CSA TN: 5005361111111111111111111111111111111111111111
`
`ODY I Maxcost:
`
`I 1111111111111111111111111111111 IIIII IIIII IIIII IIII 1111
`13765927
`Patron:
`
`I L :-
`
`RAPID:CS1 ILLiad 2195561111111!11111111111111111111111111111111
`
`NOTICE:
`This material may be protected
`by copyright law (Title 17 U.S. Code)
`
`Borrower: RAPID:CS1
`NEW: Main Library
`Lending String:
`
`3 October 2018
`
`.
`
`Location: 3S
`can#: Journal of microscopy
`
`Journal Title: Journal of microscopy
`
`Vol.: 165 No: 1 Mon/Yr: 1992
`Pages: 103-117
`
`Article Title: Colour confocal reflection microscopy
`using red, green and blue lasers
`Odyssey: 144.37 .5.98
`
`Search Info:
`
`C
`
`...J
`
`>-
`(1) z
`.30 Not on Shelf
`~w
`~~ Lacking
`~~ Not as Cited
`Su:
`.!: u..
`0 Notes:
`
`1st Date:
`2nd Date:
`
`Tight Binding
`Poor Condition
`
`(cid:143)
`(cid:143)
`(cid:143)
`
`Initials:
`Initials:
`(cid:143)
`(cid:143)
`
`Pulled by:
`Updated by:
`Scanned by:
`Delivered by:
`
`/
`
`41600101
`Interlibrary Loan
`Sacramento State University Library
`2000 State Univ. Dr.
`SACRAMENTO CA 95819-6039
`
`LIBRARY MAIL
`RETURN POSTAGE GUARANTEED
`
`NEW: Main Library
`San Marcos
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`Journal of Microscopy, Vol. 165, Pt !,January 1992,pp. 103-117.
`Received 7 December 1990; revised and accepted 13 May 1991
`
`Colour confocal reflection microscopy using red,
`green and blue lasers
`
`by C. J. COGSWELL*, D. K. HAMILTONt and C. J. R. SHEPPARD*,
`*Physical Optics Department, School of Physics, University of Sydney, NSW 2006,
`Australia and tDepartment of Engineering Science, University of Oxford, Parks Road,
`Oxford, U.K.
`
`KEY WORDS. Confocal microscopy, colour imaging, scanning laser microscopy, three(cid:173)
`dimensional ~ging, biological microscopy, immunogold, reflection bright-field.
`
`SUMMARY
`To obtain colour reflected confocal images we have incorporated three lasers (HeNe:
`633 nm; NdY AG: 532 nm; HeCd: 442 nm) and three photomultiplier detectors into our
`on-axis scanning system then adjusted the registration of the simultaneous output
`signals to produce full-colour images on a video monitor. Colour confocal images were
`produced from multi-stained fixed tissue as well as from natural pigments in fresh plant
`material. Rayleigh scattering properties' of immunogold-labelled specimens were
`studied to show how variations in colour response can be utilized to identify
`subwavelength gold particles. Colour stereo pairs were produced to illustrate the
`accuracy with which the three-laser microscope system can record depth information
`without incurring problems due to chromatic aberration effects.
`
`/
`
`INTRODUCTION
`The confocal scanning optical microscope is well known to have useful imaging
`properties not possessed by conventional microscopes: it has somewhat improved
`lateral resolution and, perhaps most importantly, it has the property of optical
`sectioning (improved axial resolution). For viewing thick objects, this latter feature
`allows the microscopist to produce a series of images from consecutive planes of focus
`which can then be combined digitally to form three-dimensional reconstructions
`(Carlsson et al., 1985).
`In the biological sciences, the term confocal microscopy is often assumed to be
`strictly synonymous with confocal fluorescence microscopy. There exists, however, a
`wide range of alternative confocal techniques such as reflected bright-field (Cogswell &
`Sheppard, 1990a), differential phase contrast (Benschop, 1987) and differential
`interference contrast (Cogswell & Sheppard, 1990b) which can retrieve amplitude and/
`or phase information and are based on the analogous methods of conventional
`microscopy. In general, these techniques, properly utilized, offer the usual advantages
`of confocal microscopy for observing the absorbing and reflecting properties of samples
`at some pre-determined, usually monochromatic, illuminating wavelength. A problem
`
`J;; 1992 The Royal Microscopical Society
`
`103
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`104
`
`C. J. Cogswell, D. K. Hamilton and C. J. R. Sheppard
`
`arises in confocal microscopy, however, when attempts are made to image accurately
`the overall spectral properties of specimens.
