acoustic sensor with precise Fabry ... - OSA Publishing

41 downloads 0 Views 194KB Size Report
Fabry-Perot cavity length control using angle polished fiber. Wenhui Wang1*, Nan Wu1, Ye Tian1, Xingwei Wang 1, Christopher Niezrecki2 and Julie Chen2.
Optical pressure/acoustic sensor with precise Fabry-Perot cavity length control using angle polished fiber Wenhui Wang1*, Nan Wu1, Ye Tian1, Xingwei Wang 1, Christopher Niezrecki2 and Julie Chen2 1

Department of Electrical and Computer Engineering, University of Massachusetts Lowell, 1 University Ave., Lowell MA, 01854, USA 2 Department of Mechanical Engineering, University of Massachusetts Lowell, 1 University Ave., Lowell, MA, 01854, USA *[email protected]

Abstract: This paper presents a novel Fabry-Perot (FP) optical fiber pressure/acoustic sensor. It consists of two V-shaped grooves having different sized widths, a diaphragm on the surface of the larger V-groove, and a 45° angle-polished fiber. The precision of FP cavity length is determined by the fabrication process of photolithography and anisotropic etching of a silicon crystal. Therefore, the cavity length can be controlled on the order of ten nm. Sensors were fabricated and tested. Test results indicate that the sensors’ cavity lengths have been controlled precisely. The packaged sensor has demonstrated very good static and dynamic responses compared to a commercially available pressure sensor and a microphone. ©2008 Optical Society of America OCIS codes: (060.2370) Fiber optics sensors; (120.2230) Fabry-Perot; (230.4685) Optical microelectromechanical devices

References and links 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

G. Zhou, and L. M. Sim, “Damage detection and assessment in fibre-reinforced composite structures with embedded fibre optic sensors,” Smart Mater. Struct. 11(6), 925–939 (2002). T. Rice, R. Duncan, D. Gifford, and B. Childers, “Fiber optic distributed strain, acoustic emission, and moisture detection sensors for health maintenance,” in IEEE Systems Readiness Technology Conference(2003), pp. 505– 514. X. Wang, B. Li, Z. Xiao, S. H. Lee, H. Roman, O. L. Russo, K. K. Chin, and K. R. Farmer, “An ultra-sensitive optical MEMS sensor for partial discharge detection,” J. Micromech. Microeng. 15(3), 521–527 (2005). J. Xu, X. Wang, K. L. Cooper, and A. Wang, “Miniature all-silica fiber optic pressure and acoustic sensors,” Opt. Lett. 30(24), 3269–3271 (2005). K. K. Chin, Y. Sun, G. Feng, G. E. Georgiou, K. Guo, E. Niver, H. Roman, and K. Noe, “Fabry-Perot diaphragm fiber-optic sensor,” Appl. Opt. 46(31), 7614–7619 (2007). D. C. Abeysinghe, S. Dasgupta, J. T. Boyd, and H. E. Jackson, “A novel MEMS pressure sensor fabricated on an optical fiber,” IEEE Photon. Technol. Lett. 13(9), 993–995 (2001). X. Wang, J. Xu, Y. Zhu, K. L. Cooper, and A. Wang, “All-fused-silica miniature optical fiber tip pressure sensor,” Opt. Lett. 31(7), 885–887 (2006). Z. Yizheng, K. L. Cooper, G. R. Pickrell, and W. Anbo, “High-temperature fiber-tip pressure sensor,” J. Lightwave Technol. 24(2), 861–869 (2006). E. Cibula, and D. Donlagić, “Miniature fiber-optic pressure sensor with a polymer diaphragm,” Appl. Opt. 44(14), 2736–2744 (2005). E. Cibula, S. Pevec, B. Lenardic, E. Pinet, and D. Donlagic, “Miniature all-glass robust pressure sensor,” Opt. Express 17(7), 5098–5106 (2009). E. Cibula, and D. Donlagic, “In-line short cavity Fabry-Perot strain sensor for quasi distributed measurement utilizing standard OTDR,” Opt. Express 15(14), 8719–8730 (2007). D. Donlagic, and E. Cibula, “All-fiber high-sensitivity pressure sensor with SiO2 diaphragm,” Opt. Lett. 30(16), 2071–2073 (2005). M. J. Gander, W. N. MacPherson, J. S. Barton, R. L. Reuben, J. D. C. Jones, R. Stevens, K. S. Chana, S. J. Anderson, and T. V. Jones, “Embedded micromachined fiber-optic Fabry–Perot pressure sensors in aerodynamics applications,” IEEE Sens. J. 3(1), 102–107 (2003). R. A. Wolthuis, G. L. Mitchell, E. Saaski, J. C. Hartl, and M. A. Afromowitz, “Development of medical pressure and temperature sensors employing optical spectrum modulation,” IEEE Trans. Biomed. Eng. 38(10), 974–981 (1991).

