1. Introduction

Together with the scientific work on technology and applications of fiber gratings, much effort has been spent to find methods of photosensitivity enhancement of optical fibers for gratings writing purposes. Various approaches have been proposed, such as hydrogen treatment, strain applying, co-doping. Other methods of shaping the fiber profiles have been also developed for special purposes. Generally, the photosensitivity of optical fibers can be defined as susceptibility to permanent refractive index change induced by the UV radiation [1].

Photosensitivity enhancement by hydrogen treatment is mainly used for low germanium content optical fibers, e.g. SMF-28 [2]. The most popular is room temperature hydrogen loading of optical fibers under the high-pressure (between tens to hundreds of bars). When such a hydrogen soaked standard telecommunication fiber is UV irradiated, the increase of the core refractive index by 0.011 can be obtained [3]. However other variants of hydrogenation have been also developed, such as high temperature hydrogen treatment of preforms [4], flame brushing [5] or OH flooding [6].

The role of strain applied during gratings inscription has been also analyzed. Depending on the fiber type (especially doping level) as well as tensile forces very different results have been obtained in dynamics of Bragg grating spectrum formation [7]. However decrease in type I grating saturation index has been observed in strained fibers with respect to unstained ones [8]. However, when large strain was applied to different fibers (standard, highly Ge and Ge/B doped) during UV irradiation a significant increase in photosensitivity was reported [9]. On the other hand, for B/Ge-co-doped fibers where stresses are formed due to the large difference between the core–cladding compositions, UV exposure relieves these stresses, contributing with according sign to the refractive index changes [10,11].

Another important technique for photosensitivity shaping in optical fibers/preforms is co-doping. In the case of standard fused silica optical fibers, the GeO₂ dopant in fiber core is associated with germanium base defects, and thus absorption band at 240 nm [12]. The number of these defects (and thus photosensitivity) increases with germanium concentration in the fiber core. Typical GeO₂ concentration in SMF-28, the most often used optical fiber, is as low as c.a. 3.5%mol, thus does not allow for obtaining FBGs with relatively high or even moderate reflectivity [13]. Therefore, the most intuitive method of photosensitivity enhancement is to increase the GeO₂ concentration in the fiber core. It has been shown, that absorption at 242 nm in a preform increases with the germanium concentration [14]. A similar relation is depicted in [13], where maximum induced refractive index changes and maximum reflectivity of FBGs strongly depends on GeO₂ concentration in the fiber core. Moreover, it can be assumed that the linear dependence of the thermo-optic coefficient on the GeO₂ dopant concentration leads to the temperature sensitivity improvement of FBGs written in highly Ge doped optical fibers [15].

Co-doping silica fibers with B₂O₃ and GeO₂ results in high photosensitivity enhancement [16]. However the main drawback of boron co-doping is a lower thermal stability of in-written grating compared to only GeO₂ doped fibers [17].

It was also found that, when the GeO₂ fiber is Sn doped, the 240 nm absorption band is slightly shifted towards 250 nm, and thus increases photosensitivity for 248 nm excimer lasers. Another advantages of Sn co-doping compared with B₂O₃ are better thermal stability and much less effect on fiber loss at the important telecommunication window of 1.55 µm [18]. Additional Na₂O co-doping is used to increase the concentration of SnO₂ while avoiding crystallization [19].

Although the fluorine doping decreases absorption band at 240 nm [20], and thus photosensitivity, it has been used for cladding refractive index increasing of fibers with photosensitive core and cladding [21]. Similarly, both aluminum and phosphorous dopants added to the GeO₂ silica fibers reduces photosensitivity at 240 nm. However they also produces absorption bands centered at 220 nm and 210 nm respectively [20]. Moreover both P and Al doping is often used to reduce the formation of clusters in active germanosilicate fibers with high concentration of rare-earth-ions for lasing applications [22]. Furthermore, the photosensitivity of preforms with Er³⁺/Al doped silica core and Al/P doped cladding have been proposed [23].

