The study of optical fiber sensors has passed almost thirty years, and various ideas and techniques have been developed [Byoungho, 2003].
However, One of the most important sensor theories is “evanescent wave”, and it also plays a crucial role in biosensor and optical coupler. The basic condition to generate evanescent wave is using the total internal reflection, and the methods to fabricate total internal reflection have been studied in many structure like optical prism, waveguide, optical fiber, and etc. Because of the principle of the transmission of the lightwave, the application of evanescent wave in optical fiber is much suitable and developable.
1.2 Research Destinations and Motivations
Though the evanescent wave has existed due to the propagation of the lightwave by the total internal reflection in the optical fiber, the effective penetration depth of the evanescent wave is about one wavelength that is too small to be obtained. Therefore, there are many methods like etching, polishing, and pulling are researched to remove the cladding to let the energy of the evanescent wave throughout the cladding. In this thesis, the commercially available D-shaped optical fiber is applied to fabricate evanescent wave optical fiber sensor to solve the difficulty of side-polished optical fiber, and it is much easier to etch and control the residual cladding depth. Besides, though the characteristic of the evanescent wave like energy
abortion and phase shift is effectively applied to measure the surrounding refractive index change, its sensitivity is not good enough. Therefore, the electric field enhancing by surface plasmon resonance is used to increase the sensitivity of the evanescent wave sensor. In the figure 1.1 and figure 1.2, the development flow and basic structure for the application of surface plasmon resonance is shown. It can see that in order to reduce the sensor volume, increase the sensitivity, and more applicable, the optical fiber sensor is developed. In order to solve the difficulty of the fabrication in removing the cladding, the commercially available D-shaped optical fiber can be used, and not only the accuracy of cladding removing depth can be increased but the roughness on the interface can be decreased. Besides, the polarization unstable that exists in general single mode fiber SPR sensor can be solved by the polarization-maintaining characteristic of D-shaped optical fiber in this thesis. From the figure 1.2, the prism structure has been generally developed and much studied. In the figure 1.3, the developable relation between the evanescent wave, surface plasmon resonance, and optical fiber are shown.
Finally, the D-shaped optical fiber surface plasmon resonance sensor is chosen as the research destination.
1.3 Research Methods
The research methods in this thesis are shown in figure 1.4. In order to research the D-shaped optical fiber surface plasmon resonance sensor, SPR measurement characteristic and D-shaped optical fiber evanescent wave must be discussed first. Then, the metal cladding characteristic must be studied due
to the metal must be coating properly. If all the parameters are obtained, the D-shaped optical fiber SPR sensor can start to work. Therefore, there are four main subject can be developed, and them are shown in figure 1.5. First, the phenomenon of the surface plasmon resonance can be check and measured by using the I. optical-type surface plasmon resonance sensor, and the phenomenon of the evanescent wave refractometer can be check by using II.
D-shaped optical fiber evanescent wave refractometer. Secondly, the phenomenon of the surface plasmon resonance for D-shaped optical fiber can be check by III, and the fiber-type polarizer can be developed. If III can be obtained, the phase measurement of IV D-shaped optical fiber surface plasmon resonance sensor may be able to be detected.
1.4 Overview of Chapters
In this article, including this chapter, it is divided into nine chapters to discuss, and a brief introduction of other eight chapters is stated in below, respectively:
Chapter 2 History Review, it introduces the history review of the surface plasmon resonance, and the optical fiber grating.
Chapter 3 Theory Analysis, in this chapter, four basic theories that are evanescent wave, surface plasmon resonance, long period fiber grating, and Heterodyne Interferometry will be proved and analyzed.
Chapter 4 Optical-Type Surface Plasmon Resonance Sensor, the primary measurement methods and parameters of surface plasmon resonance will be
discussed. Besides, the relative sensitivity and resolution will be compared also. Finally, the basic Kretschmann coupler will be used to as the surface plasmon resonance sensor, and the relative error will be discussed.
Chapter 5 Fabrication of The D-Shaped Optical Fiber Surface Plasmon Resonance Sensor, the most important etching technology and how to measure the etching depth are discussed and the full fabrication process of the D-shaped optical fiber surface plasmon resonance sensor is described.
