Comparative Study of γ- and e-Radiation-Induced Effects on FBGs Using Different Femtosecond Laser Inscription Methods

This work presents an extensive, comparative study of the gamma and electron radiation effects on the behaviour of femtosecond laser-inscribed fibre Bragg gratings (FBGs) using the point-by-point and plane-by-plane inscription methods. The FBGs were inscribed in standard telecommunication single mode silica fibre (SMF28) and exposed to a total accumulated radiation dose of 15 kGy for both gamma and electron radiation. The gratings’ spectra were measured and analysed before and after the exposure to radiation, with complementary material characterisation using Fourier transform infrared (FTIR) spectroscopy. Changes in the response of the FBGs’ temperature coefficients were analysed on exposure to the different types of radiation, and we consider which of the two inscription methods result in gratings that are more robust in such harsh environments. Moreover, we used the FTIR spectroscopy to locate which chemical bonds are responsible for the changes on temperature coefficients and which are related with the optical characteristics of the FBGs.


Introduction
Fibre Bragg gratings (FBGs) are widely used in many applications as sensing elements [1,2]. They are often preferred to electrical sensors due their small size and flexible design, offering immunity to electromagnetic interference and their unique multiplexing capabilities. FBGs are wavelength selective filters that reflect light from the laser-induced periodic refractive index modulations in the core of the fibre that satisfy the Bragg condition at a specific resonant wavelength, λ Bragg , as shown below, where n e f f is the effective refractive index and Λ is the modulation period. These optical components are intended to be used either as radiation or temperature sensors in harsh radiation-ionising environments. The temperature response of the FBGs is denoted as where α 0 is the FBG temperature sensitivity at a temperature T 0 , which typically is ≈10 pm/ • C for gratings inscribed in germanosilicate fibres. For radiation sensing measurements, the most studied mechanism is the radiation-induced attenuation (RIA), which is related with the absorption bands of the material [3]. RIA is influenced by the amount of germanium (Ge) doping in the fibre core [4,5]. Increasing the dopant in the fibre core is a common technique for increasing the fibre photosensitivity, making the fibre more sensitive to ultraviolet (UV) laser modification of the refractive index for the inscription of the FBGs. An alternative is to combine exposure to UV laser light with hydrogenation to modify the optical fibre core, resulting in highly radiation sensitive Type IA FBGs [6,7]. Pure silica fibres have been shown robust performance and limited RIA effects, enabling applications for temperature measurements in radiation environments [8]. However, by minimising or removing the dopants in the silica core implies a non-photosensitive fibre, impeding or preventing FBG inscription, unless deep-UV lasers are used. An alternative solution to efficient grating inscription is the use of high-intensity ultrafast pulses derived from femtosecond lasers. The femtosecond laser-inscribed FBGs either using UV, visible, or infrared (IR) radiation, are characterised by having greater temperature stability and measurement capability close to 1000 • C. In addition, as Gusarov et al. [9] showed, when the femtosecond laser inscription is performed via a two-photon absorption process, the gratings under ionised radiation have very similar wavelength shifts, either using pure silica or heavily doped silica fibres [10]. The femtosecond laser inscription locally increases the material density, resulting in greater radiation resistance.
In addition to the RIA effect, another effect can also induce refractive index changes when glass is exposed to high-energy particles or ionising radiation, causing the radiationinduced refractive index change, which is well known as RIRIC [11,12]. One of the physical explanations regarding this effect was due to the compaction phenomenon and appearance of defect-related absorption bands via Kramers-Kronig relations [13].
There are four key femtosecond laser inscription methods. The first method is inscription using a phase mask [14,15], the second is the point-by-point (PbP) method [16][17][18], where high-intensity pulses generate periodic points in the core of the fibre, the line-by-line method [19,20] and the plane-by-plane (PlbPl) method [21][22][23][24]. The PbP method employs tightly focused fs laser pulses to form microscale voids in the centre of the core. The gratings are realised by a series of voids with high-precision positioning in the fibre's core, with a distance between two consecutive voids equal to the grating period. On the other hand, using the PlbPl method, three-dimensional refractive index planes are created with uniform modification distributed across the core with lower pulse energy. This does not affect the optical profile of the FBGs, but it affects the material modification and the interaction of the radiation with the laser-modified material. A schematic diagram showing the induced refractive index modification in the fibre core for the PbP and PlbPl is presented in Figure 1. the permanent changes of the FBGs after the exposure. In addition, we proceed a step further and give a first consideration of how these changes occur at a molecular level, influencing the temperature sensitivity of the FBGs, with respect to Fourier Transform infrared (FTIR) spectroscopy results. To the best of our knowledge, this is the first work comparing the post-irradiation effects of femtosecond written gratings on exposure to γand e-radiation. These comparative studies are necessary to understand the possibilities of optimising sensors for their possible use in harsh environments.

