1. Introduction

Long period fiber gratings consist of a periodic modulation of the core refractive index in a single mode fiber (SMF) with a much longer period than the optical wavelength. This enables a co-propagating coupling between the fundamental core mode and higher order cladding modes. Each coupling of light transmitted through the core to a specific cladding mode leads to the generation of an attenuation band in the transmission spectrum.

In 1996, it was proved that it is possible to obtain both a refractometer and a strain sensor based on the attenuation band shift as a function of refractive index and strain [1].

Since that moment, many articles have been published with the aim of increasing the sensitivity of the optical fiber devices. To this purpose, three major highlights have been achieved during the last years. The first one is the dispersion turning point (DTP). Attenuation bands experience typically either a blueshift or a redshift as the grating period increases. However, for specific periods it is possible to split the optical resonance in two bands, which respectively experience a blueshift and a redshift function of the external refractive index. Gratings in this situation (the turn dispersion point [2]) have an optimal sensitivity to physical or chemical perturbations.

A second improvement can be obtained thanks to the transition to guidance of a cladding mode in the overlay [3], a phenomenon that is usually called mode transition (MT) [4]. In order to generate the MT it is necessary to deposit a thin-film over the fiber, something that was done for the first time in 2002 [5]. Adjusting the characteristics of the thin-film (i.e. refractive index and thickness), it is possible to optimize the device sensitivity [6].

The first combination of thin-film deposition with an LPFG tuned at the DTP was done by Wang et al. [7]. After that, the sensitivity was enhanced with a thin-film optimized for MT deposited on an LPFG at DTP [8], and a record sensitivity of 9100 nm/RIU was obtained [9].

A third way of increasing the sensitivity of LPFGs is by reducing the cladding diameter, typically with an etching process [10–13], and in a recent work it has been theoretically proved that a sensitivity around 143000 nm/RIU can be attained by combining adequately the three phenomena [14]. However, the experimental fitting of the phenomena seems complicated, especially due to the high precision demanded for the deposition of the thin-film.

This work is focused on the combination of the dispersion turning point and the reduction of the cladding diameter with the purpose of obtaining a maximum sensitivity without the need of a thin-film. This permits to avoid using costly deposition machines like atomic layer deposition and the cross-sensitivity of the thin-film, in case the device is used for detection of biological or chemical purposes.

Long period fiber gratings written in 80 microns diameter fibers were etched to set the turning point of low order cladding modes at the telecommunication bands. The gratings were calibrated as refractive index sensors and the results show an improved sensitivity and a reasonable agreement with theoretical calculations.

2. Experimental setup

Gratings with period 210 µm were written in a hydrogen loaded fiber from Fibercore (product code SM125(9/80), core diameter 9 μm, cladding diameter 80 μm and NA = 0.1195). This fiber with reduced diameter permits to enhance the sensitivity of the device [15], since the tail of the cladding modes spread further in the surrounding medium. Moreover, the use of an 80 μm structure permits to induce less imperfections during the etching process than in a standard fiber with 125 μm diameter. The gratings were photoinscribed point by point with a doubled argon laser; the fibers were stored for 48 hours to allow the out-diffusion of the remaining hydrogen, hence the grating stabilization. Two gratings were written for this experiment. The first one (LPFG 1) had a length of 9 mm and coupled the LP_0,6 mode at 1125nm with a resonance depth of 5.5 dB after dehydrogenation. The second one (LPFG 2) had 7.5 mm in length and exhibited a resonance of 4 dB at 1125nm. No resonances corresponding to higher order modes were observed in the wavelength range of fabrication rig. A resonance of a next lower order mode was observed at 1000nm but it could have been hardly measured because the fiber works in multimode regime at that wavelength.

