We present a micro-electro-mechanical systems (MEMS) based Fabry-Perot (FP) sensor along with an optical system-on-a-chip (SOC) interrogator for simultaneous pressure and temperature sensing. The sensor employs a simple structure with an air-backed silicon membrane cross-axially bonded to a 45° polished optical fiber. This structure renders two cascaded FP cavities, enabling simultaneous pressure and temperature sensing in close proximity along the optical axis. The optical SOC consists of a broadband source, a MEMS FP tunable filter, a photodetector, and the supporting circuitry, serving as a miniature spectrometer for retrieving the two FP cavity lengths. Within the measured pressure and temperature ranges, experimental results demonstrate that the sensor exhibits a good linear response to external pressure and temperature changes.
© 2013 Optical Society of America
Miniature fiber-optic pressure sensors are desirable in many applications owing to their advantages, such as small size, light weight, immunity to electromagnetic interference (EMI), multiplexibility, and remote sensing capabilities. Current miniature fiber-optic pressure sensors are mainly based on three types of sensing techniques: intensity-modulated techniques, fiber Bragg gratings (FBG), and Fabry-Perot (FP) interferometers. Among these techniques, FP pressure sensors are the most widely used because of their easy construction and high sensitivities [1–3].
Most FP pressure sensors employ a co-axial configuration [4–6], in which the FP cavity is formed between a reflective diaphragm and a fiber endface, and the FP cavity shares the same optical axis with the optical fiber. A co-axial FP sensor is usually simple in design and easy to be fabricated. On the other hand, a cross-axial FP sensor [7,8] can be constructed so that the optical axis of the FP cavity is steered to be perpendicular to the axis of the optical fiber. The cross-axial configuration makes it easier to mount the sensor on a surface when the sensing diaphragm needs to be aligned parallel to the substrate. This configuration is also helpful in measuring static pressures without picking up the dynamic pressures due to the surface flows . However, cross-axial FP sensors are usually more complicated in design and difficult to fabricate.
For most FP pressure sensors, temperature drift may become an important issue, especially for sensors designed to work in a highly fluctuated temperature environment. A traditional way to address this problem is to use a separate temperature sensor to obtain the temperature information and compensate for the temperature effect in the pressure readings, which adds the complexity to the measurement system. On the other hand, simultaneous measurements of temperature and pressure are required for a broad range of applications. To address this need, a number of miniature fiber-optic sensors for simultaneous pressure and temperature measurements have been investigated. Some of these sensors utilized a FP pressure sensor along with a FBG temperature sensor for simultaneous pressure and temperature measurement [10,11]. In another approach, an all-fiber dual FP cavity structure with two distinctive cavity lengths was utilized, which allows for spectrally resolving the pressure and temperature information . For all these methods, the pressure sensing and temperature sensing happen at different locations; the temperature measurement is performed at least 1 mm away from the pressure sensing location. Moreover, although these existing FP pressure sensors may have miniature sizes, their optical interrogation systems are usually bulky, which limits the usage of these sensors in applications that require sensor instrumentation in a confined space.
In this paper, a micro-electro-mechanical systems (MEMS) based fiber-optic FP sensor that employs the cross-axial configuration for simultaneous temperature and pressure sensing is presented, which is interrogated by using an optical system-on-a-chip (SOC) [12,13]. Different from the conventional FP sensors, the sensor simply employs an air backed silicon membrane to achieve a dual-cavity (an air cavity and a silicon cavity) structure, whose reflection spectrum will exhibit two distinct frequencies resulted from the two cavities. Both of the two cavity lengths are sensitive to pressure and temperature variations, which will induce changes to the reflection spectrum. By performing a wavelength tuning with the SOC optical system and carrying out proper data processing of the spectrum information, each optical frequency component can be distinguished and monitored, enabling simultaneous pressure and temperature measurements in a small volume.
