Open Access
Add to BibSonomyAbstract
One promising new application for optical fiber sensors is in monitoring superconducting magnets that are, inevitably, operated at cryogenic temperatures. The cryogenic thermal sensitivity of commercially available optical fibers is depressed by the low coefficient of thermal expansion of the constituent materials. Here, single mode, telecommunication-grade optical fibers are coated with different metals to alter their sensitivity to thermal perturbation at temperatures as low as 4.2 K. Commercially available fibers with acrylate coating are compared to fibers with only metallic coatings and those with acrylate-metal composite coatings, in terms of their sensitivity to thermal perturbations in the temperature range from 4.2 to 61 K. The metals considered include Sn, PbSnAg and InBi, both on cladding and on an acrylate primary coating. Results show that the 4.2 K thermal sensitivity can be enhanced significantly by a composite coating approach.
© 2017 Optical Society of America
1. Introduction
Optical fiber sensors are a relatively mature technology and find use in numerous applications, most commonly to measure strain, temperature or pressure, but also tailored to sense the concentration of specific chemical species [1, 2]. The main advantages of optical fiber sensors are immunity to electromagnetic noise, lightweight, relative robustness and durability, sensitivity, cost, and the fact that they can be tailored in many ways to suit specific applications. Their aspect ratio, particularly in applications where distributed sensing over long distances is needed, is also an important feature of optical fiber sensors that is not offered by any competitive technique.
Recently, a new potential area of application of optical fiber sensors is as a monitoring tool in superconducting magnets (SCMs). In this application, their main task is to detect local transitions to the normal state. In fact, a local transition to the normal state causes a local temperature increase (hot-spot), which is the driver for material degradation. Therefore, an undetected hot-spot, which is the primary failure mechanism in SCMs, may permanently damage the conductor and induce an irreversible loss of superconductivity. The additional complication in monitoring superconductor technologies is the cryogenic environment in which they operate. In fact, due to the need to cool superconducting materials below their critical temperatures, all superconducting devices are operated at cryogenic temperatures ranging from the 1.9 K for the NbTi superconducting dipole magnets of the Large Hadron Collider (LHC) [3] to the higher temperature magnet applications that employ high temperature superconductors (HTS). In some HTS applications, such as electric motors and generators, the operating temperature may be as high as 77 K [4].
All superconducting systems must be protected from a failure event referred to as quench. Superconductors exhibit superconductivity only under certain conditions of temperature and magnetic flux density, and within that operating space there is a limit to lossless superconducting transport known as the critical current density, Jc. Each of these quantities has an upper bound that, if crossed, causes the transition of the superconducting material to the normal state (in the case of temperature and magnetic field) or to a “mixed state” in which some of the current is transported by the metallic matrix that is in intimate contact with the superconductor. These transitions to the normal state typically occur locally and then propagate, driven by the joule heating created by the initial normal zone [5–8]. Therefore, to protect the magnet form its own stored energy, it is necessary to detect a normal zone as quickly as possible to prevent it from spreading uncontrollably throughout the magnet, resulting in a very high local maximum temperature that can cause irreversible damage to the superconducting material [9–11].
The joule heating in a normal zone is created by the current flowing in a resistive material, a normal zone is accompanied by both a temperature and voltage rise. Typically SCMs operate at a current density well below Jc, providing a “current margin”, which is the difference between the critical current density and the operating current density, and it is a safety margin adopted during magnet design. Thus, at the beginning stages of a quench, the temperature increases without any corresponding increase in voltage; the voltage does not begin to increase until the temperature has risen above the “current sharing temperature,” Tcs, which is the temperature at which Jc has decreased to equal the operating current density (because dJc/dT is negative).
At present, normal zone detection during the operation of SCMs is via voltage measurements across relatively large segments of the superconducting winding; this approach has proven effective for NbTi and Nb3Sn based low temperature superconductor (LTS)-based magnets for decades. In the HTS magnets, however, the normal zone propagation velocity is orders-of-magnitude slower than that in LTS magnets, resulting in much higher peak temperatures for the same voltage [5–8]. Thus, the sole reliance of voltage-based sensors puts HTS magnets at risk of irreversible material degradation. Furthermore, many SCMs operate in environments with high electromagnetic noise, particularly in nuclear fusion reactors. For these reasons, quench detection techniques that are not based on voltage measurement have been proposed and studied [12–19]. Some of them are based on Brillouin or Rayleigh scattering interrogated optical fibers and others take advantage of acoustic emission during a temperature transient to detect local thermal perturbations.
