ALIGNED EMBOSSED DIAPHRAGM BASED FIBER OPTIC SENSOR

Abstract

The present invention is a diaphragm-fiber optic sensor (DFOS), interferometric sensor. This DFOS is based on the principles of Fabry-Perot and Michelson/Mach-Zehnder. The sensor is low cost and is designed with high efficiency, reliability, and Q-point stability, fabricated using MEMS (micro mechanic-electrical system) technology, and has demonstrated excellent performance. A DFOS according to the invention includes a cavity between two surfaces: a diaphragm made of silicon or other material with a rigid body (or boss) at the center and clamped along its edge, and the endface of a single mode optic fiber. By utilizing MEMS technology, the gap width between the diaphragm and the fiber endface is made accurately, ranging from 1 micron to 10 microns. To stabilize the Q-point of the DFOS when in use as an acoustic sensor, a system of microchannels is built in the structure of the diaphragm so that the pressure difference on two sides of the diaphragm is kept a constant, independent of the hydraulic pressure and/or low frequency noise when the device is inserted in liquid mediums.

Description

FIELD OF THE INVENTION

This invention involves the design and fabrication of an aligned embossed diaphragm based fiber optic sensor (DFOS) using MEMS technology based on the principles of Fabry-Perot and Michelson/Mach-Zehnder.

BACKGROUND OF THE INVENTION

1. Principles of Plane Wave Interferometric Sensor

1.1. Principles of Classical Plane Wave Fabry-Perot Interferometric Sensor

The theory of Fabry-Perot sensor is based on classical Fabry-Perot interferometry of plane waves (FIG. 1). Due to interference of multiply reflected beams, the intensity of the total reflected light I^(o) is expressed as Airy function

$\begin{array}{cc}{I}^{\left(o\right)}=\frac{F{\sin }^2\frac{\delta }{2}}{1+F{\sin }^2\frac{\delta }{2}}I^{\left(i\right)}& \left(1\right)\end{array}$

where I⁽ⁱ⁾ is the intensity of the incident light, and δ the phase dependent on the optic path or interference gap width L. F is the finesse, defined by

$\begin{array}{cc}F=\frac{4R}{{\left(1-R\right)}^2}& \left(2\right)\end{array}$

where

$\begin{array}{cc}R=\sqrt{r^{\prime }r^{″}}& \left(3\right)\end{array}$

r′ and r″ being the reflection coefficient at the interface of the media with n and n′, and that with n and n″, respectively. When F is small, say F<0.2, which corresponds to R<0.046, equation (1) can be approximated as [1]

$\begin{array}{cc}\frac{I^{\left(o\right)}}{I^{\left(i\right)}}\approx F{\sin }^2\frac{\delta }{2}=\frac{F}{2}\left(1-\cos \delta \right)=\frac{F}{2}\left[1+\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)\right]& \left(4\right)\end{array}$

where λ is the nominal wavelength of the light that generates the optic path differences of the series of reflected beams 2, n the refractive index of the medium of the gap (n˜1 for air, 1.33 for water, and 1.48 for oil), L the width of interference gap, and φ_o a phase factor related to the equilibrium gap width without input signal. Note that (4) depicts I^(o) as a harmonic function of L, based on which Fabry-Perot interferometric sensor is designed.

1.2. Principles of Classical Plane Wave Michelson/Mach-Zehnder Interferometric Sensor

The principles of Michelson interferometric sensor can be depicted by FIG. 2. FIG. 2 shows source 202, beamsplitter 204, mirrors 206 and detector 208. Instead of interference of multiply reflected beams of the Fabry-Perot interferometer, the output light intensity of the Michelson interferometer is due to interference of two beams [2]. One beam contains a variable optic path under measurement, and the other is the reference. The output light intensity I^(o) of an ideal lossless Michelson interferometer is expressed as

$\begin{array}{cc}{I}^{\left(o\right)}=I_1^{\left(i\right)}+I_2^{\left(i\right)}+2\sqrt{I_1^{\left(i\right)}I_2^{\left(i\right)}}\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)& \left(5\right)\end{array}$

where I₁⁽ⁱ⁾ and I₂⁽ⁱ⁾ are the intensity of the probing beam and the reference beam, respectively. When I₁⁽ⁱ⁾=I₂⁽ⁱ⁾=I⁽ⁱ⁾, (5) is reduced to

$\begin{array}{cc}{I}^{\left(o\right)}=2I^{\left(i\right)}+2I^{\left(i\right)}\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)& \left(6\right)\end{array}$

A third type of optical interference device is Mach-Zehnder interferometric sensor (FIG. 3). In a Michelson type interferometric device there is only one beam splitter or coupler, while both the sensing leg and the reference leg reflect back from the mirrors. On the other hand, in a Mach-Zehnder type device there are two beam splitters or couplers, while neither the sensing leg nor the reference leg reflect back from the mirrors. Ideally they both observe equation (6). Therefore, these are referred to herein as Michelson/Mach-Zehnder interferometric sensors.

1.3. Comparison of Plane Wave Fabry-Perot and Michelson/Mach-Zehnder Interferometric Sensors

As shown in FIGS. 1, 2 and 3, 3 dB or 50%-50% beam splitters are necessary to construct the Fabry-Perot or Michelson/Mach-Zehnder interferometric sensors. FIG. 3 includes source 302, beamsplitters 304, mirrors 306, and detector 310. The meaningful way is to define the light source output intensity as the input intensity I_(in) of the interferometric sensor, and the light intensity received by the detector as the output intensity I_(out). Thus, an ideal Fabry-Perot interferometric sensor has

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}\approx \frac{F}{8}\left[1+\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)\right]I_{\left(in\right)}<\frac{1}{40}I_{\left(in\right)}+\frac{1}{40}\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)I_{\left(in\right)}& \left(7\right)\end{array}$

while an ideal Michelson/Mach-Zehnder interferometric sensor has

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}=\frac{1}{2}\left[1+\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)\right]<\frac{1}{2}I_{\left(in\right)}+\frac{1}{2}\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)I_{\left(in\right)}& \left(8\right)\end{array}$

The efficiency in terms of I_(out)/I_(in) of a Michelson/Mach-Zehnder type interferometric sensor is more than 20 times higher than that of a Fabry-Perot type interferometric sensor.

