Radiation resistant fiber Bragg grating in random air-line fibers for sensing applications in nuclear reactor cores

: This paper reports the testing results of radiation resistant fiber Bragg grating (FBG) in random air-line (RAL) fibers in comparison with FBGs in other radiation-hardened fibers. FBGs in RAL fibers were fabricated by 80 fs ultrafast laser pulse using a phase mask approach. The fiber Bragg gratings tests were carried out in the core region of a 6 MW MIT research reactor (MITR) at a steady temperature above 600°C and an average fast neutron (>1 MeV) flux >1.2 × 10 14 n/cm 2 /s. Fifty five-day tests of FBG sensors showed less than 5 dB reduction in FBG peak strength after over 1 × 10 20 n/cm 2 of accumulated fast neutron dose. The radiation-induced compaction of FBG sensors produced less than 5.5 nm FBG wavelength shift toward shorter wavelength. To test temporal responses of FBG sensors, a number of reactor anomaly events were artificially created to abruptly change reactor power,


Introduction
Nuclear reactors are reliable and low-cost, carbon-free energy sources to supplement intermittent renewable energy production tied to the electrical grid. High spatial resolution and multi-modal information gathering sensors through the nuclear power system could have potentials to reduce design redundancy, and subsequently the overall cost of the nuclear power systems. Thus, development and implementation of radiation-hardened advanced sensors in reactor cores is a major theme for the development of next generation nuclear power systems. To ensure operational safety, multiple-level redundancies are built into nuclear reactors. These fail-safe designs of nuclear power systems often lead to prolonged construction periods and high construction and operational costs. An alternative approach to ensure operational safety is through better sensing and control schemes.
Nuclear reactor cores are arguably the most challenging of artificial environments. High temperatures and strong neutron fluxes at levels of 1 × 10 14 n/cm 2 /s essentially put most of sensor devices out of operation. Currently, thermocouples are probably the only reliable temperature measurement devices for in-core applications. Since the 1990s, researchers have being exploring the use of optical fiber sensors for in-core monitoring for nuclear power systems [1]. Well known for their resilience in harsh environments [2], fiber Bragg grating (FBG) sensors and distributed fiber sensors can perform multiple simultaneous measurements including temperature, strain, pressure, water levels, and chemical sensing with high spatial resolutions. This makes FBG sensors attractive sensing devices for nuclear energy applications.
Study of the behavior of various types of FBGs in ionizing radiation environment has spanned for over two decades [3]. Probably the most comprehensive studies are from Fernandez and Gusarov et al., who reported their experiments on FBG sensors in harsh nuclear core environments at the BR1 and BR2 reactors (SCK.CEN) from 1999 to 2008. In 2002, they reported the use of FBGs in various optical fibers, especially photosensitive fibers and telecommunications fibers for online temperature monitoring in low neutron flux (~10 11 n/cm 2 /s) and high neutron flux (~10 14 n/cm 2 /s) environments. Moderate gamma irradiation was also presented in those studies [4,5]. FBG sensors reported in these studies were irradiated at low temperature from 10 to 80°C, while FBGs endured various radiation doses up to 160 MGy of gamma radiation and a total fast neutron fluence (> 1 MeV) up to 1.47 × 10 17 n/cm 2 . After comparative studies using a thermocouple as a reference, it was reported that FBG sensors can measure temperature in those environments with a 3°C accuracy. Studies carried out by Fernandez and Gusarov et al. reveal that both optical fiber materials and FBG sensor fabrication process have great impacts on performance of FBG under these harsh conditions [6-10]. Fielder et al. [11] also claimed an 87% survival rate of FBG sensors in neutron/gamma radiation studies, where FBG fabricated in Ge-doped and B/Ge co-doped fibers endured hours of fast neutron flux (1 × 10 13 n/cm 2 /s) irradiation with accumulated neutron fluence of ~5 × 10 19 n/cm 2 , and γ-irradiation of 2 × 10 3 MGy. Recently, Remy et al. [12] used FBG sensors to measure the longitudinal compaction effect on pure silica core and F-doped core samples exposed to fast neutron fluences of ~4.76 × 10 19 n/cm 2 at 291°C by measuring the radiation-induced length changes, namely the longitudinal length shortening of the fiber samples.