Chapter 6 D-Shaped Optical Fiber Evanescent Wave Refractometer, the characteristic of evanescent wave will be utilized to measure the phase shift due the surrounding refractive index various by D-shaped optical fiber.
Chapter 7 D-Shaped Optical Fiber Metal-Clad Polarizer, two methods that are thick metal cladding and thin metal cladding (SPR) are used to fabricate the fiber-type polarizer, and the extinction ratio and insertion loss will be measured.
Chapter 8 D-Shaped Optical Fiber Surface Plasmon Resonance Sensor, the phase measurement method with fixed resonance angle will be used to measure the change of the surrounding refractive index by fiber-type surface plasmon resonance sensor.
Chapter 9 Conclusion and Future Work, the conclusions of this thesis are discussed and the future work are suggested to solve the problem of fiber-type SPR sensors at present.
5
e Predicted y Ferrell And Stern 1968 Kretschmann、 Otto coupler (Prism)Optical Fiber Sensor
In order to : 1. Reduce Volume 2. Increase Sensitivity Cladding Remove
1. Etch 2. Polish 3. Pull D-Shaped Fiber In order to improve the method to remove the cladding
1967 Teng、Stern coupler (Grating) … … Figure 1.1 The Development of SPR Biosensor Conformation in ideal
6
SPR Biosensor D-Shaped Optical Fiber SPR Sensor
[Homola, et al., 1999]
[Kano, et al., 1994] [Homola, 2003] [Jorgenson et al., 1993]
[Johnstone et al., 1992] Figure 1.2 The Development of SPR Biosensor Conformation in example
Evanescent wave SPR
R ∆φ θsp R ∆φ ∆ω
Kretschmann ˇ[a] ˇ[b] ˇ[c] ˇ[d]
Prism
Otto (Structure is too hard to be used.)
Grating ˇ[e] ˇ[f]
Evanescent wave SPR
R ∆φ θsp R ∆φ ∆ω
Figure 1.3 Developed Refraction Index Sensors for Evanescent Wave and SPR. “ˇ” means it had been developed, and “☆” means it is the developing goal in this thesis. θspis SPR resonance angle, R is reflectivity, ∆ is φ Phase shift between p-wave and s-wave, and ∆ is wavelength shift. ω
*
[a]=[Matsubara et al., 1988]
[b]= [Nylander et al., 1982]
[c]=[Nelson et al., 1996]
[d]=[Zhang et al., 1988]
[e]=[Cullen et al., 1987]
[f]=[Vukusic, 1992]
SPR Measurement Characteristics Discussion
D-Shaped Optical Fiber Evanescent Wave Discussion
D-Shaped Optical Fiber SPR Sensor
D-Shaped Optical Fiber Metal Cladding Discussion
Figure 1.4 The research method and flow chart
9
Figure 1.5 The D-Shaped Optical Fiber Development Structure and Design Flow
I. Optical-Type Surface Plasmon Resonance Sensor II. D-Shaped Optical Fiber Evanescent Wave Refractometer IV. D-Shaped Optical Fiber Surface Plasmon Resonance III. D-Shaped Optical Fiber Metal-Clad Polarizer
He-Ne Laser Z
X Y
(,) 2
oQλϕ(45)oEO (90)oP(0)oPPhotodetecter Lock-In Amplifier Function GeneratorDC-Bias Fiber-CouplerD-Shaped Optical Fiber polarizerCondenser Lens He-Ne Laser Z
X Y
θ (,) 2
oQλϕ(45)oEO (90)oP (45)oP
Photodetect
Lock-In Amplifier Function GeneratorDC-Bias He-Ne Laser Z
X Y
(,) 2
oQλϕ(45)oEO (90)oP(45)oPPhotodetect
Lock-In Amplifier Function GeneratorDC-Bias Fiber-Coupler
D-Fiber Sensing Section
Condenser Lens
Chapter 2 The History Review
2.1 History Review of SPR Sensors
In the early 20th century, Wood [Wood, 1902] observed the phenomenon of anomalous diffraction on diffraction gratings due to the excitation of surface plasma waves and it was first described. In 1930, Rudberg found out the phenomenon of discontinuous energy loss due to an electron throughout the thin metal film. In 1951, Bohm and Pines expressed it as “Plasma Oscillations”. In 1958, Ferrell integrated above two phenomena and predicted
“Radiative Surface Plasmon Oscillations” by theory. In 1960 Steinmann confirmed “Radiative Surface Plasmon Oscillations” by experiments.