SUMMARY OF THE INVENTION

One embodiment of the present invention is a diaphragm-fiber optic sensor (DFOS), interferometric sensor based on the principles of Fabry-Perot and Michelson/Mach-Zehnder. More specifically, the present invention is an aligned embossed diaphragm based fiber optic sensor fabricated using MEMS (micro mechanic-electrical system) technology. The DFOS includes a cavity between a mechanically clamped diaphragm and the endface of a single mode optic fiber. More specifically, the diaphragm may be embossed and may contain microchannels. The present invention is also a method of fabricating these sensors in their multiple embodiments. The present invention can be used for optical, mechanical, pressure, temperature, chemical, biometric or acoustic sensing. One specific application is the detection of on-line acoustic signature of sparking and arcing in a multitude of applications including: large electric utility transformers, auto-transformers, tap-changers, phase angle regulators, voltage regulators, reactors, circuit breakers, pipe-type high voltage cables, and other oil insulated utility and electric equipment.

BRIEF DESCRIPTION OF THE DRAWINGS

To assist those of ordinary skill in the relevant art in making and using the subject matter hereof, reference is made to the appended drawings, wherein:

FIG. 1 depicts the classical Fabry-Perot interferometer.

FIG. 2 depicts the classical Michelson/Mach-Zehnder interferometer.

FIG. 3 depicts a Mach-Zehnder interferometric sensor with its sensing branch and reference branch symmetric and merging at a second beam splitter or coupler.

FIG. 4 is a graph depicting the characteristics of a Gaussian beam in a homogeneous medium.

FIGS. 5A-C relate to a profile of a Gaussian beam and its reflection (Michelson/Mach-Zehnder) and multiple reflections (Fabry-Perot) in a diaphragm-fiber optic sensor (DFOS).

FIG. 6 depicts the structure of the embossed center of the diaphragm-fiber optic sensor (DFOS).

FIG. 7 shows the lateral diaphragm-fiber misalignment of a diaphragm-fiber optic sensor (DFOS) without the embossed center in contrast with the alignment of a diaphragm-fiber optic sensor (DFOS) with the embossed center.

FIG. 8 depicts the principles of Q-Point stabilization of DFOS using microchannels along with labels for components of the diaphragm-fiber optic sensor (DFOS).

FIG. 9 shows the design of the Q-Point stabilized diaphragm-fiber optic sensor (DFOS).

FIG. 10 is an optical microscopic image of the fabricated diaphragm of a Q-Point stabilized diaphragm-fiber optic sensor (DFOS) showing the embossed center and microchannels.

FIG. 11 depicts the static measurement of the output optic intensity as a function of pressure, which is in unit of centimeter of water column, and the optical output power is in arbitrary unit.

FIG. 12 is a graph of the dynamic measurement of the ultrasonic acoustic signal in water. The green waves represent the measurements of the Piezoelectric Acoustic Sensor (PZT) of PAC while the yellow waves represent the measurements of the diaphragm-fiber optic sensor (DFOS) of NJIT and PSE&G. A PAC Transducer generated the Acoustic Signal at 150 kHz.

DETAILED DESCRIPTION OF THE INVENTION

Definition of Symbols

• V_FP the output voltage of the DFOS
• V_FPo the maximum of output voltage of the DFOS
• n refractive index of the medium; for air n˜1
• λ wavelength of the light used for the DFOS
• L the width of the narrow gap between the back of the diaphragm and the end surface of the single mode optic fiber
• L_o the equilibrium width of the narrow gap diaphragm and the optic fiber
• φ_o the Q-point phase factor determined by the equilibrium width of the interference gap
• E Young's modulus of the material of the diaphragm
• ν Poisson coefficient of the material of the diaphragm
• η the constant of proportionality in the equation of displacement versus pressure, which is dependent on the geometric shape of the diaphragm
• u the thickness of the silicon wafer (or other material) used for the fabrication of the DFOS diaphragm
• t the thickness of the diaphragm of the DFOS
• a the square silicon chip (or other material) size of the DFOS
• b the size or length of the square diaphragm of the DFOS
• c the size of the embossed square center
• e the length of the microchannel
• f the width of the microchannel
• f′ the width of the narrow bottleneck of the microchannel
• D_out the external diameter of the stainless steel tube for the assembling of the DFOS
• D_in the internal diameter of the stainless steel tube for the assembling of the DFOS, which is equal to the diameter of the ferrule
• P_o the pressure needed for the diaphragm to bend ⅛ of the wavelength of the light λ/n
• P_atm the atmospheric pressure
• P_a the maximum pressure of the acoustic wave
• P_f the pressure at the front side of the diaphragm of the DFOS
• P_b the pressure at the back side of the diaphragm of the DFOS
• P_l the pressure at the lateral side of the diaphragm of the DFOS
• P_cap capillary pressure of the liquid in the microchannel
• P₁ the initial air pressure of the cavity, or at the backside of the diaphragm before the DFOS is immersed in the liquid
• P₂ the final air pressure of the cavity, or at the backside of the diaphragm after the DFOS is immersed in the liquid
• V₁ the initial air volume of the cavity or the backside of the diaphragm before the DFOS is immersed in the liquid
• V₂ the final air volume of the cavity or at the backside of the diaphragm after the DFOS is immersed in the liquid
• ρ the density of the liquid the DFOS is immersed in
• g gravitational acceleration, ˜9.8 m/s²
• h the depth of the liquid
• NA numerical aperture of the fiber
• θ_beam angle of spreading of the Gaussian beam
• w_o waist of the Gaussian beam
• z_o Rayleigh length of the Gaussian beam
• R wave front radius of the Gaussian beam
• n_f refractive index of the core of the step-index fiber
• n_c refractive index of the cladding of the step-index fiber