One technique that has shown great potential in HTS SCMs is based on Rayleigh-backscattering interrogated optical fibers (RIOF) [14, 15, 20]. Recent results show that the spectral shift signal measured by an optical fiber integrated into either a SCM [14] or directly embedded in the HTS conductor [20] is a more effective and more rapid quench detection system. Furthermore, RIOF-based quench detection is immune to electromagnetic noise and provides localization of normal zones with high spatial resolution.
One challenge that must be overcome to exploit the full potential of RIOF as quench detection and monitoring tool in SCMs is the low thermal sensitivity at cryogenic temperatures. Furthermore, there are no systematic studies of the behavior in terms of sensitivity and integrity of optical fibers at temperatures as low as 4.2 K. In this work, optical fibers have been coated with a set of different coating materials and their thermal sensitivity at temperature as low as 4.2 K has been compared when interrogated by Rayleigh scattering. Note that, although Rayleigh backscattering has been the interrogation used in this investigation, the same results would hold in case the fibers were interrogated by fiber Bragg gratings (FBG).
There are a number of ways to use an optical fiber as a sensor and each results in different sensing capabilities and sensor characteristics. The ones that have been considered for superconductor applications are FBGs [18], Brillouin scattering [12, 13], and RIOF. FBGs are intrinsically point sensors, and although multiple gratings can be inscribed and multiplexed on the same fiber, when the length of the fiber to be interrogated becomes of the order of the kilometer, interrogating an optical fiber via FBGs with spatial resolution < 1 cm is prohibitive. The need for distributed sensing with high spatial resolution originates from the fact that the position of the incipient perturbation is unknown and unpredictable, thus the time needed for the normal zone to propagate from its origin to the sensing location would greatly slow down the detection. Brillouin scattering also showed insufficient spatial resolution, so owing to its mm-range spatial resolution, RIOF is the best choice for SCM applications.
Here we report on the fabrication of metal-coated fibers and their thermal sensitivity at temperatures as low as 4.2 K. The thermal sensitivity is herein defined as the maximum spectral shift experienced by a fiber sample due to a fixed thermal perturbation. Therefore, the definition is meaningful when comparing different samples, but is also dependent on the energy released by the perturbation and therefore is not an absolute measure. Here we subject each sample to the same thermal perturbation and define the resulting spectral shift as the sensitivity of that sample; all other experimental conditions are identical.
RIOF sensing is based on quantifying the spectral shift, which is defined relative to a reference state. Therefore the spectral shift relates to changes in temperature and strain relative to a starting condition. The thermal sensitivity of RIOF can be expressed as the sum of two components: the physical elongation of the core-cladding structure upon temperature changes and the change in index of refraction due to the same temperature variation. The first effect is quantified by the thermal expansion coefficient of the optical fiber and the second by the thermo-optic coefficient. In formulae:
where is the thermal expansion coefficient, the thermo-optic coefficient, and the wavelength shift. The thermo-optic coefficient can be expressed aswhere n is the index of refraction of the fiber core and T is the temperature. Note that both contributions ( and ) are temperature dependent. The thermo-optic coefficient is usually considered constant for sensors operating at or above room temperature. While this is regarded as a good approximation in literature, for cryogenics temperatures, the thermo-optic coefficient cannot be considered constant. A typical value for the ( + ) at room temperature is , with and ranges between and (since n = 1.4469 for silica fibers and ranges to ) [21, 22]. Although at room temperature the thermo-optic coefficient is the major contribution to the thermal sensitivity of the spectral shift, this behavior is reversed at lower temperature. NASA Ames research center studied the temperature dependence of the thermo-optic coefficient and found experimentally that it dramatically decreases with temperature, approaching zero at 4.2 K. Specifically, a of about at room temperature becomes about at 77 K and approaches zero for temperatures of 4.2 K or below [23]. Therefore, at cryogenic temperatures, thermal expansion becomes the dominating contribution to the thermal sensitivity of RIOF. It is worth noting that the value of the coefficient of thermal expansion (CTE) used in the literature for comparison with the thermo-optic contribution is that of pure silica; this may not hold for coated optical fibers since the effective thermal expansion of the core-cladding-coatings composite structure can be higher than the thermal expansion of pure silica, if the coating materials have a higher CTE than silica.Assuming perfect coupling between layers, in a composite structure like that of an optical fiber, the maximum elongation of the cladding per unit of temperature change is obtained when the product of Young’s modulus (E) and the thermal expansion coefficient (CTE) of the coating material is maximized.