In this paper, we report in-core studies of FBG sensors at radiation levels similar to those occurring inside commercial nuclear reactors. FBG irradiation tests were performed at >600°C with fast neutron fluxes above 1.2 × 10 14 n/cm 2 /s for a 55-day continuous irradiations with 5.7 × 10 20 n/cm 2 accumulated dose, at least a 10-fold higher dose than previously reported. To evaluate the potential capability of the fiber sensors to monitor reactor anomalies, multiple events were implemented during the irradiation to create abrupt changes of reactor temperatures and reactor powers. FBG responses to these abnormal conditions were characterized to assess thermal-optic properties of FBGs at different stages of irradiation and their temporal response during abrupt temperature changes. Experimental results presented in this paper provide a quantitative evaluation of FBG performance in commercial reactor cores. It also provides cues on how to mitigate FBG property drifts in reactor cores in order to improve both reliability and accuracy of FBG sensors for in-core applications.

Fiber selection, FBG fabrication, and experimental setup
The in-core radiation experiments reported in this paper were carried out in the MIT nuclear reactor lab. Given the limited space for sensor insertion, two types of optical fibers were selected for the experiment. These two optical fibers were selected based on previous literature studies on radiation-hardened fibers under both gamma and neutron irradiation. They were also selected based on the cost of fibers for potential large scale deployment. These two fibers include pure-silica random air-line (Corning's RAL) fibers and pure silica core, F-doped cladding single-mode fibers (Corning's Vascade) [13]. According to Morana et al [14], F-doped single mode fibers exhibited the lowest radiation-induced attenuation (RIA) at 1550 nm operation, in a study where fs-FBGs inscribed on F-doped, pure-silica core, and Ge-doped fibers were exposed to an accumulated gamma dose of 1 MGy. However, upon exposure to radiation, F-doped fibers had shown a slightly larger Bragg wavelength shift (BWS) towards the blue (−0.03pm/kGy for F-doped) compared to pure silica core fibers (−0.02pm/kGy for pure silica core). Ge-doped fibers exhibited by far the largest radiation-induced BWS by approximately two order of magnitudes, which is undesirable for harsh environment sensing.
The pure-silica RAL fiber used in this paper is a nano-engineered optical fiber. The fabrication process is similar to Corning's ClearCurve fiber [15,16] where the fiber was fabricated using outside vapor deposition (OVD) process. A pure silica core cane was inserted into a silica soot blank made with the OVD process to form a preform. The preform was consolidated in a consolidation furnace at a pressurized nitrogen atmosphere around 1450 °C to form a glass cladding containing air bubbles. The preform was then drawn into fiber. During the draw process, the air bubbles were stretched into thin elongated airlines. The optical confinement of the guided mode within the fiber was achieved by the surrounding random air-line structures with a lower average refractive index. Comparing with other airline microstructural fibers, RAL fibers can be made at much lower cost. Figure 1(a) shows a scanning electron microscope (SEM) cross-section photo of the RAL fiber, where the 8 m diameter fiber core is surrounded by a random air-line ring 45 m in diameter to provide optical confinement. Figure 1(b) shows detailed air-line distributions in the fiber cladding. Size of air-line ranges from 15 nm to 1 m. The size and distribution of the air-line can be controlled by the nitrogen pressure during the fiber preform consolidation process.
