1. Introduction

In the early 1970’s a liquid core with a solid cladding optical fiber was first developed as part of an objective to produce a low-loss optical fiber for communication systems [1,2]. More recently, the importance of having a water core photonic crystal fiber has been theoretically discussed [3]. Chemical and biological sensing of species such as those required in medical diagnostics and environmental contamination studies appears to be the most prominent application of a microfluidic channel water-core optical fiber since these species can be introduced readily into water. Obviously, the solvents that can be used are not limited to water. In many other specific applications a tailored organic host may be preferable.

Additionally, there remain numerous possibilities of developing novel active devices such as those based on the introduction of a soluble non-linear medium for potential switching and modulation applications. Water itself has a notable high non-linear refractive index [4], and other solutions include those using dyes or more complex species such as porphyrins, which can enable the introduction into solvent form of many metals, for example. The high nonlinear coefficients will be beneficial to processes such as four wave mixing, second and third harmonic generation and others [5]. In summary, some work has been published discussing the use of photonic crystal fibers, such as photonic bandgap fibers, for biosensing using liquids [3,6], but none have reported experimental results on the practical implementation of a liquid core photonic crystal fiber. In this paper, we present the development of a water-silica core guiding Fresnel fiber where a strong interaction between the liquid core and the propagating optical mode is observed.

Fresnel optical fibers are unique in that they are designed to tailor the phase scattering of light by placing air holes in appropriate positions exploiting the phase zone boundaries of the waveguide. In this case it is a cylindrical waveguide [7], though it is not limited to cylindrical geometry given the considerable flexibility in tailoring and distributing the air holes within the fiber. Characteristic of such a process, in contrast to other waveguides, is the multiple foci observed at the output of the waveguide as the particular optical fields interfere with each other as a function of distance [8]. This process also occurs within the waveguide [9]. Thus our particular Fresnel fiber design may be described as having the peculiarities of a photonic crystal fiber where there is an effective step index contribution traveling within the ring of silica around the central hole, and a photonic bandgap fiber where there is an equivalent optical mode traveling within most light in the air hole and which leads to an effective bandgap. Both fields interfere as they propagate analogous to two interfering super-modes of the structure. It is therefore important to determine how having water in the core alters these properties.

2. Water core Fresnel fiber fabrication

Figure 1 shows a cross section of a typical Fresnel fiber with only air in the core used in these experiments. It is produced by drilling air holes positioned across a preform rod according to a classical Fresnel distribution of zone plates and details of its design and fabrication can be found in previous work [8]. The central hole is slightly off centre indicating it was drilled after the cladding holes were made. In order to fill only the central hole of the Fresnel fiber and not those delineating its cladding (though one may wish to do this) a single-hole hollow core optical fiber is spliced at both ends of the Fresnel fiber using custom tailored electric arc energies and fusion times. The methodology is summarised in Fig. 2.

 

Fig. 1. SEM image of the cross-section of the Fresnel fiber.

Download Full Size | PDF

After splicing the fibers one end is dipped in doubly distilled and de-ionised water, and through capillary action the hollow cores of the two fibers are filled with water. The single-hole hollow core fiber spliced at both ends of the Fresnel fiber also permits ready observation of the water when it has filled both fiber types - the cladding holes in the Fresnel fiber make it difficult to observe the water progression within it. This is monitored from the side using a compact CCD spy camera, objective lens and imaging software. The general advantage of this approach is that any liquid or gas can enter the central hole but not the cladding holes making it easier to interpret changes in spectral properties due to changes in the core. Finally once the water is inside the Fresnel fiber, we cleave away the single-hole fiber close to the Fresnel fiber ends, leaving behind compact air-caps that introduce a minimal insertion loss when the Fresnel fiber is butted or spliced onto standard telecommunications fiber. With practice, we have been able to make these end caps consistently as thin as 10µm and believe further reductions are possible. For the purposes of this study, the end cap at one end of the fiber is removed to enable analysis of the near field properties of the fiber. Finally, given the precision that is possible with the fabrication of air-structured optical fibers, end caps with strategically position air holes that match desired air-holes in the Fresnel fiber or another photonic crystal fiber both at the core or cladding can be custom made. This will be advantageous, for example, in the development of interferometers using dual core photonic crystal fibers where one core is the sensor arm and the other is the reference arm of the interferometer. In the pictures of the fiber end face shown in Fig. 2(c) and 2(d) it is possible to see that the central hole of the Fresnel fiber is completely open, and therefore selective filling is possible.

