CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH …CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH...

CONFOCAL MICROSCOPY - LINE SCANNING SETUP

WITH HIGH BRIGHTNESS LED

A Thesis Presented

by

Ali Vakili

to

Department of Electrical and Computer Engineering

in partial fulfillment of the requirements

for the degree of

Master of Science

in

Electrical Engineering

Northeastern University

Boston, Massachusetts

April 2014

NORTHEASTERN UNIVERSITY

Graduate School of Engineering

Thesis Title: Confocal Microscopy - Line Scanning Setup with High Brightness

LED.

Author: Ali Vakili.

Department: Electrical and Computer Engineering.

Approved for Thesis Requirements of the Master of Science Degree:

Thesis Supervisor: Prof. Charles DiMarzio Date

Thesis committee: Dr. Milind Rajadhyaksha Date

Thesis Committee: Prof. Mark Niedre Date

Department Chair: Prof. Sheila S. Hemami Date

Director of the Graduate School: Prof. Sara Wadia-Fascetti Date

Acknowledgments

I would like to express my special appreciation and thanks to my advisor Professor

Charles A. DiMarzio, who has been a tremendous mentor for me and gave me the

opportunity to work under his supervision in Optical Science Laboratory. I also want

to thank Dr. Milind Rajadhyaksha who co-advised me during this research. Also I

want to express my appreciation towards Professor Mark Niedre and Dr. Milind Ra-

jadhyaksha for serving on my thesis committee. I would like to thank all my friends

and colleagues in optical science laboratory for all their support and contributions.

Many thanks to Joseph Hollmann for his contribution in the projects with his great

ideas and amazing MATLAB skills and also, Ali Golabchi for providing many con-

structive feedback on this thesis. At the end, I would like to thank my family and all

my friends who always supported me.

iii

Contents

Acknowledgments iii

1 Introduction 1

1.1 Optical Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Optical Sectioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3 Confocal Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.3.1 Point Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.3.2 Line Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.4 Components in Confocal Microscope . . . . . . . . . . . . . . . . . . 16

1.4.1 Pinhole or Slit . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1.4.2 Beam Splitter . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.4.3 Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2 Light Sources 25

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.1.1 Brightness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.1.2 Wavelength . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.1.3 Coherence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.2 Laser Light Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

iv

2.2.1 Introduction to Laser . . . . . . . . . . . . . . . . . . . . . . . 30

2.2.2 Laser Diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.3 Non-Laser Light Sources . . . . . . . . . . . . . . . . . . . . . . . . . 36

2.3.1 Arc Lamps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

2.3.2 Light Emitting Diode . . . . . . . . . . . . . . . . . . . . . . . 37

2.3.3 High Brightness LED . . . . . . . . . . . . . . . . . . . . . . . 39

2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3 Scattering Medium and Pupil Configurations 40

3.1 Scattering Media and Gating Mechanism . . . . . . . . . . . . . . . . 40

3.2 Pupil Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3.2.1 Full Pupil Configuration . . . . . . . . . . . . . . . . . . . . . 44

3.2.2 Half Pupil Configuration . . . . . . . . . . . . . . . . . . . . . 46

3.2.3 Divided Pupil Configuration . . . . . . . . . . . . . . . . . . . 46

3.2.4 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4 Experiments and Results 50

4.1 Dual Wedge Confocal Microscope . . . . . . . . . . . . . . . . . . . . 51

4.1.1 Dual Wedge Scanner . . . . . . . . . . . . . . . . . . . . . . . 51

4.1.2 Fluorescence Modes . . . . . . . . . . . . . . . . . . . . . . . . 54

4.1.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

4.1.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

4.2 Line Scanning Confocal Microscope . . . . . . . . . . . . . . . . . . . 61

4.2.1 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . 62

4.2.2 Resolution Measurement . . . . . . . . . . . . . . . . . . . . . 64

4.2.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

v

4.3 High Brightness LED as Light Source . . . . . . . . . . . . . . . . . . 68

4.3.1 LED Specifications . . . . . . . . . . . . . . . . . . . . . . . . 69

4.3.2 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4.3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

5 Conclusion 76

6 Future Work 79

6.1 Quantify Multiply Scattered Photons Rejection . . . . . . . . . . . . 79

6.2 Image Larger Area with Dual-Line Scanning . . . . . . . . . . . . . . 80

6.3 Laser-Driven Light Sources . . . . . . . . . . . . . . . . . . . . . . . . 82

Bibliography 83

vi

List of Figures

1.1 Concept of numerical aperture. a) specimen is illuminated by a collimated beam,

then numerical aperture is defined as, NA = nsin(α). b) A condenser has a

numerical aperture the same as that of the objective lens. In this case, working

aperture is the sum of numerical aperture of objective lens and condenser, therefore

NA = nsin(2α). Figure taken from ZEISS microscopy online campus . . . . . . 4

1.2 Point spread function and Airy disk definition. a) Airy disk pattern generated form

light diffracted in specimen. b) 3D representations of the diffraction pattern on the

image plane, known as the point-spread function. c) An Airy disk is the region

enclosed by the first minimum of the Airy pattern and contains approximately

84% of the energy. Figure taken from ZEISS microscopy online campus . . . . . 5

1.3 Two Airy disks close to each other. dmin is the minimum distance that each of

the two Airy disks can be resolved by the human eye. Figure taken from ZEISS

microscopy online campus . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.4 Rayleigh criterion, two Airy disks are considered resolvable if the valley between the

peaks is about 20%-30% of the maximum. Figure taken from Nikon Instruments,

Inc. website . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

vii

1.5 Illumination method comparison between conventional and confocal microscopes.

In wide-field illumination, a large volume of the specimen is illuminated. However

in confocal microscopy, a very small volume of the specimen is exposed [1]. . . . 9

1.6 Comparison of wide-field(upper row) and confocal microscopy(lower row). (a) and

(b) mouse brain hippocampus thick section. (c) and (d) Thick section of rat smooth

muscle. (e) and (f) Sunflower pollen grain [1] . . . . . . . . . . . . . . . . . 10

1.7 Typical confocal microscope. Excitation light is directed to the sample by a dichroic

mirror or beam splitter. Two galvanometer mirror scan the beam on the sample in

two direction and objective lens, focuses the beam onto a diffraction limited spot

on specimen. Light coming back from the specimen, travels the same path to the

dichroic mirror which transmit the beam to the confocal pinhole and the detector.

A/D which is synchronized with scanner is then used to collect information and

store in the computer. Image is then reconstructed in computer and is displayed

on the monitor [1]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.8 Rotating polygon mirror used as scanner. A motor rotates the mirror and contin-

uously scans the beam on the specimen. Unlike galvo scanner, it is not needed to

return the mirror to the initial position. Therefore it is easier to synchronize the

scanner with data acquisition board. Backscattered light returns the same path to

the polygon mirror. Figure taken from LEYBOLD Photonics educational kit website 13

1.9 The axial response for (a) point-scan and (b) line-scan for different pupil configu-

ration. [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1.10 Images of human epidermis in vivo. Scale bars, 50 µm. (a) the stratum corneum

(SC), granular (GR), and spinous (SP) cells. (b) smaller basal cells with dark nuclei

(arrows), arranged in ring-shaped clusters (arrowheads) [3]. . . . . . . . . . . . 15

1.11 Cylindrical Lens focuses the beam onto a line. . . . . . . . . . . . . . . . . . 16

viii

1.12 Field plane and pupil planes for coherent and incoherent line of light source. In

the case of coherent light source, in both image and pupil planes light is focused

on a line, while in the case of incoherent light sources, source can be separated as

many point sources. Therefore, in field planes, light is focused on a line, but in

pupil planes light is not focused. . . . . . . . . . . . . . . . . . . . . . . . 17

1.13 Typical slit aperture. Only a small portion of light goes through the aperture. It

generates a line illumination on the other side. . . . . . . . . . . . . . . . . . 18

1.14 Diagram of pinhole application in confocal microscope. Pinhole or confocal detec-

tor aperture rejects light coming from below or above the focal plane therefore it

improves contrast and resolution. . . . . . . . . . . . . . . . . . . . . . . . 19

1.15 Typical beam splitters. Flat beam splitter on the left. It splits the incident

beam(from bottom) into two beams. On the right, a cube beam splitter which

splits the incident beam into half.It transmits half and reflects the other half. . . 20

1.16 Polarizing beam splitter. incident beam(black) is not polarized. The beam splitter

will splits the incident beam into vertical(transmission) and horizontal(reflection)

polarization. Figure taken from Thorlabs Inc. . . . . . . . . . . . . . . . . . 21

1.17 Diagram of a photomultiplier tube(PMT). Incident photon excites photo-cathode to

generate primary electron. Primary electron then excite dynodes to create multiple

electrons in a cascade process. At the end of the tube, anode receives electrons and

generates output signal. Figure taken from scintillator materials group at Stanford

University . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.18 Comparison of different detectors spectral sensitivity. . . . . . . . . . . . . . . 23

ix

2.1 Radiance in an Image. The image area is related to the object area by A2 = mA1,

and the solid angles are related by Ω2 = Ω1

m2 ( nn′ )

2. Where m is the magnification

and n and n′ are the index of refraction of materials on two sides of the imaging

system, respectively [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.2 The emission profile of mercury arc lamp. Wavelengths in the visible region are use-

ful in fluorescence microscopy. Emission profile of the xenon arc lamp is presented.

Xenon arc lamp emission is almost flat in visible region [5]. . . . . . . . . . . . 28

2.3 Spatial coherence and temporal coherence. An incoherent light source emits light

in all direction with different wavelengths. Waves that pass the pinhole aperture

are spatially coherent, but in order to make a coherent beam, a wavelength filter

is required. Figure taken from ZEISS microscopy online campus . . . . . . . . 29

2.4 Laser cavity and lasing process. 1. Cavity consists of one fully reflective mirror,

one partly reflective mirror that couples the output beam to the optical system

and gain medium. 2.The pumping mechanism excites the atoms inside the gain

medium. 3 to 5. Spontaneous emission cause the stimulated emission and lasing

process continues to generate coherent laser beam. . . . . . . . . . . . . . . . 31

2.5 Power profile of two classes of lasers. Although the average power of both profiles

are the same, but the pulsed laser contains higher power within a very small period

of time, whereas in the continues wave laser the output power is constant for all

the time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.6 Structure of an Hg arc lamp. The electric discharge between the cathode and the

anode ionizes the gas between them. Excited atoms get back to their stable state

and release the energy in the form of light. The wavelength of the emitted photon

depends on the difference of the energy levels. Figure taken from ZEISS microscopy

online campus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

x

2.7 Diagram of a light emitting diode. Applying voltage to the LED will results in

current flow within the junction. When electrons recombine with holes, they release

their energy in the form of light. This phenomenon is called electro-luminescence

effect. Figure taken from Department of Physics at Warwick University. . . . . . 38

3.1 Photons coming from the green region pass through the pinhole. These are desired

photons. On the other hand, there are some photons coming from entrance window,

but will be rejected by the aperture stop (shown in yellow). There is also another

type of photons, called multiply scattered photons (shown in red). These photons

seem to come from the entrance window. These photons pass through the pinhole

and destructively contribute in the image. These photons needs to be rejected in

order to obtain a better image. . . . . . . . . . . . . . . . . . . . . . . . . 41

3.2 Transmission through a medium. Green photon transmits without scattering (de-

sired). Red photon undergoes multiple scattering events which increases the trav-

eling distance, hence the arrival time increases. Detected signal is shown in Firgure

3.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.3 Arrival time for photons shown in Figure 3.2. Green signal corresponds to trans-

mitted photons without scattering. Red signal corresponds to multiply scattered

photons reached the detector traveled longer distance. The detector is turned on

for the gating window, therefore, red signal is rejected. . . . . . . . . . . . . . 43

3.4 Basics of interferometry. Based on the optical path length, beams from reference

and source arms interfere constructively (on the right) or destructively (on the left). 44

xi

3.5 Polarization gating. Light coming from the light source is unpolarized. After it

passes through the polarizer, only the portion of the light that its polarization

matches with the polarizers orientation goes through. In case of microscopy and

rejecting multiple scattered photons, the light coming from the specimen is de-

plorized after multiple scattering events. Therefore, by utilizing a polarizer with

the right orientation, one can reject unwanted photons. Figure taken from Ameri-

can Polarizers Inc. website. . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.6 Image of the surface of the water. Reflection form the surface of the water is S

polarized. In order to reject the S polarization, a polarizer is used (image on the

right) which improves the contrast. . . . . . . . . . . . . . . . . . . . . . . 46

3.7 Multiple scattering photons in full pupil and half pupil configuration. It is shown

that in full pupil configuration (on the left), unwanted multiple scattered photons

can pass through the whole pupil. In case of using the half pupil configuration (on

the right), statistically, half of thee unwanted multiple scattered photons is rejected

by pupil stop ,but it leads to losing resolution since only half of the numerical

aperture is being used. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.8 Divided pupil configuration uses half of the pupil for illumination and the other

half for detection. Multiple scattered photons are drawn for two cases. In case

(a) the photon does not contribute in the image since it is not in the illumination

path. In case (b) the photon reaches the detector. Therefore, statistically, divided

pupil configuration works the same as half pupil configuration in terms of rejecting

unwanted multiple scattered photons. . . . . . . . . . . . . . . . . . . . . . 48

xii

4.1 Basic of the dual wedge scanning system. (a) Refraction by one prism. Rotating

the prism will cause the beam to travel a circular pattern. (b) Refraction by two

rotation prism. Each prism will generate a circular pattern. (c) Combination of

the patterns. Based on the direction and speed of rotation, different patterns of

scanning can be obtained [6]. . . . . . . . . . . . . . . . . . . . . . . . . . 52

4.2 Two general patterns of scanning. (a) is achieved when prisms rotate in the same

direction. (b) is achieved when prisms rotate in the opposite direction [6]. . . . . 53

4.3 Diagram of the dual wedge scanning reflectance confocal microscope. . . . . . . 54

4.4 Images of (a) cellulose fibers within paper and (b) the cellular structure within a

plant leaf taken by dual wedge point scanning confocal microscope. . . . . . . . 55

4.5 Fluorescence process. In 1-photon fluorescence, electron absorbs a photon and goes

to a higher level of energy. This is a real state. The excited electron returns to its

stable level and emit a photon with slightly higher wavelength in comparison to the

primary photon. In 3-photon excitation, electron absorbs 3-photon and stepwise.