2. System configuration and working principle
The schematic of the MEMS FP sensor along with the optical SOC is shown in Fig. 1. The optical SOC is developed through hybrid integration of a superluminescent light emitting diode (SLED), a MEMS tunable FP filter, and a photodiode, as shown in Fig. 1(a). The tunable FP filter is controlled by using a MEMS comb-actuator. The SLED combined with the tunable FP filter can be used as a tunable laser to scan the spectrum over the entire working range of the FP filter. The on-chip photodiode is used to record the intensity signal at each scanned wavelength. The optical SOC can therefore be used as a miniature spectrometer with a specific wavelength range defined by the tunable FP filter for obtaining the reflection spectrum of the FP sensor.
The FP sensor is composed of a single mode fiber with a 45 degree polished endface, a silicon housing structure, and a silicon sensing diaphragm, as illustrated in Fig. 1(b). There are two FP cavities in this structure: one being a 60-μm air cavity formed between the fiber top surface and the bottom surface of the membrane, and the other being a 90-μm silicon cavity formed between two surfaces of the silicon membrane itself. Note that even though the lengths of two cascaded are not significantly different, since the reflection spectrum of the sensor is determined by the optical path difference (OPD) (OPD = 2nL, where L is the cavity length and n is the refractive index), and silicon has a much larger refractive index than air, the OPDs defined by the two cavities will be readily distinguishable, which renders the sensor reflection spectrum to exhibit two distinct frequency components, as shown in Fig. 1(c). Owing to this unique dual cavity design with the two cavities closely placed together, simultaneous pressure and temperature sensing can be achieved at extremely closed locations (within 150 μm along the optical axis).
The working principle of the sensor system for pressure and temperature measurement is elaborated as follows. For each single FP cavity, when fast Fourier transform (FFT) is performed to the wave number spectrum (i.e., intensity versus wave number (1/λ, where λ is the wavelength)) , in the FFT result, there will be a peak exhibiting at the corresponding OPD. As illustrated in Fig. 1(c), three peaks will exhibit in the FFT result, which are corresponding to the OPDs of the air cavity, the silicon cavity, and the combination of the air and silicon cavities, respectively.
It should be noted that under an externally applied pressure, the silicon membrane will deform, which will alter not only the OPD through the air cavity but also that through the silicon membrane, because the light incident angle to the silicon membrane surface will be changed. If the deformation of the silicon membrane is not large, a linear relationship between the pressure variation and OPDs changes in both cavities can be obtained. Similarly, due to the thermal expansion of the two cavities, the OPDs through the two cavities will also exhibit a linear relationship to temperature. Assuming the influences of temperature and pressure on the OPDs are independent, it can be obtained thatEq. (1) can be determined. In the experiment, for simplicity, values of A and C were determined with a fixed temperature and varying pressures, while the values of B and D were determined with a fixed pressure and varying temperatures. With an inverse operation of Eq. (1), this equation can be rewritten asEq. (1), P is the measured absolute pressure, P0 is the original pressure at the calibration, T is the measured absolute temperature, T0 is the original temperature at the calibration, OPDair and OPDsilicon are the OPDs of the air cavity and silicon cavity at the measured P and T, respectively, and OPDair0 and OPDsilicon0 are the original OPDs of the air cavity and silicon cavity at P0 and T0, respectively.
Based on the sensor’s reflection spectrum that is obtained with the optical SOC, the two OPDs (OPDair and OPDsilicon) can be determined, and used to predict the measured pressure and temperature (P and T) based on Eq. (2). Therefore, simultaneous pressure and temperature sensing can be achieved.
3. Sensor fabrication
The fabrication process of the MEMS FP sensor is illustrated in Fig. 2. The first step is to carry out thermal oxidation of 1 μm on a 4-inch wafer of 300 μm thickness. This wafer is used as the upper wafer that forms the diaphragm. The oxidation layer (SiO2) later serves as a mask for both wet etching and dry etching. After carrying out reactive-ion etching (RIE) of SiO2 to open the etching window, potassium hydroxide (KOH) etching is used to form a flat surface of the air cavity. RIE is then performed on the front side of SiO2 and 150-μm thick Si is etched away via deep reactive-ion etching (DRIE) to form the 2500μm × 2500μm × 90μm silicon diaphragm [see Fig. 3(a)]. Note that the size and thickness of the silicon membrane can be tailored to meet the requirements of different pressure and temperature sensing needs.
