SCHOOL OF ELECTICAL AND ELECTRONIC ENGINGEERING. Automated Reflectometry System. By Sufyan Samara. Supervisor Professor Mike Somekh

Transcription

1 SCHOOL OF ELECTICAL AND ELECTRONIC ENGINGEERING Individual Project Automated Reflectometry System By Sufyan Samara Supervisor Professor Mike Somekh September 2005

2 Intentionally left as blank page 2

3 Table of Contents List of Figures... 5 Preface... 6 Abstract... 8 Project Aim and Objectives... 8 Project specification... 8 Applications and benefits of the SPR... 9 History of SPR... 9 Introduction (SPR theory and configurations) Definition of SPP/SPR How and when a surface plasmon resonance occurs, what do we need? SPR Dependencies Configuration to Excite SPR Commercial Instruments used to sense the SPR Mechanical Design Basic Design and Operation How to build this design? Building the Actual real system Problem in the System Angle-Control Optics and Optical Components Laser diode Diaphragm Polarizer CCD linear Camera Filter Prism Lenses Solving the problem of the limited range in incident angles ( ) Resolution and Beam shape Software and results Linear CCD camera program (Spectra) General SPR Sensor (GSPRS) program

4 The limitation caused by the CCD camera Conclusion and ideas to improve the design Appendix A: Mechanical design first idea, and why it is not feasible?...52 Appendix B: Mechanical components used in the design Appendix C: The GSPRS Code Appendix D: What are the contents of the Included CD References...79 Bibliography

5 List of Figures Fig. 1: surface plasmon polariton (SPP) Fig. 2: k-vector of the surface plasmon Fig. 3: Surface Plasmon Resonance (or SPR) occurs when Fig. 4: deferent angles of incident light and their phase diagram Fig. 5: SPR comparison between aluminium and silver Fig. 6: The metal film thin and its effect on SPR Fig. 7: grating coupled systems Fig. 8: the grating k-vector kg adds to the incident light k-vector Fig. 9: Optical waveguide system Fig. 10: Otto arrangement Fig. 11: Otto enhanced configuration Fig. 12: Kretchmann configuration Fig. 13: the possible range of angles for SPR Fig. 14: The rhombus shape like mechanical design Fig. 15: the basic operation for rhombus shape like mechanical design Fig. 16: a 3D module showing the first idea of the system Fig. 17: A 3D module of the system before it was build Fig. 18: A 3D module of the complete mechanical system Fig. 19: real pictures of the complete system Fig. 20: Angle-control (threaded rod) "real picture" Fig. 21: Angle-control (path router) "real picture" Fig. 22: The laser beam from the laser diode Fig. 23: Right Angle Prism and the incident angles Fig. 24: Right angle and half sphere prisms Fig. 25: The optical lenses system used in this system Fig. 26: Beam expander Fig. 27: The linear CCD camera program from Thorlabs (Spectra) Fig. 28: GSPRS program Fig. 29: GSPRS Imports Spectra saved files Fig. 30: The GSPRS MAF Fig. 31: The GSPRS Polynomial fitting Fig. 32: The SPR as it reflects of the CCD camera

6 Preface This Project thesis submitted in part fulfillment of the requirements for the degree of Master of Science in Electronic Communications and Computer Engineering, University of Nottingham, September This thesis is about building a general surface plasmon resonance sensor passed on TIR configuration and it is arranged to include the following five sections; Introduction: this section provides the introduction for surface plasmon resonance theory; it begins by defining this phenomenon then continues to exploit the basic configurations and setups needed in order to excite the SPR. Finally it introduces some commercial instruments used to measure the SPR. Mechanical Design: this section provides an overview of the steps taken to build the mechanical design. It shows the pictures, modules, and the ways to operate the system. Optics and Optical Components: This section the information about the optical components used in the system, such as the lenses system, and the reasons for using these components. Software: This section gives the information about the software needed for reading the data from the CCD camera and the software developed to process these data. Conclusion and ideas to improve the system: this section exploits the weak points of the system and gives some ideas to improve the system. 6

7 Acknowledgment I thank GOD who gives me the power and knowledge to do this project. I thank my parents and my family for their support. All the thanks and appreciations to those who help me to complete this project, with special thanks to Professor Mike Somekh, Dr. Shugang liu, and Dr. Mark Pitter. 7

8 Automated Reflectometry System Abstract Surface plasmon resonance (SPR) is an important phenomenon used for building sensors especially in the Biological fields. Many sensors based on this phenomenon have been designed for specific purposes, however, in this report a general SPR sensor based on the, so-called, TIR configuration will be designed which allow us to scan the incident angles to monitor with an angular precision to 1/100. Project Aim and Objectives Building a SPR system based on TIR configuration with sufficient angular precision to monitor the small changes in binding reactions. This involves finding/building the proper mechanical/optical design, using the suitable optical components (CCD camera, laser, lenses), and building the necessary software to process data. Project specification Design a general SPR sensor with the following specification: Mechanically stable. Relatively small, simple and portable (50 cm height x 40 cm width x 20 cm depth). Can be assembled/disassembled easily. Can be modified (add/remove/mount the optical components) easily using Thorlabs cages mount and rods. Can find the SPR with an angular precision of 1/100 of a degree. Can scan a range of incident angles to find the SPR from 40 to 80 degrees. 8

