Open Access
Add to BibSonomy
Abstract
This work illustrates, to the best of our knowledge, the first non-adiabatic tapered single-mode zirconium fluoride optical fiber sensor in the mid-infrared spectral range. It is designed and fabricated via pulling and heating technique. A waist diameter dw = 25 µm with no visible crystallization is achieved, overcoming the typical fluoride glass challenges associated with crystallization, narrow temperature fabrication window, and low glass transition temperature. The performance of the non-adiabatic tapered optical fiber is theoretically and experimentally investigated, demonstrating its high potential for a wide range of sensing applications in the mid-infrared spectral range.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The Mid-InfraRed (Mid-IR) wavelength range is attracting significant scientific interest because most molecules present rotational-vibrational absorption lines in this spectral range, thus exhibiting a characteristic fingerprint [1]. Mid-IR sensing is useful in a plethora of applications, which include environmental monitoring, spectroscopic analysis, imaging, and medical diagnosis and therapy [2–4]. Therefore, there is a high demand for Mid-IR components development.
In recent years, fiber-optic sensors have received ever increasing attention, due to their well-known intrinsic advantages, e.g., immunity to electromagnetic interference, compact size, and low cost [5]. Large classes of optical fiber sensors can be roughly identified, by considering the employed operation principle/technology, e.g., long period gratings [6], fiber Bragg gratings [7,8], plasmon resonances [9], in-line interferometers [10], and tapers [11]. The fabrication techniques for silica glass have reached a state of maturity. However, silica glass exhibits high optical losses in the Mid-IR, being the attenuation magnitude order of about $36\; \textrm{dB}$ over a $40\; \textrm{cm}$ segment at the wavelength $\; \lambda \; = 3.392\; {\mathrm{\mu} \mathrm{m}}$ [11]. For Mid-IR applications, the zirconium fluoride glass, commonly referred to as fluorozirconate or ZBLAN (ZrF4-BaF2-LaF3-AlF3-NaF) glass, is one of the most employed materials, thanks to its transparency up to $\lambda = 4.5\; {\mathrm{\mu} \mathrm{m}}$ [12].
The fabrication of Mid-IR compatible devices with fluoride optical fiber is recently gaining significant attention [2,9,12–14]. In the context of fiber gratings, it is worth noting that zirconium fluoride glass is virtually non-photosensitive and requires high concentration of cerium to induce a large degree of photosensitivity [1]. For instance, She et al. have demonstrated the possibility to fabricate a long period grating in fluoroindate glass using a femtosecond laser, in 2021 [15]. In 2022, Goya et al. have reported a fluoride optical fiber sensor for refractive index measurement based on side-polished structure [16].
Fluoride optical fiber sensors with heating-based fabrication technique have not been reported in literature until now, presumably because of the challenges associated with this type of glass, including i) its tendency to crystallize (leading to high insertion loss), ii) the narrow fabrication temperature window (due to the steep viscosity-temperature profile), and iii) the low glass transition temperature, i.e., ${T_g} = 265{\; }^\circ \textrm{C}$ for ZBLAN, while ${T_g} = 1200{\; }^\circ \textrm{C}$ for silica [2].
Moreover, silica all-fiber interferometers have received considerable attention for their high sensitivity, absolute detection with wavelength codified information, low fabrication cost, broad measurement range, and compact size [5,17]. A possible implementation is based on non-adiabatic optical fiber taper, to excite cladding modes and exploit the different optical path lengths of the electromagnetic modes. This allows to obtain a comb spectrum at the output, that shifts with strain, environmental refractive index change, and temperature variation [18,19].
In this paper, for the first time, a non-adiabatic tapered ZBLAN optical fiber (Le Verre Fluoré, Bruz, France) is designed and characterized. The feasibility of temperature sensing, based on Mach-Zender interferometry of the optical modes in fluoride taper, is demonstrated. The unique properties of ZBLAN glass pertaining to the thermo-optic and the thermal expansion coefficients are exploited. The comb spectrum and the temperature sensitivity are predicted via 3-D Beam Propagation Method (BPM) simulation and electromagnetic modal analysis. To verify the operation of the fabricated proof of concept, a measurement is carried out via power readout in the Mid-IR at the wavelength $\lambda = 3.34\; {\mathrm{\mu} \mathrm{m}}$. Furthermore, a broadband characterization of the non-adiabatic tapered zirconium fluoride optical fiber confirms the Mach-Zender interferometer effect. The employment of fluoride glasses is promising for different reasons: i) the higher operating wavelength $\lambda $ leads to higher sensitivity, ii) the fluoride properties are different from those of silica, thus a more efficient temperature compensation can be performed in the Near Infrared by employing both glasses and exploiting their different thermal behavior, in case of multiparameter sensing, iii) the Mid-IR is the fingerprint region of several molecules and it can be enabled by the use of fluoride glass, iv) the sensing mechanism, based on Mach-Zender interferometry of the optical modes in fluoride taper, exhibits a temperature sensitivity very high if compared with the state of the art. In addition, the work has allowed the knowledge improvement of the theoretical principles that regulate the shift of the output spectrum with practical implications for an accurate design.
The paper is structured as follows: in Section 2, the operating mechanism of fiber in-line interferometer is mentioned, highlighting the advantages of using zirconium fluoride glass; in Section 3, the electromagnetic design of the optical fiber sensor, via BPM simulation and electromagnetic modal analysis, and the simulated sensitivity; in Section 4, the fabrication and the characterization of the device; in Section 5, the conclusions and prospects for future work.
