Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A piezoelectric fiber scanner for reflectance confocal imaging of biological tissues Fan, Ran 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2019_september_fan_ran.pdf [ 3.05MB ]
Metadata
JSON: 24-1.0380521.json
JSON-LD: 24-1.0380521-ld.json
RDF/XML (Pretty): 24-1.0380521-rdf.xml
RDF/JSON: 24-1.0380521-rdf.json
Turtle: 24-1.0380521-turtle.txt
N-Triples: 24-1.0380521-rdf-ntriples.txt
Original Record: 24-1.0380521-source.json
Full Text
24-1.0380521-fulltext.txt
Citation
24-1.0380521.ris

Full Text

A PIEZOELECTRIC FIBER SCANNER FOR REFLECTANCE CONFOCAL IMAGING OF BIOLOGICAL TISSUES by Ran Fan B.E.Sc., Western University, 2017  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Mechanical Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   August 2019  © Ran Fan, 2019 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  A Piezoelectric Fiber Scanner for Reflectance Confocal Imaging of Biological Tissues  submitted by Ran Fan in partial fulfillment of the requirements for the degree of Master of Applied Science in Mechanical Engineering  Examining Committee: Dr. Ryozo Nagamune, Mechanical Engineering Supervisor Dr. Mu Chiao, Mechanical Engineering Co-supervisor  Dr. Haishan Zeng, BC Cancer Agency Supervisory Committee Member Dr. Xiaoliang Jin, Mechanical Engineering Supervisory Committee Member  iii  Abstract This thesis describes a hand-held confocal optical scanner for cellular imaging. In current clinical practice, the detection of disease like colorectal cancer is performed using endoscopy and biopsy. Random biopsies and oversampling are usually required to reduce false results, so potentials exist for non-invasive optical diagnostic techniques. Optical microscopy techniques such as confocal laser scanning microscopy provide images for biological tissues at a micrometer level. The challenge is to miniaturize the system into a form of hand-held devices or catheters. The developed system in this thesis consists of a portable scanner probe and a reflectance confocal imaging unit. The imaging unit was constructed in an all-fiber configuration for convenient packaging. Confocal scanning was performed at 785 𝑛𝑚 laser illumination with a piezoelectric fiber scanner probe. The fiber-based probe was constructed by mounting a fiber directly on a two-dimensional piezoelectric bender. Horizontal image scanning has been successfully achieved, and the developed device can provide image resolution of 1.16-1.41 𝜇𝑚 in the lateral direction and can resolve cell structures. The operating scanning speed is 1.25 frames per second with 88  ×  88 𝜇𝑚2  field of view, potentially applicable to real-time imaging. Image results were presented with onion epidermis and optical paper samples in comparison to galvanometer mirrors based confocal laser scanning unit.   iv  Lay Summary Cancers have become a leading health problem worldwide, especially colorectal cancers. An endoscope, typically a tubular instrument, is used to look deep into the body as a diagnostic tool in clinical practice for early colorectal cancer detection. Tissue samples (biopsies) are often obtained with endoscopy procedures. Oversampling is required for specific diseases but has adverse effects on patients. Non-invasive optical imaging technologies could be used to improve cancer detection. The research in this thesis is aimed at designing a miniaturized scanning system with appropriate imaging technology that could assist early cancer detection by providing images that could resolve cell structures in real-time. Physicians could reduce the number of unnecessary biopsies with the miniaturized microscopic imaging units. v  Preface The research presented was conducted at the University of British Columbia (UBC) Microelectromechanical Laboratory in the Department of Mechanical Engineering under the supervision of Dr. Mu Chiao and Dr. Ryozo Nagamune. A portion of the research was also conducted in the imaging unit of the Integrative Oncology Department in the British Columbia Cancer Agency under the supervision of Dr. Haishan Zeng. A version of Chapter 2 and 3 are based on the following conference paper that has been published: R. Fan*, A. Moallemi*, Y. Zhang, L. Chen, Y. Zhao, Z. Wu, R. Nagamune, K. Chou, H. Zeng, and M. Chiao, “A Hybrid Pneumatic and Piezoelectric 3D Micro Scanner for Cancer Imaging,” 20th International Conference on Solid-State Sensors, Actuators and Microsystems & Eurosensors 33, pp.1568-1571, Berlin, Germany, June 23-27, 2019. The manuscript is equally contributed by Ali Moallemi and myself. The paper describes the design of a 3D micro scanner using pneumatic and piezoelectric actuators for cellular imaging. The design and testing for the piezoelectric actuators were conducted by myself.    vi  Table of Contents  Abstract ......................................................................................................................................... iii Lay Summary ............................................................................................................................... iv Preface .............................................................................................................................................v Table of Contents ......................................................................................................................... vi List of Tables ................................................................................................................................ ix List of Figures .................................................................................................................................x List of Symbols ........................................................................................................................... xiii List of Abbreviations ................................................................................................................. xiv Acknowledgements .................................................................................................................... xvi Chapter 1: Introduction ................................................................................................................1 1.1 Colorectal Cancer and Current Detection Method .......................................................... 1 1.1.1 Endoscopy and Endoscopic Biopsy ............................................................................ 1 1.1.2 Endomicroscopy ......................................................................................................... 2 1.2 Optical Microscopy Technologies .................................................................................. 3 1.3 Scanners Probes .............................................................................................................. 5 1.3.1 Electromagnetic Fiber Scanners ................................................................................. 6 1.3.2 Electrothermal Fiber Scanners .................................................................................... 7 1.3.3 Piezoelectric Fiber Scanners ....................................................................................... 7 1.3.4 Concluding Remarks for Fiber Scanners .................................................................... 8 1.4 Research Objectives ........................................................................................................ 9 vii  1.5 Thesis Overview ............................................................................................................. 9 Chapter 2: Design and Methods .................................................................................................11 2.1 Optical System Design .................................................................................................. 11 2.1.1 Reflectance Confocal Microscopy ............................................................................ 11 2.1.2 Desktop Optical System Design ............................................................................... 12 2.1.3 All-Fiber Optical System Design .............................................................................. 16 2.2 Mechanical Design of the Probe ................................................................................... 18 2.2.1 Piezoelectric Bender Actuation Theory .................................................................... 18 2.2.2 Piezoelectric Bimorph Scanner Design .................................................................... 20 2.2.3 Piezoelectric 2D Bender Scanner Design ................................................................. 21 2.2.4 Probe Assembly ........................................................................................................ 22 2.3 Signal Control and Imaging Acquisition ...................................................................... 23 2.3.1 Signal Control ........................................................................................................... 23 2.3.2 Scan Patterns ............................................................................................................. 24 2.3.2.1 Spiral Scan Pattern ............................................................................................ 24 2.3.2.2 Lissajous Scan Pattern ...................................................................................... 25 2.3.2.3 Raster Scan Pattern ........................................................................................... 26 2.3.2.4 Scan Pattern for the Probe ................................................................................. 27 Chapter 3: Results and Discussions ...........................................................................................29 3.1 Optical System Characteristics ..................................................................................... 29 3.1.1 Lateral Resolution ..................................................................................................... 29 3.2 Scanner Mechanical Characteristics ............................................................................. 32 3.2.1 Bimorph Scanner Characteristics .............................................................................. 32 viii  3.2.2 2D Bender Scanner Characteristics .......................................................................... 35 3.3 Imaging Results ............................................................................................................ 37 3.3.1 Real-Time Image Acquisition ................................................................................... 37 3.3.2 High Reflective Index Surfaces ................................................................................ 41 3.3.3 Resolution Targets .................................................................................................... 43 3.3.4 Optical Paper and Onion Epidermis ......................................................................... 45 3.3.5 Background Patterns in Images ................................................................................ 47 Chapter 4: Conclusion and Future Work ..................................................................................49 4.1 Summary ....................................................................................................................... 49 4.2 Future Works ................................................................................................................ 50 Bibliography .................................................................................................................................52 Appendices ....................................................................................................................................58 Appendix A Operating Procedures for Imaging System .......................................................... 