There is a physical limitation to maximizing the value of the E*CTE product with a unique material because, in general, the higher the bond strength, the larger the depth of the interatomic potential well. For any class of materials, the deeper the potential well, the narrower and more symmetric it becomes, generating a lower thermal expansion because of the lower increase in interatomic separation. In other words, as a general rule, as E increases, CTE decreases. For this reason, a composite coating structure is considered with a higher CTE in one component and a higher E in the other, yielding a higher E*CTE product.
2. Experimental approach
A coating method is developed and used to coat short samples of optical fibers. The thermal sensitivity of the different coated fibers is measured as a function of temperature from 4.2 to 61 K, and all samples are subsequently characterized via scanning electron microscopy (SEM) and energy dispersive x-ray spectroscopy (EDS) maps to assess the thickness and uniformity of the coating and the degree of bonding between the coating and fiber surface.
2.1 Materials studied
All the samples have a 125 µm cladding of a single-mode Corning® SMF-28e + ® optical fiber. Samples studied include as-received acrylate-coated fibers (“A”), as well as samples coated with InBi, Sn, and PbSn. Table 1 summarizes each sample type; note that A-InBi, A-Sn and A-PbSn refer to samples for which the respective metallic coating was deposited atop the acrylate, whereas in the other samples the acrylate was first removed. The precise composition of the InBi alloy is 66.3 wt% In and 33.7 wt% Bi, corresponding to the eutectic composition with melting point of 72 °C. The PbSn alloy is 62 wt% Sn, 36 wt% Pb and 2 wt% Ag, also corresponding to a eutectic composition that melts and solidifies at 179 °C. The Sn coating is 99.95% pure by weight and melts at 232 °C.
Table 1. Samples Studied
2.2 Coating method
A coating technique has been developed to deposit the metal coatings on optical fibers with and without acrylate coatings. The coating process is a melt-coating based approach because it takes advantage of the phase transformation of solidification, where the optical fiber is the substrate on which the metal solidifies and grows. The coating method is therefore a physical process, unlike the common chemical or electrochemical coating methods often used for optical fibers. The benefits of this coating process are the simplicity and low-cost as well as the high coating speed and throughput in comparison to chemical methods [24, 25].
The coating system is illustrated in Fig. 1. The optical fiber is drawn through a reservoir containing the melted coating material. Due to the temperature difference between the fiber surface and the liquid phase of the coating material, a coating layer solidifies onto the fiber. The coating material is kept at the desired temperature via a dedicated heater powered by a controllable power supply. A number of interdependent parameters must be controlled to engineer the coating quality, including the drawing speed, liquid level in the reservoir, the orifice size through which the fiber is drawn, and the temperature of the liquid metal and the fiber. Typically, the metal temperature is kept 1-5 K higher than the melting point of the coating material and the fiber enters the coating device at room temperature. A higher temperature margin would disable the coating process because the initially solidified layer would melt before the fiber exits the coating element, leaving the fiber uncoated. For the same reason, the coating speed has a lower bound whereas the liquid level of the coating material an upper bound, because both a lower coating speed and a higher liquid level cause every point on the fiber surface to reside longer in the liquid metal bath and therefore favor a higher fiber temperature while in contact with the liquid metal; those conditions may re-melt the initially solidified layer before the fiber exits the coating device.
Fig. 1 Illustration of the simple preliminary setup for fiber coating. H is the liquid level, Dor the orifice diameter and Vd the drawing speed. Note that the drawing can be in either direction.