In 1962, Ferrell and Stern predicted “Surface Plasmon Wave” could be excited by plane polarization wave. Then just in 1967, Teng and Stern brought forward optical excitation of surface plasmon by the method of grating, and in 1968, Kretschmann [Kretschmann et al., 1968] and Otto [Otto, 1986] brought forward optical excitation of surface plasmon by the method of attenuated total reflection. Sine then, Surface Plasmons have been intensively studied and their major properties have been assessed [Raether, 1988], [Boardman, 1982].
Up to 1982, SPR started to popular in optical measurement after Nylander and Liedberg first used SPR for gas detection [Nylander et al., 1982], [Liedberg et al., 1983], [Liedberg et al., 1993]. Since then SPR sensing has been receiving continuously growing attention from scientific community.
Since 1992, integrated optical waveguide for SPR sensors [Lambeck, 1992] is pioneered, and then various groups have developed SPR-sensing devices using slab and channel single-mode integrated optical waveguides. In 1993, Jorgenson and Yee [Jorgenson et al., 1993] first used optical fibers for SPR sensing that was based on wavelength modulation in multimode optical fibers with partly removed cladding and metal film deposited symmetrically around the exposed section of the fiber core.
2.2 History Review of Evanescent Wave Fiber Optic Sensors
In [Lamb, 2002], Hirschfed first introduced the concept of the evanescent wave in 1965. In 1975, Kronick and Little put this concept into practice for immunoassays. At the same time 1974, Hesse used a fiber optic-based sensor for O2 and iodide. In 1984, the first evanescent wave fiber optic immunosensor was developed by Hirschfeld, and it was further optimized by Andrade et al. in 1985 and Sutherland et al. in 1984. In 1991, the ellipsometry was used to fabricate the optical fiber immunosensor [Kooyman et al, 1991].
The relative phase retardation due the surrounding refractive index various can be measured by detecting the polarization change. In the same year, simple interferometer for evanescent field refractive index sensing was designed [Rene et al, 1991], and the theoretic sensitivity was calculated. In 1993, the D-shaped optical fiber was made as polarimetric optical-fibre sensor for biochemical measurements [Heideman et al., 1993], and it also used the ellipsometry. In 1998, the D-shaped optical fiber was made as polarimetric D-fiber sensor for chemical applications by using fringes counting
[Muhammad, 1998].
2.3 History Review of Fiber-Type Polarizers
In 1980, the first single-mode fiber-optic polarizer was fabricated by using birefringent crystal [Bergh, 1980], and the characteristic of different refractive index for different polarization was used to separate the TE wave and TM wave. In order to overcome the large volume from birefringent crystal, the metal characteristic of the attenuation difference between the TE mode and TM mode was used to fabricate the metal-clad fiber-optic polarizer [Eickhoff, 1980], [Hosaka et al., 1982], [Hosaka et al.,1983]. By this way, the extinction ratio can be increased with decreasing the cladding thickness between the metal and the core. However, the insertion loss will become large when the deep etching depth is arrived. Therefore, it was hard to obtain high extinction ratio and low insertion loss at the same time by the thick metal-clad method.
In 1986, because the SPR had started to popular in optical measurement, the SPR phenomenon was used to fabricate the thin metal-clad fiber-optic polarizer [Feth, 1986]. By SPR method, the high extinction ratio and low insertion loss could be obtained at the same time with special matching condition. Since then, the fiber-type SPR polarizer had been widely studied.
In 1995, a Hi-Bi single-mode fiber experiencing a periodic perturbation was used to fabricate in-line fiber polarizer [Wang et al., 1995]. The periodic perturbation could be produced by a periodic microbending, acoustic waves, or a photorefractive index grating. Recently, resonant tunneling [Arun Kumar
et al, 1997], LPG in birefringent optical fibers [Ortega et al., 1997], and FBG with locally pressed [Torres et al. 2003] were individual brought forward, and all these designs used the characteristic of the fiber grating different mode coupling due the different core refractive index.