Not only the selection of particular dopants and their concentration, but also modification of dopant profile can be used for optimizing photosensitivity and optical fibers propagation properties. Gradient profiling of optical fibers has been first proposed by Bell Labs., by the modified chemical vapor deposition (MCVD) process [24], where the core material, of germanium-doped silica composition, has been deposited in many layers. The index of refraction has been graded by varying the composition of the layers in a prescribed manner during deposition [25].

Lyytikäinen et. al. have shown that dopant profile change due to diffusion can be induced during drawing in both Ge and F-doped silica fibres [26]. In turn, Gibson et. al. have demonstrated that the refractive-index profile of a commercially available silica-based optical fiber can be accurately reconfigured. The process relies on the controlled relocation of the silica glass dopants across the fiber cross section through heat treatment and the accurate measurement of the resulting dopant redistribution with scanning probe microscopy and differential etching techniques [27]. Elsmann et al. also proposed sapphire-derived all-glass optical fiber with parabolic core profile determined by the aluminium content for high temperature resistant fiber Bragg gratings inscription. However the fiber exhibit multimode guidance behaviour, that limits its usefulness in temperature sensing applications [28].

Another method for photosensitivity shaping is based on developing optical fibers with multi-layered cross-sections. For example a kind of double-layer core erbium doped photosensitive silica optical fiber (DLC-EDPF) has been proposed, whose core has been made of two layers: photosensitive layer and erbium doped layer. The double-layer core design can perfectly overcome the difficulties in fabrication of single-core design EDPF through MCVD method combined with solutions doping technology, where high erbium and germanium doping is the cause of undesirable clustering [29].

A multi-layer approach has been used in SiO₂–ZrO₂ nanostructured optical fiber developed by means of chemical sol-gel method, where zirconia nanocrystals are dispersed inside an amorphous silica core. An optical fiber obtained from eight silica-zirconia sol layers deposited on the inner wall of the preform exhibit the core fiber refractive index with gradient profile [30]. Similar refractive index profile has been further obtained in SiO₂/SnO₂ based nanostructured optical fibers fabricated using two ways based on the chemical sol-gel method: the “inverse-dip coating” and the “powder in tube” processes [31]. In the case of above fibers the term of “nanostructured” is associated with nanoparticles with the size of few nanometers used as dopants.

A novel fabrication technology for nanostructured graded index preforms, optical fibers and micro-optical components, based on the stack-and-draw method used for photonic crystal fibres has been proposed [32,33]. The main advantage of proposed method is breaking the limit of MCVD technique. i.e. the axial symmetry, and the possibility of obtaining elliptical gradient index structures [34], and micro-optics components, such as gradient index microaxicons, suitable for integration with optical fibers [35]. Recently, we have also proposed active and passive fused silica based Ge doped nanostructured optical fibers for gratings inscription purposes [36–38].

It is worth to mention, that progress in photonic crystal fibers (PCFs) / microstructured optical fibers (MOFs) optimization in terms of fiber Bragg grating inscription purposes is also noticeable. Bragg gratings and long-period gratings have been successfully recorded into microstructured optical fibers containing Ge-doped silicate core using standard ultraviolet laser sources and in several MOFs of different geometry and dopants, including excitation of cladding modes [39–41]. However the capillary structure surrounding the solid guiding core is the cause of beam scattering for side-illuminations during grating recording [42,43].

Among all wide range of photosensitive optical fibers that have been developed in recent decades, those that have been commercialized and are commonly used for fiber gratings inscription are step-index highly Ge doped or Ge/B doped fibers [44,45]. However in highly Ge doped fibers, with the increase of germanium doping level of the fiber core, its refractive index rises as well, and the reduction of core diameter is needed to maintain single-mode operation, which impairs their compatibility with commonly used SMF-28 based devices and components [44].

To maintain the compatibility with SMF-28, photosensitive GeO₂ doped fibers are additionally co-doped with B₂O₃ that reduces the refractive index of the fiber core. Moreover it has been proven that Ge/B doped fibers are even more photosensitive than only Ge doped fibers [45]. However, the main drawbacks of B₂O₃ co-doped fiber are their high attenuation and reduction of thermo-optic coefficient, limiting its applications in temperature sensing and long-distance sensing or transmission systems [45,46].