58 Appendix B LabVIEW program for scanner control and image acquisition ............................ 62  ix  List of Tables  Table 1.1 Comparison of various optical microscopy technologies. .............................................. 4         x  List of Figures  Figure 1.1 Graphic layout of optical endomicroscopy. ................................................................... 3 Figure 1.2 Graphic layout of an endo-microscope scanner probe. ................................................. 5 Figure 2.1 Working principle of reflectance confocal microscopy (adapted from [27]). ............. 12 Figure 2.2 Overall schematic of desktop imaging system design. ................................................ 13 Figure 2.3 Optical schematic diagram of the scanner probe design in a desktop setup. .............. 14 Figure 2.4 Overall schematic for all-fiber imaging system design. .............................................. 16 Figure 2.5 Scanner probe optical configuration for the all-fiber imaging system. ....................... 17 Figure 2.6 Piezo bimorph bender in a cantilever configuration and parallel electrical connection........................................................................................................................................................ 19 Figure 2.7 (a) A schematic of the scanner probe design with piezoelectric bimorphs and (b) a prototype of the piezoelectric bimorph scanner. ........................................................................... 20 Figure 2.8 (a) A schematic of the scanner probe design with the 2D piezoelectric bender and (b) a prototype of the 2D piezoelectric bender scanner. ..................................................................... 22 Figure 2.9 Fully assembled (a) desktop scanner probe and (b) all-fiber scanner probe. .............. 23 Figure 2.10 An illustration of spiral scan pattern with 𝑥 = 𝐴(𝑡)𝑠𝑖𝑛(2𝜋𝑓𝑡), 𝑦 = 𝐴(𝑡)𝑠𝑖𝑛(2𝜋𝑓𝑡 +𝜋/2) with 𝐴(𝑡) = 𝐴𝑡. ................................................................................................................... 25 Figure 2.11 An illustration of spiral scan pattern with 𝑥 = 𝐴𝑠𝑖𝑛2𝜋𝑓𝑡, 𝑦 = 𝐴𝑠𝑖𝑛2𝜋(𝑓 ± 𝛿𝑓)𝑡. . 26 Figure 2.12 An illustration of the raster scan pattern with a fast axis (𝑋) and slow axis (𝑌). ...... 27 Figure 2.13 An illustration of voltage signals applied onto 𝑋 axis (blue) and 𝑌 axis (red) of scanner probe versus pixel sequence. ........................................................................................... 28 xi  Figure 3.1 Two-dimensional layout of the scanner probe optical design. .................................... 30 Figure 3.2 MFT diagram of the confocal scanner probe with C230 objective lens for on-axis and off-axis (100 𝜇𝑚) performances using Zemax. ............................................................................ 30 Figure 3.3 (a) Image representation of Ronchi ruling target with 110 𝑙𝑝/𝑚𝑚, (b) intensity profile obtained by scanning the target in (a) and (c) the intensity profile scanning at the center and edges of the FOV. ......................................................................................................................... 31 Figure 3.4 Piezoelectric bimorph scanner frequency performances under  𝑉𝑝𝑝 = 74 𝑉 triangular waveform in both (a) 𝑋 and (b) 𝑌 directions. ............................................................................... 33 Figure 3.5 Deformation distribution of the bimorph scanner (a) in 𝑋 (maximum ~400 𝜇𝑚) and (b) in 𝑌 direction (maximum ~700 𝜇𝑚) at static operations. ...................................................... 34 Figure 3.6  Piezoelectric bimorph scanner characteristics under constant driving frequencies in both (a) 𝑋 (200 𝐻𝑧) and (b) 𝑌 (1 𝐻𝑧) directions. ......................................................................... 34 Figure 3.7 Scan patterns of the piezoelectric bimorph actuator with 200 𝐻𝑧 in 𝑋 axis and 60 𝐻𝑧 in 𝑌 axis under 74 𝑉𝑝𝑝. .................................................................................................................. 35 Figure 3.8 Deformation distribution of the 2D scanner at (a) maximum of  ~279.39 𝜇𝑚 in 𝑌 and (b) maximum of ~273.01 𝜇𝑚 in 𝑋 direction under static operations. ......................................... 36 Figure 3.9 Piezoelectric 2D scanner frequency performances under triangular modulated voltage (𝑉𝑝𝑝 = 60 𝑉) in both (a) 𝑋 and (b) 𝑌 directions. ........................................................................... 36 Figure 3.10 Piezoelectric 2D scanner characteristics under constant frequencies in both (a) 𝑋 (320 𝐻𝑧) and (b) 𝑌 (1 𝐻𝑧) directions. ........................................................................................... 37 Figure 3.11 Phase delay in 2D bender with the driving signal in blue and the actuator response in red. ................................................................................................................................................ 38 xii  Figure 3.12 Pixel delay of the imaging data due to response delay of the actuator (blue for data and red for non-useful data signals). ............................................................................................. 38 Figure 3.13 Voltage applied onto 𝑋 (fast) axis and 𝑌 (slow) axis of the scanner probe versus pixel sequence with pixel delay correction. .................................................................................. 39 Figure 3.14 Image of a resolution target (a) before and (b) after pixel delay correction. ............. 40 Figure 3.15 Horizontal scan of a mirror using the 2D bender scanner and the all-fiber confocal system. .......................................................................................................................................... 42 Figure 3.16 Horizontal scan of aluminum deposited glass slide using the 2D bender scanner and the all-fiber confocal system. ........................................................................................................ 43 Figure 3.17 Horizontal scan of a USAF resolution target group 7 for (a) element 1-3 and (b) element 3-6.................................................................................................................................... 44 Figure 3.18 Horizontal scan of a 10 𝜇𝑚 grid on the resolution card. ........................................... 45 Figure 3.19 Paper image (a) using developed 2D bender scanner and all-fiber confocal system and (b) using desktop galvanometer mirror confocal system. ...................................................... 46 Figure 3.20 Onion epidermis image (a) using developed 2D bender scanner and all-fiber confocal system and (b) using desktop galvanometer mirror confocal system. .......................................... 46 Figure 3.21 Background noise pattern in images. ......................................................................... 47  xiii  List of Symbols 𝑑31 Transverse piezoelectric large-signal deformation coefficient 𝐷 Laser beam size 𝐷𝑒𝑓𝑓 Effective flexural stiffness of the piezoelectric bender beam 𝐸𝑝 Young’s modulus of piezoelectric material 𝑓 Focal length ℎ𝑐 Height of the substrate plate in a piezoelectric bender ℎ𝑚 Distance between the neutral axis and the piezoelectric layer for a piezoelectric bender ℎ𝑝 Height of the piezoelectric plate in a piezoelectric bender 𝐿 Free length of a cantilever beam 𝑅𝑙𝑎𝑡𝑒𝑟𝑎𝑙 Lateral resolution 𝑉𝑝𝑝 Peak to peak voltage 𝜆 Laser wavelength 𝜔 Mode field diameter of a fiber      xiv  List of Abbreviations APD Avalanche photodiode a.u. Arbitrary unit CCD Charge-coupled device CM Confocal microscopy DAQ Data acquisition FCM Fluorescence confocal microscopy FEA Finite element analysis FL Focal length fps Frames per second FOV Field of view LDV Laser doppler vibrometer lp Line pair MEMS Microelectromechanical system MPM Multi-photon microscopy MTF Modulation transfer function  NA Numerical aperture ND Neutral density xv  NIR Near-infrared OCT Optical coherence tomography PBS Polarization beam splitter PM Polarization-maintaining PZT Piezo zirconate titanate RCM Reflectance confocal microscopy SHG Second-harmonic generation SMF Single-mode fiber SNR Signal-to-noise ratio TPEF Two-photon excitation fluorescence USAF U.S. air force   xvi  Acknowledgements I am most grateful to my research supervisors, Dr. Mu Chiao, and Dr. Ryozo Nagamune, for their support and guidance over the last two years. Their vision and expertise have significantly inspired me not only on my research but in my future professional career development. It has been a great honor to study and work in UBC with such kind and caring individuals and join their extraordinary research groups. I would like to thank Dr. Haishan Zeng and Dr. Xiaoliang Jin for dedicating their valuable time serving on my examining committee and for all their comments that have improved my thesis.  Special thanks to Dr. Haishan Zeng for the help with my research carried in BC Cancer Agency.  I want to express my enduring gratitude to the faculty, staff, and my fellow students at UBC, who have inspired me in the field of my study and helped me with my work. I would like to thank all members of the Microelectromechanical System Laboratory and Control Engineering Laboratory family for sharing knowledge and having an excellent time together. I owe particular thanks to Dr. Lei Chen, Dr. Yuan Zhao, Dr. Hadi Mansoor, Zhenguo Wu, and Yunshan Zhang, for answering my endless questions in the field of optical engineering.  Finally, I would like to thank my parents and my boyfriend, Siran Jia, for their encouragement, support, and love in both my study and life.     1  Chapter 1: Introduction 1.1 Colorectal Cancer and Current Detection Method Colorectal Cancer has become a leading health problem, with over 1 million new cases diagnosed and 551,629 deaths occurred worldwide in 2018 [1]. In Canada, it is the second cause of mortality from cancers, and the projected deaths with both sexes combined are 9,400 in 2018 [2]. Cancer stage, which is how far cancer has progressed, is the most critical factor for early colorectal cancer detection. The probability of surviving from colorectal cancer will be significantly improved if it can be diagnosed at an early stage with a 5-year survival rate as high as 90% [3]. In current clinical practice, endoscopy and biopsy are the main tests used by doctors to diagnose colorectal cancer.  1.1.1 Endoscopy and Endoscopic Biopsy Endoscopy is a procedure that physicians conduct visual inspections for the lining of the entire colon using an endoscope. The endoscope is a flexible, thin, and tube-like tool with several channels to examine or treat organs when advances in the body cavities. In a typical clinical endoscope, there are channels for video camera and light, a channel for water irrigation, a channel for air, and a channel for instruments. Special tools such as tiny forceps, snares, or wired loops can be passed through the instrument channel to remove polyps or take tissues out. An endoscopic biopsy is usually collected with the endoscopy. During a biopsy procedure, cells, tissues, or tumors are collected as samples using cutting tools passed in the instrument channel. After taking the biopsy, the samples are sectioned and stained to obtain histology images used for further cancer detection. Clinical decisions are made based on the samples from abnormal 2  morphological or color appearance region. Random sampling and oversampling are sometimes required to lower the chance of false results and increase diagnostic yields [4]. Side effects of biopsy include bleeding, infection around biopsy site, and discomfort and pain in patients. Therefore, there is a high potential for non-invasive diagnose procedure for early colorectal cancer detection.  