2. Theory
Tapering an optical fiber involves the reduction of the core and cladding diameters, by heating and pulling the fiber till obtaining a waist region [20]. In single-mode optical fibers, a taper can be considered adiabatic if most of the guided power is kept in the fundamental mode LP01 [20]. In a non-adiabatic taper, the power of the fundamental mode LP01, propagating in the transition and in the waist region, is coupled into the higher order modes, see Fig. 1(a). Due to the different optical paths of the electromagnetic modes, a Mach-Zender interferometer effect is obtained, leading to a comb-like spectral pattern at the output, see Fig. 1(b) [21]. Based on two-beam optical interference equation, the output intensity can be expressed by [18,19]:
where ${I_{out}}$ is the output light intensity; ${I_1}$, and ${I_2}$ are respectively the intensities of the fundamental mode LP01 and of the cladding mode LP0m in the waist region; $\phi = 2\pi /\lambda \; \times \; \mathrm{\Delta }{n_{eff}} \times L\; = \; 2\pi /\lambda \; \times ({{n_{eff,L{P_{01}}}} - {n_{eff,L{P_{0m}}}}} )\times L $ is the phase delay of the electromagnetic modes [19]; $\mathrm{\Delta }{n_{eff}} $ is the effective refractive index difference between the fundamental mode LP01 and cladding mode LP0m; $L$ is the cavity length, i.e., the waist length. ${I_{out}}$ depends on the refractive index change, strain, and environmental temperature change [18,19].
Fig. 1. (a) Sketch of a non-adiabatic tapered optical fiber, with the down-taper, waist, and up-taper regions. (b) Schematic diagram of a non-adiabatic tapered optical fiber operating as a Mach-Zender interferometer.
The attenuation peak ${\lambda _m}$ of the m-th Mach-Zender interference can be found by satisfying the condition $\phi = 2\pi /\lambda \times \; \mathrm{\Delta }{n_{eff}} \times L = ({2m + 1} )\pi $ [5]:
The temperature sensitivity is affected by the thermo-optic effect and the thermal expansion coefficient [18]. In particular, the refractive index change is $\mathrm{\delta }n = dn/dT\; \times \mathrm{\Delta }T$, while the length change is $\mathrm{\delta }L = \alpha L\mathrm{\Delta }T$ [22]; $dn/dT$ is the thermo-optic coefficient; $\alpha $ the coefficient of thermal expansion; $\mathrm{\Delta }T$ the temperature variation [23]. For the zirconium fluoride glass, the thermo-optic coefficient is $dn/dT ={-} 1.475 \times {10^{ - 5}}\; {\textrm{K}^{ - 1}}$, and the thermal expansion coefficient is $\alpha = 1.72 \times {10^{ - 5}}\; {\textrm{K}^{ - 1}}$ [24]. These values are very different from those of the silica glass $dn/dT = 1.2 \times {10^{ - 5}}{\; }{\textrm{K}^{ - 1}}$ and $\alpha = 5.5 \times {10^{ - 7}}\; {\textrm{K}^{ - 1}}$ [25].
3. Electromagnetic design
3.1 Design of the non-adiabatic taper
The electromagnetic design of the non-adiabatic taper is carried out on a single mode ZBLAN optical fiber, ZFG SM [1.95] 6.5/125 from Le Verre Fluoré (Bruz, France). The core diameter is ${d_{co}} = 6.5\; {\mathrm{\mu} \mathrm{m}}$, the cladding diameter is ${d_{cl}} = 125\; {\mathrm{\mu} \mathrm{m}}.$ The core refractive index is ${n_{co}} = 1.49537$, the cladding refractive index is ${n_{cl}} = 1.47758$ (i.e., numerical aperture $NA = 0.23$). A 3-D BPM simulation is exploited to calculate the sensor output spectrum. The mesh grid size is ${M_x} = {M_y} = 0.25\; {\mathrm{\mu} \mathrm{m}}$ for the optical fiber cross-section and ${M_z} = 0.5\; {\mathrm{\mu} \mathrm{m}}$ for the longitudinal direction. To obtain a non-adiabatic transition, the designed down-taper ${L_{dt}} $ and the up-taper ${L_{ut}} $ lengths are ${L_{dt}} = {L_{ut}} = 2\; \textrm{mm}$. Moreover, to prevent the adjacent minima of the interferogram from being too close, the designed waist length is ${L_w} \simeq \; 8\; \textrm{mm}$ [26,27]. The simulations are here reported for ${L_w} = \; 8.05\; \textrm{mm}$, according to the fabricated proof of concept and the experimental findings. The waist diameter is ${d_w} = 25\; {\mathrm{\mu} \mathrm{m}}$. By means of pathway monitor, the excitation of the LP02 mode in the down-taper region is verified.