It can go up to three real states as in the plot on the right. In our case, the melanin

can absorb three photon and stepwise go up three virtual states and then emit a

photon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.6 New design of the dual wedge point scanning confocal microscope. New parts are

the PMT, the dichroic mirror and the filter and the lens in front of the PMT. Red

lines represent light path for reflection mode and yellow lines represent fluorescence

mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

xiii

4.7 Transmission of the filter in front of the PMT and reflection of the dichroic mir-

ror. The goal was to block any leak from the incident beam to the PMT. The

fluorescence wavelength is about 450nm. The spectrum of the combination of the

filter and the dichroic mirror shows that we were successful in rejecting 839nm

wavelength of the laser. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

4.8 Images of enhanced emission of Sepia melanin in atmosphere. (a) confocal re-

flectance image. (b) three-photon image. (c) merged image. Scale bar is 10µm . . 59

4.9 Images of a dark human hair in atmosphere. (a) confocal reflectance image. (b)

three-photon image. (c) merged image. Scale bar is 10µm. . . . . . . . . . . . 60

4.10 Schematic of the line scanning confocal microscope. Note image and pupil planes.

We have vertical lines in image planes and horizontal lines in pupil planes. Cylin-

drical lens focuses the laser to a line. First pupil stop is applied at the cylindrical

lens focal spot. Galvo scanner is used to swipe the line on the sample, and backscat-

tered light travels the same path to the beam splitter. Second pupil stop is applied

after the beam splitter at focal spot of the telescope lens #1. Detector lens focuses

the backscattered light on the detector array. . . . . . . . . . . . . . . . . . 63

4.11 USAF target under the line scanning confocal microscope. Note the size of the

smallest bars in the image is about 2.19µm. . . . . . . . . . . . . . . . . . . 64

4.12 90%-10% method to find the transverse resolution. Upper red dashed line shows the

90% of the difference between the maximum and the minimum; lower red dashed

line shows the 10%. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.13 Measurement of the axial resolution. The axial distance between two green dashed

lines is defined as axial resolution. . . . . . . . . . . . . . . . . . . . . . . . 66

xiv

4.14 Knife-edge images for three different pupil configurations. a. divided pupil config-

uration, b. half pupil configuration and c. full pupil configuration. Small part of

each image is magnified to show the transition from completely reflecting(bright)

to completely transparent(dark) surfaces. Faster transition indicates better reso-

lution. Fluctuations are result of synchronization signals in the data acquisition

which happens in the direction of orthogonal to scanning direction. . . . . . . . 67

4.15 Axial scan for three different pupil configurations. Full pupil has a higher value

at the peak, because using the full pupil for illumination; more power is delivered

to the sample. In half pupil configurations, we lose sectioning ability as plot is

flattened. Utilizing divided pupil configuration will recover some of the loss of

sectioning ability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

4.16 Image of a business card fibers under the line scanning confocal microscope using

the full pupil configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4.17 Emission spectrum of the high brightness LED. Red dashed lines indicate FWHM

of the spectrum. The maximum emission occurs at 812nm and the FWHM is about

32nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.18 Geometry of area and solid angle for the source and the detector. . . . . . . . . 71

4.19 Line-scanning confocal microscope design. a) Illumination path to sample. b)

Detection path from the sample to the detector. c) First design option to place

LED in the field plane. d) Second design option is to place LED in pupil plane. . 73

4.20 US air force target under microscope. LED is used as the light source. Field of

view is about 70µm in vertical direction and 88µm in horizontal direction. . . . 74

xv

4.21 Image of the horizontal knife-edge with the corresponding averaged detected signal

from the detector. The 90%-10% criterion is applied on the signal, to measure

the resolution. The resolution is defined as the transverse distance between the

corresponding spots detected by the green dashed lines.. . . . . . . . . . . . . 75

4.22 Image of the vertical knife-edge with the corresponding averaged detected signal

from the detector. The 90%-10% criterion is applied on the signal, to measure

the resolution. The resolution is defined as the transverse distance between the

corresponding spots detected by the green dashed lines. . . . . . . . . . . . . 75

6.1 Weights for applying Laplacian transform. The center pixel has the weight of 20,

its close neighbors has -4 and far neighbors has -1. . . . . . . . . . . . . . . . 80

6.2 Dual-line scanning probe. Blue arrows shows the direction of scanning the probe

on the large field of view. Paired green and red dotted lines are used together to

form and register the image. . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.3 Principle of operation of the laser-driven light source. Figure taken from Energetiq

website. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

xvi

Abstract

”Confocal microscopy is an important tool in biomedical imaging. It can provide

images of sub nuclear particles. Marvin Minsky invented confocal microscopy on

1955, but it did not attract attention until recent developments in semiconductor

electronics and optics. Utilizing a high numerical aperture objective and a pinhole

aperture in front of the detector, enables us to illuminate a very small spot on the

specimen and detect the light coming only from the illuminated spot. This provides

background light rejection and the optical sectioning ability.

The invention of the laser increased the application of the confocal microscopy.

Lasers are the popular light sources in confocal microscopy since they provide high

radiance required for confocal. Also they can provide high degree of monochromaticity

with wide range of output wavelengths.

This thesis analyzes the properties of confocal microscopy in both point scanning

and line scanning systems. The most important part of this thesis is the utilization

of a recently developed high brightness LED as the light source. The laser light

source in our line scanning confocal microscope is replaced by a high brightness LED,

in reflectance mode. We have shown that high brightness LED is able to provide

enough radiance for confocal microscopy. Properties of images obtained by the high

brightness LED is compared with the same setup utilizing a diode laser as the light

source.”

Chapter 1

Introduction

Confocal microscopy was invented in 1955 by Marvin Minsky. He had the idea of

illuminating only small spot on the specimen unlike conventional microscopy. He

managed to avoid most of the unwanted light from any location of the specimen except

for the illuminated spot. Also on the detection path, he added a pinhole to collect

only the light coming from the focal spot on the specimen. His invention, confocal

microscopy, scanned the specimen point by point and the image was reconstructed by

scanning the diffraction limited spot on the specimen. The first confocal microscope

used a moving stage for the specimen in order to scan the beam. The specimen was

moved in two directions with the beam stationary. He was able to obtain an image

every 10 seconds.

Minsky’s invention did not receive attention for imaging until advances in optics

and electronic instruments. After invention of laser, confocal microscopy rapidly be-

came popular since lasers provide high radiance with the ability to focus the beam

onto a diffraction limited spot. Development of the detectors helped confocal mi-

croscopy to detect very weak light emission. Different materials with fluorescence

1

CHAPTER 1. INTRODUCTION 2

emissions made a great impact on the application of confocal microscopy. Confocal

microscopy is now being used in so many medical and industrial applications. Fluo-

rescence confocal microscopy utilizing wide range of fluorophore materials, provides

images of sub nuclear structure of cells in biomedical imaging with good contrast.

Characteristization of the confocal microscopy, as one of the most popular mi-

croscopy methods in biomedical imaging, is the main topic of this thesis. In the

following chapter, we are going to discuss basic concepts of microscopy with the main

concentration on confocal microscopy. We are going to discuss point scanning and

line scanning systems and most important components of a confocal microscope.

Since part of our research is to use a high brightness light emitting diode as a

non-laser light source, Chapter 2 discusses characteristics of a light source in general

and specifically light sources usually being used in confocal microscopy. We start

with the radiance theorem and continue the discussion on laser and non-laser light

sources.

In Chapter 3, we are going to discuss different pupil configurations and describe

divided pupil configuration as one of the gating mechanisms to reject unwanted mul-

tiply scattered photons. We have examined three different pupil configurations and

compared them. Results suggest that a divided pupil configuration can recover some

of the lost resolution that results from use of smaller area than the whole aperture,

by rejecting more of unwanted multiply scattered photons.

With all the material we needed for our experiments about confocal microscopy

discussed in Chapters 1 to 3, Chapter 4 discusses our experiments in three sections.

Section 4.1 represents our research on a point scanning confocal microscope in fluo-

rescence mode which was used to image pure melanin and melanin in the human skin.


Section 4.2, represents our study on different pupil configurations for the line scan-

ning confocal microscope. Section 4.3, discuss the high brightness LED as the light

source of our line scanning confocal microscope. At the end of this thesis, Chapter 5

summarize results of our experiments and Chapter 6 discuss the possible future work.

1.1 Optical Imaging

The basic concept of the modern microscopy was established about a hundred years

ago by Ernst Abbe. He showed that the image resolution is determined by the objec-

tive lens, condenser, wavelength of the light and index of refraction of the material

between the objective lens and specimen.

dmin =1.22λ0

NAobj +NAcond(1.1)

where dmin is the minimum resolvable spacing of 2 point objects within a specimen

which is expressed as lateral resolution, λ0 is the utilized wavelength, NAobj and

NAcond are numerical apertures of the objective lens and condenser respectively. Nu-

merical aperture, as it is shown in Equation 1.2, is the product of sine of the half angle

(α) and refractive index of the material between the objective lens and specimen (n).

Figure 1.1 demonstrate how numerical aperture is defined;

NA = nsin(α). (1.2)

As can be seen in Figure 1.1, by utilizing a condenser a higher working numerical

aperture is obtained, therefore, referring to Equation 1.1, the minimum resolvable

spacing decreases. There is also another factor that can help reduce dmin in Equation

1.1. Equation 1.2 demonstrates that increasing n, the index of refraction between


Figure 1.1: Concept of numerical aperture. a) specimen is illuminated by a collimated beam,then numerical aperture is defined as, NA = nsin(α). b) A condenser has a numerical aperture thesame as that of the objective lens. In this case, working aperture is the sum of numerical apertureof objective lens and condenser, therefore NA = nsin(2α). Figure taken from ZEISS microscopyonline campus

the specimen and objective lens, increases the numerical aperture, therefore leads to

smaller dmin in Equation 1.1. dmin, known as the minimum resolvable distance on the

specimen, is defined based on the Airy disk created by diffraction of light within the

optical components. As shown in Figure 1.2, light interacts with the specimen and

diffracts, resulting in the diffraction pattern known as the point spread function(PSF).

Based on the Airy disk pattern showed in Figure 1.3, the Rayleigh criterion implies

that two Airy disks can be resolved when separated by the minimum distance, dmin.

This distance measures how well the microscope can resolve details of the specimen

in the image. Based on this, lateral resolution of the microscope is defined.

Similar to lateral resolution definition, axial resolution is defined as the minimum

distance in the axial direction which two point spread functions can still be seen


Figure 1.2: Point spread function and Airy disk definition. a) Airy disk pattern generated formlight diffracted in specimen. b) 3D representations of the diffraction pattern on the image plane,known as the point-spread function. c) An Airy disk is the region enclosed by the first minimum ofthe Airy pattern and contains approximately 84% of the energy. Figure taken from ZEISS microscopyonline campus

as two. Linfoot and Wolf (1956) [7], calculated the 3D pattern of the point spread

function and based on that, the minimum axial distance from center of the 3D point

spread function is given as Equation 1.3. [5]

zmin =2λ0n

(NAobj)2(1.3)

Comparing Equations 1.1 and 1.3 shows that unlike lateral resolution, axial reso-

lution changes with the inverse of numerical aperture squared. Since NA < 1, axial

resolution is worse than lateral resolution.


Figure 1.3: Two Airy disks close to each other. dmin is the minimum distance that each of the twoAiry disks can be resolved by the human eye. Figure taken from ZEISS microscopy online campus

1.2 Optical Sectioning

Optical sectioning refers to the ability of a microscope to produce images of the

focal plane within the sample. Although physical sectioning provides good resolution

and sensitivity, it is technically difficult and invasive. On the other hand, optical

sectioning is technically easier and provides multiple slices, but it requires complex

computing. In order to obtain a thin section, it is required that only light coming

from the focal plane passes through the microscope. Ideally a microscope is set to

detect light only coming from the focal plane, but after passing through the sample,

light scatters, especially in high scattering medium like biological tissue. Therefore, in

practice light from planes other than the focal plane reaches the detector and affects

optical sectioning.


Figure 1.4: Rayleigh criterion, two Airy disks are considered resolvable if the valley between thepeaks is about 20%-30% of the maximum. Figure taken from Nikon Instruments, Inc. website

There are many techniques to improve optical sectioning. In confocal microscopy

illumination is focused on to a point/line and by adding a pinhole/slit in front of

the detector, only light coming from the focal plane passes through the microscope.

Therefore, confocal microscopy can provide a good optical sectioning. Multi photon

fluorescence microscopy also improves optical sectioning. The fact that signal is

generated only from the region that has been stimulated to fluoresce, means that if

the illumination is specific to the focal plane, the image is obtained based on the signal

only coming from the focal plane. Therefore, it provides a good optical sectioning.

In two-photon excitation and similar modalities in order to produce a signal, atoms

need to absorb two or more photons simultaneously. Therefore, this phenomena will

happen only in region very close to the focal plane, which improves optical sectioning.

