sensors Article
An Improved Metal-Packaged Strain Sensor Based on A Regenerated Fiber Bragg Grating in Hydrogen-Loaded Boron–Germanium Co-Doped Photosensitive Fiber for High-Temperature Applications Yun Tu 1 , Lin Ye 2 , Shao-Ping Zhou 1 and Shan-Tung Tu 1, * 1
2
*
Key Laboratory of Pressure Systems and Safety (Ministry of Education), School of Mechanical and Power Engineering, East China University of Science and Technology, Shanghai 200237, China;
[email protected] (Y.T.);
[email protected] (S.-P.Z.) Laboratory of Smart Materials and Structures, Centre for Advanced Materials Technology, School of Aerospace, Mechanical and Mechatronic Engineering, The University of Sydney, Sydney 2006, Australia;
[email protected] Correspondence:
[email protected]; Tel.: +86-21-6425-3425
Academic Editor: Vittorio M. N. Passaro Received: 30 November 2016; Accepted: 25 January 2017; Published: 23 February 2017
Abstract: Local strain measurements are considered as an effective method for structural health monitoring of high-temperature components, which require accurate, reliable and durable sensors. To develop strain sensors that can be used in higher temperature environments, an improved metal-packaged strain sensor based on a regenerated fiber Bragg grating (RFBG) fabricated in hydrogen (H2 )-loaded boron–germanium (B–Ge) co-doped photosensitive fiber is developed using the process of combining magnetron sputtering and electroplating, addressing the limitation of mechanical strength degradation of silica optical fibers after annealing at a high temperature for regeneration. The regeneration characteristics of the RFBGs and the strain characteristics of the sensor are evaluated. Numerical simulation of the sensor is conducted using a three-dimensional finite element model. Anomalous decay behavior of two regeneration regimes is observed for the FBGs written in H2 -loaded B–Ge co-doped fiber. The strain sensor exhibits good linearity, stability and repeatability when exposed to constant high temperatures of up to 540 ◦ C. A satisfactory agreement is obtained between the experimental and numerical results in strain sensitivity. The results demonstrate that the improved metal-packaged strain sensors based on RFBGs in H2 -loaded B–Ge co-doped fiber provide great potential for high-temperature applications by addressing the issues of mechanical integrity and packaging. Keywords: regenerated fiber Bragg grating (RFBG); metal-packaged; strain sensor; photosensitive fiber; high temperature; strength degradation; structural health monitoring
1. Introduction As a result of the global energy deficit and environmental deterioration, most plants tend to be larger scale in operations at higher operating parameters in order to improve energy conversion efficiency and productivity with reduced environmental impact. This could, unfortunately, lead to unexpected and fatal industrial accidents due to material deterioration at high temperatures leading significant loss in assets, and at times, even human life. It is most improbable that the integrity of a component before operation could be guaranteed merely using a conventional design due to the
Sensors 2017, 17, 431; doi:10.3390/s17030431
www.mdpi.com/journal/sensors
Sensors 2017, 17, 431
2 of 18
time-varying service conditions, particularly at a high temperature [1]. Structural health monitoring is thus an essential procedure to ensure the safety of the plant by providing real-time, reliable and accurate information about the condition of critical components. Since the local strain has been used as an indicator of the creep condition of high-temperature components, given the monotonic relationship between strain and creep life, strain measurements have been considered as the most reliable method for structural health monitoring of these components. However, traditional strain gauges are not reliable for prolonged measurements at high temperatures [2], in addition to nonlinearity of thermal-induced apparent strains and susceptibility to electromagnetic interference (EMI). Measurement accuracy with optical strain gauges and digital image correlation (DIC) is also adversely affected by insufficient image quality due to the inevitable degradation of markers exposed to high temperatures with time [3]. Therefore, strain measurements at high temperatures, where it is difficult to guarantee reliability and durability of the sensors, have been a long-standing challenge, with numerous potential industrial applications over decades, e.g., in the processing, energy and aerospace industry. Optical fiber sensors (OFSs) are well suited to structural health monitoring due to a number of advantages over their electrical counterparts, such as small size, light weight, electrically passive operation, high sensitivity, and resistance to electromagnetic interference and corrosion. In particular, wavelength encoded fiber Bragg grating (FBG) sensors have inherent self-referencing and wavelength multiplexing capabilities that allow them to be easily serialized in a single optical fiber and spliced to telecommunication fibers for remote, distributed and multi-parameter sensing [4]. However, conventional type-I FBGs exhibiting a strong decay at high temperatures can only operate in principle up to 300 ◦ C for lengthy periods [5]. Many of research efforts have been expended towards the investigation of thermally stable gratings at high temperatures, including formation of type-In (type-IIA) gratings [6,7] and type-II gratings [8], writing by femtosecond lasers [9], formation of surface relief FBGs [10], and tailoring of glass composition [11–13]. In the past decade, another variant, regenerated fiber Bragg gratings (RFBGs), with superior high-temperature stability has been found and considered as the essential potential for high-temperature applications [14,15]. As fiber-optic strain sensors must monitor strains and ensure long-term reliability in adverse environmental conditions, appropriate sensor packaging and attachment are crucial for their operation and lifetime. A common solution is to encapsulate an optical fiber sensor with an epoxy, which is indeed a simple method at room temperature, but most epoxies degrade after exposure to high temperatures over 400 ◦ C. To overcome the limitation of this conventional encapsulation method, an all-metal packaging process has been proposed for protecting the bare RFBGs from environmental attack as well as easily attaching the RFBG-based strain sensors to metallic high-temperature components [16,17]. The process generally includes two steps. In the first step, one- or two-layer metallic films are typically deposited on the bare RFBGs as an adhesive and/or conductive layer by low temperature processes such as electroless plating [16,18], physical vapor deposition (e.g., magnetron sputtering [17,19] and evaporation deposition [20]), and laser-assisted maskless micro-deposition [21]. This occurs in order to achieve reliable bonding between glass and metal, in addition to allowing to electroplate nickel coating on the metallic films as a protective layer. After that, the metal-coated RFBGs are embedded into the metallic substrate of a high melting temperature by using brazing [16], ultrasonic consolidation [22], laser additive manufacturing [19,23], or nickel electroplating [17,21]. Careful selection of the packaging materials and processes is critical to ensure the long-term survivability of the package itself under high temperatures, and to achieve a strong and reliable glass-to-metal bond without mechanical or thermal damage. Besides the packaging issue, the mechanical integrity of optical fibers after annealing at a high temperature for regeneration is another vital issue which impedes the practical utility of the RFBG-based strain sensors, given that the silica fibers are found to become mechanically clearly weaker upon annealing, as reported by the authors and other researchers [24–27]. It is observed that cooling to room temperature after heating the silica fibers to a high temperature induces an additional significant
Sensors 2017, 17, 431
3 of 18
reduction in mechanical strength compared to those tested at this high temperature at which annealing occurs. The degree of weakening becomes more severe with an increased temperature of annealing. Accordingly, the metal-packaged strain sensors based on the use of the RFBGs fabricated in standard telecommunication silica fibers (Corning Inc., SMF-28, Corning, NY, USA) can only be used up to 400 ◦ C [17], as the mechanical strength of the standard SMF-28 silica optical fibers in which the RFBGs require the regeneration temperature of around 900 ◦ C was observed to considerably degrade after annealing at 900 ◦ C [26]. However, annealing at a high temperature is necessary for the fabrication of the RFBGs, although it usually leads to mechanical strength degradation of the silica optical fibers. The annealing temperature thus needs to be carefully chosen as a function of the anticipated operating temperature and duration in order to extend the application of RFBG to high-temperature conditions. In this paper, the objective is to develop an RFBG-based strain sensor capable of application in higher temperature environments, e.g., fossil-fuel power plants operating at temperatures up to 540 ◦ C. A metal-packaged strain sensor prototype is developed using an RFBG fabricated in hydrogen (H2 )-loaded boron–germanium (B–Ge) co-doped photosensitive fiber (Fibercore Ltd., PS1250/1500, Southampton, UK). The PS1250/1500 fiber, in which the RFBGs require a relatively lower regeneration temperature (of 500 ◦ C) will lead not only to better mechanical integrity but also to better high-temperature stability than the SMF-28 fiber that was used in fabricating the high-temperature strain sensors [17]. Thus, the RFBGs in PS1250/1500 fiber are chosen as base-sensing elements in the present work, to develop high-temperature strain sensors. The fabrication process of the strain sensor prototype is elaborated, and the sensor prototype and its corresponding bare RFBG are characterized at high temperatures under uniaxial tensile loading. Numerical simulation of the sensor is carried out based on the basic principle of strain measurements using FBGs and three-dimensional (3-D) finite element (FE) modeling to analyze the mechanical response of the metal-packaged strain sensor. 2. Strain Sensing Principles of Fiber Bragg Gratings An FBG is formed as a periodic variation in the refractive index of the core of a photosensitive single-mode optical fiber. When a broadband light source is coupled to the optical fiber containing an FBG, the grating diffractive properties promote that only a very narrow wavelength band is back-reflected. Thus, the use of FBG sensors relies on the determination of the center wavelength of the back-reflected narrow band, called the Bragg wavelength, λB , defined by the Bragg condition [28] λB = 2neff Λ
(1)
where neff is the effective refractive index of the fiber core and Λ is the grating period. Both neff and Λ are affected by changes in deformation and temperature. Using Equation (1), the shift in the Bragg wavelength, ∆λB , due to axial strain and temperature changes, is given by [29]:
∂neff ∂Λ ∂neff ∂Λ ∆λB = 2 Λ + ne f f ∆l + 2 Λ + ne f f ∆T ∂l ∂l ∂T ∂T
(2)