Figure 2 shows the normalized output power ${P_{out}}$ versus the wavelength $\lambda $ for different temperature changes $\mathrm{\Delta }T = 0{\; \textrm{K}}$, $\mathrm{\Delta }T = 50\; \textrm{K}$, $\mathrm{\Delta }T = 100\; \textrm{K}$. There is a blueshift of the dips for increasing temperature T and the simulated temperature sensitivity at the wavelength $\lambda \simeq 3310\; \textrm{nm}$ is ${S_T} ={-} 85.3\; \textrm{pm}/\textrm{K}$, calculated as the ratio between the wavelengths $\lambda $ where the normalized output power ${P_{out}} = 1$ and the associated temperature variation $\Delta T$. This result may seem counterintuitive with respect to (2), especially considering the high positive coefficient of thermal expansion $\alpha $ that increases the waist length L.
Fig. 2. Normalized output power ${P_{out}}$, simulated via BPM, versus the wavelength $\lambda $ for different temperature variations $\mathrm{\Delta }T$.
In silica tapers, the red shift of the m-th interference dip is generally explained in terms of the positive variation of $\mathrm{\Delta }{n_{eff}}$ and of the length L, which is small due to the low thermal expansion coefficient $\alpha $ [2,18,19,28].
Anyway, it is worth noting that, in silica optical fiber SMF-28 at the wavelength $\lambda \simeq 1550\; \textrm{nm}$, we have computed, via Finite Element Method (FEM), that the positive thermo-optic coefficient $dn/dT$ leads to a lower value of $\mathrm{\Delta }{n_{eff}}$ with temperature increasing. Therefore, a deeper investigation is performed to better understand the described effect.
3.2 Analysis of the results
In order to improve the understanding of the theoretical principle that determines the shift of the output spectrum, a detailed electromagnetic investigation with FEM is performed. In particular, a commercial FEM software has been employed to predict the behavior of the comb-like spectral pattern at the output, shifting with temperature T. The 2-D electromagnetic modal analysis simulation is computed in the waist region of the optical fiber ZFG SM [1.95] 6.5/125, i.e., cladding diameter ${d_{cl}} = {d_w} = 25\; {\mathrm{\mu} \mathrm{m}}$ and core diameter ${d_{co}} = 1.3\; {\mathrm{\mu} \mathrm{m}}$, for temperature variations $\mathrm{\Delta }T = 0\; \textrm{K}$, $\mathrm{\Delta }T = 50\; \textrm{K}$, $\mathrm{\Delta }T = 100\; \textrm{K}$, in the wavelength range from $\lambda = 3280\; \textrm{nm}$ to $\lambda = 3350\; \textrm{nm}$, with a step $\mathrm{\Delta }\lambda = 0.1\; \textrm{nm}$. The effective refractive indices ${n_{eff}}$ of the fundamental LP01 and of the cladding mode LP02 are calculated considering the influence of the fluoride glass thermo-optic coefficient $dn/dT ={-} 1.475 \times {10^{ - 5}}\; {\textrm{K}^{ - 1}}$ on the core and cladding refractive indices.
Figure 3 shows the effective refractive index difference $\Delta {n_{eff}}$ between the LP01 mode and the LP02 mode versus the wavelength $\lambda $ for different temperature variations $\Delta T$. The effective refractive index difference $\Delta {n_{eff}}$ between the LP01 mode and the LP02 mode increases with temperature T. In the inset, the normalized electric field norm E is reported for both the polarizations, x and y, of the LP01 mode and of the LP02 mode at the wavelength $\lambda = 3.34\; \mu m$, for temperature variation $\Delta T = 0\; \textrm{K}$.
Fig. 3. Effective refractive index difference $\mathrm{\Delta }{n_{eff}}$ between the LP01 mode and the LP02 mode, simulated via FEM, versus the wavelength $\lambda $ for different temperature variations $\mathrm{\Delta }T$. In the inset, the normalized electric field norm E is reported for both the polarizations of the LP01 mode and of the LP02 mode at the wavelength $\lambda = 3.34\; {\mathrm{\mu} \mathrm{m}}$, for temperature variation $\Delta T = 0\; \textrm{K}$.
Figure 4 shows the phase delay of the modes $\phi $ versus the wavelength $\lambda $, depending on: i) the temperature dispersion of the effective refractive index difference $\Delta {n_{eff}}$ between the LP01 mode and the LP02 mode and ii) the change in length of the waist region L, varying with the fluoride glass thermal expansion coefficient $\alpha = 1.72 \times {10^{ - 5}}\; {\textrm{K}^{ - 1}}$.
Fig. 4. Phase delay of the modes $\phi $ versus the wavelength $\lambda $, simulated via FEM, for different temperature variations $\mathrm{\Delta }T$.
Solving the equation $\phi = 2\pi /\lambda \times \; \mathrm{\Delta }{n_{eff}} \times L = ({2m + 1} )\pi $ for finding the minima, the intersection between the horizontal line $\phi = ({2m + 1} )\pi $ and the line $\phi = 2\pi /\lambda \times \; \mathrm{\Delta }{n_{eff}} \times L$ shifts to shorter wavelengths $\lambda $ (i.e., blueshift) for increasing temperature T.
Figure 5 reports the cosine of the phase delay $\phi $ versus wavelength $\lambda $ for different temperature variations $\Delta T$. The blueshift of the comb-like spectral pattern is confirmed. For a positive temperature change in zirconium fluoride glass: i) the negative thermo-optic coefficient $dn/dT$ leads to an increase of the effective refractive index difference $\Delta {n_{eff}}$ between the fundamental mode LP01 and the cladding mode LP02; ii) the positive thermal expansion coefficient $\alpha $ leads to an increase of the cavity length L. Both these effects cause the simulated blueshift. The temperature sensitivity simulated with FEM, at the wavelength $\lambda {\; }\sim {\; }3310\; \textrm{nm}$, is ${S_T} ={-} 85.3\; \textrm{pm}/\textrm{K}$. This value is in optimum agreement with 3-D BPM simulation. Moreover, the 3-D BPM simulation also gives information about the intensity I of the modes, allowing the calculation of ${I_{out}}$.
