Open Access
Add to BibSonomy
Abstract
In this study, a sapphire-derived fiber (SDF)-based Fabry-Pérot interferometer (FPI) is proposed and experimentally demonstrated as a high-temperature sensor using the arc discharge crystallization process, forming a region with disordered mullite grains. This shows that the disordered mullite grains are related to the gradual temperature distribution during the arc discharge process, which results in a larger refractive index (RI) modulation of the SDF near the fusing area, forming a reflection mirror. An FPI was obtained by combining the optical fiber end facet. Considering the high-temperature resistance of the fiber, the fabricated FPI was used for high-temperature sensing. This shows that the device can operate at temperatures of up to 1200 °C with a sensitivity of 15.47 pm/°C, demonstrating that the proposed devices have potential applications in high-temperature environments.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
It is well known that high-temperature monitoring is vital in many applications, such as internal combustion engines, gas turbines, deep oil, and geothermal drilling [1]. Owing to the harsh environments and limited space, there are few high-temperature sensors that can effectively work for a long time. Fortunately, optical fiber sensors are becoming a favorable choice for temperature sensing owing to their merits of electromagnetic interference, corrosion resistance, and small size [2]. In general, the telecom standard Ge-doped silica fibers (SMFs) based sensors are not suitable for high temperatures, particularly above 800 °C, because the creep of the silica glass material and the diffusion of the dopants (e.g., Ge) inside the SMF core could limit the practicality of the sensor [3]. Hence, it is necessary to develop a special optical fiber sensor for high-temperature sensing.
Various methods have been proposed for fabricating such sensors, which are mostly divided by thermal processing, for example, thermal regenerated gratings (RGs) [4–7] and materials with higher melting points [8]. Thermal RGs have attracted considerable interests from researchers owing to their excellent optical properties at ultra-high temperatures [4]. Canning et al. observed that the internal stress of optical fibers was high at high temperatures, and this could induce glass crystallization, which is the main reason for realizing the ultra-high-temperature sustainability of fiber-optic sensors [7]. Optical fiber sensors with thermal processing can typically withstand temperatures of up to 1000 °C. Yang et al. applied this method to a high-temperature sensing fiber Bragg grating (FBG) design, and the sensing range was increased to a temperature of 1400 °C [5]. However, the performance of such optical fiber temperature sensors decreased as the temperature further increases. Compared with other processing methods, materials with higher melting point could be a good choice. Much attention has been paid to single-crystal sapphire fibers that are grown using the laser-heated pedestal-growth method with a high melting point (2040 °C) and low losses [8]. Generally, the sapphire fibers have large cores and do not contain cladding layers, which can lead to multiple propagation modes and are particularly vulnerable to interference. Nevertheless, a relatively clear reflection spectrum can be still obtained if a micro FPI sensor is fabricated on the sapphire fiber tip. On the other hand, such a microcavity usually needs to be fabricated by a femtosecond laser micromachining technology or introduce a single crystal sapphire diaphragm as an interference cavity, which undoubtedly increases the complexity of sensor preparation [9,10]. In addition, if a relative longer access sapphire fiber is used, e.g., tens of centimeter, it will introduce multimode problem which impacts the FPI interference spectrum.
Based on pure alumina (Al2O3) composition of sapphire fiber, it can be directly fabricated as an optical fiber preform rod and drawn as sapphire-derived fibers (SDFs), which have drawn much attention because of their highly alumina-doped core, outstanding advantages of controllable core diameter, confined modes by cladding protection, and high-temperature resistance [11]. It was observed the composition of the high alumina dopant concentration core changed and the mullite crystallization region occurred when the temperature reached the crystallization temperature [12–14]. The size distribution of the mullite grains is related to the gradual temperature distribution. Owing to the high-temperature resistance property, the region with mullite grains can be used for device design and high-temperature sensing [14–20]. Recently, Liu et al. demonstrated that the interface between the mullite crystallization region and non-crystallization region in SDF could be used as reflection mirrors for designing Fabry-Pérot interferometers (FPIs), which are based on their high refractive index (RI) modulation characteristics [17]. However, the material and scattering properties are still not well known. It is observed that the RI distribution in a disordered scattering medium is inhomogeneous, and the sizes of the nanoparticles are related to the temperature, which can induce strong scattering and influence the transmission and reflection.