2.3 Thermal sensitivity measurement
A custom setup for the measurement of thermal sensitivity at cryogenic temperature is shown in Fig. 2. A G-10 frame, capable of holding four fiber samples, serves as the samples holder. The frame is designed to minimize any factors that could artificially alter the thermal sensitivity of the fibers. Each fiber sample (about 15 cm in length) is held in place by the frame and equally spaced (1 cm), like the strings of a guitar. Each sample is inserted into a stainless-steel capillary tube surrounded by a heating element that is used to provide the thermal perturbation that generates a change in spectral shift; Fig. 3 illustrates the heating element. A nichrome wire is used as the heating element, with a resistance of 2.38 Ohms. For each experiment, the heater is pulsed for 500 ms at 2 W, corresponding to 1 J of thermal energy.
Fig. 2 Photograph of the G-10 frame holding four fiber samples for cryogenic thermal sensitivity measurements.
The fiber samples are spliced to a pigtail with a LC/APC connector and the fiber is terminated with a 30 cm coreless fiber to reduce the light reflected by terminal interface. During the experiments, the sample holder is held in a cryostat where the fibers are cooled by gaseous helium evaporating from a liquid reservoir at the bottom of the cryostat.
The Rayleigh backscattering interrogation is performed with an Optical Distributed Sensor Interrogator by LUNA Technologies, remotely controlled by LabVIEW software. Since the Rayleigh interrogation requires reference state to calculate the spectral shift, a reference scan is taken during the steady state preceding the release of the thermal perturbation by the heater, at the desired temperature of the experiment. Therefore, before the heater fires, the spectral shift is, by definition, zero everywhere.
2.4 Microscopy
Each fiber sample is prepared for SEM (JEOL 6010LA) and EDS analyses by vertically mounting it in epoxy, so that the fiber axis is perpendicular to the polished surface and the cross section can be imaged. The epoxy mount containing the samples is then polished to a mirror finish and sputter coated with a few nm layer of a Au-Pd alloy that provides a conductive surface. Micrographs were taken in secondary electrons mode and element distribution maps were taken with EDS.
3. Results and discussion
The primary acrylate coating that is commercially available with the SMF-28e + optical fiber is approximately 60 µm in thickness. The thickness of the metallic coatings deposited here are estimated using the SEM micrographs. Specifically, the coating thickness has been estimated in four different positions within the fiber cross section, a right angle apart from one another. For all samples, averages and standard deviations are calculated from these four values and included in Table 1. The standard deviation is indicative of the uniformity and concentricity of the coating relative to the cladding, or the primary coating.
Figure 4 shows cross-sectional SEM micrographs and corresponding EDS maps of the A-InBi fiber sample. Note that, although the InBi secondary coating thickness is not uniform around the fiber surface (quantified by the high standard deviation in Table 1), the InBi satisfactorily bonded to the acrylate primary coating around the entire fiber surface. The Si and C signals are included with the EDS maps to show the extension of the silica cladding and acrylate coating, respectively. Both In and Bi EDS signals are shown in Fig. 4, confirming the presence and extension of the InBi secondary coating.
Fig. 4 SEM micrograph and corresponding EDS maps of an A-InBi fiber sample comprised of an acrylate primary coating and a InBi secondary coating. Note the uniform bonding on the entire fiber. The carbon signal collected outside of the InBi coating is due to the epoxy mount (which surrounds the fiber and holds it in place in the sample holder to allow for polishing and imaging).
The SEM micrographs and EDS maps of the other samples are shown in Figs. 5-9. In all the samples, the metallic coating wets the entire fiber surface, except for the InBi coating on silica cladding (i.e., without the acrylate coating). Figure 9 indicates that the InBi did not bond uniformly to the cladding surface and about half of the fiber surface is uncoated. Multiple attempts were made to coat InBi directly to the fiber without acrylate but with the same result.
Fig. 5 A-Sn fiber sample comprised of an acrylate primary coating and a Sn secondary coating. Note the uniform bonding on the entire fiber. Similar to Fig. 4, the carbon signal collected outside of the Sn coating is due to the epoxy mount.
Fig. 6 A-PbSn fiber sample comprised of an acrylate primary coating and a PbSnAg secondary coating. Note the uniform bonding on the entire fiber, similarly to Fig. 4 and 5. The carbon signal collected outside of the tin coating is due to the epoxy mount.
Fig. 7 PbSn fiber sample comprised of a PbSnAg coating on the SiO2 cladding. Note the uniform bonding of PbSnAg to the entire fiber surface. The Si and O EDS maps identify the SiO2 cladding whereas the Sn and Pb signals are the PbSn coating.