DETAILED DESCRIPTION

2. Principles of Gaussian Beam Interferometric Diaphragm-Fiber Optic Sensor (DFOS)

2.1 Principles of Gaussian Beam Interferometry

In recent years there has been extensive effort to develop Fabry-Perot type and Michelson/Mach-Zehnder type interferometric sensors [3-7], which utilize a single mode optic fiber to deliver the probing or interrogating light as well as to receive the measured interferometric signal. However, many published works completely ignored the fact that the light delivered and received by a fiber is not in the form of plane waves. Others dealt with the difference between the plane waves, based on which the classic theory of interferometry as well as equations (7) and (8) are established, and the Gaussian beams, which adequately and approximately describe the lights delivered and received by the fiber, albeit in an over-simplified way [8]. Lack of basic understanding of behavior of the Gaussian beam interferometry has prevented the creation of such sensors.

As a paraxial approximation of the solution of the Maxwell equations describing light propagating in uniform medium, the Gaussian beam (as shown in FIG. 4) is expressed by its field component [9]

$\begin{array}{cc}E\left(x,y,z\right)=E_o\frac{w_o}{w\left(z\right)}{}^{\mathrm{kz}}{}^{k\left(x^2+y^2\right)/2R\left(z\right)}{}^{-\mathrm{\eta }\left(z\right)}{}^{-\left(x^2+y^2\right)/w^2\left(z\right)}& \left(9\right)\end{array}$

where E_o is a constant, and η(z), R(z), w(z), w_o, and z_o are defined as follows

$\begin{array}{cc}\eta \left(z\right)={\tan }^{-1}\left(z/z_o\right)& \left(10\right)\\ R\left(z\right)=z+z_o^2/z& \left(11\right)\\ w\left(z\right)=w_o\sqrt{1+z^2/z_o^2}& \left(12\right)\\ w\left(z=0\right)=w_o& \left(13\right)\\ z_o=\frac{\pi w_o^2}{\lambda }& \left(14\right)\end{array}$

z_o is knows as Rayleigh range and w(z) is called spot size of a Gaussian beam. Spot size w(z) at z=0 is w_o, called beam waist. R(z) is the wavefront radius of curvature after propagating a distance z. R(z) is infinite at z=0, passes through a minimum at some finite z, and rises again toward infinity as z is further increased.

Note that in equation (9), the first three exponential terms determine the phase of the Gaussian wave. The total phase is

$\begin{array}{cc}\Psi ={\Psi }_1+{\Psi }_2+{\Psi }_3=\mathrm{kz}+\frac{1}{2}k\left(x^2+y^2\right)/R\left(z\right)-\eta \left(z\right)& \left(15\right)\\ {\Psi }_1=\mathrm{kz}& \left(16\right)\\ {\Psi }_2=\frac{1}{2}k\left(x^2+y^2\right)/R\left(z\right)& \left(17\right)\\ {\Psi }_3=-\eta \left(z\right)=-{\tan }^{-1}\left(z/z_o\right)& \left(18\right)\end{array}$

Ψ₁ is the phase of a typical plane wave; Ψ₂ is the phase depends on z and also the distance from the z-axis in the plane normal to the propagation direction. Ψ₃ is called Gouy phase. It is shown [9] that for the same z

Ψ₁>>Ψ₂,Ψ₃  (19)

Therefore, the interference of Gaussian beam propagating in uniform medium or free space can be treated as plane wave. The only modification is the loss of intensity when the Gaussian beam couples back into the fiber due to its angular spreading.

2.2 Comparison of Gaussian Beam Interferometric Fabry-Perot and Michelson/Mach-Zehnder Diaphragm-Fiber Optic Sensors (DFOS)

As in the case of plane waves, the Gaussian beam in a Fabry-Perot interferometer has multiple reflections, while the Gaussian beam in a Michelson/Mach-Zehnder interferometer reflects from the diaphragm only once. As shown in FIGS. 5A-C, optical power loss due to angular spreading of the Gaussian beam can be calculated using mirror symmetry. FIG. 5A shows diaphragm 502, diaphragm support 504, Gaussian Beam Profile 506, and endface 508 and FIG. 5B shows diaphragm surface 512 and imaginary surface 514. FIG. 5C shows integration at Z=2 L. The loss of the beam power returned back into the fiber after the first reflection from the diaphragm is due to angular spreading of the Gaussian beam's axial propagation of a distance of 2 L. The loss of the beam power returned back into the fiber after the second reflection from the diaphragm is due to angular spreading of the Gaussian beam's axial propagation of a distance of 4 L. As in plane wave interferometric sensors, the Gaussian beam diaphragm-fiber optic sensor (DFOS) can operate either as a Fabry-Perot interferometric device or as a Michelson/Mach-Zehnder interferometric device. The only difference in the diaphragm-fiber structure of the two types of Gaussian beam interferometric devices is in the surface coating of the diaphragm and the fiber endface. The Fabry-Perot type follows equation (3), while the Mach-Zehnder type mandates |r′|=0 and |r″|=1. The basic equations (7) and (8) can be modified with a loss factor α_F-P for the Fabry-Perot type and α_M for the Michelson/Mach-Zehnder type. Thus, equations (7) and (8) become

