1. Introduction
The Fabry–Perot interferometer (FPI) has the advantages of small size and high sensitivity, and it plays an important role in pressure sensing [
1,
2,
3,
4]. The basic principle of the fiber optic fiber sensor is multi-beam interference; since the invention of multi-beam interferometer, many kinds of Fabry–Perot (FP) sensors have been gradually developed. Until now, a variety of micro-optical fiber FP sensors have emerged, among which micro-optical fiber pressure sensors which can be used in aviation, oil fields and other pressure measurement fields have received high attention.
In recent years, interferometric pressure-sensing based on air micro-cavity in fiber is one of the research hotspots. Before that, there were researchers who reported an open micro-cavity at the end of a single-mode optical fiber fusion silicon tube for pressure sensing. The cavity can be compressed by ambient pressure, obtaining a sensitivity greater than 1000 nm/kPa. This is at least an order of magnitude higher than currently reported diaphragm-based fiber optic probe sensors. However, during the measurement, one of the reflecting surfaces in the FP chamber was formed by a liquid, and the difficulty in implementing encapsulation techniques has limited the use of structures in most environments. Therefore, most of the research focuses on the closed FP micro-cavity for pressure sensing. In a closed-cavity structure, the length of the FP cavity varies with the change of pressure [
5,
6,
7,
8]. For example, FPI based on a glass tube has a pressure sensitivity of 23.4 pm/MPa [
3]. The pressure sensitivity of FPI based on fiber optic microbubbles is 315 pm/MPa [
7]. An effective way to improve the pressure sensitivity was to reduce the thickness of the FP cavity wall or diagram, and ultra-high-pressure sensitivity can be up to 39.4 and 70.5 nm/kPa [
9]. However, the measuring range of the device is limited, the film attached to the end of the fiber is easy to crack, and the mechanical strength is weak. In 2014, Liao et al. used expansion arc discharge technology to form a bubble micro-cavity with a film thickness of 500 nm on the end face of the fiber and used it for pressure sensing [
5]. The mechanical strength was increased due to its spherical structure and submicron film thickness, thus enabling pressure detection with a sensitivity of 1036 pm/MPa in high-pressure environments. However, the sensitivity of the sensor was greatly affected, since the reflective surface of FP cavity of the structure is an uneven spherical microbubble membrane. In 2015, an FPI structure for gas pressure sensing was reported that consists of two fused capillary tubes with different inner diameters. The sensitivity is 4147 pm/MPa in the pressure range of 0~1.52 MPa [
10]. In addition, nuclear offset fusion splicing technology was used in the manufacturing process, to reduce the fringe angle, so that the detection limit of sensor parts was lowered to 4.81 kPa.
In this study, we used the optimized arc discharge technology to prepare the end-face bubble micro-cavity and prepare the fiber end-face uniform microbubble cavity with a thickness of micrometer scale. Furthermore, the end-face film of the bubble micro-cavity was transferred to the end face of the glass tube, to form the fiber FPI and apply to the field of pressure sensing. Due to the high uniformity of the silicon microbubble cavity prepared by this method, the thickness of the whole film layer can be considered to be the same under certain conditions, which can reduce unnecessary interference in the process of interference with high measurement sensitivity. More importantly, the special stable structure and high temperature stability of the end face bubble micro-cavity make it more suitable for high-pressure applications. The relationship between the structural parameters and the pressure response of thin films is theoretically and experimentally demonstrated.
3. Sensor Structure Principle Analysis and Model Simulation
The basic structure of the fiber optic FP pressure sensor is a low fineness Fabry–Perot interferometer. The inner surface of the diaphragm and the end face of the single-mode fiber constitute two cavity surfaces of the Fabry–Perot cavity. The light emitted by the laser source is coupled into the fiber core and is partially reflected when it reaches the two cavity surfaces of the diaphragm sensor, and it forms a dual-beam FP interference. The schematic diagram of the sensor probe structure is shown in
Figure 6.
According to the elastic deformation principle of thin film, pressure-sensitive film deforms under the action of external pressure, thus changing the length of FP cavity and causing the change of interference spectrum. By measuring the interference spectrum and demodulating the interference spectrum, the pressure change acting on the thin film can be obtained. The relationship between pressure change and cavity-length change can be expressed by the following formula:
where Δ
P is the change of pressure difference between inside and outside the diaphragm;
h is the thickness of the film;
r is the effective radius of the thin film, that is, the inner radius of the capillary silicon tube; Δ
L is the variation of cavity length;
μ is the Poisson ratio of the thin film; and
E is the Young’s modulus of the thin film.
We used the ANSYS16.0, according to the above theory and Equation (1); the deformation of thin films is studied in the software. Here, Δ
P was set to 1000 kPa,
h to 0.8 μm,
r to 30 μm, Poisson’s ratio
μ to 0.17 and
E to 73,000 MPa (see
Figure 7). It seems that the central position of the membrane is most affected by air pressure. As the membrane is derived from the microbubble cavity, its structural characteristics also determine if the thickness of the central position of the membrane is relatively thinner, enabling the measurement of higher sensitivity.