where ∆l and ∆T are the changes in grating length and temperature, respectively. The first term in Equation (2) represents the strain effect on an optical fiber. When the fiber is only axially strained, the Bragg wavelength varies due to the changes in the grating period and the photoelastic-induced changes in the refractive index. In this case, the strain effect term in Equation (2) can be expressed as ( ) n2eff ∆λB = λB ε z − (3) [ p12 ε z + ( p11 + p12 )ε r ] 2 where p11 and p12 are the components of the strain-optic tensor, and ε z and ε r are the axial and radial strains in the optical fiber respectively. Typical values for the strain-optic constants and effective refractive index for typical B–Ge co-doped photosensitive fibers are p11 = 0.113, p12 = 0.252 and
Sensors 2017, 17, 431
4 of 18
neff = 1.455 [30]. For an FBG with Bragg wavelength of 1550 nm, the typical strain sensitivity is approximately a 1.2-pm change in Bragg wavelength as a result of applying a strain of 10−6 to the grating. 3. Metal-Packaged Strain Sensor Prototype The study and development of the improved metal-packaged strain sensor prototype using an RFBG fabricated in H2 -loaded PS1250/1500 photosensitive fiber as the sensing element are articulated in three aspects. 3.1. Fabrication of Strain Sensor Prototype The fabrication of the metal-packaged RFBG strain sensor prototype using the metal-coated RFBG embedded in the steel substrate could be divided in to four steps: Step 1: Customized type-I seed FBGs used in the present work were written in H2 -loaded PS1250/1500 photosensitive fiber including 10 mol % of GeO2 and 14–18 mol % of B2 O3 through a phase mask with a reflectivity of ~80%, a 3-dB reflection bandwidth of less than 0.3 nm and a grating length of 8 mm, which were produced by a commercial company. Step 2: RFBGs were regenerated by annealing process. The type-I seed FBG was loosely placed in a capillary quartz tube (1 mm inside diameter and 2.5 mm outside diameter) horizontally inserted into the central area of a horizontal miniature tube furnace in order to maintain the temperature uniform along the grating and avoid any mechanical or thermal strain on the fiber. Then, a calibrated armored N-type thermocouple with a precision of ±0.5 ◦ C was assembled together with the capillary quartz tube, and its tip was adjacent to the grating in order to control the temperature. It was built into the furnace’s feedback during the annealing throughout which the spectral behavior of the grating was monitored in reflection by a commercial FBG interrogator (Micron Optics, Inc., Sm125-500, Atlanta, GA, USA). The tabular furnace is custom-made with a length of 40 mm to anneal a short length of the fiber, as the mechanical strength of the fiber is considerably reduced by the annealing treatment and polymer coating is also removed via thermal stripping. Coating is essential to protect the fiber surface from handling damage during subsequent fabrication. The furnace is equipped with a quartz tube with an inside diameter of 8 mm and a length of 40 mm to homogenize the temperature inside the furnace. The furnace temperature was raised from room temperature to 500 ◦ C within 50 min, and subsequently held constant at the temperature of 500 ◦ C for ~120 min, which is the temperature triggering regeneration. To the best of the present authors’ knowledge, this temperature is the lowest regeneration temperature observed for the B–Ge co-doped photosensitive fiber, where a typical annealing process used for regeneration of RFBGs is shown in Figure 1. As the temperature approached 500 ◦ C, the reflection peak power of the seed FBG started to decay drastically until it fell below the noise floor in a few minutes after the temperature reached 500 ◦ C. After that, a regenerated grating appeared at longer wavelength after ~8 min at the temperature of 500 ◦ C where the isothermal annealing was performed for ~120 min, followed by its reflection strength being increased gradually to a maximum until stability. No variation was observed in the strength of the RFBG as the temperature was decreased to room temperature over 90 min. The reflection spectra of the RFBG and its corresponding seed FBG was recorded at room temperature of 21 ◦ C by the FBG interrogator. Step 3: The multilayer metal-coated optical fiber containing the RFBG was fabricated by depositing a titanium (Ti) film on the fiber as an adhesive layer followed by a silver (Ag) film on the adhesive layer as a conductive layer by magnetron sputtering, and subsequently a nickel (Ni) coating on the conductive layer as a protective layer by electroplating. The details of this step have been described in our previous work [31]. Step 4: A series of experiments was carried out for successfully packaging the multilayer metal-coated RFBG into a P91 steel substrate to obtain a metal-packaged strain sensor prototype by using the RFBG fabricated in H2 -loaded PS1250/1500 photosensitive fiber by means of the same
Sensors 2017, 17, 431
5 of 18
process of all-metal packaging used for the RFBG fabricated in standard SMF-28 telecommunication fiber in our previous work [17]. Sensorsdescribed 2017, 17, 431 5 of 17
Figure 1. Typical evolution of the reflection peak power and the Bragg wavelength shift of one grating Figure 1. Typical evolution of the reflection peak power and the Bragg wavelength shift of one grating with the corresponding temperature profile during annealing for fabrication of the regenerated fiber with the corresponding temperature profile during annealing for fabrication of the regenerated fiber Bragg grating (RFBG) in H2-loaded PS1250/1500 fiber. Bragg grating (RFBG) in H2 -loaded PS1250/1500 fiber.
3.2. Characterization of Strain Sensor Prototype 3.2. Characterization of Strain Sensor Prototype For the determination of strain characteristics of metal-packaged RFBG strain sensors, uniaxial For the determination of strain characteristics of metal-packaged RFBG strain sensors, uniaxial tensile tests were carried out on a P91 steel sheet tensile test specimen with the attached sensor tensile tests were carried out on a P91 steel sheet tensile test specimen with the attached sensor prototype. The specimen has a gauge length of 70 mm with other dimensions shown in Figure 2. The prototype. The specimen has a gauge length of 70 mm with other dimensions shown in Figure 2. The sensor was mounted at the center of the gauge section along the center line using electrical resistance sensor was mounted at the center of the gauge section along the center line using electrical resistance spot welding at both ends, minimizing adverse thermal effects on the specimen. An spot welding at both ends, minimizing adverse thermal effects on the specimen. An electromechanical electromechanical universal test machine (MTS-SANS, CMT5504) instrumented with a furnace was universal test machine (MTS-SANS, CMT5504) instrumented with a furnace was used to produce used to produce strains in the test specimen by applying tensile loading, as described in detail in our strains in the test specimen by applying tensile loading, as described in detail in our previous work [17]. previous work [17]. The specimen was pinned to grips, taking care to avoid bending or torsion The specimen was pinned to grips, taking care to avoid bending or torsion loading to the sensor pigtail. loading to the sensor pigtail. The sensor was connected to the Sm125-500 FBG interrogator to assess The sensor was connected to the Sm125-500 FBG interrogator to assess the Bragg peak in the reflection the Bragg peak in the reflection spectrum measured during the mounting procedure. After that, the spectrum measured during the mounting procedure. After that, the temperature was increased at a temperature was increased at a uniform rate to the desired temperatures of 100 °C, 200 °C, 300 °C, uniform rate to the desired temperatures of 100 ◦ C, 200 ◦ C, 300 ◦ C, 400 ◦ C, 500 ◦ C and 540 ◦ C, and kept 400 °C, 500 °C and 540 °C, and kept constant for around 20 min before tensile testing. In order to constant for around 20 min before tensile testing. In order to maintain a nearly uniform temperature maintain a nearly uniform temperature over the gauge length of the specimen, the temperature over the gauge length of the specimen, the temperature gradient defined by the difference in three gradient defined by the difference in three calibrated N-type thermocouples was measured, with one calibrated N-type thermocouples was measured, with one mounted at the opposite side of the sensor mounted at the opposite side of the sensor and the other two mounted near two clamping fixtures. and the other two mounted near two clamping fixtures. The test temperature was controlled within The test temperature was controlled within a tolerance range of ±2 °C and the temperature differences a tolerance range of ±2 ◦ C and the temperature differences in the three points did not exceed 3 ◦ C in the three points did not exceed 3 °C during tensile testing. The tensile force was applied to the during tensile testing. The tensile force was applied to the specimen with a load interval of 1 kN or specimen with a load interval of 1 kN or 0.5 kN and held constant at each load level for 2 min to 0.5 kN and held constant at each load level for 2 min to obtain multiple measurements for defining the obtain multiple measurements for defining the average value. The maximum tensile force did not average value. The maximum tensile force did not exceed 8.0, 8.0, 8.0, 7.0, 6.5, 6.0, and 5.5 kN at the exceed 8.0, 8.0, 8.0, 7.0, 6.5, 6.0, and 5.5 kN at the corresponding test temperatures of room corresponding test temperatures of room temperature (26.5 ◦ C), 100 ◦ C, 200 ◦ C, 300 ◦ C, 400 ◦ C, 500 ◦ C temperature (26.5 °C), 100 °C, 200 °C, 300 °C, 400 °C, 500 °C and 540 °C, to avoid undesirable plastic and 540 ◦ C, to avoid undesirable plastic deformation of the P91 steel specimen. deformation of the P91 steel specimen. In order to compare the strain characteristics of bare and packaged RFBG sensors, uniaxial tensile tests were also performed on the bare RFBG by using the test apparatus described in our previous work [17]. The fiber containing the RFBG was carefully wrapped around the two capstans and its ends were mechanically griped. The fiber was heated to the temperatures of room temperature (21 ◦ C), 100 ◦ C, 200 ◦ C, 300 ◦ C, 400 ◦ C, 500 ◦ C and 600 ◦ C and kept at each temperature for at least 20 min before being tested. During the test, the fiber was stretched at a load interval of 0.2 N and kept constant at each load level for 2 min during which the temperature was held constant to obtain multiple measurements for defining the average value. The maximum force of 4 N, corresponding to approximately 0.4%
Figure 2. Pin-loaded sheet tensile test specimen with 70-mm gauge length.
maintain a nearly uniform temperature over the gauge length of the specimen, the temperature gradient defined by the difference in three calibrated N-type thermocouples was measured, with one mounted at the opposite side of the sensor and the other two mounted near two clamping fixtures. The test temperature was controlled within a tolerance range of ±2 °C and the temperature differences in the2017, three did not exceed 3 °C during tensile testing. The tensile force was applied to Sensors 17,points 431 6 ofthe 18 specimen with a load interval of 1 kN or 0.5 kN and held constant at each load level for 2 min to obtain multiple measurements for defining the average value. The maximum tensile force did not strain, to 7.0, the fiber thekN RFBG the rupture test of the fiber, considering the exceedwas 8.0,applied 8.0, 8.0, 6.5, containing 6.0, and 5.5 at to theavoid corresponding temperatures of room ◦ C not exceeding 8 N as fracture force of silica optical fiber after annealing at a temperature of 500 temperature (26.5 °C), 100 °C, 200 °C, 300 °C, 400 °C, 500 °C and 540 °C, to avoid undesirable plastic discussed in our previous work [26]. deformation of the P91 steel specimen.