Fig. 5. Cosine of the phase delay of the modes $\phi $ versus the wavelength $\lambda $, simulated via FEM, for different temperature variations $\mathrm{\Delta }T$.
The main aspect that has brought other authors to give a different explanation of the observed redshift in silica glass is that, generally, the effective refractive index difference $\mathrm{\Delta }{n_{eff}}$ has a positive gradient with wavelength $\lambda $. In silica optical fiber, the redshift is due to a decrease of $\mathrm{\Delta }{n_{eff}}$ by increasing the temperature, which predominates over the length increasing effect, depending on the small thermal expansion coefficient $\alpha $. In zirconium fluoride optical fiber, the increase of $\mathrm{\Delta }{n_{eff}}$ and L gives a blueshift by increasing the temperature.
As further validation, we have considered the case of tensile strain instead of temperature variation. We have verified, considering the photoelastic effect [29], that the observed blueshift in silica optical fiber is due to both i) an increase of the effective refractive index difference $\mathrm{\Delta }{n_{eff}}$, and ii) an increase of L with tensile strain $\varepsilon $ [30,31].
4. Fabrication and experimental results
4.1 Fabrication
A single mode ZBLAN optical fiber, ZFG SM [1.95] 6.5/125 from Le Verre Fluoré (Bruz, France), is employed for the fabrication of the non-adiabatic taper. The segment to be tapered is stripped by using a gel to remove the polyacrylate coating. The optical fiber is then cleaned with isopropyl alcohol and cleaved with Vytran LDC-400, with $95g$ tension. Then, the optical fiber is mounted on the fiber holding blocks of the Vytran GPX-2400. Pretension F is applied before the starting of the process, moving one fiber holding block.
A commercial graphite filament is chosen by considering the cladding diameter ${d_{cl}}$ of the optical fiber. It is used for heating the fiber close to the glass softening point ${T_g} = 265\; ^\circ \textrm{C}$. This temperature is accurately maintained during the entire fabrication process. Excessive heating may lead to crystallizations and induce losses. On the other hand, if the glass is not softened, fiber breaking, or big waist diameter can be obtained. The fabrication parameters are reported in Table 1. The initial filament power ${P_f}$ needs to be finely tuned, since it is slightly dependent on ambient temperature especially when working with soft glasses [32]. Due to the hygroscopic nature of fluoride glass, heating should not be performed in ambient air, but a pure non-reactive gas should be considered [2]. Therefore, argon is employed with a flow rate ${F_{Ar}} = 0.35\; \textrm{L}/\textrm{min}$. The pre-heating, applied prior to the movement of the fiber holding block, has been optimized to avoid optical fiber bending due to gravity. The room temperature is $Tr = {\; }25{\; }^\circ \textrm{C}$, the room humidity is $RH \simeq 30{\%}$. The tapering process takes about $30\; \textrm{s}$.
Table 1. Fabrication parameters for Vytran GPX-2400
Figure 6 shows the micrographs of the fabricated non-adiabatic taper, taken with Vytran GPX-2400 CCD camera. A smooth transition is observed and a very good agreement between the designed sensor geometry and the measured one is apparent.
Fig. 6. Micrographs of the longitudinal section of the non-adiabatic tapered optical fiber, taken with Vytran GPX-2400 CCD camera; each micrograph is long $800\mathrm{\; \mu m}$.
In addition to the microscopic pictures, the measurement of the output power ${P_{out}}$ during the fabrication process of a non-adiabatic tapered zirconium fluoride optical fiber is valuable to investigate the quality of the device. For this reason, the optical fiber to be tapered is excited via laser source Thorlabs HLS635, emitting at wavelength $\lambda = 635\; \textrm{nm}$ (i.e., red light), by means of a Bare Fiber Terminator (BFT). Figure 7 shows the measurements pertaining to the output power ${P_{out}}$ and the tension F versus the drawing time t. In particular, the down-taper region is completely fabricated at time $t \simeq 17\; \textrm{s}$. The non-adiabatic taper fabrication ends at $t \simeq 37\; \textrm{s}$. Thus, by considering the power meter response time and the visual inspection of the non-adiabatic tapered optical fiber (i.e., where red light is scattered), it is possible to attribute most of the losses to the down-taper. The measured insertion loss is $IL{\; } \simeq 0.8\; \textrm{dB}$ at the wavelength $\lambda = 635\; \textrm{nm}$, suggesting a good quality of the taper.
Fig. 7. Output power ${P_{out}}$ at wavelength $\lambda = 635\; nm$ and tension F measured during the fabrication of the non-adiabatic tapered optical fiber.
