GO
2. Foundations of confocal microscopy
2.1 Optical principles of the confocal
microscope
2.2 Types of confocal microscopes
2.2.1 Tandem scanning Nipkow disk
based confocal microscope
2.2.2 One-sided Nipkow disk confocal
microscope
2.2.3 Scanning slit confocal microscope
2.2.4 Laser scanning confocal
microscope
2.3 Physical limitations of confocal
microscopes
2. FOUNDATIONS OF CONFOCAL MICROSCOPY
2.1
Optical Principles of the Confocal Microscope
What is a confocal microscope? A confocal
microscope is a type of microscope in which a thick object such as the
cornea is illuminated with a focused spot of light. The same microscope
objective that is used to illuminate a point, P1, in the object is
used to collect the scattered and reflected light from same point P1.
This is illustrated in Fig. 1. For simplicity, we show two
separate microscopes placed on opposite sides of a thick translucent object.
The microscope on the left side is used to illuminate a point (P1)
in the object; and the microscope on the right is used to detect
the illuminated point within the object. A point source
of light at the aperture S1 is focused by lens L1 onto a small spot
in the focal plane within the object. The second lens L1 is positioned
so its focus is also in the focal plane, and coincident (confocal)
with the spot of illumination. Lens L2 will form an image of the
illuminated spot at the aperture S2. All of the light that lens L2
collects from the illuminated spot will enter the aperture at S2 and reach
the detector.
How does a confocal microscope discriminate
against light that is not in the focal plane? For real, thick,
scattering objects the point of light imaged by lens L1 onto the thick
object will focus the light as a spot in the focal plane; however
there will be light of lower intensity in the double cone on both sides
of the focal plane. In Fig. 1. the light paths that are from
an out of focus plane, are shown as dotted lines. In this case, some
of the scatted light from an out of focus plane is collected by lens L2.
In this case the collected light is spread out at the aperture S2.
Only a small amount of the light that is spread out enters the aperture
S2. Therefore, the detector will detect only a small amount of the
light from the out of focus planes. This is the physical basis for
strong discrimination against out of focus light in a confocal microscope.
In Fig. 1. the diameter of the apertures are greatly enhanced for visual clarity. Clinical confocal microscopes have a single microscope objective. The single objective is typically used for point illumination of the object and to simultaneously image the point of illumination on the aperture of the light detector. The optical principle of the confocal microscope has been simply illustrated for a single point of the object being illuminated and the same point being simultaneously imaged on the detector. Both the aperture S1 and the aperture S2 are co-focused (confocal) on the same point in the focal plane. This is the origin of the word confocal.
Practical implementations of confocal microscopes
do not form an image of a single point within the object, but form
a two-dimensional image. Various scanning systems have been developed
to synchronously scan both the point illumination and the point of image
detection in order to form an image of a two dimensional optical section
of the thick object. Minsky described in his patent two methods of
point scanning. The image is built up of a number of points which
correspond to the illumination and detection volumes. These volumes
may be scanned sequentially as in a raster scan, or in parallel as in a
line scan. There are two general methods of scanning in a confocal
microscope: (1) stage scanning, and (2) beam scanning. Stage
scanning requires that the specimen be scanned across a stationary beam
of light and is not suitable for in vivo ocular observation. Beam
scanning requires that the illumination beam of light is scanned across
the eye. Beam scanning is used with clinical confocal microscopes
and with the scanning laser ophthalmoscope.
A confocal microscope has enhanced transverse
( x and y coordinates in the plane of the specimen) and axial
resolution (z coordinate, which is orthogonal to the
plane of the specimen) in comparison with a nonconfocal microscope
using the same wavelength of light and the same microscope objective.
It is the enhanced axial resolution in a confocal microscope which permits
the improved optical sectioning of thick specimens and their three-dimensional
reconstruction (Masters, 1997).
The transverse resolution of a confocal
microscope is proportional to the numerical aperture (NA) of the microscope
objective. However, the axial resolution is more sensitive to the
numerical aperture of the microscope objective. Therefore to obtain
the maximum axial resolution, and hence the best degree of optical
sectioning, it is necessary to use a microscope objective with
a large numerical aperture. For work with living cells and tissues
we recommend long working distance, water immersion microscope objectives
with high numerical aperture such as the Leitz 50X, NA 1.0 microscope objective.