Figure 2. Pin-loaded sheet tensile test specimen with 70-mm gauge length. Figure 2. Pin-loaded sheet tensile test specimen with 70-mm gauge length.
The load and the corresponding Bragg wavelength were recorded simultaneously from the load cell and the FBG interrogator during tensile tests of both the P91 steel specimen with the RFBG sensor and the bare fiber with the RFBG. 3.3. Numerical Modelling The 3-D FE modeling approach, which has been proposed and detailed in our previous work [17], was applied to modeling of the metal-package RFBG strain sensor prototype in the present work in order to find the state of stress and strain in the embedded optical fiber. The mechanical properties of the materials used in the 3-D FE analysis are listed in Table 1. One half of the structural model of a specimen with a metal-packaged RFBG strain sensor attached was discretized with 134,952 hexahedral elements (SOLID185) due to the symmetry of both the geometry and loading. One half of the metal-packaged RFBG sensor was selected for mesh refinement and discretized with 125,952 hexahedral mesh elements, as shown in Figure 3a. As seen, finer mesh sizes were chosen at the location of the optical fiber and the sputtered and electroplated metallic layers. The surface to be connected on the specimen was defined as the target element (TARGE170), whilst the surface to be connected on the substrate was defined as the contact element (CONTA173). Structural loads are transferred from the surface of the specimen to the surface of the substrate via the spot-weld connection points. The boundary conditions and the load were applied to the one half of the structural model, as shown in Figure 3b. Sensors 2017, 17, 431
7 of 17
Figure Three-dimensional (3-D) (3-D) finite finite element element (FE) (FE)model: model:(a) (a)one onehalf halfofofa ameshed meshedmodel modelofofa Figure 3. 3. Three-dimensional ametal-packaged metal-packagedRFBG RFBGstrain strainsensor, sensor,and and(b) (b)boundary boundaryconditions conditionsand andapplied appliedload. load.
4. Results and Discussion 4.1. Mechanical Strength Degradation of Silica Optical Fibers after Annnealing at High Temperatures Annealing at a high temperature is necessary for the fabrication of the RFBGs, although it usually
Sensors 2017, 17, 431
7 of 18
Table 1. Material parameters for three-dimensional (3-D) finite element (FE) analysis [32–34]. Temperature (◦ C)
Parameter Efiber (GPa) νfiber Etitanium (GPa) νtitanium Esilver (GPa) νsilver Enickel (GPa) νnickel EP91 (GPa) νP91 µP91
26.5
100
200
300
400
500
540
72.9 0.17 116 0.34 76 0.37 217 0.31 220 0.29 0.15
73.8 0.17 112 0.34 71 0.37 201 0.31 216 0.29 0.15
74.95 0.17 106 0.34 65 0.37 180 0.31 210 0.29 0.15
76.04 0.17 100 0.34 59 0.37 194 0.31 204 0.30 0.15
77.036 0.17 95 0.34 52 0.37 204 0.31 195 0.29 0.15
77.936 0.17 89 0.34 46 0.37 195 0.31 185 0.30 0.15
78.28 0.17 87 0.34 43.5 0.37 191 0.31 179 0.29 0.15
The axial strains of the optical fiber were obtained from the 3-D FE simulation by obtaining an average strain for the nodes along the symmetry axis of the fiber and over 8-mm gauge length equivalent to the grating length of RFBG. Accordingly, the numerical results of the shifts in Bragg wavelength were calculated by substituting the axial and radial strains of the optical fiber into Equation (3). 4. Results and Discussion 4.1. Mechanical Strength Degradation of Silica Optical Fibers after Annnealing at High Temperatures Annealing at a high temperature is necessary for the fabrication of the RFBGs, although it usually leads to mechanical strength degradation of the silica optical fibers. It is thus essential to quantify the effects of annealing on the tensile strength of silica optical fibers. A summary of the results from the tensile tests of the silica optical fibers reported in our previous work [26] is shown in Figure 4. The fracture stress of all annealed fibers decreases precipitously after annealing at 500 ◦ C and 900 ◦ C. The mean tensile strengths determined by the fracture stresses collected from the tensile testing from 15 samples are shown in Figure 5. The results show that the annealing treatment leads to a significant reduction in the strength of the silica optical fibers after annealing at high temperatures. A particularly interesting result is that the strength of the annealed fibers is not only very much lower than that of the samples without annealing tested at room temperature, but it is also much lower than those of the samples tested at the temperatures of 300 ◦ C and 540 ◦ C, which is very similar to the behavior observed by others [24,27]. Figure 5 also shows that the higher temperature of annealing at 900 ◦ C could lead to a larger reduction in strength of the fibers compared to those annealed at 500 ◦ C. Silica fibers were found to become mechanically weaker after being annealed in air but the cause of such weakening was not well defined. Very recently, it was found that surface crystallization is probably responsible for the mechanical weakening observed in silica glass fiber surface after annealing at temperatures in excess of around 800 ◦ C, while water diffusion-controlled virtual pitting of the glass surface is likely the source for the strength degradation at lower temperatures [27]. The results of the tensile tests indicate that the regeneration process of the RFBGs would considerably reduce the mechanical strength of silica optical fibers, even if careful preparations are integrated during the regeneration process. Moreover, the mechanical strength decreased significantly with the increased temperature of annealing at which the regeneration occurs. In order to extend the application of RFBG to high-temperature conditions, the annealing temperature needs to be carefully chosen as a function of the anticipated operating temperature and duration. To the best of the present authors’ knowledge, the annealing temperature of around 900 ◦ C is the lowest temperature at which regeneration occurs for seed FBGs in H2 -loaded SMF-28 fiber, whereas the annealing temperature of around 500 ◦ C is the lowest temperature at which regeneration occurs for seed FBGs in PS1250/1500
Sensors 2017, 17, 431
8 of 18
photosensitive fiber. Accordingly, the mechanical strength of the PS1250/1500 fiber annealed at the lower temperature of 500 ◦ C is expected to be higher than that of the SMF-28 fiber annealed at the higher temperature of 900 ◦ C. Therefore, The H2 -loaded PS1250/1500 fiber in which the RFBGs require a relatively lower regeneration temperature of 500 ◦ C is a better choice as the material of sensing elements to develop high-temperature strain sensors. Sensors 2017, 17, 431
8 of 17
Sensors 2017, 17, 431
8 of 17
Figure 4. Weibull plot of the tensile strengths of the bare silica optical fibers annealed at 500 °C and
Figure900 4. Weibull plot of the tensile strengths of the bare silica optical fibers annealed at 500 ◦ C and °C respectively and tested at room temperature (26 °C), compared with those of the bare fibers Figure 4. Weibull plot of the tensile strengths of the bare silica optical fibers annealed at 500 °C and ◦ ◦ 900 Cwithout respectively andtested tested at room temperature (26 compared with the bare fibers annealing at room temperature (26 °C), 300 C), °C and 540 °C, as wellthose as theof as-received 900 °C respectively and tested at room temperature◦ (26 °C), ◦compared with those of the bare fibers without annealing at room temperature (26room C), temperature 300 C and 540 ◦ C, as well as the as-received fibers tested attested room temperature (26 °C) [26]. RT: without annealing tested at room temperature (26 °C), 300 °C and 540 °C, as well as the as-received fibers tested at room temperature (26 ◦ C) [26]. RT: room temperature fibers tested at room temperature (26 °C) [26]. RT: room temperature
Figure 5. Mean tensile strengths of the bare silica optical fibers annealed at 500 °C and 900 °C respectively and tested at room temperature (26 °C), compared with those of the bare fibers ◦ Cwithout 5. Mean tensile strengthsofofthe the bare bare silica silica optical annealed at 500 900 FigureFigure 5. Mean tensile strengths opticalfibers fibers annealed at °C 500and and°C900 ◦ C annealing tested at room temperature (26 °C), 300 °C and 540 °C [26]. respectively and tested at room temperature (26 °C), compared with those of the bare fibers without ◦ respectively and tested at room temperature (26 C), compared with those of the bare fibers without annealing tested at room temperature (26◦ °C), 300 °C ◦ and 540 °C◦ [26].
