Combined Radiation and Temperature Effects on Brillouin-Based Optical Fiber Sensors

Abstract

The combined effects of temperature (from −80 °C to +80 °C) and 100 kV X-ray exposure (up to 108 kGy(SiO₂)) on the physical properties of Brillouin scattering and losses in three differently doped silica-based optical fibers, with varying dopant type and concentration (4 wt%(Ge), 10 wt%(Ge) and 1 wt%(F)), are experimentally studied in this work. The dependencies of Brillouin Frequency Shifts (BFS), Radiation-Induced Attenuation (RIA) levels, Brillouin gain attenuation, Brillouin frequency temperature (C_T) and strain (C_ε) sensitivity coefficients are studied under X-rays in a wide temperature range [−80 °C; +80 °C]. Brillouin sensing capabilities are investigated using a Brillouin Optical Time Domain Analyzer (BOTDA), and several properties are reported: (i) similar behavior of the Brillouin gain amplitude decrease with the increase in the RIA; (ii) the F-doped and heavily Ge-doped fibers do not exhibit a temperature dependence under radiation for their responses in Brillouin gain losses. Increasing Ge dopant concentration also reduces the irradiation temperature effect on RIA. In addition, Radiation-Induced Brillouin Frequency Shift (RI-BFS) manifests a slightly different behavior for lower temperatures than RIA, presenting an opportunity for a comprehensive understanding of RI-BFS origins. Related temperature and strain sensors are designed for harsh environments over an extended irradiation temperature range, which is useful for a wide range of applications.

1. Introduction

The exploitation of Brillouin scattering in silica-based optical fibers has attracted much interest in temperature and strain monitoring in harsh environments such as those encountered in the nuclear industry [1,2]. Measuring the fiber Brillouin signature with a reflectometer (either in the time or frequency domain [3,4]) opens the way to distributed temperature and strain in situ measurement along the fiber length over long distances of tens of kilometers [5]. This work investigates the sensing possibilities offered by Brillouin-based optical fiber sensors in harsh environments with combined radiation and temperature constraints. In the literature, this approach to combined temperature/irradiation effects via Brillouin sensing has never been systematically studied.

This study investigates both losses and Brillouin signature behaviors under X-rays on Germanium-doped (Ge-doped) and Fluorine-doped (F-doped) single-mode (SM) optical fibers. Ge-doped single-mode fibers (SMFs) are commonly employed for such applications, constituting the majority of deployed single-mode fibers in the telecommunications industry. F-doped optical fibers are nowadays considered as radiation-hard SMFs for environments associated with steady-state irradiation and doses exceeding 10 kGy [6], for which they present the lowest degradation due to the Radiation-Induced Attenuation (RIA) phenomenon at Telecom wavelengths. This work is performed using a Stimulated Brillouin Scattering (SBS) interrogator known as Brillouin Optical Time Domain Analyzer (BOTDA) at 1550 nm. The counter-propagation of the two-pump pulse and continuous probe along the waveguide allows Brillouin backscattering, containing the Brillouin spectrum signature, when the two counter-propagative waves match the Brillouin frequency. In fact, through the electrostriction effect [7], the stimulation of acoustic waves in the fiber core and cladding allows Brillouin spectrum measurement along the fiber length. Brillouin frequencies of the fibers (their specific and intrinsic responses) (${\nu }_B$) are dependent on both temperature and stress applied and present a cross-sensitivity of both temperature ($C_T$) and strain ($C_{\epsilon }$) coefficients for the standard types of fibers of 1 MHz/°C and 0.05 MHz/$\mu \epsilon$ [8,9], respectively, originating from the silica core glass density and refractive index profile of the fiber, which are both temperature (T) and strain (ε) dependent. The central Brillouin Frequency Shift follows relation (1):

$$\Delta {\nu }_B=C_T\Delta T+C_{\epsilon }\Delta \epsilon$$

(1)

The frequency shifts due to temperature and strain variations according to the fiber sensitivities. The central Brillouin Frequency Shift $\left(\Delta {\nu }_B\right)$ follows relation (1), where $\Delta T$ and $\Delta \epsilon$ are temperature and strain variations [10].