9 Can be operated manually and automatically. Applications and benefits of the SPR Surface plasmons are known for a long time, but the applications that use this phenomenon are new to develop and many research still needed. Surface waves and surface plasmons are becoming incredibly important for sensor applications; they represent a sensitive and cost-effective means for examining chemical and biological reactions. Surface wave sensors, for example, are becoming very important for antibody/antigen reactions, and biological sensors drug discovery. History of SPR It was During the investigation of the energy losses by fast electrons passing through thin metal foils In the beginning of the 20 th century, when Pines and Bohm[1] pointed that these energy losses are due to plasma oscillations or plasmons in the sea of conduction electrons. This sea of conduction electrons refer to bulk materials, which have properties, that allow the assumption of the conduction electrons acting as a free electron gas such as Aluminium. This can be characterized by the dielectric function ω ρ ε ( k, ω) = 1, ω( ω + i / τ ) 2 Where k is the wave number vector, ω is the angular frequency, τ for the dissipative effects in the system, ωρ is the plasmons frequency and its given by 2 n. e ω ρ =, ε 0m e Where n is the free electron density of the material, e is the electron charge, and me is the effective mass of an electron. Gabor attempted to develop a theory to describe the energy loss characteristics of electrons passing through finite metal crystals. As Ritchie [2] points out, Gabor 9

10 mistakenly assumed that the electric field is always zero at the surface of the crystal, which led to unrealistic results. Ritchie applied more realistic boundary conditions and calculated the electron interaction probability in extremely thin films. His work resulted in the prediction of surface plasmons and consequently pioneered a whole new field of interest in surface physics. Introduction (SPR theory and configurations) Definition of SPP/SPR Surface plasmon polariton (SPP) is an electromagnetic excitation that propagates in a wave like fashion along the planar interface between a metal and a dielectric medium, often vacuum, and whose amplitude decays exponentially (evanescent wave) with increasing distance into each medium from the interface, figure 1, (normally it decays over a distance of about one light wavelength, and for it to be useful, a measurement should be done within 300nm [3] from the sensor surface). Fig. 1: surface plasmon polariton (SPP) The wavelength of the evanescent wave is the same as of the incident light and the energy of the evanescent wave is dissipated by heat. 10

11 Thus, a SPP is a surface electromagnetic wave, whose electromagnetic field is confined to the near vicinity of the dielectric metal interface, and the plasmon is the particle name of the electron density waves. How and when a surface plasmon resonance occurs, what do we need? According to Maxwell s theory, surface plasmons can propagate along a metallic surface and have a spectrum of angular frequencies ω related to the wave-vector (k) by a dispersion relation [4] ε ε 1 2 k sp =, ε1 + ε 2 ' " 2 ε 2 + iε 2 Whereε =, and ε1 are the dielectric constant of the metal and of the medium in contact with it, respectively. The propagation constant, β, of the surface plasma wave propagating at the interface between a semi-infinite dielectric and metal is given by the following expression: ε mε β = k, ε + ε m Where k denotes the free space wave number, ε m the dielectric constant of the metal (ε m = ε mr + i ε mi, real and imaginary components, respectively) and ε is the dielectric constant of the dielectric (the refractive index of the dielectric n= ε ). Wave vector k1 of light at frequency ω traveling through the medium ε1 is described by ω k 1 = ε 1, c Where c is the speed of light in vacuum. 11

12 Fig. 2: k-vector of the surface plasmon Since the SP's dispersion relation (curve SP in Fig.2) never intersects the dispersion relation of light in air kc, they cannot be excited directly by a freely propagating beam of light incident upon the metal surface. However, it is possible to "turn down" the light line in Fig.1 to the point where both lines cross each other [5]. In other words the excitation or resonance (surface plasmon resonance SPR) occurs when the momentum (k-vector) of the incoming light is equal to the k- vector of the plasmons (k-vector of resonance). At that point (the resonance) the photon can interact with the free electron at the metal surface, the incident light photons are absorbed and converted into surface plasmons. The basic setup (based on kretchmann configuration) to excite the surface plasmon is shown in fig.3. 12

13 Fig. 3: Surface Plasmon Resonance (or SPR) occurs when a thin conducting film is placed at the interface between the two optical media. At a specific incident angle, greater than the TIR angle, the surface plasmons (oscillating electrons at the edges of the metal) in the conducting film resonantly couple with the light because their frequencies match. Since energy is absorbed in this resonance, the reflected intensity, I, shows a drop at the angle where SPR is occurs The k-vector of the photons and plasmons can be described by a vector function, figure 4, with both magnitude and direction. The relative magnitude of the component changes when the angle or the wavelength of the incident light changes. However, plasmons are confined to the plane of the metal film, so for SPR to be excited, the incident light k-vector component parallel to the surface needs match the SPR k-vector. Thus, the energy, polarity, and the angle of incident light must be right in order to form a surface plasmon resonance [6]. The following figures show a simple representation of the incident light at deferent angles and their corresponding phase diagrams. 13