1.1.2 Endomicroscopy Optical microscopy is an imaging technology that could provide detailed morphological and histological information of biological samples. By miniaturizing optical microscope into a  scanning probe and combing to clinical endoscope instrument channel, a non-invasive method named endomicroscopy could better facilitate early colorectal cancer detection. The tissue morphology is observed in-vivo (in a living organism) rather than ex-vivo from the endoscopic biopsy, which involves cutting and staining tissues from the organism in an external environment. Horizontal sectional images (imaging planes that are parallel to tissue surface) and cross-sectional images (so-called vertical sectioning images) could be obtained with the optical images from endomicroscopy. Endomicroscopy will significantly reduce the time required and increase the efficiency for the diagnostic tests, and reduce the pain in patients as well. There are two essential parts for an optical endo-microscope; an optical system with appropriate imaging techniques to generate tissue image and a scanner probe that could fit into the instrument channel of a clinical endoscope with miniaturized scanning mechanism (Figure 1.1).  3   Figure 1.1 Graphic layout of optical endomicroscopy. The following sections will review the state-of-art technologies for both units.  1.2 Optical Microscopy Technologies For early cancer detection, studies mainly focused on the mucosa epithelial cells [5] (first layer of the colon), and the typical cell sizes are 8-33 𝜇𝑚 [6]. Various emerging optical technologies have been developed for non-invasive biomedical imaging at the cellular resolution level. Multiple groups have demonstrated imaging techniques, including optical coherence tomography (OCT) [7], multi-photon microscopy (MPM) [8] and confocal microscopy (CM) [9] for in-vivo assessment. Images obtained by these technologies could provide comparable morphological information as the histology images from an endoscopic biopsy.  Optical coherence tomography is an imaging technique using low coherence near-infrared light to capture micrometer-resolution images. The interference patterns between the reference light source and the backscattered light from samples are used as signals for image reconstruction [10]. OCT can perform cross-sectional, three-dimensional images of microstructures with relatively large penetration depth but generally low resolution and contrast.  Multi-photon microscopy is a fluorescence imaging technique based on two-photon excitation fluorescence (TPEF) and second harmonic generation (SHG) to perform high-resolution tissue 4  imaging. Electrons in the sample are excited by absorbing two low energy photons, and a higher energy photon is then emitted. Cross-sectional imaging could be performed with demanding optics like femtosecond lasers (lasers with ultra-short pulses) and high numerical aperture (NA) objectives, which significantly increase the cost of the system.  Confocal microscopy uses apertures or pinholes to reject out-of-focus light, enabling sectioning at specific depth with micro resolution and high contrast. Imaging is realized based on the variations in the reflective indexes of samples. CM is usually performed in fluorescence-mode confocal microscopy (FCM) and reflectance confocal microscopy (RCM). FCM offers high contrast and high signal-to-noise ratio (SNR) images by introducing exogenous fluorophores while RCM relies on the native tissue signals [11].  A comparison of all the optical technologies mentioned with their characteristics is summarized in Table 1.1[12].  OCT MPM FCM RCM Resolution ~10 𝜇𝑚 ~1 𝜇𝑚 ~1 𝜇𝑚 ~1 𝜇𝑚 Penetration Depth ~2 𝑚𝑚 ~1 𝑚𝑚 ~0.5 𝑚𝑚 ~0.5 𝑚𝑚 Contrast Low High High High Contrast Agent No No Yes No Cost Medium (~50K) High (~200K) Medium (~50K) Medium (~50K)  Table 1.1 Comparison of various optical microscopy technologies. Among the mentioned technologies, OCT provides considerable penetration depth, but the resolution is not as good as others. RCM could provide high-contrast and high-resolution images without exogenous fluorescence agents and expensive optics like FCM and MPM. Although the imaging depth is limited in RCM, it is good enough for detecting epithelial cell with layer thickness 5  around 200 𝜇𝑚 to 300 𝜇𝑚 [5]. For this reason, RCM is chosen as the optical technology for the design of the imaging system in this thesis.  1.3 Scanners Probes Another essential component in endo-microscope is a scanning unit. A scanning probe usually consists of lenses and optical fibers (Figure 1.2).  Figure 1.2 Graphic layout of an endo-microscope scanner probe. Scanning is realized by tilting the illuminating path in an optical setup to change the focal point relative to the samples. Optical fibers are used as the form of the optical path to make the system flexible. Optical fibers have small footprints that are well suitable for compact applications. Multiple groups have demonstrated scanners probes such as micromirrors scanners [13], lens scanners [14], and fiber scanners [15]. The micromirror scanners work as manipulating the illuminating path by rotating the mirrors in the scanners. Two-dimensional micromirrors, which 6  are inserted between the fiber end and the objective lens serves as the light path manipulators. This approach has complexity in design and manufacturing in nature. The lens scanners shift the focus point by moving the lenses in the setup. However, the lens weight limits the mechanical scanning speed by putting burdens on the probe structure. The working principle of fiber scanners is to deflect the illumination fiber directly by attaching the fiber on actuators. Any movement of the fiber end translates to displacements of the focusing spot on samples. Fiber scanners have the most compact forms [16] and have been selected as the scanner probe design in this thesis. Microelectromechanical System (MEMS) is a technology that can integrate miniaturized sensors and actuators with control circuits into a small device. Compact endo-microscope fiber scanners are often fabricated using MEMS technology with actuators such as electromagnetic [17] [18], piezoelectric [19], and electrothermal [20] types. The following section reviews fiber scanners based on their actuation methods.  1.3.1 Electromagnetic Fiber Scanners Electromagnetic actuation uses magnetic fields to move an object. When a ferromagnetic subject is placed in an external magnetic field, a magnetic force is exerted on the ferromagnetic material if there is a misalignment of magnetic moments of domains that magnetizes the ferromagnetic material. Our group has previously demonstrated a magnetically driven fiber scanner [21] using coils. Scanning is achieved by mounting a fiber onto two ferromagnetic cantilever beams driven by electromagnetic forces. The size of the probe is 4.5  𝑚𝑚 ×  2.5 𝑚𝑚 ×  22 𝑚𝑚  with a mechanical scanning area of 0.4× 0.6 𝑚𝑚2 and the driving power for the system is about 168 𝑚𝑊. 7  Electromagnetic actuators could provide large displacements but require high driving currents and high electric power at the same time. They also present disadvantages in miniaturization since the scanning range is limited by the magnetic forces drop in the miniaturized scanner [21]. 1.3.2 Electrothermal Fiber Scanners Electrothermal actuators operate based on the thermal expansion of materials. Thermal expansion is caused by the joule heat as electric currents passed through the structure in the design [22].  Seo et al. [20] reported an optical fiber scanner with electrothermal silicon micro-actuator in a bimorph beam structure. A 20 𝑚𝑚  fiber is directly mounted on a double hot arm and a cold arm in asymmetric structure and the total footprint of the scanner is 1.28 𝑚𝑚 × 0.44 𝑚𝑚 × 28 𝑚𝑚. Two-dimensional scanning is realized by a current flow in the double hot arm. The joule heat in the double hot arm results in a thermal expansion difference with both the cold arm and the cold optical fiber. The scanner operates at 239.4 𝐻𝑧 and 218.4 𝐻𝑧 for both scan directions with a scan area of 451 × 558 𝜇𝑚2 and 90 𝑚𝑊 power consumption. The advantage of using electrothermal actuators is the broad actuation range they provide, but the drawback is the slow operating speed resulted from the low thermal dissipation rate. Most electrothermal actuated scanners operate on a few hundred hertz. Another drawback is the high-power consumption for driving the actuators with large currents. 1.3.3 Piezoelectric Fiber Scanners Piezoelectricity phenomenon is known as when a force applied on piezoelectric materials such as lead zirconate titanate (PZT), electric charges are generated in the crystalline materials. Piezoelectric actuators use the converse effect, i.e., an applied electric field yields mechanical 8  response. Rivera et al. [23] demonstrated a 2D piezoelectric fiber scanner with two piezoelectric bimorphs. Area scanning is realized by aligning the two bimorphs with their bending axes perpendicular to each other. With a 4 𝑐𝑚 length scanner probe, the device can achieve 110 𝜇𝑚 × 110 𝜇𝑚 area scanning and 4.1 frames per second scanning rate. Schulz-Hildebrandt et al. [24] developed a piezoelectric fiber scanner using a customized piezo tube with two pairs of radially attached electrodes serving as the two scanning axes. Fiber is positioned centrally inside the tube to form a fiber cantilever structure. Bending oscillations of the fiber cantilever are generated by applying sinusoidal voltages of 38.7 𝑉𝑝𝑝 and 46.1 𝑉𝑝𝑝 for both axes respectively. The scanning frequency is 1.22 𝑘𝐻𝑧 with a 13.5 𝑚𝑚 probe length. The mechanical scanning area is a 0.8 𝑚𝑚 diameter circle. Piezoelectric actuators operate at very high speed, usually kilohertz but with limited deflections compared to other types of actuators. Operating voltages for piezoelectric actuators are typically high; however, the total electrical power consumption is usually lower ( ~10 𝑚𝑊 ) than electromagnetic and electrothermal actuators ( ~100 𝑚𝑊 ) due to small driving currents (~100 𝜇𝐴).  1.3.4 Concluding Remarks for Fiber Scanners Endoscopic scanners are categorized based on their actuator types (micromirror, lens scanners, and fiber scanners) and actuation methods (electromagnetic, electrothermal, and piezoelectric types).  Among them, fiber-based piezoelectric scanners have relative fast scanning speed, compact size, and low power consumption. A fast scanning rate is required for in-vivo applications. Although 9  the displacement range for piezoelectric actuators is typically small, it can be improved if a fiber cantilever structure is designed in the scanner.  1.4 Research Objectives The research hypothesis is that a compact, portable, and non-invasive endo-microscope to obtain real-time cellular images for biological tissues in horizontal section planes is achievable and can be developed. The objective of the thesis is to develop an optical endo-microscope that could fit into the instrument channel (c) of a clinical endoscope [25]. For real-time image, the imaging speed should be fast enough (ideally more significant than 6 fps). The image size should be larger than 100 × 100 𝜇𝑚2 and the lateral resolution of the system needs to reach around 1 𝜇𝑚 to resolve cell structure and provide enough details for high-quality images. 1.5 Thesis Overview This thesis is organized with background introduction in Chapter 1, followed by design methods in Chapter 2, simulation and experimental results in Chapter 3, and finally concluded with future work discussions in Chapter 4. Chapter 1 introduces the current detection method for early colorectal cancer and the need for non-invasive techniques to facilitate detections. Current optical technologies and scanner probes are introduced and compared as well.   