However, when irradiated, point defects are generated in the fiber glass matrix, leading to the creation of additional spectral absorption bands in the fiber transmission window responsible for an excess of optical losses called Radiation-Induced Attenuation (RIA) [11]. Depending on the nature of the involved optically active defects, RIA usually grows with the total ionizing dose and can decay after irradiation as some defects recover [12]. RIA also leads to a decrease in the Brillouin spectrum signal amplitude. Moreover, the Brillouin spectrum is also subjected to radiation-induced modification of its spectrum through Radiation-Induced Brillouin Frequency Shift (RI-BFS) [13]. RI-BFS leads to systematic temperature and strain errors when measuring Brillouin central frequency in a harsh environment as the BFS is used to monitor those parameters. RIA and Brillouin spectrum modifications under irradiation have been shown to depend on the fiber composition (dopant and concentration), on the operating wavelength and on environmental conditions such as the irradiation temperature [2].

The irradiation effect on Brillouin signature has been studied [14] on both Ge-doped and F-doped fibers under γ-rays at room temperature up to 10 MGy, exhibiting an RI-BFS of 4 MHz and 2.3 MHz, respectively. These frequency shifts can be interpreted as variations of 4 °C or 80 µε for the Ge fiber, which results in errors being induced in the measurements by the irradiations. In our case, no significant radiation effect on $C_{\epsilon }$ was pointed out for the tested fibers, and a decrease of 6% was noted in $C_T$ after a 10 MGy accumulated dose for the F-doped fiber. Online SBS measurements were conducted during γ-rays exposure up to a cumulative dose of 1 MGy in Ge-doped single-mode fibers (SMF) at room temperature, +80 °C, +100 °C and +120 °C. This first combined thermal and radiation effect study on Brillouin sensing revealed a dependence of the Brillouin peak amplitude decrease on the irradiation temperature.; lower losses were observed at higher temperatures [15] and showed an RI-BFS of 4 MHz, experiencing a sublinear trend. It was then not obvious that the irradiation temperature had an impact on the RI-BFS considering the uncertainties of the measurement. This phenomenon needs a deeper investigation.

RIA in Ge- and F-doped fibers has been extensively investigated under X-rays at room temperature [11]. RIA in F-doped fibers exhibits low levels and low irradiation temperature dependence [6] from −80 °C to +80 °C at 1550 nm. Similar studies have been conducted on Ge-doped fibers, exhibiting higher RIA levels and a high sensitivity to the temperature of irradiation in the same temperature range [16]. In fact, RIA and its temperature sensitivity also depend on the Ge dopant concentration [17], demonstrating a reduced sensitivity to temperature with an increase in germanium dopant concentration (>7 mol%).

This work aims to study both radiation and temperature contributions to Brillouin sensing capabilities for different fiber compositions in a large temperature range. Three types of fibers were tested in this work: a low GeO₂ dopant concentration (4 wt%) (SMF28e+), a higher GeO₂ concentration (10 wt%) and a radiation-hardened F-doped fiber (1 wt%). Samples were X-ray irradiated up to a cumulated dose of 108 kGy(SiO₂) at a dose rate of 6 Gy/s from −80 °C to +80 °C, providing optical fiber Brillouin-based sensing capabilities in a harsh environment to adapt the fiber type to the desired application.

2. Materials and Methods

2.1. Investigated Samples

To investigate the effect of the dopant nature and concentration on the Brillouin signature, three different types of single-mode fibers at 1550 nm were considered in this study: one commercial standard fiber from Corning^® (New York, NY, USA), SMF28e+, an acrylate coated and Ge doped in the core (~4 wt%) with a pure silica cladding; the second fiber type had higher Ge doping in the core (~10 wt%), named CMS, manufactured by Exail (Paris, France) [18], also an acrylate coated with a pure silica cladding; the last fiber type was a pure-silica-core fiber, polyimide coated with an F-doped cladding (~1 wt%), manufactured by Exail. Silica-based optical fibers can be classified as three types depending on their RIA vulnerability: radio-sensitive ones (P doped and Al doped), radiation tolerant (Ge doped) and radiation hardened (F doped, pure silica). In fact, the F-doped fiber type is known to be radiation hardened and exhibits low RIA levels [6] and RI-BFS levels [14] in the Telecom window for MGy accumulated dose levels. On the other hand, Ge-doped fibers are known to be radiation tolerant and highly sensitive to environmental conditions such as temperature. They exhibit higher RIA and RI-BFS levels than the F-doped fiber type.