annealing tested atCharacteristics room temperature (26 C),Fiber 300 Bragg C andGratings 540 C [26]. 4.2. Regeneration of Regenerated
4.2. Regeneration Characteristics of Regenerated Fiber Bragg Figure 1 shows the evolution of the reflection peakGratings power and the Bragg wavelength shift of a
4.2. Regeneration Regenerated Fiber Bragg Gratings type-I FBG Characteristics in PS1250/1500offiber as a function of duration of the annealing. In addition, the Figure 1 shows the evolution of the reflection peak power and the Bragg wavelength shift of a temperature measured with the N-type thermocouple is also shown in Figure 1. The behavior of the type-I 1FBG in PS1250/1500 fiber function of duration of and the the annealing. In addition, the Figure shows the evolution of as thea reflection peak power Bragg wavelength shift of a gratings in H2-loaded PS1250/1500 fiber is significantly different in comparison with that of the temperature measured with the N-type thermocouple is also shown in Figure 1. The behavior of the type-I FBG in PS1250/1500 fiber as a function of duration of the annealing. In addition, the temperature gratings in H2-loaded SMF-28 fiber whose behavior has been described in our previous work [31]. gratings inthe H2-loaded PS1250/1500 fiber isalso significantly different in comparison withofthat ofgratings the measured with N-type thermocouple shown in Figure The the in The type-I FBG in H2-loaded PS1250/1500 is fiber started to slowly decay1. after thebehavior temperature exceeded gratings in H2-loaded SMF-28 fiber whose behavior has been described in our previous work [31]. H2 -loaded is °C significantly in comparison withthethat of of the in 150 °CPS1250/1500 (correspondingfiber to ~200 for FBG in H2different -loaded SMF-28 fiber). However, decay thegratings FBG The type-I FBG in H2-loaded PS1250/1500 fiber started to slowly decay after the temperature exceeded stopped at the temperature of behavior ~300 °C, and subsequently the reflection peak power of the FBG started H2 -loaded SMF-28 fiber whose has been described in our previous work [31]. The type-I 150 °C (corresponding to ~200 °C for FBG in H2-loaded SMF-28 fiber). However, the decay of the FBG toHincrease slightly. Above ~430 °C, astarted rapid decrease in reflection peakthe power was observed up to the150 ◦ C FBG instopped fiber to slowly decay after temperature exceeded 2 -loaded at the PS1250/1500 temperature of ~300 °C, and subsequently the reflection peak power of the FBG started regeneration of 500 °C forH the-loaded FBG in H2-loaded fiber). PS1250/1500 fiber. The of the ◦ C for (corresponding totemperature ~200Above FBG However, thebehavior decay ofthe the FBG to increase slightly. ~430 °C, in a rapid in reflection peak power was observed up to 2 decreaseSMF-28 FBGs in H2-loaded PS1250/1500 fiber exhibiting two regeneration temperature regimes is very similar ◦ regeneration temperature 500 °C thesubsequently FBG in H2-loaded fiber.power The behavior thestarted stopped at the temperature ofof ~300 C,for and the PS1250/1500 reflection peak of the of FBG to that of the FBGs in H2-loaded GF1B fibers reported by Polz et al. [35]. In contrast, for the FBGs in ◦fiber FBGs in H2-loaded PS1250/1500 two regeneration temperature regimeswas is very similar up to to increase slightly. Above ~430one C, aexhibiting rapid decrease in reflection peak power observed H2-loaded SMF28 fiber, only regeneration regime above 900 °C was observed by the authors and to that of the FBGs in H2-loaded GF1B fibers reported by Polz et al. [35]. In contrast, for the FBGs in ◦ the regeneration temperature of 500 C for the in H2with -loaded PS1250/1500 fiber. The of other researchers [31,35]. Such behavior can beFBG associated the special type of fiber used. In behavior order H2-loaded SMF28 fiber, only one regeneration regime above 900 °C was observed by the authors and to achieve extremely high photosensitivity and to match mode field diameter (MFD) of the SMF-28 other researchers [31,35]. Such behavior can be associated with the special type of fiber used. In order to achieve extremely high photosensitivity and to match mode field diameter (MFD) of the SMF-28
Sensors 2017, 17, 431
9 of 18
the FBGs in H2 -loaded PS1250/1500 fiber exhibiting two regeneration temperature regimes is very similar to that of the FBGs in H2 -loaded GF1B fibers reported by Polz et al. [35]. In contrast, for the FBGs in H2 -loaded SMF28 fiber, only one regeneration regime above 900 ◦ C was observed by the authors and other researchers [31,35]. Such behavior can be associated with the special type of fiber used. In order to achieve extremely high photosensitivity and to match mode field diameter (MFD) of the SMF-28 fiber, the core of PS1250/1500 fiber is co-doped with boron and germanium. The two regimes of regeneration may thus be related to the different structural relaxations that occur 9inof the Sensors 2017, 17, 431 17 core and cladding where the dopant composition and concentrations differ, which has not been clarified to fiber, theinvestigations core of PS1250/1500 fiber isto co-doped withrelationship boron and germanium. The two regimes of with date. Further are needed clarify the the two regeneration regimes regeneration may thus be related to the different structural relaxations that occur in the core and the specific fiber structure, dopant composition and concentrations. cladding where the dopant composition and concentrations differ, which has not been clarified to In addition to the decay of the grating, a shift of the Bragg wavelength has also been observed, date. Further investigations are needed to clarify the relationship the two regeneration regimes with as shown in Figure 6.structure, The stepdopant of isothermal annealing at a temperature of 500 ◦ C is zoomed. In this the specific fiber composition and concentrations. step, as theInreflection power of the seedaFBG fast to the inflection point of complete addition topeak the decay of the grating, shift of thedecreases Bragg wavelength has also been observed, erasure, the Bragg wavelength is of abruptly shifted to shorter wavelengths, indicates a strong as shown in Figure 6. The step isothermal annealing at a temperature of 500which °C is zoomed. In this step, the reflection peak power the seed FBG fast decreases to the inflection pointhigh of complete decrease inasboth the refractive index of modulation and average index change at this temperature, Bragg wavelength is abruptly shifted to shorter wavelengths, which indicatesfiber a strong and iserasure, typicalthe behavior of thermal decay for a normal type-I FBG in PS1250/1500 [36]. As decrease in both the refractive index modulation and average index change at this high temperature, the reflection peak power of the RFBG increases from the inflection point, its Bragg wavelength is and is typical behavior of thermal decay for a normal type-I FBG in PS1250/1500 fiber [36]. As the much longer than that of its seed grating, consistent with the similar behavior reported in [37], which reflection peak power of the RFBG increases from the inflection point, its Bragg wavelength is much may be explained by the RFBG formed with changes (such as stress) at the core-cladding interface. longer than that of its seed grating, consistent with the similar behavior reported in [37], which may After that, a significant negative shift in changes the Bragg wavelength of RFBG indicating a reduction in be explained by the RFBG formed with (such as stress) at the core-cladding interface. After the average change is also observed until the gradually stabilizes that, a refractive significant index negative shift in the Bragg wavelength of wavelength RFBG indicating a reduction in the at the refractive annealing index change also◦ C. observed until of theawavelength gradually at the end for end ofaverage the isothermal ofis500 The trend negative shift in thestabilizes Bragg wavelength of thegrating isothermal of 500 °C.grating The trend of a negative shift inthe the grating Bragg wavelength for both both seed andannealing its regenerated is similar to that for in H2 -loaded SMF-28 seed grating and its regenerated grating is similar to that for the grating in H 2-loaded SMF-28 fiber fiber reported in our previous work [31] and also similar to that observed for regenerated type-IIA reported in our previous work [31] and also similar to that observed for regenerated type-IIA gratings gratings [38]. However, it is inconsistent with other researchers’ observations that have shown a trend [38]. However, it is inconsistent with other researchers’ observations that have shown a trend of a of a positive shift in the Bragg wavelength for the gratings in photosensitive fibers with different positive shift in the Bragg wavelength for the gratings in photosensitive fibers with different dopant dopantcomposition composition concentrations [35,37,39]. andand concentrations [35,37,39].
