1. Introduction

Air-silica microstructured optical fibers, also known as photonic crystal or holey fibers (HFs) are a new class of optical fibers with non conventional propagation characteristics that have been largely investigated. HFs consist of a pure silica core surrounded by a regular array of air holes that run along the length of the fiber and that are arranged in a hexagonal structure around the core [1,2]. The structure of HFs enables new possibilities for optical sensing. The most common approach consists of making the sample to interact with the evanescent field of the HF guided modes [3,4]. To do so one has to fill the holes with the sample, a gas or liquid, for example, and then the analysis or detection is carried out. In some situations, such a process may be inconvenient or impractical. That is why, two new approaches for sensing with HFs have been reported recently. One of them consists of tapering the HF, preserving the structure, to a point in which the core has sub-wavelength diameter [5]. The other alternative consists of tapering the HF and collapsing the air-holes over a localized region [6]. In both cases the tapering process is adiabatic and makes the guided mode of the HF to spread out. Thus, the tapered HF becomes sensitive to the external environment and suitable for sensing purposes.

In this paper we demonstrate, for the first time to our knowledge, a refractometric sensing device based on a uniform-waist tapered HF with collapsed air-holes. The fabrication of such a taper consists of stretching the holey fiber slowly while it is heated with an oscillating high-temperature flame torch. Such a non adiabatic tapering process causes that for certain taper diameters the air holes of the HF get collapsed. As a consequence, the core mode of the fiber, in the region where the holes are open, couples to the multiple modes of the taper waist, which is a solid multimode fiber. The beating between the multiple modes of the zone with collapsed air holes causes the transmission spectrum of the taper to exhibit several interference peaks. This is similar to what is observed in abruptly tapered standard single-mode fibers [7]. In our case, however, the interference peaks shift to longer wavelengths as the external refractive index augments. One interesting feature of the sensor presented here is that the multiple interference peaks can be used simultaneously to monitor the external index. This property combined with a fitting procedure may enable more accurate measurements. The estimated maximum resolution of the proposed device is about 1x10^-5. Such a resolution is typical for refractive index sensors based on conventional tapered or core-exposed fibers [8–11], Bragg and long period gratings [12–15], and interferometers [16–18].

The applications of the refractometric device presented here can be used for a variety of applications since many physical, biological, and chemical parameters can be known throughout the measurement of the refractive index.

2. Sensor fabrication and operation principle

The fiber employed in our experiments was fabricated in our laboratories. It is a large-mode-area HF with a few air-hole rings in the cladding and is described in more detail elsewhere [19,20]. A cross section of the untapered HF and a schematic representation of the taper are shown in Fig. 1. The diameter of the solid core is 11 μm, the average hole diameter d is 2.7 μm, and the average hole spacing (pitch) Λ is 5.45 μm. As one can see from Fig. 1, the HF consists of four full rings of air holes arranged in a hexagonal structure around the core. Such a HF has a high strength and low bending losses [21].

 

Fig. 1. Image of the cross section of an untapered HF used in our experiments (left) and illustration of a uniform-waist tapered HF (right). L₀ is the length of the uniform waist and ρ is the taper waist diameter. The outer diameter of the fiber was 125 μm and the relative hole diameter d/Λ = 0.5.

Download Full Size | PDF

To taper the HF we first inserted it into a standard single-mode fiber by fusion splicing both fibers. This allowed us to seal the ends of the HF. The length of HF was approximately 30 cm. Then, we stretched the HF while it was being heated with an oscillating flame torch. The tapering conditions, such as elongation speed, temperature of the flame, etc., are described elsewhere [22]. The length of oscillation of the torch, and also the length L₀ of the uniform waist of the taper, were set to 5 mm. We fabricated several samples with waist diameters ρ between 20 to 50 μm under similar conditions. After the tapering process the tapers were cleaved under tension. Later they were examined using a commercial atomic force microscope (AFM) operated in contact mode.

In Figs. 2(a), (b), and (c) we show, respectively, AFM images of tapers with waist diameters ρ = 50, 39, and 31 μm. It can be seen in the photographs that in the 50 and 39 μm-thick taper the air holes are still present. It is worth noting that the holey structure in both tapers is also preserved. In the 31 μm-thick taper, however, the holes are totally collapsed and the holey structure cannot be distinguished. In this case, part of the tapered section of the HF becomes a solid fiber (with infinite cladding) which can support multiple modes. However, but not all of them are necessary excited. The multiple modes of the solid section of the taper give rise to multiple interference peaks due to the beating between them. Such interference peaks are sensitive to the external environment since the propagation constant of the modes depend on it. This property can be exploited for refractive index sensing, or for sensing any other parameter that modifies the propagation constant of the interfering modes. It is necessary to mention that modal interference, as well as its sensing properties, are well known in non adiabatic tapers made of standard single-mode fibers [7,8], but the registration of changes of the interference fringes under the action of an external parameter are more difficult.

 

Fig. 2. AFM images of three tapered HFs with waist diameters of 50 μm (a), 39 μm (b), and 31 μm (c). The scan sizes of the AFM images are, respectively, 13.3, 11.9, and 3.8 μm.

Download Full Size | PDF

Note that the tapers discussed here are different to those reported in [6], since the collapsing there of the holes was made adiabatically and no interference peaks were observed in the transmission spectra of the tapers. It should be pointed out that the losses introduced by the tapering of our HFs were below 3 dB (measured at 1550 nm).