14 Fig. 4: deferent angles of incident light and their phase diagram. In Fig.4 (A) the angle of the incident light is not the right angle to excite the SPR, as the x-component of the incident light does not match the SPR vector. However, as the angle change in part (B) the incident light will excite the SPR as its x-component matches the SPR vector. SPR is a very good sensor, because it is very sensitive to any chemical change in the metal surface, this is due to the change in k-vector of the plasmons when the composition of the medium changes, so the angle of incident light at which the resonance occurs changes with large amount due to the fact that the field at the interface between the metal (gold) and the upper dielectric is greatly enhanced, and this can be measured very precisely. From the above, we can conclude that SPR depends on three things: SPR Dependencies The properties of the metal film. The wavelength of the incident light. The refractive index of the media on either sides of the metal film, which is sensitive to temperature. The metal film: 14

15 The metal must have conduction band electrons capable of resonating with the incoming light at a suitable wavelength [7, 8] (The visible and near-infrared parts of the spectrum are particularly convenient because optical components and high performance detectors appropriate for this region are readily available). Metals that satisfy to this condition are silver, gold, copper, aluminium, sodium and indium. In addition, the metal on the sensor surface must be free of oxides, sulphides and not react to other molecules on exposure to the atmosphere or liquid. Of the metals, indium is too expensive, sodium too reactive, copper and aluminum are too broad in their SPR response and silver is too susceptible to oxidation. This leaves gold as the most practical metal. Gold is very resistant to oxidation and other atmospheric contaminants but is compatible with a lot of chemical modification systems. An example for the SPR deference between metals is shown in figure 5; silver and aluminium are chosen for this example. Fig. 5: SPR comparison between aluminium and silver The thickness of the metal layer is of great importance, above the optimum value the dip in the reflective light becomes shallow, and below it the dip becomes broader (for the gold should be nearly 50 nm). This can be shown using a simulation program [6] as in the following figure: 15

16 Fig. 6: The metal film thin and its effect on SPR The light source should be monochromatic and p-polarized (polarized in the plane of the surface) to obtain a sharp dip. All the light, which is not p-polarized, will not contribute to the SPR and will increase the background intensity of the reflected light [8]. Configuration to Excite SPR So far we introduced one method to excite the SPR which is the TIR (total internal reflection) using either prism or half sphere (recommended for reasons yet to be discussed), which is the method to be used in this design. However, there is other ways to excite the SPR, in all them the basic idea is the same, for each configuration we introduce something to improve the k-vector incident light or add another k-vector to it in order to match the k-vector of the surface plasmon resonance so the excitation occurs, see figure 4. These configurations are as follows: 16

17 Grating coupled systems This configuration is shown in figure 7. Fig. 7: grating coupled systems This configuration uses the grating to add a k-vector to the incident light k-vector to match the SPR k-vector; this is shown in figure 8. Fig. 8: the grating k-vector kg adds to the incident light k-vector kp to match the SPR k-vector ks. For the grating coupled systems, the sinusoidal grating is the optimal grating [10]. The period (top-to-top) and amplitude (topto-trough) of the grating will determine the wavelength of resonance; the overall uniformity of the sinusoidal geometry determines the strength of the SPR signal. Therefore, sensor-tosensor reproducibility depends upon the fidelity of grating production. This configuration is normally used for lower production costs and more latitude in selection of construction materials. 17

18 Optical waveguide systems This configuration is shown in figure 9. Fig. 9: Optical waveguide system. The optical waveguide systems [4] have some attractive features like the simple way to control the optical path (efficient control of properties of the light, suppression of the effect of stray light, etc.), small size and ruggedness. By varying the angle of incidence of the light, a light wave is guided by the waveguide. On entering the region with a grating (2400 lines/mm) and a thin metal overlay, it evanescently penetrates through the metal layer. At the end of the wave guide the out coming light is detected by photodiodes. An algorithm is used to model the adsorbed material linearly to surface concentrations. Prism coupled attenuated total reflection system (TIR) o Otto arrangement This works as shown in figure 3, however, There is a distance between the metal and the TIR surface, which is filled with a lower refractive index medium. The evanescent waves from the TIR are used to excite the evanescent waves in the Metal in enhanced way. This is useful in study of SPR in solid phase media, but it is less useful for applications with solution because the distance between the metal and the TIR surface reduces the SPR efficiency. This configuration is shown in figure

19 Fig. 10: Otto arrangement Another configuration looks like the Otto arrangement but uses a special layer to enhance TIR is shown in figure 11. Fig. 11: Otto enhanced configuration. The coupling of the TIR light to plasmons is done by the resonant mirror (RM) principle. On a prism in which the light is in TIR a small layer of silica (~ 1 mm) is deposited. On top of the silica, there is a titania layer. The silica layer is thin enough to allow the evanescent field generated in the prism to couple into the high refractive index titania. This so called frustrated total internal reflection (FTIR) allows the titania layer to function as an optical waveguide. Repeated total internal reflection of the guided mode within the wave guiding titania layer results in the production of an evanescent field at the titania-adlayer interface. The exact angle of incident light at which the resonance between the wave guided mode and the evanescently coupled light occurs is directly dependent upon the refractive index of the surface adlayer. Unlike the SPR device however, there is virtually no loss of the reflected light intensity associated with the resonance condition. Rather, 19