Inscription Details
The FBGs used for this work were all inscribed using a femtosecond laser (HighQ) operating at 517 nm with 220 fs laser pulse duration. We used standard Corning SMF28  In 2015, Morata et al. [25] studied the response of type II femtosecond laser-written FBGs using the phase mask and the PbP methods when exposed to X-ray radiation. The PbP-FBG have been found to have more stable performance when compared with the phase mask FBGs with respect to Bragg wavelength shift and amplitude variation of the gratings. This work was the first one that studied the radiation hardness of PbP-FBGs. The same group confirmed their results in 2017 [26], concluding that the specific gratings showed the best hardness to radiation environments. In this paper, we compare the PbP and PlbPl methods when the gratings were exposed to γand e-radiation, and we study the permanent changes of the FBGs after the exposure. In addition, we proceed a step further and give a first consideration of how these changes occur at a molecular level, influencing the temperature sensitivity of the FBGs, with respect to Fourier Transform infrared (FTIR) spectroscopy results. To the best of our knowledge, this is the first work comparing the post-irradiation effects of femtosecond written gratings on exposure to γand e-radiation. These comparative studies are necessary to understand the possibilities of optimising sensors for their possible use in harsh environments.

Inscription Details
The FBGs used for this work were all inscribed using a femtosecond laser (HighQ) operating at 517 nm with 220 fs laser pulse duration. We used standard Corning SMF28 fibre with an 8.2 µm core diameter, 125 µm cladding diameters, and 3% mol germanium content in the core. We inscribed gratings operating at ≈1560 nm using two different inscription methods, PlbPl and PbP. It is noted that all the inscriptions were performed through the polymer fibre coating.
For the PbP inscriptions, the pulse energy was set at ≈900 nJ, and the repetition rate was set at 1 kHz. The fibre samples were mounted on an air-bearing translation stage (Aerotech) with nanometre accuracy for accurate movement during the inscription while the laser pulse focused through a long-working distance objective, with NA 0.42. The translation speed was calculated and set according to the period of the gratings; for a fourth-order FBG with a resonance wavelength at 1560 nm, the equivalent period is ≈2.15 µm. A single pulse train was inscribed at the centre of the fibre core on the fibre axis, with a total grating length of ≈7 mm, as shown in Figure 2.
the permanent changes of the FBGs after the exposure. In addition, we proceed a step further and give a first consideration of how these changes occur at a molecular level, influencing the temperature sensitivity of the FBGs, with respect to Fourier Transform infrared (FTIR) spectroscopy results. To the best of our knowledge, this is the first work comparing the post-irradiation effects of femtosecond written gratings on exposure to γand e-radiation. These comparative studies are necessary to understand the possibilities of optimising sensors for their possible use in harsh environments.