4.2 Characterization
The excitation source consists of an interband cascade laser (ICL), Nanoplus Nanosystems and Technologies GmbH (Wurzburg, Germany), working at the wavelength $\lambda = 3.34\; {\mathrm{\mu} \mathrm{m}}$. It has been pigtailed with indium fluoride optical fiber IFG (0.30) 9.5/125, Le Verre Fluoré (Bruz, France). The fabricated sensor is excited via means of BFT. The farfield beam profile is captured via Pyrocam IIIHR before the down-taper, and in the waist region at the wavelength $\lambda = 3.34\; {\mathrm{\mu} \mathrm{m}}$, and is reported in Fig. 8. As confirmed by 3D-BPM simulation, also in the waist region the LP01 mode is the one with the highest intensity, see Fig. 8(b).
Fig. 8. Measured farfield beam profile, captured with Pyrocam IIIHR at the wavelength $\lambda = 3.34{\; }\mu m$; (a) before the down-taper, (b) in the waist region.
Figure 9 reports the sketch of the experimental setup adopted for the sensing performance versus temperature T. The fabricated device is positioned in an electric furnace and the output power at the wavelength $\lambda = 3.34\; {\mathrm{\mu} \mathrm{m}}$ is logged as the temperature varied from $T = 20\; ^\circ \textrm{C}$ to $T = 80\; ^\circ \textrm{C}$ with a step $\mathrm{\Delta }T = 20\; ^\circ \textrm{C}$. At each step, 3 minutes waiting time is employed to stabilize the temperature T. The output power ${P_{out}}$ is read by a power meter.
Fig. 9. Schematic of the experimental set-up. The interband cascade laser (ICL) power is coupled into the tapered ZFG SM [1.95] 6.5/125 by means of BFT. The output power is read by a thermal power sensor (PM) for different temperature T, supplied by the electric furnace.
Figure 10 reports the comparison between the experimental results (circle markers) and the BPM simulations (cross markers). The output powers of the BPM simulation have been normalized considering ${P_{out}}({T = 20\;^\circ \textrm{C}} )= 1.1\; \textrm{mW}$. The best fit line of the measured points (dash-dotted line) is also reported. A good agreement is found between the experimental results and the simulations.
Fig. 10. Output power ${P_{out}}$ at the wavelength $\lambda = 3.34{\; }\mu m$, for different temperatures T; measurements (circle markers), best fit line of the measured points (dash-dotted line), and BPM simulation (cross markers).
In order to demonstrate the interferometry between the LP01 and LP02 electromagnetic modes, a second experimental set-up has been considered, see Fig. 11. In particular, it consists of a broadband halogen lamp, a monochromator Horiba iHR550 to sample the wavelength $\lambda $, an InSb detector, cooled with liquid nitrogen, and a lock-in amplifier. The measurements are carried out in the wavelength range from $\lambda = 3.00\mathrm{\;\ \mu m}$ to $\lambda = 4.00\mathrm{\;\ \mu m}$ with a resolution $\Delta \lambda = 10\; \textrm{nm}$. Firstly, the power guided by the ZFG 6.5/125 optical fiber is measured. Then, the ZFG 6.5/125 optical fiber is aligned with the non-adiabatic tapered optical fiber based on zirconium fluoride glass and the power at the output port is measured. The measured spectrum is reported in Fig. 12. The measurement demonstrates the successful construction of the device since a comb-like spectral pattern is obtained at the output.
Fig. 11. Schematic of the experimental set-up used for the spectrum characterization of the non-adiabatic tapered optical fiber.
Fig. 12. Measured normalized output power ${P_{out}}$ of the non-adiabatic tapered optical fiber versus the wavelength $\lambda $.
This demonstrates for the first time the possibility to obtain an in-line interferometer optical fiber sensor, which exploits the peculiar characteristics of ZBLAN glass to improve the temperature sensitivity ${S_T}$. Moreover, it is expected that the temperature sensitivity ${S_T}$ could be further improved working near the upper wavelength limit of the glass transparency (i.e., $\lambda = 4.5\; {\mathrm{\mu} \mathrm{m}}$ for the zirconium fluoride glass).
4.3 Performance comparison
Table 2 reports a comparison between the proposed optical fiber interferometer and other temperature optical fiber sensors reported in literature. In case of multiparameter sensors, only the temperature sensitivity is reported, for the sake of simplicity. The proposed non-adiabatic optical fiber taper sensor exhibits temperature sensitivity ${S_T}$ performance higher than several silica sensors, even when based on more complex structures.
Table 2. Performance comparison with literature
The temperature sensitivity ${S_T}$ of the proposed sensor is negative (i.e., blueshift) instead of positive (i.e., redshift), as for most of the other devices reported in literature. This feature can be fundamental for multiparameter sensing: the condition number can be lowered, leading to a better-conditioned problem [33]. The silica sensors with lateral splicing/offset splicing require a great care to obtain repeatability. In addition, the proposed sensor is low-cost with respect to most of the devices reported in Table II, based on more sophisticated geometries. The proposed solution could be employed also as evanescent field refractive index sensor in the Mid-IR where the absorption lines of most molecules occur [34,35].
4.4 Discussion
In this paper, a non-adiabatic tapered zirconium fluoride optical fiber is proposed. Different experimental set-ups have been employed to demonstrate its feasibility, proving the low losses and the comb spectrum.
To give an example of the device potential, it is investigated for sensing applications. In particular, the fabricated sensor has been characterized versus temperature T, making use of an ICL and a thermal power sensor.