Depending on the fiber RIA sensitivity and the temperature of irradiation, different sample lengths were used to characterize RIA and acquire the Brillouin signature. F-doped fibers 30 m in length were used for all temperatures, while lengths ranging from 5 m to 30 m were used for Ge-doped fiber samples from −80 °C to +80 °C in order to adjust the RIA levels to the optical budgets of our setup; exact values are reported in

Table 1. Sample characteristics, lengths and related central Brillouin frequencies.

Fiber Samples	SMF28e+				CMS				F Doped
Intrinsic attenuation @1550 nm	0.20 dB/km				0.31 dB/km				0.48 dB/km
Numerical apperture	0.14				0.20				0.135
Temperature of irradiation (°C)	−80	−40	+20	+80	−80	−40	+20	+80	−80	−40	+20	+80
Sample length (m)	5	15	30	30	10	30	30	30	30	30	30	30
Brillouin central frequency before irradiation (GHz)	10.705	10.757	10.803	10.857	10.134	10.175	10.225	10.280	10.974	10.995	11.049	11.114

. The fundamental Brillouin peak responses are reported in Figure 1, at distinct frequencies for the three tested fiber types. Knowing that some fibers such as SMF28e+ present additional peaks [19], this study focuses on the fundamental peak since CMS and the F-doped fibers present only one distinct peak.

2.2. Irradiation Tests

To study the combined effects of both temperature and X-ray irradiation on Brillouin signature and losses of the studied samples, irradiations up to 108 kGy(SiO₂) at a dose rate of 6 Gy(SiO₂)/s were performed at −80 °C, −40 °C, +20 °C and +80 °C. For each irradiation temperature, OTDR (Optical Time Domain Reflectometry) and BOTDA (Brillouin Optical Time Domain Analysis) measurements were performed during different runs to acquire the RIA kinetics as well as the Brillouin spectrum evolution during and after the X-ray exposure. OTDR measurements were performed with the VIAVI MTS-400 interrogator working at 1310 nm, 1550 nm and 1625 nm to evaluate RIA levels at these specific wavelengths. Brillouin spectra were acquired using a BOTDA from OzOptics operating at 1550 nm as presented in Figure 2. Optical fiber samples were coiled as a monolayer spiral of 10 cm outer diameter to homogenize the deposited dose over the entire sample length, while the temperature was monitored using a thermal plate. Such a curvature diameter does not induce losses in silica-based optical fibers, and no particular contribution of the coiling process to the measurement was expected. This thermal plate (Instec HCP204SC) stabilizes the sample temperature (within ±0.2 °C) during the irradiation run. The irradiation source consists of an X-ray tube with a tungsten target operated at 100 kV with a current of 30 mA generating photons with a mean energy fluence of ~40 keV [20].

Thermocouples (black stars) were placed on the thermal plate to control the temperature. Shielded transport fibers (SMF-28) connected the OTDR interrogator (single-ended technique) or the BOTDA (double ended) to the irradiated sample (optical fiber splices represented by red crosses).

2.3. Post Mortem Measurements

To characterize the irradiation effect on Brillouin temperature ($C_T$) and strain ($C_{\epsilon }$) sensitivities of our studied samples, measurements according to the setup presented in Figure 3a were performed on both pristine and 108 kGy(SiO₂) irradiated samples at −80 °C, −40 °C, +20 °C and +80 °C. For $C_T$ characterization, temperature steps were uniformly applied to the sample placed in a heating/cooling chamber (BINDER MK115). Multiple temperature cycles (increase and decrease) were applied at 2 °C/min for 10 min ramps followed by 1 h of stabilization before each measurement for 20 °C steps from −40 °C to +80 °C. The Brillouin spectrum was acquired every 5 min with the BOTDA interrogator to track the induced temperature BFS. Samples were loosely cooled with no constraint applied to avoid any strain contribution. It should be remembered that the BFS linearly evolves with both temperature and strain for our application range [21]: −40 °C to +80 °C and 0 µ$\epsilon$ to 2500 µ$\epsilon$.