The pure silica core fibers are commonly known to have low susceptibility to RIA and hence are the most radiation resistant fibers, especially in operation in near IR range [17]. The entire RAL fiber is made of a single fused silica material with identical thermal expansion coefficient in both fiber core and cladding regions. Both fiber core and cladding also have similar responses under both neutron/gamma radiation. Thus, it is conceivable that the fiber will have even better resistance under radiation exposure. Comparing with other radiation hardened pure silica fibers such as photonic crystal fiber [18], which is also ideally suited for FBG inscription [2], RAL can be manufactured at much lower cost.
To fabricate the Bragg grating in the RAL fiber, a Ti:sapphire regenerative amplifier operating at a central wavelength of = 800 nm was used. The fs-beam diameter was ~7 mm at the 1/e 2 level. The full width at half maximum (FWHM) transform-limited width of the pulses was ~80 fs, as measured using non-collinear auto-correlation. However, the pulses were pre-chirped to 500 fs in order to facilitate Type II-IR grating formation [19] and the associated nano-grating formation [20]. The pulse repetition rate was kept at 5 Hz. The linearly polarized laser pulses were focused into the fiber core using a 12 mm focal length plano-convex acylindrical lens after passing through a silica phase mask, which has a 3.21 µm pitch. The fiber was positioned approximately 300 m away from the phase mask along the beam propagation direction. The acylindrical lens was dithered ± 10 µm at 0.1 Hz with a piezo actuator in order to scan the beam across the fiber core. The resulting grating structure produced a high-order resonance at ~1525 nm in the telecom band as shown in Fig. 1(c). FBGs were inscribed in 25 m section of fibers. The locations of the 1 cm long FBGs are approximately 3.5 cm from one end of the fibers. The sensor section of the fiber was slipped into a 6 m long 316 stainless-steel tubing with an outer diameter of 1.59 mm (1/16"), and a wall thickness of 0.25 mm (0.010") for protection and mechanical support of the sample as it descends into the reactor core. The rest of the fiber, which was not subject to any radiation outside the reactor, was sheathed with polyvinylchloride (PVC) tubing. One end of the stainless-steel tube was sealed by welding in argon gas ambient, while the other end of the stainless steel tube was left open to allow the fiber to be fusion spliced to a fiber patch cable with angled physical contact connector to connect with an FBG interrogation system. The FBG interrogation system comprised a TUNICS T100S-HP tunable laser and a CT400 optical switch (both from Yenista Optics) with four interrogation channels. FBG spectra were recorded every 20 seconds as wavelengths of the tunable laser were swept in the range from 1440 nm to 1640 nm with a wavelength resolution of 2 pm.
The MIT reactor (MITR) is a 6 MW light water-cooled, heavy water-reflected reactor with in-core positions for flow loops and instrumented experiment facilities. The core design provides a high power density with a neutron flux and spectrum similar to that of a typical light water power reactor. MITR operates 24/7 with 10-week cycles; typical in-core experiments will operate under steady-state conditions for the duration of each cycle. The fibers are irradiated in MITR's In-Core Sample Assembly (ICSA) facility. The ICSA consists of a dry, pressurized (inert-gas filled) Grade 2 titanium (ASTM Grade-2) thimble that extends from the bottom of the reactor core to the top of the reactor primary core tank, and permits simplified installation of instrumented experiments desiring temperatures up to 900°C. The experiment capsule and instrumentation leads are all contained within this thimble and are routed at the top through pressure fittings and out of the reactor biological shielding. The heating in this facility is supplied entirely by nuclear heating of the materials under neutron and gamma irradiation, with some fine control available by manipulating the ratio of helium to neon in the facility cover gas. The temperature is actively monitored by multiple thermocouples within the core and feeds back into the gas control system, generally permitting ± 2°C stability during the cycle. Local temperatures are set by the reactor power, materials, and geometry of the heat transfer pathway -the ultimate heatsink is the 50°C reactor primary coolant flowing at the exterior of the ICSA thimble. Figures 2(a) and 2(b) show a photo and 3D schematic sketch of the test capsule which is placed inside the ICSA thimble. It consists of a graphite sample holder in a 20.5 cm long grade 2 titanium cylinder cup. Stainless tubes containing various fiber sensors are inserted in 4 holes drilled into the graphite sample holder as shown in Fig. 2(a). Three K-type 1/16" thermocouple with Inconel sheaths (from Omega) were also inserted into the reactor core to perform temperature measurements at three different locations in the reactor core region. The locations of the thermocouples are −18 cm, −12.5 cm, and −5 cm from the center of the titanium cup. The titanium cup, which was placed on a grade 2 titanium pedestal or spacer, was positioned in the middle section of the reactor core to receive the maximum neutron flux.
