directly onto optical fiber tips by using a self-assembly approach. The presented fabrication technique relies on the use of the Breath Figure methodology directly operating onto the optical fiber tip by adopting a judicious fiber preparation. In order to adapt the self-assembly procedure, allowing to correctly operate onto non-conventional substrates as the case of optical fibers, we specifically prepared a customized Aluminum holder, designed to house the optical fiber by exactly fitting the ceramic ferule external diameter so as to enlarge the available deposition surface, facilitating the formation of regular and ordered lattice structures in controlled conditions. Once created the holes pattern in the polystyrene layer can be used also as a template for successive depositions of functional materials, able to confer advanced functionalities to the final device.

It is remarkable that, for the first time to the best of our knowledge, the proposed approach enables the creation of periodic structures on an unconventional substrate such as the optical fiber end facets by using a self-assembling approach offering a high production throughput without using sophisticated and expensive technologies.

The effectiveness of the proposed technological process has been confirmed through the realization of specific optical fiber platforms based on the integration of hybrid metallo-dielectric crystals onto the optical fiber tip. The miniaturized fiber-tip devices are composed of a thin polystyrene layer patterned with a hexagonal lattice of holes, then further covered by a conformal layer of gold at nanoscale. Specifically, we selected gold as functional material in order to confer plasmonic features and to judiciously exploit them for sensing applications.

A deep morphological analysis has been carried out by means of SEM and AFM tools, in order to retrieve detailed information on the geometrical composition of the realized samples as well as to demonstrate the successful creation of well-ordered self-assembled lattice structures onto optical fibers.

Paralleling the experimental activities, we also developed a numerical tool able to design arbitrary structures conceived to be integrated onto optical fiber end facets, by taking into account both the field distribution of the fundamental mode in single mode optical fiber as well as the finite size of the crystal structure. As far as the production of structures on the optical fiber tip is finding increasing interest in the scientific community, the availability of a numerical tool able to faithfully describe and predict the spectral features of the realized samples can be considered as a beneficial side effect of this activity, useful also for further development in the design of all fiber devices based on the integration of functional materials at nano scale.

The results of the full-wave numerical analyses are in agreement with the experimental data and also demonstrated that the spectral features of the realized sensing probes are essentially dominated by interferometric effects significantly assisted and modified by plasmonic interactions occurring at the metal interfaces.

In order to show the functionalities of the realized optical probes, we focused our attention on investigating the sensing performances. Taking advantage of the plasmonic character of the proposed nanoprobes , we investigated the sensing performances in terms of sensitivity to the surrounding refractive index, revealing unprecedented performances (as high as about 2,000 nm/RIU) when compared to other plasmonic nanoprobes realized on the optical fiber tip using traditional nanotechnologies. The excellent sensitivities versus SRI changes make these structures promising candidates for novel miniaturized label free biological nanoprobes.

We point out that no attempts at this stage have been made to optimize the platform performances in terms of refractive index sensitivity. However, by exploiting the large set of degrees of freedom offered by the composite metallo-dielectric structures, combined with the versatility of the proposed fabrication technique, further optimization margins exist.

The achieved experimental results can be considered as a significant step forward in the Lab on Fiber technology roadmap, demonstrating the feasibility of the proposed technological approach to attain advanced miniaturized sensors by exploiting an easy and low cost fabrication process, suitable to be employed in rapid prototyping and mass production. Additionally, since the technique is intrinsically suitable to create periodic pattern on multiple fibers simultaneously, further efforts will be devoted to:

  • enlarging the set of employable materials in order to add new functionalities to the final device

  • demonstrating the effectiveness of this method for high throughput fabrication

  • changing the pattern features to obtain regular cavities patterns with pitches down to hundreds of nanometers, enabling a better use of enhanced light matter interaction occurring in confined domains.