The strain sensitivity coefficient $C_{\epsilon }$ measurement setup involved two magnetic holders placed at a distance of 3 m, maintaining the optical fiber and preventing any slippage. A force sensor was positioned at one of the holders to monitor any potential slippage of the fiber during the drawing process. A micrometric displacement of the pre-strained sample on the support enabled the monitoring of the strain-induced BFS up to 2500 µ$\epsilon$ using the BOTDA interrogator. The temperature of the room was kept constant (within ±0.2 °C) to avoid any temperature variation contribution to the BFS. In the case of irradiated samples, it is important to note that the coatings can be damaged and induce slipping of the fiber on the support [22]; the two contact points with the magnetic holder were replaced by pristine samples spliced to the irradiated one. This procedure allowed us to suppress the effect of the degraded coating [23] on the $C\epsilon$ measurement.

2.4. Data Analysis

OTDR traces were recorded before, during and after irradiation. Data analysis consisted of calculating the slope of the trace over the FUT length. The reference trace was subtracted from the measured slope over time to overcome the intrinsic attenuation of the fiber and to follow the evolution of the RIA during both the irradiation and the recovery phases.

The evolution of the Brillouin Gain Spectrum (BGS) was recorded using the BOTDA interrogator. Data post treatment consist of a Lorentzian fit of each BGS signature providing its Brillouin frequency and peak amplitude over time during and after the irradiation. In addition, thermocouples provide the temperature measurement (near the sample), which helps in correcting the measured BFS by the temperature-induced frequency shift in order to extract the RI-BFS (Radiation-Induced BFS) from the measured BFS response, which can be affected by temperature variations (~1 °C) in the irradiation chamber. Characterizing Brillouin temperature and strain sensitivity coefficients also results from the same Lorentzian BGS fitting procedure, over temperature or strain. A linear fit response to BFS with temperature or strain gives the related sensitivity of the sample.

3. Results

In this section, we present for all the studied samples the measured RIA dependence at different temperatures of irradiation. In addition, BGS behavior under irradiation at different temperatures was monitored, providing results on the BGS attenuation and BFS, as well as Brillouin temperature and strain sensitivities.

3.1. Radiation-Induced Attenuation

In this section, we present the measured RIA dependence on temperature of irradiation for SMF28e+, CMS and F-doped fibers obtained at the unique dose rate of 6 Gy (SiO₂)/s and accumulated dose of 108 kGy(SiO₂). Each fiber type was irradiated at −80 °C, −40 °C, +20 °C and +80 °C. The online measurements were carried out with an OTDR at three probe wavelengths (1310 nm, 1550 nm and 1625 nm) according to the irradiation configuration shown in Figure 2. The evolution of the RIA both during the irradiation run (5 h) and recovery phase (1 h) is presented in Figure 4 for all the studied fibers and at the three wavelengths.

RIA corresponds to the increase in the fiber attenuation caused by point defects. The F-doped fiber exhibited the lowest RIA levels at any measured wavelength, with a maximum of 30 dB/km, 17 dB/km and 5 dB/km at −80 °C, −40 °C, +20 °C and +80 °C, respectively, at 1550 nm. We can highlight that for all our samples and at all probed wavelengths, RIA levels decreased versus temperature. It is known that pure-silica-core and fluorine-doped fibers are radiation-hardened fibers exhibiting low RIA levels [24] in the NIR (Near-Infra-Red) domain. The RIA difference between −80 °C and +80 °C is about 48 dB/km, 24 dB/km and 17 dBm at 1310 nm, 1550 nm and 1625 nm, respectively. According to RIA spectral studies conducted on both pure-silica-core fiber and F-doped fibers [24,25], RIA levels increase when decreasing the wavelength in the NIR domain for a lower temperature of irradiation. Similar results were obtained from −80 °C to +80 °C, with losses equal to 43 dB/km and 6 dB/km, respectively, at 1550 nm up to an accumulated dose of 100 kGy [6]. In this same study, the limited RIA temperature dependence of the F-doped fiber is explained by the invariance of the defect generation rate with temperature. The main defect contributors are identified as two precursor sites from the drawing processes and one from the matrix [6].