The two LPFGs were subjected to the etching process. The diameter of the LPFG 1 was reduced up to the moment when LP_0,6 band at DTP was observed. To this purpose, a slow etching process was performed (the LPFG was immersed in a 10% hydrofluoric acid solution), whereas LPFG 2 was etched up to 35 μm with a faster etching process (the LPFG was immersed in 40% hydrofluoric acid – HF solution) in order to obtain the LP_0,3 DTP band.

After stopping the etching process in both fibers, the wavelength shift of the attenuation bands was monitored when immersing the sensor in different glycerol solutions. The refractive index was measured with a 30GS refractometer from Mettler Toledo Inc operating at 589.3 nm. Since the measurement range in our experiments is located in the range 1300-1700 nm, reduction in 0.012 refractive index units due to dispersion of glycol and water in the infrared region was considered [16,17].

In Fig. 1 the transmission configuration setup is depicted. It consists of a broadband source (Agilent 83437A) that launches light into one end of the fiber where the grating is written. The second end is connected to an Agilent 86140B optical spectrum analyser, which monitors the transmission spectrum.

 

Fig. 1 Experimental setup.

Download Full Size | PDF

3. Numerical and experimental results

In order to analyze the response of the device, and to design the gratings, we used the finite difference method FDM of FIMMWAVE® (Photon Design Inc.)

According to [14], the sensitivity progressively increases as the cladding diameter is reduced. Simultaneously, the DTP experiences a blueshift and the coupling coefficient increases (the attenuation band depth increases). In addition to this, the sensitivity increases if the DTP is located at longer wavelengths. Consequently, in order to meet all these requirements, gratings of period 210 µm and length 9 (LPFG 1) and 7.5 mm (LPFG 2) were written as it has been explained before. These period and lengths permit to observe deep attenuation bands respectively for the LP_0,6 and LP_0,3 modes at their dispersion turning points.

LPFG 1 was etched and the temporal evolution of the grating spectra is shown in Visualization 1. Measurements as a function of time were correlated with the numerical simulation as a function of the fiber diameter. The etching rate was 0.14 µm/min.

When the gratings are etched, the effective refractive index of the cladding mode decreases and the coupling wavelength shifts to larger wavelengths. Furthermore, the etching process increases the overlap between the core and the cladding mode, hence the depth of the resonance of this undercoupled gratings.

When etching LPFG 1, of length 9 mm, the LP_0,6 band was redshifted until the DTP was obtained at 1575 nm, as Visualization 1 shows.

After that, the LPFG 2 was progressively etched up to the turning point of the LP_0,3 coupling band. The length of the device was lower, 7.5 mm, in order to consider the progressive increase of the attenuation band depth due to coupling coefficient increase due to the diameter reduction. The evolution of the spectrum up to the LP_0,6 DTP is just the same as the one obtained with LPFG 1, except for the depth because the length of LPFG 2 is different. Also LP_0,5, LP_0,4, and LP_0,3 DTPs can be seen in Visualization 2. The etching rate, 1.25 µm/min, was higher than in LPFG 1 because 40% HF solution instead of 10% HF was used, as indicated in section 2. In Visualization 2 it can be observed that the DTP of LP_0,6, LP_0,5, LP_0,4 and LP_0,3 is progressively blue shifted at the same time the depth of the attenuation band increases up to 15 dB.

In Fig. 2 numerical and experimental results of LPFG 1 at LP_0,6 DTP and LPFG 2 at LP_0,3 DTP are shown. In addition to this, the effective indices as a function of wavelength for LPFG 1 and LPFG 2 are displayed. In LPFG 2, with a lower diameter, there is a higher separation. Two surrounding refractive indices have been analyzed: 1.0 and 1.413. The variation in the effective index is higher for the cladding modes of LPFG 2. The most interesting point is that even LP_0,6 in LPFG 2 experiences a higher variation than LP_0,6 in LPFG 1. In other words, the same mode in a structure with a smaller diameter has a higher sensitivity to the refractive index. However, the attenuation bands that can be tracked in the optical spectrum, as the cladding diameter is progressively reduced, belong to lower order modes up to the lowest one LP_0,2 [14]. It has been verified that the diameter of the optical sensor is so small that, due to bending, it is not possible to track the resonances in a reliable fashion. Consequently, we considered that LP_0,3 DTP is the lowest order band that can be used for sensing.