`The ability of the human visual system to detect colour is, and has been for decades,
`one of the primary factors underlying the design of conventional microscopes and the
`preparation of specimens (be they utilized in biological, geological or materials science)
`which may span applications as diverse as the location of particular biological tissues by
`incorporation of vital stains or techniques to determine crystalline mineral phases in
`polarizing microscopy. To date, however, there are only a few commercially available
`confocal instrument designs that have the ability to produce full-colour images and
`these certainly have limitations. For example, confocal microscopes which use white(cid:173)
`light sources and employ rotating Nipkow discs are easy to operate and can produce
`useful full-colour images; however, these images frequently display chromatic
`aberration effects which can create an undesirable loss in resolution. On the other hand,
`these chromatic effects may sometimes be put to good use to provide additional
`information. For example, Boyde (I 987) has used the natural chromatic aberration ofa
`microscope objective to colour-code images in the confocal tandem-scanning micro(cid:173)
`scope. Similarly, Molesini et al. (1984) have introduced chromatic aberration in a laser
`system in order to form profiles of surface topography. However, in the tandem(cid:173)
`scanning designs, because the different wavelengths of light pass through the same pin(cid:173)
`holes, it is not possible to adjust independently the foci of the various colours in order to
`cancel out or selectively introduce chromatic aberration at will. (This will be shown to
`be one of the advantages of our three-colour confocal microscope described below.)
`Another design of confocal microscope, which utilizes three laser light sources in a
`beam-scanning configuration, is produced by at le,ast one commercial manufacturer
`(Lasertec model 2LM11; Awamura et al., 1987). However, due to the inherent
`properties of beam-scanning designs (Cogswell & Sheppard, 1990a), it is extremely
`difficult to control the chromatic path differences which occur as the beam scans across
`the field. In addition, determining the optimum axial position for each of the three (red,
`green and blue) detector apertures is further complicated by the fact that different
`objectives must be utilized when a magnification change is desired.
`In view of these limitations, in this study we have chosen to utilize an on-axis 1
`specimen-scanning confocal microscope configuration with three laser illuminators and
`three detectors. This microscope design allows us to eliminate off-axis chromatic effects
`and to fine-tune the system to one single high-NA objective through the-addition of
`small-correction opticians' lenses (Cogswell et al., 1990). If a change in magnification is
`then required, we can continue using the same objective and simply alter the distance
`over which the object is scanned.
`
`THE CONFOCAL SCANNING OPTICAL MICROSCOPE
`Figure I (a) illustrates the basic principles of an on-axis scanning optical microscope
`operating in reflection. The laser beam is expanded to illuminate the objective lens,
`which focuses a diffraction-limited spot onto the object plane. The object is scanned
`across the spot in a pattern in the x-y plane similar to a video raster. Light reflected from
`the specimen passes back through the objective lens, is reflected from the half-silvered
`mirror and falls onto the photodetector, whose output, after amplification, modulates
`the brightness of a video display (scanned simultaneously with the object) to form a::
`image.
`Depending upon whether the detector is of large area or a point, there are two fonns
`of the scanning optical microscope, often referred to as Type 1 (conventional) and·
`Type 2 (confocal) (Sheppard & Choudhury, 1977). In the Type 1 microscope, the
`imaging performance is determined by just the diffraction-limited spot of ligh1
`projected onto the specimen by the objective lens, with the detector responding to a:: ·
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`tal
`
`(b)
`
`Colour confocal reflection microscopy
`
`105
`
`objective lens
`
`point detector
`
`objective lens
`
`object
`
`IC!l!l!::I point detector
`
`tel
`
`I (z)
`
`z
`
`;;ig.1. (a) (b) Schematic diagrams of the reflection confocal scanning optical microscope, showing the change
`position of the focused image with respect to the detector pin-hole as the object (mirror) is moved out of the
`cal plane. (c) A typical defocus response curve with intensity I(z) plotted along the vertical axis and the
`ocus position of the objective (z position) along the horizontal axis, with z0 and z 1 corresponding to the focus
`ositions in (a) and (b).
