1. Introduction

Several studies have been presented on the photosensitivity and index engineering of phosphate glasses using high intensity ultraviolet [1–3] and infrared [4] laser sources. These studies are of particular interest due to the increased number of applications of rare-earth codoped phosphate glasses in the development of high efficiency planar and fibre active components [5]; but also due to the specific micro-structure and physical properties of the phosphate glass. The pentavalent phosphorus defines the micro-structure of phosphate glass, leading to a linear-like polymerisation chain built in combination with oxygen and other ion modifiers; a micro-coordination structure which is substantially different from that of common silicate glasses [6]. The specific microscopic properties define the photosensitivity characteristics of phosphate glasses: the refractive index changes induced in those glasses are rather low (10⁻⁵-10⁻⁴) by utilizing standard ultraviolet lasers [1,2]. The photosensitivity origin of phosphate glasses has been investigated employing a number of diagnostics [1,2,4,7]; however, there is still an open question related to the sign of the refractive index changes induced in phosphate glasses using different laser wavelengths and exposure conditions.

In the past, transient and low strength gratings have been recorded in phosphosilicate glass fibres using nanosecond 193nm and 240nm laser radiation [8]. However recently, strong Bragg gratings were inscribed in a rare-earth doped all-phosphate glass optical fibre, using standard phase mask approach and 193nm excimer laser radiation, by Albert, et al [9]. The Type I refractive index changes presented in [9] are of the order of 4x10⁻⁴, while the inscribed gratings exhibited interesting thermal regeneration characteristics at relatively low annealing temperatures. Grobnic, et al, have inscribed refractive index changes greater than 1.5x10⁻³ in a similar phosphate glass fibre [10]; while Bernier, et al, have recorded negative refractive index changes in ZBLAN glass fibres [11], both of them using 800nm femtosecond radiation.

Herein, we present results on the photosensitivity and Bragg grating recording behavior of an all-phosphate glass rare earth doped optical fibre, utilizing 248nm, 500fs laser radiation. In the grating inscription results presented, non-monotonic refractive index changes are observed versus accumulated energy density dose. This specific behavior resembles that of the Type IIA (belonging to the greater family of Type In grating types [12]) photosensitivity mechanism that usually occurs during grating recording in high-Ge doped silicate glass optical fibres using deep ultraviolet laser sources; however, it has never been observed before in grating recording in soft glasses [11], including phosphate glass fibres. The results presented refer to average and modulated refractive index changes, accompanied by a specific thermal study of the inscribed gratings. Moreover, the origin of the non-monotonic refractive index changes inscribed is further investigated utilizing Knoop hardness (HK) micro-indentation measurements, performed in irradiated fibre samples.

2. Experimental

The Bragg grating inscription was performed using a 248nm KrF excimer-dye laser system, Lambda Physik EMG 150MSC, with 500-fs pulses at a repetition rate of 10 Hz. The experimental set-up used for grating recording was a double phase mask interferometer, described in detail elsewhere [13]. The first phase mask used had a period of 1084 nm, while the second had a period of 535 nm, and they are both made from fused silica and being optimized for 248nm wavelength operation. The resulting grating period recorded in the fibres exposed was 528.1nm. Before the two phase gratings interferometer, a 0.4cm long x 1cm wide rectangular, metallic aperture was used for selecting the most uniform part of the beam, while a 6cm focal length CaF₂ cylindrical lens that was placed after, was used for focusing this beam along the longitudinal axis of the fiber. The distance between the second phase mask and the fiber was 2.5cm [13]. The grating transmission was monitored online using a broadband superluminescent source and an optical spectrum analyzer.

A phosphate glass single clad, Er/Yb codoped phosphate glass optical fiber, manufactured by INO, was used for the grating inscription. The concentration of the rare earth dopants in the fiber are 0.90wt% for Er, and 7.0wt% for Yb. Energy dispersive X-ray spectroscopy (EDX) measurements of the fibre glass showed that both core and cladding areas exhibit a similar stoichiometric composition, as well as, revealed the existence of Al and Ba ion modifiers in those. These two ion modifiers were traced at higher concentrations in the fiber core area. A short length of ~10cm of the Er/Yb codoped phosphate fibre was spliced at both ends to SMF28 telecom fiber, using a standard arc fusion splicer. The splicing process resulted in an additional transmission loss of 15dB per splice, as that was measured at the wavelength of 1580nm, and the generation of a smooth and long range Fabry-Perot spectral ripple of 4dB extinction ratio.

For performing Knoop hardness measurements, the phosphate glass optical fibres were exposed using an un-modulated beam, embedded in a hard resin, and then they were side polished to optical quality until the centre of the fibre core was revealed. Knoop hardness measurements were performed using a Matsuzawa, MXT70, digital indentation microscope, and by applying a 100gf load for 20s indentation time. Under these micro-indentation conditions the HK of the pristine glass was measured to be ≈292, for ten imprinting samples.