Optical sectioning ability is related to the wavelength, NA of the objective lens

and index of refraction of the material between objective lens and the sample. As


mentioned above in Equation 1.3, optical sectioning improves with the inverse of the

NA squared.

1.3 Confocal Microscopy

Conventional wide-field illumination optical microscopy has been widely used since

well before Abbe’s work. In comparison, confocal microscopy, developed about half a

century ago, offers background rejection from outside the focal plane and the ability

to control depth of field. Therefore, it provides better resolution and contrast. [1] The

main idea in the confocal approach is utilizing a spatial filter to reject background or

what is called ”out-of-focus” light. The illumination method in confocal microscopy

is completely different from conventional microscopes. As Figure 1.5 illustrates, in

a conventional microscope, a light source intensely illuminates the whole specimen

continuously and uniformly. Then the objective lens gathers an enormous amount of

light which is a combination of in and out-of-focus light. This degrades resolution and

contrast of the image. However, in confocal microscopy, light from the light source

is first expanded to fill the objective aperture. The objective lens then focuses the

beam to a very small spot on the focal plane. The size of the spot depends on the

design of the microscope, working wavelength and working numerical aperture. [8]

Since the light coming from above and below the focal plane is not confocal with

the pinhole in front of the detector, it will form larger Airy disk on the pinhole plane.

Hence, only a small portion of this out-of-focus light will be delivered to the detector

and most of this out-of-focus light does not contribute to the image [9]. In conven-

tional wide field microscopes, the specimen is illuminated by an incoherent mercury

or xenon arc-discharge lamp. The image can be viewed either directly in the eyepiece


Figure 1.5: Illumination method comparison between conventional and confocal microscopes.In wide-field illumination, a large volume of the specimen is illuminated. However in confocalmicroscopy, a very small volume of the specimen is exposed [1].

or utilizing an array detector or traditional films. On the other hand, image formation

in confocal microscopy is fundamentally different. Unlike conventional microscopes,

a confocal microscope consists of an excitation source, scanning system and detector

connected to a computer for acquisition, processing, analysis, and displaying the im-

age [10]. In the case of a point scanning system, only one point is imaged at a time

and the computer constructs the image by scanning the point on the specimen.

One of the important components of the scanning system is the pinhole aperture

in front of the detector. It rejects out of focus light from above or below the corre-

sponding conjugate image plane on the specimen [11]. Out-of-focus light projects an

Airy disks with a diameter larger than those from the focal plane on the aperture

pinhole. Since their Airy disks are larger than those coming from the focal plane,

and they are spread out over a larger area, only a small portion of their power goes

through the pinhole. The other benefit of using a pinhole is to reject most of the

stray light that passes through the optical system.

Confocal microscopy rapidly became popular, being relatively easy to implement

while providing extremely high quality images. It is now widely used in many in-

dustrial and clinical applications. For instance, confocal microscope plays a very


important role in cell biology and imaging living cells and tissues. [5, 12–15]. Figure

1.6 illustrates the difference between wide-field and confocal fluorescence microscopy.

Beside fluorescence, reflectance confocal microscopy is being used in biomedical imag-

ing such as in vivo imaging of human skin as it is presented in Figure 1.10. Reflectance

confocal microscopy is the subject of our research. In Chapter 4, we have discussed

our research on confocal reflectance microscopy in both point scanning and line scan-

ning systems.

Figure 1.6: Comparison of wide-field(upper row) and confocal microscopy(lower row). (a) and (b)mouse brain hippocampus thick section. (c) and (d) Thick section of rat smooth muscle. (e) and(f) Sunflower pollen grain [1]

Although the confocal microscope provides better resolution compared to conven-

tional microscopes, its resolution is poor compared to transmission electron micro-

scope (TEM) [5]. However, confocal microscopes are considerably cheaper and easier


to use since TEM requires sample preparation before imaging. The process of prepar-

ing the sample takes time and cost and make it impossible for TEM to be used as an

in-vivo imaging method.

1.3.1 Point Scanning

As mentioned before, the confocal microscope focuses the beam to a very small spot

on the specimen. In order to form an image, scanners are used to scan the focal point

in two dimensions. Microscopes that use such a scanning method are known as point

scanning confocal microscopes. Figure 1.7 shows a typical design for a point scanning

confocal microscope. The laser beam is first expanded to fill the aperture of the

objective lens, which focuses the beam to a diffraction-limited spot on the specimen.

The spot is then scanned on the specimen utilizing two galvanometer mirrors. There

are several patterns to scan the spot on the specimen, but most typical pattern is the

raster pattern across the specimen plane.

Hence, the image is generated by scanning the focused beam across the specimen

with a predefined pattern of scanning. In the raster pattern, two galvo mirrors move

the focused beam. One is responsible for moving the beam from left to right in the x

direction, and the other mirror moves the spot in the y direction. After each row in

the x direction, the first galvo returns to its first position rapidly, and the other one

moves the beam in the y direction. During this transition, no data is saved. Since

the speed of light is much higher than scanning speed, the light coming back from the

specimen travels the same path as the excitation light, during the scanning process.

This process is termed descanning in the literature [8, 16, 17]. The problem of using

a galvo mirror is that it takes time to return to initial position. This process makes

it challenging to synchronize the scanner with the data acquisition board. In order


Figure 1.7: Typical confocal microscope. Excitation light is directed to the sample by a dichroicmirror or beam splitter. Two galvanometer mirror scan the beam on the sample in two directionand objective lens, focuses the beam onto a diffraction limited spot on specimen. Light coming backfrom the specimen, travels the same path to the dichroic mirror which transmit the beam to theconfocal pinhole and the detector. A/D which is synchronized with scanner is then used to collectinformation and store in the computer. Image is then reconstructed in computer and is displayedon the monitor [1].

to scan the diffraction limited spot faster, the galvo scanner can be replaced by a

rotating polygon mirror. Figure 1.8 shows how a rotating polygon can scan the spot

on the specimen faster. Unlike a galvo scanner, the rotating polygon does not need

to return to its initial position after each scan. Therefore it is easier to synchronize

it with the data acquisition board. The descan process is the same as in utilizing a

galvo scanner.

After the scanning mirrors, the reflected light passes through a dichroic mirror or

a beam splitter and then to a lens which focuses the light on the detector through the

pinhole. Although the focused spot on the specimen is moved by the scanner, focused


Figure 1.8: Rotating polygon mirror used as scanner. A motor rotates the mirror and continuouslyscans the beam on the specimen. Unlike galvo scanner, it is not needed to return the mirror tothe initial position. Therefore it is easier to synchronize the scanner with data acquisition board.Backscattered light returns the same path to the polygon mirror. Figure taken from LEYBOLDPhotonics educational kit website

light on the pinhole is stationary and scanning process just changes the intensity of

the light focused on the detector according to the amount of backscattered light

from the scanning spot. Light is then converted into an analog electrical signal with

variation in voltage that contains the information of the images. The signal is then

periodically sampled utilizing an analog to digital (A/D) converter. Sampling speed

is synchronized with the scanning system in order to reconstructed the image point-

by-point inside the computer. Unlike wide-field microscopy, in confocal microscopy

the image never exists as a real image and cannot be seen by a microscope eyepiece.

1.3.2 Line Scanning

Using a point scanning confocal microscope requires scanning the diffraction limited

spot in two directions. This is done usually by two scanners each one responsible for

moving in one direction. Utilizing scanners in two dimension makes the design, data

acquisition and image formation more complicated. In order to simplify the scanning


process, line scanning configurations have been developed. In a line scanning system,

the beam is focused to a line instead of a point. Therefore, it takes only one scanner

to sweep the line in the direction perpendicular to the line direction. Since, a line

scanning confocal microscope scans the line only in one direction, it is much easier to

work with in comparison to point scanning system with two scanners.

Although a line scanning confocal microscope is only confocal in one direction, it

provides optical sectioning and resolution comparable to similar point scanning con-

focal system. Figure 1.9 present axial response for point scanning and line scanning

confocal microscope with different pupil configurations which are discussed in Chap-

ter 3. Dwyer et el. reported that they were able to obtain images of human epidermis

in vivo, using a line scanning confocal microscope [3]. Figure 1.10 shows images of

human epidermis obtained with a line scanning confocal microscope as reported by

Dwyer et al.

Figure 1.9: The axial response for (a) point-scan and (b) line-scan for different pupil configuration.[2]

In order to use a line scanning configuration, first it is required to make a line in

the image plane. Depending on the illumination source, different designs can be used


Figure 1.10: Images of human epidermis in vivo. Scale bars, 50 µm. (a) the stratum corneum(SC), granular (GR), and spinous (SP) cells. (b) smaller basal cells with dark nuclei (arrows),arranged in ring-shaped clusters (arrowheads) [3].

to create the line illumination. In the case of a coherent collimated laser source, a

cylindrical lens focuses the beam onto a line parallel to its axis. Figure 1.11 shows

how a beam is focused onto a line utilizing a cylindrical lens. Since the beam is

collimated and coherent, on the conjugate pupil plane, the beam is also focused onto

a line but perpendicular to the direction of the primary line.

In the case of an incoherent illumination source, a slit aperture is placed in the

image plane in front of the light source. The slit aperture transmits only a small

portion of the total source power. In this case, there will be a line on all the field

planes of the microscope, but on the pupil planes, the beam is not necessarily a line.

Figure 1.12 shows the difference between coherent and incoherent cases. As further

discussed in Chapter 4, the source can be placed either in the pupil or field plane.

In any case, different design is required. Figure 1.13 shows a typical slit aperture to

generate a line illumination for line scanning confocal microscope.


Figure 1.11: Cylindrical Lens focuses the beam onto a line.

1.4 Components in Confocal Microscope

1.4.1 Pinhole or Slit

Based on the theory behind the confocal microscopy, the pinhole in point scanning,

or the slit in line scanning, is one of the most important components of the confocal

microscope. In theory, a confocal microscope illuminates very small diffraction limited

spot within the specimen and collects the light coming only from that spot. In order

to make sure that detected light is coming from the diffraction limited spot, a pinhole

or slit is used in front of the detector. Since numerical aperture on the sample should

match with the numerical aperture on the detector, the size of the pinhole depends

on the diffraction limited spot size and magnification of the confocal microscope. The

size of the pinhole is chosen by the size of the image of the diffraction limited spot

on the image plane of the detector. Utilizing a smaller pinhole leads to rejection of

light coming from the focal spot. This results in optical power loss. On the other

hand, a larger pinhole leads to leakage of unwanted light from out-of-focus planes


Figure 1.12: Field plane and pupil planes for coherent and incoherent line of light source. In thecase of coherent light source, in both image and pupil planes light is focused on a line, while in thecase of incoherent light sources, source can be separated as many point sources. Therefore, in fieldplanes, light is focused on a line, but in pupil planes light is not focused.

to the detector which leads to poor contrast and resolution. Figure 1.14 shows how

a pinhole rejects light coming from below or above the focal plane. Theoretically,

pinhole size is the same size as the point spread function diameter, but in practice,

in order to deliver more power to the detector, the pinhole size is chosen to be about

3 times greater than the point spread function diameter.

In the case of using line scanning confocal microscope, since it is only confocal in

one direction, a linear array detector is being used. Therefore, a slit aperture, same

as Figure 1.13, is usually being used in front of the detector.


Figure 1.13: Typical slit aperture. Only a small portion of light goes through the aperture. Itgenerates a line illumination on the other side.

1.4.2 Beam Splitter

A splitter is an optical device that splits the beam into two different beams. In so

many applications such as interferometry, it is important to split the beam into two

beams. In interferometry, one part of the beam goes through the setup and interacts

with the sample. The other part of the beam is reserved as source beam. Since these

two beams go through different paths, recombining them provides information about

the optical path traveled by beams. This is the fundamental idea of interferometry,

which is used in so many applications such as optical coherence tomography. Most

common beam splitters are made of two prisms glued together at their bases. The

glue between prisms is chosen for a particular wavelength. Therefore, light going

through the beam splitter at the middle point splits into two beams. One part of the

beam will pass the beam splitter and the other part is reflected and goes through the

other prism. Figure 1.15 on the right represents typical cube beam splitter.

Another type of beam splitter consists of a glass or a piece of plastic, coated

with one or more thin transparent dielectric. The thickness and compositions of the

coatings are designed for different transparency. Figure 1.15 on the left,shows a flat

beam splitter. In some applications, it is required to split beams based on their

polarization. In this case, polarizing beam splitters are necessary. These types of


Figure 1.14: Diagram of pinhole application in confocal microscope. Pinhole or confocal detectoraperture rejects light coming from below or above the focal plane therefore it improves contrast andresolution.

beam splitters use birefringent materials, which allow them to split the beam based

on its polarization. Figure 1.16 illustrates how a polarizing beam splitter reflects and

transmits the beam based on its polarization. A polarizing beam splitter is usually

used along with a quarter-wave plate. Therefore, most of the incident beam will

transmit through the beam splitter and pass the quarter-wave plate twice. When the

reflected beam goes through beam splitter, its polarization is changed by the quarter

wave plate and, therefore, most of it will reflect by the beam splitter. This process is

very useful since it helps reduce power loss.


Figure 1.15: Typical beam splitters. Flat beam splitter on the left. It splits the incident beam(frombottom) into two beams. On the right, a cube beam splitter which splits the incident beam intohalf.It transmits half and reflects the other half.

1.4.3 Detector

For each imaging method, the detector is chosen based on the imaging method, range

of the wavelength and required sensitivity. Most common detectors for different

microscopy methods are photomultiplier tubes, photodiodes and solid-state charged-

couple devices (CCDs). In the case of confocal microscopy, light coming from the

specimen goes through the pinhole aperture in front of the detector in order to reject

light coming from out-of-focus spots. Therefore, the amount of light that reaches the

detector is usually exceedingly low, and a detector with high sensitivity is required.

