1. Introduction

Cryogenic-temperature measurements are crucial especially in scientific research and development, aerospace applications, and superconducting fields. Various kinds of cryogenic temperature sensing techniques based on electrical and optical methods have been reported⁽¹⁾. Traditional electrical sensors have shown a relatively good response time, accuracy, and repeatability. However, the applicability is limited if the cryogenic temperature environment to be monitored is conductive, potentially explosive, and radioactive. For its inherent advantages such as the dielectric nature, immunity to electromagnetic interference, radiation resistance, optical fiber sensors is appropriate for these environments⁽²⁾.

Optical fiber sensors, especially fiber Bragg gratings (FBG) temperature sensors are an ideal distributed temperature sensor for real-time monitoring of temperature. Nevertheless, FBG exhibits poor sensitivity within the cryogenic temperature environment. The temperature sensitivity of a standard FBG is nonlinear and is known to be five times weaker at the boiling point of liquid nitrogen, -196 ℃⁽³⁾. In order to compensate the drawbacks, several FBG cryogenic sensors have been developed in the past few years^(4-6). For instance, cryogenic temperature sensors using the FBG bonded to PMMA, Aluminum, and Teflon substrates⁽⁴⁾; aluminum precoated FBG using various secondary metal coatings^(5,6). And all the above-mentioned methods require an excessive cost of production and sophisticated coating techniques. This creates a requirement for the alternative optical sensors with simple low-cost design, multi-point capability, and miniature size.

Fiber-optic temperature sensors, which measure the Fresnel reflection from the tip of a fiber coated in a polymer with a temperature-dependent refractive index, have recently been proposed for cryogenic conditions⁽⁷⁾. To the best of our knowledge, such sensors have not been tested for multipoint cryogenic temperature measurements. In this letter, we report the simple and precise multipoint fiber-optic temperature sensor for cryogenic environments. The present sensor system has shown a potential for cryogenic temperature sensing with simple demodulation technique, ease of operation, good repeatability, and stability.

2. Operating Principle

The proposed sensor system is based on Fresnel-reflection from the interface of optical fiber end and the interfacing material, here epoxy resin and polymethyl methacrylate (PMMA). Temperature can be determined by measuring the intensity variation from the change in refractive indices of the interfacing material. Figure 1 shows the Fresnel reflection phenomenon occurring at the fiber end-polymer interface. Where $\theta_{i}$ and $\theta_{t}$ are the angle of incidence and transmission.

Fig. 1. Fresnel reflection at fiber end/ Polymer interface

3. Experimental Design

The proposed fiber sensor, consisting of an array waveguide grating (AWG) with 16 channels, arranged in linear array equidistantly at a wavelength pitch of 0.3 nm. A broadband light source (BBS) which has 70 nm of spectral width and 12 dBm output power is used as a light source. Light beam is demultiplexed into a number of beams with various wavelengths by an AWG. These beams pass through couplers to the fiber ends coated with polymers, respectively. Real-time Fresnel reflection signals from the sensing heads are synchronously received by an array of photo-detectors PD enabling fifteen different sensing points. Figure 2 presents the experimental setup of the multi-point cryogenic temperature sensor.

Fig. 2. Optical sensor configuration (ASE BBS: Amplified spontaneous emission broadband source, PD: Photodetector array, AWG: Array waveguide grating)

Fig. 3. Scanning electron microscope imaging of sensor head (a) before coating (b) after polymer coating

The light source of BBS goes to an AWG. Each sensing head consists of fiber tip is cured with a polymer material (epoxy resin) except the 1st channel. Figure 3 shows the scanning electron microscope images of sensor heads before coating (a) and after coating (b). To compensate the optical fluctuations caused by the light source and other environmental effects, 1st channel of the AWG was used as a reference signal $I_{air}$ which was placed in the air. Assuming the paraxial path in standard single mode fiber, the measured intensity $I_{m}$ and $I_{air}$ on the mth channel and 1st channel with polymer material and air can be expressed as follows⁽⁷⁾:

The calculated output as follows.

where $I_{o}$, $k$ are the intensity of incident light and transmission loss. $n_{core}$, $n_{m}$ are the refractive indices of the fiber core and the coating material. The feasibility of the constructed temperature sensor was substantiated by comparing the experimental results with a reference thermocouple.