The SMF28e+ fiber exhibited the highest RIA temperature sensitivity [6], with RIA level differences of 2055 dB/km at 1310 nm, 878 dB/km at 1550 nm and 736 dB/km at 1625 nm between +80 °C and −80 °C. This high RIA vulnerability increased at lower wavelengths. In fact, in the NIR region, Kashaykin et al. [17] suggested that the GeY color center is the main defect responsible for the absorption of low-Ge-dopant-concentration optical fibers such as SMF28e+. Even though this defect structure remains unknown, its concentration behavior with the temperature of irradiation has been identified in [17], showing an increase in the GeY-related attenuation with decreasing temperature. In addition, it is known that the GeY absorption band is centered at 1.38 eV (~898 nm) with a Full-Width Half Maximum (FWHM) of 0.71 eV [24]. In the NIR domain, for low-Ge-doped fibers such as the SMF28e+, RIA levels and kinetics mainly depend on the GeY defect concentration behavior, having a high sensitivity to temperature and being responsible for the high RIA temperature sensitivity. In addition, both RIA levels and their temperature sensitivities decrease at higher wavelengths due to the reduction in the GeY center contribution.

Increasing the GeO₂ dopant percentage in the fiber core has been shown to increase the GeX and decrease the GeY concentrations during the irradiation [17]. For highly Ge-doped fibers, such as our CMS fiber, the RIA depends on both GeY and GeX defects in the NIR domain. This same study showed that GeX-related RIA contribution decreases with temperature. In addition, the contribution of GeY reduces for a higher dopant concentration in the fiber core. We observed a drastic diminution of the RIA temperature sensitivity for the CMS samples in the NIR domain due to the increase in the GeX concentration and the reduced contribution of the GeY center. Results exhibited no major RIA temperature dependency on CMS samples but a higher sensitivity for higher wavelengths. In fact, at higher wavelengths, the contribution of the GeX spectral band, centered at 2.6 eV (~477 nm) with a FWHM of 0.71 eV, decreased, and the GeY band became preponderant, inducing a higher temperature sensitivity in the NIR domain.

3.2. Brillouin Gain Attenuation

BGS was acquired using a conventional BOTDA in order to perform online measurement as described in Section 2.2. The procedure outlined in Section 2.4 enables monitoring of the peak amplitude evolution throughout both the irradiation and recovery phases. Figure 5 presents the Brillouin peak amplitude over accumulated dose and time for the three tested fibers at −80 °C, −40 °C, +20 °C and +80 °C. Brillouin spectrum measurements were conducted under identical conditions to the RIA experiments, although during separate irradiation runs.

The Brillouin spectrum’s amplitude decreases with the accumulated dose. As for the RIA procedure, radiation-induced absorption bands are generated in the fiber core and cladding due to X-rays exposure. Consequently, the counter-propagating pump pulse, CW probe of the BOTDA interrogator and the backscattered signal from the sample are affected by RIA. This process results in a decrease in the BGS amplitude over the accumulated dose. The fiber composition changes the sensitivity of the sample to BGS attenuation due to the dopant-related defects’ absorption bands in the fiber core and cladding. All three tested fibers with different dopant types and concentrations presented different Brillouin gain attenuation behaviors under the same irradiation conditions. The evolution of the Brillouin peak amplitude under irradiation depends on both linear losses due to RIA effects and Brillouin central gain ($g_{B0}$). The latter depends on both the refractive index profile (RIP) and Acoustic Velocity Profile (AVP) along the fiber core and cladding [26], defined as:

$$g_{B0}=\frac{2\pi {\gamma }^2}{n_{eff}c{\lambda }_p{}^2\rho V_{A_{eff}}\Delta {\nu }_B}$$

(2)

$$V_{A_{eff}}=\sqrt{\frac{E\left(1-\nu \right)}{\left(1+\nu \right)\left(1-2\nu \right)\rho }}$$

(3)

where $n_{eff}$ and $V_{A_{eff}}$ are the effective refractive index of the guided mode and the effective acoustic velocity in the core, respectively, ${\lambda }_p$ is the pump wavelength, $\rho$ is the density of the fiber core, E is Young’s modulus, $\nu$ is the Poisson number and $\gamma$ is the electro-strictive constant in the core. The modification of the fiber matrix under irradiation by the introduction of point defects in the core and cladding results in a change of the acoustic velocity and the refractive index [27,28].

