FR2826736A1 - Confocal optical scanner for biological microscope, has optical system to focus light from object to form beam reflected by rotary mirror for scanning - Google Patents

Abstract

The confocal optical scanner for a microscope has an optical system to focus an illuminating beam (FX) from a light source to illuminate a pint on an object (1107). The system focuses the light from the object to a point (FEO) in a first image plane.. A rotary mirror (1104) reflects the light beam to scan the object. A spatial filter (1203) in the first image plane filters the light to obtain a beam to be detected.

Description

<Desc / Clms Page number 1>

Confocal optical scanning device Technical area
The invention relates to a confocal optical scanning device used for example in a fast confocal microscope, more particularly a confocal fluorescence microscope intended to operate in real time.

Prior art.

 The principle of the confocal scanning microscope is to illuminate the sample with a light beam focused on a point, and to detect only the light coming back from this point. The confocal microscope therefore comprises an illumination beam possibly filtered by a microscopic hole, and the wave coming from the illuminated object itself crosses a microscopic hole selecting the light coming from the illuminated point. The wave coming from the object is then detected.

 To generate an image of the object, it is necessary to move the illuminated point in a plane. This scanning can be carried out in various ways: i) by moving the observed sample relative to the objective. It is the simplest method but it is excessively slow. ii) -by making the microscopic holes on a disc traversed in one direction by the lighting light and in the other direction by the light coming from the object. When the disc is rapidly rotating in the image plane, the microscopic holes move and the sample is scanned. This is the Nipkow disc technique. Since the disc has many holes, this scanning technique is very fast. Detection can be done directly on a camera, and it is also possible to observe the image directly with the naked eye. On the other hand, 95% of the available light power is lost, which limits the effectiveness of this technique when the samples are weakly fluorescent. Zeiss, however, offers such a microscope (CARV) for the observation of fluorescent samples. iii) -by using galvanometric mirrors to deflect the beam. The sample is then scanned point by point and a single microscopic hole is used. A photomultiplier tube allows the detection of photons having passed through this hole. This technique is the most used in confocal fluorescence microscopy because there is little loss of light energy. On the other hand, the imaging speed is limited by the saturation of the fluorescence and / or the frequency of oscillation of the galvanometric mirrors. The image cannot be observed directly and must necessarily be reconstructed by computer from the data acquired by the photomultiplier tube. The reconstitution of a horizontal plane of the sample requires knowing exactly the position of the galvanometric mirrors corresponding to the signal sampled at each instant on the photoreceptor. The good functioning of the system requires a very precise control of the galvanometric mirrors and a perfect synchronization between the movement of the galvanometric mirrors and the sampling of the signal coming from the photoreceptor.

<Desc / Clms Page number 2>

 To improve the imaging speed, a Nipkow disc microscope has been designed comprising a collecting disc made up of a network of microlenses, and integral with the disc carrying the microscopic holes. Each microlens is placed in front of a microscopic hole and focuses on this hole the light coming from the lighting beam. This technique is described for example in the American patent number 5,162, 941 of the university of Wayne, as well as in the American patent number 5,579, 157 of the company Yokogawa Electric Corporation. The corresponding microscope is marketed under the name "Ultraview" by Perkin-Elmer. This technique avoids the loss of light power resulting from the use of a Nipkow disc, while retaining the high imaging speed inherent in Nipkow disc systems.

 The technique used by Yokogawa Electric Corp. poses difficult problems of mechanical production and alignment and is costly to implement. The disc carrying the microscopic holes and the collecting disc must be aligned with each other with great precision. The assembly constituted by these two discs must be in rapid rotation around an axis which must be perfectly fixed, perfectly orthogonal to the plane of the discs, and perfectly orthogonal to the direction of propagation of the incident light. In practice, it is extremely difficult to keep the position of the axis fixed, and this results in image defects, typically small clear or dark lines superimposed on the image. American patent number 5,579,157 to Yokogawa describes a solution making it possible to attenuate this type of defects, however this solution is not very effective. Other alignment problems exist, described for example by US patent number 5,717,519 of Yokogawa Electric Corp.

 In the Yokogawa Electric Corp. system or in Zeiss CARV, part of the lighting wave is reflected on the plate carrying the microscopic holes and then goes towards the CCD sensor. This wave being very intense compared to the wave corresponding to the fluorescence, it is difficult to eliminate using dichroic filters, which leads to a significant attenuation of the beam.

 In the Yokogawa Electric Corp. system or in Zeiss CARV, the same microscopic holes are successively crossed by the lighting beam directed towards the sample, and by the beam coming from the sample. If the diameter of these holes is reduced, the loss of intensity of the beam which results therefrom is much greater than that which results from the simple reduction in the diameter of the microscopic hole in a confocal laser scanning microscope by galvanometric mirrors. For the intensity to remain reasonably high, the microcopic holes must have a diameter close to the width of the Airy spot. This causes a sharp decrease in resolution compared to the confocal laser scanning microscope with galvanometric mirrors, in which the resolution can be increased by reducing the diameter of the microscopic hole.

 In the Yokogawa Electric Corp. system or in Zeiss CARV, the image of a microscopic hole on the CCD camera, when this image has moved a length equal to half the width of Airy's spot, covers quite the same image before moving. Therefore the resolution is limited only by the fluorescent emission wavelength, and the resolution limit

<Desc / Clms Page number 3>

 is greater than it is in a confocal scanning microscope with galvanometric mirrors, in which the short excitation wavelength is used.

 Another solution for improving the imaging speed is described by American patent number 5,351,152 from the company Zeiss. The system described in this patent comprises a fixed network of microlenses which separates the laser beam into sub-beams which are each filtered by a microscopic hole situated in an image plane. The objective focuses in the object the beams coming from each of these microscopic holes. The beam re-emitted by the object is then redirected to a CCD sensor, each pixel of the CCD sensor being the image of a point on the object on which the beam from a corresponding microscopic hole is focused.

Scanning is performed by method (i) of moving the sample relative to the objective, although other methods are not expressly excluded. The point of the sample which is illuminated by one of the sub-beams and whose image is obtained on a corresponding point of the CCD sensor scans a reduced area of the sample. The image of the sample must be reconstructed by computer from a series of images obtained successively on the CCD sensor during the scanning operation of the sample. The scanning speed of this microscope is therefore limited by the reading speed of the CCD sensor, which must be read several times to obtain a single image.

 In general, 3D deconvolution of the image, which can significantly improve the resolution, is difficult using existing confocal microscopes. In devices with galvanometric mirror scanning, it is made difficult by sampling errors. In Nipkow disc devices, it is made difficult by the spatial inhomogeneities of the lighting.

Description of the invention
The subject of the invention is a confocal optical scanning device with the aid of which the previously indicated defects of confocal scanning microscopes by galvanometric mirrors or by Nipkow disc with or without collector are resolved. In particular, the invention allows direct observation of the image and acquisition of the image using a CCD sensor, and simplifies the problems of synchronization and alignment. This is compatible, in the present invention, with point-to-point scanning of the type used by scanning microscopes with galvanometric mirrors. The invention also makes it possible to observe the image in real time, at a speed equal to or greater than that attained by the Nipkow disc microscopes with collector, but without the difficulties of alignment and mechanical production associated with the use of these devices, at a very reduced cost, and with improved resolution by the absence of reflection on the non-transparent parts of the assembly carrying the microscopic holes.

 In a confocal scanning microscope with galvanometric mirrors of the usual type, it is possible to directly observe with an eyepiece the image plane or the beam reflected by the object reaches directly. However, the image thus observed is not a confocal image because it has not been filtered by the microscopic hole, and only the image acquired by sampling the signal passing through the microscopic hole is truly confocal. To optically obtain a confocal image, it is necessary to use a means equivalent to filtering through the microscopic hole. However, if the microscopic hole or equivalent is fixed, the image formed behind this hole does not directly constitute an image of the object: we are brought back to the

<Desc / Clms Page number 4>

 system described in US patent number 5,239,178 to Zeiss. If the microscopic hole or equivalent is mobile, we are brought back to the Nipkow disc system with or without collector, and to its known defects.