Recently improved FBG inscription efficiency in nanostructured optical fibers with parabolic core refractive index profile has been observed in comparison to SMF-28 fiber with similar (even 20% higher) GeO₂ concentration [38]. However, due to the insufficient amount of various optical fiber samples, the reason for better gratings growth and their higher reflection efficiency has not been clearly identified.

In this paper we experimentally analyze the role of radially shaping the GeO₂ spatial distribution in the fiber core as an efficient method for photosensitivity tailoring. For this purpose, FBGs inscription process in specially designed and fabricated optical fibers, i.e., in effectively parabolic profiled core with low Ge content nanostructured graded index and highly Ge doped multi-step index optical fibers was investigated and analyzed in detail. The scope of research also includes the study of commercially available step-index SMF-28 and photosensitive optical fibers as references for comparison purposes. Temperature responses of FBGs written in optical fibers with radially shaped GeO2 concentration (and thus refractive index) in the core were also examined to verify their application potential for sensing purposes.

2. Investigated optical fibers

The optical fibers analyzed within this paper were nanostructured graded-index (nGRIN) fiber and multi-step-index (MSIN) fiber. A common feature of these optical fibers is their variable and axisymmetric refractive index profile in the fiber core. In both cases, the variability of refractive index is obtained by modification in amount of GeO₂ concentration within the fiber core. It is known that GeO₂ not only increases the refractive index of SiO₂ based fibers, but also increases their photosensitivity [1]. Playing with concentration and distribution of GeO₂ in the fiber core is then a method for tailoring the photosensitivity of such a fiber. Therefore, to the best of our knowledge, FBGs inscription process in optical fibers with radially shaped photosensitivity was examined for a first time.

Photosensitivity shaping here, i.e. in nGRIN and MSIN fibers is a result of intentional cumulation of higher GeO₂ doped glass in the center of the core. Intended GeO₂ dopant distribution in the fiber is considered as a method for enhancement of FBG reflectivity due to two factors: locally increased refractive index modulation Δn of FBG and better overlap of propagating core mode with grating transverse profile. Because the parameter known as “photosensitivity” is closely related to concentration of photosensitive ions in the bulk glass resulting in appropriate Δn change after UV exposition, the enhancement in FBGs spectral response should not be confused with photosensitivity enhancement itself. Photosensitivity tailoring should be then understood as shaping the GeO₂ distribution in the fiber core.

2.1 Optical fibers with radially shaped profile of core refractive index

The first analyzed optical fiber was fabricated by means of recently invented nanorods based nanostructuration method [33]. This fiber is characterized by a discrete core structure, which has effectively gradient refractive index profile, achieved by proper distribution of GeO₂-doped and pure silica nanorods (Fig. 1) [37]. Effective refractive index profile yields from the effective medium approximation (EMA) [47], as long as the nanorods in the fiber core have a size smaller than λ/5. For considered fiber at 1550 nm the relation between wavelength and nanorods size is better than λ/8. This means that the refractive index in the fiber core is effectively continuous, but its chemical composition is discrete. Doped nanorods are distributed nonuniformly in the core, but the concentration of each GeO₂ doped nanorod is constant and equals 4.9% mol. Since such a dielectric material is exposed on UV radiation the refractive index increase of constant value is induced, while the RI of pure silica nanorods remain the same. Again effective refractive index profile will be achieved accordingly with EMA, resulting in parabolic profile increased approximately with a constant value. It can be then said, that the fiber photosensitivity profile is also varied, having the same parabolic distribution. The maximum GeO₂ content in the core structure is about 4.9%mol and is located in the central part of the optical fiber. The concentration decreases parabolically as it moves away from the center of the fiber, reaching 0%mol (pure silica) at the core/cladding interface. Thanks to parabolic index profile the nGRIN fiber is single-mode, as its cut-off wavelength is around 1 µm, and its mode profile matches with optical fiber used in telecommunication, i.e. SMF-28. It is worth to noting that the fiber with the same core radius and step-index profile with Ge-doping level of 4.9%mol, is few-mode and its modal characteristic is not compatible with telecom systems. Thus an average concentration of GeO₂ in the core of nanostructured optical fiber of about 2.7%mol was calculated and used as equivalent to concentration in SMF-28, which is about 3.5%mol. Although the maximum germanium dioxide concentration (in the core center) is higher than in SMF-28, due to the parabolic index distribution, an average GeO₂ doping level in the nGRIN fiber is slightly smaller than in the SMF-28. The nGRIN fiber has a cladding made of pure silica with a refractive index of 1.444 at 1550 nm and a standard diameter of approximately 125 µm.