The neutron flux profile in the reactor core at 5.6 MW, where the test capsule was located, is shown in Fig. 2(c). The test capsule, which is depicted in the figure using a grey bar, is placed in the region with the maximum fast neutron flux, which are in excess of 1.2 × 10 14 n/cm 2 /s for fast neutrons with >0.1 MeV kinetic energy and 0.6 × 10 14 n/cm 2 /s for fast neutrons with >1 MeV kinetic energy.

Results and discussion
When the experiment commenced, the capsule temperature was raised by 50°C from room temperature at the first day allowing a test for the response of the sensors before exposure to neutron flux and to confirm that the sensors were safely and accurately inserted inside the nuclear reactor core. The reactor power was then gradually raised over a period of 3 days to be kept steady at 5.62 MW with the capsule reaching a final temperature of 622°C with a combination of nuclear heating and helium/neon regulation. Figure 3(a) shows the FBG spectra every 5 days during the test for FBG in RAL fibers. Figure 3(b) shows the spectra of FBG inscribed on Vascade fibers from Day 0 to Day 7. The FBGs in Vascade fibers were damaged after 7 days in the reactor. All data presented thereafter was acquired from FBG in RAL fibers. To simulate FBG sensor behaviors during various reactor anomaly scenarios, several events were staged to assess the response of the FBG prolonged neutron radiation at various accumulated fluences. The reactor power and the capsule temperature measured by TC3 are presented in Fig. 3(c). Reactor power was reduced to 4 MW and 0 MW at Day 18 and Day 32, respectively. A small reactor power reduction also occurred at Day 41 from 5.6 MW to 5.59 MW. This small reactor power reduction did not have significant impacts on FBG. An additional reactor event also occurred at Day 36 as shown in Fig. 3(d), which caused a temporal drop of the FBG wavelength. It is not clear what was the cause of this event. We therefore did not include this event for subsequent data analysis. The temperature and reactor power was then increased at various rates. High-temperature performance of FBG inscribed on this RAL fiber by the femtosecond laser was confirmed independently without radiation [21]. In these experiments, no degradation in reflectivity of the FBG was observed after 5 days at 800°C. The test temperature of 800°C was significantly higher than the reactor temperature of 620°C.
The FBG peak wavelength and reflectivity, measured from peak to baseline, were recorded every 20 seconds. The overall radiation effect on the FBG wavelength and strength is shown in Fig. 3(d). Although reactor anomalies affect the FBG performance, the overall trend of the radiation effect is to cause a blue shift of the FBG resonant wavelength. At constant temperatures (622°C) and neutron flux (5.6 MW), the FBG wavelength drifts slowly towards shorter wavelength, with an average wavelength drift rate of 0.096 nm/day. The FBG wavelength drifts appear to be linear to the accumulated neutron flux. Over the test period of 55 days, FBG wavelength shifted 4.47 nm towards the shorter wavelength. This is consistent with the previous observation of irradiation effects on FBG under gamma and x-ray radiation, as prolonged irradiation (accumulated dosage > 100 kGy) produce blue BWS [3,22]. Although FBGs behavior can be qualitatively explained, the lack of models that accounts fully for the underlying physics governing the behavior of various types of FBGs in radiation environments remains a problem [3,4].