According to a version of the invention, these problems are solved by a confocal optical scanning device comprising: - a beam deflector A, which deflects the light beam as it reaches the observed sample, - a beam deflector B, which deflects the beam coming from the observed sample, characterized in that it comprises a beam deflector C, which deflects, by an angle proportional to the angle of deflection by said deflector B, the beam coming from the sample observed and having been deflected by said beam deflector B.

 Deflection by deflectors A and B is the technique commonly used in confocal laser scanning microscopes. The deflectors can typically be galvanometric mirrors, or acousto-optical deflectors as in the microscope marketed by the company Noran Instruments. The deflection on the deflector A makes it possible to move each illuminated point in the object. The deflection on the deflector B makes it possible to bring back to a fixed point of an image plane PB the wave coming from an illuminated point, so as to be able for example to filter this wave by a microscopic hole. Usually the wave is then detected and the confocal image is reconstructed by computer. The image formed in the plane PB is movable in the sense that the image of a fixed point of the object moves in this plane. However to detect this image with a camera or to observe it directly it would be necessary that it be fixed. The invention consists in using a deflector C which has the role of moving this image so as to compensate for its movement due to the deflector B, and to make it fixed relative to the sample. The deflectors not being placed in image planes, the deflection essentially results in changes in the direction of the beam at the exit of the deflectors, which themselves cause displacements of the image point in the image plane.

The control of the deflector C to compensate for the action of the deflector B is, in the general case, difficult. The simplest solution is to use movable mirrors or pairs of movable mirrors as deflectors satisfying the following conditions: - each movable mirror of said deflector B is merged with a corresponding movable mirror of said deflector A, or is integral with a mirror corresponding mobile of said deflector A, - each mobile mirror of said deflector C is merged with a corresponding mobile mirror of said deflector A, or is integral with a corresponding mobile mirror of said deflector A,
Indeed, in this case, and by means of the realization of a correctly dimensioned optical system, the movement of the mirrors of the deflector C can automatically compensate for that of the mirrors of the deflector B, so that after deflection by B and C the direction of the beam from a fixed point on the object is effectively constant. Compensation therefore does not require any particular precaution in controlling the deflectors. It requires at most an adjustment of the optical systems. The design of an appropriate optical path is facilitated if the mirrors are in afocal zones, or the wave coming from an observed point of the object is plane.

 The scanning device preferably comprises at least one microscopic detection hole or at least one microscopic reflecting point crossed by said beam coming from the observed sample

<Desc / Clms Page number 5>

 and having been deflected by deflector B, before this beam is deflected again by said deflector C. This allows, as in a conventional confocal microscope, to filter the wave in a plane or the image of an illuminated point is fixed, i.e. does not move when the illuminated point moves in the observed sample. The most conventional solution is the use of a microscopic hole. The role of a reflecting point is similar to that of a microscopic hole, the beam forming the image then being the reflected beam instead of the beam passing through the hole.

 The scanning device preferably comprises means for focusing at one or more points of an image plane the wave coming from the observed sample and having been deflected by said deflectors B and C. This focusing makes it possible to effectively form an image at from the wave whose deviations have been compensated for.

 The scanning device preferably comprises several detection microscopic holes or detection reflecting points, and means for dividing the light beam of illumination into several sub-beams, each focused at a point different from the sample observed. Each of said microscopic holes is preferably the image of a point on the object on which one of said sub-beams is focused. This allows multiple points to be scanned in parallel, which improves the speed of the system and makes it comparable to that of a Nipkow disk microscope.

 In the case where the scanning device comprises several microscopic holes or reflecting points, the deflector C must preferably compensate exactly for the action of the deflector B so that the direction of a beam coming from a fixed point of the object is , after deflection by said deflectors B and C, independent of the direction of this beam after passage of said deflector B. Indeed, in the contrary case, a point of the image plane would be likely to be illuminated successively by beams coming from different points of the object and passing through different microscopic holes.

 The acquisition of the confocal images on a camera involves, in the case where the excitation wavelength is notably lower than the wavelength of fluorescent emission, a loss of resolution compared to the acquisition on a sensor punctual. This is the case for example with the Nipkow confocal disc microscope, whose resolution is lower than that of the confocal laser scanning microscope. This is due to the fact that the image of a microscopic hole on the CCD camera, when this image has moved a length equal to half the width of the spot of Airy, covers quite largely the same image before moving. According to a version of the invention, it is possible to solve this problem by reducing the size of the microscopic holes so as to attenuate the phenomenon of overlap, however this solution causes a reduction in the useful light intensity. In order to avoid this loss of light intensity, and according to a version of the invention, each of said microscopic holes is associated with a converging lens placed before this hole on the path of the wave coming from the object observed, and focusing on this hole the beam coming from a corresponding point of the object. This lens makes it possible to reduce the width of the Airy spot and therefore to use a hole of smaller diameter without loss of intensity.

 When laser lighting is used, the scanning device preferably comprises means for varying the phase of each sub-beam. Indeed, in this case, the waves constituting each sub-beam are correlated with each other, which is liable to create interference in the object which can

<Desc / Clms Page number 6>

 disturb the image. The correlation between sub-beams must therefore preferably be broken, which can be done by independently modifying, over time, the phase of each sub-beam. The means for varying the phase of each sub-beam is for example a transparent pane carrying extra thicknesses and rotating around an axis.

 An important advantage of the present scanning device is that it makes it possible to obtain homogeneous lighting throughout the sample, or at least whose spatial periodicity is a submultiple of the sampling distance. In addition, it does not generate sampling errors. This makes it particularly well suited to the use of deconvolution algorithms, which allow an improvement in the quality and resolution of the image.

Quick description of the figures.

 FIG. 1 represents the preferred embodiment, using multi-point laser lighting and an array of microscopic holes. FIG. 2 represents a second embodiment, using incoherent multi-point lighting and an array of microscopic holes. FIG. 3 represents the image on a sensor of a network of microscopic holes used in the preferred embodiment, and the trajectory of a point of this image. FIGS. 4 to 6 illustrate the command applied to a beam attenuator as a function of the position and the speed of movement of the image on the camera of a microscopic hole. FIG. 7 represents a cube integrating two networks of microscopic holes and a dichroic mirror, simplifying the second embodiment. FIG. 8 represents the image on a sensor of a network of microscopic holes used in the second embodiment, and the trajectory of a point on this image, a single galvanometric mirror being used. FIG. 9 represents an array of lenses and an array of microscopic holes used for lighting the sample. FIG. 10 represents an array of lenses and an array of microscopic holes used to detect the wave coming from the sample. Figure 11 shows a side view of a set of mirrors also visible in Figure 1. Figure 12 shows a simplified version of the first embodiment. FIG. 13 represents a fifth embodiment, using incoherent lighting and a network of microscopic reflecting points. FIG. 14 represents an array of microscopic holes used in the first embodiment. FIG. 15 represents a network of reflecting points used in the fifth embodiment. Figure 16 shows a simplified version of the second embodiment. Figure 17 shows an improved version of the fifth embodiment. FIG. 18 represents a version with two galvanometric mirrors of the fifth embodiment. Figure 19 shows a third embodiment, using point-to-point scanning and a microscopic reflecting point. This third embodiment, represented by FIG. 19, is best suited to an immediate understanding of the operating principle. Figure 20 shows a fourth embodiment, using point-to-point scanning and a microscopic hole.

Preferred embodiment.

<Desc / Clms Page number 7>

 This embodiment is shown in FIG. 1. It constitutes the most efficient embodiment in terms of image quality and speed. In fact: - multipoint laser lighting is used, which allows, in particular on small areas, to use intense lighting without saturating the sample.

 - Networks of microscopic holes are used, which filter the beam perfectly well and limit the difficulties linked to possible parasitic reflections.

 The beam from a laser 100 passes through a beam attenuator 140 which can for example be electro-optical or acousto-optical. It then passes through a beam expander comprising for example the lenses 101 and 102. The beam then passes through a network of microlenses 103.