20 resonance is accompanied via a 2 pi change in phase of the reflected light, which is recorded interferometically [10]. o Kretchmann configuration The metal layer is directly on top of the TIR surface enabling a more efficient plasmon generation. This configuration is useful for solution. This configuration is shown in figure 12. Fig. 12: Kretchmann configuration. Commercial Instruments used to sense the SPR BIACORE: This system is fixed angle equipped with a continuous flow system in which four channels are coupled in series. It has an automatic sample needle to deliver buffer and sample to the sensor chip surface. The continuous flow ensures that no changes in analyte concentration occur during the measurement [11]. 20

21 Spreeta Evaluation Module: From Texas Instruments this is one of the smallest surface plasmon resonance biosensors and all the optics are integrated in the Spreeta sensor. The evaluation modules are designed to allow users to experiment with surface plasmon resonance technology in a variety of applications. The flow cell is made of Teflon with supplied with gaskets for sealing [12]. SPRimager: From GWC instruments is a somewhat different instrument in the way that is not giving a graph as result but a picture of the sensor chip surface. A collimated, polychromatic light source passes through a polarizer and impinges on a prism/thin gold film sample assembly at an incident angle above the critical angle. The reflected light then passes through a narrowband interference filter and is detected with an inexpensive CCD-camera. Real-time video images are acquired as the binding reaction occurs. The prism/thin gold film sample assembly is mounted on a rotation stage so that the incident angle can be manipulated; this allows for ease in maximizing the contrast of the SPR image. A liquid flow cell is attached to the assembly permitting images to be collected in-situ. Simple removal of this cell attachment allows for the collection of ex-situ images [13]. 21

22 Other instruments exist such as IBIS Biosensor, IAsys system, Plasmonic system. 22

23 Mechanical Design Basic Design and Operation To build a general SPR sensor we began exploiting the possibilities and the configurations that may needed for that purpose. We were searching for a relatively simple to build system yet rigid, flexible, and provide us with all the specification we are aiming for. We needed a system that can operate in vertical and horizontal positions, relatively portable (weight around 10 kg, size (height 50 cm x width 40 cm x depth 20 cm), can be operated manually and automatically, and can be modified and extended easily. We found that we can use the TIR configuration (krechmann like configuration); because it is easy to mount the prism, easy to put samples under test on that configuration, easy to scan for SPR, and can be modified and replaced (samples and prism) without affecting the whole design. After that we start thinking of a method we can use to scan the TIR configuration for the entire range of angles to find the SPR possible angle. Before that we needed to know what is that range. To know that we used a simulation program [6] that gave us the results shown in figure 13. This simulation is done by varying the upper dielectric which is the main variable may exist in this system. We changed the refractive index between the two possible upper dielectrics this system may be used in (air and solution); this shows that we need a system that enables us to scan the incident and reflected angles between 40 and 80 degrees. Also we need a mechanical system that enables us to mount the optical components easily. We started to exploit each idea that serves the need, to find the suitability and the possibility of each one. We find that the best idea [6] is to design and build a rhombus shape like mechanical arms and mount all the optics on it. The rhombus shape gives us the flexibility, accuracy, and ease of operation and controlling for the whole system. It allows us to change the angles by only changing the position of one vertex linearly, without much concern on adjusting the angle of each arm in order to match the incident angle with the reflected angle. 23

24 Fig. 13: the possible range of angles for SPR. The basic idea of the rhombus shape like mechanical design is shown in figure 14 with each arm written its name on it. 24

25 Fig. 14: The rhombus shape like mechanical design In this design, four arms are needed with the same lengths (L) and they are named as follows: Upper left arm: CCD camera arm for detecting the reflected light. Upper right arm: laser arm. The two lower arms: Control arms. Also for vertices: Upper vertex: Prism. Lower vertex: angle-control. Left and right vertices: joints. By moving the Angle-control up and down the angle of the arms to the normal (and so the incident and the reflected angles) will change equally, allowing the possibility to scan the prism (sample) for all the angles to find the SPR angle. This is shown clearly in figure

26 Fig. 15: the basic operation for rhombus shape like mechanical design. How to build this design? Two options were introduced, the first one is to design and build all the mechanical system in the workshop. This was not convenient for different reasons like the time lacking, the limitation of replacement parts, the errors may occur in the manufacturing process which require remanufacturing and that consumes a lot of time (we have limited time to design and build this project). The second option was to search and buy general assembly parts that can be used in this system and use the workshop for making the parts we design in order to connect the components we buy to match our needs. This was the option we follow because the parts needed to be designed and build in the workshop for this option were relatively simpler than the first option and consumes less time than building the whole design. 26

27 We begin to search for the parts we want to buy, in the mean time we build a module of the system we are planning to build using a 3D modulation program; this is shown in figure 16. Fig. 16: a 3D module showing the first idea of the system. This design was not suitable because it will be large in size and hard to mount optical parts on it, for more information see appendix A. After some search we found it that it is convenient to buy our components from Thorlabs because they provide a large variety of general purpose components with suitable price/quality match. We found that we need the following items to build the design: Swivel Mount for 60mm Cage System. 60mm/30mm Cage Plate Adapter. 30mm Cage Plate. Cage Assembly Rods. 90 Degree "T" Extension. All the information, dimensions, and diagrams about these parts are included in Appendix B. 27

