1.

INTRODUCTION

A variety of fiber-based refractive index (RI) sensors have been studied due to their excellent characteristics, such as high sensitivity, ultra-fast response, compact size, corrosion resistance, and immunity to electromagnetic interference. As a kind of commonly used fiber-optic RI sensing devices, Mach-Zehnder interferometers (MZIs) have been attracting increasing attention. Several schemes of MZI-based RI sensors have been demonstrated, including MZI with two cascaded single-mode fiber (SMF) abrupt-tapers ^1-2, MZI with two collapsed regions or two waist-broadened tapers in a photonic crystal fiber (PCF) ^3-4, MZI using two short double cladding fibers cascaded in a SMF ⁵, and single S-tapered fiber MZI ⁶, etc. One of the two tapers, two collapsed regions, or two short double cladding fibers works as a beam splitter and the other works as a beam combiner to form light interference. Currently, structure parameters of beam splitters and combiners can’t be online tuned using only one interferometer, such as offset structures and peanut-shape structures ^7-8. Optimization of their structure parameters and interference spectrum requires making a lot of fiber interferometers, which is inconvenient and costly ⁸. Besides, the abrupt-tapers of interferometers have to be very thin which results in poor mechanical strength and interferometers using special fibers such as PCFs are costly ^2-4.

In this study, a low-cost RI sensor based on an asymmetrical MZI with two different step-like tapers is proposed and demonstrated. In addition, a convenient experimental method to study the influence of step-like taper parameters on interference spectrum properties is proposed. It is realized with only one interferometer by applying offset and arc discharge time after time to the step-like taper in a fusion splicer, and monitoring the corresponding interference spectrum under different times of discharge in an optical spectrum analyzer (OSA). This method makes it easy to obtain an interference spectrum with a high extinction ratio (ER). The new simple and flexible method of on-line adjusting parameters of the taper can speed up the optimization process of interference spectrum. By simulation and Fast Fourier Transform (FFT), the change of spectrum properties in the taper parameter on-line adjustment process is analyzed and interpreted. In RI sensing experiment, the sensor could achieve a minimum RI resolution of 5.33×10^-6 RIU if the resolution of OSA were 1pm. The proposed RI sensor has advantages of high sensitivity, great mechanical strength, low-cost, and ease of fabrication.

2.

SENSOR STRUCTURE AND FABRICATION

The asymmetrical MZI is formed by cascading a step-like taper1 with small gradient and a step-like taper2 with large gradient in a SMF, as shown in Fig. 1. The two step-like tapers work as a beam splitter and a beam combiner, respectively. In step-like taper1, the input light is split into two paths: fundamental mode in fiber core and dominating cladding mode in fiber cladding. In step-like taper2, the core mode and the cladding mode combine and form interference. According to the dual-mode coupling theory, the light intensity of interference is described as:

Where I₁ and I₂ respectively represent the intensity of the core mode and the cladding mode, and ϕ is the phase difference between the two modes. The ratio between I₁ and I₂ mainly depends on the length, the minimum waist diameter and the gradient of the step-like taper. Therefore, ER of interference spectrum can be adjusted by changing parameters of the step-like taper. This has been experimentally verified using the proposed on-line spectrum adjustment method in this study.

Fig. 1.

Schematic and structure diagram of the asymmetrical MZI with two different step-like tapers.

Fig. 2. illustrates the fabrication process of the step-like taper and Fig. 3. is the corresponding photos. Fabrication process of the asymmetrical MZI and on-line adjustment of interference spectrum is as followed. Firstly, arc discharge (intensity of 100 unit and time of 1000 ms) is applied to fabricate two ellipsoidal half tapers. Secondly, an offset (6μm) is set by adjusting the motor of the fusion splicer (Fitel, S178). Finally, the step-like taper1 with a small gradient (the left taper in Fig. 1.) is completed by manually fusion splicing the two half tapers using the SMF to SMF splicing program. The other step-like taper is firstly fabricated using the same process as above and a symmetrical MZI is formed. It is worth noting that the interference spectrum can be online tuned by further applying offset and discharge (intensity of 100 unit and time of 200 ms) to the last-fabricated step-like taper and monitoring the interference spectrum in an OSA. As a result, the step-like taper2 with large gradient (the right taper in Fig. 1.) and the asymmetrical MZI are formed. With the number of offset and arc discharge increasing, interference spectrum properties such as insertion loss and ER vary correspondingly.

Fig. 2.

Schematic diagrams of the step-like taper fabrication process.

Fig. 3.

Photos of the fabrication process.

To obtain an interference spectrum with high ER, optimization experiments are carried out using the online adjustment method mentioned above. Interference spectra under consecutive times of offset and weak arc discharge to the step-like taper2 are shown in Fig. 4. As the number of offset and discharge times increases, insertion loss increases and ER increases firstly and then decreases, as shown in Fig. 5. To clearly interpret this phenomenon, interference spectra are further analyzed in spatial frequency domain by FFT. From Fig. 6, we can see that the intensity of the dominating cladding mode located at 0.0499 nm^-1 increases before the fifth discharge, resulting in the increase of ER; after that, the intensity of the dominating cladding mode decreases and that of minor cladding modes increases, causing the decrease of ER and the inhomogeneity in spectrum. To verify this phenomenon theoretically, simulation is carried out using Rsoft software. Simulation results (Fig. 7.) show that as offset between two half tapers of step-like taper2 increases, insertion loss increases and ER increases firstly and then decreases, and inhomogeneity of interference spectrum increases. Therefore, experiment results are in consistent with simulation results.

Fig. 4.

Spectra under different times of offset and arc discharge.

Fig. 5.

Corresponding ERs of spectra in Fig. 4.

Fig. 6.

Corresponding FFT results of spectra in Fig. 4.

Fig. 7.

Simulation results of MZI when step-like taper2 has different offset.

3.

RI SENSING EXPERIMENT AND RESULTS

RI sensing experiment is carried out using the system shown in Fig. 8(a). The MZI with a length of 2 cm is fixed in a liquid cell and connected to an ASE and an OSA, which provide broadband light and measure the output optical spectrum, respectively. Fig. 8(b) is the photo of the liquid cell made of polymethyl methacrylate (PMMA). An optimized MZI which has a high-quality interference spectrum with an ER of 12 dB is used in the experiment. With the increase of RI of the liquid to be measured, the interference spectrum has a red shift, as shown in Fig. 6. By linear fitting, a high sensitivity of around 187 nm/RIU is obtained in the RI range of 1.3409-1.3736.

Fig. 8.

(a). Schematic diagram of the RI sensing experiment system. (b). Photo of the liquid cell.

Fig. 9.

Interference spectra under different RIs and linear fitting results of RI sensitivities.

4.

CONCLUSIONS

In this study, a highly-sensitive and low-cost RI sensor based on a MZI is proposed and experimentally demonstrated. The MZI is fabricated by cascading two step-like tapers whose structure parameters can be online tuned to obtain a high-quality interference spectrum. RI sensing experiment results show that RI sensitivity is 187.67 nm/RIU and a resolution of 5.33×10^-6 RIU could be obtained if the resolution of OSA were 1pm. This MZI has advantages of high RI sensitivity, low-cost, and on-line adjustment of spectrum, making it a good candidate for RI sensing in many fields such as Biology Engineering, Medical Science and Environmental Science.