 

Fig. 2. Fabrication process of the end caps to selectively fill the central hole of the air-core Fresnel fiber. a) Schematic of the fabrication process; b) side image of the end caps showing the length of the end cap; c) cross section image with the focal point in the central hole of the Fresnel fiber showing that central hole is completely open; d) Cross section image of the end face of the water-core Fresnel fiber.

Download Full Size | PDF

3. Fiber characterization

 

Fig. 3. Experimental setup to characterize the water core Fresnel fiber.

Download Full Size | PDF

The setup employed to characterize light propagation in the water core Fresnel fiber both in the near and far fields is show in Fig. 3. The broadband white light from a Xenon lamp is coupled into a single mode fiber and guided up to the Fresnel fiber. Through a 3-axis nanometer precision stage the light is launched into the Fresnel fiber. In order to adjust the coupling a camera is used to certify that the waveguide is being correctly excited. The lens in front of the fiber is used to take the images of the fiber modal fields. After adjusting the coupling condition the camera is removed and the light is launched in the photodetector of an optical spectrum analyser (OSA). The same procedure is applied for both fibers, pristine and water core. The lengths of the fiber samples are of 30cm. The Fresnel fiber propagation characteristics are sensitive to the launch conditions. Depending on the coupling it is possible to excite either a mode traveling only in the silica ring around the central hole (shown in Fig. 1) or a mode with peak intensity in the hole. Once the latter is excited, we observe that periodic coupling occurs between it and the ring mode consistent with interference between two supermodes of the waveguide. A simple ray tracing analogy is that of a series of Fresnel lenses running the length of the fiber. When water is inserted in the core, the refractive index of the core is raised by ~0.3 and according to the Fresnel equation for light transmission through interfaces [10], for antiresonant-like guidance, to have the same amount of light in the core, as for the pristine case, the incident light must reach the silica-water interface with larger angles from the surface normal. This suggests that the launch conditions are altered and therefore so too must the propagation condition. The measured transmission loss for the water-core Fresnel fiber is ~7dB/m at 1550nm.

A practical problem that arises is water evaporation close to the ends of the fiber, creating an air-liquid interface inside the core. Hence, at the air-liquid interface, due to surface tension, there is the formation of a concave meniscus, which is responsible for strong reflection and refraction of the launched light away from the core. In order to remove or reduce this problem, one end cap at one end of the fiber is maintained and filled with index matching liquid that diffuses into both fibers, which helps the coupling of light inside the Fresnel fiber. Although a small bubble sometimes forms between the water and the gel, the total losses are sufficiently reduced. Due to evaporation at the other end, the water stays inside the core for a period of approximately 20 hours. For practical device implementation methods for sealing the ends, various methods can be employed, including additional “post” end caps made from single mode step index fiber matched to the water-core Fresnel fiber, CO₂ laser collapsing of the ends, using thicker non-evaporative gels, and using liquid solvents with higher boiling points.