In this study, an FPI based on mullite crystallization-induced disordered mullite grains is proposed and experimentally characterized in terms of its high-temperature sensitivity and long-term stability. The disordered mullite grains were generated during the arc discharge crystallization process of the SDF, and its major composition was approximately 3:2 mullite. The mechanism of the effect of the crystallization region on the FPI spectrum is comprehensively explained. The gradual distribution of the disordered mullite grain sizes formed an equivalent reflection mirror owing to the scattering of light. The results showed that the FPI could operate at high temperatures of up to 1200 °C with good linearity, stability, and repeatability after an annealing process. The SDF FPI can be conveniently fabricated by arc discharge relative to the complex fabrication process of sapphire FPI [9,10]. Comparing with the FPI sensor with double crystallization regions [17], only one crystallization reflection mirror is introduced in this work, which makes it easier to control the FPI interference spectrum pattern. An intrinsic FP cavity can be fabricated based on SDF by only one arc discharge. Therefore, the proposed device has great potential for applications in various high-temperature conditions.
2. FPI device
2.1 Fabrication
Figure 1(a) shows the fabrication process of the FPI with a section of SDF with a mullite crystallization region and an SMF. The core and cladding diameters of the SDF were 16 and 125 μm, respectively. To avoid bubble generation, the intensity and duration of the arc discharge were set as standard mode and 3000 ms, respectively. A conical and gradually indistinct mullite crystallization region was formed along the axial direction of the SDF owing to the gradient temperature distribution during the arc discharge process [14]. The experimental results showed that the RI of the mullite crystallization gradually changed, and a strong reflection interface was equivalently formed between the tip and the uniform area. Hence, an FPI was formed, which contained three reflection interfaces labeled M1, M2, and M3, as shown in Fig. 1(b). M1 and M3 form FPI1 with a cavity length of L1. M2 and M3 form FPI2 with a cavity length of L2. The multiple reflections in the FPI can be neglected because of the low reflection of each reflection mirror [1,21]. The normalized reflection spectrum IR can be expressed as follows:
To comprehensively study the impact of the mullite crystallization region, we fabricated several FPIs with different SDF lengths from 20 to hundreds of microns. Figures 2(a)–(c) show microscopic images of the fabricated devices with SDF lengths of 47.5, 96.0, and 409.7 µm, respectively. Figures 2(d)–(f) show the corresponding reflection spectra of the fabricated FPIs. The free spectral ranges (FSRs) are 22.52, 11.70, and 1.96 nm, respectively. The corresponding spatial frequency spectra were obtained, as shown in Figs. 2(g)–(i). Two peaks were observed in each spatial frequency spectrum, and then the frequency was calculated based on the equation $FSR = {1 / \xi } = {{{\lambda ^2}} / {2nL}}$ (where $\xi$, $\lambda$, n, and $L$ represent the spatial frequency, operating wavelength, RI, and cavity length of the corresponding FPI [23], respectively). It can be calculated that the dominant influence factor on the reflection spectrum is at the frequency point of 0.050, 0.100, and 0.513, respectively. To explore the relationship between the two main frequency peaks, the cavity length can be calculated according to the FSR equation. Comparing the results from the experimentally obtained values, the differences between the calculated cavity length in each FPI were all approximately 25.0 µm. In other words, one dominant cavity determines the FPI reflection spectrum, while the other cavity with a 25.0 µm difference forms a Vernier effect on the dominant cavity, which leads to the inhomogeneity of the reflection spectrum.
Fig. 2. (a)–(c) Microscopic images; (d)–(f) reflection spectra of the FPI with the SDF length of 47.5, 96.0, and 409.7 µm, respectively; and (g)–(i) spatial frequency spectra of the FPI in (a)–(c).
2.2 Experimental analysis of the FPI device
To analyze the fabricated device, the RI of the crystallization region was studied. Here, an FPI with an SDF length of 96.0 µm was used for mapping the RI distribution within the crystallization region. Figure 3 shows the microscopic images and RI profiles at different positions, which were measured by an RI profiler (SHR-1602). As shown in Fig. 3(b), the RI difference between the core and cladding of the SDF near the splicing point was 0.016 (1, red curve). The reflection at the interface between the SMF and SDF was calculated to be only 0.0014%, which can be neglected. However, the RI difference increased sharply from the SDF core after the heating treatment by the arc discharge (2, green curve, approximately 0.028) to the mullite crystallization region (3, pink curve, 0.067, 4, cyan curve, 0.070, 5, blue curve, 0.073 and 6, violet curve, 0.076). Owing to the difference between the areas of 2 and 3, it can cause relatively large reflection, which can be equivalent to a reflection mirror M1 for the FP cavity.