3. Results and discussion

The average and modulated refractive index change results obtained for a grating exposure using 93mJ/cm² average energy density (0.37TW/cm² intensity at the bright fringe) are shown in Fig. 1 .Lower energy densities (by a ratio of 25%) led to observation of monotonous, photo-induced refractive index engineering process, for similar accumulated energy density doses. No evident cladding or core damage was revealed after inspecting the exposed fiber using optical microscopy, thus, no Type II heat-induced damage changes were inscribed. After the inscription was initiated a spectral peak corresponding to the Bragg reflection was observed at the spectral vicinity of 1625nm, scattering far away from the Er-doped absorption peak. The maximum refractive index changes measured are of the order of 1.2x10⁻³, manifold greater than those induced using 248nm nanosecond [1] and 213nm, 150ps exposures [3] in a similar phosphate glass matrix. For an accumulated energy density of 4.5KJ/cm² a characteristic turning point similar to that of the Type IIA photosensitivity behavior is observed in the modulated refractive index change data; while in the corresponding average refractive index changes, a significant blue shift singularity appears. At this characteristic turning point, the grating does not reach zero strength (namely becomes 3.1dB). There are two possible factors that can be related with this non-zero turning point. The first is associated with irregularities in the intensity profile of the beam (herein measured to be between ~30%), which in turn affect the progression of index engineering and the photosensitivity type along the grating length [14]. Namely, spatial irregularities in the beam intensity profile affect the accumulated energy dose at different points of the grating length; triggering positive or negative refractive index changes photosensitivity mechanisms asynchronously during exposure. The second factor is related with the underlying photosensitivity mechanisms occurring in the phosphate glass under 248nm, 500fs irradiation, and specifically the actual generation and progression of positive and negative refractive index changes. The non-zero turning point of Fig. 1 can be associated with a positive refractive index changes residue which is formed during the whole duration of the grating exposure. This specific point will be discussed in the following paragraphs.The aforementioned considerations related with the laser beam quality may be also supported by the broad shape transmission spectrum of the grating inscribed (see Fig. 2 ). The shape of the grating spectra presented in Fig. 2 indicates the recording of a rather chirped grating with possible phase imperfections due to the spatial distribution of the beam intensity and slight misalignment with the interference pattern. The final grating strength is 6.3dB approximately, over a length of 4mm, while the red-band transmission notch is believed to be related with strong phase imperfection and not with radiation induced birefringence.

 

Fig. 1 Index modulation Δn_mod (red points) and average refractive index Δn_ave changes (purple points) versus accumulated energy density, for grating exposure of the phosphate glass fibre using 248-nm 500-fs excimer laser radiation. The blue cross-points denote the exposure instances (number of pulses for fixed energy density) where Knoop micro-indentation measurements were performed.

Download Full Size | PDF

 

Fig. 2 Transmission (black line) and reflection (red line) spectra of a 4mm long Bragg grating fabricated in a phosphate glass fibre using 248-nm, 500-fs excimer laser radiation, for an accumulated energy density of 6.5KJ/cm².

Download Full Size | PDF

An annealing study was performed six months after the inscription, while the 93mJ/cm² grating was kept at room temperature and low humidity conditions. Comparative spectral measurements show that the grating remained spectrally stable after this prolonged storage period. Then, the grating inscribed was placed into a copper capsule resting onto a hot plate and was isothermally annealed up to 377°C, at increasing steps of 30min duration. Grating decay results are presented in Fig. 3 . The annealing process was interrupted at 310°C, letting the grating cool down to room temperature and measuring its relaxed spectral characteristics. Accordingly, the grating was then re-annealed to 377°C for a 30 min slot and then left to cool down at room temperature. The data of Fig. 3 show that the grating retains the 75% of its strength at the end of the annealing cycle at 377°C, thus, a significant part of the refractive index changes inscribed (see Table 1 ). The annealing data presented for the 93mJ/cm² exposure, are rather similar to those obtained for Type IIA grating recordings, exhibiting a slow demarcation trend versus increased temperature, with a minor regeneration point at 100°C. For comparison, gratings recorded using ~25% lower energy densities, were erased faster at lower temperatures; however, they exhibited prominent regeneration effects at the aforementioned annealing vicinity of 100°C (see Fig. 3), similarly to the results presented by Albert, et al [15]. This specific behavior may be associated with slow structural relaxation process of defects that are not transformed/exhausted during the Type I photosensitivity regime. Since in a saturated Type IIA grating these defect states are rather transformed/exhausted, such regeneration effects are expected to be of lower magnitude [15].The non-monotonous index engineering data presented here come to augment studies previously presented on the photosensitivity of phosphate glasses [2,3]. We have shown that an alumino-phosphate glass matrix undergoes substantial, non-monotonic Knoop hardness changes upon irradiation using 248nm [16] and 193nm excimer lasers [7]. Similarly in this work, a pristine fibre sample was exposed to 186mJ/cm² energy density 248nm, 500fs radiation, resembling the energy flux conditions occurring in the bright fringe of interference. We repeated such Knoop micro-hardness measurements here, however performed onto the exposed and side-polished, optical fibre core (see Fig. 4 ). The accumulated energy density figures employed in those fiber exposures for performing Knoop hardness indentations were determined after considering the modulated refractive index change data of Fig. 1. The targeted accumulated energy densities/number of pulses points refer to the initial stage of the exposure (where refractive index changes progress rapidly), the saturation plateau of the “Type I” regime, the turning point in grating strength, and finally, the saturation point of the “Type IIA” regime (see cross-points in Fig. 1).