#113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16613

15. K. Hsieh, and E. Smela, “Micromachined optical fiber based Fabry-Perot pressure sensor,” presented at MEMS Alliance Symposium, Laurel, MD, 18–19 Apr. 2005. 16. E. Pinet, A. Pham, and S. Rioux, “Miniature fiber optic pressure sensor for medical applications: an opportunity for intra-aortic balloon pumping (IABP) therapy,” Proc. SPIE 5855, 234–237 (2005). 17. A. J. Nijdam, J. G. E. Gardeniers, C. Gui, and M. Elwenspoek, “Etching pits and dislocations in Si {111},” Sens. Actuators A Phys. 86(3), 238–247 (2000).

1. Introduction Acoustic and pressure sensors have been widely used in fields such as process control, environmental monitoring, nondestructive structural health monitoring, blast characterization, and voice communication [1, 2]. A Fabry-Perot (FP) based optical sensor utilizing a thin diaphragm as a sensing element is one of the common structures due to its high sensitivity, compact size, and survivability in harsh environments [3–10]. For high bandwidth applications requiring rapid response time, the spectrum demodulation method is usually more expensive or not fast enough to demodulate the acoustic waves of high frequency or rapidly fluctuating pressure signals compared to the intensity demodulation approach. For the intensity demodulation method, the sensor is usually operated within a quasi-linear range around the quadrature point Q and the cavity length should be fabricated precisely to meet this requirement if the working wavelength is fixed [4, 11]. In addition, the diaphragm is the essential part of the sensor and must be fabricated with high quality to ensure appropriate sensitivity and repeatability. However, it is difficult to fabricate sensors featuring accurate cavity length control and a high quality thin diaphragm, as well as a high manufacturing yield coupled with low cost. Very accurate cavity length control of 3 nm precision was reported by Xu et al. [4], but the diaphragm thickness in their study was only about 5 µm. The thickness of the diaphragm can be controlled precisely by polishing and in-line monitoring during fiber etching, which was also reported [12]. However, combined precise cavity length and diaphragm thickness control has not been reported yet. Another precise cavity length tuning method using heating and pulling to change the cavity length permanently and precisely was also demonstrated [11]. But this method cannot be applied to a diaphragm based fiber sensor which does not have the lead out fiber for pulling. Deposition and etching technologies used in microelectromechanical systems (MEMS) have the ability to fabricate high quality diaphragms for optical fiber sensors [3, 13–16]. The sensor described in this paper also takes the advantages of MEMS technology, and has a novel cavity design that possesses a high quality diaphragm and precise cavity length control. The sensor has the potential to be manufactured repeatedly on a large scale with a high manufacturing yield. Within this paper the new sensor design is presented and validated empirically. 2. Principle of design and fabrication For the sensor design, two V-shaped grooves with different sized widths were fabricated by using lithography and anisotropic wet etching on a (100) silicon substrate as shown in Fig. 1. After deposition of a thin film (such as silicon nitride, shown in a green color in Fig. 1), a diaphragm was fabricated on the side wall of the larger V-groove by etching away the silicon (gray color area in section view of Fig. 1 (b) and(c)) from the backside using deep reactiveion etching (DRIE). A 45° angle-polished fiber was firmly pressed and fixed in the smaller groove to deliver and collected light to and from the diaphragm. The optical fiber side wall and the diaphragm form the FP cavity as shown in Fig. 1 (b). Because the fiber side wall surface is firmly tangent to the smaller V-groove and the diaphragm is on the larger V-groove, these two V-grooves’ positions determine the cavity length.

#113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16614

Fig. 1. 3D schematic model of the sensor. (a) oblique view the angle polished fiber and two Vgrooves within the silicon chip; (b) horizontal section view (perpendicular to the diaphragm) with details about the FP cavity. (c) vertical section view with the details about the hole etched to release the diaphragm.