Inscription Details
The FBGs used for this work were all inscribed using a femtosecond laser (HighQ) operating at 517 nm with 220 fs laser pulse duration. We used standard Corning SMF28 fibre with an 8.2 μm core diameter, 125 μm cladding diameters, and 3% mol germanium content in the core. We inscribed gratings operating at ≈1560 nm using two different inscription methods, PlbPl and PbP. It is noted that all the inscriptions were performed through the polymer fibre coating.
For the PbP inscriptions, the pulse energy was set at ≈900 nJ, and the repetition rate was set at 1 kHz. The fibre samples were mounted on an air-bearing translation stage (Aerotech) with nanometre accuracy for accurate movement during the inscription while the laser pulse focused through a long-working distance objective, with NA 0.42. The translation speed was calculated and set according to the period of the gratings; for a fourth-order FBG with a resonance wavelength at 1560 nm, the equivalent period is ≈2.15 μm. A single pulse train was inscribed at the centre of the fibre core on the fibre axis, with a total grating length of ≈7 mm, as shown in Figure 2. The same inscription setup was used for the PlbPl inscription method, using a pulse energy of ≈100 nJ, inscribing refractive index modification planes perpendicular to the fibre core, as presented in Figure 3. The rest of the inscription parameters were as follows: repetition rate of 5 kHz and 1000 grating periods, which is equivalent to ≈2.58 mm total grating length. A fundamental difference between the PlbPl and PbP method is required

8.2μm
Objectve lens Figure 2. Microscope image of an FBG inscribed using the PbP femtosecond laser inscription method.
The same inscription setup was used for the PlbPl inscription method, using a pulse energy of ≈100 nJ, inscribing refractive index modification planes perpendicular to the fibre core, as presented in Figure 3. The rest of the inscription parameters were as follows: repetition rate of 5 kHz and 1000 grating periods, which is equivalent to ≈2.58 mm total grating length. A fundamental difference between the PlbPl and PbP method is required for higher energy pulses for the inscription of strong gratings in the latter method. All the FBG samples were spliced with APC/FC silica pigtails and characterised in transmission, prior to subjecting them to any radiation experiments, using a broadband light source, ASE730 (Thorlabs) and optical spectrum analyser (OSA), Advantest Q8384 with 10 pm optical resolution (Figure 4a). for higher energy pulses for the inscription of strong gratings in the latter method. All the FBG samples were spliced with APC/FC silica pigtails and characterised in transmission, prior to subjecting them to any radiation experiments, using a broadband light source, ASE730 (Thorlabs) and optical spectrum analyser (OSA), Advantest Q8384 with 10 pm optical resolution (Figure 4a).  for higher energy pulses for the inscription of strong gratings in the latter method. All the FBG samples were spliced with APC/FC silica pigtails and characterised in transmission, prior to subjecting them to any radiation experiments, using a broadband light source, ASE730 (Thorlabs) and optical spectrum analyser (OSA), Advantest Q8384 with 10 pm optical resolution (Figure 4a).

Radiation Exposure Details
The gamma irradiation of the samples was performed using the 60 Co GC-5000 (BRIT, India) irradiator of the "Horia Hulubei" National Institute of Physics and Nuclear Engineering, having an irradiation chamber volume of 5000 cm 3 . The setup is described in detail in [27]. For the irradiation requirements, all samples were concentrically spooled to a diameter of about 10 cm, separately mounted between two cardboard sheets, and placed in the middle of the cylindrical irradiation chamber, 10 cm from the base of the chamber. The maximum temperature during the exposure was 32 • C, increasing by a mean slope of 0.1 • C/min, starting from room temperature. This did not affect the experiments, as we investigated the results following exposure. The dose rate was 5.5 kGy/h, which was measured with one standard deviation of 3.3%. The total accumulated doses for the four optical fibre samples were 15 kGy. The dosimetry system employed was of the Ethanol-Chlorobenzene (ECB) type with oscillometric readout, traceable at the National Physical Laboratory, by RISOE HDRL.
The electron beam irradiation was performed in several steps, using a "travellingwave" linear accelerator. The accelerator was driven by a 2 MW peak power tunable EEV M5125 type magnetron operating in the S-band (2992-3001 MHz). The optimum values of the electron beam (EB) peak current I EB and EB energy E EB to produce maximum output power P EB for a fixed pulse duration t EB and repetition frequency f EB were as follows: E EB = 5.5 MeV; I EB = 130 mA; P EB = 134 W (fEB = 50 Hz, tEB = 3.75 µs). The working parameters were adjusted to obtain a dose rate of ≈3.2 kGy/min at 60 cm from the electron exit window using for dosimetry a graphite calorimeter calibrated at Risø DTU National Laboratory for Sustainable Energy. The total accumulated doses for samples were 15 kGy.
Regarding the FBGs, non-temperature treatment or any other treatment was performed before and after the exposure to radiation.