Figure 6. Evolution of the reflection peak power and the Bragg wavelength shift of the grating shown
Figure 6. Evolution of the reflection peak power and the Bragg wavelength shift of the grating shown in Figure 1 during the isothermal annealing at 500 °C. in Figure 1 during the isothermal annealing at 500 ◦ C. In sensing applications, it is desirable to have a good spectra profile and low reflectivity of
Insidelobes sensingwith applications, it ismain desirable haveFigure a good spectra profile andof low reflectivity of respect to the lobe ofto FBGs. 7 shows a comparison the reflection sidelobes with respect to theseed main lobe of FBGs. in Figure 7 showsfiber a comparison of theofreflection spectrum of the type-I grating inscribed PS1250/1500 to the spectrum its corresponding RFBG at room temperature (21 °C). The spectrum of the grating spectrum of the type-I seedobserved grating inscribed in PS1250/1500 fiber to the spectrum of seed its corresponding a broad bandwidth and noticeable with a Bragg of 1549.813 nm,aabroad RFBG presents observed at room temperature (21 ◦ C).sidelobes, The spectrum of thewavelength seed grating presents 3 dB reflection bandwidth of ~0.281 nm, a sidelobe suppression ratio of ~15.5 dB, and a reflectivity of bandwidth and noticeable sidelobes, with a Bragg wavelength of 1549.813 nm, a 3 dB reflection ~76.8%. In contrast to the seed grating, the RFBG provides a low reflectivity of ~9.6%, but has a good bandwidth of ~0.281 nm, a sidelobe suppression ratio of ~15.5 dB, and a reflectivity of ~76.8%. In spectral profile with a better-defined peak (1550.450 nm), a lower 3 dB reflection bandwidth contrast to the seed grating, the RFBG provides a low reflectivity of ~9.6%, but has a good spectral (~0.190 nm) and a larger reduction in sidelobe reflectivity (~−25.16 dB) compared to the main lobe. profileThe withspectral a better-defined (1550.450 nm), a lower 3 dB reflection bandwidth (~0.190 nm) and a quality of peak the RFBG is significantly improved, arising from the high-temperature annealing treatment, which is sufficiently suitable for the built-in peak detection algorithm of most commercial FBG interrogators and is well applicable for multiplexing. The reason for this is that the thermally-activated defects induced by the high intensity pulses are annealed out, leaving a very smooth interface between the structurally altered and the unaffected region with the material. The
Sensors 2017, 17, 431
10 of 18
larger reduction in sidelobe reflectivity (~−25.16 dB) compared to the main lobe. The spectral quality of the RFBG is significantly improved, arising from the high-temperature annealing treatment, which is sufficiently suitable for the built-in peak detection algorithm of most commercial FBG interrogators and is well applicable for multiplexing. The reason for this is that the thermally-activated defects induced by the high intensity pulses are annealed out, leaving a very smooth interface between the Sensors 2017, 17, 431 10 of 17 structurally altered and the unaffected region with the material. The reduction in overall reflectivity Sensors 2017, in 17, 10 of 17 of the reduction overall reflectivity and bandwidth is directly to the reduction of the and bandwidth is431 directly proportional to the reduction of proportional the refractive index modulation refractive index modulation of the grating [5]. It is also observed that the Bragg wavelength of the grating [5]. It is also observed that the Bragg wavelength of the RFBG is longer than that of its seed reduction in overall and bandwidth is directly proportional to◦ thewith reduction of the is longer than reflectivity that consistent of its seed grating room temperature, the consistent behavior gratingRFBG at room temperature, with theatbehavior observed consistent at 500 C but not with refractive index modulation of the grating [5]. It is also observed that the Bragg wavelength the observed at 500 °C but not consistent with the behavior of the RFBG in H2-loaded SMF-28offiber the behavior of the RFBG in H -loaded SMF-28 fiber observed in our previous work [31]. The reason 2of its seed grating at room temperature, consistent with the behavior RFBG is longer than thatwork observed in our previous [31]. The reason for the apparent spatial mismatch between the seed for theand apparent spatial mismatch between thethe seed andisregenerated the regenerated observed at 500 °C but not consistent with behavior of thetoRFBG in H2-loaded SMF-28 fiber regenerated gratings is that the regenerated grating likely have gratings formed atis thethat core-cladding observed in our previous work [31]. The reason for the apparent spatial mismatch between the seed gratinginterface is likely to have formed at the core-cladding interface boundary. boundary. and regenerated gratings is that the regenerated grating is likely to have formed at the core-cladding interface boundary.
7. Reflection spectra of the typicaltype-I type-I seed grating (FBG) and its corresponding FigureFigure 7. Reflection spectra of the typical seedfiber fiberBragg Bragg grating (FBG) and its corresponding RFBG in H2-loaded PS1250/1500 fiber measured at room temperature (21 °C). ◦ RFBG in H2 -loaded PS1250/1500 fiber measured at room temperature (21 C). Figure 7. Reflection spectra of the typical type-I seed fiber Bragg grating (FBG) and its corresponding
RFBGCharacteristics in H2-loaded PS1250/1500 fiber measured at room temperature (21 °C). 4.3. Strain of Sensor Prototype
4.3. Strain Characteristics of Sensor Prototype
FigureCharacteristics 8 shows a laboratorial prototype of a metal-packaged strain sensor based on use of the 4.3. Strain of Sensor Prototype
RFBG 8 fabricated H2-loaded PS1250/1500 bare RFBG, sputter-coated withbased titanium Figure shows ainlaboratorial prototypefiber. of a The metal-packaged strain sensor on and use of the Figure 8 shows a laboratorial of a metal-packaged strain sensor on use ofwith the films with thickness of prototype approximately 0.6 µm, andRFBG, electroplated withbased nickelwith coating RFBG silver fabricated in Ha2total -loaded PS1250/1500 fiber. The bare sputter-coated titanium and RFBG fabricated in H 2 -loaded PS1250/1500 fiber. The bare RFBG, sputter-coated with titanium and a thickness of around 200 µm, is embedded in P91 steel substrate by nickel electroplating. silver films with a total thickness of approximately 0.6 µm, and electroplated with nickel coating with silver films with a total thickness of approximately 0.6 µm, and electroplated with nickel coating with a thickness of around 200 200 µm,µm, is embedded substratebyby nickel electroplating. a thickness of around is embeddedininP91 P91 steel steel substrate nickel electroplating.
Figure 8. A laboratorial prototype of a metal-packaged strain sensor based on the RFBG in H2-loaded PS1250/1500 fiber. Figure 8. A laboratorial prototype of a metal-packaged strain sensor based on the RFBG in H2-loaded Figure 8. A laboratorial prototype of a metal-packaged strain sensor based on the RFBG in H2 -loaded PS1250/1500 fiber.the response to applied strain, the metal-packaged strain sensor based on the To characterize
PS1250/1500 fiber. RFBG fabricated in H2-loaded PS1250/1500 fiber was mounted on the P91 steel specimen by spot To characterize the was response to and applied strain, at theconstant metal-packaged strainofsensor on the welding. The specimen loaded unloaded temperatures room based temperature To(26.5 characterize thein response toPS1250/1500 applied strain, the metal-packaged sensor onspot the RFBG RFBG fabricated H°C, 2-loaded fiber was540 mounted on 9a–g the strain P91 steel specimen by °C), 100 °C, 200 300 °C, 400 °C, 500 °C and °C. Figure illustrates thebased wavelength welding. The specimen was RFBG loadedstrain and unloaded constant ofcalculated room temperature fabricated inofHthe PS1250/1500 fiber was mounted onfunction thetemperatures P91ofsteel specimen byfrom spotthe welding. shifts metal-packaged sensor as at a linear strains 2 -loaded (26.5 °C),forces, 100 °C,the 200cross-sectional °C, 300 °C, 400area °C, and 500 °C and 540 °C. Figure illustrates the wavelength applied Young’s modulus of 9a–g the P91 steel specimen at the shifts of the metal-packaged RFBG strain sensor as a linear function of strains calculated from the applied forces, the cross-sectional area and Young’s modulus of the P91 steel specimen at the
Sensors 2017, 17, 431
11 of 18
The specimen was loaded and unloaded at constant temperatures of room temperature (26.5 ◦ C), 100 ◦ C, 200 ◦ C, 300 ◦ C, 400 ◦ C, 500 ◦ C and 540 ◦ C. Figure 9a–g illustrates the wavelength shifts of the metal-packaged RFBG strain sensor as a linear function of strains calculated from the applied forces, the Sensors 2017, 17, 431 11 of 17 cross-sectional area and Young’s modulus of the P91 steel specimen at the corresponding temperature (as listed in bare TableRFBG 1). The responses the bare RFBG 2-loaded (ascorresponding listed in Tabletemperature 1). The responses of the in H PS1250/1500 fiberintoHthe strains 2 -loadedof PS1250/1500 fiber the strains determined from the forces also shownThe in Figure 9a–gshifts for determined from thetoapplied forces are also shown in applied Figure 9a–g forare comparison. observed The observed shifts in Bragg wavelength and strains show for both metalin comparison. Bragg wavelength and strains show linearity for both metal-packaged andlinearity bare RFBG sensors, with 2) higher than bare RFBG sensors, with the adjusted coefficients of determination thepackaged adjustedand coefficients of determination (adj R2 ) higher than 0.999 for the former(adj andR0.9999 for latter. 0.999 for the former and 0.9999 for latter. Thebehavior metal-packaged itsintegrity linear behavior The metal-packaged sensor preserves its linear implyingsensor a goodpreserves interfacial between implying a good interfacial integrity between every two layers and elastic deformations each every two layers and elastic deformations in each material. The strain sensitivities derived in from the material. The strain sensitivities derived from the slope of the straight lines in Figure 9a–g are 2.10, − 1 slope of the straight lines in Figure 9a–g are 2.10, 2.15, 2.12, 2.17, 2.15, 2.12 and 2.11 pm µε for the 2.15, 2.12, 2.17,RFBG 2.15, 2.12 and 2.11 loading pm µε−1 at forconstant the metal-packaged sensor under loading metal-packaged sensor under temperaturesRFBG of room temperature (26.5 ◦atC), constant temperatures of room temperature (26.5 °C), 100 °C, 200 °C, 300 °C 400 °C, 500 °C and 540 100 ◦ C, 200 ◦ C, 300 ◦ C 400 ◦ C, 500 ◦ C and 540 ◦ C, respectively, which is slightly higher than those −1 °C, respectively, which is slightly higher than those of 2.08, 2.10, 2.08, 2.15, 2.04, 2.09 and 2.06 pm µε of 2.08, 2.10, 2.08, 2.15, 2.04, 2.09 and 2.06 pm µε−1 under unloading, as elastic hysteresis occurs in under unloading, as elastic hysteresis occurs in the relatively flexible structural substrate. In addition, the relatively flexible structural substrate. In addition, at corresponding test temperatures, the values at corresponding test temperatures, the values are ~30% higher than the values of the metal-packaged are ~30% higher than the values of the metal-packaged strain sensor fabricated based on the RFBG strain sensor fabricated based on the RFBG in H2-loaded SMF-28 fiber reported in our previous work in H2 -loaded SMF-28 fiber reported in our previous work [17]. This could be mainly attributed to [17]. This could be mainly attributed to the differences in geometrical dimensions (the thickness of the differences in geometrical dimensions (the thickness of the electroplated nickel coating, the depth the electroplated nickel coating, the depth of the fiber embedded into the substrate, etc.) of the of packaged the fiber embedded into the substrate, the packaged structure fabricated manually, and structure fabricated manually,etc.) andofinaccuracy in the material parameters (Young’s inaccuracy in the material parameters (Young’s modulus, etc.) used to calculate the strains to which modulus, etc.) used to calculate the strains to which the steel specimen is subjected. The thinner thecoating steel specimen is subjected. The the thinner coating of of electroplated nickel and location of electroplated nickel and deeper location the fiber embedded intothe thedeeper substrate wouldof theresult fiberin embedded into the substrate would result in the higher strain sensitivity of the sensors. The the higher strain sensitivity of the sensors. The value of Young’s modulus for P91 steel used value of Young’s modulus for P91 steel used to calculate the strains to which the specimen is subjected to calculate the strains to which the specimen is subjected may be slightly greater than the true value. may be slightlythe greater than the trueare value. Accordingly, thestrains calculated strains than the Accordingly, calculated strains smaller than the true to which the are P91smaller steel specimen true strains to leading which the P91higher steel specimen subjected, leading the highershift sensitivity that is the is subjected, to the sensitivityisthat is the ratio of theto wavelength to the calculated ratio of the wavelength shift to the calculated strain. strain.