3. Experimental results and discussion

Once we tapered the fibers we proceeded to make refractive index measurements. The tapered section of the HF was immersed into Cargille oils with calibrated refractive index (calibrated at 519 nm and 25 °C). In the left-hand column of Fig. 3 we show the normalized transmission spectra of three tapered HFs immersed in Cargille oils with indexes indicated in the figure. All the spectra were normalized with respect to the maximum of the unique peak, or with respect to the maximum of the highest peak. However, the normalization was not crucial in our experiments since we monitored the position of the interference peaks. In the right-hand column of the figure, we show the position of the maxima of the peaks as a function of the external refractive index. The diameters of the tapered fibers, from top to bottom of Fig. 3, were, respectively, 39, 31, and 20 μm. It is important to point out that the devices were tested in a simple transmission measurement setup consisting of a low power LED, with peak emission at 1300 nm and 80 nm of spectral width, and a high-resolution optical spectrum analyzer. Also between consecutive experiments we cleaned the devices with acetone and then they were dried with air. This process ensured us similar conditions in all the measurements.

We can see from Fig. 3 that the 39 μm-thick taper, i.e., the taper in which the holes are not totally collapsed, exhibits a spectrum that is basically the output spectrum of the LED. This indicates that there is no interference between modes. However, the spectra of the tapers with ρ = 31 and 20 μm exhibit a series of maxima and minima. Note the thinner taper exhibits more, and sharper interference peaks that the thicker one. In these two tapers the air holes of the cladding got collapsed. The interference peaks arise owing to the beating between the multiple modes of the solid section of the taper. The beat length of such modes decrease as the taper diameter decreases since the separation between the propagation constants of the modes increase. That is to say, the interference peaks become closer in wavelength separation as ρ diminishes. For this reason the number peaks increase, and become sharper, with reducing the taper diameter, see Fig. 3. The later can be controlled easily during the tapering process [22]. It is important to point out that the multiple interference peaks also observed when we lunched light from a LED with peak emission at 1540 nm, and 40 nm of spectral width, to the tapers. Therefore, the sensor can operate with any broadband optical source (and a suitable spectrum analyzer). In this regard, the tapers discussed here are similar to biconic tapers made of conventional single-mode fibers, see [7,8].

 

Fig. 3. Transmission spectra (left) and position of the maxima of the peak or peaks (right) as a function of the external refractive index of three tapered HFs with ρ = 39 (top plots), 31 (middle plots) and 20 μm (bottom plots). The peaks are numbered to show the shift they suffered when the external index changes.

Download Full Size | PDF

We can also see from Fig. 3 that all interference peaks shift to longer wavelengths as the external index augments. The shift of the peaks is more remarkable for indexes higher than 1.440. In that range of indexes the estimated maximum resolution of the sensor was found to be around 1x10^-5, considering that the resolution of the spectrum analyzer was 2 nm. Note also from the figure that the intensity of the peaks changes with the index, but their shape remains constant. For the applications presented here this is not an issue since one monitors the position of the maxima of the peaks rather than their intensity.

Our experiments revealed that the interference peaks always appeared for tapers with diameters thinner than 31 μm. However, the position of the maxima of such peaks varied slightly. One interesting feature of tapered HFs is that the interference peaks can be monitored during the tapering process, or with the method reported in [23]. Thus, one may stop the process when the desirable numbers of peaks are obtained. We observed that the interference peaks were insensitive to temperature, but very sensitive to bending. The former property is important since temperature compensation, -a familiar problem in optical sensors, - may not be needed. Other interesting feature of the tapers discussed here are the multiple interference peaks themselves. All such peaks can be used simultaneously to monitor the refractive index of the medium surrounding the taper. This property may enable more accurate measurements if it is combined with a fitting process. For example, it is not difficult to calculate that the use four peaks instead of one may improve the accuracy of the measurements by a factor of two [24]. It is important to point out that intensity-modulated refractive index sensors based on core-exposed or tapered fibers [8–11], and wavelength-modulated sensors based on Bragg gratings [12–14], or interferometers [16–18] do not allow the simultaneous monitoring of the refractive index with several peaks. Therefore, our devices may be better in terms of sensitivity and resolution than many existing fiber-based refractive index sensors. However, a comparison, or the advantages of the sensor proposed here to, or over existing index sensor is beyond the scope of the present paper.

Finally, we would like to point out that the applications of refractive index sensors are diverse since many physical, chemical, and biological parameters, such as, liquid level, temperature, gas concentration, bacteria activity, etc., can be determined through measurements of the refractive index. Tapered holey fibers with collapsed air holes can be exploited for such applications. In addition, they can be coated with thin films made of variable-index materials which may lead to the development of novel sensors.

4. Conclusions

In this paper, we demonstrate the use of uniform-waist tapered HFs with collapsed air holes in the cladding for refractive index sensing. We sealed the HF in both ends by fusion splicing it with standard single-mode fiber. Then, the tapering was carried out by stretching the fiber slowly while it was being heated with a high-temperature oscillating flame torch. We fabricated tapers with waist diameters between 20 to 50 μm and inspected them under a commercial AFM. We found that in samples with waist diameters thinner than 31 μm the air holes of the cladding got collapsed. A section of such tapers becomes a solid multimode fiber which can support multiple modes. The beating between such modes gives rise to multiple interference peaks. These peaks are sensitive to the medium that surrounds the taper since the propagation constants of the interfering modes depend on it. As a result, a HF taper with collapsed air holes can be exploited for refractive sensing, as well for sensing other parameters. One of the features that make the sensor proposed here unique is the fact that all the interference peaks can be used simultaneously to monitor the external refractive index. This property combined with a fitting procedure may allow high-accuracy refractometric measurements. The refractometric device reported here can be useful in a variety of applications of practical interest, since many physical, chemical, and biological parameters can be known through measurements of the refractive index.

We are investigating theoretically the transmission properties of tapered large-mode-area HFs to exploit them for sensing applications. The results will be reported in the future [24].