Results
After the irradiation, the samples were characterised using the same source and optical spectrum analyser (Figure 4a), as in Section 2, and the spectra were compared with measurements made prior to irradiation. We analysed the samples in terms of reflectivity, full width at half maximum (FWHM) bandwidth, coupling coefficient (k), refractive index modification (∆n mod ), and resonance wavelength with respect to the equation below, where L is the length of the grating, m is the order of the grating, and η is the mode overlap parameter between the cladding and core region. Figure 4b,c show the spectra of the PbP and the PlbPl inscribed FBGs, respectively, before and after the exposure to γ-radiation. We analysed the grating profiles, and the results are summarised in Tables 1 and 2, for the FBG inscribed using the PbP and the PlbPl method, respectively. For the gamma radiation, we observe a negative wavelength shift of −60 pm for the grating inscribed using the PlbPl method, whereas the PbP counterpart was only 6 pm. The PbP-FBG shows a decrease in the coupling coefficient (k) and the grating reflectivity. On the other hand, for the PlbPl-FBG, following the irradiation, both the reflectivity and k increased.
Tables 3 and 4 summarise the characteristics of the FBG samples exposed with electron radiation for PbP and PlbPl inscription methods, respectively, while the spectra are presented in Figure 5a,b, before and after their exposure. Again, we observe a decrease in reflectivity and k for the PbP-FBG and an increase with the PlbPl-FBG. However, it is noted that the PbP-FBG experiences a greater impact to the electron radiation with a −224 pm wavelength shift of the Bragg peak compared to 52 pm for the PlbPl. For the calculations of Tables 1-4, we calculated η = 0.82 and the relevant equations from reference [28].

Thermal Response
After the exposure of the FBGs to electron and gamma radiation and their spectral characterisation, the samples were placed in a computer-controlled climate chamber for temperature measurements (Figure 6a). The temperature of the chamber was increased from 30 to 90 • C with steps of 5 • C, and the reflection spectrum of the gratings was measured using a Micron Optics Hyperion si155 with 1 pm resolution.

Thermal Response
After the exposure of the FBGs to electron and gamma radiation and their spectral characterisation, the samples were placed in a computer-controlled climate chamber for temperature measurements (Figure 6a). The temperature of the chamber was increased from 30 to 90 °C with steps of 5 °C, and the reflection spectrum of the gratings was measured using a Micron Optics Hyperion si155 with 1 pm resolution.
As a reference sample, one FBG inscribed using the PbP and one using the PlbPl method was characterised under the specific temperature range, and their sensitivities are presented in Figure 6. The response for each of the two gratings for the same temperature range was exactly the same with a slope of 10.22 pm/°C. The thermal response of the FBGs was anticipated to be the similar, since the temperature response of the gratings reflects the properties of the material. As a reference sample, one FBG inscribed using the PbP and one using the PlbPl method was characterised under the specific temperature range, and their sensitivities are presented in Figure 6. The response for each of the two gratings for the same temperature range was exactly the same with a slope of 10.22 pm/ • C. The thermal response of the FBGs was anticipated to be the similar, since the temperature response of the gratings reflects the properties of the material.
In Figure 7, we present the temperature responses of the electron and gamma irradiated samples, for the grating inscribed with the PlbPl and PbP method, respectively. Comparing the results in two figures (Figure 7a,b), we observe a small increase in the value of the temperature coefficient of the PlbPl-FBG (10.45 pm/ • C), whereas the PbP FBGs maintain their pre-irradiation values regarding the γ-radiation. On the other hand, the electron-radiated samples note a significant increase in the temperature coefficient. There is a 9.19% increase in value for the PlbPl-FBG and 12.91% for the PbP-FBG. In Figure 7, we present the temperature responses of the electron and gamma irradiated samples, for the grating inscribed with the PlbPl and PbP method, respectively. Comparing the results in two figures (Figure 7a,b), we observe a small increase in the value of the temperature coefficient of the PlbPl-FBG (10.45 pm/°C), whereas the PbP FBGs maintain their pre-irradiation values regarding the γ-radiation. On the other hand, the electronradiated samples note a significant increase in the temperature coefficient. There is a 9.19% increase in value for the PlbPl-FBG and 12.91% for the PbP-FBG.