 There is shown in solid lines the sub-beam from one of these microlenses and in dotted lines the sub-beam from another microlens. The directions of the excitation beam FX and of the beam re-emitted by the fluorescent object FE have been indicated by arrows.

 The beam from the microlens array passes through a window 106 movable in rotation about an axis 107, then an optional lens 135 which may not be used, then an array of microscopic holes 105. The array of microscopic holes 105 can consist of by depositing a reflective layer on a transparent pane by a "litographic" type method, the holes then being interruptions of the reflective layer. In this case a neutral filter placed at the laser output can be used to attenuate the return effects of the laser light. The network of microscopic holes can also consist of a frosted metal plate in which the holes are drilled by means of a laser. This solution avoids the problems of return of the laser beam. The part of the beam which passes through a microlens of the grating 103 constitutes a sub-beam focused on a microscopic hole of the grating 105. The surface of the glass 106 is divided into a set of sub-surfaces, for example the sub-surfaces 120 and 121 Half of these subsurfaces carry an extra thickness generating a phase shift of 180 degrees in the laser beam passing through them. The sub-surfaces carrying extra thicknesses are distributed in a pseudo-random manner in all of the sub-surfaces. Each subsurface is approximately square. The width of the side of the square is equal to the distance between two adjacent microscopic holes in the array 105. The glass 106 is positioned so that each microscopic hole in the array 105 is placed below a separate subsurface. The rapid rotation of the window 106 makes it possible to generate pseudo-random phase shifts of all of the sub-beams, so that the spatial coherence of the beam is broken after crossing the glass 106. Indeed, the coherence between beams is likely to slightly disturb the image. The window 106 is however of little use if the microscopic holes in the network 105 are sufficiently spaced, in which case it can be eliminated. The beam having passed through the window 106 and the network of microscopic holes 105 then passes through the dichroic mirror 104 and the tube lens 108 and is then reflected by a mirror 109. It is then reflected by a galvanometric mirror 110 movable in rotation around a axis located in the plane of the figure, by a mirror 111, and by a galvanometric mirror 112 movable in rotation about an axis orthogonal to the plane of the figure. It crosses the microscope objective 113 and reaches the sample 114. The objective 113 is an objective forming an image of the observed sample at infinity. The array of microscopic holes 105 is in the image plane (plane where a stigmatic image of

<Desc / Clms Page number 8>

 the object observed). The array of microscopic holes 105 is not absolutely essential. It is useful especially to make the beam "cleaner" in the case where the microlens array 103 is of average quality. If the microlens array 103 is of very good quality, better image quality can be obtained by eliminating the array of microscopic holes 105.

 Preferably, the image focal plane of the objective 113 is in the object focal plane (in the objective direction towards lens 108) of the lens 108 and the array of holes 105 is in a focal plane of the lens 108. This makes it possible to maximize the useful opening of each sub-beam. In this case the lens 135 is not useful.

 The illuminated sample in return emits a non-coherent light beam. This beam passes through the objective 113, is reflected by the galvanometric mirror 112, the mirror 111, the galvanometric mirror 110, the mirror 109. It passes through the tube lens 108, is reflected by the dichroic mirror 104, then passes through the array of microlenses 131 and the array of microscopic holes 130. The pitch (distance between the centers of two adjacent holes) of the array of microscopic holes 130 is identical to the pitch of the array 105. The diameter of a microscopic hole of the array 130 may be equal to or smaller than the diameter of a microscopic hole in the array 105. Each microscopic hole in the array 130 corresponds to a microlens of the microlens array 131. The arrays 131 and 130 are positioned so that each hole of the array 130 is the image of a point of the object on which the lighting beam from a hole in the array 105 is focused. This means that each microscopic hole in the array 130 is in an image plane, by a microlent the corresponding eye, from the focal plane of the tube lens 108. The beam is then reflected by the mirror 115. It is then successively reflected by the mirrors 143, 144, 145, 146, 147 constituting the assembly 142 represented by a block in FIG. 1, which are shown in FIG. 11 in a view in the direction V indicated in FIG. 1. The assembly 142 serves to reverse the angle of the beam with respect to a plane containing the optical axis and situated in the plane of FIG. . The direction P indicated in FIG. 11 shows the direction of observation in which FIG. 1 is made. The beam then crosses the lens 116 identical to the tube lens 108, and whose object focal plane is on the network of microscopic holes 130. It is reflected by the second face of the galvanometric mirror 110, by the mirror 117, by the second face of the galvanometric mirror 112. It passes through the monochromator filter 141. It then passes through the lens 118 which forms in its image focal plane the image of the network 105, and therefore of the sample observed. It reaches the focal image plane of the lens 118. A CCD sensor 119 can be placed in this plane, however it is also possible to directly observe the image formed in this plane, using an eyepiece. The focal distance of the lens 116 must be exactly equal to the focal distance of the lens 108, and to allow precise adjustment the lens 116 can be replaced by a pair of lenses, the adjustment of the distance between lenses allowing an adjustment of the distance focal length of the whole.

 The networks of microscopic holes 130 and 131 are for example of the type shown diagrammatically in FIG. 14. They can be produced by litography on a glass pane, a metallic deposit as absorbent as possible constituting the absorbent layer.

 FIG. 9 shows the principle of the focusing of the laser beam on the array of microscopic holes 105, by the array of microlenses 103. The angle indicated in this figure is worth

<Desc / Clms Page number 9>

F 0 ; = oMV. La largeur d'un trou microscopique du réseau 105 est par exemple égale à la largeur de la F108 tache d'Airy soit L105 = 1, 22ou F ; 08 est la distance focale de la lentille 108, Test la ouv FU3

distance focale de l'objectif 113, ouv est l'ouverture numérique de l'objectif 113, Z,,, est la longueur d'onde du laser. La distance D entre deux trous microscopiques adjacents est de préférence au moins 10 fois la largeur de chaque trou. Les microlentilles du réseau 103 sont des microlentilles sphériques limitées par des carrés et adjacentes les unes aux autres. La largeur de chaque microlentille (coté du carré la limitant ou distance entre les centres de deux microlentilles adjacentes) est égale à la distance D entre deux trous microscopiques adjacents. La distance focale F de chaque microlentille du réseau de microlentilles 103

D F D F108 1 et sa largeur D sont en outre liés par la relation = a = ouv 113, soit F103 = 2F103 F108 2 Fl13 ouv

Par exemple on peut avoir : F =2mm 113 ouv = 1, 25 F, 08 = 200 mm D = 2 mm Alas = 488 nm (laser argon) F = 80 mm L105 == 47/an Si la largeur des trous microscopiques du réseau 105 est égale au diamètre de la tache d'Airy,

soit 1, 22-as ---, il n'y a pas de perte d'énergie mais le faisceau n'est pas gaussien, ce qui se traduit par OMV/ une diminution de la résolution. Un faisceau approximativement gaussien peut être obtenu en diminuant le 1 7 diamètre du trou microscopique, par exemple à 1, 22-----. Toutefois ceci entraîne une diminution 20uv FI 13

de l'intensité utile. La meilleure solution pour obtenir un faisceau gaussien sans perdre inutilement de l'intensité utile est de ne pas utiliser le réseau de trous microscopiques, dont le rôle est surtout de pallier aux imperfections du réseau de microlentilles. Dans ce cas le faisceau est parfaitement gaussien, ce qui assure une résolution maximale. Cette solution nécessite toutefois des microlentilles de bonne qualité.