The two-beam interference equation is expressed as follows:
where
and
are the intensities reflected by the glass–air interface and air–glass interface;
represents the initial phase, and
is the optical phase difference between the two reflected light beams, which can be described as follows:
where
n is the refractive index of air,
L is the length of the air-cavity and
λ is the optical wavelength in vacuum.
At positions of dips in the interference spectrum, the phase difference between the two reflected light beams meets the following condition:
where
m is the integer, and
λm is the wavelength of the
m-th order interference dip. When axial contact force is applied to the FPI, the air-cavity length decreases, generating a wavelength shift of the interference dip. The wavelength shift amount of the
m-th order interference dip can be expressed as follows:
4. Sensitivity characterization and Analysis
After the preparation of the sensor structure, the sensor performance was characterized and analyzed. The experimental instruments in this experiment include the table pressure pump with high-precision barometer, optical fiber welding machine, ASE (Amplified Spontaneous Emission) broadband light source and optical spectrum analyzer (OSA). An experimental test schematic diagram is shown in
Figure 8. An ASE (Amplified Spontaneous Emission) light source was used to provide incident light, with a wavelength range from 1525 to 1610 nm for the FP interference chamber. Pressure table pumps and high-precision gauges provided and calibrated pressure changes. In this way, we could provide a linearly varying pressure value with a resolution of 10 Pa. The interference spectrum signal was demodulated by OSA (Optical spectrum analyzer).
In the experiment, OSA (Optical spectrum analysis) adopted AQ6370D of Yokogawa company, with the spectral measurement range of 600~1750 nm and the minimum measurement resolution of 20 pm. BBS (broadband source) is the KG-ASE series from Beijing, Beijing KangGuan century photoelectric technology co., LTD. The spectral width was 1525~1610 nm. In the experiment, 3–5 obvious interference peaks were formed in the broadband spectral range of 85 nm. With the change of pressure, the interference peak moved toward to the longer wave direction.
As shown in
Figure 9b, with the increase of pump pressure, the peak value of the interference spectrum showed a redshift state. When the applied pressure increased from 100.0 to 500.0 kPa, the peak value was offset by 2.7 nm.
We define the sensitivity as
α:
where Δ
λ is the offset of the spectral peak from the spectrometer measurement, and
Δp is the change in pressure from the precision barometer measurement. According to the test data, the peak value offset-pressure sensitivity was calculated to be about 6750 pm/MPa.
We attempted to conduct a pressure test on it in the range of 600 to 1000 kPa, and we obtained the results shown in
Figure 10a. We enlarged the local spectrogram to get
Figure 10b, in which we observed an obvious spectral redshift, with a shift of 2.89 nm, and calculated that its peak offset-pressure sensitivity was about 7225 pm/MPa. These results show that we can apply it to the detection of pressure change in high-pressure environments.
According to the data analysis of the upper experiments, it is believed that that fabricated sensor will perform well in the high-pressure environment. The results of the prepared sensor in the pressure range of 1100 to 1600 kPa were obtained, as shown in
Figure 11b. When the pressure increased by 100 kPa, spectral peak was also redshifted at an offset of 6.66 pm. The calculated sensitivity was 6660 pm/MPa.
We analyzed the offset of the wavelength of the central interference peak in the case that the pressure variation ranged from 100 to 1600 kPa. The image was fitted linearly. As shown in
Figure 12, we calculated that the slope of the curve is 0.00679, which means that our sensor’s pressure sensitivity was 6790 pm/MPa.
The detection limit (
η) is determined by the smallest detectable wavelength shift of the reflection spectrum (
m), and the sensitivity (
α) by
η =
m/
α.
m can be assumed to be the standard deviation of the total noise (
σ) of the output signal resulting from amplitude noise, thermal noise and OSA spectrum resolution, which can be expressed as follows:
In this paper, m can be calculated to be ~22.55 pm, and hence the detection limit η is ~3.21 kPa.
In
Table 1, we compare the sensitivity of pressure sensors of the same material and similar structure in recent years. Using the sensor preparation method in this paper, the sensitivity of the fabricated sensor was nearly doubled, compared with the recent research. The curve shows good linearity, and we can think that our sensor structure has good test performance under high-pressure environments.
5. Conclusions
In this paper, an improved method of preparing microbubbles based on pressure-assisted discharge method was studied. By controlling the arc discharge intensity, time and fiber position, to reduce the thickness of the microbubble film, the micrometer-scale all-silicon film was prepared. By controlling arc discharge and tension, the end film of the bubble micro-cavity was transferred to the structure of fusing single-mode fiber and a glass tube. The FP interference cavity at the end of the fiber could be formed. The structure was applied in the field of air-pressure sensing, and the sensitivity of the structure to air-pressure response was tested up to 6790 pm/MPa. Our structure showed great potential in the pressure environment. Further optimization of the thickness and surface area of the membrane, as well as the length of the FP cavity, is expected to achieve higher sensitivity. In this way, we can provide more sensitive structures and methods for pressure monitoring.