Chapter 2 describes the optical design, mechanical design, and software design of the endo-microscope system. 10  Chapter 3 presents the optical and mechanical characteristics of the developed system as well as various image results obtained using the system. Chapter 4 summaries the work has been done and discuss the potential future improvements for the developed imaging system. 11  Chapter 2: Design and Methods The design for the endo-microscope is divided into two parallel parts; one is the reflectance confocal optical system design, and the other is a moveable scanner probe design. The two parts are connected by a flexible optical fiber.  2.1 Optical System Design One method for the reflectance confocal optical system design is to use optics devices with desktop setup while the other is to use all-fiber optical components. Our group has previously demonstrated the design of a desktop reflectance confocal system [26]. A reconstruction, according to previous work with slight modifications, is demonstrated first. However, the desktop system lacks stability in experiments. At the same time, aligning components in the system requires skills and is also time-consuming. Thus, an all-fiber optical system is designed later to conquer the problems.   2.1.1 Reflectance Confocal Microscopy Reflectance confocal microscopy provides high-resolution tissue images. Figure 2.1 shows a typical configuration of reflectance confocal microscopy.  12   Figure 2.1 Working principle of reflectance confocal microscopy (adapted from [27]). The system consists of a light source, a beam splitter, an objective lens, two pinholes, and a detector. Laser illuminated from the light source is limited by the source pinhole and focused on the samples by the objective lens. Backscattered light from the focal plane enters the detector through the detector pinhole. The point light source at the source pinhole, the illuminated spot on the sample and the detector pinhole form three conjugate parts to reject the out-of-focus light scattering [27]. Because only focused light is collected, RCM has high SNR and can reduce image blur [13].  2.1.2 Desktop Optical System Design The overall desktop optical system schematic is shown in Figure 2.2.  13   Figure 2.2 Overall schematic of desktop imaging system design. A 785 𝑛𝑚 near-infrared laser diode (Starbright 785XM, Torsana Laser Technologies) with 500 𝑚𝑊 output power is used as the light source for the system. The 785 𝑛𝑚 wavelength laser is selected as it can provide sufficient energy for light scattering in tissue and enough penetration depth [28] at the same time. The laser first passes through a 10x beam expander to expand the beam size from 1 𝑚𝑚  to approximately 10 𝑚𝑚 . Larger inputting beam size yields smaller collimated beam at a large distance as it passes through focusing optics [29]. The small beam size is necessary for fiber coupling later. The expanded laser then passes through a half-wave plate to the polarization beam splitter (PBS). The half-wave plate alters the polarization state of the light traveling through it. By orienting the half-wave plate, the amount of the reflected light by PBS can be controlled. If the half-wave plate fast axis is appropriately aligned with the PBS, the maximum amount of vertically polarized light is being reflected, and non-vertical polarized light passes through the PBS and absorbed by light stoppers. The PBS reflected light then passes through a 14  fiber coupling objective lens (Newport, M-10x) and is focused into a single-mode polarization-maintaining (PM) fiber (Thorlabs, PM780-HP) with a mode field diameter of 4.5 𝜇𝑚. To couple light efficiently into a fiber, the beam size of the laser needs to match the fiber core size. The coupling objective lens (focal length 𝑓 = 16 𝑚𝑚) is selected based on the following equation [30]:  𝑓 = 𝐷𝜋𝜔4𝜆 , (2.1) where 𝜔 is the mode field diameter of the PM fiber, 𝐷 is the laser beam size, and 𝜆 is the laser wavelength.  The fiber acts as a flexible connection between the desktop laser setup and the scanner probe. The optical schematic diagram of the scanner probe design is shown in Figure 2.3.  Figure 2.3 Optical schematic diagram of the scanner probe design in a desktop setup. In the scanner probe and at the distal end of the PM fiber, light is linearly and vertically polarized. The laser light coming out from the PM fiber is collimated by a commercial collimating lens (Thorlabs, C220TMD-B) and then directed through a quarter-wave plate. Orienting the fast or slow 15  axis of the quarter-wave plate at 45° to the incident light will change the light polarization from linear to circular. Circularly polarized light could better facilitate imaging for biological samples [31]. The light is finally focused onto the sample by an objective lens (Thorlabs, C230TMD-B) and scattered. Scattered photons from the sample retain the polarization state as before [32] and pass the quarter-wave plate again with another 45° rotation. In total, the light has rotated 90° according to the incident light and is horizontally polarized now. The scattered light is then coupled back to the PM fiber, and the PM fiber acts as the confocal pinhole, rejecting the light that is not from the focal point. Only horizontally polarized light coming back from the PM fiber can pass through the PBS again and be detected by the avalanche photodiode (HAMAMATSU, C10508-01) at the detector side of the PBS. The PBS blocks unwanted light polarization states as well as light reflections from the surface of optical components, which is referred to as polarization effects to improve imaging quality and contrast [33].  On the detector side of the PBS, a 100 𝜇𝑚 pinhole is placed before the avalanche photodiode (APD) to block out ambient light noise. The light intensity signal is converted to electrical signals by the APD and collected using a data acquisition card (DAQ-card). The card is also used to generate driving signals for the scanner as well as digitalize the light signal for image reconstruction.  In our previous design, there is a metallic neutral density filter (ND filter) before the beam expander to control the inputting power to the system. However, due to the low efficiency in the fiber laser coupling in experiments, the power of light delivered to the sample is not enough for imaging, so the ND filter is removed. The system also lacks stability, and the image quality depends mostly on the pinhole (the one located in front of the APD) size and its precise position, making aligning between the components time-consuming.  16  2.1.3 All-Fiber Optical System Design The desktop system involves many optics like the half-wave plate, the PBS, the quarter-wave plates, and the detector pinhole. These components need to be fixed relative to each other as a whole, and time and skills are required for aligning them properly. A flexible all-fiber optical system is then designed to replace some critical parts of the desktop system, allowing operating and packaging the system with ease. The schematic of the all-fiber system is shown in Figure 2.4.   Figure 2.4 Overall schematic for all-fiber imaging system design. The 785 𝑛𝑚 laser is still used as the light source. Instead of using the polarization effects and PBS to separate the incoming light and sample scattered light, a 3-port circulator (OF-LINK, PICIR-785-H7-L-10-FA) is used instead. A circulator is a passive device, in which a light signal entering into any port of the device is transmitted into the next port in a rotation fashion. Light from the laser passes through the beam expander as before but is coupled into the Port 1 fiber of the circulator. The light coming from Port 1 goes out to Port 2, and the Port 2 fiber is coupled with the scanner probe fiber using fiber coupler. PM fiber is no longer needed since no polarization effect 17  is involved in the design and single-mode fiber (SMF) at the laser wavelength is used instead. The scanner probe optical design is the same as before but with the quarter-wave plate removed (Figure 2.5). Scattered light comes back to the circulator through Port 2 and goes out to Port 3. The Port 3 fiber is coupled directly with a photodiode (OPEAK, PD-M-TBPIN-SW) for intensity signal sensing. Same DAQ card is used for signal control, collection, and imaging reconstruction.  Figure 2.5 Scanner probe optical configuration for the all-fiber imaging system. Compared with desktop design, the all-fiber design is more portable and more comfortable to operate. Using all-fiber components reduce the environmental noise by packaging the parts together, and the scattered signals are more stable for future in-vivo measurement since not much alignment issues involved in the design. One drawback of the all-fiber system is slightly higher signal loss in the circulator since the circulator itself has an attenuation (~25%). Despite the signal loss, the system could still provide high-quality images with enough resolution since the detector has a very high amplification gain for signal sensing. Time for aligning the components is reduced since most connections are designed as simple plug-in mechanisms using fiber couplers.  18  The beam expander and fiber coupling lens could also be replaced by a fiber port collimator, which could fit directly onto the laser source. A customized 780 𝑛𝑚 collimator has been made (OPEAK) but not yet applied to the system. A polarization controller (OPEAK) could also add to the system to regulate the inputting light power to the system in order to protect the circulator and other optic components.  2.2 Mechanical Design of the Probe The probe is designed as a piezoelectrically actuated fiber raster scanner. Piezoelectric actuators provide high-speed scanning with less power consumption, and raster scan pattern does not need complicated post-imaging processing procedures.  2.2.1 Piezoelectric Bender Actuation Theory Piezoelectricity is a well-known effect in which pressure on piezoelectric materials yields an electric field. Piezoelectric actuators utilize inverse piezoelectric effect that an electrical voltage applied on the materials induce internal stress and yields mechanical response as a change in length [34]. The most common types of piezoelectric actuators used in endomicroscopic applications are piezoelectric tubes [23], and other types of actuator like piezoelectric benders have also been applied [23]. In general, piezoelectric benders usually have more significant deflection (hundreds of micrometers) than piezoelectric tubes (several micrometers). In order to have a massive deflection and a broad imaging field, piezoelectric benders are considered for the scanner probe design.  Benders are mainly in unimorph and bimorph types. Unimorph benders have one layer of piezoelectric plate bonded to a non-piezoelectric layer while bimorph benders have a sandwich 19  structure (Figure 2.6). Two piezoelectric plates are joined together by a third elastic layer in between to yield more mechanical reliability [35]. Benders are usually mounted as cantilever beam in applications.  