Fig. 3. (a) Microscopic image of the FPI with SDF length of 96.0 µm (cross section of an SDF); (b) RI profiles of the SDF (black) and the FPI at typical locations with 1-z:-304.0 µm (red), 2-z:0 µm (green), 3-z:3.0 µm (pink), 4-z:27.0 µm (cyan), 5-z:59.0 µm (blue), and 6-z:96.0 µm (violet), respectively.
To further study the fabrication, the material characteristics of the mullite crystallization region were analyzed using a scanning electron microscope (SEM, JSM-7500F), which was pretreated by hydrofluoric acid (HF) etching for 30 s, and then the small grains of the mullite crystallization region could be clearly obtained. Figures 4(a), (c), and (e) show the SEM images; Figs. 4(b), (d), and (f) show the grain size distributions of the mullite crystallization region at 27.0, 59.0, and 96.0 µm, respectively. It can be clearly observed that the mullite grain size decreased dramatically with an increase in the mullite crystallization region length. The average diameters are approximately 0.28, 0.18, and 0.13 µm, respectively. The difference in the mullite grain size distribution can be explained by the principle of spinodal nucleation [12,14]. Owing to the high concentration alumina dopant, the core material is modified depending on the Gaussian-like temperature distribution along the axial direction of the SDF by arc discharge. When the core temperature is close to the crystallization temperature, a metastable immiscibility region is formed, and the phase separation processes are accompanied by the nucleation and growth of the mullite [24,25]. The nucleation and growth rates are diverse at different temperatures and heating and cooling rates [25]. When the temperature is higher than 1650 °C and exceeds the crystallization temperature, the crystallization hardly happened because the nucleation and growth rates significantly decrease. As the temperature decreases and approaches the crystallization temperature, the nucleation and growth rates increase. However, with a continuous decrease in temperature, the viscosity of the glass increases gradually, the ions in the glass gradually lose the ability of mutual movement, and finally, mullite stops nucleating and growing [26]. The mullite crystallization region at 27.0 µm has higher temperature and faster heating and cooling rates, which results in shorter nucleation times and fewer nucleation grain numbers. Simultaneously, a longer time in the range of the crystallization temperature results in a faster growth rate, and as a result, the mullite grain size is larger. On the contrary, the mullite crystallization region at 96.0 µm has a lower temperature and slower heating and cooling rates, which leads to longer nucleation times and more nucleation numbers. However, the temperature soon decreases below the crystallization temperature, and mullite stops growing. Therefore, a smaller mullite grain size was observed here.
Fig. 4. (a), (c), and (e) SEM images; (b), (d), and (f) grain size distributions of the mullite crystallization region at 27.0, 59.0, and 96.0 µm, respectively.
Concerning material composition, the crystallization region was analyzed by Raman scattering. Corresponding Raman spectra were obtained and recorded using a Raman microscope (HR800, Horiba Jobin Yvon) with a 50× objective and a confocal Raman spectrometer (LabRAM HR Evolution, Horiba), respectively [27]. A 532 nm laser with a power of 50 mW was used as the excitation light with an acquisition time of 1 s. Figure 5 shows the normalized Raman spectra of the mullite crystallization region at 59.0 and 96.0 µm. Six characteristic vibration modes were detected, corresponding to the mullite signature: 305, 415, 600, 800, 950, and 1130 cm-1 [28], which were in accordance with the mullite data obtained in the literatures [29–33]. These typical Raman peaks are similar to those of 3:2 mullite, which indicates that the major composition is approximately 3:2 mullite.
In addition, the length of crystallization region can be precisely controlled through the heating process. The crystallization process of the SDF core is explained through a spinodal nucleation effect, which is related to temperature and the rates of heating and cooling process [24,25]. If we can provide heating process zone which satisfies the spinodal nucleation condition, the designed length of the crystallization region can be obtained within the zone accordingly. In order to demonstrate the crystallization evaluation process, a CO2 laser as heating source was utilized to heat the SDF. The microscope photography under different laser powers of 0.65 W, 0.75 W and 0.86 W are shown in Fig. 6. The laser spot presents a Gaussian profile. With increasing the laser power, different length of the heating process zone can be controlled. As a result, the length of the crystallization region got longer with increasing laser power, as shown in the dash line box of Fig. 6.