 

Fig. 1. Nanostructured graded-index optical fiber: a) distribution of nanorods in the core (binary structure), b) scheme of effective refractive index profile at 1550 nm, c) SEM image of core area (dark areas – fused silica glass, bright areas – Ge doped silica rods).

Download Full Size | PDF

Multi-step-index (MSIN) fiber, presented in this paper, belongs to the series of HNA (high numerical aperture, NA ≈ 0.4) fibers manufactured through the MCVD method. Such a fiber has already been demonstrated for other applications [48]; however, highly doped fiber core makes it also an excellent candidate for FBG inscription. High GeO₂ concentration enhances fiber photosensitivity but also increases the risk of preform breaking due to thermal expansion mismatch stress. The mismatch stress arises at the core-cladding interface when the fabricated preform is cooled down. The MSIN fiber was specially designed – it has a simple, three-step refractive index profile, which causes the thermal stress is reduced by half, compared to the single-step profile. It exhibits single-mode operation above 1050 nm. The refractive index profile of such a multi-step-index fiber and SEM image of core cross-section are presented in Fig. 2. The key parameters are listed in Table 1.

 

Fig. 2. Multi-step-index optical fiber: a) refractive index profile (a – central area, b – middle area, c – peripheral area), b) SEM image of the fiber core.

Download Full Size | PDF

Table 1. Fundamental parameters of multi-step-index optical fiber.

View Table | View all tables in this article

The various refractive indices of the a-c concentric areas of the fiber core result from different GeO₂ concentrations in each layer, i.e., 41.8, 14.1, and 4.0%mol, respectively for the central area “a”, middle area “b” and peripheral area “c” as shown in Fig. 2 and follows approximately parabolic index profile. Although the maximum GeO₂ concentration in the center of the fiber core is as high as 41.8%mol, an average germanium dioxide in the whole core region is 18.7%mol. This value is similar to the commercially available SM1500(4.2/125) (Fibercore) step-index photosensitive optical fibre having c.a. 19%mol GeO₂ concentration. The reduced core diameter of MSIN fiber ensures single-mode operation at 1550 nm spectral range. Effective mode diameters for MSIN and SM1500 are similar, around 3 µm and 4.2 µm respectively.

It is worth to notice that using average values of GeO₂ concentration for both nGRIN and MSIN fibers for comparison with their counterparts instead of using their maximum doping levels is justified as, because of GeO₂ distribution, total amount of dopant ions in the core volume is smaller (averaged) than in step-index fibers, and, what is more, step-index fibers with this level of GeO₂ doping in the core and the same size of the cores as in nGRIN and MSIN fibers would be few- or multimode fibers, which makes them not convenient for telecom systems and most FBG based sensing applications.

The most important parameters of all investigated optical fibers are given in Table 2.

Table 2. Summary of the most important parameters of investigated optical fibers.

View Table | View all tables in this article

2.2 Reference step-index optical fibers

To compare the results for both nGRIN and MSIN fibers, the following commercially available optical fibers have been selected:

3. Experimental analysis

3.1 FBG inscription experiments and photosensitivity analysis

Before the gratings inscription, all the optical fibers were hydrogen loaded under the pressure of 135 bars for 14 days at room temperature. Both SMF-28 and nGRIN fibers due to the low average GeO₂ concentration exhibit very low intrinsic photosensitivity that should be enhanced to obtain high enough reflectivity.