Figure 2.6 Piezo bimorph bender in a cantilever configuration and parallel electrical connection. Piezoelectric plates are manufactured with polling directions. Controlling the electric field across the plates to the polling directions will make the beam bend upwards or downwards. Plates in the bimorph can be polarized in the same direction (parallel configuration) or the opposite direction (series configuration). The deflection of a parallel electrical configured piezoelectric bender is given as the following equation [36]:  𝛿 = 𝑑31𝐸𝑝ℎ𝑚𝑉𝐿22𝐷 , (2.2) where 𝑑31 is the transverse piezoelectric large-signal deformation coefficient, 𝐸𝑝 the is Young’s modulus of the piezoelectric material, ℎ𝑚  is the distance between the neutral axis and the piezoelectric layer for a bender, 𝑉 is the operating voltage, 𝐿 is the free length of the beam and 𝐷𝑒𝑓𝑓 the effective flexural stiffness of the beam is [37], 20   𝐷𝑒𝑓𝑓 =2𝐸𝑝3ℎ𝑝3 +𝐸𝑐ℎ𝑐312+𝐸𝑝ℎ𝑐2(2ℎ𝑝2 + ℎ𝑐ℎ𝑝) , (2.3) where 𝐸𝑐  is the Young’s modulus of the middle substrate material, ℎ𝑝  is the height of the piezoelectric material and ℎ𝑐 is the height of the substrate. To get a sense for the magnitude of the deflection, a commercially available piezoelectric bender (PiezoDrive, BA3502) has a dimension of 35 (L) × 2.5 (W)× 0.7 (H) 𝑚𝑚 and the quoted 𝑑31 is −270 × 10−12 𝑚/𝑉 with a stiffness of 220 𝑁/𝑚. The quoted maximum deflection is 0.7 𝑚𝑚 at 74 𝑉.  2.2.2 Piezoelectric Bimorph Scanner Design  Two-axis scanner is required to operate lateral scanning in both 𝑋 and Y directions to perform a horizontal scan. The scanner is designed with two types of piezoelectric actuators. One is to use two piezoelectric bimorphs mounted together while the other is two use a commercial 2D piezoelectric bender. Bimorph benders could provide more displacements but are operated at a slower speed, and they are generally larger than 2D benders.  A schematic of the scanner actuator design with piezoelectric bimorphs is shown in Figure 2.7.  Figure 2.7 (a) A schematic of the scanner probe design with piezoelectric bimorphs and (b) a prototype of the piezoelectric bimorph scanner. 21  The scanner consists of a fiber and two cantilever beams (𝐵1 and 𝐵2). Both beams are made of commercially available piezoelectric bimorphs (PiezoDrive, BA3502). 𝐵2 is clamped at one end with 3D printed mold (MIICRAFT, Ultra) and is compliant to bend in 𝑌 direction. 𝐵1 is mounted on the free end of 𝐵2 using UV curable adhesive (Norland, Optical Adhesive 63) such that their bending axes are perpendicular to each other, making 𝐵1 flexible in 𝑋 direction. The mounted part between the beams has a length of 5 𝑚𝑚. The piezoelectric bimorph has a dimension of 35 (L) × 2.5 (W)× 0.7 (H) 𝑚𝑚, and the total length in this design is 65 𝑚𝑚. Fiber is mounted along the neutral axis of both beams using the same UV adhesive, with the fiber tip aligning the free end of 𝐵1. The alignment is to make sure the fiber tip movement matches the beam movements, and no phase shift occurs between them. If fiber is hanging with a free length, it will form another cantilever structure and resonant at a different frequency, making scanning pattern distorted. Scanning is realized by actuating both beams with patterned voltage signals using an open-loop piezoelectric controller (Thorlabs, MDT693B). Both beams are designed to bend in their intended bending axis (~700 𝜇𝑚) since the stiffness in other directions are much higher, leading little imaging artifacts.  2.2.3 Piezoelectric 2D Bender Scanner Design  The bimorph benders yield massive displacement with the long beam length. However, to make the scanner probe more suitable for an endo-microscope, the rigid length of the scanner needs to be further reduced. A long rigid length will make the probe hard to advance in body curvatures and bring pain in patients. In the same time, by mounting two benders together, 𝐵1 scanning speed is limited by the weight of the 𝐵2, and any miss alignment (like the beams are not mounted at the right angle) between the two bending axes will distort the scan pattern. A commercially available 22  2D bender (Noliac NAC2710) is used as the second design (Figure 2.8) to reduce the rigid probe length.   Figure 2.8 (a) A schematic of the scanner probe design with the 2D piezoelectric bender and (b) a prototype of the 2D piezoelectric bender scanner. Similar as before, the 2D piezoelectric bender (36.5 (L) × 1.75 (W) × 1.75 (H) 𝑚𝑚) is used as a cantilever beam with one end clamped by a 3D printed fixture (MIICRAFT, Ultra). Fiber is mounted directly on the actuator with the fiber end aligning the free end of the actuator. The 2D bender itself offers movement in two directions (𝑋  and 𝑌 ) with a displacement of 180 𝜇𝑚 , respectively. Scanning is realized by applying patterned voltages to the actuator through an open-loop piezoelectric controller. The 2D bender scanner has limited scanning range but is more portable in size.  2.2.4 Probe Assembly Lenses and waveplates are positioned and secured using retaining rings, cage plates, and posts for experiment setup. With the 3D printed fixture, the actuators are mounted on an adjustable cage plate to align with the optical axis of the collimating and objective lens couple. The fiber tip is 23  positioned at the focal point of the collimating lens. Figure 2.9 shows the assembled scanner probe. If the probe is used for all-fiber system, the quarter-wave plate is removed.   Figure 2.9 Fully assembled (a) desktop scanner probe and (b) all-fiber scanner probe. The size of the lenses (6.33 𝑚𝑚 diameter) in the current design determines the probe diameter. The assembled probe dimension could be further reduced if miniaturized lenses are used, but with a trade-off in image resolution.  2.3 Signal Control and Imaging Acquisition 2.3.1 Signal Control A scanner control and image reconstruction program are developed in LabVIEW to record real-time image data. The DAQ-card (National Instruments, DAQ 6363) is used to generate the driving signals. The reflected light signal from the samples is detected by the photodiode. The photodiode itself converts the intensity light signals to voltages signals. The DAQ-card reads the inputting voltage signals with synchronization to the driving signals. Synchronizations avoid the programmable phase shift between different scanning sections.  24  A real-time image is constructed in LabVIEW to display the horizontal section scanning area. The image is converted to grayscale for better viewing based on the amplitudes of the voltages at the entire scanning area.  2.3.2 Scan Patterns The fiber scanners move in two directions with modulated waveforms to scan an area. Depending on the frequency of the waveforms, scanners may be in resonant or non-resonant modes with scanning patterns such as spiral, Lissajous, and raster. An ideal scan pattern should be able to sample a scan area thoroughly with a balanced sampling density. Depending on the applications and actuator types, different scan patterns are available.  2.3.2.1 Spiral Scan Pattern A spiral scan uses resonant sinusoidal motion in both axes. A triangular or sinusoidal modulated waveform scan a circular area with increasing diameter. Figure 2.10 shows a spiral scan pattern. Since both orthogonal axes are driven in resonant modes, the deflections in scanners are maximized with much lower driving energy. However, post-imaging processing is required to map the scan from angular form to Cartesian form for image reconstruction. Each circle in one scan requires the same amount of time to finish, leading a densely sampled center area and lightly sampled circumference. A constant phase shift between the two scanning axes is also required. However, any small phase shift that may come from the mechanical design or environment noise will distort the scanning pattern. Typically, a lookup table is created by scanning a known pattern to obtain the image reconstruction algorithm.  25   Figure 2.10 An illustration of spiral scan pattern with 𝒙 = 𝑨(𝒕)𝒔𝒊𝒏(𝟐𝝅𝒇𝒕), 𝒚 = 𝑨(𝒕)𝒔𝒊𝒏(𝟐𝝅𝒇𝒕 + 𝝅/𝟐) with 𝑨(𝒕) = 𝑨𝒕. 2.3.2.2 Lissajous Scan Pattern A Lissajous scan is also a resonance scan with a small frequency shift between the two scanning axes. Compared with a spiral scan, the scan area has a more balanced sample density and more stability since no phase shift presented between the two orthogonal axes. However, the mechanical coupling between the two scanning axes may lead to distortions in a scan pattern, and precise tracking of the fiber is needed [38]. Figure 2.11 shows a Lissajous scan pattern.  26   Figure 2.11 An illustration of spiral scan pattern with 𝒙 = 𝑨𝒔𝒊𝒏(𝟐𝝅𝒇𝒕), 𝒚 = 𝑨𝒔𝒊𝒏(𝟐𝝅(𝒇 ± 𝜹𝒇)𝒕). 2.3.2.3 Raster Scan Pattern The raster scan is a non-resonance scan. The two scanning axes are modulated as a fast axis and a slow axis. The scan is realized by reading consecutive lines on a rectangular area. The scan can be unidirectional or bi-directional. In unidirectional scan, the beam scan across the fast axis and rapidly return to start the next line while in the bi-directional scan, both forward and backward lines are read. More sample data is collected in a bi-directional scan. Raster scan yields the authentic sample section images without post-imaging reconstruction. However, it could not offer large deflection and fast scanning speed as spiral and Lissajous scan. Figure 2.12 shows a raster scan pattern.  27   Figure 2.12 An illustration of the raster scan pattern with a fast axis (𝑿) and slow axis (𝒀). Lissajous and spiral fiber scanners have faster speed than raster fiber scanners in general, but they both require a complicated algorithm for image reconstruction. A raster patterned fiber scanners is a good option for the scanner design in this thesis.  2.3.2.4 Scan Pattern for the Probe Bi-directional raster scanning is used to obtain imaging data. The bi-directional scan is preferred than unidirectional because more data point is collected for image reconstruction. However, the backward and forward scans may not be symmetric due to actuator performance [39]. Figure 2.13 illustrates a scanning scheme. 28   Figure 2.13 An illustration of voltage signals applied onto 𝑿 axis (blue) and 𝒀 axis (red) of scanner probe versus pixel sequence. The blue plot represents the voltage applied onto the 𝑋 (fast) scanning axis versus time sequence, and the red plot is the voltage applied onto the 𝑌 (slow) axis versus time sequence. Data collection is synchronized with the time sequence of the scan signals. The procedures of setting up the optical system, scanner probe, and software control are given in Appendix A and B.   29  Chapter 3: Results and Discussions 3.1 Optical System Characteristics 3.1.1 Lateral Resolution Resolution is an essential characteristic of a microscopy imaging system. High resolution is needed to resolve cell structure. Resolution performance of a confocal microscope at the instrument’s limit is calculated using the Abbe diffraction limit and the Rayleigh criterion. This limit represents the maximum resolution the system could reach without considering any optics imperfections [40]. The theoretical lateral resolution limit is calculated as 0.87 𝜇𝑚 using the following equation [41]:  𝑅𝑙𝑎𝑡𝑒𝑟𝑎𝑙 =0.61𝜆𝑒𝑥𝑐𝑁𝐴𝑜𝑏𝑗 , (3.1) where 𝜆𝑒𝑥𝑐 is the excitation wavelength of the laser (785 𝑛𝑚) and  𝑁𝐴𝑜𝑏𝑗 is the numerical aperture of the objective lens (0.55). Modulation Transfer Function (MTF) of a confocal system could also provide contrast and resolution at a given spatial frequency [42]. The optical performance of the system is predicted and simulated in Zemax optical software (Radiant Zemax LLC). The light fields are simulated at both on-axis (fiber is positioned at the center of the lens couple) and off-axis (fiber is actuated to the maximum deflection) to represent the scanning range. Figure 3.1 illustrates the layout of the optical configuration for the scanner probe simulated in Zemax, and Figure 3.2 shows the MTF plot of the system.  30   Figure 3.1 Two-dimensional layout of the scanner probe optical design. The vertical axis in Figure 3.2 represents the image contrast percentage, and the horizontal axis shows the spatial frequency in line pairs per mm. Resolution is calculated at the spatial frequency where the contrast is 10% of its maximum value [42]. The on-axis and off-axis spatial frequencies are 577 and 559 𝑙𝑝/𝑚𝑚, corresponding to 0.86 𝜇𝑚 and 0.90 𝜇𝑚 resolution. The off-axis position is chosen according to the scanner actuation range.  Figure 3.2 MFT diagram of the confocal scanner probe with C230 objective lens for on-axis and off-axis (100 𝝁𝒎) performances using Zemax. 