When light is coupled into the FPI cavity with the above crystallization grains, it scatters owing to the disordered mullite grains. Some scattering light could return to a location it has previously visited after successive scatterings, and there always exists a time-reversed path with an identical phase delay, which enhances the energy density at the location and eventually leads to localization [34]. This effect can be explained by the self-consistent (SC) theory by introducing a position-dependent renormalized coefficient D [35]. The amount of D is determined by the return probability, and the maximum renormalization occurs at a location where the return probability is the highest, which depends on the grain size, grain boundary, density, and RI of the grains [34,36]. It is worth mentioning that the scattering light decreases as the mean grain size is small, and the renormalization of the scattering light is proportional to the density and RI of the grain [37–39]. As can be observed from Figs. 3 and 4, with an increase in the mullite crystallization region length, the average mullite grain size decreased, but the density and RI of the grain increased. Therefore, as the RI of the grain and grain boundaries increases and the average mullite grain size decreases, we can reasonably infer that the enhanced reflection is concentrated in one specific area of the mullite crystallization region, thus forming another reflection mirror M2 .
In addition, the scattering light is strongly affected by the overall effective RI of the mullite crystallization region as follows:
where $f(z )$ and z represent the effective RI and the length from the origin position of the mullite crystallization region, respectively. The cavity acts as an FPI, rendering the presence of a certain effective RI, which is detectable by a phase change of the interferometric response caused by the change in the optical path length of the cavity [40]. Hence, the optical path of the cavity obtained as a function of z with an effective RI can be defined as follows:Consequently, owing to the inhomogeneous and gradually changed effective RI of the mullite crystallization region in each FPI, a fixed optical path of approximately 25.0 µm between the two reflection mirrors was formed.
3. High-temperature sensing results
Mullite has been widely used in harsh high-temperature environments because of its high melting point (1850 °C) and excellent high-temperature creep resistance [41,42]. Therefore, we focus on the potential applications of the proposed FPI for high-temperature sensors. Figure 7 shows the schematic diagram of the high-temperature experimental system which consists of a furnace (SXL-1700C, SIOMM) for temperature control with a constant temperature space of approximate 40 cm×30 cm×30 cm, and an optical sensing interrogator. The wavelength range and the resolution of the interrogator are 1510∼1590 nm and 5 pm, respectively [20]. In order to avoid the damage of the SMF within the high temperature zone, the SDF FPI sensor was placed vertically in the furnace chamber, which ensured the sensing fiber free from bending stress.
When the SDF FPI is subjected to a temperature variation (ΔT), the effective RI (neff) and L change, and the wavelength shift (Δλm) of interference spectrum dips (λm) can be described as follows:
First, an annealing process was conducted by repeating the heating and cooling cycle once to relax the internal residual stress (RS) of the SDF (Round 1). Next, the heating and cooling cycles were repeated three times to evaluate the repeatability of the FPI (Round 2-Round 4). Here, the FPI with an SDF length of 409.7 µm was used, and the spectra in the fourth heating and cooling cycle are displayed in Fig. 8(a). Clear monotonic red and blue shifts with increasing and decreasing temperatures, respectively, were observed. According to Eq. (4), the temperature sensitivity is related to the TOC and CTE of FPI cavity in the SDF core. Generally, the TOC and CTE of the glass material are almost unchanged under different temperature [43]. Therefore, the interference wavelength shift is linearly proportional to the temperature change. The sensitivity of 15.47 pm/°C is obtained from the linear fitting results of the measured data in Fig. 8(b) with a high linearity of R2 > 0.99, showing well-matched temperature sensitivity in the three heating and cooling cycles. It is worth emphasizing that the extinction ratio (ER) of the FPI does not change even after four heating and cooling cycles, as shown in Fig. 8(c). The slight difference in the reflection spectrum before and after four heating and cooling cycles is mainly attributed to the relaxation of the internal RS of SDF [16,44,45]. If the SDF FPI is heated to higher temperature (>1200 °C), mullite grains would begin to dissolve in the crystallization region, grain boundary would deform and alumina began to precipitate [17]. Consequently, a decrease of intensities from the reflection mirrors would lead to the change of the reflection spectrum.
Fig. 8. (a) Reflection spectra of the FPI with SDF length of 409.7 µm at different temperatures in the fourth temperature cycle (Round 4); (b) dip wavelengths of the FPI at different temperatures in four temperature cycles (Round 1-Round 4); and (c) reflection spectra of the FPI before and after four temperature cycles (Round 1 and Round 4).