In Fig. 4 it is possible to see the axial evolution of the modal interference for two different fibers, the air-core Fresnel fiber (Fig. 4(a)) and the water-core Fresnel fiber (Fig. 4(b)), at three different positions: within, at and beyond the fiber end face. The π/6 rotation between the ring and hole modes is clear, indicating that the hole mode is analogous to an antiresonant mode within the waveguide [8]. Evidently, the field distributions for both fibers are similar and the observed asymmetry is a result of the structural asymmetry of the Fresnel fiber arising from the slightly off-centre hollow core [8]. On the other hand, a small difference in the beat length (arising from periodic interference) appears where the water-core Fresnel fiber has a longer period. This variation is expected because of the considerable increase of the refractive index of the core (from 1 to 1.3), which changes the effective index of the mode traveling through the hole (for the air core the estimated effective index is ~1,23 and for the water core ~1.39).

 

Fig. 4. Modal evolution for 1550nm of the air-core Fresnel fiber (a) and of the water core Fresnel fiber (b) showing the mode in three different axial positions of the near field of the fibers.

Download Full Size | PDF

Since the mode still travels quite clearly in the water, strong interaction of the mode with the water takes place making this an ideal sensing fiber. The air-core and water-core Fresnel fibers’ transmission band spectra are shown in Fig. 5. For the air-core fiber it is possible to observe a very broad transmission spectrum where the apparent band edges fall off at ~400nm and ~1650nm. In contrast, the water-core fiber has a transmission spectrum that is quite distinct. The strong attenuation from 400nm to 900nm is the result of several contributing factors the details of which remain under investigation. Among them, we can cite the absorption of the organic index matching gel (Methyl-Phenyl-Polysiloxane) in the 400nm region, the shorter wavelength band edge within resolution of the spectrum analyser where the signal sensitivity drops substantially below 400nm (similar to that of the air-core fiber), and the role of the small bubble between the index matching gel and water inside the core. Due to the complex structure at end of the fiber, the coupling loss becomes an important factor. It is estimated to be ~12dB. However, it also appears that the transmission bandgap has shifted to longer wavelengths as would be expected since the average index has been raised by adding water. Considering the absorption of the organic index matching gel, and the sharp change in signal transmission below 900nm notwithstanding, the overall bandgap width appears to have increased. Finally, the absorption bands associated with water, particular those around 1250nm and 1400nm [11], are readily observed (marked by the arrows in Fig. 5) indicating strong overlap of the optical mode within the core. The cutoff at ~900nm means that losses in the visible are higher than in the near IR. For applications where the absorption and emission of biomolecules fall within the visible region of the spectrum, a new fiber redesigned for these wavelengths, where the band gap is at shorter wavelengths, primarily reducing the hole spacing and adjusting the Fresnel rings accordingly, is required [9]. Since the fiber is fabricated with pure silica the transmission properties at shorter wavelengths are excellent and the gradual degradation seen at short wavelengths in germanosilicate fibers is avoided.

 

Fig. 5. Transmission bands of the air-core Fresnel fiber (a) and the water-core Fresnel fiber (b). The main absorption bands of water are observed.

Download Full Size | PDF

4. Conclusion and discussion

Light guidance in water inside a Fresnel fiber was demonstrated and advantages of the Fresnel fiber over other fibers have been discussed. Other advantages over more conventional photonic bandgap fibers includes the ability to inscribe structures such as fiber Bragg gratings into the silica ring, further enhancing sensor opportunities in these fibers [12]. Even increasing the refractive index of the core by as much as 0.3, the Fresnel regime was maintained indicating the flexibility of using the Fresnel fiber to produce liquid core fibers over the fibers where the transmission is based on total internal reflection. To demonstrate the light propagating through the core, and therefore the strong interaction between the light and the water, the absorption bands of the water were measured. This indicates that applications such as biomedical diagnostics that involve detection of samples using water as a solvent is possible. For example, this fiber can be used in the study of complex systems like bacteria colonies where their size can be monitored using scatter measurements [13]. Their oxygen uptake and generation within water can be directly measured as a function of time using water-soluble porphyrin cages [14]. Further, with appropriate design of future Fresnel fibers applications in UV transmission over reasonable lengths are possible since purified and deionised water is measured to have a transmission down to ~185nm - at such wavelengths conventional fibers photodarken rapidly.