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}\approx {\alpha }_{F-P}\frac{F}{8}\left[1+\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)\right]I_{\left(in\right)}& \left(20\right)\end{array}$

and

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}={\alpha }_M\frac{1}{2}\left[1+\sin \left(\frac{4\pi n}{\lambda }L+{\phi }_o\right)\right]I_{\left(in\right)}& \left(21\right)\end{array}$

respectively. Note that the Gaussian beam of Fabry-Perot type interferometry undergoes multiple reflections, while that of the Michelson/Mach-Zehnder type interferometry reflects only once. Considering mirror symmetry of the Gaussian beam, each reflection is associated with beam size expansion due to angular spreading, and therefore induces coupling loss. Thus

α_F-P<<α_M<1  (22)

Furthermore, the Gaussian beam multiple reflection in a DFOS occurs between the diaphragm surface and the endface of the fiber, which is composed of a core, cladding and ferrule.

2.3 Comparison of Near Field and Far Field Gaussian Beam Interferometric Fabry-Perot and Michelson/Mach-Zehnder Diaphragm-Fiber Optic Sensors (DFOS)

The interference gap width of a Gaussian beam DFOS, either Fabry-Perot type or Michelson/Mach-Zehnder type, plays an important role in determining the performance of the device, due to loss when the Gaussian beam, the size of which was expanded due to angular spreading, couples back into the fiber. Two categories of the mode of operation of the Gaussian beam DFOS are defined for the purposes of this application—near field (NF) and far field (FF). The DFOS, either Fabry-Perot type or Michelson Mach-Zehnder type, operates in NF mode if the interference gap width under equilibrium condition (no signal) follows equation L_o<z_o. The DFOS, either Fabry-Perot type or Michelson/Mach-Zehnder type, operates in FF mode if the interference gap width under equilibrium condition (no signal) follows equation L_o>z_o. It follows that

α^(NF)_F-P>α^(FF)_F-P,α^(NF)_M>α^(FF)_M  (23)

In addition to efficiency, probable misalignment of the fiber, with respect to the diaphragm favors near field operation rather than far field operation. In each of the three types of misalignment—axial, lateral, and angular, more light is lost when the fiber is farther away from the diaphragm.

To summarize, the Michelson/Mach-Zehnder DFOS is preferred over the Fabry-Perot DFOS, but more difficult to implement. Near field mode is preferred than far field mode unless the endface of the fiber should be kept at a distance greater than the Rayleigh range z_o away from the diaphragm.

3. Design of Gaussian Beam DFOS

3.1 Design Consideration for Fabry-Perot Type and Michelson/Mach-Zehnder Type Gaussian Beam DFOS

There are two differences in Fabry-Perot type and Michelson/Mach-Zehnder type DFOS. The Fabry-Perot type DFOS has only one fiber delivering and receiving light, while the Michelson/Mach-Zehnder type DFOS has two fibers preferably of the same length, one for measuring the optic path of the interference gap width, and the other, the endface of which is coated with gold or aluminum for 100% reflection, for reference.

The second difference is in the surface coating. For Fabry-Perot type DFOS, if neither the endface of the fiber nor the diaphragm is coated, then

$\begin{array}{cc}R=r^{\prime }r^{″}=\frac{n^{\prime }-n}{n^{\prime }+n}·\frac{n^{″}-n}{n^{″}+n}=0.187\times 0.551=0.103>0.046& \left(24\right)\end{array}$

which characterizes the air (n=1) gap between the fiber (SiO₂, n′=1.46) and the diaphragm (Si, n″=3.45). Therefore, in order for the Airy function of Fabry-Perot interferometry to be approximated by a harmonic function for sensor application, the fiber endface can be coated with antireflection coating to reduce its r′ from 0.187 to 0.0835. For Michelson/Mach-Zehnder type DFOS, the fiber endface is coated with antireflection coating so that the transmission coefficient is close to 100%, while the Si surface of the diaphragm is coated with Au or Al so that the reflection coefficient is close to 100%.

3.2 Design Consideration for Near Field and Far Field Gaussian Beam DFOS

Device assembly is the key difference between a near field DFOS and a far field DFOS.

In one embodiment, for NF sensor, the interference width gap or the distance between the diaphragm and the fiber endface L_o is kept at 1˜10 microns by using MEMS technology. In another embodiment for FF sensor, L_o is kept much larger, and therefore its assembly is not as demanding as its NF counterpart.

3.3. Design of the Lateral Size and Diaphragm of the Gaussian Beam DFOS

Under the condition of small bending where the displacement of the diaphragm L_o−L is much smaller than the thickness t of the diaphragm, the displacement is proportional to the pressure acted on the diaphragm

$\begin{array}{cc}\Delta L=L_o-L=\frac{b^4\left(1-v^2\right)}{16\eta {\mathrm{Et}}^3}P& \left(25\right)\end{array}$

where b denotes the size of the clamped or rigidly supported diaphragm, t the thickness of the diaphragm, ν and E the Poisson coefficient and Young's modulus of the diaphragm material, respectively. For a circular diaphragm with uniform thickness of diameter b, η=5.33. For a square diaphragm with the side length b, η=4.82 [10]. In one embodiment of a circular diaphragm according to the invention including a circular center emboss (FIG. 6) [11]

$\begin{array}{cc}\eta =\frac{1}{3\left(1-\frac{c^4}{b^4}-4\frac{c^2}{b^2}\ln \frac{b}{c}\right)}& \left({25}^{\prime }\right)\end{array}$

In one embodiment of a square diaphragm with a square center emboss according to the invention may require a numerical calculation using ANSIS simulation software.