Figure 5 shows the combined effects of the dopant nature and the temperature of irradiation for the three tested fibers, which exhibited a similar behavior with RIA. The Brillouin gain of the slightly Ge-doped fiber (SMF28e+) exhibited a high temperature dependence, experiencing a gain attenuation of 32% at +80 °C and 90% loss at −80 °C compared to non-irradiated samples. The gain of the highly Ge-doped CMS fiber showed no temperature dependence up to 108 kGy(SiO₂), experiencing 50% gain attenuation for all tested temperatures. Given that the primary factor contributing to Brillouin gain loss is RIA, similar behaviors were observed such as: the loss increases at lower temperature, the stabilization effect of the temperature dependence for the highly Ge-doped fiber core, as described in Section 3.1, and the effect of the dopant nature. In fact, Ge-doped fibers exhibited higher gain losses than F-doped fibers due to their spectral absorption bands in the NIR domain, inducing greater RIA. The F-doped fiber exhibited low Brillouin gain irradiation temperature sensitivity from 12% losses at +80 °C to 20% losses at −80 °C compared to pristine samples. As the temperature decreases, the efficiency of thermal annealing diminishes and the Brillouin gain exhibits higher attenuation at −80 °C. Since the Brillouin gain attenuation is subjected to RIA, refractive index and acoustic velocity changes, it is not possible yet to decorrelate the contribution of each phenomenon to Brillouin gain attenuation.

3.3. Radiation-Induced Brillouin Frequency Shift

X-ray irradiation not only affects losses in the fiber and Brillouin gain losses, but also induces a BFS. The Lorentz–Lorenz equation for molecular refractivity links the refractive index change to the density change [27,29], showing a linear increase in the refractive index change with the density. On the other hand, the Kramers–Kronig equation links an increase in the refractive index change to an increase in the RIA. For X-rays exposure up to 108 kGy(SiO₂) absorbed dose, no significant densification or dilatation are expected [11]. No particular relation between X-ray irradiation and a silica-based glass compaction has been demonstrated, but a significant effect from γ-rays on Ge-doped fibers irradiated at 13 MGy(H₂O) [30] has been reported, showing a positive refractive index change of $5.7\times {10}^{-3}$.

A significant RI-BFS at this dose level can then be seen as a multi-contribution of the refractive index change due to RIA, a change in the acoustic velocity due to Young’s modulus or a change in the Poisson coefficient respective to relations (3) and (4).

$${\nu }_{B0}=\frac{2n_{eff}{V}_{A_{eff}}}{{\lambda }_p}$$

(4)

where ${\nu }_{B0}$ is the Brillouin central peak frequency depending on the effective refractive index, the effective acoustic velocity and the pump wavelength.

Figure 6 shows the RI-BFS evolution as a function of the accumulated dose up to 108 kGy(SiO₂) at −80 °C, −40 °C, +20 °C and +80 °C for the three tested fiber samples. The presented results have to be considered with a ±0.4 MHz uncertainty margin resulting from temperature variations inside the irradiation chamber, which can become critical when differentiating low RI-BFS levels. The sole contribution of X-rays to BFS was obtained from the measured BFS temperature correction procedure described in Section 2.4. The overall behavior was similar to the RIA and Brillouin gain loss behaviors. Larger BFS were found at lower temperatures for all tested fibers. SMF28e+ demonstrated the highest RI-BFS at −80 °C, reaching up to 6.8 MHz, and, at +80 °C, it achieved 0.6 MHz. For a higher Ge dopant concentration, the CMS fiber exhibited a lower temperature of irradiation sensitivity, experiencing a 4.2 MHz RI-BFS at −80 °C and a 1.4 MHz RI-BFS at +80 °C. Concerning the F-doped fiber, no RI-BFS could be seen from +80 °C to −40 °C, whereas, as for the RIA and BGS losses, the F-doped fiber exhibited a particular temperature sensitivity, experiencing an RI-BFS of 1.8 MHz at −80 °C only. For low temperatures, the known radiation-hardened F-doped fiber reaches equivalent RI-BFS levels to a Ge-doped fiber at room temperature.