`
`he light reflected from the object. In this configuration, as far as the reflected light
`omponent is concerned, the objective lens acts merely as a collector. The imaging
`erformance of this form of scanning system is identical with that of a conventional
`icroscope. In the confocal configuration, which utilizes a pin-hole in front of the
`etector, the object is still illuminated with a diffraction-limited spot, but reflected light
`collected only over the area of the pin-hole. The point detector can be thought of as
`efining a region on the object which is, since the same objective is used for both the
`cident and reflected beams, another similar diffraction-limited spot. The effective
`oint-spread function (PSF) of this arrangement is thus narrower than in the
`onventional microscope.
`In addition to an improved PSF, the confocal microscope has the property of optical
`ctioning. If the object is moved out of the focal plane as shown in Fig. l(b), reflected
`ght is no longer brought to a focus on the point detector, with the result that the
`rength of the detected signal drops significantly. Out-of-focus information from the
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`106
`
`C. J. Cogswell, D. K. Hamilton and C. J. R. Sheppard
`
`"1-------------1HeNe laser
`633nm
`
`,J!f:.=-++----....,.;.-----1HeCd laser
`442nm
`
`.l4:~~~1t:====-1rn
`
`point
`detector
`
`~:r======tl=t::==:::::hl.1
`
`point
`detector
`
`M mirrors
`D dichroic mirror
`B beam splitters
`L
`lenses
`0 objective lens
`
`Fig. 2. The optical arrangement of the three-colour confocal scanning optical microscope.
`
`object is thus suppressed, rather than appearing blurred as happens in the conventional
`microscope. The optical-sectioning property is usually characterized by the so-called
`defocus response. (This is sometimes referred to as the V(z) response which is the
`analogous technique in acoustic microscopy.) The defocus response is measured with a
`perfect reflector as object, and gives the variation of the detector intensity output as the
`object is scanned axially through focus. A typical defocus response curve is shown in :
`Fig. l(c) with intensity(/) plotted along the vertical axis and the focus position of the:
`objective (z position) along the horizontal axis. The two z positions indi~ed on this
`plot correspond to the focus positions in Fig. l(a, b). A standard method for evaluating/
`performance using the defocus response technique is to measure the full width of the :
`peak at half the maximum signal intensity (FW.HM). For exampl~,...with red light and an i
`objective lens with a numerical aperture of 1 ·3, the defocus response typically has a /
`FWHM value ofless than 0·5 µm which is a reasonable indicator of the axial resolution I
`obtained using this configuration.
`
`1
`
`I
`
`
`1'
`
`A THREE-COLOUR CONFOCAL REFLECTION MICROSCOPE
`Experimental apparatus
`The optical arrangement of the three-colour confocal microscope is shown in Fig. 2.
`The laser light sources are HeNe: 633 nm; frequency-doubled Y AG: 532 nm and
`HeCd: 442 nm. Beam expansion and collimation are achieved for the red light by lensei
`LI and L3 (adjusted to give an accurately parallel beam), for the blue light by L2 andLJ
`and for the green light by L4 and LS. Lenses L2 and L4, which are low-power
`microscope objectives, are mounted so that their axial and transverse positions witl:
`respect to the optical axis are adjustable. The red and blue beams are combined witl:
`minimal power loss by the dichroic mirror D, and the green beam is introduced via tl:e
`beam splitter BI which also passes part of the reflected beam to one of the detector1
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`Colour conj ocal reflection microscopy
`
`107
`
`The objective lens is a Zeiss planapochromat with an NA of 1 ·4 which has been
`optimized for an optical tube length of 160 mm. Since the system was designed to
`operate with infinity optics, the diverging lens L9 provides the appropriate correction
`so that this objective can be used with parallel light.