The radiation also weakened the FBG strength. The grating used in these experiment has ~20% (~1 dB) peak reflectivity. The reduction of FBG strength could be the combined effect of radiation-induced absorption and loss of reflectivity of the FBG. On average, during the first 30 days of radiation, the FBG amplitude dropped by 0.125 dB/day, or a reduction of 0.12 dB per 1.0 × 10 19 n/cm 2 accumulated fast neutron flux, which is plotted together with the FBG wavelength shifts in Fig. 3(c). However, the reduction of FBG strength appears to saturate as the test progressed. Overall, the FBG strength drops by 6 dB over a period of 55 days of active radiation. Vascade fibers. The in-core (c) temperature measured by TC3 and reactor power profile in the first 55 days of nuclear reactor operation, and (d) the RAL FBG response in the form of wavelength and intensity amplitude shifts. The FBG peak wavelength systematic shift due to neutron radiation induced index (RII) changes at constant temperature. The intensity peaks also slightly drop due to the effect of RIA.
Due to three simulated anomaly interrupts of reactor operations at Days 18, 32, and 41, reactor power was varied to 4 MW, 0.1 MW, and 5.59 MW and 0 MW, respectively, from the targeted 5.6 MW at steady state. Behaviors of FBG were also influenced by these interrupts. Results are presented in Figs. 4(a)-4(d). The interrupts of the reactor operation have impact on both FBG wavelength and FBG strength. The continuous irradiation of FBG from Day 6 to Day 18 produced 1.2 nm FBG wavelength blue-shift from 1527.27 nm to 1526.08 nm. The FBG shift follows a linear relationship with the accumulated neutron dosage or irradiation time. This amounts to an FBG linear shift rate of −0.096 nm/day. This wavelength shift rate remains largely consistent without significant saturation. During the subsequent constant irradiation periods between Day 22 and Day 31, Day 38 and Day 41, and Day 42 and Day 55, FBG wavelengths shift rates are measured to be −0.090 nm/day, −0.098 nm/day, and −0.101 nm/day, respectively. FBG wavelength shifts were not characterized during the nuclear power interrupts. This is because the power profiles during the interrupts are complex and relatively short. The relative steady FBG wavelength shift summarized in Table 1 suggests that it could come from a single dominating underlying mechanism. This is consistent with previous reports on the radiation effects on FBG under gamma and x-ray radiation [3,22]. A linear and steady FBG wavelength drift might be mitigated or calibrated on-the-fly to improve the FBG measurement accuracy for static measurement applications (e.g. temperature or strain). For example, a temperature insensitive in-fiber interferometer might be used to calibrate the radiation induced index (RII) change. On the other hand, irradiation-induced FBG peak reduction shows strong signs of saturation. During the first period of steady radiation between Day 6 and Day 18, FBG peak reduced by 2 dB with a −0.136 dB/day decay rate. The FBG peak decayed at a lower rate during the second period of steady radiation between Day 22 and Day 31 at −0.113 dB/day. The decay rates show strong saturation after 38 days with −0.04 dB/day and −0.023 dB/day during the last two steady irradiation periods, respectively. Intensity amplitudes of the FBG peaks are reduced by less than 0.6 dB between Day 38 and Day 55.  To evaluate FBG responses to reactor anomalies while temperature and reactor power levels experience abrupt changes, a number of events were staged on Day 18, 32, and 41.
