Methods for localization and spectral imaging of particles and related systems

Abstract

The invention relates to a method for spatial localization and spectral imaging of particles (202) in a sample (200). The method comprises the steps of: A) illuminating the sample through an annular light cone (124) from a light beam partially reflected by the annular mirror (13) and focused by a peripheral region (142) of a dark field microscope objective (14), B) recording a dark field microscopic image collected by a central region (144) of the objective, C) positioning a point of interest in the image, D) extracting coordinates of the point of interest, and D) determining whether the point of interest is a point of interest. The method comprises the steps of: A) extracting a sample from a laser beam to form a list of points, E) moving the sample, scanning one of the extracted points, F) irradiating the points on the sample by the laser beam transmitted through the ring mirror and focused by the central region of the objective lens of the dark field microscope, G) collecting the emitted Raman spectrum through the central region of the objective lens. The invention therefore relates to an associated system.

Description

Technical Field

[0001] The present invention generally relates to the field of particle analysis by spectroscopic imaging.

[0002] More specifically, the present invention relates to methods for positioning and spectral imaging of transparent microparticles, and related systems.

[0003] The present invention has particularly advantageous applications in the tracking and rapid spectral imaging of transparent microparticles dispersed within macroscopically sized samples. Background Art

[0004] The detection and characterization of the chemical properties of micron-sized solid particles (often referred to as microparticles) is a subject of interest in many industries and / or applications. For example, the study of microplastics, including their detection, quantification, and identification, is essential in fields such as environmental science and the food industry. Within the framework of such studies, macroscopic samples of a few square millimeters or even square centimeters are removed from a medium to provide a representative overview of the medium. The sample can be, for example, in liquid or gaseous form and contain suspended solid particles or in the form of a solid particle powder.

[0005] To meet this analytical need, numerous measurement systems have been developed. Spectral imaging systems have been proposed, combining a so-called full-field microscopy modality for spatial localization of microparticles with a spectroscopic analysis modality, most commonly recording the spectrum of each microparticle by point analysis using a laser source, which allows the chemical composition of the analyzed microparticle to be tracked.

[0006] However, when using high-magnification microscope objectives, the image field of view of such spectroscopic imaging systems is still limited to a rectangular area of ​​a few millimeters by a few millimeters, or even smaller to spatially resolve the micronized particles of interest. Therefore, scanning the entire surface of a sample to locate microparticles requires creating a mosaic of the sample, which is then moved using a stage, with each tile or piece of the mosaic having the size of the imaging system's image field of view, whose dimensions are determined, among other things, by the microscope objective and / or image sensor.

[0007] According to the method, which is often described as static, a set of tiles representing a mosaic of the sample is acquired by full-field microscopy before a second point analysis is performed by laser spectroscopy. Thus, the sample is first fully imaged before spectral analysis.

[0008] Macroscopic samples typically require the acquisition and processing of mosaics containing millions of tiles. This image size necessitates compression of the mosaic's image quality to store it for extraction of the coordinates of the individual particles. This extraction is followed by point analysis, where spectra are acquired at the extracted coordinates. However, this image compression reduces spatial resolution.

[0009] A dynamic approach has been proposed to remedy the deterioration in spatial resolution. This dynamic approach involves extracting particle coordinates block by block and acquiring spectra in real time for each block before processing to the next. This dynamic approach offers numerous advantages, including reduced image size, as only the coordinates of interest are retained, rather than the entire image block. However, this approach increases the acquisition time for each image block when switching from the imaging modality to the spectral analysis modality.

[0010] Known solutions for observing single particles are described, for example, in the paper Rapid ultrasensitive monitoring the single-particle surface-enhanced Raman scattering (SERS) using a dark-field microspectroscopy assisted system [J]. Chin. Chem. Lett., 2020, 31(2): 473-475 by Shaochuang Liu, Yilun Ying, and Yitao Long, while a microspectroscopy measurement method is described in WO 2018 / 138098 A1. Summary of the Invention

[0011] In order to overcome the above shortcomings of the prior art, the present invention proposes to adopt a dynamic method to perform the positioning and spectral analysis of particles without the need for switching of motorized mechanical parts, thereby reducing the acquisition time.