 

Fig. 3 Isochronal annealing results for Bragg gratings recorded in the phosphate glass fibre, using 248-nm, 500fs laser radiation. Red circles: 93mJ/cm² energy density, 6.5kJ/cm² accumulated energy dose. Blue triangles: 78mJ/cm² energy density, 3.6kJ/cm² accumulated energy dose.

Download Full Size | PDF

Table 1. Summary of the Bragg Wavelength Shift and Erased Average Refractive Index for Different Annealing and Cooling Cycles of the Phosphate Glass Fibre Bragg Reflector for 93mJ/cm² Energy Density Exposure

View Table

 

Fig. 4 Knoop micro-hardness indentation measurements performed in the core of the exposed phosphate glass fibres, for different number of irradiation pulses, and 186mJ/cm² energy density per pulse.

Download Full Size | PDF

As the data of Fig. 4 show, the Knoop hardness of the glass increases during the early stages of exposure, while after follows a clear declining trend towards lower values. The actual decrease of the Knoop hardness due to the irradiation processing is greater than 10%. That significant change of the Knoop hardness is associated with volume densification/or dilation effects [2,7,16] and corresponding changes to Young’s modulus of the material that are dependent upon irradiation conditions [7,17].The initial increase of the hardness (which is associated with positive refractive index changes) may be attributed to the dissociation of the O-P-O bond and its transformation to either a double P = O bond or an Al-O-P of higher energy and size. Also, P-O hole centers are expected to contribute to the positive refractive index changes saturated at the early stages of the irradiation [18]. Accordingly prolonged exposures using high intensity ultraviolet radiation, induce extended de-polymerisation of the phosphate glass matrix irrespectively of defect type while increasing glass randomness; facts which have been also verified using micro-Raman spectroscopy [2,7]. Volume dilation will progress faster at the later phase of the exposure after the P-O defect transformations have been saturated. This assertion appears taking place during grating inscription (see Fig. 1), where the modulated refractive index changes turning point occurs at an energy density close to that of the saturation of the grating.

After considering the refractive index change and Knoop hardness data presented above, a hypothesis can be made for the anomalous photosensitivity behavior of the phosphate glass fiber, under 248nm, 500fs irradiation. The density decrease of the exposed phosphate glass that was revealed using Knoop hardness measurements, takes place at the bright fringes of interference, while contributing a negative refractive index change component in the grating recording process. That glass softening effect occurs in an analogous extend along the whole transversal section of the phosphate glass fibre, since both the core and cladding are of quite similar stoichiometric composition, while being subjected to energy density doses of almost the same magnitude. Subsequently, the localized density decrease and volume dilation, can lead to compressive axial stress generation in the dark fringes of the exposure [19,20]. That compressive stress generation leads to positive refractive index changes. Together with the other positive refractive index change mechanism that of the glass hardening (see Fig. 4) and P-O hole centers, saturated at the initial phases of the ultraviolet radiation exposure, they can form an overall positive refractive index changes background. This non-saturated positive refractive index background can be associated with the non-zero turning point of the Δn_mod refractive index change data [14], as well as, with the overall positive, average refractive index changes, presented in Fig. 1. The positive refractive index change mechanisms are initiated and progress at different instances of the exposure; and in parallel with the underlying volume dilation, which dominates the later phase of the exposure and drives the non-monotonous refractive index engineering. The above photosensitivity hypothesis may be interpreted as the mirror counterpart of the Type IIA of the highly doped germanosilicate fibres, where the exposed regions undergo compaction and the induced stress effects are tensile [19].

4. Summary

Summarising, we report the observation of non-monotonic photosensitivity and refractive index engineering of a phosphate glass fibre during Bragg grating recording using a double phase mask interferometer and 248nm, 500fs laser radiation. The results obtained show a photosensitivity behavior with similar growth characteristics to the Type IIA observed for exposures in high-Ge doped silicate glass fibres. The results presented are discussed according to the volume dilation model that has been proposed before for phosphate glasses.