In the experiment, the phase mask method was used due to its best repeatability among all the FBGs writing methods. As a UV light source, KrF pulsed excimer laser was used that operates at 248 nm. UV writing beam emitted by the laser was focused on optical fiber through a cylindrical lens with a focal length of f = 17.5 cm. A phase mask with a period of 1067 nm optimized for operation at 248 nm was used to ensure Bragg wavelengths λ_B of inscribed gratings within the spectral range of 1530–1560 nm. The width of the focused UV beam on optical fibers was L = 3.5 mm. During the gratings inscription their spectral transmission characteristics were on-line monitored and acquired with use of optical spectrum analyzer (OSA) AQ6370B (Yokogawa). As a broadband light source superluminescent diode (SLED) IPSDD1503-1114 (InPhenix) was used. All the gratings were inscribed using the same irradiation conditions, i.e. pulse energy 1 mJ and pulse repetition rate 100 Hz. The process of FBGs growth dynamics in various optical fibers consisted of the repetitive procedure of optical fibers irradiation with a series of N = 100 pulses (or a multiple of N depending on grating growth rate) and measurement of spectral transmission response using OSA with the 0.02 nm resolution. This procedure was repeated until the transmission minimum Tr_min at Bragg wavelength reached at least -20 dB, which corresponds to the reflectivity of 99%. In the case of highly GeO₂ doped optical fibers (MSIN and SM1500) the transmission spectra were measured until the sum of N = 1000 pulses was reached. Illustration of grating growth in both nGRIN and MSIN optical fibers is presented in Fig. 3. Similar results were obtained for reference optical fibers.

 

Fig. 3. Illustration of grating growth in a) nGRIN and b) MSIN optical fiber during the inscription examination.

Download Full Size | PDF

The transmission minimum (Tr_min) for FBGs written in all the optical fibers versus the number of pulses (N) are presented in Fig. 4(a). It illustrates the dynamic of gratings growth during the inscription. Similar, more intuitive relation in linear scale is depicted in Fig. 4(b), where percentage reflection coefficient of FBGs (r_max) derived from Tr_min is plotted versus number of pulses (N).

 

Fig. 4. Evolution of: a) transmission minimum (Tr_min) and b) reflection coefficient (r_max) versus the number of pulses (N) for FBGs inscription.

Download Full Size | PDF

It can be seen that for step-index (reference) optical fibers with uniform GeO₂ distribution in the fiber core, gratings growth efficiency improves as the germanium dioxide doping level increases. However for FBGs written in nGRIN and MSIN optical fibers with radially shaped core refractive index, and thus GeO₂ concentration profile, reflection coefficients increases much more rapidly than in corresponding reference step-index fibers, i.e. SMF-28 and SM1500 respectively. Therefore, it can be concluded that the improvement in the efficiency of FBGs formation in optical fibers with radially shaped refractive index in the fiber core results from the fact that both RI profile and GeO₂ dopant distribution fit well with fundamental mode profile that propagates through the fibers core with in-written gratings. We suppose that mode-grating interaction is enhanced because the strongest gratings region formed in the center of the fiber core with highest GeO₂ concentration overlaps with central part of the fundamental mode where light intensity is maximal.

In the case of FBGs written in highly GeO₂ doped optical fibers MSIN and SM1500, more than 90% reflectivity was obtain after few hundreds (i.e. 200 and 500 respectively) of UV light pulses, while for lower photosensitive fibers the number of pulses had to be an order of magnitude higher. These relation is similar when reflection coefficients reach 100%. However, the predominance of MSIN and nGRIN fibers are especially evident in the range of linear increase of reflectivity as a function of number of pulses. In particular, gratings inscription by means of 2000 pulses results in almost 3 times higher reflection coefficient of FBG written in nGRIN fiber as compared with its counterpart SMF-28. Similar advantage of MSIN is observed as compared with SM1500.

Although the Tr_min and r_max versus N illustrate the dynamics of FBGs formation, photosensitivity defined as susceptibility to refractive index changes due to the UV irradiation is usually described in the form of amplitude of refractive index changes Δn_AC in function of pulses N. To plot this relation Δn_AC can be derived from Tr_min (and thus r_max) using the formula that is determined on the basis of coupled-mode theory [51]:

 

Fig. 5. Illustration of a) amplitude of refractive index changes and b) Bragg wavelength shift of FBGs versus the number of pulses.