The V-groove fabricated on the (100) silicon wafer by anisotropic wet etching of the silicon crystal is an ideal structure to hold standard optical fiber. The edge position of the Vgroove is determined by the position of the edge of the protection mask which is usually a thin film of silicon nitride. The protection mask is patterned by a photolithography method which is capable of providing ultra-high repeatable and accurate patterns. The accuracy of the relative positions of the V-grooves can be controlled to less than 10 nm based on current lithography technology. There are several outstanding advantages of this design: 1) Precise cavity length control: The cavity length is solely determined by precision photolithography which can be better than 10 nm. Because the grooves are fabricated at the same time, variation of the following etching process will influence the grooves in the same manner. Therefore, the repeatability within the wafer and among the wafers is guaranteed. 2) High sensitivity: The diaphragm thickness can be fabricated down to submicron level and precisely controlled based on the well-developed deposition process for MEMS and semiconductor technology. This will lead to high sensitivity. 3) High yield and low cost: Benefited from the MEMS fabrication technology, this design is potentially for high volume and low cost fabrication. Different sizes of diaphragms and V-grooves can be fabricated at the same time. They can be selected during the assembly with different diaphragms having different sensitivities. This flexibility is an important feature for mass production due to the simplified fabrication process. Multiple sensors (~1000) can be fabricated from a single wafer each having identical properties or differing properties. As an example, a sensor was fabricated by the following process: (1) A double sided polished (100) silicon wafer with 0.5 µm silicon nitride was photolithographied to open the etching window for the grooves. (2) V-grooves were fabricated by etching in (40% by weight) potassium hydroxide (KOH) solution at 70 °C. (3) 1.5 µm silica dioxide and 0.5 µm silicon nitride were deposited by chemical vapor deposition (LPCVD), respectively. (4) Backside cavities were etched by inductive coupled plasma (ICP) etching to release the diaphragm. (5) Finally, the silicon oxide was etched using 49% hydrofluoric acid (HF). Figure 2 shows the two grooves with different sizes after KOH etching. The step between these two grooves determines the cavity length. Figure 3 shows the photo of the released diaphragm taken by a dark field microscope. In this device, six diaphragms with different diameters were fabricated simultaneously. The released rings show the areas where oxide between silicon nitride and silicon has been removed. Due to the over etching of silicon at the edge near the center of the groove, the diaphragm is not a perfect circle. This may increase the sensitivity slightly but not affect the measurement too much. The etching of silicon oxide during the last step in the process increases the size of the diaphragm with a small amount. When the chip is packaged with the angle polished fiber, any one of these diaphragms can be chosen depending on the application requirements such as sensitivity and dynamic range. In this paper, a diaphragm #113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16615

with 80 µm diameter and grooves with 31.1 µm cavity length were used and a 45° angle polished fiber was fixed in the groove by epoxy. The hole at the bottom side was sealed by another silicon chip using epoxy. The single sensor size after packaging is about 11 mm×1.5 mm×0.5 mm and can be further reduced by using a thinner sealing material. For absolute pressure sensors, the manufacturing must be done in a vacuum. An epoxy-free packaging method, such as thermal bonding, is under development and may achieve better thermal stability.

Fig. 2. Scanning Electron Microscope (SEM) photo of the grooves etched by KOH.

Fig. 3. Microscopic photo of the large groove after the backside silicon has been etched away.

3. Experimental validation The optical sensor was tested side by side with a commercially available pressure transducer (Omega, PX303.) in a pressure chamber. The output of the reference pressure sensor was assumed to be the true pressure value applied to the testing chamber to evaluate the performance of the optical sensor. The spectrum shift and the output of the reference sensor were collected by a data acquisition system. Figure 4 shows the test result of sensitivity, linearity, and hysteresis. The measured sensitivity of the diaphragm for this sensor is 3.1 nm/kPa, which is less than the designed value of 14 nm/kPa. The difference between the calculated and measured sensitivity is likely attributed to variations in the material properties of the diaphragm (which are related to the fabrication method and process parameters), the thickness or geometry of the diaphragm (due to manufacturing variations), and possible flaws in the predictive model itself. All of these possible explanations are currently under investigation, but are beyond the scope of this paper which is focused on empirically validating the performance of the novel sensor. The residual tensile stress in silicon nitride may also reduce the sensitivity. The linearity is good with a correlation coefficient (R) of 0.99987. The maximum shift difference for bidirectional running for a given pressure is approximately 1 nm, which means maximum hysteresis is 0.3%. In order to make the sensor work in a linear zone when the intensity demodulation method is used, the maximum allowable diaphragm deflection should be limited within ±λ/8, where λ is the working wavelength. For the sensor above, it is approximately ±194 nm (equal to approximately ± 65 kPa) when the center wavelength is 1.55 µm.

#113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16616

Fig. 4. Linearity and hysteresis of the optical sensor. Black squares and red circles are the measurement data from the sensor when the applied pressure was increased and decreased, respectively.