Figure 9. Cont.
Sensors 2017, 17, 431 Sensors 2017, 17, 431
12 of 18 12 of 17
Figure 9. Shift in the Bragg wavelength as a function of strain obtained from tensile tests at constant Figure 9. Shift in the Bragg wavelength as a function of strain obtained from tensile tests at constant temperatures of 26.5 °C (a); 100 °C (b); 200 °C (c); 300 °C (d); 400 °C (e); 500 °C (f) and 540 °C (g). FE: temperatures of 26.5 ◦ C (a); 100 ◦ C (b); 200 ◦ C (c); 300 ◦ C (d); 400 ◦ C (e); 500 ◦ C (f) and 540 ◦ C (g). FE: finite element. finite element.
The experimental results of the tensile tests performed on the bare RFBG in H2-loaded The experimental results of the tensile tests performed on the bare RFBG in H2 -loaded PS1250/1500 fiber are plotted in Figure 9a–g for comparison with those of the metal-packaged RFBG PS1250/1500 fiber are plotted in Figure 9a–g for comparison with those of the metal-packaged−1RFBG strain sensor. We can obtain the strain sensitivities of 1.21, 1.22, 1.22, 1.23, 1.25 and 1.28 pm µε for strain sensor. We can obtain the strain sensitivities of 1.21, 1.22, 1.22, 1.23, 1.25 and 1.28 pm µε−1 for the bare RFBG sensor under loading at constant temperatures of room temperature (21 °C), 100 °C,◦ ◦ the200 bare sensor loading at constantintemperatures of room temperature °C,RFBG 300 °C, 400 °Cunder and 500 °C respectively, contrast to those of 1.21, 1.21, 1.22, (21 1.23, C), 1.25100 andC, ◦ ◦ ◦ ◦ 200 300µε−1 C,under 400 C and 500 They C respectively, contrast tovalues those of 1.21,RFBG 1.22, in 1.23, 1.25 and 1.28C,pm unloading. are slightlyinhigher than of 1.21, the bare H2-loaded −1 under unloading. They are slightly higher than values of the bare RFBG in H -loaded 1.28 pm µε 2 SMF-28 fiber reported in our previous work [17]. The strain sensitivity of the metal-packaged RFBG SMF-28 reported ourthat previous workRFBG [17]. sensor The strain of the metal-packaged RFBG sensorfiber is ~70% higherin than of the bare due sensitivity to the flexible structure of the substrate sensor is ~70% higher that the of the bare RFBG sensor due to the flexible structure the substrate specially designed tothan enhance strain sensitivity. An approximation of the nominalofgauge length specially designed the strainRFBG sensitivity. approximation of theaxis nominal length (see Figure 8) of to theenhance metal-packaged strain An sensor in measurement is the gauge distance of (see Figure 8) of the metal-packaged RFBG strain sensor in measurement axis is the distance of between between the weld spots. The true gauge length of the strain sensitive section of the sensor is the length theofweld spots. The gauge length of the strain sensitive section offlexible the sensor is thewhich lengthisof metal-coated fibertrue containing the RFBG, measured by the inside of the structure, metal-coated containing the RFBG, measured the inside of flexible structure, which is shorter than fiber the nominal gauge length. The structuralby deformations to the be measured over the nominal gauge length mainly result in elongation over the true gauge length rather than elongation over the shorter than the nominal gauge length. The structural deformations to be measured over the nominal nominal gauge length, sinceinthe rigidity over section of the substrate without the flexible structure gauge length mainly result elongation overthe the true gauge length rather than elongation over the is muchgauge greater than since that over the section the flexible structure. the strain sensitivity of nominal length, the rigidity overwith the section of the substrateHere, without the flexible structure sensor is the ratio between the shift in Bragg wavelength of the strain sensor induced by the average is much greater than that over the section with the flexible structure. Here, the strain sensitivity of unit elongation the true length and the nominal determined the applied force, sensor is the ratioover between thegauge shift in Bragg wavelength ofstrain the strain sensor from induced by the average cross-sectional area and Young’s modulus thethe structure to strain be measured, corresponding to the unit elongation over the true gauge length of and nominal determined from the applied average unit elongation over the nominal gauge length. As a result, the strain sensitivity of the metalforce, cross-sectional area and Young’s modulus of the structure to be measured, corresponding to is higher than the baregauge RFBG length. sensor due the flexible structure. thepackaged average sensor unit elongation overthat theof nominal As atoresult, the strain sensitivity of the For sensor applications, the Bragg wavelength of a RFBG-based strain sensor must be stable and metal-packaged sensor is higher than that of the bare RFBG sensor due to the flexible structure. repeatable when subjected to loading at high temperatures. Accordingly, the tensile tests were For sensor applications, the Bragg wavelength of a RFBG-based strain sensor must be stable conducted at the same test temperatures for three times to determine its stability and repeatability. and repeatable when subjected to loading at high temperatures. Accordingly, the tensile tests Figure 10a–g shows the temporal evolution of the shift in Bragg wavelength during the mechanicalwere conducted at the same test temperatures for three times to determine its stability and loading cycles for the sensor prototype at constant temperatures of 26.5 °C, 100 °C, 200 °C, 300 °C, repeatability. Figure 10a–g shows the temporal evolution of the shift in Bragg wavelength during 400 °C, 500 °C and 540 °C, with the same scale in both the wavelength shift and time. It is observed the mechanical-loading cycles for the sensor prototype at constant temperatures of 26.5 ◦ C, 100 ◦ C, that the sensor prototype presents no obvious drift in its Bragg wavelength. Each shift in Bragg
Sensors 2017, 17, 431
13 of 18
200 ◦ C, 300 ◦ C, 400 ◦ C, 500 ◦ C and 540 ◦ C, with the same scale in both the wavelength shift and time. It is observed that the sensor prototype presents no obvious drift in its Bragg wavelength. 13 Each shift Sensors 2017, 17, 431 of 17 in Bragg wavelength of the sensor prototype is calculated by averaging the shifts measured as the wavelength of thewithin sensor prototype is calculatedits bystrain averaging the shiftsThese measured as the load is held load is held constant 2 min to determine sensitivity. values obtained from the constant within 2 min totests determine strain These from mechanical mechanical loading-cycling agree its well withsensitivity. one another, asvalues shownobtained in Figure 11. the Slight fluctuations testsofagree well with one another, as shown in Figure 11. disturbance Slight fluctuations in the in theloading-cycling wavelength shift the sensor prototype are related to temperature at temperatures wavelength shift of the sensor prototype are related to temperature disturbance at temperatures ◦ ◦ of 300 C and 400 C. These results thus highlight the importance to compensate the temperatureofeffect 300 °C and 400 °C. These results thus highlight the importance to compensate the temperature effect to enhance the accuracy. The experimental results demonstrate that the metal-packaged strain sensor to enhance the accuracy. The experimental results demonstrate that the metal-packaged strain sensor based on the RFBG fabricated in H2 -loaded PS1250/1500 fiber has good stability and repeatability. based on the RFBG fabricated in H2-loaded PS1250/1500 fiber has good stability and repeatability. It It hashas alsoalso been proven betweenthe the optical fiber titanium is been proventhat thatthe theinterfacial interfacial bonding bonding between optical fiber andand the the titanium layerlayer is strong, as well as the nickel layer and the strong, as well as the nickel layer and theP91 P91steel steelsubstrate. substrate.
Figure 10. Temporal evolution of the wavelength shift during the mechanical loading cycles for the
Figure 10. Temporal evolution of the wavelength shift during the mechanical loading cycles for metal-packaged strain sensor based on the RFBG in H2-loaded PS1250/1500 fiber at constant the metal-packaged strain sensor based on°C the PS1250/1500 fiber at constant 2 -loaded temperatures of 26.5 °C (a); 100 °C (b); 200 (c);RFBG 300 °Cin (d);H400 °C (e); 500 °C (f) and 540 °C (g). temperatures of 26.5 ◦ C (a); 100 ◦ C (b); 200 ◦ C (c); 300 ◦ C (d); 400 ◦ C (e); 500 ◦ C (f) and 540 ◦ C (g).