[0012] More specifically, according to the present invention, a method for spatially locating and spectrally analyzing particles in an image using a microscope system is proposed, the method comprising the following steps:

[0013] A) illuminating the sample area with an annular light cone originating from a white light beam partially reflected by a toroidal mirror and focused by the peripheral region of a darkfield microscope objective,

[0014] B) recording a first darkfield microscope image over a field of view contained in the illuminated area of ​​the sample, the first darkfield image being collected by a central area of ​​a darkfield microscope objective,

[0015] C) locating a plurality of microparticles in the first dark field image,

[0016] D) extracting the coordinates of points corresponding to a plurality of microparticles within the recorded dark field image to form a point list,

[0017] E) Move the sample so that the optical axis of the darkfield microscope objective coincides with a point in the point list,

[0018] F) illuminating a measurement point on the sample with an excitation laser beam that is transmitted through a ring mirror and focused by the central area of ​​a dark-field microscope objective, said measurement point coinciding with a point in the point list,

[0019] G) Collection of the Raman spectrum emitted from the measurement point, which is collected by the central area of ​​the dark field microscope objective.

[0020] Thus, thanks to the present invention, in the context of dynamic methods, switching from one microscopy modality to another does not require mechanical switching. Within the context of the dynamic methods defined in the introduction, this property is advantageous for samples of large size relative to the microscope's field of view. Such dynamic methods were introduced, in particular, to avoid the use of spatial resolution compression (which is necessary to process large amounts of data given the size of the sample being studied). However, in order to acquire a wide-field image and then a point spectrum within the image, motorized mechanical components need to be switched within each tile, resulting in an incompressible idle time of approximately 5 seconds per tile. When scaling to millimeter or even centimeter samples, this idle time represents hours. Thanks to the present invention, the advantages of dynamic methods—precise positioning at the sample scale, high spatial resolution, and measurements that are virtually unaffected by environmental changes—are retained while eliminating the undesirable mechanical switching times. The elimination of mechanical switching times thus removes a significant obstacle to the use of dynamic methods for samples, particularly large ones.

[0021] Furthermore, thanks to the invention, the mechanical reliability of the system is increased. Indeed, moving parts are a potential cause of misalignment and even of system obsolescence.

[0022] The proposed solution not only eliminates these idle times, thus significantly saving time and improving system reliability, but also advantageously combines the characteristics of two selected modalities: dark-field microscopy and Raman microspectroscopy, also known as Raman microspectroscopy. On the one hand, this results in a solution that requires no mechanical switching and no signal loss: within the light loss range, the entire Raman signal collected during epitaxial detection reaches the spectrometer. On the other hand, this facilitates the study of large samples containing microparticles with low light absorption.

[0023] Other non-limiting and advantageous features of the process according to the invention are the following (alone or in all technically possible combinations):

[0024] - steps A) and F) are performed simultaneously, the area on the sample being illuminated by the peripheral area of ​​the dark field microscope objective, while the measuring points within the area are illuminated by the central area of ​​the dark field microscope objective,

[0025] -The measurement point is located on the optical axis in the central area of ​​the dark field microscope objective,

[0026] - reconstructing a field of view larger than the image field of view by using a mosaic composed of a plurality of said dark field microscopy images, wherein said plurality of dark field microscopy images correspond to a plurality of image blocks of said mosaic.

[0027] - the positioning of each tile of the mosaic is predefined upstream of the process,

[0028] - steps E), F) and G) are repeated until all points in the list of points extracted on a given tile have been scanned, followed by a step of moving the sample in order to focus the field of view of the dark field microscope objective on another tile of the mosaic and restarting the process from step A) in order to record another dark field microscope image,

[0029] - for each tile of the mosaic, the point list includes a certain number of points,

[0030] - the positioning of at least one tile of the mosaic is established as the coordinates of a point in said list of points,

[0031] - After step E) of moving so that the optical axis of the dark field microscope objective (14) coincides with a point on the point list, steps A), B), C), D), F) and G) are performed to record a new dark field microscope image corresponding to a new tile (40) of the mosaic (4) and a Raman spectrum emitted from the measurement point, these steps being followed by another movement (54) of the sample (200) so that the optical axis of the dark field microscope objective (14) coincides with another point on the point list.

[0032] - Additional points extracted from said new dark field microscope image recorded after the movement in step E) are added to the list of points to be scanned.