`The light reflected from the mechanically scanned object passes back through the
`objective and beam-splitters B4, B3 and Bl, to give three resulting beams. These pass
`respectively through 442-, 633- and 532-nm bandpass interference filters and are then
`focused onto point detectors formed by pin-holes in front of photomultipliers. After
`amplification, the signals from the detectors pass to a digital frame store controlled by
`an LSI 11/23 minicomputer. The signals from the red, green and blue detectors are
`stored separately, and are then fed to the three channels of a colour video monitor to
`form a full-colour picture.
`
`Alignment
`The essential requirement for correct operation of a confocal imaging system is that,
`in the object plane, the focused spot formed from the source by the objective lens must
`coincide with the back-projected image of the point detector. This must be true for all
`three colours, and furthermore, all three foci must coincide, axially and transversely.
`Although chromatic aberration from the objective lens was known to be minimal, a
`small amount was present, due to the lens L9, and allowance had to be made for this by
`making small changes in the parallelism of the green and blue beams entering L9. The
`microscope ~s known to be accurately aligned for red light, and the object was
`positioned to be in focus with this colour. The pin-holes were removed from the blue
`and green detectors so that they operated in the Type 1 mode. The axial positions of
`lenses L2 and L4 were then adjusted to bring the blue and green images into focus: this
`meant that the focused spots illuminating the object coincided axially. With the pin(cid:173)
`holes replaced, the axial positions of lenses L6 and LS were adjusted to make the
`maxima of the blue and green defocus responses occur at the same axial position of the
`object as the response maximum with the red beam.
`Even though the alignment was begtin by setting the various mirrors and beam
`splitters to make the three beams follow, as accurately as possible, the same path to the
`object, fine adjustment was needed to make the three foci coincide precisely in the
`transverse plane and so give exact registration between the three images on the display
`screen. The microscope was set to a very high magnification on an object with fine
`features, scratch marks on an evaporated metal film being ideal. The blue image was
`brought into registration with the red, shifting its lateral position by moving the lens L2
`laterally a small amount at a time and each time repositioning the pin-hole in front of the
`blue detector to re-centre it on its focused spot until the images coincided. This was
`repeated with the green beam, making the adjustment with the lens L4.
`
`The defocus response
`One of the main aims in developing the three-colour confocal microscope was to use
`its optical sectioning property on biological specimens. To avoid an object changing
`colour as it goes out of the focal plane and begins to be suppressed from the image, it is
`important that the defocus responses not only have maxima which coincide, but are as
`similar in shape as possible. Such a colour change would be seriously detrimental, for
`example, if a through-focus series of images were to be processed into a stereo pair. As
`an example, consider the effect when the blue defocus response is narrower than those
`for the red and green on a feature in the object which reflects all three colours equally. In
`focus, the feature would appear white (assuming the gains of the three channels to be
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`108
`
`C. J. Cogswell, D. K. Hamilton and C. J. R. Sheppard
`
`correctly balanced), but as the focus was moved in either direction, it would become
`yellow (red+ green) as it faded in intensity. Such a situation is shown in Fig. 6(a) which
`is an image, produced while the defocus responses were still unmatched, of a mirror
`tilted with respect to the optic axis. The change in colour from white through brownish(cid:173)
`yellow away from focus is clearly visible. There is also a noticeable colour variation in
`the bottom part of the picture, which is due to a pronounced side-lobe in the green
`response.
`In an ideal aberration-free system, the widths of the defocus responses should be
`proportional to the wavelengths of the light: the blue should be the narrowest and the
`red the widest. Two possible methods exist for equalizing them, both of which require
`widening the green and blue responses to the size of the red one. The first alternative is
`to increase the diameter of the pin-hole in front of the detector. In a limited range, above
`a certain diameter which depends on the wavelength and the angle of convergence of the
`detector beam, the width of the defocus response increases with the pin-hole diameter
`(Sheppard & Cogswell, 1990). The second method is to stop down the diameter of the
`detector beam just before its converging lens. Both methods will also reduce the
`transverse resolution, but as this also depends on wavelength, the effect should be
`beneficial, tending to reduce differences that may otherwise occur between the colours.