31  The lateral resolution is also characterized experimentally using edge response by scanning across a series of Ronchi ruling targets (110 𝑙𝑝/𝑚𝑚) along the 𝑋 direction (Thorlabs, R1L1S1P). Lateral resolution for a Gaussian spot may be defined as 0.78 times the distance, where the light intensity changes from 10% to 90% of its maximum values when scanning across a sharp edge [43]:  𝑅𝑙𝑎𝑡𝑒𝑟𝑎𝑙 = 0.78 (𝑋10−90) . (3.2) Figure 3.3(a) shows a graph representation of the Ronchi ruling target, with a line width of 4.55 𝜇𝑚 (110 𝑙𝑝/𝑚𝑚). Figure 3.3(b) shows the intensity profile obtained when scanning the target. The resulting later resolution is 1.16 𝜇𝑚 at the center (on-axis position) of the field of view (FOV) and 1.45 𝜇𝑚 at the edges of the FOV (off-axis position).   Figure 3.3 (a) Image representation of Ronchi ruling target with 110 𝒍𝒑/𝒎𝒎, (b) intensity profile obtained by scanning the target in (a) and (c) the intensity profile scanning at the center and edges of the FOV. 32  The lens couple (collimating lens and objective lens) in the system also has a magnification factor, meaning that the scanning range on the imaging plane is a factor of the actuation range (fiber mechanical movement in 𝑋 and 𝑌 directions) at the light source. Lateral magnification factor for this simple confocal system could be calculated as the effective focal length ratio between the collimating and objective lenses [44]:  𝑀𝑎𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 =𝐹𝐿𝑜𝑏𝑗𝐹𝐿𝑐𝑜𝑙𝑙𝑖𝑚𝑎𝑡𝑒 , (3.3) where the collimating and objective lenses are aspheric lenses with focal lengths of 11 𝑚𝑚 and 4.51 𝑚𝑚, respectively. The system magnification is calculated as 0.41. System magnification could also be predicted from the simulation in Zemax, showing in Figure 3.1. When the light field is positioned at ±100 µ𝑚 , image size in the transverse direction on imaging plane is about ±38 µ𝑚, indicating a magnification factor of 0.38.  3.2 Scanner Mechanical Characteristics The mechanical deflections of the scanners are simulated in COMSOL Multiphysics and measured experimentally using a Laser Doppler Vibrometer (LDV) with the peak to peak displacement of the piezoelectric actuators in both 𝑋 and 𝑌 directions. The actuation range determines the scanning range and further determines the final image size.  3.2.1 Bimorph Scanner Characteristics For piezoelectric actuators, the deformation depends on the magnitude and frequency of the driving voltages. The frequency response of the bimorph scanner after mounting the fiber was 33  predicted by COMSOL Multiphysics and measured using LDV (Figure 3.4). Both bimorph benders are driven under a triangular voltage signal with 74 𝑉𝑝𝑝 to show the frequency response. The predicted first mode resonant frequencies of the beams were 210 𝐻𝑧 and 90 𝐻𝑧 in 𝑋 and 𝑌 directions, respectively. The measured resonant frequencies were 190 𝐻𝑧 and 80 𝐻𝑧, respectively. The measured resonant frequencies are lower than the predictions, mainly due to the mass of the UV adhesive that used to mount the fiber and beams together was ignored in the calculations.  Figure 3.4 Piezoelectric bimorph scanner frequency performances under  𝑽𝒑𝒑 = 𝟕𝟒 𝑽 triangular waveform in both (a) 𝑿 and (b) 𝒀 directions. Figure 3.5 shows the deformation analysis in COMSOL Multiphysics of the scanner by treating the piezoelectric bimorphs as two cantilever beams with stiffness. The fiber mounted on the beam adds stiffness to the whole structure. From the finite element analysis (FEA) in COMSOL Multiphysics, the maximum displacements at the 𝐵2 end are predicted as 742.52 𝜇𝑚 in 𝑌 direction and 446.48𝜇𝑚 in 𝑋 direction for static operation (under maximum constant driving voltage). The cantilever beam has more deflection in 𝑌 direction because of the longer beam length in that direction.  34   Figure 3.5 Deformation distribution of the bimorph scanner (a) in 𝑿 (maximum ~𝟒𝟎𝟎 𝝁𝒎) and (b) in 𝒀 direction (maximum ~𝟕𝟎𝟎 𝝁𝒎) at static operations. Since the displacement of piezoelectric actuators is also determined by the magnitude of the driving voltage, the tip deflection in both 𝑋 and 𝑌 directions with respect to the amplitude of  driving voltages (peak to peak sinusoidal voltage) are also measured by LDV (Figure 3.6) under constant driving frequencies of 200𝐻𝑧 and 1 𝐻𝑧 for 𝑋 and 𝑌 directions.  Figure 3.6  Piezoelectric bimorph scanner characteristics under constant driving frequencies in both (a) 𝑿 (200 𝑯𝒛) and (b) 𝒀 (1 𝑯𝒛) directions. 35  As shown in Figures 3.6, the beam displacement is proportional to the peak to peak driving voltage, so the maximum driving voltage is used to reach large deflection. With the maximum driving voltage of 74 𝑉𝑝𝑝 (specified by the manufacturer) and scanning at 200 𝐻𝑧 and 1 𝐻𝑧 for both 𝑋 and 𝑌  axes, the scanning range is 775 𝜇𝑚  and 364 𝜇𝑚 , respectively (measured by LDV). Raster scanning pattern could be formed with 𝐵1 beam as the fast axis (𝑋 direction) and 𝐵2 beam as the slow axis (𝑌 direction), but the resulting image will have a 2:1-pixel ratio. Extra image processing steps are needed to calibrate the scanned images. Figure 3.7 shows a scan pattern captured by Charge-coupled Device (CCD) camera (BASLER, A102fc-2), with scanning range for 0.775 𝑚𝑚 and 0.729 𝑚𝑚 at 200 𝐻𝑧 and 60 𝐻𝑧 for the two axes under 74 𝑉𝑝𝑝.  Figure 3.7 Scan patterns of the piezoelectric bimorph actuator with 200 𝑯𝒛 in 𝑿 axis and 60 𝑯𝒛 in 𝒀 axis under 74 𝑽𝒑𝒑. 3.2.2 2D Bender Scanner Characteristics The 2D bender scanner deformation is also predicted in COMSOL Multiphysics by treating the actuator as a cantilever beam (Figure 3.8).  36   Figure 3.8 Deformation distribution of the 2D scanner at (a) maximum of  ~𝟐𝟕𝟗. 𝟑𝟗 𝝁𝒎 in 𝒀 and (b) maximum of ~𝟐𝟕𝟑. 𝟎𝟏 𝝁𝒎 in 𝑿 direction under static operations. The maximum displacements are predicted as 279.39 𝜇𝑚 in 𝑋  direction and 273.01 𝜇𝑚  in 𝑌 direction. The 2D scanner has similar properties as the bimorph benders since it is manufactured like two benders put together. Figure 3.9 shows the scanner movement in 𝑋 and 𝑌 directions under triangular modulated voltage ( 60 𝑉𝑝𝑝) measured using LDV and Figure 3.10 shows the deflections under constant frequencies (320 𝐻𝑧 in 𝑋 direction and 1 𝐻𝑧 in 𝑌 direction) with increasing 𝑉𝑝𝑝 in the driving signals.   Figure 3.9 Piezoelectric 2D scanner frequency performances under triangular modulated voltage (𝑽𝒑𝒑 = 60 𝑽) in both (a) 𝑿 and (b) 𝒀 directions. 37   Figure 3.10 Piezoelectric 2D scanner characteristics under constant frequencies in both (a) 𝑿 (320 𝑯𝒛) and (b) 𝒀 (1 𝑯𝒛) directions. Both 𝑋 and  𝑌 axes have resonance frequencies around 550 𝐻𝑧 and 490 𝐻𝑧, respectively, and the beam deflection is proportional to the magnitude of the driving voltages. At the maximum driving voltage (60 𝑉𝑝𝑝 ) and a frequency of 320 𝐻𝑧  and 1 𝐻𝑧  in both 𝑋  and 𝑌  directions, the displacements are measured as 201 𝜇𝑚 and 204 𝜇𝑚. If the 2D scanner is driven at this condition, the resulting scan range is square, indicating a 1:1-pixel ratio and no further pixel calibration is needed.   3.3 Imaging Results 3.3.1 Real-Time Image Acquisition All actuators in the scanner probe present a response delay with the driving signals. As an illustration, Figure 3.11 shows the 2D bender movements in 𝑋 direction under a triangular wave with a frequency of 490 𝐻𝑧 at 60 𝑉𝑝𝑝.  38   Figure 3.11 Phase delay in 2D bender with the driving signal in blue and the actuator response in red. The driving signal was plotted in blue, and the actuator response was plotted in red. As seen, the actuator presents a response delay of ~1290 𝜇𝑠. As an example, if the signal is used for a raster scan with image acquisition speed at 2 frames per second with 512 × 512 pixels, the 1290 𝜇𝑠 delay will cause 2460 pixel delays in the imaging data. It is calculated using the following equation:  𝑃𝑖𝑥𝑒𝑙 𝐷𝑒𝑙𝑎𝑦 𝑁𝑢𝑚𝑏𝑒𝑟 =  𝐷𝑒𝑙𝑎𝑦 𝑇𝑖𝑚𝑒𝑓𝑝𝑠 × 𝑡𝑜𝑡𝑎𝑙 𝑝𝑖𝑥𝑒𝑙𝑠 . (3.4) An illustration of the pixel delay caused by the response delay of the actuator is shown in Figure 3.12. The blue color represents the useful imaging data, while the red-colored data contains useless information and is collected before the actuator responded to the driving signals.  Figure 3.12 Pixel delay of the imaging data due to response delay of the actuator (blue for data and red for non-useful data signals). 39  The control scheme is adjusted to have a fast axis pre-ramp, allowing extra time for the actuator to response (Figure 3.13). A series of control signals are also added to the slow axis to let it return to the origin after a frame scan and get ready for generating the next frame of a image. In image reconstruction, the pixel delay data and the pre-ramp data are deleted. Figure 3.14 shows a pair of images before and after the pixel delay correction.   Figure 3.13 Voltage applied onto 𝑿 (fast) axis and 𝒀 (slow) axis of the scanner probe versus pixel sequence with pixel delay correction. 40   Figure 3.14 Image of a resolution target (a) before and (b) after pixel delay correction. As seen in Figure 3.14, after the phase delay correction, there is still an image overlap. The overlap is mainly due to the hysteresis presented by the piezoelectric actuators. The actuator has a slightly different path for the forward and backward lines in the bi-directional raster scan. The scanner is operated in open-loop now, and the hysteresis might be corrected by design a closed-loop control system for the scanner. The actuator also response non-linear at the edge of the scanning range, leading image distortion at the circumference. The stretch in the left part of the image shown in Figure 3.14 (b) is an indication of nonlinearity response of the 2D actuator. Nonlinearity could be solved by post-image processing in the future or with a closed-loop controller.  41  3.3.2 High Reflective Index Surfaces Horizontal scanning is performed by actuating the 2D scanner in both 𝑋 and 𝑌 directions in a raster pattern with a frame rate of 1.25 fps (roughly 320 𝐻𝑧 in 𝑋 direction and 1 𝐻𝑧 in 𝑌 direction) with 512 × 512 pixels using the all-fiber system setup. This setup is the most stable one and yields better images among all other combinations. The bimorph scanner results non-square pixels thus is not used for imaging. The specific frame rate is selected because, at other frame rates, the pixel delay in both scanning axes could not be corrected by merely deleting the delayed pixels since hysteresis plays an essential part in the scanning. The scan area is also square at this scanning speed, indicating there is no need to correct pixel ratio distortion. The scan area is approximately 200 × 200 𝜇𝑚2, which results in an 80 × 80 𝜇𝑚2 image size due to the magnification of the lens couple. The scanner could be operated at higher scan frequencies like 600 𝐻𝑧 in fast axis and 5 𝐻𝑧 for slow axis (5 fps), where the deflection is still good enough for imaging (~170 𝜇𝑚), but extensive post-imaging processing is needed to correct the pixel delays and hysteresis presented at that speed. Imaging is first conducted with high reflective surfaces like a mirror (Thorlabs, PF10-03-F01) to show the feasibility of the system (Figure 3.15).  42   Figure 3.15 Horizontal scan of a mirror using the 2D bender scanner and the all-fiber confocal system. For a mirror scanned with a confocal system, it will have a bright center and dark outside circumference, since the reflection is highest at the on-axis position and becomes lower at off-axis positions. In Figure 3.15, the right middle plane indicates the brightest area, and the lower-left area is the off-axis dark region. There is a slight miss alignment between the fiber tip and the optical axis of the lens couple. The black patterns in the image indicate scratches on mirror.   