2.4 History Review of Single Mode Optical Fiber SPR Sensors
In 1993, Jorgenson and Yee [Jorgenson et al., 1993] first used optical fibers for SPR sensing that was based on wavelength modulation in multimode optical fibers with partly removed cladding and metal film deposited symmetrically around the exposed section of the fiber core. In 1995, Homola [Homola, 1995] used intensity modulation in single mode optical fibers which were side-polish and coated with a thin metal film. These SPR sensors suffered from modal noise and polarization instability. In order to overcome the polarization instability, two methods were developed. In the first method [Homola et al., 2001], light from a polychromatic light source passes a Lyott fiber optic depolarizer which produces unpolarized light which was then coupled into a fiber optic SPR sensing element, and then the transmitted light was analyzed with a spectrograph. The second method [Piliarik et al., 2003]
uses a polarization-maintaining fiber to control polarization of light in the fiber-optic SPR sensing element. However, the SPR measurement dynamic range of single mode optical fiber was restricted near the refractive index of cladding. Therefore, in 1998, a thin tantalum pentoxide overlayer was used to tune the operation range of the sensor to desired region of the refractive indices [Slavý´k et al., 1998]. In 2001, a tapered fiber was used to require
simper technology than the polishing technique, and the different surface plasma modes supported by the metal film can be excited by the fiber mode [Diez et al., 2001].
Chapter 3 Basic Theory
In this chapter, there will be four crucial basic theories that will be proved and analyzed, and these four basic theories will be the foundation stone to establish the follow experimental principles and conformations.
3.1 Evanescence Wave [趙凱華 et al.,1992]
As is known to everyone, if the angle of the incidence light is large than the critical angle of the total reflection, the energy of the incidence light will be reflected totally. However, this phenomenon doesn’t mean that there will not exist refraction light. In fact, there will exist a special phenomenon that is the total energy of the refraction light will decay exponentially with the distance far from the interface, and the prove are as follows:
First, the mathematic model of incidence wave, reflection wave, and refraction wave (like figure3.1) can be described by the electromagnetic wave of Maxwell’ theory as (3.1.1), (3.1.2), (3.1.3).
1 1 1 1
light frequency, reflective light frequency, and refractive light frequency respectively. Then researching for k , and due to the continuum of the 2 z boundary condition in the interface, k can be represented as follows: 2 z
2 2 2 2 2 2 2 2
and by introducing the geometry relation k can be rewritten as follows: 2 z
2 2 2 2 2 2
Then, there are some key points to be discussed:
if n2 > then n1 k2 z∈ R
if n2 < then n1 k may be I2 z ∈
By the Snell’s Law, we can find the critical angle condition:
2 1
sin c n
n = θ (3.1.6)
Then k2 z can be rewritten as follows in critical angle condition:
2 2
Now, it can clear to see that the magnitude of the refraction wave will decay exponentially with the distance far from the interface, and it is called
“Evanescent Wave”.
3.2 Surface Plasmon Resonance [Raether, H., 1988]
In 1962, Ferrell and Stern predicted “Surface Plasmon Wave” could be excited by plane wave, and it exists a resonance condition [Raether, H., 1988]:
“Resonance” means that if the incidence light satisfies some conditions, the total energy will couple into the plasmon on the metal interface, and the evidence can be derived as follows:
First, for p-wave, the mathematic model of incidence wave and refraction wave (like Figure3.2) can be described by the electromagnetic wave of Maxwell’ theory as (3.2.2), (3.2.3) [Raether, H., 1988].
For z > 0:
2 = 02exp (i k x k zx2 + z2 −ωt) (= Ex2,0,Ez2)exp (i k x k zx2 + z2 −ωt) E E
(3.2.3)
By fulfilling Maxwell’s equation, for free space and two infinite plane:
1 i and the continuum relation on the boundary:
1 2
;
Then now, we only care for the real part because the imaginary part just means a damping term or internal absorption. So, k′ describes the x phenomenon (wave vector) of the surface plasmon wave (SPW) in the x-direction. Therefore, the resonance condition of SPW can be written as:
2 1
Equation (3.2.17) is called the resonance condition of surface plasmon wave.