The transverse resolution of a conventional and a confocal microscope are
now compared following the analysis of Wilson (Wilson, 1990). The
image of a single point specimen is viewed in reflected light. The
image intensity is given by eq. (1).
eq. (1)
where Iconv is the intensity of light from the object, J1 is the first-order Bessel function, and is a coordinate which is related to the transverse distance in the focal plane r, by eq. (2).
eq. (2)
The symbol is the wavelength, and the numerical
aperture of the objective is
For the confocal case in the presence of the
pinhole, the image is now given by eq. (3).
eq. (3)
for the confocal case the image is sharpened by
a factor of 1.4 relative to the conventional microscope. With a confocal
microscope the resolution is about 40% better than in a conventional microscope.
An experimental method to measure the axial resolution of a given microscope
objective in a confocal microscope is to measure the variation of the intensity
of the light reflected from a front surface mirror as it is scanned through
the focal plane. One measure of this resolution is the width of the
curve at one half maximum intensity. A non-confocal or standard microscope
would not show any variation in the intensity as the mirror is scanned
through the focal plane.
Wilson has shown the following treatment
for the axial resolution in a confocal microscope for imaging both points
and planes (Wilson, 1990). A confocal microscope is scanned axially
so that the intensity of light reflected from a plane mirror is detected
as a function of the distance that the mirror moves towards the focal plane.
At the focal plane the intensity of the reflected signal is maximum.
The intensity of the reflected light is given by simple paraxial theory
as eq. (4).
eq. (4)
the symbol is a normalized axial coordinate which is related to the real axial distance z by eq. (5),
eq. (5)
These equations are for plane reflectors. For point or line reflectors the equation becomes eq. (6).
eq. (6)
The optical sectioning is weaker for a point or
a line than for a plane. All of these equations refer only to bright
field imaging in the reflection mode. For fluoresce imaging, which
is incoherent light imaging, all of the equations are different.
Image quality is not only dependent on
resolution, but also is very dependent on the contrast of the image (section
2.3). In addition to resolution, the background rejection of
light out of the focal plane, and the signal-to-noise of various
types of confocal microscopes have been analyzed for various types of confocal
microscopes (Sandison and Webb, 1994).
2.2
Types of Confocal Microscopes
This section discusses a new paradigm to
visualize the living cells and tissues: the real-time confocal microscope.
The confocal microscope, invented by Minsky in 1957, is
being developed as a new optical technique for the clinical examination
of the human eye. The first time that a clinician uses the clinical
confocal microscope to observe a patient’s cornea there is wonder and excitement.
The ability to observe cellular details, nerves, scar tissue,
and keratocytes, in unstained, living ocular tissue is remarkable
and forms a new paradigm for ocular examination.
The observer will notice two improvements
in the imaging characteristics of a confocal microscope: (1) enhanced transverse
resolution, and (2) enhanced axial resolution as compared to a standard
microscope. The former improvement results in higher resolution in
the plane of the specimen. The latter effect results in the superb
capability of a confocal microscope to optical section a thick, highly
scattering specimen. This is the main advantage of a confocal
microscope.
The confocal microscope provides ”en face”
images of the cornea; the plane of the image is orthogonal to the thickness
of the cornea. These images are very different, and are oriented
perpendicular, from the typical sections obtained in histopathology
in which the tissue is cut along the thickness of the cornea (vertical
sections).
In contrast to the conventional light microscope, which images all of the points in the specimen in parallel, a confocal optical microscope optimizes illumination and detection for only a single spot on the specimen. In order to form a two-dimensional image with a confocal microscope, it is necessary to scan the illumination spot over the area of the specimen. Several generic types of confocal light microscope are now described. Details of their optical and mechanical designs are given in reprints of the original papers and patents (Masters, 1996b).
2.2.1
Tandem scanning Nipkow disk based confocal microscope
A real-time tandem scanning confocal microscope, in which the image could be observed with the naked eye, was developed by Petran and Hadravsky. They acknowledged the contribution of Nipkow who invented the Nipkow disk in 1884 to provide real-time, point illumination and point detection.