Sensors 2017, 17, 431 Sensors 2017, 17, 431
14 of 18 14 of 17
Figure Strain sensitivity of the Figure 11. 11. Strain sensitivity of the metal-packaged metal-packaged strain strain sensor sensor based based on on the the RFBG RFBG in in H H22-loaded -loaded PS1250/1500 fiber obtained obtained from from mechanical mechanical load-cycling load-cycling tests tests as as aa function function of of temperature. temperature. PS1250/1500 fiber
4.4. Numerical Results The numerical numerical results resultsestimated estimatedbyby implementing developed 3-Dmodel FE model the implementing the the developed 3-D FE of the of metalmetal-packaged RFBGsensor strain are sensor are also in Figure 9a–g, for with comparison with the packaged RFBG strain also shown in shown Figure 9a–g, for comparison the experimental experimental results. shifts in the Bragg wavelength calculated by substituting theradial axial and radial results. The shifts in The the Bragg wavelength calculated by substituting the axial and strains of strains optical fiber obtained the FE simulation into Equation and strain determined optical of fiber obtained from thefrom FE simulation into Equation (3), and(3), strain valuesvalues determined from from the forces applied the specimen are linear. The numerical sensitivities of 1.85, 1.84, 1.86, 1.85, 1.85, 1.86, the forces applied to theto specimen are linear. The numerical strainstrain sensitivities of 1.84, −1 are derived from the slope of the solid straight blue lines shown in 1.85, 1.83, 1.83, 1.83, 1.83, and and 1.831.83 pm pm µε−1µεare derived from the slope of the solid straight blue lines shown Figure 9a–g, respectively. respectively. Comparisons of the experimental results and the numerical results show a satisfactory agreement with a relative error less than 15.7%, which may be primarily attributed to the inaccuracy inin 3-D FEFE model and thethe errors in inaccuracy in in the the material materialparameters parameters(particularly (particularlythe theP91 P91steel) steel)use use 3-D model and errors the measurement of structural dimensions. in the measurement of structural dimensions. To deformation of the P91P91 steelsteel specimen, the specimen was only To avoid avoidundesirable undesirableplastic plastic deformation of the specimen, the specimen wastested only up to ~0.08% strain, corresponding to ~0.12% in the RFBG obtained the FEfrom simulation. tested up to ~0.08% strain, corresponding tostrain ~0.12% strain in the RFBGfrom obtained the FE For further For loading, theloading, verifiedthe linear trendlinear may trend be maintained up to the up strain limit of the bare simulation. further verified may be maintained to the strain limit of RFBG (i.e., ~0.4% 4 N as in previous sections, corresponding to ~0.26% strain in the the bare RFBG (i.e.,at~0.4% at 4discussed N as discussed in previous sections, corresponding to ~0.26% strain in specimen) which restricts the strain measurement rangerange of theof sensor. However, this is also the specimen) which restricts the strain measurement the sensor. However, this largely is also dependent on the behavior of the substrate. At a strainAt of a0.26%, Mises occurring in largely dependent on the behavior of the substrate. strainthe of von 0.26%, thestresses von Mises stresses the substrate aresubstrate determined the FE modeling at the room temperature, as shown in Figure 12. occurring in the arefrom determined from the FE modeling at the room temperature, as shown Assuming theAssuming spot weldsthe with sufficient strength to transfer the structural from the specimen to in Figure 12. spot welds with sufficient strength to transferloads the structural loads from the sensor, theto maximum von Mises stress occurring the nickel layerin farthe exceeds yield specimen the sensor, the maximum von Misesinstress occurring nickelthe layer far strength exceeds of MPa [32] inof addition to yielding occurring the region of the spot welds, which that the59.0 yield strength 59.0 MPa [32] in addition to in yielding occurring in the region of theconfirms spot welds, the strain measurement rangemeasurement of the sensor range is limited not only is bylimited the strain rangebyofthe thestrain RFBG, but which confirms that the strain of the sensor not only range also byRFBG, the strength ofby thethe metallic packaging materials and the materials spot welds. of the but also strength of the metallic packaging and the spot welds.
Sensors 2017, 17, 431
Sensors 2017, 17, 431
15 of 18
15 of 17
Figure Von Mises stress distributionininthe themetal-packaged metal-packaged RFBG strain sensor at a strain 0.26%of 0.26% Figure 12. Von12.Mises stress distribution RFBG strain sensor at a of strain applied to the test specimen at room temperature. applied to the test specimen at room temperature.
5. Conclusions 5. Conclusions In this paper, the results of the development and characterization of a laboratorial prototype of a
In metal-packaged this paper, thestrain results of the development characterization a laboratorial prototype of sensor based on the use of and an RFBG fabricated in Hof 2 -loaded B–Ge co-doped a metal-packaged strain based on the useconclusions of an RFBG fabricated infrom H2-loaded B–Ge co-doped photosensitive fiber sensor have been reported. Some have been drawn the investigation as follows: photosensitive fiber have been reported. Some conclusions have been drawn from the investigation as follows: 1.
1.
2.
3.
4.
Regenerated fiber Bragg gratings fabricated in fibers (e.g., B–Ge co-doped photosensitive fiber) which a relatively lower regeneration temperature of fiber) Regenerated require fiber Bragg gratings fabricated in fibers (e.g., (e.g., B–Geregeneration co-doped temperature photosensitive ◦ C) are preferred as the sensing elements to develop high-temperature strain sensors. This 500 which require a relatively lower regeneration temperature (e.g., regeneration temperature of takes into consideration that the regeneration process considerably reduces the mechanical 500 °C) are preferred as the sensing elements to develop high-temperature strain sensors. This strength of the silica optical fibers for which degradation becomes more severe after annealing takes into consideration that the regeneration process considerably reducesduring the mechanical at higher regeneration temperatures even if careful preparations are integrated the strength of the silica optical fibers for which degradation becomes more severe after annealing regeneration process. at 2. higher even based if careful preparations are integrated during the The regeneration metal-packaged temperatures strain sensor prototype on the use of the RFBG fabricated in H2 -loaded PS1250/1500 fiber exhibits good linearity, stability and repeatability when exposed to constant regeneration process. high temperatures up tosensor 540 ◦ C, prototype which is higher thanon thethe upper temperature limit in H2The metal-packaged strain based useoperating of the RFBG fabricated ◦ of 400 C for the strain sensor based on the RFBG in H2 -loaded SMF-28 fiber reported in our loaded PS1250/1500 fiber exhibits good linearity, stability and repeatability when exposed to previous work [17]. Strain sensitivity of the metal-packaged sensor is ~70% higher than that of constant temperatures up todue 540to°C, theof upper operating temperature thehigh corresponding bare RFBG the which supportis ofhigher flexible than structure the metallic substrate. limit 400 °C fordecay the behavior strain sensor basedtwo onregeneration the RFBG regimes in H2-loaded reported in 3. ofAnomalous of exhibiting has beenSMF-28 found forfiber the FBGs our previous [17]. Strain sensitivity thereflectivity metal-packaged is ~70% higher at than that written work in H2 -loaded PS1250/1500 fiber, of with showing asensor small but clear increase ◦ ◦ temperatures between ~300 C andto~430 C, interpreted as a first regeneration regime. Above of the corresponding bare RFBG due the support of flexible structure of the metallic substrate. ◦ C, the gratings started to decay up to 500 ◦ C where the FBGs regenerated, which is ~430 Anomalous decay behavior of exhibiting two regeneration regimes has been found for the FBGs interpreted as a second regeneration regime. This is similar to the behavior of the FBGs in written in H2-loaded PS1250/1500 fiber, with reflectivity showing a small but clear increase at H2 -loaded GF1B photosensitive fiber observed by Polz et al. [35]. In contrast to that, for the FBGs temperatures between ~300 and ~430 °C, interpreted as a900 first ◦ Cregeneration in H2 -loaded SMF28 fiber,°C only one regeneration regime above was observed. regime. The two Above ~430 °C, the gratings started tobedecay to different 500 °C structural where the FBGs regenerated, which is regimes of regeneration may relatedup to the relaxations that occur in the core and the dopantregime. composition interpreted as acladding secondwhere regeneration Thisand is concentrations similar to thediffer. behavior of the FBGs in H24. Comparisons of the experimental results and numerical results of strain sensitivity forthe theFBGs in loaded GF1B photosensitive fiber observed by the Polz et al. [35]. In contrast to that, for metal-packaged strain sensor prototype show a satisfactory agreement with a relative error H2-loaded SMF28 fiber, only one regeneration regime above 900 °C was observed. The two less than 15.7% in the range of the test temperature, which may be primarily attributed to the
regimes of regeneration may be related to the different structural relaxations that occur in the core and cladding where the dopant composition and concentrations differ. Comparisons of the experimental results and the numerical results of strain sensitivity for the metal-packaged strain sensor prototype show a satisfactory agreement with a relative error less than 15.7% in the range of the test temperature, which may be primarily attributed to the
Sensors 2017, 17, 431
16 of 18
inaccuracy in the material parameters, particularly of the P91 steel, used in 3-D FE model and the errors in the measurement of structural dimensions. The FE simulation also shows the operational strain range of the sensor is limited not only by the strain measurement range of the RFBG, but also by the strength of the metallic packaging materials and the spot welds. The metal-packaged strain sensors using silica optical fibers provide great potential for strain measurements in high-temperature environments in ways that were not possible before. However, the strength degradation of silica optical fibers after annealing at high temperatures may also impedes the practical utility of the RFBG-based strain sensors. The mechanical integrity and packaging remain as the key challenges. In future work, it is crucial to improve material systems to meet higher temperature challenges. Acknowledgments: The authors would like to acknowledge the financial supports provided by the National Natural Science Foundation of China (Nos. 51505150 and 11472105), the Natural Science Foundation of Shanghai (No. 15ZR1409100), the China Postdoctoral Science Foundation (No. 2015M580298), the Fundamental Research Funds for the Central Universities (No. WG1514032), and the 111 Project (No. B13020). Author Contributions: S.-T. Tu and Y. Tu conceived and designed the experiments; Y. Tu performed the experiments and analyzed the data; Y. Tu, L. Ye, S.-P. Zhou and S.-T. Tu interpreted the experimental results; Y. Tu, L. Ye, S.-P. Zhou and S.-T. Tu wrote and revised the paper. Conflicts of Interest: The authors declare no conflict of interest.
References 1.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14.