Here, the spatial frequency of the interference spectrum was also used to analyze the FSR of the cavity [46]. In our case, the sensor consisted of two main FPI cavities due to the crystallization region. The temperature-induced variations in the spatial frequency spectrum of the FPI at different temperatures in the fourth temperature cycle are shown in Fig. 9(a). The two peaks at frequencies of 0.487 and 0.513 were not shifted, but the intensities were changed by the temperature, as shown in Fig. 9(b). The dominant frequency of 0.513 corresponded to the cavity of FPI1, whereas the frequency point 0.487 corresponded to the cavity of FPI2. Figures 9(c) and (d) show the spatial frequency spectra of the FPI at 52 °C and 1200 °C in the experiments of the four heating and cooling cycles, respectively, demonstrating the excellent repeatability of the high-temperature measurements.
Fig. 9. (a) Spatial frequency spectra; (b) intensities of the frequency points 0.487 and 0.513 of the FPI at different temperatures in the fourth temperature cycle (Round 4); spatial frequency spectra of the FPI at (c) 52 °C and (d) 1200 °C in four heating and cooling cycles (Round 1-Round 4), respectively.
In addition, as shown in Figs. 9(a), (c) and (d), the peak intensities at 0.513 and 0.478 corresponding to FPI1 and FPI2 cavities increase and decrease with the increase in temperature, respectively, indicating that the contribution of TOC δ is more predominant than that of CTE α in the proposed FPI cavities. Therefore, it is necessary to consider δ in disordered mullite grains to analyze the FPI sensing performance at high temperatures.
Early studies by Schneider et al. showed that δ was mainly depended on the temperature coefficient of the electronic polarizability ϕ and the volumetric CTE β, which stems from the differences in the associated vacancies and specific linkage of octahedral and tetrahedral chains in the mullite structure [47–51]. Therefore, the disordered mullite grains with different sizes, grain boundaries, densities and RIs exhibit diverse δ, which leads to the different scattering and enhanced reflection in FPI, and finally change the FPI1 and FPI2 cavities intensities.
Additionally, the long-term stability of the proposed SDF FPI was evaluated. The FPI with SDF length of 96.0 µm was monitored continuously for a long time at a high temperature of 1200 °C. The interference spectra in 1 hour, 5 hours and 9 hours are illustrated in Fig. 10, indicating that the mullite crystallization region could be used for ultrahigh-temperature sensors.
Fig. 10. Interference spectra of the FPI with SDF length of 96.0 µm as maintaining at the temperature of 1200 °C.
4. Discussion
In the FPI scheme, interference only occurs between two optical waves when their optical path difference is within the coherence length of the optical light source [52]. Therefore, we compared the reflection parameters of the FPI with an SDF length of 302.1 µm using different optical light sources. Here, an optical sensing interrogator (SM125) with a better coherence swept laser source and an amplified spontaneous emission (ASE) optical light source (SL3200-C42, Jiahui, 1530 nm to 1610 nm) were employed as the optical light source. Additionally, the experimental system included a circulator and an optical spectrometer analyzer (OSA, 70, Yokogawa) with a minimum spectral resolution of 0.02 nm. Figures 11(a) and (c) show the reflection spectra without the influence of the background light, and Figs. 11(b) and (d) show the corresponding maximum ERs of the FPI. It can be clearly observed that the interference fringe of the FPI was heavily dependent on the coherence of the optical light source. When a swept laser source with better coherence was used, clear localization was formed in the mullite crystallization region; thus, the reflection spectra had a high maximum ER of 13.64 dB. In contrast, the reflection spectrum of the FPI using the ASE optical light source owned a maximum ER of 0.51 dB, indicating that the coherence of the optical light source had a great influence on the ER of the FPI. The better coherence of the optical light source, the stronger scattering light in the mullite crystallization region, thereby forming a reflection mirror of the FPI.
Fig. 11. (a), (c) Reflection spectra; (b), (d) the corresponding FSRs and maximum ERs of the FPI using the SM125 swept laser source and ASE optical light source, respectively.
5. Conclusion
An FPI based on mullite crystallization region disordered mullite grains was constructed and characterized in terms of its high temperature sensitivity and long-term stability. The disordered mullite grains were generated during the arc discharge crystallization process of the SDF, and its major composition was approximately 3:2 mullite. The gradual distribution in the disordered mullite grain sizes formed an equivalent reflecting mirror owing to the scattering of light. The proposed FPI sensor is capable of withstanding high temperatures of up to 1200 °C and shows good linearity, stability, and repeatability after an annealing process, which makes it promising for high-temperature sensing applications.
Funding
National Natural Science Foundation of China (61635006, 61735009, 61975108).