FTIR Measurements
Fourier transform infrared (FTIR) spectroscopy was employed to analyse the molecular structure of the samples following irradiation. This technique evaluated the resultant spectra under changes in the vibrational modes of the molecules, where such information is used to verify and/or estimate the molecular structure by analysing and comparing the vibration in the frequency region of functional groups [29]. The source of the FTIR lies in the infrared (IR) range, and only the vibration modes that change the dipole moment of a molecule appear in the FTIR spectrum [30].
For the FTIR analysis, only the FBGs region of the fibre was used. The samples were crushed into powder and analysed using an attenuated total reflection (ATR) probe, which comprises a crystal plate and has an area of 1 mm 2 . In this region, the light source is focused, and the evanescent wave resulting from the total internal reflection on the sample is analysed.

FTIR Measurements
Fourier transform infrared (FTIR) spectroscopy was employed to analyse the molecular structure of the samples following irradiation. This technique evaluated the resultant spectra under changes in the vibrational modes of the molecules, where such information is used to verify and/or estimate the molecular structure by analysing and comparing the vibration in the frequency region of functional groups [29]. The source of the FTIR lies in the infrared (IR) range, and only the vibration modes that change the dipole moment of a molecule appear in the FTIR spectrum [30].
For the FTIR analysis, only the FBGs region of the fibre was used. The samples were crushed into powder and analysed using an attenuated total reflection (ATR) probe, which comprises a crystal plate and has an area of 1 mm 2 . In this region, the light source is focused, and the evanescent wave resulting from the total internal reflection on the sample is analysed.
To evaluate the influence of the irradiation in the samples' molecular structure, the results of the FTIR for irradiated samples were compared with non-irradiated SMF28 optical fibre. In the region of 3200 to 340 cm −1 , we have nine principal absorbance peaks, To evaluate the influence of the irradiation in the samples' molecular structure, the results of the FTIR for irradiated samples were compared with non-irradiated SMF28 optical fibre. In the region of 3200 to 340 cm −1 , we have nine principal absorbance peaks, ≈1729.86, 1608.36, 1510, 1456, 1375, 1245, 1103, 829, and 457 cm −1 . We can divide this spectrum in three regions, <500 cm −1 , 500 to 1200 cm −1 , and >1200 cm −1 . The first region is related to the Germanium (Ge) bonds such as GeO 2 and Ge-Si [31], the second region is related to silicon-oxygen bonds such as Si-OH and Si-O-Si for peaks higher than 1100 cm −1 , while the last region is related to carbon bonds such as aliphatic CH groups and CO. In Figure 8a-d, we present the FTIR spectrum of the PbP-FBG and PlbPl-FBG samples irradiated with γand e-radiation, respectively, in comparison with a reference sample that was not exposed to any radiation. of the reference spectrum. A comparison between the peak transmittance at 457 cm −1 between the γ-and e-radiation samples indicates that the electron radiation has a greater negative impact for both samples. Particularly, we note a peak decrease in the transmittance at 457 cm −1 for electron exposed samples of >12% and ≈5% respectively for the PbPand PlbPl-FBG when compared with their FTIR spectra when exposed to γ-radiation. In addition, at 800 cm −1 , when comparing the PbP-and PlbPl-FBG for the electron radiation, we see similar bandwidth and frequency peaks but difference transmittance. (c) (d) Figure 8. FTIR spectra of the irradiated (and reference) PbP-and the PlbPl-FBG samples when exposed to (a) γ-radiation and (b) e-radiation. The zoomed FTIR spectra of the exposed samples to (c) γ-radiation and (d) e-radiation.