The principle of the tandem scanning confocal microscope is as follows. Sets of conjugate pin holes (40-60 microns in diameter) are arranged in several sets of Archimedes spirals. Each pinhole on one side of the disk has an equivalent and conjugate pinhole on the other side of the disk. The illumination light passes through a set of pinholes (about 100 at a time) and is imaged by the microscope objective to form a diffraction limited spot on the specimen. The reflected light from the specimen passes through a conjugate set (about 100 at a time) of pinholes on the other side of the disk and can be observed in the eye piece of the microscope. Both the illumination and the reflected light are scanned in parallel over the specimen to generate the two-dimensional image of the focal plane by spinning the Nipkow disk. This microscope is called a tandem scanning reflected light microscope.
Since the ratio of the area of the holes to the area of the disk is usually only about 1-2 percent, only a small fraction of the illumination reaches the sample, and a similar small fraction of the light reflected from the sample passes the disk and reaches the detector. Therefore, the illumination must be very bright (a xenon or mercury arc lamp is usually required). The advantages of the real-time tandem scanning confocal microscope include the following: (1) real-time operation; image acquisition is at video rates (defined as 30 images per second) or faster, and (2) true color confocal imaging, and (3) the specimen can be viewed with the eye.
These systems are best suited for reflected light confocal imaging. However, even in the reflected light mode, confocal microscopes based on a Nipkow disk containing pinholes have a very poor light throughput. The tandem scanning Nipkow disk based confocal microscope is a poor choice for weakly reflecting specimens such as living cells, tissues, and organs. The low intensity of light that reaches the detector (the eye of the observer, the film plane of a camera, or the CCD camera) results in an image with marginal image quality.
2.2.2
One-sided Nipkow disk confocal microscope
It is possible to use the set of pinholes
on the same side of the Nipkow disk for both the illumination and the detection.
Xiao, Corle and Kino invented a real-time, one sided, Nipkow disk
based confocal microscope ( Xiao et al., 1988; 1990). This design
has several advantages over the tandem scanning confocal microscope:
it is less sensitive to vibration of the Nipkow disk, it has a simplified
optical design, and it is easier to align the microscope.
In order to reduce the reflected light from
the surface of the Nipkow disk three techniques were implemented.
The disk is tilted so that the light reflected from the surface of the
disk is reflected into a beam stop. The surface of the disk is blackened
to reduce the surface reflections. A polarizer is placed between
the light source and the disk; which illuminates the disk with polarized
light. A quarter wave plate is placed between the Nipkow disk and
the microscope objective, and an analyzer is placed between the Nipkow
disk and the detector. The combination of polarizer, quarter
wave plate and analyzer effectively separates the light from the specimen
and the light reflected from the surface of the disk. This optical
arrangement sharply discriminates light reflected from the surface
of the disk; however, it also slightly reduces the light reflected
from the object that reaches the detector.
An advantage of the one-sided Nipkow disk confocal microscope is a simpler optical design as compared to the tandem scanning Nipkow disk confocal microscope. A disadvantage is that since the illumination and reflected light follow the same optical path it is not easy to correct for chromatic aberrations in the microscope. This design, as with the tandem scanning Nipkow disk based microscope, has the disadvantage of the low transmission of the disk which also makes the microscope a poor choice for weakly reflecting specimens such as living cells, tissues, and organs.
2.2.3
Scanning slit confocal microscope
An alternative to point scanning, as exemplified
in the designs of confocal microscopes based on the Nipkow disk,
is to use a slit of illumination which is scanned over the back focal plane
of the microscope objective. The advantage of this optical arrangement
is that since many points on the axis of the slit are scanned in parallel,
the scanning time is markedly decreased. Another very important advantage
is that scanning slit confocal microscopes have superior light throughput
as compared to point scanning Nipkow disk systems. The disadvantage is
that the microscope is truly confocal only in the axis perpendicular to
the slit height. In comparison to a pinhole based confocal microscope,
a slit based confocal microscope provides lower transverse and axial resolution.
This comparison is for the same wavelength of illumination and reflected
light and the same microscope objective in each case. However,
for confocal imaging of weakly reflecting living biological specimens,
the trade off between lower resolution and higher light throughput
is acceptable.