The radiation-induced frequency shift temperature sensitivity was reduced for a higher GeO₂ dopant concentration between the SMF28e+ and CMS samples. Having a more major contribution of the RIA to the refractive index change than the density has to both the refractive index and acoustic velocity, the main contribution to the RI-BFS is the RIA through the Kramers–Kronig dispersion relation. The RI-BFS did not present any particular recovery after irradiation for all tested temperature from −80 °C to +80 °C.

A systematic frequency measurement error is induced under irradiation due to RI-BFS, leading to temperature and strain systematic errors when used as a sensor. This systematic error depends on both the fiber composition and the temperature of irradiation, as shown in Figure 6. After the 108 kGy(SiO₂) accumulated dose, temperature- and strain-related radiation-induced errors from −80 °C to +80 °C were as presented in

Table 2. Radiation-induced temperature and strain errors.

Fiber Samples	SMF28e+				CMS				F Doped
Temperature of irradiation (°C)	−80	−40	+20	+80	−80	−40	+20	+80	−80	−40	+20	+80
Temperature error (°C)	+6.21	+4.92	+1.07	+0.51	+3.82	+3.02	+1.70	+1.29	+1.48	<+0.3	<+0.2	<+0.2
Strain error (µε)	+143	+112	+38	+28	+97	+77	+43	+32	+33	<+5	<+5	<+5

. The F-doped fiber exhibited the lowest temperature and strain radiation-induced errors with no errors for temperatures higher than −40 °C.

3.4. Post Mortem Results

Post-irradiation temperature and strain Brillouin sensitivity measurements were performed for the three tested fibers irradiated up to 108 kGy(SiO₂) at −80 °C, −40 °C, +20 °C and +80 °C. Temperature and strain measurements were performed following the procedure described in Section 2.3.

It has been shown that increasing the GeO₂ dopant weight percentage in the fiber core reduced the fiber Brillouin central peak temperature and strain sensitivities [31], reaching about a −1.48% and −1.61% decrease for 1 mol% of dopant measured on a 3 mol% and 8 mol% GeO₂-doped fiber, respectively [32]. It was also reported that the Brillouin central frequency depends on the molar concentration, with shifts of −87.3 MHz/mol% at 1550 nm [32] and −97 MHz/wt% at 1320 nm [31].

Results are presented in Figure 7 and corroborate the behaviors of both $C_T$ and $C\epsilon$ from the previous studies presented above concerning GeO₂ doping effects. Figure 7 reveals no particular temperature of irradiation dependency after X-ray exposure from −80 °C to +80 °C for either F-doped or CMS samples within the measurement uncertainties. A higher temperature dependence of the $C_T$ radiation response can be seen for the SMF28e+ sample, exhibiting a decrease of 3.6% between +80 °C and −80 °C. Knowing that $C_T$ decreases with the absorbed doses for slightly GeO₂-doped fibers [22], a tendency for lower temperatures can be seen where the thermal annealing is less efficient. However, this last result is very close to the measurement uncertainties.

Post-irradiation $C\epsilon$ measurements showed no particular irradiation temperature dependence for all three tested fiber types from −80 °C to +80 °C, after 108 kGy(SiO₂) X-rays exposure, within our measurement uncertainties. Both $C_T$ and $C\epsilon$ are not affected by the irradiation temperature in the investigated range and uncertainties associated with our experimental setup. Compared to pristine fibers, no significant modification of either sensitivity was found for the irradiated tested samples considering uncertainty margins.