Download Full Size | PDF

Results presented in Fig. 5(a) show that for step-index fibers UV induced amplitude of refractive index changes linearly, increasing with the number of pulses used for FBG inscription. Additionally, the Δn_AC growth rate (or grating growth coefficient) defined as a slope of the Δn_AC=f(N) increases with the GeO₂ concentration, as shown in Table 3. However, in the case of FBGs in MSIN and nGRIN fibers, this relation is no longer valid, because the fibers reveal significantly higher grating growth coefficient when compared to step-index ones with similar average GeO₂ concentration, i.e., SM1500 and SMF-28 accordingly. This effect is especially noticeable for MSIN fiber, where for 1000 pulses, the induced Δn_AC is almost twice the value obtained for SM1500.

Table 3. Grating growth coefficient, defined as a slope of Δn_AC=f(N).

View Table | View all tables in this article

It is worth to mention that Eq. (1) allows determine the amplitude of refractive index modulation Δn_AC assuming that the transverse profile of each grating plane is uniform. This assumption is fulfilled for FBGs inscribed in fibers with uniform distribution of GeO₂ in the core cross-section as in the case of SMF-28, SM1500 and GF1B. Whereas MSIN and nGRIN optical fibers have radially varying GeO₂ concentration, therefore Δn_AC estimated from Eq. (1) for our fibers is “effective” (or averaged) amplitude of refractive index modulation. In fact Δn_AC of in-written grating is the highest in the center of the core and smaller near the core/cladding interface. Nonetheless experimentally obtained higher Δn_AC values mean that we have improvement in grating strength (and thus reflectivity). As this model results in effective values of Δn_AC for our fibers, then it is justified to use average values of concentration in terms of comparison to their counterparts.

A linear red-shift of Bragg wavelengths with the increasing number of pulses is also observed for each grating, as shown in Fig. 5(b). This is caused by the cumulative energy growth, subsequently supplied to the optical fiber core during the FBG inscription and increasing the average refractive index, and thus the effective refractive index of the fiber core n_eff. Generally, the wavelength shift also indicates the photosensitivity level of the fiber. It can be noticed that for highly GeO₂ doped MSIN and SM1500 fibers, these Bragg wavelengths change is significantly higher than for less doped fibers (for certain number of pulses), which manifests in steeper slopes of these two curves in Fig. 5(b).

3.2 Temperature response measurements

To minimize inaccuracies in temperature characterization of FBGs due to the thermal decay effects, all the gratings were thermally stabilized in 80°C for 42 hours prior to the experiment. This causes that weakly bounded hydrogen is effectively removed from the fiber core, thereby ensuring that the FBGs parameters gain long-term stability and no longer change permanently when working at elevated temperatures.

Then all the gratings were placed in the oven, and their spectra were measured at various temperatures within the range of 30°C - 80°C with 10°C step. The temperature was measured by means of K–type thermocouple, enabling better accuracy (±1.5°C) than the build-in temperature controller of the oven.

Results of the thermal test are plotted as Bragg wavelength shift Δλ of each grating versus temperature. The uncertainty of Δλ determination equals to ± 5 pm. The reference Bragg wavelengths λ_B0 were measured in room temperature 25.4°C before placing the FBGs in the oven. Experimental data with the linear fits for low Ge doped parabolic nGRIN and reference step-index SMF28 with in-written FBGs are shown in Fig. 6(a). Similar results for high Ge doped MSIN and reference step-index SM1500 optical fibers are presented in Fig. 6(b). The slopes of linear fits determine the FBGs temperature sensitivities K_T. To precisely compare obtained results for each grating, generally having different resonance wavelengths, the normalized Bragg wavelength sensitivities to temperature were calculated using formula: k_T=K_T/λ_B. Due to the fact that thermo-optic effect in silica optical fibers is dominant with respect to the thermal expansion, k_T can be expressed as normalized effective refractive index change. The results of this experiment are summarized in Table 4.

 

Fig. 6. Temperature response of FBGs inscribed in a) low Ge doped nGRIN and SMF-28 and b) high Ge doped MSIN and SM1500 optical fibers.