Several arrangements have been developed to provide the scanning of the slit of illumination over the specimen, and the synchronous descanning of the reflected light from the object. Many of the original papers and patents, which are described in this paper, have been republished (Masters, 1996b). The simplest design is a two sided mirror mounted on a single oscillating shaft. A design of real-time, slit scanning microscope with focal-plane specific illumination was developed by Baer. Another form of a scanning slit aperture confocal microscope was developed by Burns. Koester developed a scanning mirror microscope with optical sectioning characteristics for observation of the living eye. The Svishchev design of a two sided mirror is the technique used in several modern designs of real-time confocal microscopes with bilateral scanning (Svishchev, 1969, 1971).
Fig. 2 illustrates the optical design
of the real-time scanning slit in vivo confocal microscope developed
by Dr. A. Thaer. The design consists of two adjustable slits placed
in conjugate planes of the confocal microscope. Both scanning of
the illumination slit over the back focal plane of the microscope objective
and descanning of the reflected light from the object is accomplished with
an oscillating two-sided mirror.
There are several advantages to scanning
slit confocal microscopes (Masters, 1989; 1991a). The slit height
can be adjusted which allows the user to vary the thickness of the optical
section. A more important feature is that the user can vary the slit
height, and therefore control the amount of light that reaches the
sample as well as the amount of reflected light that reaches the detector.
This is important for samples that are very transparent and therefore can
be imaged with the slit height very small; more opaque samples require
that the slit height is increased. The microscope can operate in
real time; that is at video rates. The light throughput is much greater
for a slit scanning confocal microscope than for a confocal microscope
based on the Nipkow disk containing sets of conjugate pinholes.
The advantage of slit scanning confocal microscope over those based on Nipkow disks containing pinholes is shown in the following example. For cases of weakly reflecting specimens, such as living, unstained, cells and tissues, the advantage of the much higher light throughput from the slit scanning systems is crucial for observation. The basal epithelial cells of the normal, in vivo, human cornea can not be observed with a tandem scanning confocal microscope. However, corneal basal epithelial cells are always observed in vivo, in normal, human subjects when they are examined with a real-time, slit scanning, in vivo confocal microscope. The reason for this discrepancy is that although the tandem scanning confocal microscope has higher axial and transverse resolution the very low light throughput of the disk does not pass enough reflected light from the specimen to form an image on the detector (in a single video frame) which has sufficient signal-to-noise and, therefore, contrast to show an image of the cells.
2.2.4
Laser scanning confocal microscope
The original patent of Minsky contained the concepts that are implemented in the commercial laser scanning confocal microscopes that are used for both laboratory investigations, and also in the scanning laser ophthalmoscope (Davidovits and Egger, 1971, 1973; Young and Roberts, 1951). A laser is used as a high intensity light source and the laser beam is scanned over the back focal plane of the microscope objective by a set of galvanometer scanning mirrors. These mirrors provide scanning in both the x and y axes. A diffraction limited spot of light is scanned over a region of the specimen. The reflected light that is imaged by the microscope objective is descanned by the galvanometer scanning mirrors and imaged onto an iris diaphragm, or a pinhole aperture located in front of the photomultiplier detector.
2.3
Physical Limitations of Confocal Microscopes
As shown in Sec. 2.1 the transverse
and axial resolution in a confocal microscope is dependent on both the
wavelength of light and the numerical aperture of the microscope objective.
Shorter wavelengths of light and higher numerical apertures yield greater
resolution. As the wavelength of the light is reduced towards the
blue the scatter of the light will be increased. This can be an advantage
to image weakly scattering structures in the specimen. However, in
a highly scattering specimen, such as scar tissue in the cornea, the increased
light scatter will result in decreased penetration of light. Ultraviolet
light can excite cellular autofluorescence, and can result in cell and
tissue phototoxicity, and possible cell death.
The microscope objective is a critical element;
it determines contrast, magnification, field of view, resolution,
image quality and image aberrations. The use of high numerical aperture
microscope objectives restricts both the field of view on the specimen
and the free working distance of the microscope objective. Higher
numerical aperture objectives have higher magnifications, therefore the
area that can be observed on a specimen is reduced. They also have
shorter free working distances. The latter feature reduces the distance
that the microscope can focus into the specimen from the surface.