Disclosures
The authors declare that there are no conflicts of interest related to this study.
Data availability
The data underlying the results presented in this paper are not publicly available at this time, but may be obtained from the authors upon reasonable request.
References
1. X. Q. Lei and X. P. Dong, “High-sensitivity Fabry-Perot interferometer high-temperature fiber sensor based on Vernier effect,” IEEE Sens. J. 20(10), 5292–5297 (2020). [CrossRef]
2. Y. Han, B. Liu, Y. F. Wu, Y. Y. Mao, J. Wu, L. L. Zhao, T. Nan, J. Wang, R. Tang, Y. L. Zhang, and Y. Liu, “Ultra-compact silicon-microcap based improved Michelson interferometer high-temperature sensor,” Opt. Express 29(5), 6703–6713 (2021). [CrossRef]
3. J. Huang, X. W. Lan, Y. Song, Y. J. Li, L. W. Hua, and H. Xiao, “Microwave interrogated sapphire fiber Michelson interferometer for high temperature sensing,” IEEE Photonics Technol. Lett. 27(13), 1398–1401 (2015). [CrossRef]
4. J. Canning, M. Stevenson, S. Bandyopadhyay, and K. Cook, “Extreme silica optical fibre gratings,” Sensors 8(10), 6448–6452 (2008). [CrossRef]
5. H. Z. Yang, X. G. Qiao, S. Das, and M. C. Paul, “Thermal regenerated grating operation at temperatures up to 1400 °C using new class of multimaterial glass-based photosensitive fiber,” Opt. Lett. 39(22), 6438–6441 (2014). [CrossRef]
6. S. Bandyopadhyay, J. Canning, M. Stevenson, and K. Cook, “Ultra-high temperature regenerated gratings in boron codoped germanosilicate optical fibre using 193 nm,” Opt. Lett. 33(16), 1917–1919 (2008). [CrossRef]
7. J. Canning, S. Bandyopadhyay, M. Stevenson, and K. Cook, “Fibre Bragg grating sensor for high temperature application,” the Australian Conference on Optical Fibre Technology (ACOFT) & Opto-Electronics Communications Conference (OECCC) (2008).
8. R. K. Nubling and J. A. Harrington, “Optical properties of single-crystal sapphire fibers,” Appl. Opt. 36(24), 5934–5940 (1997). [CrossRef]
9. S. Yang, Z. Feng, X. T. Jia, G. Pickrell, W. Ng, A. B. Wang, and Y. Z. Zhu, “All-sapphire miniature optical fiber tip sensor for high temperature measurement,” J. Lightwave Technol. 38(7), 1988–1997 (2020). [CrossRef]
10. B. T. Wang, Y. X. Niu, S. W. Zheng, Y. H. Yin, and M. Ding, “A high temperature sensor based on sapphire fiber Fabry-Perot interferometer,” IEEE Photonics Technol. Lett. 32(2), 89–92 (2020). [CrossRef]
11. P. Dragic, T. Hawkins, P. Foy, S. Morris, and J. Ballato, “Sapphire-derived all-glass optical fibers,” Nat. Photonics 6(9), 627–633 (2012). [CrossRef]
12. J. W. Cahn, “Phase separation by spinodal decomposition in isotropic systems,” J. Chem. Phys. 42(1), 93–99 (1965). [CrossRef]
13. S. Aramaki and R. Roy, “Revised phase diagram for the system Al2O3-SiO2,” J American Ceramic Society 45(5), 229–242 (1962). [CrossRef]
14. L. Hong, F. F. Pang, H. H. Liu, J. Xu, Z. Y. Chen, Z. W. Zhao, and T. Y. Wang, “Refractive index modulation by crystallization in sapphire-derived fiber,” IEEE Photonics Technol. Lett. 29(9), 723–726 (2017). [CrossRef]
15. T. Elsmann, A. Lorenz, N. S. Yazd, T. Habisreuther, J. Dellith, A. Schwuchow, J. Bierlich, K. Schuster, M. Rothhardt, L. Kido, and H. Bartelt, “High temperature sensing with fiber Bragg gratings in sapphire-derived all glass optical fiber,” Opt. Express 22(22), 26825–26833 (2014). [CrossRef]
16. J. Xu, H. H. Liu, F. F. Pang, L. Hong, Z. W. Ma, Z. W. Zhao, N. Chen, Z. Y. Chen, and T. Y. Wang, “Cascaded Mach-Zehnder interferometers in crystallized sapphire-derived fiber for temperature-insensitive filters,” Opt. Mater. Express 7(4), 1406–1413 (2017). [CrossRef]
17. H. Liu, F. Pang, L. Hong, Z. Ma, J. Wen, Z. Chen, and T. Wang, “Crystallization-induced refractive index modulation on sapphire-derived fiber for ultrahigh temperature sensing,” Opt. Express 27(5), 6201–6209 (2019). [CrossRef]
18. D. Grobnic, S. J. Mihailov, J. Ballato, and P. D. Dragic, “Type I and II Bragg gratings made with infrared femtosecond radiation in high and low alumina content aluminosilicate optical fibers,” Optica 2(4), 313–322 (2015). [CrossRef]
19. P. F. Chen, F. F. Pang, Z. W. Zhao, L. Hong, N. Chen, Z. Y. Chen, and T. Y. Wang, “Fabry-Perot cavity based on sapphire-derived fiber for high temperature sensor,” the 24th International Conference on Optical Fibre Sensors (OFS), 96347T–1(2015).