During these staged reactor anomalies, reactor power and temperature were raised and reduced at various steps for different periods of times, which creates opportunities to evaluate thermal optic properties of the FBG at various stages of neutron irradiations and different temperature and radiation flux conditions. Some of the transient responses of the FBGs are shown in Figs. 5(a) and 5(b). Results presented in Fig. 5(a) show a staged event while temperature and reactor power was fluctuated several times between 110 and 160°C and 0.5-1 MW, respectively. This was designed to test responsivity of FBG sensors at low temperatures. The temperatures presented in Figs. 5(a) and 5(b) were measured by a thermocouple (TC3 in Fig. 2(b)). Figure 5(b) shows the temporal responses of FBG sensors for another event while the reactor temperature was raised stepwise from 165 to 540°C, and correspondingly the reactor power from 1 to 5.5 MW, while observing the FBG wavelength shift that mimics the change in the aforementioned parameters. The FBG wavelength responses to reactor power changes presented in Figs. 5(a)-5(b) are compared with the thermocouple measurement results. If FBG wavelength shifts are directly and linearly related to reactor temperature changes under intense neutron flux, it is interesting to note that FBG wavelength has faster response to the reactor power changes than those measured by the thermocouple devices as the calculated time constants for the sensors' thermal responses were 239.3 seconds for the FBG sensor, and 269.6 seconds for the thermocouples. This is more pronouncedly illustrated in Fig. 5(b). Given that thermocouples are the only temperature measurement devices within the reactor cores, it is not clear what is the cause of these different response time. To evaluate FBG temperature sensitivity coefficients at different stages of irradiation, four temperature cycles were executed and plotted in Figs. 6(a)-6(d). FBG wavelength changes vs temperature was recorded when the reactor temperature was gradually dropped before quick restoration to the target 622oC. Figures 6(a) and 6(b) show FBG wavelength changes vs. temperature changes during events occurred between Day 32 and Day 33. Figures 6(c) and 6(d) show FBG wavelength changes vs. temperature changes recorded on Day 55, when two events were staged for FBG temperature sensitivity coefficient measurements. All of the measured data are linearly fitted to evaluate the FBG sensor's thermo-optic coefficient that would allow calibration of the sensor and conversion of FBG wavelength shifts to temperature readings. The wavelength changes were measured with respect to the initial wavelength at the start of each cycle, taken at constant temperature and neutron flux. The fitting specifications and results are also summarized in Table 2. The results show a changes in the temperature sensitivity coefficient due to irradiation. Prior to neutron radiation, the FBG temperature sensitivity coefficient is 14 pm/°C. Prolong radiation has reduced the FBG temperature coefficient to 5.97 pm/°C at Day 55.

Summary and conclusion
This paper presents nuclear irradiation experimental results of temperature-stable FBG sensors fabricated by a femtosecond laser in pure silica random airline fibers. The results show that when appropriate fibers are chosen and suitable laser fabrication techniques are implemented, FBG sensors can survive extreme environments of the nuclear reactor cores under intense gamma and neutron irradiation at high temperatures. The FBG peak decay stabilizes after approximately 34 days of irradiation. However, the FBG wavelength shows a consistent drift toward shorter wavelength at almost constant linear rate of −0.1 nm/day. The reduction in grating strength is related to a reduction in the index modulation while the wavelength shift is due to a change in the average index of the glass under neutron irradiation. The use of FBGs for static measurement could be impacted by this linear FBG wavelength drift. A self-calibration method might be needed for FBG to perform static measurements such as absolute temperature measurements. FBG Wavelength drift and changes of the FBG temperature sensitivity coefficients induced by strong radiation present significant challenges to use FBG to perform static measurements. Therefore, applications of FBG, in current forms, might be limited by these radiation-induced parameter drifts unless an in-pile calibration technique can be developed. On the other hand, results presented in Table 2 shows that after initial large changes of FBG properties, FBG temperature sensitivity coefficient tends to stabilize. A radiation-harden fiber with multiple FBG could be used to detect significant temperature swings and temperature profile changes over a short period of time. Radiation resilient FBGs presented in this paper can also be used to perform dynamic measurements such as distributed acoustic and vibrational measurements in the reactor core while FBGs are used as reflectors. In summary, results presented in this paper reveal that temperature-stable Type-II FBGs fabricated in radiation-hardened fibers could survive harsh in-pile conditions. Despite large parameter drift induced by strong neutron radiation, further engineering and innovation on both optical fibers and fiber devices could lead to useful fiber sensors for various in-pile measurements to improve safety and efficiency of existing and next generation nuclear reactors.