28 Using the above components we build another 3D module for the system after modification introduced by using the "90 Degree "T" Extension" before we buy the components to see the feasibility of these components. We used the real dimensions of each component and we got the system shown in figure 17. Fig. 17: A 3D module of the system before it was build. From the 3D module we see that the system was convenient, feasible and easy to assemble and mount optical components on it, also the problem of the system being large and the arm length/angle dependency (see appendix) was removed by using the 90 Degree "T" Extension component. Now we need to design and build the components that allow us to modify this design to match our aim (rhombus like shape design). This implies the need for the following components. 28

29 Vertical to horizontal Rod connector. Connector position adjuster. 90 Degree "T" Extension adjuster. Threaded Rod. Vertical to horizontal Threaded Rod connector. Threaded Rod Holder. Path router. All the information, dimensions, and diagrams about these parts are included in the Appendix B. Inserting the precious components into the 3D module we get the complete mechanical system shown in figure 18. Fig. 18: A 3D module of the complete mechanical system. Figure 18 shows the complete system with incident angle of 80 degrees. 29

30 From the 3D module the design dimensions were estimated to be (LxWxH)=(50x40x20) mm. Building the Actual system After we show (from the 3D module) that the system is feasible to be build, we began buying, designing and building the necessary components. After assembling these components the actual system appears as shown in figure 19. Fig. 19: real pictures of the complete system. 30

31 Problem in the System After building the system we found that the system can operate over a range of incident angles from 45 to 85 degrees (because of the Swivel Mount for 60mm Cage System) not from 40 to 80 degrees. This problem can be solved using optical components as described in Optics and Optical components section. Angle-Control There is two ways to control the angle in this system; one is by using the threaded rod with a motor (automated). The other one is by using the path route method, which can be controlled manually or automatically. However, because of the CCD camera, this system can not be fully automated; this will be discussed in the Optics and Optical components section, also in the software section. The first method in changing angle is shown in figure 20; it is made from the threaded rod, Vertical to horizontal Threaded Rod connector, and Threaded Rod Holder. Fig. 20: Angle-control (threaded rod) "real picture". 31

32 As the threaded rod rotates the Vertical to horizontal Threaded Rod connector will move across the rod, which causes the changing in angle. By connecting a motor to the end of the rod (see figure 18) we can control the angle automatically using micro controller such as microchip PIC16F84 or using the PC computer. The second method is done using the path router, Cage Assembly Rods, 90 Degree "T" Extension, and two vertical to horizontal Rod connectors as shown in figure 21. Fig. 21: Angle-control (path router) "real picture". The path router can be labeled to indicate the exact angle position; this can be done easily in three steps as follows. 1. Pull the arm until the incident angle is 45 degrees and mark the path router on that angle. 2. Push the arm until the incident angle is 85 degrees and mark the path router on that angle. 3. Then we divide the distance in between the two marks equally to cover the range of degrees we want. 32

33 Optics and Optical Components The optical components used in this system are: Laser diode. Polarizer. Diaphragm. Lenses. Prism. Filter. CCD linear camera. Laser diode The laser we are using is a 1mW, 635nm wavelength with 4 mm (round) diameter from Thorlabs. The laser should have power fluctuation as small as possible because any fluctuations in the power will appear as noise while reading the data. Diaphragm This is needed to eliminate the unnecessary beam light incident from the laser diode. This is needed because the laser beam that comes from the laser diode is not all parallel; it contains unwanted light around the parallel beam which is shown in figure 22. Fig. 22: The laser beam from the laser diode. We are only interested in the middle circle. 33

34 Polarizer This is needed to make the entire incident light p-polarized, which is the useful component of light that will contribute in SPR excitation; this is also needed so the movement in the laser diode will not affect the polarization of the beam. CCD linear Camera We are using a linear CCD camera from Thorlabs. It uses the parallel port as an interface to the PC and it has separate software to pull the data from it (this is one of the limitations for this CCD camera, more information in the software section). It has the following specifications: This CCD camera uses the ILX526A rectangular reduction type CCD linear image sensor with a built-in timing generator and clock-drivers. This CCD camera has 3000 effective pixels, with a pixel size of 7µm (pitch) x 200µm, and a maximum operation frequency of 1MHz. Filter We used a filter in front of the CCD camera for two reasons. Minimize the noise from the other light sources. Allow the device to operate in the daylight (no need to operate in dark environment). The filter we used is a Laser Line Filter (FL635-10), with CWL=635nm ±2, FWHM=10nm ±2 from Thorlabs. Prism Using TIR configuration we may use either the right angle prism or half sphere prism. Although the right angle prism was used in this design (because of the availability of that kind of prism), the half sphere is recommended. This is because it does not introduce any changes to the range of incident angles (because all the incident angles will be perpendicular to the surface of the half 34

35 sphere prism). However, in the case of right angle prism only one ray of the incident angles will be perpendicular to the surface, and the others will hit the surface with angle less than 90 degree causing the ray angle to be reduced as it change from one medium(air) to another (prism medium) as shown in figure 23 Fig. 23: Right Angle Prism and the incident angles. In figure 23, L1 is the beam radius; L2 is the focal length of the cylindrical lens, L3 is the distance from the right angle prism surface to the actual focal point, θ1 is the outer ray incident angle to the right angle prism, and θ2 is the angle after the ray have be refracted. For this design: L1=2 mm, L2=50 mm. 1 L1 θ 1 = tan ( ) = L 2 Assuming the refractive index of the prism n0=1.5; 1 sinθ1 θ 2 = sin ( ) = , n 0 Hence, the total range of incident angles in one shot is 2x1.527=