Even though the fluoride glass cannot be employed for high temperature sensing as silica, it can be a potential candidate for cryogenic temperature measurement. Silica glass is characterized by a low coefficient of thermal expansion $\alpha $ even at room temperature and its thermo-optic coefficient $dn/dT$ drastically decreases at cryogenic temperatures, resulting in a temperature sensitivity that is too low for practical applications [49,50]. On the contrary, the use of fluoride glass may improve the temperature sensitivity at cryogenic temperatures, with no need to attach materials characterized by high thermal expansion coefficient to the glass [51]. As an example, the refractive index dispersion with wavelength $\lambda $ of a BGZA fluoride glass has been measured for different temperature T, demonstrating that the thermo-optic coefficient is still remarkably high ($dn/dT{\; } \simeq {\; } - 1 \times {10^{ - 5}}{\; }{\textrm{K}^{ - 1}}$) also in the temperature range from $T = 25\; \textrm{K}$ to $T = 100\; \textrm{K}$ [52].
The numerical analysis has demonstrated that, for a positive temperature change in zirconium fluoride glass, the effective refractive index change $\Delta {n_{eff}}$ between the fundamental mode LP01 and the cladding mode LP02 increases and the cavity length L increases. These are both responsible for the negative temperature sensitivity, i.e., blueshift, differently from non-adiabatic tapered silica optical fiber, which exhibits a positive temperature sensitivity. A feasible monitoring system can be constituted by both silica and fluoride optical fiber sensors. Their combination can be strategically exploited in case of multiparameter sensing, e.g., simultaneous temperature and strain monitoring, for temperature compensation.
It is worthwhile noting the possibility to extend the proposed sensor for use in multiple potential applications. Future work will be devoted to exploit the peculiar properties of the Mid-IR light in terms of absorption lines.
5. Conclusion
A single-mode non-adiabatic tapered optical fiber based on zirconium fluoride glass is designed and fabricated for the first time. This low-cost microdevice is very attractive for Mid-IR sensing. In particular, by exploiting the zirconium fluoride thermo-optic and thermal expansion coefficients, and the possibility to operate at longer wavelength, a temperature sensor with higher sensitivity than those made of silica, to parity of structure, is demonstrated. The sensitivity operation principle is investigated via FEM and 3-D BPM simulations and a comprehensive explanation is proposed. A blueshift of the dips is simulated for increasing temperature value. The simulated sensitivity is ${S_T} ={-} 85.3\; pm/K$. The characterization is performed in the Mid-IR, by monitoring the output power variation versus temperature. The experimental findings are in good agreement with the simulations. The broadband characterization of the device has confirmed the expected comb-like spectral pattern at the output of the non-adiabatic tapered zirconium fluoride optical fiber.
Funding
Ministero dell'Università e della Ricerca (CUP: I53D23005710 001, I53D23005710 001, P2022ZF9P2, ARS01_00254); Horizon 2020 Framework Programme (101016956); Italian National Recovery and Resilience Plan (CUP: D93C22000910001, PE00000001).
Acknowledgments
The authors would like to thank Nanoplus Nanosystems and Technologies GmbH (Wurzburg, Germany) and Dr. Robert Weih for the ICL source optimization, Le Verre Fluoré (Bruz, France) and Dr. Sebastien Venck for the ICL source pigtailing.
Disclosures
The authors declare no conflicts of interest.
Data availability
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. A. Fuerbach, G. Bharathan, and M. Ams, “Grating Inscription Into Fluoride Fibers: A Review,” IEEE Photonics J. 11(5), 1–11 (2019). [CrossRef]
2. A. Séguin, R. Itzamná Becerra-Deana, S. Virally, et al., “Fabrication and characterization of indium fluoride multimode fused fiber couplers for the mid-infrared,” Opt. Express 31(20), 33670–33678 (2023). [CrossRef]
3. F. Attivissimo, A. Di Nisio, A. M. L. Lanzolla, et al., “Analysis of position estimation techniques in a surgical EM tracking system,” IEEE Sensors J. 21(13), 14389–14396 (2021). [CrossRef]
4. R. De Palo, Arianna Elefante, Gabriele Biagi, et al., “Quartz-Enhanced Photoacoustic Sensors for Detection of Eight Air Pollutants,” Adv. Photonics Res. 4(6), 1 (2023). [CrossRef]