Substituting (25) into (21) and (21), it follows that

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}\approx {\alpha }_{F-P}\frac{F}{8}\left[1+\sin \left(\frac{\pi }{2P_o}P+{\phi }_o\right)\right]I_{\left(in\right)}& \left(26\right)\end{array}$

and

$\begin{array}{cc}{I}_{\left(\mathrm{out}\right)}\approx {\alpha }_M\frac{1}{2}\left[1+\sin \left(\frac{\pi }{2P_o}P+{\phi }_o\right)\right]I_{\left(in\right)}& \left(26\right)\end{array}$

for Fabry-Perot and Michelson/Mach-Zehnder DFOS operating as a pressure or acoustic sensor, respectively. P_o is the pressure for the interference gap between the diaphragm and the fiber endface to reduce by ⅛ of the wavelength

$\begin{array}{cc}{P}_o=\frac{2\eta {\mathrm{Et}}^3\lambda }{b^4\left(1-v^2\right)n}& \left(28\right)\end{array}$

3.4 Design of Embossed Center in the Diaphragm of the Gaussian Beam DFOS

In one embodiment of a piezoresistive pressure sensor, the sensing element diaphragm is designed with an embossed center to make the structure stable. In DFOS the distance between the diaphragm and the fiber endface may be kept within 1˜10μ for near field operation. Micro electro-mechanic system (MEMS) technology should be used. In one embodiment of the present invention, the incorporation of a preferably rigid embossed center in the diaphragm design (FIG. 6) of the DFOS has the following advantages:

3.4.1 A near field 1˜10μ interference gap is much easier to process and maintain for a small area—the surface of the embossed center—than the whole area of the diaphragm.

3.4.2 In one embodiment, the incorporation of a small embossed center reduces considerably the back pressure from the backside of the silicon diaphragm, which will reduce the sensibility of the DFOS. Backpressure is defined as the dynamic pressure change in the sensor cavity, which is bound by the back surface of the diaphragm and the surface of the fiber with ferrule and includes the interference gap, due to the dynamic pressure variation at the front of the diaphragm. By the law of ideal gas,

$\begin{array}{cc}\Delta P_b=-\frac{\Delta V_b}{V_b}P_b& \left(29\right)\end{array}$

where P_b is the equilibrium back pressure, V_b the equilibrium back volume, ΔP_b the back pressure increase due to the diaphragm bending ΔL caused back volume decrease of ΔV_b. By using equation (25), it follows that

$\begin{array}{cc}\Delta V_b=\alpha b^2\Delta L=-\alpha \frac{b^6\left(1-v^2\right)}{16\eta {\mathrm{Et}}^3}\Delta P_f& \left(30\right)\end{array}$

where ΔP_f is the pressure increase at the front of the diaphragm, and α depends on the shape of bended diaphragm. α is assumed to be ˜0.5 if the diaphragm keeps straight during bending, and >0.5 for a more realistic and curved diaphragm bending. As shown in FIG. 6, the following is true

V _b=(b²−c²)(u−t−L)+b²L  (31)

Substituting (9) in (7) and (8), it follows that

$\begin{array}{cc}\frac{\Delta P_b}{\Delta P_f}=\alpha \frac{16\eta {\mathrm{Et}}^3P_b}{b^6\left(1-v^2\right)}\frac{1}{\left(b^2-c^2\right)\left(u-t-L\right)+b^2L}& \left(32\right)\end{array}$

In a DFOS with a rigid embossed center, u−t−L is a few orders greater than L, while in the DFOS without an embossed center u−t=L. Therefore,

$\begin{array}{cc}{\left(\frac{\Delta P_b}{\Delta P_f}\right)}_{u-t>>L}<<{\left(\frac{\Delta P_b}{\Delta P_f}\right)}_{u-t=L}& \left(33\right)\end{array}$

Since the sensitivity of the DFOS is proportional to the pressure difference of front side and back side of the diaphragm

$\begin{array}{cc}\frac{\Delta P_f-\Delta P_b}{\Delta P_f}=1-\frac{\Delta P_b}{\Delta P_f}& \left(34\right)\end{array}$

Thus the sensitivity of the DFOS with an embossed center is greater than that without it.

3.4.3 Among the three types of misalignment—axial, lateral, and angular—of the fiber with respect to the diaphragm, the lateral is the most severe one for a flat diaphragm, since the reflected light may completely miss the fiber core (see fibers 704 in FIG. 7). For one embodiment, as shown in FIG. 7, the embossed center 706 can avoid the lateral misalignment (which occurs from non-embossed diaphragm 702) which caused low optical efficiency as well as the noise. But using MEMS technology, the embossed center can also keep the two surfaces of the gap parallel at an exact distance to avoid any misalignment-caused loss of efficiency, noise, non-uniformity and low reliability of the DFOS. This is accomplished because embossed center 706 reflects directly back to fiber core 708.

3.4.4 In one embodiment, as shown in FIG. 8, diaphragm 800 made by MEMS technology can be clamped onto the fiber (804)-ferrule (802)-stainless steel (806) (or other material) frame by mechanical, chemical, soldering, glueing, anodic bonding, and other method. As shown in FIG. 8, in mechanical bonding, a spring tightened by a screw is used to give pressure to the diaphragm so that it is clamped onto the fiber-ferrule-stainless steel (or other material) frame; and by adjusting the pressure applied to the diaphragm, the interference gap width, or quadrant- or Q-point, can be adjusted due to the elastic yielding of the bottom material of the diaphragm.