20. Z. Wang, H. Liu, Z. Ma, Z. Chen, T. Wang, and F. Pang, “High temperature strain sensing with alumina ceramic derived fiber based Fabry-Perot interferometer,” Opt. Express 27(20), 27691–22818 (2019). [CrossRef]
21. Z. L. Ran, Y. J. Rao, W. J. Liu, X. Liao, and K. S. Chiang, “Laser-micromachined Fabry-Perot optical fiber tip sensor for high-resolution temperature independent measurement of refractive index,” Opt. Express 16(3), 2252–2263 (2008). [CrossRef]
22. M. Q. Chen, H. M. Wei, Y. Zhao, X. H. Lei, and S. Krishnaswamy, “Temperature insensitive air-cavity Fabry-Perot gas pressure sensor based on core-offset fusion of hollow-core fibers,” Sensors and Actuators A: Physical 298, 111589 (2019). [CrossRef]
23. B. Xu, Y. L. Yang, Z. B. Jia, and D. N. Wang, “Hybrid Fabry-Perot interferometer for simultaneous liquid refractive index and temperature measurement,” Opt. Express 25(13), 14483–14493 (2017). [CrossRef]
24. J. F. MacDowell and G. H. Beall, “Immiscibility and crystallization in Al2O3-SiO2 glasses,” J. Amer. Ceram. Soc. 52(1), 17–25 (1969). [CrossRef]
25. B. Saruhan, W. Albers, H. Schneider, and W. A. Kaysser, “Reaction and sintering mechanisms of mullite in the systems cristobalite/α-Al2O3 and amorphous SiO2/α-Al2O3,” J. Eur. Ceram. Soc. 16(10), 1075–1081 (1996). [CrossRef]
26. S. H. Risbud and J. A. Pask, “Calculated thermodynamic data and metastable immiscibility in the system SiO2-Al2O3,” J American Ceramic Society 60(9-10), 418–424 (1977). [CrossRef]
27. N. Chen, P. Ding, Y. Shi, T. Y. Jin, Y. Y. Sun, H. Y. Wang, and Y. He, “Portable and reliable surface-enhanced Raman scattering silicon chip for signal-on detection of trace trinitrotoluene explosive in real systems,” Anal. Chem. 89(9), 5072–5078 (2017). [CrossRef]
28. R. G. Carvalho, A. J. S. Fernandes, F. J. Oliveira, E. Alves, N. Franco, C. Louro, R. F. Silva, and F. M. Costa, “Single and polycrystalline mullite fibres grown by laser floating zone technique,” J. Eur. Ceram. Soc. 30(16), 3311–3318 (2010). [CrossRef]
29. P. McMillan and B. Piriou, “The structures and vibrational spectra of crystals and glasses in the silica-alumina system,” J. Non-Cryst. Solids 53(3), 279–298 (1982). [CrossRef]
30. T. Wermelinger, C. Borgia, C. Solenthaler, and R. Spolenak, “3-D Raman spectroscopy measurements of the symmetry of residual stress fields in plastically deformed sapphire crystals,” Acta Mater. 55(14), 4657–4665 (2007). [CrossRef]
31. N. Bost, S. Duraipandian, G. Guimbretière, and J. Poirier, “Raman spectra of synthetic and natural mullite,” Vib. Spectrosc. 82, 50–52 (2016). [CrossRef]
32. C. H. Riischer, “Phonon spectra of 2:1 mullite in infrared and Raman experiments,” Phys. Chem. Minerals 23(1), 50–55 (1996). [CrossRef]
33. C. P. Zhang, Y. J. Jiang, and Y. F. Ma, “Optical floating zone growth and dielectric constants of near-3:2 mullite crystals,” J. Eur. Ceram. Soc. 36(3), 577–581 (2016). [CrossRef]
34. R. Sarma, T. Golubev, A. Yamilov, and H. Cao, “Control of light diffusion in a disordered photonic waveguide,” Appl. Phys. Lett. 105(4), 041104 (2014). [CrossRef]