Fig. 8 Sn fiber sample comprised of a Sn coating on the SiO2 cladding. Note the uniform bonding of Sn to the entire fiber surface.
Fig. 9 InBi fiber sample comprised of an InBi coating on the SiO2 cladding. Note the only partial bonding to the fiber surface.
The results of sensitivity measurements on all coated fibers from 4.2 – 61 K are plotted in Fig. 10. The A-InBi sample shows the highest sensitivity throughout the investigated temperature range. The acrylate single layer becomes more sensitive than the A-PbSn and A-Sn composite coatings in the range 5 – 20 K. All coated samples follow the same trend: their thermal sensitivity increases as the temperature increases from 4.2 K to 50 K, and then decreases from 50 K to 60 K. Due to the lack of CTE and E data for the coatings materials as a function of temperature, in the temperature range of interest, the explanation of this sensitivity temperature behavior may only be speculative. Note however that the heat capacity of the tested fibers increases with temperature, as expected because the heat capacities of Sn, Pb and In increase with temperature [26], so the measurable signal, which represents the temperature change, decreases with increasing temperature. This implies that the increased spectral shift with increasing temperature must be due to an increased thermal sensitivity of the fibers. Moreover, the increase in thermal sensitivity with temperature that is common to all samples with an effective coating atop the silica cladding is accounted for by the thermal expansion of Sn, In and Pb increasing with temperature whereas the decrease in sensitivity at 45-50 K, which is also common to all samples, correlates with a decrease in thermal expansion of silica [27, 28], which is the core and cladding material of all samples.
Figure 11 presents the spectral shift results at 4.2 K for the different coated fiber samples. The sensitivity values shown are each an average of 5 measurements. These results show that for InBi and PbSn, the composite coating with the metal deposited atop the acrylate, results in a much higher spectral shift. For the Sn coating, the presence of the acrylate appears to have no effect, as the A-Sn and Sn coated fibers have nearly identical spectral shifts. We interpret these results as indicators that the combination of high CTE and high E is the underlying cause for the increase in sensitivity at low temperature, and that as the temperature decreases the more significant the increased sensitivity. Note that the low sensitivity in the InBi sample at 4.2 K, and the lack of a significant increase in sensitivity with increasing temperature, is consistent with the poor coating quality seen in Fig. 9 and quantified by the standard deviation of the coating thickness in Table 1.
Fig. 11 Thermal sensitivity results at 4.2 K for all coated samples. The value shown is the average of 5 measurements.
Another characteristic of the optical fiber sensor altered by coating is the recovery time, defined here as the time required for the spectral shift to return to zero; results for each coating at three temperatures are shown in Table 2. When analyzed as a function of temperature, the recovery time increases for all samples as the temperature increases. In nearly every case, when comparing at the same temperature, the fibers with metallic coatings have faster recovery times than the acrylate-only sample, and the difference is greater at 4.2 K than at higher temperatures. At 30 K, the acrylate no longer has the longest recovery time; in fact, A-PbSn and A-InBi have a longer recovery time than acrylate, which in turn takes longer to recover than PbSn and Sn, with InBi being the fastest. Therefore, at 30 K the recovery time correlates more with the thermal mass of the sample, since the InBi, which is only coating half the fiber, is the fastest, the acrylate is intermediate, and the composite coatings are the slowest. The average recovery time on all samples is 1.68 s at 4.2 K and becomes 3.87 s at 10 K and 7.09 s at 30 K. The increase in recovery time with increasing temperature is true for all samples and indicates that the product of thermal resistivity and heat capacity is increased. This is in agreement with available heat capacity and thermal conductivity data for Sn, Pb, and In. The heat capacity of these elements, as for most crystalline solids, increases with increasing temperature. Their thermal conductivities, however, decrease with increasing temperature in the range 4-50 K [26]. Therefore, the product of the heat capacity and the thermal resistivity (which is the inverse of the thermal conductivity), increases with temperature, explaining the observed trends in the recovery time with temperature for all the samples.