[0033] The present invention also relates to a system for spatial positioning in image and spectral analysis of microparticles, comprising an optical microscope, a laser source, a white light source, an image sensor, a spectrometer, and a control unit, wherein the control unit includes an image processing unit, the optical microscope includes a sample holder mounted on a translation stage, and the sample holder is adapted to accommodate a sample. The system is characterized in that it includes an optical system, the optical system includes an annular reflector, a dark field microscope objective lens, the optical system is arranged between the laser source, the white light source, and the sample holder, and the dark field microscope objective lens has a central area and a peripheral area.

[0034] - a ring mirror (13) configured to reflect a portion of the white light beam (120) emitted by the white light source toward the peripheral region (142) of the dark field microscope objective (14), and to transmit the excitation laser beam (160) emitted by the laser source (16) toward the central region (144) of the dark field microscope objective (14),

[0035] - the dark field microscope objective (14) is adapted to transmit the portion of the white light beam to the sample holder (20) via its peripheral region (142) and to transmit the excitation laser beam (160) to the sample holder (20) via its central region (144),

[0036] - the darkfield microscope objective (14) is adapted to focus a portion of the light beam to illuminate a first area of ​​the sample (200) on the sample holder (20) through an annular light cone (124), and the darkfield microscope objective (14) is further adapted to focus the excitation laser beam (160) to a measurement point contained in the first area,

[0037] - the central area (144) of the dark field microscope objective (14) is adapted to collect a dark field microscope image over the image field comprised by the first illumination area, the dark field microscope image being recorded by the image sensor (15),

[0038] - the image processing unit (60) is adapted to locate a potential plurality of microparticles (202) and extract coordinates of points associated with the plurality of microparticles (202) from the dark field microscope image to create a point list,

[0039] - the translation stage (11) is configured to move the sample holder (so that the measurement point coincides with a point in the point list, or to set the image field of the dark field microscope objective (14) to another area that is at least partially different from the first area;

[0040] - the central area (144) of the dark field microscope objective (14) is also adapted to collect a Raman spectrum generated from the measurement point, said Raman spectrum being recorded by a Raman spectrometer (19). BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Furthermore, various other features of the present invention will become apparent from the accompanying description made with reference to the accompanying drawings which illustrate non-limiting embodiments of the invention, in which:

[0042] Figure 1 Schematically depicts an optical microscope system adapted to implement a method for spatial localization and spectroscopic analysis of particles in a sample,

[0043] Figure 2 is a block diagram of a first embodiment of the method of the present invention,

[0044] Figure 3 An example mosaic is shown, comprising a plurality of tiles, each tile corresponding to a Figure 2 The method records the image,

[0045] Figure 4is a block diagram of a second embodiment of the method of the present invention,

[0046] Figure 5 Shown with Figure 4 Related methods, in particular, show an example of recording multiple consecutive image tiles to form a mosaic according to the second embodiment,

[0047] Figure 6 Shows Figure 3 or Figure 5 The image of the sample, inside the image, is Figure 3 or Figure 5 The mosaic includes some particles that locate the positions of the particles.

[0048] It should be noted that in these figures, structural and / or functional elements that are common to different alternatives may have the same reference numerals.

[0049] Various other modifications of the invention are possible within the scope of the appended claims. DETAILED DESCRIPTION

[0050] Figure 1 A system 1 for spatial localization and spectral analysis of particles in an image is shown, based on an optical microscope, an image sensor 19, an image processing system 60, and a Raman spectrometer 19. The system is referred to herein as an "optical microscope system" and is based on light-matter interaction to image and / or analyze a sample 200 having micrometer-sized structures. These structures are in particular in the form of particles 202, i.e., solid particles having a size between a few tenths of a micrometer and a few hundred micrometers.

[0051] Here, the microparticles considered are microparticles 202 of plastic material, i.e., polymeric microparticles 202, for example composed of long carbon chains and derived from fossil fuels. However, the term "microparticles" also refers, in a non-limiting manner, to mineral particles of micrometer size, or also to particles of pharmaceutical powders or solid particles suspended in air.

[0052] These particles 202 are contained in a sample 200 having a size ranging from a few hundred micrometers to several tens of centimeters. The sample 200 has, for example, macroscopic dimensions and comprises particles 202 suspended in water or air, or also in the form of a powder. The sample 200 is, for example, taken from the environment for analysis and / or research in various fields, such as the food industry, the pharmaceutical industry, geology, and environmental science, to name a few. Here, the collected sample 200 is placed on a sample holder 20, in particular a glass microscope slide, and may be protected by the microscope slide.