`In practice, we found it necessary to use only the first of these two alternatives and
`two versions of this method were employed, one utilizing large pin-holes and another
`utilizing small pin-holes. The larger pin-holes were used with an object of low
`reflectivity when as much light as possible was needed on the detectors, and also when
`only moderate magnification was required so that some sacrifice in resolution was
`tolerable. Optimum matching of the defocus responses occurred with pin-holes of
`100 µm diameter on the red and green detectors, and one of 150 µm diameter on the
`blue. Figure 3(a-c) shows these responses, which match quite closely. The widths of the
`defocus responses (FWHM) were 0·93, l · 14 and l · 12 µm for the red, green and blue
`wavelengths respectively. For maximum resolution at high magnification, and when
`the light level was less of a problem, 10-µm pin-holes, which were small ,S9-:ough to give
`close to true confocal operation, were used on all three detectors. In this case the
`performance of the system differed from the ideal because of the presence of
`aberrations, and it was found that the precise axial position of the correcting lens L9 had
`little effect on the shape of the green response, but considerably-more on that of the blue
`response. This was optimized, and the resulting responses are shown in Fig. 3(d-f).
`The widths of the defocus responses are now 0·65, 0·66 and 0·65 µm for the red, green
`and blue wavelengths respectively. These should be compared with the theoretical
`widths for an aplanatic system of NA 1 ·4 which are 0·32, 0·27 and 0·22 µm respectively
`(Sheppard & Wilson, 1981). Since it was not possible to balance simultaneously the
`aberrations for all three colours, the performance of our experimental system was a
`factor of two worse than the ideal.
`Two other oil-immersion objectives (as well as various dry objectives) were
`investigated on our confocal colour system. The first was an infinity-corrected Leitz
`achromat objective of NA 1 ·36 (in this case lens L9 was, of course, removed). Although
`this objective was not an apochromat, it was possible to register the three colours. The
`second objective tested was a Zeiss Neofluar of NA l ·3 which was corrected for its 160-
`mm tube length. This gave a defocus response width of 0·32 µm with blue light, which
`was very close to the theoretical limit, but, unfortunately, due to the presence of
`chromatic aberration, could not be registered adequately for all three colours, owing to
`the dominance of the lens L9 in defining the divergence of the beam entering the
`objective. The fact that the defocus response was considerably sharper for this fluorite
`objective than for the planapochromat (NA l ·4) used in Fig. 3(f) confirms a trend we
`have observed previously (Cogswell et al., 1990).
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`Colour conj ocal reflection microscopy
`
`109
`
`Large Pinholes
`
`Small Pinholes
`
`red
`100µm
`pinhole
`
`la)
`
`green
`100µm
`pinhole
`
`lb)
`
`blue
`150µm
`pinhole
`
`le)
`
`red
`10µm
`pinhole
`
`(d)
`
`green
`10µm
`pinhole
`
`le)
`
`blue
`10µm
`pinhole
`
`(f)
`
`;ig, 3, Measured defocus responses for large pinholes: (a) 100 µm, (b) 100 µm, (c) 150 µm, and for small
`,inholes (10 µm; d-f). In both cases, the responses correspond to the red, green and blue lasers respectively,
`nd were made using a planapochromat 63 x, l ·4-NA oil-immersion objective. Total width of the horizontal
`z) scale= 5 µm in each plot.
`
`/
`
`fUMAN COLOUR VISION RESPON'SE
`In order to produce images which provide a good full-colour visual response for the
`1verage human observer, some consideration must be given as to what constitutes an
`1ppropriate choice of wavelengths for the red, green and blue laser illuminators.
`>Uitable choices of laser wavelengths for colour reproduction have been discussed by
`fobel & Ward (1989). Their results agree substantially with work done by Thornton
`1971) on colour rendition with fluorescent lights. The blue wavelength should be less
`han 460 nm in order to get good reproduction of blues and purples. More importantly,
`he green wavelength should be greater than 520 nm, and preferably greater than 530
`1m, in order to reproduce yellows and oranges well. Our choice of lasers (633, 532 and
`l42 nm) satisfies these requirements. It should be noted that if we were to utilize an
`1rgon ion laser (which is commonly found in commercial confocal fluorescence
`nicroscopes) for the blue (488 nm) and/or the green (514 nm) wavelengths, we would
`;et a restricted colour rendition. It is also important that the colours match those of the
`,hosphors in the video monitor. In particular, the green laser wavelength should be
`:lose to 540 nm. Again, our choice of laser wavelengths satisfies this requirement.