36 This means that the right angle prism reduces the range of incident angles by 33%, which is a large amount, especially if we want to use a small beam (see Conclusion and ideas to improve the system). This will limit moving the Anglecontrol to smaller steps which may slow the process and increase the errors. In the case of the half sphere prism all the incident light will enter the prism perpendicularly, and so the range of angles will stays the same. Figure 24 shows both systems. Fig. 24: Right angle and half sphere prisms. Lenses In this design we are using a 4 mm round laser beam to excite the SPR on a prism, and a CCD linear camera with an effective sensor area of (200µm high x 21000µm width). This implies the need of a lenses system to concentrate the incident light on the SPR surface and to image the reflected light to cover all the effective area of the CCD camera sensor. After some research and testing we found that the simplest optical system was in using three lenses, one cylindrical lens with a focal length of 50mm on the Laser 36

37 arm, one negative spherical lens with focal length of 50 mm on the CCD camera arm, and one spherical lens with focal length of 160 mm on the CCD camera arm. The operation of this system is shown in figure 25. Fig. 25: The optical lenses system used in this system. The cylindrical lens is useful to have more than one degree of incident light in one position; because the laser beam is parallel all the light come to the prism in the same angle, however, with the cylindrical lens, the light enters the prism with a range of angles depends on the beam diameter and cylindrical lens focal length (see the prism section). The negative spherical lens is good to expand the 4 mm to cover the whole range of the CCD linear camera sensor effective area, and the spherical lens for 37

38 making the light parallel. The following figures show the light beam in every stage. The beam coming from the laser is shown in figure 22. The laser beam after the diaphragm, cylindrical lens, and before entering the prism The laser beam on the SPR prism surface 38

39 The laser beam reflected from the prism, and before entering the CCD camera arm. The laser beam on the CCD linear camera sensor The laser beam reflected from the prism with SPR, and before entering the CCD camera arm. 39

40 The laser beam on the CCD camera sensor with SPR The laser beam on the CCD camera sensor with SPR, this picture was taken with using a camera after the white balance was reduced to the minimum, this indicate that we may improve the system if another filter used to minimize the laser beam intensity before coming to the CCD camera. Solving the problem of the limited range in incident angles ( ) As discussed in the prism and lenses system sections, the incident light hits the prism with a range of incident angles nearly equal to three degrees. So at an incident angle of 45 degrees, the range of angles between 43.5 degrees and 46.5 degrees is covered. However we need the incident angles to be covered from 40 and above, this can be achieved either by modifying our system to begin 40

41 from 40 degrees (difficult), or to expand the beam to cover a larger range of angles (recommended). As we mentioned before, the range of angles depends on two things, the focal length of cylindrical lens (the minimum the better), and the laser beam diameter. So by changing one of them we can control the range of angles, hence, the easiest thing to do is to expand the laser beam diameter, this is possible using the laser beam expander system shown in figure 26. Fig. 26: Beam expander. With simple calculation we find that f f 1 2 = d 2 ; d1 Where D1 is the original beam diameter, D2 the final beam diameter (after expanding), f1 the focal length of the first spherical lens, f2 the focal length of the second spherical lens, d1 the radius of the original beam (D1/2), and d2 is the radius of the final beam (after expanding (D2/2)). Now assuming d1= 2mm, f1=12, and we want the final beam radius d2 to be 6mm, then using the last equation we find that we need a second lens of focal length f2=36mm. 41

42 Resolution and Beam shape As we mentioned before, we are using a linear CCD camera sensor with 3000 pixels, each with size of 7µm (pitch) x 200µm. In other words, we are using a sensor with a width of 21 mm and 200µm height. This sensor is used to detect a range of angles of three degrees, which make the resolution nearly 3 = deg/pix So we need to image a beam of rectangular shape on the sensor. However, this shape (the rectangular) will not work because as shown in the above pictures the SPR image is not a nice vertical line, which makes it very hard to detect the SPR if we want to use an optical component (cylindrical lens) to compress the beam to a rectangular shape. Also with an ellipse shape it is easier to adjust the beam to cover the entire CCD camera sensor (alignment of the beam). Software and results We are using two software; one for getting the data from the linear CCD camera (Spectra) which is provided by Thorlabs, and another one developed by me "Sufyan Samara" using matlab GUI (graphical user interface) to process the data we got from the first program. Linear CCD camera program (Spectra) This software is provided by Thorlabs. It connects to the parallel port on which the CCD camera is connected. It is used to pull the data that comes from the CCD camera at real time and maximum frequency rate of 1MHz. This program is shown in figure