[0053] Figure 1The system 1 for spatial localization in image and spectral analysis of microparticles, schematically shown in FIG, is therefore configured to study the sample 200, in particular by enabling the imaging of the microparticles 202 contained therein, thereby obtaining morphological information about these microparticles 202. For example, the area, perimeter, shape, etc. of these microparticles 202 can be observed using images recorded using the optical microscope system 1. This morphological study can also advantageously be accomplished by studying the chemical composition of each of these microparticles 202, in particular by recording spectra, such as Raman spectra or photoluminescence spectra, which characterize the chemical composition. However, the recording of these characteristic spectra typically uses point measurement methods.

[0054] For the study of large-scale samples 200 containing microparticles 202, it is particularly advantageous to determine the spatial location of these microparticles 202 in advance, particularly from microscope images, and to establish characteristic spectra at these localized points, most commonly through point measurements at micrometer-sized points. It is then possible to spatially resolve information about the chemical composition of the sample 200. The term "large-scale" as used herein refers to a sample size that is larger than the size of the microparticles 202, i.e., the sample 200 under consideration has macroscopic dimensions and has a side length or diameter of, for example, several millimeters.

[0055] This initial location of the point of interest saves time given the small size of the point at which measurements can be made relative to the overall size of the sample 200. This is particularly true if the sample 200 contains a low concentration of particles 202. Thus, recording spectra only at the point of interest rather than performing point measurements at every point on the sample 200 advantageously reduces acquisition time, processing time, and the storage space required to store the measured spectra.

[0056] Thus, the system 1 for spatial localization and spectroscopic analysis of particles in images comprises at least two different microscopy modalities, each based on a contrast mechanism, i.e., the implementation of light-matter interaction and / or different light-matter interaction mechanisms. Here, in the present disclosure, the dark-field microscopy modality is advantageously selected for imaging the sample 200, in conjunction with the Raman microspectroscopy modality, sometimes also referred to as Raman microspectroscopy.

[0057] The dark field microscopy modality is configured to illuminate the surface of a sample 200 and acquire an image of at least a portion of the illuminated surface. Dark field microscopy, or more simply dark field imaging, is based on the following contrast mechanism: the sample 200 and structures contained therein, such as particles 202, are illuminated in such a way that only light rays deflected by the sample 200 and structures present on the sample 200 can be collected by the optical objective.

[0058] Darkfield microscopy is an imaging modality particularly suitable for studying microparticles 202, particularly plastic microparticles 202. Darkfield microscopy is known to degrade contrast on transparent samples 200 (i.e., samples that absorb very little light). Certain plastic materials, particularly microplastics and / or mineral or pharmaceutical powder particles, are known to be transparent and exhibit relatively low contrast when illuminated using other imaging modalities, such as brightfield illumination.

[0059] System 1 for spatial localization in image and spectral analysis of microparticles includes a second microscopic modality corresponding to a Raman microspectroscopy modality. This Raman microspectroscopy modality detects the chemical composition of a sample 200 at a measurement point by recording a Raman spectrum emitted from the measurement point after light excitation focused on the measurement point. Raman spectra are obtained through Raman scattering, which corresponds to the inelastic scattering phenomenon of light due to energy exchange with the medium being detected. This energy change represents the molecular vibration mode of the medium being detected at the measurement point. Therefore, the emitted Raman spectrum includes Raman lines. This Raman spectrum characterizes a given chemical substance, and its measurement and analysis can provide information about the chemical composition of the medium of interest.

[0060] Therefore, the Raman microspectroscopy modality advantageously allows point-by-point mapping, i.e., scanning the chemical composition of a measurement point of the sample 200, more specifically, the chemical composition of the microparticles 202 contained within the sample 200. For example, the composition of the plastic microparticles 202 dispersed within the aqueous sample 200 can be observed.

[0061] The implementation of each of these microscopy modalities is based on different illumination devices and / or devices for collecting the signal generated by the light-matter interaction, whether in the form of images or spectra.

[0062] The present invention proposes a system 1 for spatial localization in microparticle imaging and spectroscopy, in which two microscopy modalities, namely a dark field microscopy modality and a Raman microspectroscopy modality, are implemented simultaneously or sequentially, without any mechanical switching between two different illumination and / or detection devices, in order to switch from one modality to the other. In fact, the mechanical switching from one microscopy modality to another that is usually proposed in optical microscopy systems is the cause of the slowness when using such systems. Although there are solutions to reduce the time required to switch between two microscopy modalities, there are still problems related to the high financial cost of these fast switching solutions, their potential misalignment and their limited lifespan, in particular due to the obsolescence of the motors that drive such mechanical switching.