F 0; = oMV. The width of a microscopic hole of the network 105 is for example equal to the width of the F108 Airy spot, ie L105 = 1, 22 or F; 08 is the focal length of lens 108, Test the aperture FU3

focal length of objective 113, open is the numerical aperture of objective 113, Z ,,, is the wavelength of the laser. The distance D between two adjacent microscopic holes is preferably at least 10 times the width of each hole. The microlenses of the array 103 are spherical microlenses bounded by squares and adjacent to each other. The width of each microlens (side of the limiting square or distance between the centers of two adjacent microlenses) is equal to the distance D between two adjacent microscopic holes. The focal length F of each microlens of the microlens array 103

DFD F108 1 and its width D are also linked by the relation = a = opening 113, i.e. F103 = 2F103 F108 2 Fl13 opening

For example we can have: F = 2mm 113 open = 1.25 F, 08 = 200 mm D = 2 mm Alas = 488 nm (argon laser) F = 80 mm L105 == 47 / year If the width of the microscopic holes of the network 105 is equal to the diameter of the Airy spot,

ie 1, 22-as ---, there is no loss of energy but the beam is not Gaussian, which results in OMV / a reduction in resolution. An approximately Gaussian beam can be obtained by reducing the diameter of the microscopic hole, for example to 1.22 -----. However, this results in a decrease 20uv FI 13

useful intensity. The best solution to obtain a Gaussian beam without unnecessarily losing useful intensity is not to use the network of microscopic holes, whose role is above all to compensate for the imperfections of the microlens network. In this case the beam is perfectly Gaussian, which ensures maximum resolution. However, this solution requires good quality microlenses.

. n n 9, a = OuV--. On a représenté le plan focal 132 de la lentille 108. En l'abscence du réseau de FI08

micro lentilles 131, le réseau de trous microscopiques 130 devrait être plaçé dans ce plan, et l'angle de FIG. 10 shows the operating principle of the microlenses of the array 131 and of the microscopic holes of the array 130. The angle indicated in this figure is worth, as in the figure

. nn 9, a = OuV--. The focal plane 132 of the lens 108 has been shown. In the absence of the network of FI08

micro lenses 131, the array of microscopic holes 130 should be placed in this plane, and the angle of

<Desc / Clms Page number 10>

divergence du faisceau après passage des trous microscopiques, dans la mesure ou ces trous sont suffisamment larges, serait a. Le réseau de lentilles 131 a pour utilité d'augmenter cet angle de divergence, ce qui équivaut à diminuer le diamètre des tache d'Airy formées sur chaque trou du réseau de trous microscopiques 130. Cect permet de diminuer le diamètre des trous microscopiques du réseau 130 sans perte d'intensité lumineuse, ce qui permet ensuite une réception sur un capteur CCD sans que la résolution soit diminuée par la superposition des taches d'Airy obtenues pour des points proches. L'angle de divergence du faisceau après passage des trous microscopiques devient, compte tenu des microlentilles du réseau 131, ss = ma avec typiquement m = 1, 5 à m = 4. Les distances d2 entre le réseau de microlentilles 131 et le réseau de trous microscopiques 130, dl entre le réseau de microlentilles 131 et le plan focal 132 de la lentille 108, et la distance focale F131 des microlentilles du réseau 131, sont liées par l'équation des lentilles minces (qui peut éventuellement être modifiée pour prendre en compte l'épaisseur des lentilles) :

1 1 1 d2 dl F131

et on a par ailleurs, compte tenu du coefficient m recherché :

dol - lem d2

De plus pour des raisons d'encombrement on a : D = 2da avec c > 1. Typiquement on peut prendre c = 2.

divergence of the beam after passage of the microscopic holes, insofar as these holes are sufficiently wide, would be a. The lens array 131 has the utility of increasing this angle of divergence, which is equivalent to reducing the diameter of the Airy spots formed on each hole of the array of microscopic holes 130. Cect makes it possible to decrease the diameter of the microscopic holes of the array 130 without loss of light intensity, which then allows reception on a CCD sensor without the resolution being reduced by the superposition of the Airy spots obtained for close points. The beam divergence angle after passage of the microscopic holes becomes, taking into account the microlenses of the array 131, ss = ma with typically m = 1.5 to m = 4. The distances d2 between the array of microlenses 131 and the array of microscopic holes 130, dl between the microlens array 131 and the focal plane 132 of the lens 108, and the focal distance F131 of the microlenses of the array 131, are linked by the equation of thin lenses (which can optionally be modified to take into account the thickness of the lenses):

1 1 1 d2 dl F131

and we also have, taking into account the coefficient m sought:

dol - lem d2

In addition, for reasons of space, we have: D = 2da with c> 1. Typically we can take c = 2.

La largeur des trous microscopiques sur le réseau 130 peut être égale au diamètre de la tache d'Airy sur ces trous soit :

LBO = 1, 22 Â flua F108 m Ouv FI 13 130-1, 22-----OMV-F

ou Âfluo est la longueru d'onde correspondant au maximum de l'émission fluorescente. F Taking into account also the expression of the angle <x = OM ---, we finally obtain: F108 dl = D F108 F, jDJ F 1 d 2c JOMV 2c FI 13 Ouv F131 = 1 dl 1 d2 = -dl m

The width of the microscopic holes on the network 130 can be equal to the diameter of the Airy spot on these holes, ie:

LBO = 1, 22 Â flua F108 m Ouv FI 13 130-1, 22 ----- OMV-F

or Âfluo is the wavelength corresponding to the maximum of the fluorescent emission.

Using the design example used above for networks 103 and 105, we obtain, for c = 2 and m = 4, and for luo = 518 nm (fluorescein):

<Desc / Clms Page number 11>

=2 mm 113 ouv = 1,25 F, = 200 mm D = 2 mm

F, 03 = 80 mm dl = 40 mm Fl31 = 13,33 mm d2 = 10 mm L130 = 12 ym
On peut améliorer la résolution en utilisant des trous microscopiques dont le diamètre est la moitié du diamètre de la tache d'Airy, ou encore inférieur, soit : L130=6 m
Pour une résolution maximale, avec des trous microscopiques dont le diamètre est suffisamment inférieur à la tache d'Airy, et pour permettre dans des conditions optimales une déconvolution ultérieure de

l'image confocale le pas-n d'échantillonnage sur la capteur CCD (distance entre les centres de deux pixels adjacents) doit vérifier :

1 F118 À lasÀ flua 119--,-----,---.-- 40Mn3 +0 À s + Afluo plig = 4ouv. FI 13 Â las + Â fluo soit F118 = 4ouvi9Fn3---,-lA fluo

Par exemple avec Pl 19 = 12, um et toujours dans le même exemple de dimensionnement on obtient : FI 18 = 477 mm Pour une résolution maximale mais sans déconvolution ultérieure de l'image confocale on peut se contenter

4 Â las + Â fluo. de : FUS = ouvPU9FU3 SOIt Fus = 168 mm 22 /Mo

Avec des trous microscopiques dont le diamètre est d'une tache d'airy on peut se contenter de :

4 +MO 4 =-/-OMVPii9i3---.--Fus = ouvPu9FU3 1 À 2 Alasllfluo

Il est donc être utile de pouvoir modifier la distance focale de la lentille 118, ou de la remplacer par un système avec zoom ou par un système de grossissement variable avec des éléments optiques interchangeables.

= 2 mm 113 open = 1.25 F, = 200 mm D = 2 mm

F, 03 = 80 mm dl = 40 mm Fl31 = 13.33 mm d2 = 10 mm L130 = 12 ym
The resolution can be improved by using microscopic holes the diameter of which is half the diameter of the Airy spot, or even less, that is: L130 = 6 m
For maximum resolution, with microscopic holes whose diameter is sufficiently smaller than the Airy spot, and to allow under optimal conditions a subsequent deconvolution of

the confocal image the sampling step-n on the CCD sensor (distance between the centers of two adjacent pixels) must verify:

1 F118 À lasÀ flua 119 -, -----, ---.-- 40Mn3 +0 À s + Afluo plig = 4ouv. FI 13 Â las + Â fluo or F118 = 4ouvi9Fn3 ---, - lA fluo

For example with Pl 19 = 12, um and always in the same dimensioning example we obtain: FI 18 = 477 mm For a maximum resolution but without subsequent deconvolution of the confocal image we can be satisfied

4 Â las + Â fluo. of: FUS = openPU9FU3 EITHER Fus = 168 mm 22 / MB

With microscopic holes the diameter of which is an airy spot, we can be satisfied with:

4 + MO 4 = - / - OMVPii9i3 ---.-- Fus = ouvPu9FU3 1 TO 2 Alasllfluo

It is therefore useful to be able to modify the focal distance of the lens 118, or to replace it by a system with zoom or by a system of variable magnification with interchangeable optical elements.