5. P. Lu, L. Men, K. Sooley, et al., “Tapered fiber Mach–Zehnder interferometer for simultaneous measurement of refractive index and temperature,” Appl. Phys. Lett. 94(13), 1 (2009). [CrossRef]
6. Y. Wang, X. Huang, and M. Wang, “Temperature Insensitive Birefringent LPG Twist Sensing Based on the Polarization Properties,” IEEE Photonics Technol. Lett. 27(22), 2367–2370 (2015). [CrossRef]
7. F. Anelli, A. Annunziato, M. Godfrey, et al., “Effects of Curvature on Flexible Bragg Grating in Off-Axis Core: Theory and Experiment,” J. Lightwave Technol. 41(9), 2904–2910 (2023). [CrossRef]
8. D. Tosi, E. G. Macchi, M. Gallati, et al., “Fiber-optic chirped FBG for distributed thermal monitoring of ex-vivo radiofrequency ablation of liver,” Biomed. Opt. Express 5(6), 1799–1811 (2014). [CrossRef]
9. C. Zhou, Hai-Kun Zhang, Peng Song, et al., “D-Shaped Photonic Crystal Fiber Plasmon Sensors Based on Self-Reference Channel,” IEEE Photonics Technol. Lett. 32(10), 589–591 (2020). [CrossRef]
10. Y. Ge and D. N. Wang, “Optical Fiber Interferometer Based on Inner Air-Cavity With an Open Channel for Gas Pressure Sensing,” IEEE Photonics Technol. Lett. 33(21), 1201–1204 (2021). [CrossRef]
11. H. Tai, H. Tanaka, and T. Yoshino, “Fiber-optic evanescent-wave methane-gas sensor using optical absorption for the 3.392-microm line of a He-Ne laser,” Opt. Lett. 12(6), 437 (1987). [CrossRef]
12. A. Annunziato, F. Anelli, P. Le Pays Du Teilleul, et al., “Fused optical fiber combiner based on indium fluoride glass: perspectives for mid-IR applications,” Opt. Express 30(24), 44160–44174 (2022). [CrossRef]
13. M. Rezaei, G. T. Zeweldi, M. H. Mobarok Shamim, et al., “Single-mode optical fiber couplers made of fluoride glass,” Opt. Express 31(17), 27183–27191 (2023). [CrossRef]
14. F. Anelli, A. Annunziato, A. M. Loconsole, et al., “Low-Loss Fluoride Optical Fiber Coupler for Mid-Infrared Applications,” J. Lightwave Technol. 42(7), 2463 (2024). [CrossRef]
15. L. She, “Mid-infrared fluoroindate glass long-period fiber grating by femtosecond laser inscription,” Infrared Phys. Technol. 116, 103808 (2021). [CrossRef]
16. K. Goya, Y. Koyama, Y. Nishijima, et al., “A fluoride fiber optics in-line sensor for mid-IR spectroscopy based on a side-polished structure,” Sens. Actuators B Chem. 351, 130904 (2022). [CrossRef]
17. W. Zhang, X. Lang, X. Liu, et al., “Advances in Tapered Optical Fiber Sensor Structures: From Conventional to Novel and Emerging,” Biosensors 13(6), 644 (2023). [CrossRef]
18. X. Fu, Qin Liu, Yan-li Xiu, et al., “A temperature sensor based on the splicing of a core offset multi-mode fiber with two single mode fiber,” Optoelectron. Lett. 11(6), 434–437 (2015). [CrossRef]
19. T. Zhou, Y. Zhang, B. Han, et al., “Low cost non-adiabatic tapered fiber for high-sensitive temperature sensing,” Opt. Fiber Technol. 45, 53–57 (2018). [CrossRef]
20. M. I. Zibaii, H. Latifi, M. Karami, et al., “Non-adiabatic tapered optical fiber sensor for measuring the interaction between α-amino acids in aqueous carbohydrate solution,” Meas. Sci. Technol. 21(10), 105801 (2010). [CrossRef]
21. A. Layeghi, H. Latifi, and O. Frazão, “Magnetic Field Sensor Based on Nonadiabatic Tapered Optical Fiber With Magnetic Fluid,” IEEE Photonics Technol. Lett. 26(19), 1904–1907 (2014). [CrossRef]
22. L. Rosa, S. Mckee, F. Poli, et al., “Thermo-optical numerical modal analysis of multicore fibers for high power lasers and amplifiers,” Opt. Fiber Technol. 70, 102857 (2022). [CrossRef]
23. A. Annunziato, F. Anelli, J. Gates, et al., “Design of Polarization-Maintaining FBGs Using Polyimide Films to Improve Strain-Temperature Sensing in CFRP Laminates,” IEEE Photonics J. 13(2), 1–15 (2021). [CrossRef]
24. Y. Deng, R. Flores-Flores, R. K. Jain, et al., “Thermo-optomechanical oscillations in high-Q ZBLAN microspheres,” Opt. Lett. 38(21), 4413–4416 (2013). [CrossRef]
25. L. Yuan, T. Wei, Q. Han, et al., “Fiber inline Michelson interferometer fabricated by a femtosecond laser,” Opt. Lett. 37(21), 4489–4491 (2012). [CrossRef]
26. X. Leng and S. S.-H. Yam, “Mode Interference in Non-Adiabatic Fiber Taper With a Long Uniform Region,” IEEE Photonics Technol. Lett. 31(18), 1491–1494 (2019). [CrossRef]
27. Y. Liu and D. N. Wang, “Fiber In-Line Fabry-Perot Interferometer With Offset Splicing for Strain Measurement With Enhanced Sensitivity,” IEEE Photonics J. 10(1), 1–8 (2018). [CrossRef]
28. P. Lu and Q. Chen, “Asymmetrical Fiber Mach–Zehnder Interferometer for Simultaneous Measurement of Axial Strain and Temperature,” IEEE Photonics J. 2(6), 942–953 (2010). [CrossRef]