For the lower wafer that serves as the fiber holder, thermal oxidation is performed first. The SiO2 window is then etched through RIE and 130-μm thick Si is etched away by DRIE so that a groove [see Fig. 3(b)] can be obtained for inserting a fiber with a diameter of 125 μm. After all the SiO2 is removed, silicon-to-silicon fusion bonding is performed. A polished 45° angled fiber is then inserted into the groove of the silicon sensor housing structure under the guidance of an optical profilometer (TMS 1200, Polytech). In order to ensure that the sensor can be operated over a relatively large temperature range, a high temperature polymer epoxy (T120-023-C2, Fiber Instrument Sales) that can endure a temperature up to 300 °C is selected to seal the cavity and fix the fiber to the sensor. A fully assembled sensor device is shown in Fig. 3(c).
4. Calibration and evaluation of the sensor with the optical system-on-a-chip
As shown in Fig. 4(a), the fully integrated optical SOC includes a fiber coupled SLED chip [Fig. 4(b)], a MEMS tunable FP filter [Fig. 4(c)], a fiber coupled photodiode [Fig. 4(d)], and a supporting circuit board that consists of power supply, signal modulation, data acquisition, and data processing units. The SLED chip has a center wavelength of 1310 nm, a spectrum width of 42 nm, and an optical power output of 100 µw. The MEMS tunable FP filter consists of a well-cleaved optical fiber serving as a partial mirror of the FP cavity and a curved silicon mirror serving as another mirror, which is attached to a comb actuator to facilitate tuning of the FP cavity length. Thin metal layers were sputtered to the two mirrors to improve the quality factor of the FP cavity. The light coupling to the photodiode is achieved by using a 45 degree angled fiber, which steers the optical beam by 90 degrees so that it can be received by the photodiode. The fully integrated optical SOC with circuit board has a footprint of 1.9” × 1.8” × 0.75”, which is much smaller than any traditional optical spectrometers, such as an Optical Spectrum Analyzer (OSA). The signal to noise ratio (SNR) of the SOC was measured to be 23.4 dB.
According to the working principle of the comb actuator, under an applied driving voltage (V), the cavity length changes linearly with respect to the square of the driving voltages (V2) . The SLED source combined with the FP tunable filter can function like a tunable laser; as the cavity length of the tunable FP filter changes, the corresponding dip wavelength from the reflection of the FP filter will shift. The optical SOC was first calibrated by using an OSA. When the applied driving voltage was changed from 0 to 5 V (i.e., V2 from 0 to 25), the corresponding dip wavelength obtained from the OSA was between 1292 and 1338 nm, which is linearly proportional to V2, as shown in Fig. 5(a). The tuning range is determined by the free spectrum range (FSR) of the FP filter, which is defined as the wavelength separation between the adjacent reflection dips at the initial cavity length of 15 µm. Therefore, when the sensor is connected to the optical SOC [see Fig. 1] and wavelength sweeping is performed over the tuning range of the FP filter, the optical SOC can be used as a spectrometer to obtain the reflection spectrum of the sensor.