3.5 Design of Resonant Frequency of the Gaussian Beam DFOS

One of the important applications or embodiments of the DFOS being developed is in its functioning as an acoustic sensor immersed in the insulating oil of utility transformers to detect ultrasonic signal or pressure wave P(t) due to partial discharge (PD) [5-7, 12-14]. When the DFOS operates as an acoustic sensor, the design of the diaphragm—shape, size, and thickness—determines its resonant frequency. For PD application, it is optimal to measure the acoustic emission around 150 kHz. By using either analytical or numerical method, the required single mode resonant frequency can be achieved to enhance sensitivity at a certain frequency. A diaphragm with multi mode vibrations can also be designed, which can obtain broad band response in other embodiments.

3.6 Design of Q-Point Stabilized Gaussian Beam DFOS by Using Micro-channels

3.6.1 Importance of Q-Point Stabilization

When DFOS is used as an acoustic sensor, the optic signal is turned into an electrical signal by a photodiode, followed by amplifiers, and a DC filter. The output voltage of the DFOS, either Fabry-Perot or Michelson/Mach-Zehnder, as a function of the acoustic signal is expressed as

$\begin{array}{cc}V\left(t\right)=V_o\sin \left[\frac{\pi }{2P_o}P\left(t\right)+{\phi }_o\right]& \left(35\right)\end{array}$

Sensitivity of the acoustic sensor S is defined as the ratio of the voltage output and small acoustic signal input

$\begin{array}{cc}S\equiv V^{\prime }\left(P\right){|}_{P=0}=V_o\frac{\pi }{2P_o}\cos \left(\frac{\pi }{2P_o}P+{\phi }_o\right)=V_o\frac{\pi }{2P_o}\cos {\phi }_o& \left(36\right)\end{array}$

which depends on the zero input (P=0) initial phase φ_o. The sensitivity of the DFOS is determined by φ_o. When φ_o=0, S reaches maximum, and when φ_o=π/2, S is 0—the DFOS cannot detect weak acoustic signals. In addition to sensitivity, φ_o also affects dynamic range and harmonic distortion. Under the condition that the DFOS operates in single fringe, again φ_o=0 offers the best dynamic range and best harmonic distortion. When φ_o=π/2, dynamic range is also 0, since any acoustic signal will bring the device into multi-fringe regime, rendering multi-value function caused uncertainty. Therefore, keeping φ_o=0 is important for any single fringe operation interferometric acoustic sensor.

3.6.2 Static Q-Point Stabilization Using Microchannels

Q-point is defined as the point where the sine curve depicted in equation (35) crosses the V-axis. Note that φ_o is determined by the equilibrium gap width or pressure difference ΔP=P_f−P_b with no acoustic signal input. When the DFOS is used as a hydrophone or as an acoustic sensor immersed in oil for utility transformer PD monitoring, the front pressure of the diaphragm is

P _f =P _atm +ρgh+P _a e ^{iω a t}  (37)

where P_atm is the atmospheric pressure, mostly determined by the location and weather, and ρgh the pressure of the liquid, either water or oil, with h as the depth of the liquid, and ρ as the density of the liquid (FIG. 8). P_ae^iωat is the acoustic signal for measurement, with an amplitude P_a and frequency ω_a. The static pressure difference at the front and back that determines the Q-point is

ΔP=P_f−P_b=P_atm+ρgh−P_b  (38)

If the back is sealed or connected to the air as reported in [2] and [3], the difference of the front and back pressure of the diaphragm, and therefore the Q-point, will change with the weather and/or the depth of the water or oil where the sensor is immersed. In one embodiment, the typical detectable acoustic signal is less than 1 Pa, while 1 mm of water depth gives a difference of 10 Pa in the hydraulic pressure. Thus, without Q-stabilization mechanism, the DFOS cannot function as an effective acoustic sensor.

Another embodiment of the present invention introduces microchannels in the DFOS to solve this Q-point instability problem. As shown in FIG. 9, the 4 microchannels are etched out of the clamping support of the diaphragm when the diaphragm thickness t is formed. The backside of the diaphragm is thus not sealed, but connected to the outside through the microchannels. When the DFOS is immersed in water or oil, the pressure outside the cavity or backside of the DFOS is balanced by the pressure of the liquid. Inside the microchannels the backside or cavity pressure of the diaphragm P_b plus the capillary pressure P_cap is balanced by the liquid pressure outside the cavity

$\begin{array}{cc}{P}_b+P_{\mathrm{cap}}=P_l=P_{a\mathrm{tm}}+\rho g\left(h+\frac{a+t}{2}\right)& \left(39\right)\end{array}$

The capillary pressure for a circular microchannel is expressed as [15]

$\begin{array}{cc}{P}_{\mathrm{cap}}=\frac{2{\gamma }_{\mathrm{LV}}\cos \theta }{r}& \left(40\right)\end{array}$

where γ_LV is the surface tension between the liquid (in one embodiment, water or oil) and its vapor inside the microchannel, θ the contact angle of the liquid and the solid (silicon), and r the radius of the channel. For microchannels with rectangular cross section, the pressure will be of the same order as expressed in equation (16). Using γ_LV˜25×10 N/m for oil, and assuming θ to be 20°, r=200μ, then

$\begin{array}{cc}{P}_{\mathrm{cap}}\approx \frac{2\times 25\times {10}^{-3}\times 0.94}{300\times {10}^{-6}}\approx 160\mathrm{Pa}& \left(41\right)\end{array}$

which is negligible compared to the liquid static pressure ρgh, typically on the order of 10,000 Pa. Substituting equation (15) to (14), then

$\begin{array}{cc}\begin{array}{c}\Delta P=P_f-P_b\\ =\left(P_{a\mathrm{tm}}+\rho \mathrm{gh}\right)-\left[P_{a\mathrm{tm}}+\rho g\left(h+\frac{a+t}{2}\right)+P_{\mathrm{cap}}\right]\\ =-P_{\mathrm{cap}}-\rho g\left(\frac{a+t}{2}\right)=\mathrm{const}\end{array}& \left(42\right)\end{array}$

ΔP, which determines the Q-point, is a constant, independent of the environmental atmospheric pressure and liquid pressure. Therefore, after the DFOS is assembled and tested with the desired and optimized Q-point, the Q-point remains stabilized.