A glass slide deposited with a thin aluminum film is used as the second sample. On the aluminum deposited side, a line is cut with laser machining equipment (New Wave Research, QuikLaze 50ST2) to expose the glass base. Light will transmit in the glass but will be reflected by aluminum. Figure 3.16 shows the resulting image. A dark line is seen in the figure since the light transmission in glass (low intensity in data signal), and the bright area shows the high-intensity signals reflected by the aluminum film.  43   Figure 3.16 Horizontal scan of aluminum deposited glass slide using the 2D bender scanner and the all-fiber confocal system. Both mirror and glass slide images show the feasibility of the whole unit.  3.3.3 Resolution Targets The imaging is then conducted on a standard U.S.Air Force (USAF) resolution target (Thorlabs, R1L1S1P). Figure 3.17 shows the image of the elements in group 7 of the target.  44   Figure 3.17 Horizontal scan of a USAF resolution target group 7 for (a) element 1-3 and (b) element 3-6. The smallest in group 7 has a resolution of  227 𝑙𝑝/𝑚𝑚, meaning the line in this feature is 2.20 𝜇𝑚 wide and 11.01 𝜇𝑚  long. With the standard resolution target, reference relationship between scanner actuation range and actual image size can be estimated. The image size is about 88 𝜇𝑚 in both directions, which roughly matches the predicted 80 𝜇𝑚2 imaging area. The actual image area is slightly larger since the fiber mounted on the beam have a small hanging length (~1𝑚𝑚) due to the fabrication process. This image also indicates that the resolution of the system is better than 2.2 𝜇𝑚 since group 7, element 6 feature could be resolved.  Figure 3.18 shows an image of a 10 𝜇𝑚 grid on the resolution target. From the image, there is significant distortion at the left edge due to the non-linear response of the scanner. If there is large nonlinearity at the edge, the image data is not evenly sampled within the whole scan area. Data is densely sampled at the edge, leading a longer grid line or a stretch in the figure. On the one hand, further studies need to be done to solve the problem by either linearized the actuator using control 45  circuit or post-image processing techniques. On the other hand, If the mechanical scanning range could be further improved, the nonlinear portion could be deleted rather than adding control techniques and sensors to make the scanner system complicated.   Figure 3.18 Horizontal scan of a 10 𝝁𝒎 grid on the resolution card. 3.3.4 Optical Paper and Onion Epidermis The horizontal scan was also performed on optical paper and onion epidermis to illustrate the feasibility to scan biological samples. Figure 3.19 shows the image of an optical paper scanned by the developed fiber actuator and an image scanned by a desktop confocal system using galvanometer mirrors. 46   Figure 3.19 Paper image (a) using developed 2D bender scanner and all-fiber confocal system and (b) using desktop galvanometer mirror confocal system. Figure 3.20 shows the image of a piece of onion epidermis scanned by the developed fiber actuator and an image scanned by a desktop confocal system.   Figure 3.20 Onion epidermis image (a) using developed 2D bender scanner and all-fiber confocal system and (b) using desktop galvanometer mirror confocal system. Similar cellular structures like paper fiber and cell walls are identifiable in the scanned image using the developed system. The image taken by standard RCM system appears to be brighter because 47  it collected more reflected signal light. The image range is also larger than the all-fiber system.  The circulator in the all-fiber system decreases the light intensity slightly due to the attenuation in the device itself. The contrast of the images from the developed system could be improved with the histogram equalization technique.  3.3.5 Background Patterns in Images There is also slight background noise and light interference pattern on the current images, and it is illustrated in Figure 3.21. The pattern will also change slightly in every run of the system.   Figure 3.21 Background noise pattern in images. Source of the noise may come from the optical components and the scanner probe in the system. For any fiber system, there is always intrinsic thermal noise presented [45]. A possible solution for reducing thermal noise is to change the coating of the fiber. Other backgrounds may come from the reflection noise at the fiber end face and fiber coupler [46]. The fiber is cleaved using a cleaver at 0 degrees. Due to the cleaving process, the surface of the fiber end is not always pure flat, even with the cleaning and polishing procedures, leading to reflection noise. An angle 48  cleave may reduce the reflection noise from fiber end, but it will require high precision cleaving process. The photodiode used in the setup brings sensor noise as well. However, if the SNR is high enough, the sensor noise does not influence the image quality much.  49  Chapter 4: Conclusion and Future Work 4.1 Summary In Chapter 1, an overview of the background information on colorectal cancer, and current detection methods with their side-effects were provided. Different non-invasive optical technologies for tissue imaging were also compared. Among them, OCT has the most considerable penetration depth, but lower resolution compared with others. MCM requires demanding optics and FCM need external fluorescence agents. RCM could provide enough penetration depth and high resolution for tissue imaging. Different types of miniaturized scanners were also reviewed. Piezoelectric fiber scanners have the most compact size, fast scanning speed, and low power consumption and are more suitable for endo-microscope scanner probe designs. Research objectives and thesis structure were also described. The research is aimed at designing a portable imaging system that could obtain tissue images at a cellular level.  In Chapter 2, the optical design of a reflectance confocal system in endomicroscopy applications was demonstrated in two ways. The desktop setup has a small loss in signal data but occupied a considerable amount of space and presented alignment issues. The all-fiber system, however, is easy to operate and are more stable. The size of the all-fiber system is also smaller, making it more portable than the desktop system. The design and fabrication of the scanner probes were also demonstrated. Two types of piezoelectric actuators were used in the probe design. The piezoelectric bimorphs offer massive movement with the sacrifice in the total rigid length of the probe. The 2D bender type has less deflection but is short and easy to align. The methods used for image processing, scan patterns generation, and actuator control were also discussed.  50  In Chapter 3, the optical performances of the imaging system were demonstrated as imaging range and image resolution. The scanners performances were discussed later with deflection and frequencies analysis with respect to different driving voltages. Imaging results in the horizontal plane were also shown using a mirror, a piece of onion epidermis, an optical paper, and resolution targets. The observed images were comparable to the existing confocal scanning system. In conclusion, a prototype of a real-time reflectance confocal piezoelectric fiber scanner with raster scan pattern for horizontal scanning of 88 𝜇𝑚 ×  88  𝜇𝑚  sizes at 1.25 frames per second is developed. The resolution of the system is within 1.16-1.45 𝜇𝑚 with a verification using USAF 1951 targets. Imaging of biological tissues was verified with a piece of onion epidermis and an optical paper. The developed system shows the potential to provide horizontal optical sectioning in tissues in a portable setting. The size of the probe is larger than the instrument channel size of a clinical endoscope due to the lenses used in the current design. However, the developed probe could be used in other applications like skin cancer detection. The probe now is more like a handheld and portable microscope for high-resolution images.  4.2 Future Works The 2D scanner is not running at the achievable fastest speed (5 fps, the speed at which the actuator deflection is still enough for imaging). Imaging speed could be further improved with phase correction, hysteresis correction, and data re-mapping in the current design. However, the size of the scanner probe could be further miniaturized using customized piezoelectric products and miniaturized lenses. Smaller sized piezoelectric benders or tubes have less displacement in general. Fiber could be mounted on the actuator with additional free length to form a fiber cantilever 51  structure instead to achieve the large deflection required by imaging. In this scenario, the resonance mode of the optical fiber could be used to analyze the cantilever displacement. Imaging speed could also be improved since a fiber cantilever usually has a resonance frequency around kilohertz, leading to higher scanning frame rate and imaging speed. There is also a background pattern on the current images. These noises may come from the optical components in the system. The image quality may be improved with either better optics or post-imaging processing techniques. Images also appear brighter on the right corner and darker on the top left, and patterns at the edges of the images are distorted due to the non-linearity in the piezoelectric actuators. Suitable image processing technologies may be applied to filter and even out the effects. A control algorism could also be applied to make actuators performance linear and resolve hysteresis issues, which may require adding additional displacement sensors. Furthermore, generally speaking, vertical section images could provide more useful information in cancer detections. The scanner probe may be able to combine with a 𝑍-axis driven system and be packaged to a catheter to obtain real-time 3D images.  52  Bibliography [1] F. Bray, J. Ferlay, I. Soerjomataram, R. L. Siegel, L. A. Torre, and A. Jemal, “Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.,” A Cancer Journal for Clinicians, vol. 68, no. 6, pp. 394–424, 2018. [2] Canadian Cancer Statistics Advisory Committee, “Canadian Cancer Statistics 2018,” Canadian Cancer Society, 2018. Available at: cancer.ca/Canadian-Cancer-Statistics-2018-EN. [3] F. A. Haggar and R. P. Boushey, “Colorectal cancer epidemiology: incidence, mortality, survival, and risk factors,” Clinics in Colon and Rectal Surgery, vol. 22, no. 4, pp. 191–197, 2009. [4] D. Moussata et al., “Are random biopsies still useful for the detection of neoplasia in patients with IBD undergoing surveillance colonoscopy with chromoendoscopy?,” Gut, vol. 67, no. 4, pp. 616–624, 2018. [5] S. E. Kudo, S. Tamura, T. Nakajima, H. O. Yamano, H. Kusaka, and H. Watanabe, “Diagnosis of colorectal tumorous lesions by magnifying endoscopy,” Gastrointestinal Endoscopy, vol. 44, no. 1, pp. 8–14, 1996. [6] M. Tahara et al., “Cell diameter measurements obtained with a handheld cell counter could be used as a surrogate marker of G2/M arrest and apoptosis in colon cancer cell lines exposed to SN-38,” Biochemical and Biophysical Research Communications, vol. 434, no. 4, pp. 753–759, 2013. [7] A. M. Zysk, F. T. Nguyen, A. L. Oldenburg, D. L. Marks, and S. A. Boppart, “Optical 53  coherence tomography: a review of clinical development from bench to bedside,” Journal of Biomedical Optics, vol. 12, no. 5, p. 051403, 2007. [8] Y. Kim et al., “Adaptive multiphoton endomicroscope incorporating a polarization-maintaining multicore optical fibre,” IEEE Journal of Selected Topics in Quantum Electronics, vol. 22, no. 3, pp. 171–178, 2015. [9] J. M. Jabbour, M. A. Saldua, J. N. Bixler, and K. C. Maitland, “Confocal endomicroscopy: instrumentation and medical applications,” Annals of biomedical engineering, vol. 40, no. 2, pp. 378–397, 2012. [10] S. Adabi, Z. Turani, E. Fatemizadeh, A. Clayton, and M. Nasiriavanaki, “Optical coherence tomography technology and quality improvement methods for optical coherence tomography