It should be noted that since the spectrum of the SLED light source has a Gaussian profile, background intensity normalization is needed. Before connecting the sensor to the optical SOC, wavelength sweeping using the MEMS tunable FP filter was carried out to obtain the spectrum profile of the SLED source. This procedure only needs to be performed once before any real measurements and data obtained can be stored for future usage. After connecting the sensor to the optical SOC, another wavelength sweeping was carried out and the optical power output as a function of V2 was recorded, which allows us to obtain the reflection spectrum without intensity normalization. To normalize the optical power, this un-normalized data was divided by the previously recorded spectrum profile of the SLED and the obtained normalized sensor reflection spectrum is shown in Fig. 5(b). Combined with Fig. 5(a), the reflection spectrum of the sensor in terms of wavelength can be obtained, which was then used to obtain the reflection spectrum in terms of wavenumber. Finally, after carrying out FFT to the wavenumber spectrum, the intensity as a function of OPD was obtained, as shown in Fig. 5(c). It can be clearly seen that the FFT result exhibits three peaks, which represent the air cavity OPD, the silicon cavity OPD, and the OPD of their combination. In this paper, to enhance the resolution and accuracy, the “one peak tracing” method  is used to retrieve these OPDs. Although this optical SOC has been used to perform phase demodulation of FP sensors in time domain , this is the first time that this system is demonstrated for spectrum domain signal processing for obtaining the absolute cavity lengths of multiple FP cavities.
The sensor along with the optical SOC was then used in pressure and temperature calibration experiments. As shown in Fig. 6, the sensor was placed in a customized pressure chamber with a pressure range of 1.7 to 60 psi. The pressure in the chamber was controlled by using a pressure regulator (R-68825-08, Marsh Bellofram). A conventional piezoelectric pressure sensor (LL-080-35A, Kulite Semiconductor) was used as the pressure reference. The temperature control in the chamber was achieved by using a temperature controller (CN77333, Omega Engineering Inc.), a film thermocouple (CO1-K, Omega Engineering Inc.), and two polyimide-insulated flexible film heaters (KH 103/10, Omega Engineering Inc.). The sensor was sandwiched between the two heaters that were used to control the temperature locally. The reference pressure and temperature data were recorded by the data acquisition board (DAQ) and saved in a computer.
At the room temperature of 26.1°C, the static sensor calibration curves for pressure measurement with the two cavities were obtained first by varying the pressure in the chamber from the room pressure to 35 psi, as shown in Fig. 7. For the air cavity, the average standard deviation of the pressure calibration for the cavity length is 0.02 µm (maximum value of 0.031 µm); while for the silicon cavity, it is 0.046 µm (maximum value of 0.063 µm). Further, at the room pressure of 14.68 psi, the calibration curves for temperature measurement from the two cavities were obtained by controlling the temperature inside the camber from room temperature to 250 °C, which are shown in Fig. 8. The average standard deviation of the temperature calibration for the air cavity length is 0.1 µm (maximum value of 0.13 µm), and for the silicon cavity, it is 0.04 µm (maximum value of 0.052 µm). In the experiment, no obvious hysteresis was observed for both the air cavity and the silicon cavity. For the silicon cavity, this can be attributed to the material properties of silicon. For the air cavity, the major contribution of hysteresis may come from the epoxy used to bond the optical fiber and the MEMS sensor device. However, since the sensor employs the cross-axial configuration, in which the optical axis is perpendicular to the FP cavity, the shrinkage or thermal expansion of the epoxy is expected to only introduce a shift along the horizontal optical axis, which has little impact to the change of the vertical FP cavity length. Therefore, the hysteresis and drift of the FP sensor due to the epoxy can be mitigated. Based on these calibration results, at the room temperature and pressure conditions (i.e, T0 = 26.1°C and P0 = 14.68 psi), the initial cavity lengths of the air and silicon cavities were obtained to be 67.03 µm and 87.22 µm, respectively, which are close to the designed value of 60 µm and 90 µm. Considering the silicon refractive index of 3.5 at a wavelength of 1310 nm, these cavity lengths correspond to the initial cavity lengths of OPDair0 = 134.06 µm and OPDsilicon0 = 610.56 µm.Eq. (3) as the transfer function of the sensor, after the OPDs in the air and silicon are determined by using the optical SOC with above-mentioned spectrum domain signal processing method, the measured pressure and temperature can be predicted. The pressure resolution of the sensor is determined to be 0.06 psi and the temperature resolution is 0.09 °C. To validate this transfer function, four sets of pressure and temperature measurement were carried out and the results are compared with the reference values in Table 1. In general, within the tested pressure and temperature range, the predicted pressure and temperature values have a good agreement with the reference values, with the maximum error being 5.13%. These results indicate that the MEMS FP sensor integrated with the optical SOC has been successfully demonstrated for simultaneous temperature and pressure sensing.