Table 2. Recovery Times
Figure 12 shows high resolution SEM and EDS maps of the InBi coating. A typical lamellar microstructure of an alloy with eutectic composition solidified from the melt is observed. In fact, this is exactly the process used to deposit the InBi coating material with eutectic composition. Figure 12 includes a higher magnification SEM micrograph and higher resolution EDS maps of the InBi coating that both strongly indicate the lamellar microstructure. The choice of eutectic composition of the metal alloys (InBi and PbSn) was motivated by the aim of having the lowest melting temperature possible, because lower melt temperature correlates to lower bond strength as well as a shallower, more asymmetric interatomic potential well, which translates to higher coefficient of thermal expansion [29]. In the case of InBi, this also translated to high sensitivity at low temperature.
Fig. 12 High magnification SEM micrograph (left) and EDS maps (right) showing the lamellar microstructure of the InBi solidified from the melt.
4. Conclusion
A melt coating process has been used to deposit metallic coatings on optical fibers with an exposed silica cladding and on fibers with an acrylate coating atop the silica cladding. These samples were cooled to temperatures ranging from 4.2 K to 61 K and their sensitivity to a fixed heat pulse was measured. Cross-sectional SEM micrographs and EDS maps showed successful bonding for PbSn and Sn coatings on both silica and acrylate surfaces, and for InBi on the acrylate surface. At 4.2 K, the most sensitive sample is the acrylate/InBi composite coating. All samples showed increasing sensitivity with temperature from 4.2 K to 50 K, and decreasing sensitivity as the temperature increases from 50 K to 60 K. The presence of a metallic coating reduces the recovery time at all temperatures, but in particular at 4.2 K. This effect decreases as temperature increases.
References and links
1. B. Lee, “Review of the present status of optical fiber sensors,” Opt. Fiber Technol. 9(2), 57–79 (2003). [CrossRef]
2. G. F. Fernando, D. J. Webb, and P. Ferdinand, “Optical-Fiber Sensors,” MRS Bull. 27(05), 359–364 (2002). [CrossRef]
3. V. Chohan, L. Walckiers, N. Smirnov, G. de Rijk, S. Sanfilippo, V. Remondino, P. Pugnat, M. Buzio, A. Siemko, and B. Vullierme, “Testing of the LHC magnets in cryogenic conditions: current experience and near future outlook,” 2004.
4. T. Okazaki, H. Sugimoto, and T. Takeda, “Liquid nitrogen cooled HTS motor for ship propulsion,” in 2006 IEEE Power Engineering Society General Meeting, (IEEE, 2006), p. 6.
5. X. R. Wang, A. R. Caruso, M. Breschi, G. M. Zhang, U. P. Trociewitz, H. W. Weijers, and J. Schwartz, “Normal zone initiation and propagation in Y-Ba-Cu-O coated conductors with Cu stabilizer,” IEEE Trans. Appl. Supercond. 15(2), 2586–2589 (2005). [CrossRef]
6. H. Song, K. Gagnon, and J. Schwartz, “Quench behavior of conduction-cooled YBa2Cu3O7-delta coated conductor pancake coils stabilized with brass or copper,” Supercond. Sci. Technol. 23(6), 065021 (2010). [CrossRef]
7. H. Song and J. Schwartz, “Stability and Quench Behavior of YBa2Cu3O7-x Coated Conductor at 4.2 K, Self-Field,” IEEE Trans. Appl. Supercond. 19(5), 3735–3743 (2009). [CrossRef]
8. X. Wang, U. P. Trociewitz, and J. Schwartz, “Near-adiabatic quench experiments on short YBa2Cu3O7-delta coated conductors,” J. Appl. Phys. 101(5), 053904 (2007). [CrossRef]
9. X. Wang, U. P. Trociewitz, and J. Schwartz, “Critical current degradation of short YBa2Cu3O7-delta coated conductor due to an unprotected quench,” Supercond. Sci. Technol. 24(3), 035006 (2011). [CrossRef]
10. L. Ye, D. Cruciani, T. Effio, F. Hunte, and J. Schwartz, “On the Causes of Degradation in Bi2Sr2CaCu2O8+x Round Wires and Coils by Quenching at 4.2 K,” IEEE Transactions on Applied Superconductivity 23(5), 6400811 (2013). [CrossRef]
11. L. Ye, D. Cruciani, M. Xu, S. Mine, K. Amm, and J. Schwartz, “Magnetic field dependent stability and quench behavior and degradation limits in conduction-cooled MgB2 wires and coils,” Supercond. Sci. Technol. 28(3), 035015 (2015). [CrossRef] [PubMed]
12. S. Mahar, J. Geng, J. Schultz, J. Minervini, S. Jiang, P. Titus, M. Takayasu, C.-y. Gung, W. Tian, and A. Chavez-Pirson, “Real-time simultaneous temperature and strain measurements at cryogenic temperatures in an optical fiber,” SPIE 7087, 70870I (2008).