Besides high sensitivity, it should respond quickly to small variation of light intensity,

which contains information from the specimen [18]. For instance, in order to get an

image, for a full frame about 0.1 of a second is required. Considering the case of line

scanning microscope, as the subject of our research, in order to capture 1000 lines

for a frame, it is required that the detector capture each line in 1× 10−4 of a second.

Now Considering a point scanning system with 1000 points in each dimension, the

detector is required to capture each point in 1 × 10−7 of a second. Therefore, the


Figure 1.16: Polarizing beam splitter. incident beam(black) is not polarized. The beam split-ter will splits the incident beam into vertical(transmission) and horizontal(reflection) polarization.Figure taken from Thorlabs Inc.

detector should be chosen by the imaging requirements. In the following, we are

going to discuss typical types of detectors that are popular in confocal microscopy.

Photomultiplier tubes are popular because of their sensitivity and ability in photon

counting. Photodiodes are the other types of detectors working with semiconductor

structure. Also, in our research we have used an array detector, therefore, last section

discusses array detectors.

• Photomultiplier tube (PMT)

Photomultiplier tubes are very popular in many microscopy applications be-

cause of their high sensitivity and fast response. A PMT is developed based on

a critical element called the photocathode. A photocathode can transfer energy

of a photon to an electron. This phenomenon, is termed the photoelectric effect.


A PMT consists of a photocathode followed by a chain of electron multipliers

(dynodes) within a vacuum tube. When the photocathode absorbs photon en-

ergy, it will release an electron. The electron reaches the first electron multiplier

and generates more electrons. The cascading process continues to the end of the

tube, where generated electrons are converted to the output signal of the PMT.

Therefore, the output signal is proportional to the photon energy absorbed by

the photocathode. The cascade multiplication of the electrons within the PMT

allows it to detect very small amounts of light. Figure 1.17 shows the diagram

of a PMT and its elements.

Figure 1.17: Diagram of a photomultiplier tube(PMT). Incident photon excites photo-cathodeto generate primary electron. Primary electron then excite dynodes to create multiple electrons ina cascade process. At the end of the tube, anode receives electrons and generates output signal.Figure taken from scintillator materials group at Stanford University

• Photodiodes

A photodiode is a p-n junction. When a photon strikes the diode, it generates

an electron within the junction based on the photoelectric effect. Electrons gen-

erated inside the junction, move towards the cathode and create a photoelectric

current. Photodiodes also respond to light absorption by current generation,

but they do not provide gain as it is in a PMT. Photodiodes have relatively flat


response over the entire visible spectrum with high quantum efficiency. Their

uniform response, dynamic range and response speed are excellent. However,

they produce a considerable amount of noise, mostly thermal noise, resulting in

relatively poor signal-to-noise ratios. Avalanche photodiodes with limited gain

have been developed and been utilized in confocal and wide-field fluorescence

microscopes. Although they have up to 300-fold gain, they exhibit significant

dark noise even when cooled. Figure 1.18 compares quantum efficiency of the

two main types of the detectors.

Figure 1.18: Comparison of different detectors spectral sensitivity.


• Array Detectors

Most array detectors use semiconductor materials. Usually detector arrays con-

sist of small detector elements placed close to each other, each able to convert

the incident electromagnetic wave into an electrical signal. A circuit is attached

to each element to relay and multiplex the electrical signal to output amplifiers.

In our line scanning confocal microscope, we have used a CMOS linear array

detector which consists of photo-diode pixels. When one of these pixels, absorbs

a photon, the incident photon creates a hole and an electron in the semiconduc-

tor. Recombination of the electron and the hole generates the electrical signal.

The detection bandwidth of the detector array depends on the band gap en-

ergy of the semiconductor material since only photons that have energy higher

than the band gap energy of the semiconductor material can excite electrons

and create electron and hole. Hence the longest wavelength to be detected by

the detector is determined by the band gap energy of the semiconductor as in

Equation 1.4.

λcutoff =hc

Ebandgap=

1.24× 10−6(eV.m)

Ebandgap(eV )(1.4)

In most imaging applications, the range of wavelength to be detected is from

UV to IR, therefore, silicon with band gap energy of 1.1eV , is the the most

popular semiconductor for array detector for such applications. In our research,

the working wavelength is about 830nm, in near infra-red region. Hence, for

our line scanning confocal microscope a silicone linear array detector is used.

This is further discussed in Chapter 4.

Chapter 2

Light Sources

2.1 Introduction

The Light source is one of the most important components of an optical system.

Different modes of imaging require different light sources with different characteristics.

Therefore, based on the application, the light source is chosen. Properties such as

brightness, coherence and wavelength, are the most important properties that have to

be considered to choose a proper light source for an optical system. As in any other

engineering problem, these parameters are related to each other so there should be a

trade off among all these parameters. In the following sections, important properties

of a light source are discussed.

2.1.1 Brightness

The radiance theorem implies that the power emitted from a source element with a

projected area of dAproj into a solid angle of dΩ follows the relation with radiance of

the source as in Equation 2.1,

25

CHAPTER 2. LIGHT SOURCES 26

d2Φ = LdAprojdΩ, (2.1)

where d2Φ is the power in watts, dΩ is the solid angle and L is defined as the Radiance

of the source with the unit of (W/m2/sr). The theorem implies that, through a lossless

optical system, radiance is conserved [4]. In the other words, if an optical system does

not produce absorption and light undergoes perfect reflection and refraction while

going through the system, the radiance is conserved and is equal on source, specimen

and detector. Figure 2.1 illustrates the geometry for the radiance theorem.

Figure 2.1: Radiance in an Image. The image area is related to the object area by A2 = mA1,and the solid angles are related by Ω2 = Ω1

m2 ( nn′ )

2. Where m is the magnification and n and n′ arethe index of refraction of materials on two sides of the imaging system, respectively [4].

According to the radiance theorem, in microscopy, one of the most important

specifications of the light source is not only its ability to provide enough photons

per second, but also its ability to provide them from a small etendue. For instance,

a large fluorescent tube produces about the same number of photons per second

in comparison to the short arc HBO-50 bulb that is usually used in fluorescence


microscopy, but in the short arc HBO-50 bulb, this number of photons is produced

from an area about one million times smaller than the fluorescent tube. Therefore,

although the fluorescent tube and the short arc HBO-50 provide the same power, the

short arc HBO-50 provides higher radiance which is required in microscopy [5].

2.1.2 Wavelength

The other important feature of a light source is the light wavelength emitted from the

light source. Plasma and filament light sources which provide white light have almost

uniform spectra in visible wavelengths. Unlike them, arc sources produce photons

whenever an excited electron loses energy and moves to a lower energy level in the

material. Therefore, the wavelength of the light depends on the energy levels of the

electrons within the material. Based on Equation 2.2 and Equation 2.3 if the photon

in the visible spectrum is needed, then the difference of energy between each level

should be between 3.3 to 1.59 electron volts.

λ1 = 380nm⇒ E1 =hc

λ2= 3.26eV (2.2)

λ2 = 780nm⇒ E2 =hc

λ2= 1.59eV (2.3)

Traditionally, an efficient way to produce excited electrons from a small area is to

raise the temperature. This is usually done by heating tungsten filament or Hg or

Xe plasma. Figure 2.2 shows an emission spectrum of a Hg plasma. Peaks of the

spectrum represents different levels of excitation.


Figure 2.2: The emission profile of mercury arc lamp. Wavelengths in the visible region are usefulin fluorescence microscopy. Emission profile of the xenon arc lamp is presented. Xenon arc lampemission is almost flat in visible region [5].

2.1.3 Coherence

The other important property of a light source is its coherence. Coherence and

brightness are closely related to each other since most bright light sources tend to be

coherent. Brightness is the ability of the source to produce and focus a large number

of photons into a small focal spot. On the other hand, coherence is the ability of the

light source to generate photons that are in phase as a wave. Therefore, they interfere

constructively on the focal spot. Figure 2.3 illustrates the concept of temporal and

spatial coherence. It also shows how to make a coherent light source from an inco-

herent light source. Lasers produce not just one selected wavelength; they produce

a very narrow bandwidth. These wavelengths are coherent at the source, but after a

distance and scattered while passing through the optical system, these wavelengths

become no longer in phase with each other. These out-of-phase wavelengths interfere

with each other at the detector and produce a pattern of bright and dark spots on

the image called speckle.


Figure 2.3: Spatial coherence and temporal coherence. An incoherent light source emits light inall direction with different wavelengths. Waves that pass the pinhole aperture are spatially coherent,but in order to make a coherent beam, a wavelength filter is required. Figure taken from ZEISSmicroscopy online campus

Depending on the microscopy mode coherent or incoherent light sources offer

advantages. If the the microscope is working in reflectance or backscattered light

mode, utilizing a coherent light source will lead to speckle pattern generation in image.

On the other hand, for confocal microscopy we need high radiance for imaging since it

is required to focus the light onto a very small area on the specimen, and incoherent

light sources usually can not provide enough radiance. A trade off between these

two parameters usually results in choosing laser light sources for confocal microscopy.

There are of course exceptions of using non-laser light sources as discussed in the

following section.


2.2 Laser Light Sources

2.2.1 Introduction to Laser

In 1917, Albert Einstein established the foundation of light amplification by stim-

ulated emission of radiation (LASER). Basically, he re-derived Max Plank’s law of

radiation and predicted that, under certain conditions, a photon can stimulate an

excited atom in the material and generate a second photon with the exact same en-

ergy. This implies that the secondary photon has the same wavelength as the primary

photon. Also, the secondary photon would have the same phase, polarization and di-

rection of propagation. In the other words, the secondary photon is coherent with

the primary photon which results in a coherent beam.

A laser consists of a gain medium, a pumping mechanism and an optical feedback.

The gain medium is the material within the laser cavity. The pumping mechanism

excites atoms within the gain medium. When electrons from an excited atom return to

their stable level, they release the energy difference between energy levels in the form

of emission. Emitted photons are brought back to the cavity by the optical feedback

and stimulate more excited atoms in the gain medium, and, therefore, produce more

photons with the same phase and frequency. This process continues to amplify the

light. The output coupler is used to bring the laser beam to the optical system. The

wavelength of the beam mainly depends on the gain material and the cavity design.

Cavity design results output linewidth much smaller than material gain linewidth.

Most practical lasers contain additional elements that affect properties of the emitted

light such as the polarization, the wavelength, and the shape of the beam. Figure 2.4

represents a simple laser cavity with its components.


Figure 2.4: Laser cavity and lasing process. 1. Cavity consists of one fully reflective mirror, onepartly reflective mirror that couples the output beam to the optical system and gain medium. 2.Thepumping mechanism excites the atoms inside the gain medium. 3 to 5. Spontaneous emission causethe stimulated emission and lasing process continues to generate coherent laser beam.

Laser development rapidly increased the importance and the application of con-

focal microscopy. Lasers are easier to use and provide high radiance. The number of

excitation lines is increasing which provides a larger range of excitation especially in

fluorescence confocal microscopy. Lasers offer such good qualities as a light source,

which make them one of the most important and popular light sources being used

in confocal microscopy. Below, some of the most important properties of lasers are

discussed.

• Monochromaticity


Lasers provide a high degree of monochromaticity. Most of the time, it is

important to excite the specimen with a single wavelength. Usually light sources

provide a range of wavelengths. A smaller range of wavelengths leads to a higher

degree of monochromaticity, and it means that most of the excitation is done

by a single wavelength.

• Brightness

The term brightness is defined with different meanings in different textbooks. In

the context of lasers, brightness which is referred also as radiance is the power

divided by the area in the focus and the solid angle in the far-field. Therefore, its

unit is usually defined as watts per m2 per stradian (Wm−2sr−1). Lasers offer

higher radiance, which is very important in confocal microscopy. For lasers, the

solid angle Ω is defined as a function of the divergence half angle of the laser

(θ).

Ω = πθ2 (2.4)

For area (A) we have:

A = πw20 (2.5)

where w0 is the beam waist radius. Knowing that divergence angle of a laser is

θ = λπw0

for the AΩ product we have:

AΩ = πw20 × π

λ2

(πw0)2= λ2 (2.6)

Hence, for lasers, the radiance can be defined as in Equation 2.7,

L =Φ

λ2(2.7)


• Coherence

Coherence of two waves is defined as how well they are correlated and in phase.

In this term, light coming from lasers has both temporal and spatial coherence.

Temporal coherence implies that the wave is well correlated for all times. Spatial

coherence implies that the wave at different points in space is correlated. Spatial

coherence characteristic of lasers enable us to focus it onto a tight spot. This

is very important in confocal microscopy. Temporal coherence implies a very

narrow spectrum as stated in Equation 2.8.

δt

T=δz

λ≈ f

δf(2.8)

Equation 2.8 shows the relation between the coherence length and the linewidth

in wavelengths. For instance, for a laser with peak wavelength λ = 635nm

and with linewidth of δf = 20nm, coherence length is about 20 wavelengths.

Equation 2.8 implies that if δf decreases (narrow bandwidth), the coherence

length increases.

In addition, some types of lasers provide plane polarized emission, and they usually

provide a Gaussian beam profile. Based on the output power profile of the laser, we

can classify lasers as two main types of Continuous Wave or Pulsed.

• Continuous Wave Laser

As the name states, continuous wave lasers provide continuous output power

over time. Figure 2.5 represents the output power profile for both continuous

and pulsed lasers. Theoretically, a continuous wave laser can be used to gen-

erate pulses of light. We can intentionally switch a continuous wave laser on

and off to generate a beam in pulsed form. Basically a continuous wave laser


is continuously pumped, therefore, a continuous wave laser provides constant

power over time which is essential in some applications to have constant power.