 The array of microlenses 131 is especially useful when the excitation wavelength is much less than the wavelength of the fluorescent emission and when the diameter of the microscopic holes is

<Desc / Clms Page number 12>

 of the order of an Airy spot. If the diameter of the microscopic holes is much less than an Airy spot or if the excitation wavelength is close to the wavelength of the fluorescent emission, the array of microlenses 131 can be eliminated without great consequences . In this case, the array of microscopic holes 130 is placed directly in the image focal plane of the lens 108.

focaux image distincts. Par conséquence, le plan focal objet de la lentille 108 (dans le sens de l'objectif vers la lentille 108) ne peut être confondu avec le plan focal image de l'objectif que pour un des objectifs utilisés. A practical difficulty is linked to the fact that objectives of different magnifications have plans

separate image focal lengths. Consequently, the object focal plane of the lens 108 (in the direction of the objective towards the lens 108) can only be confused with the image focal plane of the objective for one of the objectives used.

It is possible to overcome this difficulty by increasing the opening of the sub-beams coming from each microlens, which decreases the available light intensity. The lens 135 overcomes this difficulty without loss of light intensity. It must be dimensioned so that a plane wave in the object has for image a plane wave between the lens 135 and the microlens array 103.

 The galvanometric mirrors 110 and 112 are controlled so as to move the image of the network 130 on the sensor 119, as shown in Figure 3. Figure 3 shows the image of the network 130 on the sensor 119, for a position of reference for galvanometric mirrors. The line 301 superimposed on the drawing shows the path followed by the image of a point 300 of the network when the galvanometric mirrors are ordered. This trajectory is traveled alternately in both directions. When this trajectory is followed, the image of a hole in the network 130 scans a small part of the image plane, and the set of images of the holes in the network scans the whole of the image plane. A confocal image is therefore generated throughout the image plane. Contour 302 shows the limit of the useful area, in which a good quality confocal image is generated. Many variants of the trajectory traveled can be used, the essential constraint being that all of the image points of the holes in the network 130 scan the entire image plane.

balayage est dûe uniquement au miroir galvanométrique le plus rapide. La position des miroirs galvanométriques peut être caractérisée par la position du point image dans le plan image 119. Les figures 5

lias et 6 illustrent la valeur du rapport - en fonction de la position du point image d'un trou Vscan

microscopique sur l'échantillon. La figure 5 représente un ensemble de points images des trous microscopiques, dont le point 300 également représenté sur la figure 4, pour la position de référence des miroirs galvanométriques. Sur cette figure on a représenté en traits continus les arêtes de la surface

lias représentant la fonction - en fonction des coordonnées x, y du point 300 dans le plan image 119. The beam attenuator 140 must be controlled according to the position of the galvanometric mirrors and their speed. These parameters can be obtained in known manner by feedback from the galvanometers, or can be obtained without feedback from the galvanometer control system, with less precision. We note Ilas the intensity of the laser beam after crossing the attenuator, and Vscan the scanning speed, that is to say the speed of movement in the image plane 119 and along the x axis of the image point of a microscopic hole in network 130. This speed of

scanning is due only to the fastest galvanometric mirror. The position of the galvanometric mirrors can be characterized by the position of the image point in the image plane 119. Figures 5

lias and 6 illustrate the value of the ratio - depending on the position of the image point of a Vscan hole

microscopic on the sample. FIG. 5 represents a set of image points of the microscopic holes, of which the point 300 also represented in FIG. 4, for the reference position of the galvanometric mirrors. In this figure we have shown in solid lines the edges of the surface

lias representing the function - as a function of the x, y coordinates of point 300 in the image plane 119.

Sur la figure 6, on a représenté l'allure de cette fonction le long de la ligne 310 de la figure 5. Une valeur Vscan

In Figure 6, the shape of this function is shown along the line 310 in Figure 5. A value

<Desc / Clms Page number 13>

constante de sur l'ensemble de la trajectoire, c'est-à-dire une commande de l'atténuateur en fonction Vscan

de la seule vitesse de balayage, permettrait en principe de supprimer les variations d'éclairage dûes aux variations de la vitesse de balayage. Toutefois, la courbe de la figure 6 permet également d'atténuer les effets d'une variation incontrôlée de l'amplitude d'oscillation des miroirs, au prix toutefois d'une perte d'intensité lumineuse. Il est également possible de faire fonctionner l'atténuateur de faisceau en mode binaire. La commande de l'intensité se fait alors conformément à la figure 7. Seule la partie centrale de la trajectoire, sur laquelle la vitesse est à peu près constante, est utilisée. Enfin, il est également possible de ne pas utiliser d'atténuateur de faisceau. L'image reste de qualité acceptable mais peut être affectée par des variations locales de l'intensité, qui peuvent être compensées par un traitement numérique ultérieur. D'une manière

générale, plus le nombre de trous microscopiques est élevé et plus la surface de la trajectoire de l'image d'un i point du réseau 130 dans le plan image 119 est élevée, moins les variations d'intensité d'éclairage sont notables. L'utilisation d'un réseau suffisamment dense de trous microscopiques peut donc remplacer avantageusement l'usage d'un atténuateur de faisceau.

constant of over the entire trajectory, i.e. a control of the attenuator as a function Vscan

scanning speed alone would, in principle, eliminate variations in lighting due to changes in scanning speed. However, the curve of FIG. 6 also makes it possible to attenuate the effects of an uncontrolled variation of the amplitude of oscillation of the mirrors, at the cost, however, of a loss of light intensity. It is also possible to operate the beam attenuator in binary mode. The intensity is then controlled in accordance with FIG. 7. Only the central part of the trajectory, over which the speed is roughly constant, is used. Finally, it is also possible not to use a beam attenuator. The image remains of acceptable quality but may be affected by local variations in intensity, which can be compensated for by subsequent digital processing. In a way

In general, the higher the number of microscopic holes and the greater the area of the image trajectory of a point of the network 130 in the image plane 119, the less the variations in lighting intensity are significant. The use of a sufficiently dense network of microscopic holes can therefore advantageously replace the use of a beam attenuator.

 Each objective corresponds to an optimal dimensioning of the microscopic holes used, of the microlenses and of the lens 135. These elements can be designed so that they suit all of the objectives, however the properties of the system will be sub-optimal with certain goals. It is also possible to design a series of studied objectives to give the best results with a given dimensioning of the rest of the system. However, it is preferable to be able to exchange the microscopic holes used, the microlenses and the lens 135, so as to be able to optimize them according to the objectives used, the excitation and fluorescence wavelengths, the level of fluorescence, the desired acquisition speed, the desired resolution. To this end, the arrays of microlenses 103 and 131, the arrays of microscopic holes 105 and 130, the dichroic mirror 104, and the lens 135, can be made integral with each other and constitute an exchangeable block in one piece. This saves the user from having to adjust the alignment of the different elements: alignment problems are solved when the block is manufactured. The positioning of the block in the rest of the system must simply be done with sufficient angular precision (of the order of 1 miliradian).

 The present embodiment can be combined with all the imaging modes known in confocal microscopy. In particular, the object-holder can be provided with a vertical piezoelectric displacement system or with a stepping motor, so as to be able to generate three-dimensional images by modifying the focusing plane. Acousto-optical attenuator-based systems can be used to switch multiple lasers and excite different fluorophores, so as to generate information-rich images by superimposition. The system is also compatible with the use of the two-photon method, the number of microscopic holes then having to be adjusted so that sufficient intensity remains available at each focal point of the beam. Filter wheels (filter 141) can be used to modify the detected wavelengths. The dichroic mirror 104 can be replaced by a semi-transparent mirror so as to be able to change excitation and fluorescence wavelengths by switching a laser and changing the filter 141, without having to exchange the mirror

<Desc / Clms Page number 14>

 dichroic 104. The replacement of the dichroic mirror 104 by a semi-transparent mirror also makes it possible to use the apparatus for observing in reflection non-fluorescent diffractive samples.