[0063] Also described is a dynamic analysis method that benefits from the presently disclosed system 1 for spatial localization and spectral analysis of microparticles in images, and is particularly suitable for imaging and spectral analysis of large-scale (i.e., macroscopic) samples including microparticles 202. Indeed, in order to fully image sample 200, the number of mechanical switches required to switch from one microscopy modality to another in order to localize microparticles 202 prior to performing spectral analysis of microparticles 202 increases exponentially due to the size of sample 200. Thanks to the disclosed method, analysis of this type of sample 200 is accelerated while maintaining high spatial resolution and minimizing the storage space required to record the acquired data.

[0064] We will first describe Figure 1 , a system 1 for spatial localization and spectral analysis of particles in an image is shown, which is intended to analyze a sample 200 as described above. In one example, the sample 200 is a sample comprising a dispersion of particles 202 (e.g., mineral particles or microplastics). However, it can also be a sample comprising a solution or liquid (e.g., an aqueous solution) in which particles 202 of plastic material are dispersed. In any case, the sample 200 is placed on a sample holder 20. The optical microscope system includes, for example, a microscope stand on which the different elements described below are arranged and connected.

[0065] The main axis 10 of the microscope is perpendicular to the plane of the sample 200, which is considered to be flat.

[0066] The sample holder 20 is integrated with a translation stage 11 or positioning stage. The latter is adapted to move the sample holder 20 along at least two translation axes (x and y), which are orthogonal to each other and parallel to the main plane of the sample 200. The translation stage 11, for example a piezoelectric stage, is controlled by a controller ( Figure 1 The sample holder 20 is controlled by a motor (not shown), which is also motorized, allowing for automated movement of the sample holder 20. The displacement amplitude of the translation stage 11 ranges from tens of millimeters to several centimeters, with a displacement accuracy in the micrometer range. A control unit 6, such as a computer, guides and synchronizes the movement of the translation stage 11.

[0067] Dark-field microscopy mode

[0068] The system 1 for spatial localization of particles in image and spectral analysis comprises a light source for illuminating the sample 200 according to the dark field microscopy modality. For example, it is a light source emitting incoherent optical radiation, such as an incandescent lamp, a halogen incandescent lamp or a lamp comprising at least one light emitting diode (LED).

[0069] More precisely, in the first embodiment, the spectrum of the light radiation emitted by the incandescent lamp at least partially covers the visible spectrum. Such light radiation is then perceived as being close to white and is therefore referred to as a white light source 12.

[0070] Here, the light radiation emitted by the white light source 12 propagates in free space in the form of a white light beam 120 , which is a collimated light beam. It is assumed that the white light beam 120 propagates along an optical axis perpendicular to the main axis 10 of the microscope.

[0071] The optical microscope system also includes a toroidal mirror 13. The toroidal mirror 13 has a flat reflective surface, such as a polished metal deposit on a flat surface. For example, in this case, it is a silver deposit. The toroidal mirror 13 has an elliptical shape, the center of which is penetrated by an elliptical central opening. In other words, the toroidal mirror 13 has the shape of an elliptical ring. The opening is centered on the surface of the toroidal mirror 13. Therefore, the central portion of the toroidal mirror 13 does not reflect light. The toroidal mirror 13 is, for example, included in a block or cube commonly known as a darkfield cube.

[0072] The position of the ring mirror 13 remains fixed relative to the microscope stand.

[0073] The arrangement of the ring mirror 13 is centered on the main axis 10 of the optical microscope system on the one hand and on the white light beam 120 on the other hand. In addition, the reflection plane of the ring mirror 13 forms an angle of 45 degrees with the main axis 10 of the microscope system. The ring mirror 13 is then adapted to reflect part of the white light beam 122 towards the sample 200.

[0074] In fact, if the white light beam 120 is initially approximated as a solid cylinder, only the portion corresponding to the hollow cylinder is reflected toward the sample 200. This portion corresponds to a portion of the white light beam 122 incident on the reflection plane of the ring mirror 13.

[0075] The remaining part is transmitted through the elliptical central opening of the ring mirror 13, for example towards a beam blocker, Figure 1 Not shown in .

[0076] The portion of the white light beam 122 corresponding to the hollow cylinder propagates along the main axis 10 of the optical microscope system towards the sample 200 .