Download Full Size | PDF

Table 4. Temperature coefficient K_T and normalized temperature coefficient k_T of FBGs written in all examined optical fibers together with the uncertainties u(K_T) and u(k_T) of their determination.

View Table | View all tables in this article

Figure 7 presents normalized temperature sensitivities k_T of FBGs related to the average GeO₂ concentration in corresponding optical fibers’ cores. The blue line in the figure indicates linearly fitted experimental data obtained for step-index fibers having a homogenous germanium dioxide distribution in the core.

 

Fig. 7. Normalized temperature sensitivities of FBGs written in investigated optical fibers.

Download Full Size | PDF

For these fibers (i.e. SMF-28, GF1B and SM1500), temperature sensitivity linearly increases with the GeO₂ doping level, which means that thermo-optic coefficient of the fiber increases accordingly. However, in the case of fiber Bragg gratings written in both high (MSIN) and low (nGRIN) doped fibers with radially profiled core refractive index, an improvement in the sensitivity to temperature is observed. We obtained 4% higher temperature sensitivity in index-profiled MSIN fiber than in SM1500. However the uncertainties of kT estimation for both fibers are as high as obtained difference between temperature coefficients. Much lower uncertainties we obtained for low-Ge doped fibers. We estimated increase of 8% in kT for low Ge doped nGRIN fiber compared to SMF-28. Thus, based on results obtained for FBG in nGRIN fiber, we can conclude that higher temperature sensitivity is associated with the GeO₂ radial distribution in the fiber core and that nominal GeO₂ concentration (on average) is only one of the factors, which decide on the value of fiber thermo-optic coefficient. However, the issue of thermal response of FBGs inscribed in fibers with non-constant germanium distribution in the core is more complex and need to be studied in details in the future.

Nevertheless, it can be concluded, that an appropriate shaping the GeO₂ localization (and thus refractive index) within the fiber core, that matches the fundamental mode profile, gives benefits in optimization of temperature sensibility of such fibers and in-written FBG based sensors.

4. Conclusions

In this paper, we report the photosensitivity tailoring in optical fiber with a radially modified refractive index profile. In particular, we have analyzed FBGs inscription in nanostructured optical fiber with effectively parabolic refractive index profile and in multi-step fiber with the core that consists of three concentric areas, each with different GeO₂ concentration. As compared with reference step-index optical fibers having similar overall GeO₂ doping level, both radially modified optical fibers reveal significantly better efficiency in gratings inscription, and much higher UV induced amplitude refractive index changes were obtained. It confirms that we have introduced a new mechanism for improvement of grating interaction with propagating mode, which relies on an appropriate GeO₂ dopant distribution and simultaneous fiber core refractive index profiling to fit well with fundamental mode distribution that propagate in optical fiber with in-written grating.

Additionally we have verified FBG written in both radially modified optical fibers in terms of their use as temperature sensors. Similar to FBG growth dynamic tests during inscription, analyzed optical fibers with in-written gratings exhibit higher temperature coefficient than their step-index counterparts. We postulate this effect was possible due to the GeO₂ distribution shaping, especially due to the local increase germanium dioxide concentration (that causes increase of thermo-optic coefficient) in the center region of fiber cores, where mode-grating interaction is strongest.

It is worth to note, that the proposed method of improvement the grating-mode interaction efficiency is complementary to existing UV sensitization techniques like hydrogenation or B₂O₃ co-doping, and they can be used simultaneously to extremely increase both photosensitivity of the fiber and FBG spectral response, especially in the fibers which need to fulfill the telecommunication requirements. This in turn opens the possibility to obtain very short and relatively strong Bragg grating structures with slights slopes for wavelength/amplitude discrimination in sensing systems. It is also a step forward development of very high UV sensitive fibers for UV pulse single-shot FBGs writing, that exhibit relatively high reflectivity. Temperature sensitivity can be also further optimized for sensing purposes.

On the other hand the proposed technique makes the possibility for decreasing the GeO₂ doping level without losing the FBGs inscription capability, and without the need for co-doping, which allows for core-cladding refractive index contrast decreasing and can lead to the development of large mode area (LMA) fibers with the ability to FBGs UV inscription.