The intensity of light incident on the
specimen is another physical limitation. When the confocal
microscope is used in the fluorescence mode there is an intensity of the
ultraviolet or blue excitation light which saturates the absorbing fluorescent
molecules; higher intensities do not result in increased fluorescence
intensity. Photobleaching of fluorescent molecules, photodamage
and phototoxicity of cells are processes that occur, to various extents
depending on the types of cells and absorbing molecules, wavelengths and
intensities of illumination.
Spatial and temporal resolution of the scanning
system of the confocal microscope, and light detectors present additional
limitations. The Rayleigh criterion of transverse resolution is used
in imaging systems. If the center of the point spread function (Airy disk)
of one object falls on the first zero of the point spread function of the
second object, then the two objects are resolved. In this case we
can observe two distinct, adjacent points, as two separate points. The
ideal spatial resolution of an optical system is governed by the Nyquist
theorem which states that the spatial resolution of the system should be
at least two times the highest spatial frequency of the specimen.
The temporal resolution of the optical system should be sufficient to image
moving specimens without motion artifacts. Various types of confocal
microscopes have different scan speeds; that is the time to acquire
a single image of the field. Two sets of orthogonal galvanometer
mirrors usually require 1 second to acquire a 512 x 512 image. Larger
images such as 1024 x 1024 pixels require several seconds.
Slit scanning systems, solid state acousto-optic modulators, and fast rotating
polygon mirrors together with galvanometer mirrors for the slow axis
are examples of scanning systems that can acquire images at video rates
33 milliseconds per frame) or faster. High scanning speed is
important for observing moving objects such as the living human eye.
The human eye can be used as the detector
for direct view confocal microscopes, although there are advantages
to using an electronic detector. Typically an electronic detector
is used to detect the signal which is digitized and displayed on a computer
monitor. Examples of electronic detectors used with confocal microscopes
include photomultiplier tubes (PMT), avalanche photodiodes (APD),
intensified video cameras, and cameras based on a charge-coupled
device (CCD). Photomultiplier tubes have low dark noise and have
good sensitivity in the visible and the blue region of the spectrum.
Avalanche photodiodes are used at video rates in the near infrared region
of the spectrum; for example in the scanning laser ophthalmoscope.
Detectors based on charge coupled devices
are available from 512 x 512 pixels up to 2K x 2K pixel arrays. The
advantages of a CCD detector coupled to a confocal microscope include high
sensitivity (a quantum efficiency approaching 90 percent), and
high dynamic range (CCD) detectors are available with 12, and 14
bits of dynamic range (Masters, 1991b; Masters and Kino, 1993). Higher
dynamic range of the light detector increases the image fidelity, since
it can faithfully record larger increments of intensity or gray values
from the specimen.
As the number of pixels increases the time to read out the image increases. This is not a problem for imaging a stationary specimen, but for confocal imaging of the living eye rapid image acquisition is required to minimize motion artifacts. There is a tradeoff between the time to acquire a single frame and the number of pixels in the image. The signal-to-noise ratio of the detector is very important. As the number of photons detected at each pixel on the specimen approaches the noise from the detector itself there is a loss of information. For very rapid scanning systems and weakly reflecting specimens such as the eye, the number of photons detected may be insufficient to form a usable image; this is a practical limitation of the confocal microscope.
Each design of a confocal microscope is a trade off and an optimization of design parameters. The choice of light source, scanning system, intermediate optics, microscope objective and photodetector should reflect the specific requirements of the instrument (see Section 3.5). There are many sources of optical aberrations in the confocal microscope. They can be minimized but not eliminated by optimal design and selection of components. Some of the optical aberrations are specific to the optical imaging of thick specimens with varying refractive indexes, e.g. the cornea and the retina of the human eye. Confocal microscopy of the living human eye has two additional limitations. The eye is a weakly reflecting specimen with low inherent contrast. The eye is also a moving object. In addition to sufficient resolution, there must be sufficient contrast in order to observe details in the specimen. These special requirements must be taken into account in the design and operation of a clinical confocal microscope to observe the living human eye.
GO