[0077] The optical microscope system further comprises a darkfield microscope objective 14 (as known in the art). This objective must have different optical zones, a peripheral zone 142 and a central zone 144. The optical axis of the darkfield microscope objective 14 is aligned with the main axis 10 of the microscope system.

[0078] In the embodiment described, the darkfield microscope objective 14 corresponds to an objective lens with a magnification of 50 times and a numerical aperture of 0.60. The field of view diameter of the darkfield microscope objective 14 is 0.44 mm. In addition, it is an infinity-corrected objective lens.

[0079] The peripheral region 142 of the darkfield microscope objective 14 corresponds to the outer region surrounding the central region 144. The central region 144 corresponds to a conventional microscope objective. Therefore, the central region 144 is centered on the optical axis of the darkfield microscope objective 14 and is surrounded by the peripheral region 142.

[0080] The portion of the white light beam 122 corresponding to the hollow cylinder passes through the peripheral region 142 of the dark-field microscope objective 14. The dimensions of the hollow cylinder and of the ring-shaped mirror 13 are selected as a function of the dimensions of the peripheral region 142. In particular, it is ensured that the entire portion of the white light beam 122 reflected by the ring-shaped mirror 13 is transmitted through the peripheral region 142 of the dark-field microscope objective 14.

[0081] This peripheral region 142 leads to an annular optical element, which focuses the hollow cylinder into an annular light cone 124. This optical element corresponds, for example, to an annular optical lens or a concave mirror. In other words, the sample 200 is illuminated in a darkfield microscopy configuration using an annular light cone 124, which corresponds to a hollow cone. Thus, the sample 200 is illuminated using light rays that are highly inclined relative to the main axis 10 of the optical microscope system, i.e., light rays that have a large inclination angle relative to the main axis 10 of the optical microscope system. The angle of inclination of the light rays is selected so that these light rays cannot be collected by the central region 144 of the darkfield microscope objective 14. Here, the inclination angle assumes a value greater than or equal to 37 degrees relative to the optical axis of the darkfield microscope objective 14.

[0082] A region of the sample 200 is thus illuminated. This region extends parallel to the plane of the sample 200, and the size, in particular the diameter, of this illuminated region is determined as a function of the properties of the darkfield microscope objective 14, such as its magnification, its numerical aperture.

[0083] The illumination area considered here has a circular size with a diameter ranging from a few micrometers to several hundred micrometers. For example, the area considered here is a circular illumination area with a diameter of at least 100 micrometers, or even a diameter of at least 140 micrometers, or even a diameter of at least 150 micrometers. The size of this circular illumination area, in this case the diameter, is greater than the size of the image to be recorded later.

[0084] According to the dark field microscope objective 14 used in the first embodiment, the illuminated area corresponds to a circular area extending in the range of several tens of micrometers or even several hundred micrometers. For example, the illuminated area corresponds to a circular area with a diameter of 440 micrometers.

[0085] Sample 200, and in particular the structures contained therein, here particles 202, more specifically plastic particles 202 and / or mineral particles 202, scatter a portion of the oblique light from annular light cone 124. According to the known contrast mechanism of darkfield microscopy, only the light deflected by the structures of sample 200 is collected by central region 144 of darkfield microscope objective 14. Thus, only structures that at least partially deflect the incident light generate contrast and contribute to the darkfield microscopy image. Central region 144 of darkfield microscope objective 14 acts as a conventional microscope objective, characterized by magnification and numerical aperture, as described above.

[0086] The light rays deflected by the particles 202 are thus collected by the central region 144 of the dark field microscope objective 14. From a spectral point of view, these deflected light rays have the same spectrum as the white light beam 120, since the contrast mechanism is based on the phenomenon of elastic scattering.

[0087] The deviated light rays exit the darkfield microscope objective 14 in a collimated state because an infinity-corrected objective is used in the first embodiment. These deviated light rays then propagate in free space along the main axis 10 of the optical microscope system and, in particular, through the opening formed in the ring mirror 13.

[0088] The deviated light rays are imaged by an image sensor 15 , which has an optical axis aligned with the main axis 10 of the optical microscope system and whose extension is therefore aligned with the optical axis of the darkfield microscope objective 14 .