The uniformity of the cavity length was tested. Reflection spectra of thirteen sensors cut from the same wafer were measured and cavity length deviations were calculated by the spectra shift from a reference value. 6 out of 13 samples are in the ± 0.05 µm range which is equal to 1/16 period shift of the spectrum. 10 out of 13 samples are within the ± 0.1 µm range. We believe that the small inconsistency is mainly attributed to the imperfections in the KOH etching process [17]. The precision can be further improved by using high quality silicon wafer that is fabricated by using a MEMS manufacturing process with a high purity etchant. Four diaphragms with the same diameter were tested for sensitivity and repeatability. The sensitivities are 17.6 nm/psi, 22.9 nm/psi, 17.8 nm/psi and 19.0 nm/psi, respectively. The variation mainly comes from the deviation of the size and shape of the diaphragm caused by unperfected ICP etching. The etching process should be optimized to minimize the influence of uneven etching depth (the diaphragm is on the V-groove side wall which is not parallel to the wafer surface) and non-conductive stop layer. Because the position of the fiber tip is not limited at the center of the diaphragm, the sensitivity can be reduced by the offset from the center of diaphragm. Therefore, the sensitivity can potentially be tuned. The intensity measurement system was also set up to verify the dynamic response of the optical fiber sensor. A laser with fixed wavelength (New Focus, 6428-SM) and a photo detector (Thorlabs, PDA10cs) were used instead of the spectrum measurement system as shown in Fig. 5. A 38 mm long, 64 mm diameter polyvinyl chloride (PVC) tube was used to confine the acoustic wave generated by an audio speaker. The reference microphone (Bruell & Kjaer (B&K) model 4939) and the optical sensor were mounted side by side on a Polymethyl methacrylate (PMMA) sheet. The PMMA sheet was fixed at the end of the PVC tube. The speaker was fixed at the other side. The reference microphone sensor has a frequency response that is flat to approximately 20 kHz (well beyond the frequency range tested) and was calibrated to have a sensitivity of 10 mV/Pa after passing the microphone through a B&K Nexus Conditioning Amplifier. The reflection spectrum and static pressure response of the optical fiber sensor were also measured and used to convert the voltage output of photo detector to pressure. Figure 6 shows the sound pressure frequency responses for a range of frequencies through 4 kHz. The speaker was excited with a sinusoidal signal at the highest level possible to avoid acoustic distortion in the speaker. The amplitude of the pressure measurement and phase difference between the signals is presented for the optical fiber sensor and the reference microphone. The results indicate that the responses of the acoustic signals are very similar. Due to the mechanical limitations of the speaker used, the signal strength at frequency higher than 4 kHz was too small to be measured correctly by this optical fiber sensor (or the microphone) whose full scale is ± 65 kPa. The deviations in the amplitude and phase can be easily attributed to the acoustical modes of the cavity which have a varying pressure and phase distribution at different locations within the chamber. The effect caused by the #113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16617

recession of the diaphragm is likely to be negligible except for very high frequency waves in which the wavelength of sound is on the same order as the dimension of the waveguide. Because both sensors are not located at the exact same position, some differences in the measurements can be expected.

Fig. 5. Schematic test set up of the intensity testing system with a speaker and an acoustical enclosure.

Fig. 6. Comparison between the dynamic responses of the optical fiber sensor and the reference microphone.

4. Conclusions A novel FP type optical fiber pressure/acoustic sensor was designed, fabricated and tested. The FP cavity is defined by two V-grooves which are fabricated by lithography and silicon anisotropic wet etching methods. The precise and reliable fabrication method provides the outstanding ability to control the accuracy and repeatability of the FP cavity length. A batch of sensor chips were fabricated and packaged with a 45° angle polished fiber for demonstration. Static testing using the spectrum measurement method shows that the cavity length can be controlled within ± 50 nm, having approximately 50% yield of the sensors on the wafer (that can be improved further). Static tests show that the sensor has very good linearity and hysteresis compared to a commercially available reference pressure sensor. Dynamic testing with a microphone was also conducted using the optical intensity measurement method. The results from the optical sensor agree very well with those from the reference microphone sensor. Acknowledgments The authors gratefully appreciate the financial support for this work provided by the U.S. Army Research Office (000000000000661 Nanomanufacturing of Multifunctional Sensors Ref. Award Number: W911NF-07-2-0081).

#113760 - $15.00 USD

(C) 2009 OSA

Received 7 Jul 2009; revised 16 Aug 2009; accepted 24 Aug 2009; published 2 Sep 2009

14 September 2009 / Vol. 17, No. 19 / OPTICS EXPRESS 16618