However, it can be seen that as the pressure and temperature variations increase, the error between the predicted value and the reference value becomes larger. The meaurement error is believed to be due to the following reasons. One reason is due to the relatively low quality factor of the MEMS tunable FP filter, which will determine the resolution of the wavelength tuning. With the current optical SOC, the tunable FP filter has a wavelength tuning resolution of 23 pm. This relatively low tuning resoltuion will introduce an systematic calibration error to Eq. (3). Note that the quality factor of the MEMS FP tunable filter can be further improved by tuning the coating reflectivity of the fiber endface and the curved mirror surface. Ultimately, if an electro-optical filter that has a much higher quality factor can be implemented in the optical SOC, this new system is expected to achieve not only a much improved wavelength tuning resolution, but also a much faster tuning speed [17–20]. The other reason is that as the pressure and temperature become high, the linear relationship described in Eq. (3) may not be valid any more and the nonlinearity of the sensor needs to be considered, which will require a nonlinear calibration method to obtain a nonlinear transfer function. In addtion, the OPDs changes with respect to pressure and temperature may not be totally independent, which can again be addressed by adopting a more appropreate sensor calibration model that counts for the crosstalk between pressure and temperature induced changes.
In this article, we have successfully demonstrated a silicon-air dual cavity FP sensor integrated with an optical SOC for simultaneous pressure and temperature measurement. The dual FP cavities are simply achieved by using an air-backed silicon membrane. Even through the silicon cavity is not significantly larger than that of the air cavity, owing to the large refractive index of the silicon, the sensor reflection spectrum exhibits a distinctive two-frequency fringe, which is detected by using the optical SOC as a miniature spectrometer. The unique sensor design renders the sensor the capability of measuring pressure and temperature in extremely close proximity, increasing the accuracy of pressure and temperature measurement. Furthermore, the sensor employs a cross-axial configuration, which allows for easy surface mounting of the sensor and accurate static pressure measurements in the presence of surface flows, and renders the potential of achieving sensor multiplexing  for distributed pressure and temperature measurement. In addition, owing to the small footprint of the optical SOC platform and the miniature size of the sensor, the entire optical sensor system can be placed and operated in a tightly confined space. This opens the door to many new application domains of optical sensing systems, such as wireless optical sensor networks, mobile robots integrated with optical sensors, and micro-air vehicles equipped with optical sensing systems.
Support received from US Department of Energy is gratefully acknowledged. The authors would like to thank Drs. Bob Romanasky, Susan Maley, and Steve Seachman for their help and support.
References and links
1. S. Nesson, M. Yu, X. Zhang, and A. H. Hsieh, “Miniature fiber optic pressure sensor with composite polymer-metal diaphragm for intradiscal pressure measurements,” J. Biomed. Opt. 13(4), 044040 (2008). [CrossRef] [PubMed]
3. S. H. Aref, M. I. Zibaii, and H. Latifi, “An improved fiber optic pressure and temperature sensor for downhole application,” Meas. Sci. Technol. 20(3), 034009 (2009). [CrossRef]
4. S. Nesson, M. Yu, X. Zhang, and A. H. Hsieh, “Miniature fiber optic pressure sensor with composite polymer-metal diaphragm for intradiscal pressure measurements,” J. Biomed. Opt. 13(4), 044040 (2008). [CrossRef] [PubMed]
6. K. Totsu, Y. Haga, T. Matsunaga, and M. Esashi, “125 μm diameter fiber-optic pressure sensor system using spectrometer-based white light interferometry with high-speed wavelength tracking,” in 3rd IEEE EMBS Spec. Top. Conf. on Microtech. in Med. and Bio. (2005), pp. 170–173.