Discussion
The comparison between the inscription methods, PbP and PlbPl, shows that the PlbPl-FBG leads to lower molecular changes after exposure to both types of radiation (electron and gamma), since the presented frequency shifts and intensities are closer to the reference sample.
In Section 5, we note a significant change in the temperature coefficient for both samples, >9% for the PlbPl-FBG and >12% for the PbP-FBG, after the exposure to the electron radiation. On the other hand, for the samples exposed to gamma radiation, the temperature coefficient change was negligible. Given the FTIR results of the PlbPl-and the PbP-FBG samples for both electron and γ-radiation, we observe a common spectrum area that seems to be more susceptible to radiation exposure. This area is at 457 cm −1 , which is related to the Ge-bonds. It is noted that the transmittance change at 457 cm −1 was higher for the PbP-FBG compared to the PlbPl-FBG. However, it seems that this area is not related with the temperature coefficient of the FBGs, since similar changes were observed for either sample exposed to electron or to gamma radiation.
A comparison between PbP and PlbPl samples exposed to γ-radiation shows similar peak intensities and bandwidth at 1100 cm −1 , which is related to the Si-bonds, and similar temperature coefficients for the two samples. Conversely, when comparing the FTIR data for the electron-radiated samples, the transmittance at 1100 cm −1 is different for the PbP and PlbPl-FBGs. Particularly, in respect to the γ-radiated samples, we had an increasing Figure 8. FTIR spectra of the irradiated (and reference) PbPand the PlbPl-FBG samples when exposed to (a) γ-radiation and (b) e-radiation. The zoomed FTIR spectra of the exposed samples to (c) γ-radiation and (d) e-radiation.
The major differences between the irradiated samples (when compared with the nonirradiated sample) are the peak intensities at the first two regions (i) <500 cm −1 and (ii) 500 to 1200 cm −1 . Starting with the PbPand PlbPl-FBG γ-radiated samples in Figure 8a,c, we observe similar changes at the spectra for the whole range except at the peaks located at 457 cm −1 . Particularly, a small shift is observed to 453.2 cm −1 only for the PlbPl-FBG with peak intensity at 72.58%, while the peak transmittance for the PbP-FBG at 457 cm −1 dropped at 64.08% compared to 83.31% of the reference spectrum. There are also small shifts and changes on peaks at 829 cm −1 and 1100 cm −1 , compared to the reference spectrum; however, these were very similar when the two FBG samples are compared. The changes in these peaks are possibly induced by the femtosecond laser radiation during the FBG inscription process.
Similarly, Figure 8b,d present the FTIR results for both samples after exposure to electron radiation. In that case, we observe significant changes at two peaks, at ≈1103 cm −1 and at ≈457 cm −1 . The transmittance peak at 1103 cm −1 with respect to the reference was shifted at 1058 cm −1 for the PbP-FBG sample and transmittance decreased to 55.66% compared to 75.12% of the reference sample. According the PlbPl-FBG sample, a frequency shift of −4 cm −1 was observed, which was significantly smaller when compared to the PbP sample, and the peak transmittance was measured at 66.42%. On the other hand, for the peak located at 457 cm −1 with respect to the reference spectrum, we observed a shift at 455 cm −1 for the PbP-FBG and at the 458 cm −1 for the PlbPl-FBG. The peak transmittance was decreased to 56.25% for the PbP-FBG and 67.9% for the PlbPl-FBG with respect to 83.31% of the reference spectrum. A comparison between the peak transmittance at 457 cm −1 between the γand e-radiation samples indicates that the electron radiation has a greater negative impact for both samples. Particularly, we note a peak decrease in the transmittance at 457 cm −1 for electron exposed samples of >12% and ≈5% respectively for the PbPand PlbPl-FBG when compared with their FTIR spectra when exposed to γ-radiation.
In addition, at 800 cm −1 , when comparing the PbP-and PlbPl-FBG for the electron radiation, we see similar bandwidth and frequency peaks but difference transmittance.