 As indicated above, a number of the elements in Figure 1 are optional. FIG. 12 represents a simplified version of the system of FIG. 10, in which elements have been removed whose absence does not significantly degrade the quality of the image. The rotating disc (106,107) has been removed, which has little effect since the illuminated points in the object are sufficiently distant from each other. The lens 135 has been removed, which has little effect since the galvanometric mirrors and the object focal plane of the lens 108 are in the immediate vicinity of the image focal plane of the objective 113. The array of microscopic holes 105 has been removed, which is acceptable if the optical quality of the microlenses is sufficient. The microlens array 131 has been eliminated, which has little effect since the excitation and emission wavelengths are little different from each other. Due to the absence of the microlens array 131, the microscopic holes of the array 130 must be larger and can typically have the diameter used on Nipkow disk microscopes, or about 30 microns.

 It is possible, by means of relay lenses, to modify the device so that the entire scanning system is placed behind an intermediate image plane. Such a solution can be useful for making a scanning device adaptable to any type of microscope.

 It is of course possible to use only a single microscopic hole, in which case the speed characteristics are those of a confocal scanning microscope with galvanometric mirrors of the current type. However, the advantage of the system remains to be able to view the image directly and to be able to record it on a camera. Embodiments 3 and 4 illustrate in more detail embodiments with a single microscopic hole or equivalent.

 It may be useful to equip the device with a device for positioning the sample 114 in the direction of the optical axis. Indeed, this makes it possible to obtain three-dimensional images consisting of series of two-dimensional images each obtained at a different depth. The three-dimensional image obtained can then be the subject of a 3D deconvolution which improves its resolution. Prior to deconvolution, the PSF ("point spread function" in English, or 3D impulse response) must be measured, for example on a fluorescent microbead.

Second embodiment:
The second embodiment differs from the first in that non-coherent lighting is used.

This limits the imaging speed but allows the use of less expensive lighting, the wavelength of which can be easily adjusted. It is shown in figure 2.

 The lighting is provided for example by the incandescent arc 150 of a mercury vapor lamp. It passes through a collector 151, a field diaphragm 152, a monochromator filter 153 selecting the excitation wavelength of the fluorescence. It passes through the lens 154, the aperture diaphragm 155, the lens 156, and reaches the network of microscopic holes 105 directly. The assembly consisting of

<Desc / CRUD Page number 15>

 elements 150 to 156 constitute Köhler lighting and can be replaced by any other Köhler lighting system. The rest of the system is identical to the previous embodiment.

 A large part of the available light intensity is reflected by the network of microscopic holes 105, which limits the clarity of the image formed in the plane 119. So that the acquisition time and the clarity of the image remain within within reasonable limits, it is preferable to use an array 105 made up of microscopic holes very close to one another. While in the previous embodiment, the distance between two microscopic holes can typically be about 20 times the width of each hole, in the present embodiment it is preferable to limit this distance, which can for example be 2 at 4 times the width of the microscopic hole. This generates disturbances at low spatial frequency on the image, but the high spatial frequencies, which are the most information carriers, continue to benefit from the confocal effect.

 As in the previous embodiment, it is possible to eliminate the array of microlenses 131. In particular, to allow the exchange of the two arrays of microscopic holes without this exchange requiring complex adjustment, and to reduce the manufacturing cost, the array of lenses 131 can be eliminated and the networks of microscopic holes 105, 130 and the dichroic mirror 104 can be integrated into a single easily achievable element, represented in FIG. 7. The dichroic mirror 400 is integrated into a transparent cube 401. Windows 402,403 are mounted on the sides of the cube, of which they are secured by suitable parts, for example 404. Empty spaces 405,406 are left between the panes and the cube 401. The side of the panes 402,404 which looks towards the cube is covered with a thin metallic layer and a thin layer of photosensitive resin. The microscopic holes are made by exposure to the photosensitive resin using a white light projector. The appearance of the light beam from this projector has been symbolically represented in dotted lines. The projector is focused for example on the plane formed by the face of the window 402 which carries the thin layers. Each illuminated point of the window 402 will become a microscopic hole. Due to the configuration of the system, the points of the glass 403 which are lit are correctly positioned without any additional adjustment being necessary, and will become the microscopic holes corresponding to the holes in the glass 402. After exposure, an appropriate liquid is introduced in the empty spaces 405.406 so as to remove the resin in the exposed areas, then an acid is used to remove the metal at these same points. The microscopic holes were then obtained. A solvent can be used to remove the residual layer of resin, and the assembly can finally be cleaned. Finally, it may be useful to introduce an optical liquid or a transparent plastic into the empty space, preferably with the same index as the blades 402, 403. This avoids unnecessary reflections on the contact surfaces. If a material of the same index as the blades 402,403 is introduced into the empty spaces, then it is possible to use to make all the panes 402,403 frosted on their face which looks towards the cube. This allows better dispersion of the incident light by the reflecting parts. Monochromators 153 and 141 can also be integrated into this cube, which makes it possible to modify the imaging mode by exchanging a single component.

 Another solution simplifying the exchange of the networks of microscopic holes consists in using only one network of holes, which is crossed in one direction by the lighting beam FX and in the other direction

<Desc / Clms Page number 16>

 by the FE beam emitted by fluorescence from the observed sample. In this case, the system is modified as shown in FIG. 16: a single array of microscopic holes 105 is used, but the dichroic mirror 104 is placed between the lens 156 and the array of microscopic holes 105. The array of microlenses 131 is also deleted and the mechanical dimensioning is modified to take account of the replacement of the network 130 by the network 105. This configuration is particularly simplified, however this simplification results in the presence of a strong light intensity resulting from the partial reflection of the beam of FX lighting by the array of holes 105 ("stray light"), which must be eliminated by the dichroic mirror 104 and the filter 141.

 In this configuration, it is also possible to delete one of the galvanometric mirrors 110 or 112, provided that the axis of rotation of the remaining mirror is slightly modified. FIG. 8, produced according to the same conventions as FIG. 3, shows the trajectory 500 of the image of a microscopic hole 503 in the plane 119, as well as the zone 502 in which an image of good quality is obtained, the position images of the microscopic holes in the reference position being as indicated in the figure, limited by the contour 501. Because the microscopic holes are numerous and close to each other, a slightly oblique trajectory makes it possible to scan the whole of the plan without having to use a second galvanometric mirror.

 Similarly, it is possible to use the two galvanometric mirrors controlled in quadrature so that the image trajectory of a point 504 moves on a circle 505. If the microscopic holes are sufficiently numerous and close to each other the entire plan is scanned. The advantage of this solution is that the movement of the illuminated point is at constant speed, which guarantees good homogeneity of the lighting. This solution can be transformed into a two-dimensional scan if the diameter of the circle is gradually increased, which is equivalent to the image of a microscopic hole in the plane 119 moving in a spiral path.

 This second embodiment has the advantage of being less expensive than the first insofar as it does not require a laser beam. It also has the advantage of allowing easy modification of the excitation and fluorescence emission wavelengths throughout the visible and UV range. On the other hand, the weak lighting hardly allows real-time imaging.

 This second embodiment can be combined with the first in the same device, shutters that can be used to switch from one lighting mode to another.

Third embodiment:
This third embodiment is a simplified embodiment using single-point laser lighting, in which the microscopic hole is replaced by a microscopic reflecting point, and in which only one face of the galvanometric mirrors is used. Because of its simplicity, this embodiment lends itself well to an intuitive understanding of the operation.

 FIG. 18 shows a confocal laser scanning fluorescence microscope according to this third embodiment of the invention. In this figure, a beam passing through a point on the object has been indicated in fine lines and the directions of the various beams have been indicated by arrows.

<Desc / Clms Page number 17>

 An excitation beam FX from a laser 1100 passes through a beam expander or collimator formed by the lenses 1101, 1102 then is reflected by the dichroic mirror 1103 which reflects the wavelength of the laser 1100 and lets the length of the wave re-emitted by fluorescence. The beam FX is then reflected on the galvanometric mirror 1104 mobile in rotation about an axis located in the plane of the figure, then on the galvanometric mirror 1105 mobile in rotation around an axis orthogonal to the plane of the figure. It then crosses the polarizing beam splitter 1204 and then the microscope objective 1106. It is focused by the objective at a point in the fluorescent sample 1107. The FEI beam re-established by fluorescence from this point is collected by the objective 1106, passes through the polarizing beam splitter 1204, is reflected successively by the two galvanometric mirrors 1105 and 1104, passes through the dichroic mirror 1103, and is focused by the lens 1108 on a reflecting point 1203 situated on the rear face of a blade quarter wave 1202. The part of the beam which passes near the reflecting point then reaches an absorbent cavity. The reflecting point can for example be produced by optical litography.