Tu, S.-T. Life prediction and monitoring of critical industrial equipment. In Proceedings of the FM 2009—Transferability and Applicability of Current Mechanics Approaches, Chengdu, China, 16–20 October 2009; Sih, G.C., Tu, S.-T., Wang, Z.-D., Eds.; East China University of Science and Technology Press: Shanghai, China, 2009; pp. 13–22. Zhang, H.; Jia, J.; Wang, N.; Hu, X.; Tu, S.-T.; Zhou, S.; Wang, Z. Development of on-line monitoring systems for high temperature components in power plants. Sensors 2013, 13, 15504–15512. [CrossRef] [PubMed] Sposito, G.; Ward, C.; Cawley, P.; Nagy, P.B.; Scruby, C. A review of non-destructive techniques for the detection of creep damage in power plant steels. NDT E Int. 2010, 43, 555–567. [CrossRef] Mihailov, S.J. Fiber Bragg grating sensors for harsh environments. Sensors 2012, 12, 1898–1918. [CrossRef] [PubMed] Baker, S.R.; Rourke, H.N.; Baker, V.; Goodchild, D. Thermal decay of fiber Bragg gratings written in boron and germanium codoped silica fiber. J. Lightwave Technol. 1997, 15, 1470–1477. [CrossRef] Groothoff, N.; Canning, J. Enhanced type IIA gratings for high-temperature operation. Opt. Lett. 2004, 29, 2360–2362. [CrossRef] [PubMed] Prakash, O.; Kumar, J.; Mahakud, R.; Agrawal, S.K.; Dixit, S.K.; Nakhe, S.V. Enhanced temperature (~800 ◦ C) stability of type-IIa FBG written by 255 nm beam. IEEE Photonics Technol. Lett. 2014, 26, 93–95. [CrossRef] Bartelt, H.; Schuster, K.; Unger, S.; Chojetzki, C.; Rothhardt, M.; Latka, I. Single-pulse fiber Bragg gratings and specific coatings for use at elevated temperatures. Appl. Opt. 2007, 46, 3417–3424. [CrossRef] [PubMed] Smelser, C.W.; Grobnic, D.; Mihailov, S.J. High-reflectivity thermally stable ultrafast induced fiber Bragg gratings in H2 -Loaded SMF-28 fiber. IEEE Photonics Technol. Lett. 2009, 21, 682–684. [CrossRef] Lowder, T.L.; Smith, K.H.; Ipson, B.L.; Hawkins, A.R.; Selfridge, R.H.; Schultz, S.M. High-temperature sensing using surface relief fiber Bragg gratings. IEEE Photonics Technol. Lett. 2005, 17, 1926–1928. [CrossRef] Brambilla, G.; Rutt, H. Fiber Bragg gratings with enhanced thermal stability. Appl. Phys. Lett. 2002, 80, 3259–3261. [CrossRef] Pal, S.; Shen, Y.H.; Mandal, J.; Sun, T.; Grattan, K.T.V. Simultaneous measurement of strain (to 2000 µε) and temperature (to 600 ◦ C) using a combined Sb-Er-Ge-codoped fiber-fluorescence and grating-based technique. IEEE Sens. J. 2005, 5, 1462–1468. [CrossRef] Butov, O.V.; Dianov, E.M.; Golant, K.M. Nitrogen-doped silica-core fibres for Bragg grating sensors operating at elevated temperatures. Meas. Sci. Technol. 2006, 17, 975–979. [CrossRef] Fokine, M. Thermal stability of chemical composition gratings in fluorine-germanium-doped silica fibers. Opt. Lett. 2002, 27, 1016–1018. [CrossRef] [PubMed]
Sensors 2017, 17, 431
15. 16. 17.
18. 19. 20.
21. 22. 23.
24. 25. 26.
27. 28.
29. 30. 31. 32. 33. 34. 35. 36.
37.
17 of 18
Canning, J.; Stevenson, M.; Bandyopadhyay, S.; Cook, K. Extreme silica optical fibre gratings. Sensors 2008, 8, 6448–6452. [CrossRef] [PubMed] Sandlin, S.; Kinnunen, T.; Ramo, J.; Sillanpaa, M. A simple method for metal re-coating of optical fibre Bragg gratings. Surf. Coat. Technol. 2006, 201, 3061–3065. [CrossRef] Tu, Y.; Tu, S.-T. Fabrication and characterization of a metal-packaged regenerated fiber Bragg grating strain sensor for structural integrity monitoring of high-temperature components. Smart Mater. Struct. 2014, 23, 035001. [CrossRef] Xie, J.-F.; Zhang, H.; Zhu, Z.; Xu, J.-N.; Hu, R.-H.; Song, L.-F. A study of the temperature sensitivity of fiber Bragg gratings after metallization. Smart Mater. Struct. 2007, 16, 1837–1842. [CrossRef] Li, X.C.; Prinz, F.; Seim, J. Thermal behavior of a metal embedded fiber Bragg grating sensor. Smart Mater. Struct. 2001, 10, 575–579. [CrossRef] Lupi, C.; Felli, F.; Ippoliti, L.; Caponero, M.A.; Ciotti, M.; Nardelli, V.; Paolozzi, A. Metal coating for enhancing the sensitivity of fibre Bragg grating sensors at cryogenic temperature. Smart Mater. Struct. 2005, 14, N71–N76. [CrossRef] Alemohammad, H.; Toyserkani, E. Metal embedded optical fiber sensors: Laser-based layered manufacturing procedures. J. Manuf. Sci. Eng.-Trans. ASME 2011, 133, 031015. [CrossRef] Li, Y.; Liu, W.; Feng, Y.; Zhang, H. Ultrasonic embedding of nickel-coated fiber Bragg grating in aluminum and associated sensing characteristics. Opt. Fiber. Technol. 2012, 18, 7–13. [CrossRef] Havermann, D.; Mathew, J.; MacPherson, W.N.; Maier, R.R.J.; Hand, D.P. Temperature and strain measurements with fiber Bragg gratings embedded in stainless steel 316. J. Lightwave Technol. 2015, 33, 2474–2479. [CrossRef] Proctor, B.A.; Whitney, I.; Johnson, J.W. The strength of fused silica. Proc. R. Soc. Lond. A 1967, 297, 534–557. [CrossRef] Wang, T.; Shao, L.Y.; Canning, J.; Cook, K. Temperature and strain characterization of regenerated gratings. Opt. Lett. 2013, 38, 247–249. [CrossRef] [PubMed] Tu, Y.; Tu, S.-T. Tensile strength of silica optical fibers for high-temperature sensing applications. In Recent Advances in Structural Integrity Analysis, Proceedings of the International Congress (APCF/SIF-2014), Sydney, Australia, 9–12 December 2014; Ye, L., Ed.; Woodhead Publishing: Oxford, UK, 2014; pp. 158–162. Lezzi, P.J.; Evke, E.E.; Aaldenberg, E.M.; Tomozawa, M. Surface crystallization and water diffusion of silica glass fibers: Causes of mechanical strength degradation. J. Am. Ceram. Soc. 2015, 98, 2411–2421. [CrossRef] Morey, W.W.; Meltz, G.; Glenn, W.H. Fiber optic Bragg grating sensors. In Proceedings of the Fiber Optic and Laser Sensors VII, Boston, MA, USA, 5–7 September 1989; DePaula, R.P., Udd, E., Eds.; SPIE: Bellingham, DC, USA, 1989; pp. 98–107. Othonos, A. Fiber Bragg gratings. Rev. Sci. Instrum. 1997, 68, 4309–4341. [CrossRef] Cheong, Y.K.; Chong, W.Y.; Chong, S.S.; Lim, K.S.; Ahmad, H. Regenerated type-IIa fibre Bragg grating from a Ge–B codoped fibre via thermal activation. Opt. Laser Technol. 2014, 62, 69–72. [CrossRef] Tu, Y.; Qi, Y.-H.; Tu, S.-T. Fabrication and thermal characteristics of multilayer metal-coated regenerated grating sensors for high-temperature sensing. Smart Mater. Struct. 2013, 22, 075026. [CrossRef] Matweb Material Property Data. Available online: http://www.matweb.com (accessed on 10 April 2016). Farraro, R.; McLellan, R. Temperature dependence of the Young’s modulus and shear modulus of pure nickel, platinum, and molybdenum. Metall. Trans. A 1977, 8, 1563–1565. [CrossRef] Pan, J.-Z. Pressure Vessel Material Practical Handbook—Carbon Steel and Alloy. Steel (Chinese Edition); Chemical Industry Press: Beijing, China, 2000. Polz, L.; Nguyen, Q.; Bartelt, H.; Roths, J. Fiber Bragg gratings in hydrogen-loaded photosensitive fiber with two regeneration regimes. Opt. Commun. 2014, 313, 128–133. [CrossRef] Pal, S.; Mandal, J.; Sun, T.; Grattan, K.T.V. Analysis of thermal decay and prediction of operational lifetime for a type I boron-germanium codoped fiber Bragg grating. Appl. Opt. 2003, 42, 2188–2197. [CrossRef] [PubMed] Bandyopadhyay, S.; Canning, J.; Stevenson, M.; Cook, K. Ultrahigh-temperature regenerated gratings in boron-codoped germanosilicate optical fiber using 193 nm. Opt. Lett. 2008, 33, 1917–1919. [CrossRef] [PubMed]
Sensors 2017, 17, 431
38.
39.
18 of 18
Lindner, E.; Canning, J.; Chojetzki, C.; Bruckner, S.; Becker, M.; Rothhardt, M.; Bartelt, H. Thermal regenerated type IIa fiber Bragg gratings for ultra-high temperature operation. Opt. Commun. 2011, 284, 183–185. [CrossRef] Lindner, E.; Chojetzki, C.; Bruckner, S.; Becker, M.; Rothhardt, M.; Bartelt, H. Thermal regeneration of fiber Bragg gratings in photosensitive fibers. Opt. Express 2009, 17, 12523–12531. [CrossRef] [PubMed] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