Continuous wave lasers work in either single-mode or multi-mode. Single-mode

operation implies that laser linewidth is very small which results in long co-

herence length. On the other hand, multi-mode operation implies that the the

laser linewidth is multiple of the mode spacing of the laser resonator, which

provides wider range of wavelengths and shorter coherence length.

• Pulsed Laser

Generally, other types of lasers that are not continuous wave laser are classified

as pulsed operation lasers. Therefore, their output power appears to be in the

form of pulses of different duration and different frequency. Some lasers are

classified as pulsed just because they cannot work in a continuous wave mode.

On the other hand, in some applications it is required to expose a sample to

a large amount of power. Energy within a pulse is equivalent to the average

power divided by the pulse repetition frequency.

In some other applications, the amount of energy in the pulse is not important

but the peak of pulse power is required. This is usually required in nonlinear

optics effects. In this case, generating short time pulses is important. There are

techniques to make pulse duration as short as possible for these types of appli-

cations. Figure 2.5 represents typical output power profile for both continuous

and pulsed laser.


Figure 2.5: Power profile of two classes of lasers. Although the average power of both profiles arethe same, but the pulsed laser contains higher power within a very small period of time, whereas inthe continues wave laser the output power is constant for all the time.

2.2.2 Laser Diode

Our research is focused on confocal microscopy. As mentioned in Chapter 4, we have

used a continuous wave laser diode with wavelength λ = 830nm as the light source for

both line scanning and point scanning confocal microscopes. As stated in previous

section, a laser consists of a gain medium and a pumping system. A laser diode

uses semiconductor materials to create p-n junction like a light emitting diode, as

the medium. In laser diodes, the gain medium is pumped electronically. Applying

voltage to the semiconductor p-n junction will create electrons and holes. As in

light emitting diodes, recombination of the electrons and holes generates photons. In


laser diodes, the goal is to recombine electrons and holes in the intrinsic region. This

semiconductor structure is located inside a laser cavity, to create laser diodes. Typical

materials being used in laser diode structure are Ga, In,Al, As, P and N. Different

combination of these materials provide different wavelength for the laser diode. For

instance, in order to obtain wavelength in the near infrared region, GaAlAs or GaAs

are typical combinations and for infra red region, InP or InGaAsP are combinations

typically being used. Laser diodes are usually small in size and therefore easy to use

in optical systems. Besides, they provide large variety of wavelengths.

2.3 Non-Laser Light Sources

Confocal microscopy was invented before the invention of laser. Early confocal mi-

croscopes used non-laser light sources. Although a laser is the favorite light source

in confocal microscopy, it has some disadvantages which open the way for non-laser

light sources in confocal microscopy. As discussed, because of the high degree of

coherence in lasers they provide high radiance required in confocal microscopy, but

their high degree of coherence leads to speckle pattern in the image. Therefore, in-

coherent light sources or in general, non-laser light sources have this advantage over

laser light sources. Two main class of non-laser light sources such as arc lamps and

light emitting diodes (LED) are described in the following sections.

2.3.1 Arc Lamps

Light produced by arc lamps is based on electric discharge between two piece of metal

known as the cathode and anode. The cathode and anode are placed close to each

other. Applying a large electrical voltage across the cathode and anode results in


an electrical discharge and ionization of the gas between them. When the electron

of the excited atom returns to its stable level, it releases the energy in the form of

light. The wavelength of the emitted photon depends on the electron energy levels in

atoms of the gas between the cathode and anode. Usually arc lamps are named after

the gas in between the cathode and anode. Neon, xenon and mercury are examples

of the gases being used in arc lamps. Figure 2.6 shows the structure of an arc lamp.

The spectrum of the lamp is presented in Figure 2.2.

Figure 2.6: Structure of an Hg arc lamp. The electric discharge between the cathode and theanode ionizes the gas between them. Excited atoms get back to their stable state and release theenergy in the form of light. The wavelength of the emitted photon depends on the difference of theenergy levels. Figure taken from ZEISS microscopy online campus.

2.3.2 Light Emitting Diode

A light emitting diode is a semiconductor light source very similar to basic diodes.

The difference is that, in a light emitting diode, when voltage of the anode exceeds

that of the cathode by more than the diode’s forward voltage, it emits light. LEDs

work based on electro-luminescence effect. When voltage greater than forward voltage


drop is applied to the LED, current flows and electrons start to move and combine

with holes. This recombination releases energy e·Vg, where Vg is the band gap voltage

of the LED and e is the electron charge. This energy determines the frequency and

therefore wavelength of the emitted photon. Figure 2.7 shows the structure of a LED

and how electrons flow through the junction, recombine with holes and emit photon.

As incoherent light sources, LEDs are important because of their long life and high

efficiency which make them favorite in so many applications such as traffic lights.

Figure 2.7: Diagram of a light emitting diode. Applying voltage to the LED will results in currentflow within the junction. When electrons recombine with holes, they release their energy in the formof light. This phenomenon is called electro-luminescence effect. Figure taken from Department ofPhysics at Warwick University.

In microscopy, especially in confocal microscopy, a high brightness light source

is needed. Although high power LEDs have been developed that can provide high

power. They achieve high power because of their large area therefore, they can not

provide enough radiance required for confocal imaging. Recently, high brightness

LEDs have been developed which are discussed in the following section.


2.3.3 High Brightness LED

High power LEDs have different application due to their high power and lifetime.

They are being used in room lighting, mobile lighting and so many applications. But

as discussed before, high power is not enough for microscopy. Especially in confocal

microscopy where it is required to focus light onto a diffraction limited spot, high

brightness sources play a critical role. Recently, high brightness LEDs have been

developed. Not only, they can produce a large number of photons per second, but

also they can provide this large number of photons per second from a very small area.

For the first time in confocal microscopy, we have used a high brightness LED in our

line scanning confocal system. Specifications of the LED that we used, are presented

in more details in Chapter 4 with the experimental results.

2.4 Summary

In this chapter, we have discussed general properties of a light source proper for

confocal microscopy. We discussed laser and non-laser light sources. We have used

a continuous wave laser diode and a high brightness LED in our experiments. Tabel

2.1 summarize main properties of typical light sources.

Spectral Range Power Radiance (W/m2/sr) LinewidthTungsten 360− 1000nm ≈ 6× 105 wide

Hg Arc (HBO50) 250− 600nm 50W ≈ 4× 107 wideLaser Diode 830 nm 24 mW 3.5× 1010 ≈ 1nm

HE-NE Laser 632 nm ≈ 50mW 1.25× 1011 0.01 nmHB LED 750-850 nm 30 mW ≈ 9.6× 103 32 nm

LDLS 170 - 800 nm ≈ 5× 106(λ = 500nm) wide

Table 2.1: Summary of light sources properties for different light sources..

Chapter 3

Scattering Medium and Pupil

Configurations

3.1 Scattering Media and Gating Mechanism

Generally, the goal in imaging is to capture specific photons coming from the speci-

men to form the image. Multiply scattered photons, which are generally unwanted,

can reach the detector and contribute in the image formation. A multiply scattered

photon is defined as a photon which originates from somewhere other than the imag-

ing area of interest, having been scattered multiple times. Usually, the detector is

focused so that it detects photons coming from the area of interest. However, some

of the multiply scattered photons reaches the detector. These photons contribute

destructively in the image formation and reduce contrast and resolution of the image.

Figure 3.1 shows how multiple scattered photons can reach the detector and cause

image degradation.

In most biomedical applications, the medium is highly scattering. For instance,

40

CHAPTER 3. SCATTERING MEDIUM AND PUPIL CONFIGURATIONS 41

human skin is a highly scattering medium [19]. In such cases, one of the biggest

objectives is to reject multiple scattered photons. There are many gating mechanisms

to reject unwanted photons such as confocal, coherence ,and polarization gating. The

principle of each of the grating mechanisms is briefly described below.

Figure 3.1: Photons coming from the green region pass through the pinhole. These are desiredphotons. On the other hand, there are some photons coming from entrance window, but will berejected by the aperture stop (shown in yellow). There is also another type of photons, calledmultiply scattered photons (shown in red). These photons seem to come from the entrance window.These photons pass through the pinhole and destructively contribute in the image. These photonsneeds to be rejected in order to obtain a better image.

• Confocal

This gating mechanism led to the invention of confocal microscopy which is

widely used in a variety of applications. As aforementioned, the main idea of

confocal microscopy is to use a pinhole as a spatial filter to reject unwanted

(multiply scattered) photons. In this mechanism, photons are rejected based

on their place of origin. If a photon is originated from somewhere other than

focal spot, the pinhole will reject it. Therefore, it rejects photons coming from

anywhere other than the focal spot. Figure 1.14 shows how a pinhole rejects

unwanted photons.


• Coherence

The idea of coherence gating is to select only those photons coming from the area

of interest in specimen based on their arrival time. This is the base of optical

interferometry and Optical Coherence Tomography (OCT) which have a wide

range of applications in both medical and industry. As illustrated in Figure 3.4,

the beam is first separated into two beams. The reference beam reflects back

from the stationary mirror. The other part of the beam goes to the specimen

and after interaction with the specimen, comes back and both beams recombine.

Depending on the different optical path length that each beam traveled, the two

beams combine coherently (constructively) or incoherently (destructively) [20].

Figure 3.2: Transmission through a medium. Green photon transmits without scattering (desired).Red photon undergoes multiple scattering events which increases the traveling distance, hence thearrival time increases. Detected signal is shown in Firgure 3.3

• Polarization

It has been shown that multiple scattering depolarizes polarized beams [21,22].

This phenomenon is being used in polarization gating. In order to reject multiple

scattered photons, one method is reject the depolarized beam. As a depth

selective technique, polarization gating is simple and inexpensive to implement,


Figure 3.3: Arrival time for photons shown in Figure 3.2. Green signal corresponds to transmittedphotons without scattering. Red signal corresponds to multiply scattered photons reached thedetector traveled longer distance. The detector is turned on for the gating window, therefore, redsignal is rejected.

but it is unable to provide deep penetration detection. Figure 3.5 shows how

a polarizer transmits a specific polarization and reject the rest of the incident

beam. This phenomenon is the base of polarization gating.

Rejecting unwanted photons can considerably increase image contrast as is shown

in Figure 3.6. The image to the left is captured without using a polarizer. Knowing

that most of the light reflected from the surface of the water is S polarized, one can

reject the specular reflection form the surface of the water by utilizing a polarizer

that blocks S polarization.


Figure 3.4: Basics of interferometry. Based on the optical path length, beams from reference andsource arms interfere constructively (on the right) or destructively (on the left).

3.2 Pupil Configurations

As discussed, rejecting more of the multiple scattered photons leads to image improve-

ment. We have utilized different pupil configurations for our line scanning confocal

microscope in order to study the effect of different pupil configurations on rejecting

unwanted photons. In the following three sections, we have discussed full pupil config-

uration as the most popular configuration. It utilizes the whole numerical aperture of

the objective. Our goal is to study divided pupil configuration and compare it to full

pupil configuration. In divided pupil configuration, half of the numerical aperture is

being used, therefore we lose resolution, but it is able to reject more of the unwanted

photons. As an intermediate step between full pupil configuration and divided pupil

configuration, we have discussed half pupil configuration.

3.2.1 Full Pupil Configuration

In general, most microscopes use the full pupil aperture for both illumination and

detection. In the case of confocal microscopy, a high numerical aperture objective

lens is required. As mentioned earlier in Equations 1.1 and 1.3, it is critical to use full


Figure 3.5: Polarization gating. Light coming from the light source is unpolarized. After it passesthrough the polarizer, only the portion of the light that its polarization matches with the polarizersorientation goes through. In case of microscopy and rejecting multiple scattered photons, the lightcoming from the specimen is deplorized after multiple scattering events. Therefore, by utilizing apolarizer with the right orientation, one can reject unwanted photons. Figure taken from AmericanPolarizers Inc. website.

numerical aperture in order to obtain a good lateral and axial resolution. Although

using the full numerical aperture provides better resolution, it will collect more of

the unwanted multiple scattered photons which degrade the image. Since, confocal

microscopy requires high numerical aperture, we have used an oil immersed objective

lens with NA = 0.9 and 60X magnification. Resolution measurements has been

done and is discussed in Chapter 4. Figure 3.7 shows how utilizing the full pupil

configuration leads to gathering more of the unwanted multiple scattered photons.


Figure 3.6: Image of the surface of the water. Reflection form the surface of the water is Spolarized. In order to reject the S polarization, a polarizer is used (image on the right) whichimproves the contrast.

3.2.2 Half Pupil Configuration

As stated, the full pupil configuration provides good resolution since it utilizes the

full numerical aperture of the objective lens, but it collects more of the unwanted

multiple scattered photons. In comparison to the full pupil configuration, the half

pupil configuration is more successful in rejecting unwanted photons, but in the case

of the half pupil configuration, microscope uses only half of the numerical aperture

of the objective lens. Based on Equations 1.1 and 1.3, the half pupil configuration

loses lateral and axial resolution by a factor of two and four, respectively. Therefore,

by utilizing the half pupil configuration, more of the unwanted photons are rejected,

but this is done with the cost of losing lateral and axial resolution.

3.2.3 Divided Pupil Configuration

We have established that the full pupil configuration provides better resolution in

comparison to the half pupil configuration. Koester et al. proposed a combination of


Figure 3.7: Multiple scattering photons in full pupil and half pupil configuration. It is shown thatin full pupil configuration (on the left), unwanted multiple scattered photons can pass through thewhole pupil. In case of using the half pupil configuration (on the right), statistically, half of theeunwanted multiple scattered photons is rejected by pupil stop ,but it leads to losing resolution sinceonly half of the numerical aperture is being used.

the full pupil and half pupil configurations that can provide benefits of both config-

urations [23, 24]. The divided pupil configuration uses half of the objective aperture

for illumination and the other half for detection. Therefore, paths for the incident

and the reflected beams are completely separate. As reported, although utilizing half

the numerical aperture will reduce lateral and axial resolution, divided pupil config-

uration can recover some of the lost resolution by rejecting more of the unwanted

photons [25]. Figure 3.8 represents divided pupil configuration proposed by Koester

et al.