35. N. Cherroret and S. E. Skipetrov, “Microscopic derivation of self-consistent equations of Anderson localization in a disordered medium of finite size,” Phys,” Phys. Rev. E 77(4), 046608 (2008). [CrossRef]
36. R. Apetz and M. P. B. van Bruggen, “Transparent alumina: A light-scattering model,” J. Am. Ceram. Soc. 86(3), 480–486 (2003). [CrossRef]
37. J. G. J. Peelen and R. Metselaar, “Light scattering by pores in polycrystalline materials: Transmission properties of alumina,” J. Appl. Phys. 45(1), 216–220 (1974). [CrossRef]
38. F. J. P. Schuurmans, D. Vanmaekelbergh, J. van de Lagemaat, and A. Lagendijk, “Strongly photonic macroporous gallium phosphide networks,” Science 284(5411), 141–143 (1999). [CrossRef]
39. S. Rotter and S. Gigan, “Light fields in complex media: Mesoscopic scattering meets wave control,” Rev. Mod. Phys. 89(1), 015005 (2017). [CrossRef]
40. R. Flores, R. Janeiro, and J. Viegas, “Optical fibre Fabry-Pérot interferometer based on inline microcavities for salinity and temperature sensing,” Sci Rep 9(1), 9556 (2019). [CrossRef]
41. D. J. Duval, S. H. Risbud, and J. F. Shackelford, “Mullite,” in Ceramic and glass materials structure properties and processing (SpringerLink, 2008), pp. 27–39.
42. Y. M. Sung, “Kinetics analysis of mullite formation reaction at high temperatures,” Acta Materialia 48(9), 2157–2162 (2000). [CrossRef]
43. J. H. Wray and J. T. Neu, “Refractive index of several glasses as a function of wavelength and temperature,” J. Opt. Soc. Am. 59(6), 774–776 (1969). [CrossRef]
44. G. Humbert and A. Malki, “Characterizations at very high temperature of electric arc-induced long-period fiber gratings,” Opt. Commun. 208(4-6), 329–335 (2002). [CrossRef]
45. Y. Mohanna, J. M. Saugrain, J. C. Rousseau, and P. Ledoux, “Relaxation of internal stresses in optical fibers,” J. Lightwave Technol. 8(12), 1799–1802 (1990). [CrossRef]
46. H. Y. Choi, K. S. Park, S. J. Park, U. C. Paek, B. H. Lee, and E. S. Choi, “Miniature fiber-optic high temperature sensor based on a hybrid structured Fabry-Perot interferometer,” Opt. Lett. 33(21), 2455–2457 (2008). [CrossRef]
47. T. Izumitani and H. Toratani, “Temperature coefficient of electronic polarizability in optical glass,” J. Non-Cryst. Solids 40(1-3), 611–619 (1980). [CrossRef]
48. Y. Himei, T. Nagakane, K. Fukumi, N. Kitamura, J. Nishii, A. Sakamoto, and K. Hirao, “Thermo-optic properties of B2O3 doped Li2O-Al2O3-SiO2 glass-ceramics,” J. Non-Cryst. Solids 354(27), 3113–3119 (2008). [CrossRef]
49. A. Koike and N. Sugimoto, “Temperature dependences of optical path length in fluorine-doped silica glass and bismuthate glass,” Optical Components and Materials III. 6116, 61160Y (2006). [CrossRef]
50. H. Schneider, “Thermal expansion of mullite,” J. Am. Ceram. Soc. 73(7), 2073–2076 (1990). [CrossRef]
51. G. Brunauer, F. Frey, H. Boysen, and H. Schneider, “High temperature thermal expansion of mullite: an in situ neutron diffraction study up to1200 °C,” J. Eur. Ceram. Soc. 21(14), 2563–2567 (2001). [CrossRef]
52. Y. Ohtsuka, “Optical coherence effects on a fiber-sensing Fabry-Perot interferometer,” Appl. Opt. 21(23), 4316–4320 (1982). [CrossRef]