4. Discussion

The presented results show and explain the RIA behavior between −80 °C to +80 °C on a F-doped fiber and two differently doped germanosilicate optical fibers. This work explains the origin of the discrepancies between the tested samples in terms of their temperature of irradiation sensitivity, with the help of previous studies, focusing on spectral analysis of defect creation under irradiation and their contributions to RIA. The same procedure is applied to the BGS amplitude decrease, exhibiting the same temperature of irradiation sensitivity response as RIA for all tested fibers. According to this study, the temperature of irradiation differently impacts the sensing capabilities in terms of RIA and BGS losses, showing a dependence on the fiber dopant in terms of nature and concentration. The irradiation temperature sensitivity is reduced, increasing the Ge dopant concentration, its effect being negligible for the highly Ge-doped fiber between +80 °C and −40 °C.

The Brillouin central peak attenuation due to radiation exposure can lead to a drastic reduction in the sensing length from several tens of kilometers for pristine to ten meters for the most sensitive fiber SMF28e+ at −80 °C after 108 kGy accumulated dose. The temperature of irradiation is a hot topic when designing Brillouin sensing techniques to ensure an accurate measurement during the entire exposure preponderant for two-ended Brillouin sensing techniques. Even though BGS losses and RIA exhibit the same behavior, the combined temperature and irradiation effects on the gain (relation (2)) should not be neglected for high accumulated doses. In fact, the RIA impacts the gain as well as the radiation-induced refractive index change and acoustic velocity. RIA mainly contributes to both Brillouin gain losses (through light propagation of both pump and signals) and the refractive index change (through the Kramers–Kronig dispersion relation [29]). Moreover, the densification of the fiber matrix can contribute to a refractive index change through the Lorentz–Lorenz relation [29], but a significant effect would require higher accumulated doses (especially under neutron irradiation).

Desensitizing an optical fiber to its temperature of irradiation in a certain temperature range, as for the CMS fiber, can improve the calibration phase, regulating the expected losses for RIA and Brillouin responses. In addition, using an irradiation-temperature-desensitized fiber helps to correct the systematic temperature or strain measurement errors induced by the RI-BFS, which remains unchanged for any considered temperature of irradiation. In the context of nuclear-waste-monitoring Brillouin-based sensing [1], this type of fiber would present many advantages in terms of calibration of systematic errors and losses for a long exposure period in a temperature-varying environment. For such an application, an interesting procedure would be to design an irradiation-temperature-immune optical fiber. The F-doped fiber would satisfy these requirements concerning Brillouin sensing in the temperature range from −40 °C to +80 °C and +20 °C to +80 °C in terms of RIA control in a temperature-varying environment.

Deeper investigations into the origin of the RI-BFS and contributions to Brillouin gain losses during irradiation require accurate measurements of acoustic velocities before and after irradiation as well as the refractive index profile.

5. Conclusions

This work illustrates Brillouin sensing capabilities for Ge- and F-doped fibers under X-rays in harsh environments from −80 °C to +80 °C. Accumulated doses up to 108 kGy(SiO₂) are reached with a dose rate of 6 Gy(SiO₂)/s for every tested sample. RIA results at 1310 nm, 1550 nm and 1625 nm exhibit a low temperature sensitivity for F-doped fibers, a high temperature sensitivity for SMF28e+ fiber and a reduced effect adding a higher GeO₂ concentration in the fiber core. Brillouin sensing capabilities are investigated under the same conditions, showing a similar behavior of the BGS amplitude decrease with RIA. BGS evolution during the irradiation runs does not show any temperature sensitivity for either F-doped or CMS samples, whereas SMF28e+ samples present a high temperature sensitivity, exhibiting the same reduced effect as RIA when increasing the Ge dopant concentration. A similar behavior is observed for the RI-BFS temperature of irradiation dependency for both SMF28e+ and F-doped fibers, exhibiting high and low RI-BFS temperature dependencies with no possible decorrelation between RIA and radiation effects on Brillouin spectrum characteristics. CMS samples exhibit a high RI-BFS temperature sensitivity, while RIA and BGS attenuations do not show any temperature sensitivity over −40 °C. Further studies could investigate acoustic velocities and refractive index change after irradiation, looking forward to gaining a clear understanding of the origin of the Radiation-Induced Brillouin Frequency Shift.