 The part of the FEI beam which is reflected by point 1203 will be denoted FE2. This beam FE2 crosses the quarter wave plate 1202, the lens 1108, the dichroic mirror 1103, is reflected by the galvanometric mirrors 1104,1105, is reflected by the polarizing beam splitter 1204, then is focused by the lens 1205 in the plan of a CCD sensor 1206 fixed on the camera 1207.

 The polarization of the laser 1100 is chosen so that the wave coming from this laser crosses the beam splitter 1204. The quarter wave plate 1202 has its neutral axis oriented at 45 degrees from the axis passing from the polarizer 1201 and is quarter d wave for the wave re-emitted by fluorescence. Its role is to rotate the direction of polarization by 90 degrees, so that only the wave reflected by the reflecting point is then reflected by the polarizing beam splitter 1204, to the exclusion of the wave resulting from parasitic reflections on lens 1108.

 Under these conditions, the confocal image of the object is formed directly on the CCD sensor 1206 when the object is scanned by means of galvanometric mirrors. An imperfect control of the galvanometric mirrors results at worst in dark areas on the image, but in no case by a displacement of the points of the object or any geometric imprecision. The object must be scanned during the integration time of the sensor. The images can then be transferred from the CCD sensor to a sampler and a computer.

 The sensor can also be replaced by an eyepiece, possibly a binocular assembly, allowing direct observation of the image formed in the plane or on the diagram the CCD 1206.

In this case, the scanning must be done quickly enough not to be noticeable by the eye.

 This embodiment can be adapted to a multibeam system of the type used in the first embodiment. Its main advantage is that it does not require any precise adjustment apart from the correspondence of the laser focal point with the reflecting point. Its major drawback is the loss of light intensity which results from the use of semi-transparent or polarizing mirrors to allow the use of a single face of the mirrors for all of the optical paths.

 The quarter wave blade 1202 can be replaced by a transparent pane and the polarizing beam splitter 1204 replaced by a non-polarizing beam splitter. The system remains perfectly

<Desc / Clms Page number 18>

 functional under these conditions, the noise due to parasitic reflections being simply a little higher. One can then also replace one of the galvanometric mirrors with an acousto-optical deflector, which makes it possible to speed up the imaging procedure.

 The present embodiment allows a particularly easy understanding of the operation of the device. As in a confocal laser scanning microscope of the most common type (for example LSM from Zeiss), the movement of the galvanometric mirrors results in a scanning of the object 1107 by the excitation beam FX. The FEI beam emitted by fluorescence then reaches, after reflection on the galvanometric mirrors, the reflecting point 1203. In a confocal scanning microscope of the usual type, this reflecting point would be replaced by a microscopic hole behind which the signal would be detected using of a photomultiplier. In this case, the FEI beam is reflected, giving the FE2 beam. The beam FE2 follows exactly the reverse path of the beam FE1 and in the absence of the beam splitter 1204 it would return exactly to the focal point of the beam FX in the sample 1107. By introducing the beam splitter 1204, this beam is derived from so that it arrives at a point on the CCD 1206. This point on the CCD 1206 is only reached when the corresponding point of the sample 1107 is illuminated by the beam FX. Each point of the CCD 1206 therefore corresponds to a single point of the sample 1107 and when the sample 1107 is scanned by means of the galvanometric mirrors an image of this sample is formed on the CCD 1206.

Fourth embodiment:
This embodiment is shown in FIG. 19. It is a system using single-point laser lighting, but as in the first embodiment, two faces of the galvanometric mirrors are used and a microscopic hole is used. These features avoid most stray reflections.

 The beam from the laser 1300 passes through the beam expander or collimator formed by the lenses 1301,1302 and is then reflected by the dichroic mirror 1303. It is then reflected by the galvanometric mirror 1304 which is mobile in rotation about an axis located in the plane of the figure, then by the fixed mirror 1305 and by the galvanometric mirror 1306 movable in rotation around an axis orthogonal to the plane of the figure. It then crosses the objective 1307 infinitely forming an image of the sample, and is focused at a point of the observed sample 1308. The light re-emitted by fluorescence from this point crosses the objective 1307 in the opposite direction, is reflected by the galvanometric mirror 1306, the fixed mirror 1305, and the galvanometric mirror 1304.

It passes through the dichroic mirror 1303, and is focused by the lens 1309 and the fixed mirror 1310 on the microscopic hole 1311. The light having passed through the microscopic hole 1310 is reflected by the mirror 1312, collimated by the lens 1313, reflected by the mirror 1314, then by the second face of the galvanometric mirror 1304. It is then reflected by the mirror 1315 then by the second face of the galvanometric mirror 1306. It is then focused by the lens 1316 on a point of the image plane in which the sensor is located CCD 1317 attached to camera 1318. As before, the CCD sensor can be replaced by an eyepiece.

 The operation of the assembly is the same as in the first embodiment.

<Desc / Clms Page number 19>

Fifth embodiment
FIG. 13 illustrates a fifth embodiment, which has the advantage of being inexpensive.

In fact, it uses non-coherent lighting, and it requires fewer adjustments or specific elements than the second embodiment. It uses a single galvanometric mirror and a single face of this mirror, and it uses a network of microscopic reflecting points.

fluorescent observé 611. Le faisceau réémis, par fluorescence, par l'objet fluorescent, sera noté FIE 1. IL traverse l'objectif 610, la lentille de tube 609, la lentille 608, le séparateur de faisceau polarisant 602. Le faisceau FEI est réfléchi par le miroir galvanométrique 603, traverse la lame quart d'onde 604 et la lentille 605, et parvient au réseau de points réfléchissants 606. La partie du faisceau qui est réfléchie par ce réseau de points réfléchissants sera notée FE2. Le faisceau FE2 traverse la lentille 605, la lame quart d'onde 604, est réfléchi par le miroir galvanométrique 603, est réfléchi par le séparateur de faisceau polarisant 602, traverse le miroir dichroïque 601 et la lentille 612, puis parvient au capteur CCD 613 situé dans le plan focal image de la lentille 612. L'image formée dans le plan ou se trouve le capteur 613 peut également être observée à l'aide d'un oculaire. The FX1 excitation beam is a non-coherent beam produced for example by a Xenon arc lamp equipped with a collector or an optical assembly allowing the generation of Köhler lighting. It is filtered by a narrow bandpass filter which selects the excitation wavelength of the fluorescence. It reaches the scanning device at 600. It is reflected by the dichroic mirror 601 then by the polarizing beam splitter 602. The beam FX1 is then reflected by the galvanometric mirror 603, crosses the quarter-wave plate 604 whose axis neutral is 45 degrees from the plane of the figure, passes through the lens 605 and reaches the network of reflecting points 606 placed in the image focal plane of the lens 605. The network of reflecting points 606 is of the type shown in FIG. 15. It consists for example of a set of reflecting points (for example 650) forming a square matrix, on a transparent glass pane treated anti-reflection. The part of the beam FX1 which is reflected by the reflecting points will be denoted FX2. This beam FX2 crosses again the lens 605, the quarter-wave plate 604, and is reflected by the galvanometric mirror 603. The beam FX2 crosses the polarizing beam splitter 602 then the lens 608 and the tube lens 609. It crosses the objective 610 and is focused on the object

fluorescent observed 611. The beam re-emitted, by fluorescence, by the fluorescent object, will be noted FIE 1. It crosses the objective 610, the tube lens 609, the lens 608, the polarizing beam splitter 602. The beam FEI is reflected by the galvanometric mirror 603, passes through the quarter-wave plate 604 and the lens 605, and reaches the network of reflecting points 606. The part of the beam which is reflected by this network of reflecting points will be denoted FE2. The beam FE2 passes through the lens 605, the quarter-wave plate 604, is reflected by the galvanometric mirror 603, is reflected by the polarizing beam splitter 602, passes through the dichroic mirror 601 and the lens 612, then reaches the CCD sensor 613 located in the focal image plane of the lens 612. The image formed in the plane where the sensor 613 is located can also be observed using an eyepiece.