Equation (18) is derived for the case of the front of the diaphragm facing up. When the front of the diaphragm faces down or sideways, equation (18) will be modified without affecting the function of the microchannel as the Q-point stabilizer.

$\begin{array}{cc}\begin{array}{cc}\mathrm{Facing}\mathrm{down}& \Delta P=-P_{\mathrm{cap}}+\rho g\left(\frac{a+t}{2}\right)=\mathrm{const}\end{array}& \left(43\right)\\ \begin{array}{cc}\mathrm{Facing}\mathrm{sideway}& \Delta P=-P_{\mathrm{cap}}=\mathrm{const}\end{array}& \left(44\right)\end{array}$

The dimensions of the microchannel are determined by the depth of liquid in which the DFOS is immersed. For the embodiment shown in FIG. 9, the 4 microchannels all have an area of f×e, and the same depth as the cavity. The volume of the 4 narrow necks with width f′ is to protect the cavity or backside of the diaphragm from the liquid, and therefore can be neglected. Before the DFOS is immersed in the liquid, the volume of the cavity plus the 4 microchannels is initially

V ₁ =b ² −c ²+4ef  (45)

The deepest h_max of the liquid that the DFOS can go without the liquid invading the cavity is determined by Boyle's law

P₁V₁=P_maxV_min  (46)

where P₁, which is the atmospheric pressure P_atm, is the cavity or diaphragm backside pressure before the DFOS is immersed in the liquid, P_max the maximum pressure the backside can tolerate, and V_min the minimum of the backside volume. Substituting equation (21), as well as P_max=P_atm+ρgh_max and V_min=b²−c into equation (22), the equation to calculate the dimensions of the microchannels is obtained:

$\begin{array}{cc}{h}_{\max }=\frac{P_{a\mathrm{tm}}}{\rho g}\frac{V_{\mathrm{microchannels}}}{V_{\mathrm{cavity}}}=\frac{P_{a\mathrm{tm}}}{\rho g}\frac{4\mathrm{ef}}{b^2-c^2}& \left(47\right)\end{array}$

Similar equations can be readily derived for other designs of the microchannels.

3.6.3 Dynamic Q-Point Stabilization Using Low Pass Microchannels

In many cases, especially when the DFOS is used as a hydrophone in the sea, in addition to the three terms in equation (37), there is a low frequency noise term due to the water wave and other noise sources. The correct and complete expression of the pressure acted on the front of the diaphragm is

P _f =P _atm +ρgh+∫P _n(ω_n)e^iωntdω_n∫P_a(ω_a)e^iωatdω_a  (48)

where the first integral is with respect to the noise spectrum P_n(ω_n), while the second integral is with respect to the acoustic signal spectrum P_a (ω_a). It is assumed that the noise frequency is much lower than the acoustic signal under measurement

ω_n<<ω_a  (49)

and that the intensity of the acoustic signal is much smaller than that of the static pressures and low frequency noise

P _a(ω_a)<<P_n(ω_n),P_atm,ρgh  (50)

Ideally it is desired that the front pressure is expressed by equation (48), while the back pressure is

P _b =P _atm +ρgh+·P _n(ω_n)e^iωatdω_n  (51)

so that

ΔP=P_f−P_b=∫P_a(ω_a)e^iωatdω_a  (52)

and only the acoustic signal is detected. A special system of low pass microchannels can be designed so that the static pressure and low frequency noise can be transmitted to the backside of the diaphragm while the higher frequency acoustic signal cannot.

4. Fabrication of Gaussian Beam DFOS

An embodiment of a DFOS with embossed center and microchannels of various dimensions and surface coating of the diaphragm and the fiber endface have been fabricated as shown in FIG. 10. Specifically, FIG. 10 is an optical microscopic image of a fabricated diaphragm of Q-Point stabilized diaphragm-fiber optic sensor (DFOS) showing an embossed center and microchannels according to the invention.

5. Experimental Testing of Gaussian Beam DFOS

Over 20 fabricated DFOS have been tested. In one embodiment, the basic performance is as designed, thus confirming the validity and practical usefulness of the invention. FIG. 11 shows the static relationship between output optic intensity as a function of pressure, which is the first static interference spectrum in a Fabry-Perot diaphragm-fiber optic sensor. Note that the characteristic of Equation (4) is verified through the bending of the diaphragm through more than half wavelength. FIG. 12 is the dynamic measurement of ultrasonic acoustic signal in water. Note that the acoustic signal measured using the MEMS Fabry-Perot DFOS 1204 is comparable to the output 1202 of a piezoelectric sensor placed next to the DFOS. The FFT of input signal peaked at 150 kHz is shown at 1206.

Applicant has attempted to disclose all embodiments and applications of the disclosed subject matter that could be reasonably foreseen. However, there may be unforeseeable, insubstantial modifications that remain as equivalents. While the present invention has been described in conjunction with specific, exemplary embodiments thereof, it is evident that many alterations, modifications, and variations will be apparent to those skilled in the art in light of the foregoing description without departing from the spirit or scope of the present disclosure. Accordingly, the present disclosure is intended to embrace all such alterations, modifications, and variations of the above detailed description.