The divided pupil configuration has been used in line scanning confocal micro-

scopes for imaging human tissue [26]. We are going to discuss more about the divided

pupil configuration in Chapter 4 since we have used the divided pupil configuration

in our line scanning confocal microscope.


Figure 3.8: Divided pupil configuration uses half of the pupil for illumination and the otherhalf for detection. Multiple scattered photons are drawn for two cases. In case (a) the photondoes not contribute in the image since it is not in the illumination path. In case (b) the photonreaches the detector. Therefore, statistically, divided pupil configuration works the same as halfpupil configuration in terms of rejecting unwanted multiple scattered photons.

3.2.4 Comparison

Qualitatively, we mentioned that the full pupil configuration provides the best reso-

lution in between the three discussed configurations, but it is weak in the background

rejection. Therefore, in terms of resolution, it is expected that the full pupil configu-

ration works the best and the divided pupil configuration is better than the half pupil

configuration. Studies on line scanning microscope have proved the above theory.

Gareau et al. used the line scanning confocal microscope for imaging the epidermis

in human skin. It is reported that the divided pupil configuration provides lateral

resolution and sectioning ability comparable with point scanning system. Table 3.1

represent results reported by Gareau et al [27]. Later in the chapter 4, in our experi-

ments, we have reported our results for both divided and full pupil configurations in

line scanning confocal microscope. Our results are consistent with results published


by Gareau et al.

Axial resolution Lateral ResolutionFull Pupil 1.7± 0.1µm 0.8± 0.1µm

Divided Pupil 1.7± 0.1µm 1± 0.1µm

Table 3.1: Comparison between full pupil configuration and divided pupil configuration on theline scanning confocal microscope as reported by Gareau et al [27].

Chapter 4

Experiments and Results

The core of our research is confocal microscopy and we have discussed physics behind

confocal microscopy in previous chapters. In the following three sections, we discuss

our three main experiment on confocal microscopy. In section 4.1 our project on dual

wedge confocal microscope is discussed. Dual wedge confocal microscope was a point

scanning confocal microscope working in reflectance mode with two rotating prisms

as scanning system. Our goal was to equip the microscope with a PMT to detect

fluorescence signal from melanin particles. Section 4.2 presents our next project on

line scanning confocal microscope. Following our discussion in Chapter 3, we have

used three different pupil configurations mentioned before to compare their ability to

reject unwanted photons and their resolution. The last part of this chapter discusses

our third project on confocal microscopy. Section 4.3 discusses our experiment on

utilizing a high brightness LED as the light source for our line scanning confocal

microscope. For the first time, we have replaced the laser with a high brightness LED

and successfully obtain an image from our line scanning confocal microscope.

50

CHAPTER 4. EXPERIMENTS AND RESULTS 51

4.1 Dual Wedge Confocal Microscope

As mentioned in previous chapters, point scanning confocal microscopy requires scan-

ning in two dimensions. It can be done either by scanning the beam on the specimen

or moving the specimen with stationary beam. Besides the cost and complexity of

utilizing two scanners, it leads to lower frame rate, specially if the field of view is big.

Dual wedge scanning reflectance confocal microscopy is designed to overcome these

problems. Instead of two mirror scanners, it utilizes two rotating prisms to scan the

beam on the specimen. This is an easy process to find the location of the scanned

spot, although we lose the uniformity in scanning area. the following section discuss

our research on point scanning confocal microscope with dual wedge scanning system.

4.1.1 Dual Wedge Scanner

The dual wedge scanner was a point scanning confocal microscope working in re-

flectance mode designed by Warger et al [6]. The goal behind designing the dual

wedge confocal microscope was to scan the field of view faster and easier in compar-

ison to other scanning systems. As discussed, in general point scanning microscope,

two scanners scan the beam on the specimen in two dimensions, which requires more

space and also costs more. The dual wedge scanner has the same basic components

of a reflectance confocal microscope except for the scanning system. The scanning

system consists of two prisms back to back on their bases. As illustrated in Figure 4.1,

a beam passing through one prism refracts due to mismatch in the index of refraction

between the prism and air. The angle of refraction depends on indices of refraction

and the prism apex angle (α). Now if the prism rotates, the beam starts to rotate

on the pattern as in Figure 4.1 (a). If two prisms placed back to back on their bases,


are used in front of the beam, depending on their speed and direction of rotation, the

beam travels in different patterns.

Figure 4.1: Basic of the dual wedge scanning system. (a) Refraction by one prism. Rotating theprism will cause the beam to travel a circular pattern. (b) Refraction by two rotation prism. Eachprism will generate a circular pattern. (c) Combination of the patterns. Based on the direction andspeed of rotation, different patterns of scanning can be obtained [6].

Figure 4.2 shows two different scanning patterns of the dual wedge reflectance

confocal microscope scanner. If the prisms rotate in the same direction, scanning

pattern (a) is obtained, and if the prisms rotate in the opposite directions, pattern

(b) is achieved for scanning.

Pattern (a) in Figure 4.2 covers mostly the surrounding area of the field of view

and pattern (b) covers mostly the center area of the field of view. In either case, the

dual wedge does not uniformly scan the whole field of view. Considering that the dual


Figure 4.2: Two general patterns of scanning. (a) is achieved when prisms rotate in the samedirection. (b) is achieved when prisms rotate in the opposite direction [6].

wedge scanning system is much cheaper and easier to use in comparison to the raster

scan with two scanner, we sacrifice uniformity in the scanning area to achieve a faster

and simpler scanning system. Two motors are connected to the Labview software,

and their phase, and speed of rotation are controlled by the computer. Also, a data

acquisition card is connected to the software. Therefore, based on the position of

prisms, software detects the position of the scanning spot in the field of view. The

image of the field of view forms based on the position of the scanning spot and data

collected from the detector at the right time. Figure 4.3 shows a sketch of the dual

wedge confocal microscope. The microscope works in reflectance mode. The beam

from the laser diode with λ = 839nm goes through the polarizing beam splitter. The

beam then goes through a Rayleigh pair to the quarter wave plate. The quarter

wave plate changes the polarization of the beam from linear polarization to circular

polarization. Reflected light from the sample is gathered by the objective lens. The

reflected beam travels the same path to the polarizing beam splitter. Since the speed

of light is much higher than the prisms’ speed of rotation, the prisms de-scan the


reflected beam. The reflected beam is assumed to consist only the photons that have

been scattered once. Therefore, the reflected beam is still circularly polarized. After

the reflected beam has passed the quarter wave plate the second time, its polarization

is perpendicular to that of the incident beam. Hence the polarizing beam splitter

reflects the de-scanned beam to the detector. A lens focuses the de-scanned beam

onto the pinhole confocal to the focal spot on the specimen. The avalanche photodiode

converts the beam into an electrical signal which is then converted into digital data

that represents the brightness of the scanned spot on the specimen.

Figure 4.3: Diagram of the dual wedge scanning reflectance confocal microscope.

Figure 4.4 shows images taken by the dual wedge scanning reflectance confocal

microscope as reported by Warger et al.

4.1.2 Fluorescence Modes

The microscope designed and built by Warger et al. was working only in reflectance

mode. In our project, we wanted to add a fluorescence detection mode to the setup.

The goal was to use a continuous wave laser diode to excite the melanin molecule

stepwise and detect the fluorescence signal utilizing a photomultiplier tube. Melanin


Figure 4.4: Images of (a) cellulose fibers within paper and (b) the cellular structure within a plantleaf taken by dual wedge point scanning confocal microscope.

is the pigmentation particle in skin. The concentration of melanin determines the

color of the skin. Interest in melanin detection arises from its effect on different

medical problems. It has been shown that cancerous lesions in the skin has a higher

concentration of melanin. Also, degradation of melanin over time could be associated

to the development of melanoma or skin cancer [28, 29]. Therefore, detection of

melanin becomes important since concentration of melanin can be a good clue about

development of melanoma or skin cancer. Kerimo et al. reported that melanin can

be imaged by a stepwise three-photon fluorescence excitation process [30]. Figure

4.5 illustrates the activation process in three photon stepwise fluorescence activation.

Kerimo et al. used a femtosecond laser as is typical from multi-photon excitation,

but also showed that the stepwise nature of the excitation allowed the use of CW

laser source. Therefore, the combination of the confocal reflectance and the stepwise

fluorescence can differentiate melanin from other high reflectance material.

Our work was to equip the dual wedge scanner confocal microscope with the

fluorescence mode. In order to do that, we needed to add a photomultiplier tube to


Figure 4.5: Fluorescence process. In 1-photon fluorescence, electron absorbs a photon and goesto a higher level of energy. This is a real state. The excited electron returns to its stable level andemit a photon with slightly higher wavelength in comparison to the primary photon. In 3-photonexcitation, electron absorbs 3-photon and stepwise. It can go up to three real states as in the ploton the right. In our case, the melanin can absorb three photon and stepwise go up three virtualstates and then emit a photon.

detect the weak fluorescence signal. Slight modifications and new design were done

as shown in Figure 4.6. The quarter wave plate is part of the reflectance mode,

therefore, in order to detach it from fluorescence mode, it was relocated to in front of

the polarizing beam splitter. The Rayleigh telescope was modified to match the lens

in front of the detector.

The PMT used was a Hamamatsu 3907-03, which displays a fairly high sensitivity

even into the NIR region. The high sensitivity of the PMT in the NIR region can

be problematic for imaging the fluorescence. Even a small fraction of the laser light,

reflected from the dichroic to the PMT, could be more intense than the three-photon

fluorescence. This would have limited the specificity of the PMT signal to the three-

photon fluorescence. To overcome this problem, an 800nm short-pass filter was added

in front of the PMT. The filter was verified to be robust enough to block the laser


Figure 4.6: New design of the dual wedge point scanning confocal microscope. New parts are thePMT, the dichroic mirror and the filter and the lens in front of the PMT. Red lines represent lightpath for reflection mode and yellow lines represent fluorescence mode.

source at the beginning of the optical path, where the intensity is at its highest.

Figure 4.7 shows the transmission of the combination of the filter with the dichroic

mirror. This process ensured no reflection would leak into the PMT. Therefore, any

image captured by the PMT will consist of only of the fluorescence signal. In front

of the PMT, a 1mm diameter pinhole was also added. The purpose of the very wide

pinhole was to limit the amount of light, and hence noise, at the PMT side, rather

than to provide confocality.

The location of the equipment, which was added to our previously existing setup,

was chosen to be in between the two telescope lenses, in a pupil plane. This location

is optimum for collecting the fluorescence signal where it serves a dual purpose:

• The loss of fluorescence signal is minimal.

• It is before the beam de-scans, so it is stationary in the plane of the aperture.

This allows the use of a smaller aperture that helps rejection of unwanted light from

entering the PMT and interfering with the measurement.


Figure 4.7: Transmission of the filter in front of the PMT and reflection of the dichroic mirror.The goal was to block any leak from the incident beam to the PMT. The fluorescence wavelength isabout 450nm. The spectrum of the combination of the filter and the dichroic mirror shows that wewere successful in rejecting 839nm wavelength of the laser.

4.1.3 Results

The laser power at the sample was measured to be about 0.5mW . This power is

enough to activate the melanin particles in order to see the enhanced melanin emission

[30]. Figure 4.8 shows the images obtained from both confocal reflectance and three-

photon fluorescence modalities, as well as the merged image of the two modalities

(Figure 4.8 (a), (b), and (c) respectively). It can be seen from the images that even

in confocal reflectance mode, pure melanin strongly reflects the NIR light. In Figure

4.8(b) the background information is completely suppressed, and only fluorescence

from specific areas containing melanin is observed. From the merged image, it can be

seen that the locations that exhibit high three-photon emission are perfectly registered


to the location of the melanin particles.

Figure 4.8: Images of enhanced emission of Sepia melanin in atmosphere. (a) confocal reflectanceimage. (b) three-photon image. (c) merged image. Scale bar is 10µm

The suppression of all background information, when switching between confocal

reflectance and three-photon fluorescence mode, suggests that the origin of the signal

is indeed the step-wise fluorescence process. This background suppression was only

possible after adding the filter in front of the PMT. Prior to having the additional

filter, the image included the USAF target. This resulted from the very small amount

of reflected IR light that entered the PMT. This image provided the same information

as confocal reflectance, however, without the sectioning ability. That image was also

lacking the needed specificity to melanin. Similarly, Figure 4.9 displays a series of

images in the same order as in Figure 4.8, but of a dark human hair instead of melanin

particles. In this case as well, the location of the fluorescence is perfectly aligned with

the location of the hair sample, and no signal was seen elsewhere, where melanin is not

present. In order to better demonstrate the fluorescence using the hair sample, the

confocal image, Figure 4.9(a), and the three-photon excitation emission Figure 4.9(b)

are slightly misplaced along the optical axis. By focusing the confocal reflectance

on the substrate, the hair appears darker than the background. The fluorescence

image remained focused on the hair. Therefore, when merging the confocal with the


excited fluorescence images, the fluorescence sites will have a better contrast and help

demonstrate the phenomenon. As can be seen from Figure 4.8 and Figure 4.9, both in

the case of pure melanin particles from Sepia and the case of melanin from the human

hair sample, not all areas that appear to have melanin present actually demonstrate

three-photon fluorescence.

Figure 4.9: Images of a dark human hair in atmosphere. (a) confocal reflectance image. (b)three-photon image. (c) merged image. Scale bar is 10µm.