7. W. Wang, N. Wu, Y. Tian, X. Wang, C. Niezrecki, and J. Chen, “Optical pressure/acoustic sensor with precise Fabry-Perot cavity length control using angle polished fiber,” Opt. Express 17(19), 16613–16618 (2009). [CrossRef] [PubMed]
9. V. L. Streeter and E. B. Wylie, Fluid Mechanics, 8th ed. (McGraw Hill, 1985), pp. 351–353.
10. K. Bremer, E. Lewis, B. Moss, G. Leen, S. Lochmann, and I. Mueller, “Conception and preliminary evaluation of an optical fibre sensor for simultaneous measurement of pressure and temperature,” J. Phys. Conf. Ser. 178, 012016 (2009). [CrossRef]
11. Y. Liu, Z. Guo, Y. Zhang, K. S. Chiang, and X. Dong, “Simultaneous pressure and temperature measurement with polymer-coated fibre Bragg grating,” Electron. Lett. 36(6), 564–566 (2000). [CrossRef]
12. C. Pang, M. Yu, X. M. Zhang, K. Gupta, and K. M. Bryden, “Multifunctional optical MEMS sensor platform with heterogeneous fiber optic Fabry–Pérot sensors for wireless sensor networks,” Sens. Actuators A Phys. 188, 471–480 (2012). [CrossRef]
13. C. Pang, M. Yu, X. M. Zhang, A. K. Gupta, and K. M. Bryden, “Multifunctional optical MEMS sensor platform for wireless optical sensor networks,” in The 16th Int. Conf. on Solid-State Sens., Actuators and Microsys. (Transducers’11) (2011). [CrossRef]
14. Y. Ge, M. Wang, X. Chen, and H. Rong, “An optical MEMS pressure sensor based on a phase demodulation method,” Sens. Actuators A Phys. 143(2), 224–229 (2008). [CrossRef]
15. R. Legtenberg, A. W. Groeneveld, and M. Elwenspoek, “Comb-drive actuators for large displacements,” J. Micromech. Microeng. 6(3), 320–329 (1996). [CrossRef]
16. B. Qi, G. R. Pickrell, J. C. Xu, P. Zhang, Y. H. Duan, W. Peng, Z. Y. Huang, W. Huo, H. Xiao, R. G. May, and A. Wang, “Novel data processing techniques for dispersive white light interferometer,” Opt. Eng. 42(11), 3165–3171 (2003). [CrossRef]
17. P. Dong, R. Shafiiha, S. Liao, H. Liang, N. N. Feng, D. Feng, G. Li, X. Zheng, A. V. Krishnamoorthy, and M. Asghari, “Wavelength-tunable silicon microring modulator,” Opt. Express 18(11), 10941–10946 (2010). [CrossRef] [PubMed]
18. S. J. Choi and P. D. Dapkus, “Tunable narrow linewidth all-buried heterostructure ring resonator filters using vernier effects,” IEEE Photon. Technol. Lett. 17(1), 106–108 (2005). [CrossRef]
19. T. Claes, W. Bogaerts, and P. Bienstman, “Experimental characterization of a silicon photonic biosensor consisting of two cascaded ring resonators based on the Vernier-effect and introduction of a curve fitting method for an improved detection limit,” Opt. Express 18(22), 22747–22761 (2010). [CrossRef] [PubMed]
20. L. Zhang, J. Y. Yang, M. Song, Y. Li, B. Zhang, R. G. Beausoleil, and A. E. Willner, “Microring-based modulation and demodulation of DPSK signal,” Opt. Express 15(18), 11564–11569 (2007). [CrossRef] [PubMed]
21. X. Li, Q. Sun, D. Liu, R. Liang, J. Zhang, J. Wo, P. P. Shum, and D. Liu, “Simultaneous wavelength and frequency encoded microstructure based quasi-distributed temperature sensor,” Opt. Express 20(11), 12076–12084 (2012). [CrossRef] [PubMed]