REFERENCES

• [1] M. Born and E. Wolf, Principles of Optics, p. 327, 6th Edition, Pergamon Press, (1980).
• [2] E. Hecht, Optics, p. 336, 2^nd Edition, Addison-Wesley Publishing Co. (1987).
• [3] R. A. Atkins, J. H. Gardner, W. N. Gibler, C. E. Lee, M. D. Oakland, M. O. Spears, V. P. Swenson, H. F. Taylor, J. J. McCoy, and G. Beshouri, “Fiber Fabry-Perot pressure sensors for internal combustion engines”, Applied Optics, Vol. 33, No. 7, pp. 1315-1319, March (1994).
• [4] Youngmin Kim and Dean P. Neikirk, “Design for Manufacture of Micro Fabry-Perot Cavity-based Sensors,” Sensors and Actuators A 50, January 1996, pp. 141-146.
• [5] B. Yu, D. W. Kim, J. Deng, H. Xiao and A. Wang, “Fiber Fabry-Perot sensors for detection of partial discharges in power transformers”, Applied Optics, Vol. 42, No. 16, pp. 3241-3250, June (2003).
• [6] M. Yu, “Acoustic Measurements Using a Fiber Optic Sensor System”, Journal of Intelligent Systems and Structures, Vo. 14, No. 7, 409-414 (2003)
• [7] Xiaodong Wang, et al. “an ultra-sensitive optic MEMS sensor for partial discharge detection”. Journal of Micromechanics and Microengineering, Vol. 15, pp. 521-527 (2005).
• [8] Private communication, Jonathan Scot Grimsley, Litton Poly Scientific, Blacksburg Va.
• [9] Amnon Yariv, Optic Electronics, 3^rd edition, p. 29, Library of Congress Cataloging in Publication Data, (1985).
• [10] S. Timoshenko, “Strength of Materials”, Part II, 3rd Edition, p. 97, D. Van Nostrand Co., 1956.
• [11] Mario di Giovanni, Flat and Corrugated Diaphragm Design Handbook, Marcel Dekker, New York and Base, 1982.
• [12] P. M. Eleftherion, “Partial Discharge XXI: Acoustic Emission-Based PD Source Location in Transformers”, IEEE Electrical Insulation Magazines, No. 6, pp. 22-26 November/December (1995).
• [13] Tapanes, E. E., Oanca, I., Katsifolis, J., Goode, J. and Su, Q., “The Innovative Use of Optic Fibres for Condition Monitoring of High Voltage Equipment”, Proceedings: IEEE Optimization of Electric and Electronic Equipment 1996 (Optim '96), Romania, pp. 791-814, May, (1996).
• [14] M. Minhas, J. P. Reynders, P. J. de Klerk: “Failure in power system transformers and appropriate monitoring techniques”, 11th Int. Symposium on High Voltage Engineering, London, paper 1.94.S23, (1999).
• [15] CRC Handbook of Chemistry and Physics, F70, 51st Edition, the Chemical Rubber Co. (1970-1971).

Claims

1. An embossed diaphragm-based fiber optic sensor.

2. The sensor of claim 1, wherein the sensor is fabricated using micro mechanic-electrical system technology.

3. The sensor of claim 2 further comprising a diaphragm.

4. The sensor of claim 3 further comprising a single mode optic fiber and a cavity between the diaphragm and an endface of the single mode optic fiber.

5. The sensor of claim 3, wherein the diaphragm is mechanically clamped.

6. The sensor of claim 1 further comprising at least one microchannel.

7. The sensor of claim 1, wherein the sensor is Q-point stabilized.

8. A fiber optic sensor comprising: a vibrating diaphragm;a single mode optic fiber having an endface; anda Fabry-Perot type cavity between the diaphragm and the endface.

9. The sensor of claim 8, wherein the sensor is fabricated using micro mechanic-electrical system technology.

10. The sensor of claim 8, wherein the diaphragm is mechanically clamped.

11. The sensor of claim 8 further comprising at least one microchannel.

12. The sensor of claim 8, wherein the diaphragm is embossed.

13. The sensor of claim 8, wherein the sensor is Q-point stabilized.

14. The sensor of claim 8, wherein the sensor is adapted for acoustic sensing in liquid mediums.

15. The sensor of claim 8, wherein the sensor is adapted for at least one of optical, mechanical, pressure, temperature, chemical, biometric and acoustic sensing.

16. The sensor of claim 8, wherein the sensor is adapted for detecting an on-line acoustic signature of sparking and arcing in a multitude of applications including at least one of large electric utility transformers, auto-transformers, tap-changers, phase angle regulators, voltage regulators, reactors, circuit breakers, pipe-type high voltage cables, and, other oil insulated utility and electric equipment.

17. A method of fabricating a diaphragm-based fiber optic sensor, the method comprising: forming a cavity between a diaphragm and the endface of a single mode optic fiber; andembossing the diaphragm.

18. The method of claim 17 further comprising mechanically clamping the diaphragm.

19. The method of claim 17 further comprising forming microchannels in the diaphragm using micro mechanic-electrical system technology.

20. The method of claim 17 further comprising stabilizing the Q-point of the optic sensor.

21. The method of claim 17 further comprising detecting on-line acoustic signature of sparking and arcing in at least one of large electric utility transformers, auto-transformers, tap-changers, phase angle regulators, voltage regulators, reactors, circuit breakers, and pipe-type high voltage cables.