 

Fig. 2 LPFG 1 etched up to the LP_0,6 DTP and LPFG 2 up to DTP LP_0,3: experimental and numerical spectra and cladding mode effective indices for refractive indices 1 and 1.413. In Visualization 1 and in Visualization 2 the etching process up to LP_0,6 DTP and up to LP_0,3 DTP is respectively tracked.

Download Full Size | PDF

After etching, the devices where immersed in the calibrated index solutions indicated in the experimental section. Numerical and experimental spectra as a function of different refractive indices are shown in Fig. 3 for LPFG 1 and LPFG 2 (LP_0,3 and LP_0,6 DTP). The lower coupling band experiences a large blue shift with the rising refractive index of the surroundings, whereas the opposite is true for the right coupling band.

 

Fig. 3 LPFG 1 and LPFG 2 etched up to the LP_0,6 DTP and LP_0,3 DTP. Sensitivity to different refractive indices a) Numerical spectra; b) Experimental spectra.

Download Full Size | PDF

In order to know the sensitivity to SRI, the response of LPFG 1 and LPFG 2 is shown in Fig. 4(a). LPFG 1 was etched up to the LP_0,6 band at DTP, whereas LPFG 2 was etched up LP_0,3 DTP. The left resonance band is represented as function of refractive index. A variation of more than 100 nm can be observed in both cases. The sensitivity in the range 1.353-1398 for LP_0,6 DTP is 1343 nm/refractive index unit (nm/RIU), whereas the sensitivity for LP_0,3 DTP improves to 3166 nm/RIU. In the latter case the sensitivity increases by a factor of 2.36 compared to LP_0,6 DTP. The flat region in the simulations below refractive index 1.36 is due to the fact that both left and right bands in the DTP overlap each other.

 

Fig. 4 a) Wavelength shift of the left attenuation band for one LPFG etched up to the LP_0,6 DTP and for another LPFG etched up to the LP_0,3 DTP. b) Wavelength difference of bands in LPFG etched up to the LP_0,3 DTP. Numerical and experimental results are shown.

Download Full Size | PDF

The sensitivity can also be analyzed as a function of the separation between the left and right attenuation bands of the DTP [11,13], for those values that can be monitored within the optical spectrum analyzer (see Fig. 4(b)). Since the LP_0,3 DTP is located at shorter wavelengths it is possible to obtain more points than with LP_0,6 DTP. This is why only the results of LP_0,3 are shown. The simulations show that the left and right bands in the DTP overlap each other below a refractive index of 1.36, at this point the separation between the bands is zero.

It is remarkable to say that the sensitivity is 8734 nm/RIU. Moreover, the sensitivity could be enhanced if LP_0,2 DTP was tracked [14], as it was indicated above.

The sensitivity of 8734 nm/RIU obtained here improves the results of Ref. 11 (1309 nm/RIU in the range 1.333–1.393) and Ref. 13 (1847 nm/RIU in the range 1.335 to 1.360) by a factor of 6.7 and 4.7 respectively, and is similar to the 9100 nm/RIU giant sensitivity obtained with a thin-film coating in [9].

4. Conclusion

In this work it has been proved that it is possible optimize and increase the sensitivity of uncoated LPFG devices to similar values to those obtained with coating just by combining two phenomena: the dispersion turning point and the cladding etching technique. The results presented here overcome those obtain by a soft-etching of the cladding by more than a factor of four. In addition to this, the length of the device has been adequately selected in order to obtain an optimized coupling to low order cladding modes at the final diameter of the LPFG after etching. We believe the devices presented here will find application in important fields such as biological, chemical or environmental parameter detection.