Comparing Equation (3.2.11) and (3.2.17), if special case ε2 < , 0 ε1 > , and 0
2 1
|ε |> is chosen (EX: air-metal interface), the inequation ε kx c
>ω and
complex wave number kzi∈ can be obtained. Complex wave number in I
the z-direction means nonradiative surface plasmon wave, and the energy will decay exponentially with the distance far from the interface. However, if the s-wave is proved by the same way, the result of SPR cannot be obtained. In other words, the surface plasmon wave only can be excited by p-wave.
Besides, the surface plasmon wave cannot be excited directly by the lightwave from the analysis of [Raether, H., 1988]. In Figure3.3, the dispersion relation for the lightwave in air and for the metal are shown, and these two cures have no intersection. It means that the lightwave cannot excite the surface plasmon wave directly. In order to overcome this problem, the couplers have been brought forward. At present, primary three coupling structures that have been used to enhance the momentum of the lightwave are the prism coupler, (like Kretschmann-Structure and Otto-Structure), grating coupler, and waveguide coupler that can be shown in Figure 1.2. If the coupler is used and the dispersion relation for the lightwave and the metal has the intersection, it can be shown in Figure 3.4. In other words, the surface plasmon wave may have the chance to be excited by the coupling light wave, and another way to express this phenomenon is that the surface plasmon wave can be exited by the evanescent wave from the prism coupler.
3.3 Three Layers Optical Waveguide
In this section, the three layers optical waveguide theory is used to analyze and simulate the electric field of surface plasmon wave. Besides, there are two main measuring methods will be discussed. One is the resonance trough shift due to the overlayer index. Another is the phase difference that is between the
P-wave and S-wave shift due to the overlayer index. In Figure 3.5, the Kretschmann three layers optical waveguide structure is shown. By the Fresnel’s Formula, we can analyze what parameters may affect the performance of SPR (surface plasmon resonance), and the Fresnel’s Formula are as follows:[ Yu et al.,2001]
12 13 curve, it can know that only the P-wave can excite the surface plasmon resonance. By the above Formula we can understand that the parameters:
metal film thickness(d)、the frequency of the incidence light(ω )、the incidence angle(θ )、and the permittivity (ε1、ε2、ε3) will affect the reflection coefficient(Rp). However, the frequencyω has been decided by using He-Ne laser (the wave length is 633 nm), and the material of prism in this experiment is BK7 (its refractive index can be known at the Table 3.1 [Lambda]), and for the He-Ne laser 633 nm the permittivity of the Au film is
2 11.6798 1.1722i
ε = − + . Now, if we take the air as the measured sample, the reflection coefficient (Rp) will only depend on the parameters θ and d.
Therefore, if we can find the appropriate parameters θ and d, we can get the maximum sensitivity of the measurement system, and the most appropriate parameters θ and d can be found by using MATLAB to calculate and the result is shown in Figure 3.7. By the Figure 3.7 and Figure 3.8 we can known that when d=45~54(nm), system has the maximum resonance trough at
430
θ = . In other words, it has the maximum sensitivity at θ =430, and with this method we can find the optimization coating thickness on the Kretschmann-Structure. In Figure 3.9, when the overlayer index n changes, 3 the resonance trough will shift. Therefore, the overlayer index change can be detected by measuring the resonance trough shift.
However, it knows that the reflection coefficient r , p r can be defined s as (3.2.5)
From equation (3.21)~(3.24), it can get (3.2.7)
( , )
p s
r r n
r = θ , (3.2.7)
and if the overlayer refractive index change, equation (3.2.6) can be written as
Comparing equation (3.2.7) and (3.2.8), the phase shift ∆ between the φ P-wave and S-wave is the function of the incident angle θ and overlayer refractive index n : 3
( )∆ = ∆ϕ ϕ θ 、∆ = ∆ϕ ϕ( )n3 , (3.2.9)
In Figure 3.10, the theoretic curve, ∆ = ∆ϕ ϕ( )n3 , of phase shift versus overlayer refractive index change with fixed-incidence angle that is near the resonance angle of water is shown.
In Figure 3.10, the theoretic curve, ∆ = ∆ϕ ϕ( )n3 , of phase shift versus overlayer refractive index change with fixed-incidence angle that is near the resonance angle of water is shown.