`Finally, it is important to emphasize that all of the discussion in this section refers to
`he human colour vision or photometric response. It was our usual practice in this
`tudy, in order to obtain good subject'ive colour balance, to alter the gains of the red,
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`110
`
`C. J. Cogswell, D. I(. Hamilton and C. J. R. Sheppard
`
`green and blue photodetectors so that the image of a mirror object was perceived to be
`white on the colour monitor.
`
`EXPERIMENTAL RESULTS: THREE-COLOUR IMAGING OF BIOLOGICAL
`SAMPLES
`The three-colour confocal reflection microscope can provide information with
`respect to the optical properties of a wide variety of samples which can be generally
`described as possessing one or more of the following four features: (i) specimens having
`highly reflective, but otherwise opaque surfaces, (ii) subjects that are largely
`transparent and which have internal refractive-index (phase) boundaries, (iii) objects
`which have internal components which demonstrate variations in absorption and
`reflection properties with wavelength (i.e. amplitude objects), and (iv) samples labelled
`with subwavelength scattering particles such as immunogold. Objects having highly
`reflective surfaces are frequently encountered in industrial inspection and indeed the
`three-colour Lasertec confocal microscope mentioned previously was designed for
`looking at spectral properties of surface structures and coatings. Most biological
`specimens, on the other hand, generally do not have such highly reflective, but
`otherwise opaque surfaces, but instead are usually transparent to some degree, and
`possess some or all of the features listed in the remaining three categories. As little work
`has previously been done to develop colour confocal laser-scanning instruments for
`examining these types of samples, we have chosen to confine our experimental studies to
`imaging the overall optical and, more specifically, the spectral characteristics peculiar to
`biological specimens.
`
`Refractive-index boundaries within transparent objects
`Biological specimens for light microscopy often have a high degree of transparency
`and this often causes great difficulty when attempts are made to obtain accurate images
`of comple~ internal structures. With respect to the optical properties of these
`structures, they frequently occur in the form of tiny variations in refractive index,
`which are often even beyond the detection capabilities of phase or differential
`interference systems. Biological stains have historically been used to convert some of
`these internal phase boundaries into visible colour variations in the resulting images. It
`is important to note, however, that most conventional bright-field microscopy of
`stained biological specimens is performed using a transmission optical configuration.
`This is primarily because of the difficulty in subduing unwan.ted flare from the many
`refractive-index boundary layers encountered in samples (including coverglass,
`mounting medium, cell walls and membranes, etc.) when reflection conventiona!
`bright-field optical systems are utilized. Confocal bright-field microscopes, on the
`other hand, do not work well in the transmission mode since lateral and axial shifts,
`which are often produced during transmission of the beam through a complex
`specimen, can cause the focused image spot to miss the detector pin-hole. These shif11
`cancel out, however, if the confocal system is utilized in a reflection mode, and
`furthermore, subduing unwanted flare is no longer a problem due to the inherent
`optical-sectioning property of confocal configurations in general.
`As mentioned previously, for confocal reflection, objectives with the highest possible
`NA and the least aberration are preferable. In addition to correcting for chromatic
`aberration, the objective lens should be tested and further corrected to the character·
`istics of the specimen being considered, including sample thickness, refractive index of
`immersion and mounting media, ,and coverglass thickness. The on-axis, specimen(cid:173)
`scanning design of our system facilitates adding correction lenses to the optical path tc
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`Colour conj ocal reflection microscopy
`
`111
`
`ninimize the effects of aberrations caused by these sample attributes. For a more
`letailed discussion see Cogswell & Sheppard (1990a).