In Figure 4.8, the pure melanin case, some of the melanin particles stuck to the

cover-slip, while others remained on the surface of the USAF target. The gap between

the two surfaces can very easily exceed 20µm, which is approximately the length of

the excited volume along the focal axis. Therefore, some regions will remain outside of

this focal volume, where the three-photon fluorescence excitation is generated. These

regions correspond with particles that did not appear to be emitting any fluorescence

when switching modalities. In the human hair case of Figure 4.9, the sample was not

held under a cover-slip. In this case, however, changes along the focal axis, which

occurred due to the wavy nature of the hair will have the same effect. In addition, the

melanin distribution inside the focused region, is in the form of spot shaped clusters,

as typical to the distribution of melanin in the hair.


4.1.4 Conclusion

The three photon stepwise fluorescence was previously observed by Kerimo et al.,

but in this project, we were able to activate the process using a continuous wave

laser with dual wedge confocal microscope which is much cheaper and easier to use.

We have shown that a melanin-specific fluorescence channel can be easily added to a

reflectance confocal microscope. We were able to image melanin fluorescence in both

pure melanin and human hair. Results that our dual wedge scanner setup is able to

detect melanin in both reflectance and fluorescence modes.

4.2 Line Scanning Confocal Microscope

We have discussed that point scanning systems require scanning systems in two di-

mensions which makes the data acquisition time slower and adds to the complexity of

the system. Besides good resolution, in-vivo imaging requires faster image formation.

One way to form the image faster is to use line scanning confocal. Although line

scanning is confocal in only one direction, it provides comparably good resolution.

On the other hand, in reflectance confocal microscopy, we are interested in single

scattered photons. Photons that undergo many scattering effects do not provide in-

formation from the interest area of imaging. In order to study the effect of different

pupil configurations in rejecting unwanted multiply scattered photons, we have used

3 different pupil configurations on line scanning confocal microscope. The following

section represents our study on different pupil configurations.


4.2.1 Instrumentation

The advantage of using the line scanning confocal microscope is that it is faster,

much cheaper and easier to work with. Besides dealing with a more complicated

system, the points scanning confocal microscope with two scanners costs more. One

way to simplify the system is to use one scanner and instead of working in the point

scanning mode, work in the line scanning mode. The microscope was previously

designed by Dwyer et al. [3]. In our project, the main goal was to try different pupil

configurations and measure the lateral and the axial resolution of the line scanning

confocal microscope for each pupil configuration. In order to study different pupil

configurations, our setup has multiple image and pupil planes. Figure 4.10 shows

the schematic of the microscope. Note the image and pupil planes that have been

distinguished by different numbers. Aperture stops were applied on pupil planes 1

and 3 for the illumination and the receiving paths respectively.

The light source is a laser diode with the wavelength of 830nm. The beam passes

through the cylindrical lens that focuses it onto a line on pupil plane 1. Then the

beam splitter reflects half of the beam to the Rayleigh pair with unity magnifica-

tion. On pupil plane 2, a galvanometric scanner sweeps the beam on the sample to

scan the field of view (FOV). The beam then goes through another Rayleigh pair

with unity magnification which focuses it onto the back focal plane of the objective

lens. The objective lens is an Olympus LUMPlanFI/IR, 60X, NA=0.90, water im-

mersion objective lens. It focuses the beam onto a line on the sample and collects

the backscattered light. The back scattered light goes through the Rayleigh pair to

the galvo. The galvo de-scans the backscattered light. Half of the backscattered light

passes the beam splitter. A lens in front of the detector is used to focus the back

scattered light on the line detector. The detector is a linear CMOS array detector


Figure 4.10: Schematic of the line scanning confocal microscope. Note image and pupil planes.We have vertical lines in image planes and horizontal lines in pupil planes. Cylindrical lens focusesthe laser to a line. First pupil stop is applied at the cylindrical lens focal spot. Galvo scanner isused to swipe the line on the sample, and backscattered light travels the same path to the beamsplitter. Second pupil stop is applied after the beam splitter at focal spot of the telescope lens #1.Detector lens focuses the backscattered light on the detector array.

(ELIS1024, Panavision Imaging, Homer, NY). It contains 1024 pixels with each pixel

about 7× 125µm. Considering the setup magnification (about 65X), it is not neces-

sary to use a slit in front of the detector. The detector is synchronized with the galvo

scanner, and both are connected to the NI-DAQ card on the computer. The Detector

signal is being sampled with 10KHz frequency for each line of scanning and with

a trigger of 10Hz frequency, a frame is captured. The national instrument software

allows us to change the number of lines in a frame. The final image is 640 × 1024

pixels.


4.2.2 Resolution Measurement

We used a US Air Force target(USAF) to study the specifications of the microscope.

Figure 4.11 represents an image of USAF target taken by the line scanning confocal

microscope. The bright part of the image is related to the part of the target which

reflects 96% light and the dark part is related to the transparent part of the target

(uncoated glass) which is assumed to be 4% reflective.

Figure 4.11: USAF target under the line scanning confocal microscope. Note the size of thesmallest bars in the image is about 2.19µm.

Transverse Resolution Measurement

In order to measure the transverse resolution of the microscope, we have applied the

”90%-10% ” criterion [26] on the knife-edge images. Figure 4.12 illustrates how the

”90%-10% ” criterion works for the transverse resolution measurement. Transverse

resolution is defined as the spatial distance between green dashed lines.


Figure 4.12: 90%-10% method to find the transverse resolution. Upper red dashed line shows the90% of the difference between the maximum and the minimum; lower red dashed line shows the10%.

In order to measure transverse resolution of the microscope for different configu-

rations, we have used a knife-edge in both horizontal and vertical directions. Figure

4.14 represents knife-edge images for three different pupil configurations. In Figure

4.14, (a) shows the knife-edge imaged with the full pupil configuration. (b) shows

the divided pupil configuration. Utilizing the divided pupil configuration, we lose the

resolution since we are using half of the numerical aperture. Although, divided pupil

configuration is able to recover some of the lost resolution as shown in Figure 4.14

(c).

We have applied ”90%-10% ” criterion on the knife-edge images to measure trans-

verse resolution for both directions. Results are summarized in Table 4.1.


Axial Resolution Measurement

To measure axial resolution, we have used a nano-mover to move the sample in the

axial direction. We have used a mirror as sample and measure the power of the de-

tected light. Therefore, when the sample is in focus, we measure the highest detected

power, and as we move the mirror out of focus, we will lose power. Distribution of

power versus axial distance from focus will show the axial resolution. Figure 4.15

shows the power distribution versus axial distance from the focus for three different

pupil configurations.

Figure 4.13: Measurement of the axial resolution. The axial distance between two green dashedlines is defined as axial resolution.


4.2.3 Results

Results are summarized in Table 4.1. As we expected, the full pupil configuration has

the best transverse resolution and the divided pupil is next, and we expected that

the half pupil configuration to be the worst and results, confirms our expectations.

Figure 4.14: Knife-edge images for three different pupil configurations. a. divided pupil config-uration, b. half pupil configuration and c. full pupil configuration. Small part of each image ismagnified to show the transition from completely reflecting(bright) to completely transparent(dark)surfaces. Faster transition indicates better resolution. Fluctuations are result of synchronizationsignals in the data acquisition which happens in the direction of orthogonal to scanning direction.

Configuration Lateral(V) Lateral(H) Axial Pixel size Sampling distanceFull Pupil 0.5 µm 1 µm 1.75 µm 130 nm 157 nmHalf Pupil 1.1 µm 1.7 µm 4.62 µm 130 nm 157 nm

Divided Pupil 0.8 µm 1.7 µm 2.37 µm 130 nm 157 nm

Table 4.1: Summary of the resolution measurement. Pixel size is the size of each individual pixelon the sample. Sampling distance corresponds to scanning distance on the sample.

In our next step, we were able to image a business card. The image is presented

in Figure 4.16.


Figure 4.15: Axial scan for three different pupil configurations. Full pupil has a higher valueat the peak, because using the full pupil for illumination; more power is delivered to the sample.In half pupil configurations, we lose sectioning ability as plot is flattened. Utilizing divided pupilconfiguration will recover some of the loss of sectioning ability.

4.3 High Brightness LED as Light Source

As discussed in previous chapters, in order to obtain an image in confocal microscopy

a high radiance light source is required. Usually lasers provide such radiance, but

our goal in this project was to examine the performance of recently developed high

brightness LED as the light source. The LED was kindly provided by our collaborators

at (Suzhou Institute of Biomedical Engineering and Technology Chinese Academy of

Sciences, Suzhou, China). Since the LED does not have spec sheet, the first step of

the project was to find basic specifications of the LED.


Figure 4.16: Image of a business card fibers under the line scanning confocal microscope using thefull pupil configuration.

4.3.1 LED Specifications

The LED we used does not have a spec sheet, therefore, we did some experiments

to find the properties of the LED. In this section we are going to discuss important

characteristics of a light source as mentioned in Chapter 2.

• Wavelength

Figure 4.17 shows the emission spectrum of the high brightness LED. The peak

of the spectrum is at 812nm which is slightly different from the wavelength of

the diode laser we used for line scanning confocal system. The spectrum extends

from about 700 − 850nm and the full width half maximum (FWHM) is about

32nm.

The spectrum indicates that most of the LED power is in the range of wave-

lengths that laser diode provides (λ = 830nm).

• Radiance


Figure 4.17: Emission spectrum of the high brightness LED. Red dashed lines indicate FWHM ofthe spectrum. The maximum emission occurs at 812nm and the FWHM is about 32nm.

As discussed in Chapter 2, radiance of a light source is defined as the amount

of light the source provides in a unit of time, area and solid angle (Figure 4.18).

Therefore, the radiance of the LED is calculated, based on the total optical

power measured in front of the LED and the area of LED.

The area of the LED is measured to be around 1mm2. Also the power is

measured to be 30mW . Assuming uniformity and the solid angle Ω = 2π, the

half space that LED is radiating the light. The radiance of the LED is calculated

as in Equation 4.1.


Figure 4.18: Geometry of area and solid angle for the source and the detector.

L =Φ

AΩ(4.1)

Therefore, the radiance of the LED is calculated to be around 4.77×103Watts/m2/sr.

To compare the radiance of the LED with the radiance of the laser diode of the

previous setup, the same calculation for the laser diode has been done. As dis-

cussed before, the radiance of a laser is calculated as in Equation 2.7. The total

power of the laser diode is measured to be 24mW . Therefore, the laser radiance

is 3.48× 1010Watts/m2/sr.

As the calculations show, the laser diode provides higher radiance for the con-

focal microscope. Low radiance is the most important problem of the LEDs

and the main reason that LEDs have never been used as the light source for a

confocal microscope.

4.3.2 Design

In the new design, the laser and the cylindrical lens have been removed from the setup

and a high brightness LED with the peak wavelength about 850nm and a slit are used.


From the beam splitter, the setup is exactly the same as the previous design which is

presented in Figure 4.10. In the case of using a laser, the cylindrical lens will focus the

beam on a horizontal line on the pupil plane. In order to have the same line profile on

the same pupil plane, we discussed two possible designs to use the LED as the light

source. The LED can be placed in either the field or the pupil plane. In each case,

the numerical aperture at the slit should match with the numerical aperture of the

rest of the setup. In the new design, in order to keep every other element in the same

position, it is necessary to place the slit in the field plane. Considering magnification,

the numerical aperture at the sample after the objective lens and the size of the LED,

the first design is chosen, and a telescope combination with 5X magnification is used

in front of the LED. Figure 4.19 shows the design modification in the illumination

path for the line scanning confocal microscope. After the slit, the beam goes through

the beam splitter which reflects half of the beam to Rayleigh paired lenses and a

galvo scanner. The scanner sweeps the beam which is focused on a vertical line in

the field plane on the specimen. Then beam goes through another set of Rayleigh

paired lenses to the objective lens. Reflected light is then collected by the objective

lens and returns the same path to the beam splitter. Half of the reflected light goes

through the beam splitter. A lens is used to focus it on the detector. The detection

path is the same as in the previous setup presented in Figure 4.10.

4.3.3 Results

Figure 4.20 shows an image obtained from the line scanning confocal microscope

utilizing a high brightness LED as the light source. In comparison with what we had

previously with the same setup, but laser diode as the light source, we can see that

we were able to use a high brightness LED and still obtain a good image.


Figure 4.19: Line-scanning confocal microscope design. a) Illumination path to sample. b)Detection path from the sample to the detector. c) First design option to place LED in the fieldplane. d) Second design option is to place LED in pupil plane.

Measuring the optical power after the slit and considering 4% specular reflection

loss from each surface, plus 50% loss at the beam splitter, the estimated amount

of power reaches the detector is about 2.18mW . In comparison, our laser diode can

provide 25mW which about 4.1mW of it reaches the detector. We were able to obtain

an image with a 1-D CMOS linear array detector. We applied 90%-10% criterion [3]

on the knife-edge images to measure the transverse resolution for both directions.

Figure 4.21 shows the image of the horizontal knife-edge with the corresponding plot

for the pixel values. The response for each scanned line is assumed to be the same in

the horizontal direction, therefore; the plotted signal is the average over all the 640

horizontal lines.

Figure 4.22 represents the image of the vertical knife-edge and the corresponding

plot for the pixel values. Again, the response for each scanned line is assumed to be

the same in the vertical direction. The plotted signal is the average over the 1024


Figure 4.20: US air force target under microscope. LED is used as the light source. Field of viewis about 70µm in vertical direction and 88µm in horizontal direction.

pixels. The difference in the average signal between Figure 4.21 and Figure 4.22 is

the illumination pattern.

Table 4.2 shows the resolution results for the line scanning confocal microscope

for both the laser diode and a high brightness LED as the light source.