images of skin: a short review,” Biomedical Engineering and Computational Biology, vol. 8, 2017. [11] L. Thiberville, S. Moreno-Swirc, T. Vercauteren, E. Peltier, C. Cavé, and G. B. Heckly, “In vivo imaging of the bronchial wall microstructure using fibered confocal fluorescence microscopy,” American Journal of Respiratory and Critical Care Medicine, vol. 175, no. 1, pp. 22–31, 2007. [12] W. Drexler and J. G. Fujimoto, Eds., Optical Coherence Tomography, 2nd ed. Springer Cham, 2015. [13] H. Ra, W. Piyawattanametha, Y. Taguchi, D. Lee, M. J. Mandella, and O. Solgaard, “Two-dimensional MEMS scanner for dual-axes confocal microscopy,” Journal of Microelectromechanical Systems, vol. 16, no. 4, pp. 969–976, 2007. [14] H. C. Park, C. Song, M. Kang, Y. Jeong, and K. H. Jeong, “Forward imaging OCT endoscopic catheter based on MEMS lens scanning,” Optics Letters, vol. 37, no. 13, pp. 54  2673–2675, 2012. [15] C. M. Lee, C. J. Engelbrecht, T. D. Soper, F. Helmchen, and E. J. Seibel, “Scanning fiber endoscopy with highly flexible, 1-mm catheterscopes for wide-field, full-color imaging,” Journal of Biophotonics, vol. 3, no. 5–6, pp. 385–407, 2010. [16] K. Hwang, Y. H. Seo, and K. H. Jeong, “Microscanners for optical endomicroscopic applications,” Micro and Nano Systems Letters, vol. 5, no. 1, pp. 1–11, 2017. [17] P. R. Patterson, D. Hah, H. Nguyen, H. Toshiyoshi, R. M. Chao, and M. C. Wu, “A scanning micromirror with angular comb drive actuation,” in Tech.Dig. 15th IEEE Int. Conf. Micro Electro Mechanical Systems (Cat. No.02CH37266), Las Vegas, Nevada, USA, pp. 544–547. [18] G. Cao, H. Mansoor, H. Zeng, I. T. Tai, and M. Chiao, “Real-time vertical / horizontal section imaging and 3D image construction using a 3-axis electromagnetic confocal microscanner,” in 19th Int. Conf. on Solid-State Sensors, Actuators, and Microsystems (TRANSDUCERS), Kaohsiung, Taiwan, 2017, pp. 254–257. [19] T. Wu, Z. Ding, K. Wang, M. Chen, and C. Wang, “Two-dimensional scanning realized by an asymmetry fiber cantilever driven by single piezo bender actuator for optical coherence tomography,” Optics Express, vol. 17, no. 16, pp. 13819–13829, 2009. [20] Y. H. Seo, K. Hwang, H. C. Park, and K. H. Jeong, “Electrothermal MEMS fiber scanner for optical endomicroscopy,” Optics Express, vol. 24, no. 4, pp. 3903–3909, 2016. [21] H. Mansoor, H. Zeng, I. T. Tai, J. Zhao, and M. Chiao, “A handheld electromagnetically actuated fiber optic raster scanner for reflectance confocal imaging of biological tissues,” IEEE Transactions on Biomedical Engineering, vol. 60, no. 5, pp. 1431–1438, 2012. [22] Y. B. Gianchandani, L. Que, and J. S. Park, “Bent-beam electrothermal actuators-Part I: 55  Single beam and cascaded devices,” Journal of Microelectromechanical Systems, vol. 10, no. 2, pp. 247–254, 2001. [23] D. R. Rivera et al., “Compact and flexible raster scanning multiphoton endoscope capable of imaging unstained tissue,” Proceedings of the National Academy of Sciences, vol. 108, no. 43, pp. 17598–17603, 2011. [24] H. Schulz-Hildebrandt et al., “High-speed fiber scanning endoscope for volumetric multi-megahertz optical coherence tomography,” Optics Letters, vol. 43, no. 18, pp. 4386–4389, 2018. [25] ASGE Technology Committee et al., “GI endoscopes,” Gastrointestinal Endoscopy, vol. 74, no. 6, pp. 1-6.e6, 2011. [26] K. Chen, “A vertical and horizontal 3-axis hand-held confocal scanner for skin imaging applications,” Master Thesis, University of British Columbia, 2015. [27] J. Wang et al., “A Confocal Endoscope for Cellular Imaging,” Engineering, vol. 1, no. 3, pp. 351–360, 2015. [28] M. H. Niemz, Laser-Tissue Interactions, 4th ed. Springer Nature Switzerland, 2019. [29] J. E. Greivenkamp, Field Guide to Geometrical Optics. SPIE Press, 2004. [30] “Technical note: fiber optic coupling,” Newport. [Online]. Available: https://www.newport.com/n/fiber-optic-coupling. [Accessed: 19-Aug-2018]. [31] D. Keller, C. Bustamante, M. F. Maestre, and I. Tinoco, “Imaging of optically active biological structures by use of circularly polarized light,” Proceedings of the National Academy of Sciences, vol. 82, no. 2, pp. 401–405, 1985. [32] S. P. Morgan and M. E. Ridgway, “Polarization properties of light backscattered from a two layer scattering medium,” Optics Express, vol. 7, no. 12, pp. 395–402, 2000. 56  [33] J. M. Zavislan, “Imaging system using polarization effects to enhance image quality,” United States Patent: 6,134,010, Oct.17, 2000. [34] R. B. Meyer, “Piezoelectric effects in liquid crystals,” Physical Review Letters, vol. 22, no. 18, p. 918, 1969. [35] S. Jiang, X. Li, S. Guo, Y. Hu, J. Yang, and Q. Jiang, “Performance of a piezoelectric bimorph for scavenging vibration energy,” Smart Materials and Structures, vol. 14, no. 4, pp. 769–774, 2005. [36] S. A. Rios and A. J. Fleming, “A novel electrical configuration for three wire piezoelectric bimorph micro-positioners,” in 2014 IEEE/ASME Int. Conf. Advanced Intelligent Mechatronics, Besançon, France, pp. 1452–1457. [37] Q. M. Wang and L. E. Cross, “Constitutive equations of symmetrical triple layer piezoelectric benders,” IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, vol. 46, no. 6, pp. 1343–1351, 1999. [38] K. Murari, W. Liang, Y. Zhang, J. Xi, and X. Li, “Lissajous scanning fiber-optic nonlinear endomicroscope with precise position calibration,” in Biomedical Optics and 3-D Imaging, OSA Technical Digest (Optical Society of America, 2012), p. BSu3A.36. [39] P. Xi, Y. Liu, and Q. Re, “Scanning and image reconstruction techniques in confocal laser scanning microscopy,” in Laser Scanning, Theory and Applications, 2011, p. 27. [40] R. W. Cole, T. Jinadasa, and C. M. Brown, “Measuring and interpreting point spread functions to determine confocal microscope resolution and ensure quality control,” Nature Protocols, vol. 6, no. 12, pp. 1929–1941, 2011. [41] J. B. Pawley, Ed., Handbook of biological confocal microscopy, 3rd ed. Springer LLC, 2010. 57  [42] G. D. Boreman, Modulation Transfer Function in Optical and Electro-Optical Systems. SPIE Press, 2001. [43] M. Rajadhyaksha, R. R. Anderson, and R. H. Webb, “Video-rate confocal scanning laser microscope for imaging human tissues in vivo,” Applied Optics, vol. 38, no. 10, pp. 2105–2115, 1999. [44] L. S. Pedrotti, “Basic Geometrical Optics,” 2008. [Online]. Available: https://spie.org/Documents/Publications/00 STEP Module 03.pdf. [45] J. Dong, J. Huang, T. Li, and L. Liu, “Observation of Fundamental Thermal Noise in Optical Fibers down to Infrasonic Frequencies,” Applied Physics Letters, vol. 108, no. 2, p. 021108, 2016. [46] A. Inoue, R. Furukawa, M. Matsuura, and Y. Koike, “Significant Noise Reduction in Multimode Fiber Links Using Graded-index Plastic Optical Fiber with Microscopic Heterogeneous Core,” in 39th European Conf. and Exhibition on Optical Communication (ECOC 2013), London, UK, 2013, pp. 843–845.  58  Appendices Appendix A   Operating Procedures for Imaging System Turn-on procedure for the system Step 1 Connect cables and wires. • For the 2D piezoelectric actuator, a bias voltage (60 𝑉) need to be applied at the positive and negative nodes.  The bias comes from the 𝑍 axis output of the open-loop piezoelectric controller.  • Connect two output ports (defaults are AO 0&1, which are pin 15&16, and pin 31&32 pairs) on the DAQ card to the piezoelectric controller 𝑋 and 𝑌 axis external inputs. • Connect the input port (default is AI 0, pin 1&2 pair) of the DAQ card to the APD. Step 2 Connect power supplies. Connect the power supplies for all units used in the system, including: • NI-6363 DAQ Card; • Open-loop piezoelectric controller (Thorlabs, MDT693B); • APD (OPEAK, PD-M-TBPIN-SW) and there is a separate switch on the power regulator and remember to turn it on; • Laser source. Step 3 Turn on the laser source. • Switch the control from Off to Standby. • Wait until the indicators on the control board switch from one to two. • Check the light path, especially the circulator connection, to make sure the path is right, and there is no laser light leakage.  59  • Switch the control from Standby to On, and there should be four indicators on. Do not look at the laser directly at eye level.  Step 4 Check the inputting laser power. • Check at the probe side first. For samples like resolution card, the input power should be less than 3 𝑚𝑊 to protect the card features. For samples like onion epidermis and optical paper, 3-5 𝑚𝑊 is good enough to show a clear image, but higher power would yield better contrast. For skins, if possible, the power could go up to 35 𝑚𝑊.  • If the power is not enough, the optical path needs to be re-aligned. For the all-fiber system, the only part that needs an alignment is laser fiber coupler. Disconnect the circulator and the probe and connect a multimode fiber to the coupler.  • Coarsely positions the coupler until the light is coming out from the distal end of the multimode fiber. Adjust the coupler mounting stage nobs to make it away from the focal plane. Place a power meter at the distal end. Use the stage nobs in 𝑋, 𝑌, and 𝑍 directions and adjusting them in turns until the power at the distal end reaches the maximum level (~220 𝑚𝑊).  • Switch the multimode fiber back to SMF or circulator Port 1 fiber. Check at Port 2 of the circulator for power. If the desired value (higher than 5 𝑚𝑊) reaches, connect the probe. If not, repeat the positioning steps. Once desired power is achieved, connect the probe.  Step 5 Turn on NI-6363 DAQ Card. Step 6 Open LabView program. Step 7 Turn on open-loop piezoelectric controller. Step 8 Start imaging. • Use the default settings for the LabVIEW program.  60  • When imaging, samples need to be placed at the focal plane of the objective lens. Currently, a stage is used to position the samples relative to the objective lens rather than moving the objective lens. The focal length of the C230 objective lens is around 4.5 𝑚𝑚, so try to position the samples accordingly.  • The maximum value on the LabView control panel changes the contrast level in images and could be changed during the imaging process according to the intensity level for the current image. If the background noise is too high (an SNR ratio smaller than 0.5), the fiber tip might need to be re-cut, or the fiber is not positioned at the focal plane of the collimating lens. Steps for cutting fiber and positioning the fiber tip with respect to the collimating lens will be provided later.   Turn-off procedure for the system Step 1 Turn off the LabView program. Step 2 Turn off the piezoelectric controller. Step 3 Turn of the APD power switch. Step 4 Turn of the laser control from On to Standby and then to Off.  Step 5 Turn of the DAQ card.  Steps for cleave fiber If the SNR is low on images or if there is laser leakage at the fiber end, the fiber needs re-cut.  • Remove the fiber jacket using fiber stripper. • Remove the coating using fiber stripper. • Wipe the stripped fiber with methanol if available to thoroughly remove the coating. 61  • Use the fiber cleaver to cut fiber. Press the cleave as fast as possible to have a clean cut. • Clean the fiber cut end if necessary.  Steps for position fiber distal end with respect to the collimating lens • With laser on but lower power, use the cage nobs to move the fiber to the optical axis of the collimating lens. From the other side of the lens, a round spot should be seen if it has been appropriately positioned.  • Check the light spot at a long distance from the collimating lens and slide the fiber cage along the axis until the size of the light spot is consistent at all distance, meaning parallel light is coming out from the collimating lens.   • Secure the position.   62  Appendix B  LabVIEW program for scanner control and image acquisition    63      64      

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0380521/manifest

Comment

Related Items