43 (A) (B) 43

44 (C) (D) Fig. 27: The linear CCD camera program from Thorlabs (Spectra). 44

45 In figure 27 (A) the laser is off, this shows the reference level (no light), In (B) it shows the system output with the laser ON but at angle where is no SPR, (C) and (D) shows the SPR but with different "Int. Time" values (optimal value depends on the measurement but it is near 0.288). For more information about using this program, see the manual of the CCD linear camera. We used the spectra program to monitor the linear CCD camera on real time to get the angle at which the SPR occurs. We begin by changing the Angle-control vertex in the system until the SPR dip appears on the Spectra program. After that we try to improve the results by changing the parameters in the Spectra program, and then we save the data on a separate file. Spectra program saves the data on a file with extension ".spe", this file is used as an input to the program we developed (GSPRS) to process these data. General SPR Sensor (GSPRS) program This program is developed to process the data from the linear CCD camera and display it in a proper way, it has two basic operations; filtering and polynomial fitting. It was developed using Matlab 6.5 GUI and it is shown in figure 28. It contains four basic steps; Step 1: in this step one should enter the incident angle; this can be obtained from the Swivel Mount for 60 mm Cage System or from the path rout after labeling. The beam diameter also can be entered in this step, the default is 4mm. However, if a beam expander is added to the system, this value should change to the new beam diameter. Depending on Step 1, the x-axis range of angles will be determined, for more information, see the code in Appendix C. Step 2: in this step one can enter the location of the file by clicking on the "Import " button a dialog box appears to browse for the file saved by the Spectra program, see figure

46 Fig. 28: GSPRS program. Step 3: this where we can choose the way of representation. Moving Average filter: this uses the "filter" function in matlab which filters a data sequence using a digital filter. It is a direct form II transposed implementation of the standard difference equation. y(n) = b(1)*x(n) + b(2)*x(n-1) b(nb+1)*x(n-nb) - a(2)*y(n-1) a(na+1)*y(n-na). 46

47 Fig. 29: GSPRS Imports Spectra saved files. Where n-1 is the filter order, and which handles both FIR and IIR filters. Or For more info see Matlab help, "filter" function. In this program only the window size is needed to be entered which is indicated by Number of samples per average (SPA). Figure 30 shows the program with MAF in operation. 47

48 (A) (B) (C) (D) Fig. 30: The GSPRS MAF. In Figure 30 (A) it shows the original data from the Spectra program over the whole range (from 1 to 3000) without filtering (window size (Number of SPA)=1), where in (B) it show it with Number of SPA =30. In (C) it shows the original data without filtering but only over the range from (1500 to 2500), and in (D) it show it over the same range but with Number of SPA set to 30. Note that the minimum points are displayed in the right bottom corner of the diagram. 48

49 Polynomial fitting: This can be used to polynomial fit the original data from the Spectra saved file or the data after being filtered by MAF. The program enables us to choose the order of the polynomial fitting and the range on which we want the fitting to be performed. The maximum range is from 1 to 3000 thousands (see the CCD camera section in the Optics and Optical component section), because the data generated by the spectra program contains 3000 points, one for each pixel. The operation of the polynomial fitting is shown in figure 31. (A) (B) (C) (D) Fig. 31: The GSPRS Polynomial fitting. 49

50 In figure 31 the (A) the polynomial fitting was performed over the whole range (1 to 3000) on the original data obtained from the Spectra program, In (B) it is performed on the MAF filtered data over the whole rage, this is done by checking the "Polynomial fitting with MAF" check box, In (C) and (D), it is the same as (A) and (B) but only over the range of (1500 to 2500). Note that the minimum points for both the data curve and the polynomial curve are displayed in the right bottom corner of the diagram. Other Options: other options exist in this program such as the "hold" and "Plot in separate figure" check boxes. If the "hold" check box was checked the plot area will not be cleared and the new data will be plotted over the last plotted data. The "Plot in separate figure" check box causes the data to be plotted in separate figure for other operations to be performed such as saving the plots. Step 4: this is for updating the plot window. Each time a change is done, the "Update" button should be clicked in order for these changes to take place in the plot window or the separate figure. All the code for the GSPRS program is listed in Appendix C. The limitation caused by the CCD camera Because the CCD camera uses a separate program to pull the data from it. It can not be monitored automatically by a separate program, and so we can not make this system fully automated. For this to be solved, we need the API (Application programming interface) for this CCD camera to be integrated in the program we want to develop so the data pulling and analyzing is generated by one program, in that way we can make this program fully automated. Unfortunately Thorlabs were unable to provide us with such thing, and so research should be done to find another CCD camera that has an API to enable the full automation of the system. 50

51 Conclusion and ideas to improve the system We were able to design and build an effective SPR sensor with acceptable specification to meet our aims and objectives. However, this system can be improved if we make some changes like searching for new CCD camera; because as we mentioned in the Optics and Optical components, also in the software section, the CCD camera limits the system from being fully automated, also as we can see in the mechanical design that the CCD camera is big, which limits the system from being smaller. A recommended CCD camera will be smaller and also have an API files that enables us to integrate the CCD camera controlling code in any program we want to develop. Also as we see in figure 32, the SPR (the dark semi-vertical line in the middle of the picture) is not a nice vertical line; it Fig. 32: The SPR as it reflects of the CCD camera. contains some scattered light which make it looks like a zigzag line. This is possible because the gold sample is not uniform or because of the multiple reflection in the optical components. A recommendation to solve this is by using a smaller laser beam and high quality optical components. However, by using a smaller beam the range of angles will continue to be limited (from 45 to 85), which imply that we need to modify the design or change it to cover the whole range again. A Possible recommendation for future design is by using two cylinders sharing the same central point and rotating in opposite directions using a suitable gear system. Another solution is to change the linear CCD camera to a 51