13. S. B. Mahar, “Spontaneous Brillouin Scattering Quench Diagnostics for Large Superconducting Magnets,” Ph.D. dissertation, Massachusetts Institute of Technology, Department of Nuclear Science and Engineering, 2008.
14. F. Scurti, S. Ishmael, G. Flanagan, and J. Schwartz, “Quench detection for high temperature superconductor magnets: a novel technique based on Rayleigh-backscattering interrogated optical fibers,” Supercond. Sci. Technol. 29(3), 03LT01 (2016). [CrossRef]
15. F. Scurti and J. Schwartz, “Optical fiber distributed sensing for high temperature superconductor magnets,” Proc. SPIE (to be published).
16. M. Marchevsky and S. A. Gourlay, “Acoustic thermometry for detecting quenches in superconducting coils and conductor stacks,” Appl. Phys. Lett. 110(1), 012601 (2017). [CrossRef]
17. M. Marchevsky, G. Sabbi, H. Bajas, and S. Gourlay, “Acoustic emission during quench training of superconducting accelerator magnets,” Cryogenics 69, 50–57 (2015). [CrossRef]
18. F. Hunte, H. Song, J. Schwartz, R. Johnson, and M. Turenne, “Fiber Bragg optical sensors for YBCO applications,” in Particle Accelerator Conference,Vancouver, Canada, 2009.
19. M. Turenne, R. Johnson, F. Hunte, J. Schwartz, and H. Song, “Multi-purpose fiber optic sensors for high temperature superconducting magnets,” in Fusion Engineering,2009. SOFE 2009. 23rd IEEE/NPSS Symposium on (2009), pp. 1–4. [CrossRef]
20. F. Scurti, S. Sathyamurthy, M. Rupich, and J. Schwartz, “Self-monitoring “SMART” REBCO coated conductor via integrated optical fibers,” Supercond. Sci. Technol. (to be published).
21. S. Yin, P. B. Ruffin, and F. T. S. Yu, Fiber Optic Sensors, Second Edition: (CRC Press, 2008).
22. A. D. Kersey, M. A. Davis, H. J. Patrick, M. LeBlanc, K. P. Koo, C. G. Askins, M. A. Putnam, and E. J. Friebele, “Fiber grating sensors,” J. Lightwave Technol. 15(8), 1442–1463 (1997). [CrossRef]
23. M. B. Reid and M. Ozcan, “Temperature dependence of fiber optic Bragg gratings at low temperatures,” Opt. Eng. 37(1), 237–240 (1998). [CrossRef]
24. Y. Li, Z. Hua, F. Yan, and P. Gang, “Metal coating of fiber Bragg grating and the temperature sensing character after metallization,” Opt. Fiber Technol. 15(4), 391–397 (2009). [CrossRef]
25. L. Xiao Chun, P. Fritz, and S. John, “Thermal behavior of a metal embedded fiber Bragg grating sensor,” Smart Mater. Struct. 10(4), 575–579 (2001). [CrossRef]
26. J. Jensen, R. G. Stewart, W. Tuttle, and H. Brechna, Brookhaven National Laboratory Selected Cryogenic Data Notebook: Sections I–IX1 (Brookhaven National Laboratory, 1980).
27. B. United States. National Bureau of Standards. Cryogenic Engineering Laboratory, Colo and V. J. Johnson, Properties of Materials at Low Temperature (Phase 1): A Compendium (Pergamon Press, 1961).
28. R. Rajini-Kumar, M. Suesser, K. G. Narayankhedkar, G. Krieg, and M. D. Atrey, “Performance evaluation of metal-coated fiber Bragg grating sensors for sensing cryogenic temperature,” Cryogenics 48(3-4), 142–147 (2008). [CrossRef]
29. J. Dugdale and D. MacDonald, “The thermal expansion of solids,” Phys. Rev. 89(4), 832–834 (1953). [CrossRef]