4. Result and Discussions

We performed a series of tests to evaluate the cryogenic sensing characteristics of the proposed sensor head. The sensor was placed in a Dewar flask filled with liquid nitrogen. The temperature was varied by controlling the amount of liquid nitrogen in the flask. The reflected optical intensities are plotted versus temperature in Fig. 4(right Y-axis). The left Y-axis shows the thermo-optic coefficients of PMMA polymer. Owing to the negative thermo-optic coefficients of the polymer resins, the optical intensity varies inversely with the temperature change⁽⁷⁾.

Fig. 4. Optical reflection intensities vs temperature

Sigmoidal fits to the measured data yielded regression coefficients (R2) of 0.990 and 0.998 for the epoxy resin and PMMA, respectively, in the temperature range 90-298 K. The average sensitivities of the epoxy and PMMA sensors are 0.04 and 0.03 dB/K, respectively, which were measured with the optical spectrum analyzer over the same temperature range.

Figure 5 shows the output spectra of the proposed system when the 4 and 8th channel is placed in the air and aqueous ethanol. As expected, the reflection intensities from the sensing channels decreased according to the refractive index changes while the reflections from other channels in the air remained stable. By tracking the peak intensity variations of all channels, the proposed system could identify the cryogenic temperatures at multi-point.

Fig. 5. Output spectra of the proposed sensor system with channel 4 and 8 is coated with epoxy

Repeatability and stability are two other desirable properties for the practical application of a temperature sensor. An experiment to investigate the long-term stability of the proposed sensor was tested with continuous measurements taken over 100 minutes at controlled room temperature and cryogenic temperatures (Fig. 6). The fluctuation in temperature output is very limited. The standard deviations of temperature measurements at room and cryogenic temperatures are approximately 2.7% and 3.4%, respectively. This means that the Fresnel sensor could be used for applications in superconducting magnets where constant monitoring of temperature is crucial for optimizing performance.

Fig. 6. Stability of measured temperature as a function of time

Fig. 7. Repeatability test of Fresnel sensor embedded in an epoxy block

The repeatability measurement of the Fresnel sensor was tested over four cycles by exposing the sensor to liquid nitrogen regularly. Here, Fresnel sensor and the thermocouple are embedded in an epoxy slab to assure the identical temperature condition and the robustness of the sensor head. The measured temperature cycles are plotted in Fig. 6. The sensor exhibits a similar pattern for all the cycles. Moreover, the reflectance of the sensor recovered to its initial value repeatedly, which ensures that temperature measurements will be repeatable in the range 98-298 K (Fig. 7)

The response graphs of sensors in 50 second time window are presented in Fig. 8. The response time estimated for various cycles, from 298 K down to 90 K for Fresnel sensor, is less than 13 seconds.

The experiment carried out in an epoxy slab delays the heat transfer and results in a longer response time of embedded sensors. The results indicate a negligible difference in the values of response time for designed sensor, indicating that the technique is reputable.

Fig. 8. Response time comparison of Fresnel sensor for repeated cycles compared with a thermocouple embedded in the same epoxy block

4. Conclusion

This work presents the findings from one of the first multi-point Fresnel sensor for real-time monitoring of cryogenic temperature. The Fresnel sensor is a robust fiber sensor that is fabricated using the telecommunication single-mode fiber facet coated with a polymer material and interrogated with a traditional fiber optic device measurement instrumentation. Overall results illustrate that it is possible to use the presented temperature sensor technology to characterize a wide range of temperatures from 98-298 K. It is prototyped to allow in-situ measurement, remote interrogation in very small volumes and real-time monitoring.