29. F. Anelli, A. Annunziato, A. Erario, et al., “Design of Microstructured Flat Optical Fiber for Multiaxial Strain Monitoring in Composite Materials,” J. Lightwave Technol. 40(17), 5986–5994 (2022). [CrossRef]
30. M. Z. Muhammad, A.A. Jasim, H. Ahmad, et al., “Non-adiabatic silica microfiber for strain and temperature sensors,” Sens. Actuators, A 192, 130–132 (2013). [CrossRef]
31. K. Q. Kieu and M. Mansuripur, “Biconical Fiber Taper Sensors,” IEEE Photonics Technol. Lett. 18(21), 2239–2241 (2006). [CrossRef]
32. É. Ducharme, S. Virally, R. Itzamná Becerra-Deana, et al., “Viscosity of fluoride glass fibers for fused component fabrication,” Appl. Opt. 61(17), 5031–5039 (2022). [CrossRef]
33. D.-P. Zhou, L. Wei, W.-K. Liu, et al., “Simultaneous measurement for strain and temperature using fiber Bragg gratings and multimode fibers,” Appl. Opt. 47(10), 1668–1672 (2008). [CrossRef]
34. K. Ke, C. Xia, M. N. Islam, et al., “Mid-infrared absorption spectroscopy and differential damage in vitro between lipids and proteins by an all-fiber-integrated supercontinuum laser,” Opt. Express 17(15), 12627–12640 (2009). [CrossRef]
35. D. Pinto, J. P. Waclawek, S. Lindner, et al., “Wavelength modulated diode probe laser for an interferometric cavity-assisted photothermal spectroscopy gas sensor,” Sens. Actuators, B 377, 133061 (2023). [CrossRef]
36. K. Wang, Y. Mizuno, H. Lee, et al., “Experimental demonstration of offset-induced sensitivity enhancement in SMS-based temperature and strain sensing,” Appl. Phys. Express 16(5), 052003 (2023). [CrossRef]
37. M. Zhou, S. Zhen, F. Liu, et al., “Single-mode fiber sensor based on core-offset splicing,” International Conference on Optical Instruments and Technology: Optical Sensors and Applications8199, 81990G (2011).
38. X. Dong, L. Su, P. Shum, et al., “Wavelength-selective all-fiber filter based on a single long-period fiber grating and a misaligned splicing point,” Opt. Commun. 258(2), 159–163 (2006). [CrossRef]
39. T. Wei, X. Lan, and H. Xiao, “Fiber Inline Core–Cladding-Mode Mach–Zehnder Interferometer Fabricated by Two-Point CO2 Laser Irradiations,” IEEE Photonics Technol. Lett. 21(10), 669–671 (2009). [CrossRef]
40. Z. Zhang, “Hollow-Core-Fiber-Based Interferometer for High-Temperature Measurements,” IEEE Photonics J. 9(2), 1–9 (2017). [CrossRef]
41. J. Zhang, Xing Fu, Yijie Shen, et al., “Ultrasensitive Temperature Sensor With Cascaded Fiber Optic Fabry–Perot Interferometers Based on Vernier Effect,” IEEE Photonics J. 10(6), 1–9 (2018). [CrossRef]
42. Y. Zhao, Ai Zhou, Huiyong Guo, et al., “An Integrated Fiber Michelson Interferometer Based on Twin-Core and Side-Hole Fibers for Multiparameter Sensing,” J. Lightwave Technol. 36(4), 993–997 (2018). [CrossRef]
43. J. Liu, D. N. Wang, and L. Zhang, “Slightly Tapered Optical Fiber With Dual Inner Air-Cavities for Simultaneous Refractive Index and Temperature Measurement,” J. Lightwave Technol. 34(21), 4872–4876 (2016). [CrossRef]
44. F. Liu, Y. Zhang, F. Meng, et al., “Complex optical fiber sensor based on the Vernier effect for temperature sensing,” Opt. Fiber Technol. 61, 102424 (2021). [CrossRef]
45. A. D. D. Le, J. Hwang, M. Yusuf, et al., “Simultaneous measurement of humidity and temperature with cytop-reduced graphene oxide-overlaid two-mode optical fiber sensor,” Sens. Actuators, B 298, 126841 (2019). [CrossRef]
46. Z. Dong, Y. Jin, G. Zhang, et al., “Single core-offset Mach-Zehnder interferometer coated with PVA for simultaneous measurement of relative humidity and temperature,” Opt. Express 29(15), 24102–24117 (2021). [CrossRef]
47. L. Duan, P. Zhang, M. Tang, et al., “Heterogeneous all-solid multicore fiber based multipath Michelson interferometer for high temperature sensing,” Opt. Express 24(18), 20210–20218 (2016). [CrossRef]
48. B. Sun, Y. Wang, J. Qu, et al., “Simultaneous measurement of pressure and temperature by employing Fabry-Perot interferometer based on pendant polymer droplet,” Opt. Express 23(3), 1906–1911 (2015). [CrossRef]
49. H. Zhang, F. Deng, Q. Wang, et al., “Development of Strain Measurement in Superconducting Magnet Through Fiber Bragg Grating,” IEEE Trans. Appl. Supercond. 18(2), 1419–1422 (2008). [CrossRef]
50. G. Rego, “Temperature Dependence of the Thermo-Optic Coefficient of SiO2 Glass,” Sensors 23(13), 6023 (2023). [CrossRef]
51. C. Lupi, F. Felli, A. Brotzu, et al., “Improving FBG Sensor Sensitivity at Cryogenic Temperature by Metal Coating,” IEEE Sens. J. 8(7), 1299–1304 (2008). [CrossRef]
52. S. Mitachi and T. Miyashita, “Refractive-index dispersion for BaF2–GdF3–ZrF4–AlF3 glasses,” Appl. Opt. 22(16), 2419–2425 (1983). [CrossRef]