 The focal plane 607 of the tube lens 609 is also a focal plane of the lens 608. The other focal plane of the tube lens 609 is preferably coincident with the image focal plane of the objective 610. Preferably the mirror galvanometric 603 is in a common focal plane of lenses 605, 608 and 612. The network of reflecting points 606 is in a focal plane of lens 605. The CCD sensor 613 is in a focal plane of lens 612. For example, lenses 609,608 , 605.612 can have a focal length of 200 mm, the 610 lens can be a Nikon immersion lens x 100 with 1.4 aperture. In this case, the network 606 can for example be formed of reflecting points the diameter of which is 30 microns, the distance between two adjacent points being for example 150 microns. The dichroic mirror 601 reflects the wavelength of the excitation beam and lets through the fluorescent emission wavelength. The blade

<Desc / Clms Page number 20>

 wave 604 is used to modify the polarization of the beams so that, for example, the beam FE2 is reflected by the polarizing splitter 602 while the beam Fie 1 passes through it.

 The galvanometric mirror 603 can for example be movable in rotation about an axis orthogonal to the plane of the figure. In this case, the network of reflecting points 606 must be oriented so that the trajectory of the image on the CCD sensor of a reflecting point is of the type described by the trajectory 500 of FIG. 8, which makes it possible to scan the entire object using a single galvanometric mirror.

 As in the first embodiment, it would have been possible to put the galvanometric mirror 603 in the image focal plane of the objective 610 or close to this plane. Relay lenses 609 and 610 were introduced to illustrate the possibility of such a construction, which lends itself particularly well to the production of a scanning device adaptable to any type of microscope. In this case, the scanning device constitutes an independent unit delimited by the dotted lines 614, and adaptable to any microscope having for example a camera output on which the unit 614 can be adapted.

 The network of reflecting points can be produced for example by litography.

 The position of the array 606 must be adjusted in rotation about the optical axis so that the movement of the mirror 603 moves the image on the camera of the reflecting points, as shown in FIG. 8. The position of the array 606 must be adjusted in translation in the direction of the optical axis so that the network 606 is combined with the CCD 613. These adjustments are easily achievable.

 The present embodiment has the advantage of being inexpensive and requiring few adjustments. It is well suited to a low cost version of the invention. The apparatus can be equipped with a vertical displacement system of the sample, in which case the quality of the image obtained can be improved by a three-dimensional deconvolution. Such a deconvolution makes it possible to improve the resolution and to remove most of the faults due to parasitic reflections in the system, and possibly to compensate for the faults linked to the short distance separating the points of the network 606. It requires the prior measurement of PSF ("Point Spread Function" in English) which can be performed by imaging a fluorescent microbead, according to a known technique.

 The fact that the lighting beam FX1 is itself reflected by the galvanometric mirror before reaching the network of microscopic points makes it possible to make a long trajectory travel at each illuminated point, without loss of light energy. Indeed the illuminated area is fixed in the object but mobile on the network of microscopic points.

 To avoid the use of a polarizing beam splitter and the resulting loss of light intensity, the device can be modified as shown in Figure 17. In Figure 17 the various elements of the device are the same as in the Figure 13, the elements 602,603 and 605 being however replaced respectively by the elements 602b, 603b and 605b. The polarizing mirror 602 is replaced by the simple mirror 602b on which the beams FX1 and FE2 are reflected. The galvanometric mirror 603b is about twice as long as the mirror 603, and the lens 605b is also larger than the lens 605.

The optical axis of the lens 605b is offset from the optical axis of the beam FX1 coming from the galvanometric mirror, so that the beam FX1 arrives at an angle on the network of reflecting points 606. The

<Desc / Clms Page number 21>

 beam FX2 reflected by the network 606 then reaches an area of the mirror 603b distinct from that which was illuminated by FX1. After reflection on 603b, the beam FX2 passes next to the mirror 602b. The FEI beam coming from the sample follows the reverse path so that after reflection on the network of microscopic holes the FE2 beam reaches the sensor.

trajectoire 505 de la figure 8. Toutefois, du fait que le faisceau d'excitation FXl est lui-même réfléchi par les i miroirs galvanométriques, une trajectoire circulaire de grand diamètre peut alors être parcourue sans perte d'intensité lumineuse. L'avantage de cette technique est que la vitesse de déplacement des points éclairés dans l'objet est à peu près constante, ce qui évite les suréclairages ponctuels qui sont liés aux phases d'accélération et de décélération d'un miroir galvanométrique unique. The device of FIG. 17 can be completed by a second galvanometric mirror, so as to be able to carry out a two-dimensional scanning as in the second embodiment. The resulting diagram is shown in Figure 18 where a galvanometric mirror 615 was added and the other elements moved accordingly. The axis of rotation of the mirror 615 is orthogonal to that of the mirror 603b. By controlling the two mirrors in quadrature, it becomes possible to follow a circular trajectory to the image on the sensor of a point in the network of reflecting points. This trajectory is similar to the

trajectory 505 of FIG. 8. However, because the excitation beam FXl is itself reflected by the i galvanometric mirrors, a circular trajectory of large diameter can then be traversed without loss of light intensity. The advantage of this technique is that the speed of movement of the illuminated points in the object is almost constant, which avoids the point overlights which are linked to the acceleration and deceleration phases of a single galvanometric mirror.

Industrial applications
The present optical scanning device, integrated into a confocal microscope, can be used for real-time imaging of biological objects.

Claims (7)

Claims (1/2) 1-Confocal optical scanning device, comprising: - a beam deflector A, which deflects the light beam as it reaches the observed sample, - a beam deflector B, which deflects the beam coming from the observed sample, characterized in that it comprises a beam deflector C, which deflects by an angle proportional to the angle of deflection by said deflector B, the beam coming from the observed sample and having been deflected by said beam deflector B. 2-device according to claim 1, characterized in that: - said beam deflectors are movable mirrors or pairs of movable mirrors, - each movable mirror of said deflector B is merged with a corresponding movable mirror of said deflector A, or is secured to a corresponding movable mirror of said deflector A, - each movable mirror of said deflector C is merged with a corresponding movable mirror of said deflector A, or is secured to a corresponding movable mirror of said deflector A, 3-device according to claim 2 , characterized by the fact that - said movable mirrors of said deflector A are each merged with a corresponding movable mirror of said deflector B, - said movable mirrors of deflector C are each made on the opposite face of a corresponding movable mirror of said deflector A,  4-device according to one of claims 1 to 3, characterized in that said beam deflectors are placed in afocal planes. 5-device according to one of claims 1 to 4, characterized in that it comprises several microscopic holes crossed by said beam coming from the observed sample and having been deflected by said deflector B, or several reflecting points reflecting said beam coming from the sample observed and having been deflected by said deflector B. 6-device according to one of claims 1 to 5, characterized in that it comprises means for focusing in one or more points of an image plane the wave coming from corresponding points of the observed sample and having been deflected by said deflectors B and C. <Desc / Clms Page number 23>  Claims (2/2) 7-device according to one of claims 5 or 6, characterized in that it comprises a means for dividing the lighting light beam into several sub-beams each focused at a different point in the sample observed, and by the fact that each of said microscopic holes or reflecting points is the image of a point of the object on which one of said sub-beams is focused. 8-device according to one of claims 1 to 7, characterized in that said deflector C deflects the beam from a given point of the object so that the direction of this beam, after deflection by deflectors B and C , or independent of the position of said deflectors B and C. 9-device according to one of claims 1 to 8 and according to claim 7, characterized in that it comprises means for varying the phase of each of said sub-beam. 10-device according to claim 9, characterized in that said means for varying the phase of each of said sub-beams is a transparent pane carrying extra thicknesses and rotating about an axis.