`In order to test the capability of our three-laser confocal microscope for producing
`:ood quality colour images of a stained, largely transparent, biological subject, we chose
`o look at a typical botanical preparation. Figure 4 shows three reflected bright-field
`mages of a cross-section from a woody plant stem, Chodanthus puberulus, fixed,
`:mbedded and subsequently stained with alcian blue and safranin. Figure 4(a) was
`nade using the HeNe (red) laser, Fig. 4(b) with the frequency-doubled YAG (green)
`aser and Fig. 4(c) with the HeCd (blue) laser. All three images were produced at the
`;ame time and displayed in the red, green and blue channels of a colour video monitor to
`
`Fig. 4. Confocal reflected bright-field images of a cross-section from a plant stem, Chodanthus puberulus,
`stained with alcian blue and safranin. The three images, simultaneously produced with (a) the HeNe (red)
`laser, (b) NdYAG (green) laser, and (c) HeCd (blue) laser, were combined to form a full-colour image to
`detect differentially stained tissue (see Fig. 6b). Scale bar= 20 µm.
`
`3SHAPE 1020 3Shape v Align IPR2021-01383
`
`

`

`112
`
`C. J. Cogswell, D. K. Hamilton and C. J. R. Sheppard
`
`produce a full-colour representation (see Fig. 6b). This technique shows clearly the
`variations in reflectivity versus wavelength of the differentially stained cell walls. It is
`also worth noting that, although the three images were made using the larger set of
`detector pin-holes (Fig. 3a-c), the resulting colour image resolution is still reasonably
`good when utilizing a 512 x 512 pixel format.
`One additional item deserving mention is that for this and the following colour
`confocal examples we have not included a selection of corresponding images using a
`Type 1 (conventional) reflection bright-field microscope and, therefore, the obvious
`comparisons between the two techniques cannot be made. This is a direct result of the
`fact that it is extremely difficult to produce reflection bright-field images of such
`samples in a conventional microscope. In all of our examples, refractive-index
`boundary layers in the specimens produced such strong reflection signals (flare) when
`imaged in the Type I microscope, that they completely masked the weaker reflection
`components of the biological material.
`
`Objects having variations in absorption and reflection properties with wavelength
`Many other biological specimens contain natural pigments which show variations in
`reflection and absorption depending on the wavelength of the illuminating source.
`Good subjects for our colour-microscope studies are the epidermal guard cells
`surrounding a leaf stoma in Rhoeo spathacea (a purple-leaved monocot) because they
`contain pigments which reflect, to various degrees, all three laser wavelengths. The
`preparations utilized were pieces of fresh leaf peel, mounted in water on a microscope
`slide with a coverglass placed over the top, its edges being sealed with rubber cement to
`prevent drying out of the tissue. Figure 6(c) is an image from a single plane of focus and
`shows the green chlorophyll-containing chloroplasts in the pair of guard cells which
`form an oval surrounding the central pore (stoma). Adjacent to the guard cells are
`accessory cells which contain pigment granules that have precipitated out of solution ir.
`the cell vacuole, a phenomenon we observed to be quite common throughout the
`preparation amongst the stomata! accessory cells. These pigment granules <;an be seen
`in the lower left portion of the image and are easily identified by their bright magenta
`colour (i.e. they strongly reflect red and blue wavelengths). This image has been
`produced using only a 256 x 256 pixel format and the same large detector pin-holesasin
`Fig. 3(a-c), so the overall resolution is rather poor. However,.- we can still get an
`appreciation for how well the three-laser system works to distinguish meaningful
`information about the object from random shot noise. That is to say, not only does the
`presence of tiny magenta granules in the image indicate that the red and blue lasers were
`in good registration, but it indicates that these tiny reflecting objects are really part of
`the specimen and not random shot noise on any one beam, which would most often
`appear as a single primary-colour pixel corresponding to one of the laser wavelengths.
`One common problem in reflection confocal microscopy, when laser illumination is
`utilized, is that images may exhibit coherent speckle. This, we have found, can be
`reduced somewhat by averaging over several planes of focus or, alternatively, by
`averaging over the three separate images produced by the three lasers. The resulting
`monochromatic image produced by the latter technique would tend to remove the
`speckle pattern characteristic to any one laser wavelength.
`Another example of Rhoeo guard cells has been included in Fig. 6(d) to illustrate the
`capability of the three-colour microscope to reconstruct accurately the three(cid:173)
`dimensional morphology of the specimen. Here, five successive confocal bright-field
`images from 3-µm focus intervals were combined using the autofocus method to form
`the left half of a stereo pai