Discussion
The comparison between the inscription methods, PbP and PlbPl, shows that the PlbPl-FBG leads to lower molecular changes after exposure to both types of radiation (electron and gamma), since the presented frequency shifts and intensities are closer to the reference sample.
In Section 5, we note a significant change in the temperature coefficient for both samples, >9% for the PlbPl-FBG and >12% for the PbP-FBG, after the exposure to the electron radiation. On the other hand, for the samples exposed to gamma radiation, the temperature coefficient change was negligible. Given the FTIR results of the PlbPland the PbP-FBG samples for both electron and γ-radiation, we observe a common spectrum area that seems to be more susceptible to radiation exposure. This area is at 457 cm −1 , which is related to the Ge-bonds. It is noted that the transmittance change at 457 cm −1 was higher for the PbP-FBG compared to the PlbPl-FBG. However, it seems that this area is not related with the temperature coefficient of the FBGs, since similar changes were observed for either sample exposed to electron or to gamma radiation.
A comparison between PbP and PlbPl samples exposed to γ-radiation shows similar peak intensities and bandwidth at 1100 cm −1 , which is related to the Si-bonds, and similar temperature coefficients for the two samples. Conversely, when comparing the FTIR data for the electron-radiated samples, the transmittance at 1100 cm −1 is different for the PbP and PlbPl-FBGs. Particularly, in respect to the γ-radiated samples, we had an increasing transmittance for the PlbPl-FBG of 2% and >10% for the PbP-FBG sample. As a result, it can be said that the significant change on the thermal coefficients of the FBGs is related with the molecular changes on the silicon-oxygen bonds at the 1100 cm −1 region and not with the Ge-bonds at 457 cm −1 . In other words, it seems that both radiation types are breaking and re-establishing the bonds in different directions, particularly for Si-O. The rearrangement of the bonds can be verified by the frequency shift and bandwidth changes when compared to the reference sample. In addition, these variations on the molecular structures are related to changes in the physical properties and material responses, which includes the thermal parameters and temperature behavior. Moreover, changes in the 457 cm −1 region are possibly responsible for changes in the FBG profiles, since for all the gratings, we notice refractive index changes before and after the exposure. Particularly, we note changes on the FBG strength, bandwidth, and wavelength, which are all strongly related with the effective refractive index of the FBGs.
Finally, it is well known that the temperature response of an FBG is the contribution of the thermal expansion and the thermo-optic effect. Following the previous discussion, it can be said that the changes in the thermal coefficient of an FBG after exposure to electronand γ-radiation are principally related to changes in the thermal expansion of the fibre.

Conclusions
We performed electron and γ-irradiation directly on Bragg gratings-inscribed single mode fibres (SMF28) with ≈3% mol Germanium content, using the PbP and the PlbPl inscription method. The effect of the irradiation on the gratings according to the inscription method was studied for the first time to the authors' knowledge. The total accumulated radiation dose for all fibre samples was 15 kG for both radiation types. The grating spectra were characterised before and after the irradiation, and the temperature response of all the samples were measured using a high-resolution spectrometer and a climate chamber for a temperature range between 30 and 90 • C. The results show that the e-radiation was more damaging for the fibre compared to the γ-radiation-exposed samples for both inscription methods. Differences were observed with the profile characteristics of the gratings, such as effective refractive index, bandwidth, and reflectivity, but also through the temperature response of the FBGs. We note a temperature sensitivity increase of 12.91% for the FBG inscribed using the PbP method after exposure to electron radiation compared to 9.19% for the PlbPl grating. The gratings inscribed using the PlbPl method were shown to have slightly better radiation hardness compared to the PbP gratings. Summarising all the results, we conclude that the thermal coefficient changes of the FBGs when exposed to ionised environments are more related to changes on the thermal expansion of the fibre (core and cladding) and material changes on the silicon-oxygen bonds, while modifications on germanium bonds are related with the optical characteristics of the FBGs.