Vertical resolution Horizontal resolutionLaser Diode 0.5µm 1µm

LED 1.35µm 1.76µm

Table 4.2: Transverse resolution comparison between two setup. The same line scanning confocalmicroscope with a laser diode as the light source and a high brightness LED is the light source.


Figure 4.21: Image of the horizontal knife-edge with the corresponding averaged detected signalfrom the detector. The 90%-10% criterion is applied on the signal, to measure the resolution. Theresolution is defined as the transverse distance between the corresponding spots detected by thegreen dashed lines..

Figure 4.22: Image of the vertical knife-edge with the corresponding averaged detected signalfrom the detector. The 90%-10% criterion is applied on the signal, to measure the resolution. Theresolution is defined as the transverse distance between the corresponding spots detected by thegreen dashed lines.

Chapter 5

Conclusion

In this document we have discussed confocal microscopy, specifically reflectance con-

focal microscopy. In Chapter 1 basic concepts of microscopy were discussed. We

started from Ernst Abbe’s equation and definition of the numerical aperture of an

optical system. We discussed the relation between lateral resolution with the working

numerical aperture of an optical microscope. As one of the most important part a

confocal microscope, the pinhole and its functionality was reviewed. Since our re-

search was mainly concentrated on line scanning reflectance confocal microscope, at

the end of Chapter 1 we discussed slit aperture along with pinhole aperture and also

line array detectors.

Chapter 4 discussed experiments and results of our research. Starting with point

scanning confocal microscope with dual wedge scanning system. The distinct feature

of the point scanning is the scanning system which utilizes two rotating prisms to scan

the field of view. We have added fluorescence mode to the existed reflectance confocal

microscope. The confocal microscope with two imaging modalities was used to image

pure melanin and melanin in human hair. Results are published in International

76

CHAPTER 5. CONCLUSION 77

Society for Optics and Photonics [31].

Next set of our work was done on line scanning confocal microscope. We have

worked on different pupil configurations as discussed in Chapter 3. Results show that

transition from the full pupil configuration to the half pupil configuration, we lose

resolution as expected. In the half pupil configuration, we use half of the numeri-

cal aperture which results in losing transverse and axial resolution by factor of two

and four, respectively. Utilizing the divided pupil configuration, results in recovering

the lost resolution. The divided pupil configuration uses only half of the numerical

aperture like the half pupil configuration, but it utilizes one half for illumination and

the other half for detection. We realized that although divided pupil configuration is

not as good as full pupil configuration in terms of resolution, but the divided pupil

configuration is more successful in rejecting unwanted multiply scattered photons.

Therefore, we can recover some of the lost resolution utilizing divided pupil configu-

ration.

As discussed in Chapter 4 one of our research goals was to use a high brightness

LED instead of a laser light source. Therefore, Chapter 2 reviewed characteristics of

a light source. We began the discussion by general properties of a light source such

as radiance, wavelength and coherence. Radiance is the most important property

of a light source specially in confocal microscopy since it is required to deliver high

radiance to the specimen. As the most common group of light sources being used in

confocal microscopy, Chapter 2 discussed lasers and their properties. In our research

we have used a diode laser as the light source. One part of our research was focused

on non-laser light sources since we used a high brightness LED as the light source

later in our project. The third set of experiments on confocal microscopy, we have

replaced the laser light source with a recently developed high brightness LED that

CHAPTER 5. CONCLUSION 78

provides enough radiance for confocal reflectance imaging. This was the first time

that an LED is used as the light source of a confocal microscope. Unlike lasers, LEDs

are incoherent light sources. Therefore, images provided with high brightness LED

is speckle free. Also, high brightness LED is much cheaper and easier to work with

in comparison the laser diode. Results are published at 2014 40th Annual Northeast

Bioengineering Conference (NEBEC) Proceedings.

To summarize, we have discussed physics behind confocal microscopy, its limita-

tions and advantages. We have done experiment on both point scanning and line

scanning confocal microscopes, in both reflectance and fluorescence modalities. We

have discussed different pupil configurations and two different light source for the line

scanning confocal microscope and provide analysis of the resolution for each imaging

system.

Chapter 6

Future Work

With all discussion about confocal microscopy and experiments, there is still a lot

more to do in the future. Main tasks and ideas for future work are presented in this

chapter.

6.1 Quantify Multiply Scattered Photons Rejec-

tion

We discussed in Chapter 4 that full pupil configuration is poor because it utilizes

the full numerical aperture. We also showed in experiments that divided pupil is

more successful in rejecting more of multiply scattered photons, but we could not

quantify this. One idea to quantify the ability of each configuration to reject multiply

scattered photons, is to calculate Laplacian of the image for the same field of view

but in different pupil configurations. Figure 6.1 shows weight function for each cell

with its 8 neighbor cells. Close neighbors has higher weight since there is a direct

correlation between close neighbors.

79

CHAPTER 6. FUTURE WORK 80

Figure 6.1: Weights for applying Laplacian transform. The center pixel has the weight of 20, itsclose neighbors has -4 and far neighbors has -1.

This is basically the same as applying a high-pass filter to the image. It is a

fairly good assumption that unwanted multiply scattered photons contribute to the

image uniformly. Ideally, if there is no unwanted photons contributed in the image,

resulting image after applying the Laplacian transform, would be zero except for

pixels correspond to knife-edge. Therefore, the more unwanted photons leaked into

the detector, the higher the value of the elements in the resulting image after applying

Laplacian transform.

6.2 Image Larger Area with Dual-Line Scanning

We have discussed the line scanning system. Also we mentioned that a high brightness

LED can provide good resolution. The fact that an LED can replace the laser in

confocal scanning system, gives us the chance to build instruments with less cost,

since LEDs are much cheaper than lasers, which is important specially in clinical


instrumentation.

The idea of dual-line scanning system consists of two main parts. First, creating

two lines on the sample at the same time, which is easily possible by using a double

slit instead of one. Second, computing the correlation between two scanned lines to

find the scanning direction. Therefore, we can build a probe with dual-line scanning.

While the user is moving the probe on the specimen (like human skin), the first line

scans the specimen and captures the image of a line, the second line is used to register

the first line in the larger field of view to be scanned. Figure 6.2 shows how the idea

of the dual-line scanning probe.

Figure 6.2: Dual-line scanning probe. Blue arrows shows the direction of scanning the probe onthe large field of view. Paired green and red dotted lines are used together to form and register theimage.

As illustrated in Figure 6.2, user scans the probe on the large field of view to be


imaged. Paired green and red dotted lines are scanning lines of the dual-line scanning

probe. The green line is used to form the image and the red dotted line is used to

register the green line in the right place of the larger field of view.

6.3 Laser-Driven Light Sources

For the first time in confocal microscopy, we have used a high brightness LED as

the light source. In order to state a good comparison between different light sources,

one other idea of future work is to use a laser-driven light source. The light source

consists of a laser which is focused to a small spot on a high intensity plasma. The

focused beam produce high temperature in the plasma and generates light from a tiny

spot, providing high brightness and power required for most imaging and spectroscopy

application.

Figure 6.3: Principle of operation of the laser-driven light source. Figure taken from Energetiqwebsite.

In arc lamps discharge between electrodes limits the lifetime of the light source,

but the lase-driven light source, can provide high radiance and power, with lifetime

longer than arc lamps. Also, the laser-driven light source provide broad spectral

range, from 170nm through visible and beyond.

Bibliography

[1] N. S. Claxton, T. J. Fellers, and M. W. Davidson, “Laser scanning confocal

microscopy,” Department of Optical Microscopy and Digital Imaging, National

High Magnetic Field Laboratory, The Florida State University, 2006.

[2] Y. G. Patel, M. Rajadhyaksha, and C. A. DiMarzio, “Optimization of pupil

design for point-scanning and line-scanning confocal microscopy,” Biomedical

optics express, vol. 2, no. 8, pp. 2231–2242, 2011.

[3] P. J. Dwyer, C. A. DiMarzio, J. M. Zavislan, W. J. Fox, and M. Rajadhyaksha,

“Confocal reflectance theta line scanning microscope for imaging human skin in

vivo,” Optics letters, vol. 31, no. 7, pp. 942–944, 2006.

[4] C. A. DiMarzio, Optics for Engineers. Crc Press, 2011.

[5] J. Pawley, Handbook of biological confocal microscopy. Springer, 2010.

[6] W. C. Warger Ii, C. A. DiMarzio, et al., “Dual-wedge scanning confocal re-

flectance microscope,” Optics letters, vol. 32, no. 15, pp. 2140–2142, 2007.

[7] E. Linfoot and E. Wolf, “Phase distribution near focus in an aberration-free

diffraction image,” Proceedings of the Physical Society. Section B, vol. 69, no. 8,

p. 823, 1956.

83

BIBLIOGRAPHY 84

[8] S. J. Wright and D. J. Wright, “Introduction to confocal microscopy,” Cell Bio-

logical Applications of Confocal Microscopy, in Methods in Cell Biology, vol. 70,

pp. 1–85, 2002.

[9] E. H. Stelzer, “Practical limits to resolution in fluorescence light microscopy,”

Imaging neurons. Cold Spring Harbor Press, Cold Spring Harbor, pp. 12–1, 2000.

[10] J. M. Murray, “Confocal microscopy, deconvolution, and structured illumina-

tion methods,” Live Cell ImagingA Laboratory Manual. RD Goldman and DL

Spector, editors. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY,

pp. 239–279, 2005.

[11] T. Wilson and A. Carlini, “Three-dimensional imaging in confocal imaging sys-

tems with finite sized detectors,” Journal of Microscopy, vol. 149, no. 1, pp. 51–

66, 1988.

[12] A. Diaspro et al., Confocal and two-photon microscopy: foundations, applica-

tions, and advances, vol. 1. Wiley-Liss New York, 2002.

[13] B. Matsumoto, Cell biological applications of confocal microscopy. Academic

Press, 2003.

[14] M. Muller, Introduction to confocal fluorescence microscopy, vol. 69. SPIE press,

2006.

[15] A. R. Hibbs, Confocal microscopy for biologists. Kluwer Academic/Plenum Pub-

lishers New York, 2004.

[16] D. B. Murphy and M. W. Davidson, Fundamentals of light microscopy and elec-

tronic imaging. John Wiley & Sons, 2012.

BIBLIOGRAPHY 85

[17] R. H. Webb, “Confocal optical microscopy,” Reports on Progress in Physics,

vol. 59, no. 3, p. 427, 1996.

[18] K. R. Spring, “Detectors for fluorescence microscopy,” in Methods in cellular

imaging, pp. 40–52, Springer, 2001.

[19] R. Anderson and J. A. Parrish, “The optics of human skin.,” Journal of Inves-

tigative Dermatology, vol. 77, no. 1, 1981.

[20] G. Indebetouw and P. Klysubun, “Imaging through scattering media with depth

resolution by use of low-coherence gating in spatiotemporal digital holography,”

Optics Letters, vol. 25, no. 4, pp. 212–214, 2000.

[21] A. K. Fung and H. Eom, “Multiple scattering and depolarization by a ran-

domly rough kirchhoff surface,” Antennas and Propagation, IEEE Transactions

on, vol. 29, no. 3, pp. 463–471, 1981.

[22] R. K. Wang, “Signal degradation by multiple scattering in optical coherence

tomography of dense tissue: a monte carlo study towards optical clearing of

biotissues,” Physics in medicine and biology, vol. 47, no. 13, p. 2281, 2002.

[23] C. J. Koester, “Scanning mirror microscope with optical sectioning characteris-

tics: applications in ophthalmology,” Applied Optics, vol. 19, no. 11, pp. 1749–

1757, 1980.

[24] C. J. Koester, “High efficiency optical sectioning with confocal slits,” Trans. R.

Microsc. Soc, vol. 1, pp. 327–332, 1990.

BIBLIOGRAPHY 86

[25] C. J. Sheppard, W. Gong, and K. Si, “The divided aperture technique for mi-

croscopy through scattering media,” Optics express, vol. 16, no. 21, pp. 17031–

17038, 2008.

[26] P. J. Dwyer, C. A. DiMarzio, and M. Rajadhyaksha, “Confocal theta line-

scanning microscope for imaging human tissues,” Applied optics, vol. 46, no. 10,

pp. 1843–1851, 2007.

[27] D. S. Gareau, S. Abeytunge, and M. Rajadhyaksha, “Line-scanning reflectance

confocal microscopy of human skin: comparison of full-pupil and divided-pupil

configurations,” Optics letters, vol. 34, no. 20, pp. 3235–3237, 2009.

[28] S. Cotton, E. Claridge, and P. Hall, “A skin imaging method based on a colour

formation model and its application to the diagnosis of pigmented skin lesions,”

in Proceedings of Medical Image Understanding and Analysis, vol. 99, pp. 49–52,

Citeseer, 1999.

[29] J. D. Whited and J. M. Grichnik, “Does this patient have a mole or a

melanoma?,” Jama, vol. 279, no. 9, pp. 696–701, 1998.

[30] J. Kerimo, M. Rajadhyaksha, and C. A. DiMarzio, “Enhanced eumelanin emis-

sion by stepwise three-photon excitation,” in SPIE BiOS, pp. 79031X–79031X,

International Society for Optics and Photonics, 2011.

[31] Y. Mega, J. Kerimo, J. P. Robinson, A. Vakili, N. Johnson, and C. DiMarzio,

“Three-photon fluorescence imaging of melanin with a dual-wedge confocal scan-

ning system,” in SPIE BiOS, pp. 822637–822637, International Society for Optics

and Photonics, 2012.

CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH …CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH...

Documents

Transcript of CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH …CONFOCAL MICROSCOPY - LINE SCANNING SETUP WITH...