52 Matrix CCD camera, which enable us to take the 2D image of the reflected light and process it using matlab. Another improvements may be researched is making this system portable and independent from the PC computer by adding a suitable DSP controller and LCD to process and show the data, with the connection to the PC computer remains as an option. Finally it would be handy and easier if a component is added so we can adjust the laser position and direction using that component so for an accurate alignment to be possible. 52

53 Appendix A: Mechanical design first idea, and why it is not feasible? From the simulation in figure 13 we find that the rage of angels we need are from degrees. Knowing the range of angles, we need the Angle-Control (and so the arms) to move freely over the whole range, this can be calculated for minimum lengths using simple geometry as shown in figure A1. Fig. A1: simple geometry to calculate arms length. For maximum and minimum angles, Figure A2 shows the geometry representation for that. (A) Fig. A2: (A), maximum incident angle (All lengths units in mm). 53

54 (B) Fig. A2:(B) Minimum Incident angle (All lengths units in mm). We know that the prism with the holder has a vertical length of nearly 5 cm, see fig. A1, adding a safe margin of about 3 cm, this will set the limit for a minimum vertical edge of about 8 cm. knowing that the maximum incident angle is 80 degree, we can draw the triangle shown in figure A2 (A). This will make the arm length nearly equal to 8/cosin(82) 57.6 cm. For the minimum incident angle (fig. A2 (B)), we find the maximum vertical length to be nearly 44 cm. For the above we have a device with approximate dimensions (LxWxH) of (88 cm x 114 cm x 20 cm), which is very big. Also it is hard to build (a lot of work to do in the workshop), and it will be hard to fix the optical component on it, that s why we needed to modify the design. 54

55 Appendix B: Mechanical components used in the design with their drawings. Swivel Mount for 60mm Cage System. This is a prism holder system with 180º Angular Displacement Between Two Legs of a Cage System. 60mm/30mm Cage Plate Adapter. This will enable us to insert the optical components inside the arm instead of on the top. 30mm Cage Plate. This is to fix optical components like lenses. And will allow the optical components to be adjusted easily on the rods. 55

56 Cage Assembly Rods. These will allow some flexibility in the design; also, it is very easy to fix the CCD camera on the end of the CCD camera arm using these rods. The modification done on the basic idea was to make the joints any where along the rods on deferent plan so the upper two arms (laser and CCD camera arms) are on deferent plans from the lower two arms (control arms). This is was possible by using the following component. With this modification the incident angle no longer depends on the arms length, and the system becomes smaller by the half. 90 Degree "T" Extension. Needed to do the modification on the system, see the previous paragraph. 56

57 Vertical to horizontal Rod connector. This will allow a horizontal rod to be connected to a vertical rod and allow the horizontal rod to rotate freely on the vertical rod. Connector position adjuster. This will allow us to adjust the Vertical to horizontal Rod connector to a fixed height along the vertical rod. 90 Degree "T" Extension adjuster. This is needed to prevent the 90 Degree "T" Extension from rotating when we change the angle. 57

58 Threaded Rod. This is needed to move the Angle-control vertex. Vertical to horizontal Threaded Rod connector. This is the actual Angle-control vertex and it moves along the Treaded Rod. Threaded Rod Holder. This is needed to hold one end of the threaded rod, allowing it to rotate freely without changing its position. 58

59 Path router. This is needed to use another way for changing the Angle-control position (explained later). The Drawings 59

60 60

61 61

62 62

63 63

64 64

65 Appendix C: The GSPRS Code. %/********************************************************************* ****/ % CopyRight 2005 % Developer: Sufyan Samara % Tool: Matlap v6.5 GUI % Date: (C)Sebtemper 2005 % % %********************************************************************** ****/ function varargout = untitled(varargin) global SPR_DATA AngleRange Angle SPR_D_FileName SPR_D_FilePath htxtsprfilepath h; % UNTITLED M-file for untitled.fig % UNTITLED, by itself, creates a new UNTITLED or raises the existing % singleton*. % % H = UNTITLED returns the handle to a new UNTITLED or the handle to % the existing singleton*. % % UNTITLED('CALLBACK',hObject,eventData,handles,...) calls the local % function named CALLBACK in UNTITLED.M with the given input arguments. % % UNTITLED('Property','Value',...) creates a new UNTITLED or raises the % existing singleton*. Starting from the left, property value pairs are % applied to the GUI before untitled_openingfunction gets called. An % unrecognized property name or invalid value makes property application % stop. All inputs are passed to untitled_openingfcn via varargin. % % *See GUI Options on GUIDE's Tools menu. Choose "GUI allows only one % instance to run (singleton)". % % See also: GUIDE, GUIDATA, GUIHANDLES % Edit the above text to modify the response to help untitled % Last Modified by GUIDE v Aug :20:31 % Begin initialization code - DO NOT EDIT gui_singleton = 1; gui_state = struct('gui_name', mfilename,... 'gui_singleton', gui_singleton,... 'gui_layoutfcn', [],... 65