[0089] The image sensor 15 comprises an imaging optical system and a light-sensitive matrix sensor 150, which is generally referred to as a camera. The optical imaging system is adapted to form an image of the sample 200 on the light-sensitive matrix sensor 150. The light-sensitive matrix sensor 150 comprises a matrix of light-sensitive pixels, here rectangular, and is capable of converting light information into accumulated electrical charges. For example, it is a CCD sensor or a CMOS sensor.

[0090] Image sensor 15 is adapted to record a darkfield microscopy image corresponding to at least a portion of the darkfield illumination area by imaging the deflected light rays onto a light-sensitive matrix sensor. Image sensor 15 has a field of view that depends, among other things, on the magnification of the optical imaging system of image sensor 15 and the size of the rectangular matrix of pixels. This field of view of image sensor 15, along with the field of view of central region 144 of darkfield microscope objective 14, limits the area of ​​sample 200 that can be imaged in darkfield microscopy. This imaging area is referred to as the image field of view.

[0091] The image field of view is contained within the illumination area and therefore its size is smaller than that of the illumination area.

[0092] The dark field microscopy image reproduces the region of the sample 200 described above, the contrast within the dark field microscopy image representing structures capable of deflecting incident light, ie particles 202, such as particles of plastic material or mineral grains.

[0093] Since the light sensitive matrix sensor 150 is rectangular, the dark field microscope image is generally rectangular.

[0094] Each dark field microscopy image represents, for example, an area of ​​140 micrometers by 105 micrometers on the sample 200 .

[0095] Image processing unit 60, such as Figure 1 As shown, the image processing unit 60 is included in, for example, a control unit 6, here a computer, and is configured to exchange information with the image sensor 15. The image processing unit 60 is also adapted to identify and extract coordinates of points that may correspond to microparticles 202 from the dark-field microscope image. More specifically, microparticles 202 with a minimum diameter of 0.5 micrometers are spatially localized and morphological parameters are extracted therefrom. This minimum diameter depends, among other things, on the optical performance, i.e., the imaging performance, of the system. There is no upper limit to the diameter of the detectable microparticles 202.

[0096] The image processing unit 60 is further configured to interact with the user via an interface integrated into the software. The user may then select certain particles 202 of interest based on specific criteria, such as morphological parameters. For example, the user may choose to retain only particles 202 whose diameters are between two value limits.

[0097] The identification and extraction is performed using dedicated image processing algorithms, or by automatic learning or machine learning algorithms known in the art. Alternatively, the user of the optical microscope system manually or partially manually identifies and extracts the coordinates of the points of interest within each darkfield microscope image.

[0098] Raman microspectroscopy modality

[0099] Raman microspectroscopy is also shown in Figure 1 middle.

[0100] Advantageously, this Raman microspectroscopy modality can be switched from one modality to another without switching mechanical elements and can therefore be performed simultaneously with a dark-field microscopy modality.

[0101] The optical microscope system comprises a laser source 16. The laser source 16 emits an excitation laser beam 160, which is continuous here. In addition, the excitation laser beam 160 is collimated at the light outlet of the laser source 16.

[0102] In a non-limiting manner, the laser source 16 corresponds to a diode-pumped solid-state laser.

[0103] For the Raman microspectroscopy modality, the excitation laser beam 160 is monochromatic and has a wavelength selected, for example, in a spectral window between 380 nm and 1064 nm, for example, here the excitation laser beam 160 has a wavelength of 532 nm.

[0104] At the exit of the laser source 16, the excitation laser beam 160 propagates along the optical axis of the beam, where the optical axis of the laser beam is initially parallel to the main axis 10 of the optical microscope system. The first dichroic filter 17 (also known as an edge filter in the prior art) reflects the excitation laser beam 160 so that the optical axis of the laser beam is perpendicular to the main axis 10 of the optical microscope system.

[0105] This first dichroic filter 17 more precisely corresponds to a high-pass filter. This filter is adapted to reflect a spectral range below a cutoff wavelength and to transmit a different spectral range above the cutoff wavelength. The cutoff wavelength is selected as a function of the wavelength of the excitation laser beam 160, here selected to reflect the excitation laser beam 160. The first dichroic filter 17 has a cutoff wavelength set by the constructor, for example, at 532.85 nanometers.

[0106] Because it corresponds to an edge filter, also known as a laser filter, the transmission curve of the first dichroic filter 17 as a function of wavelength has a very small transition region between the transmission wavelength and the reflection wavelength. In other words, the transmission curve of the first dichroic filter has a slope that is considered steep at its cutoff wavelength. In this case, the transition between the transmission wavelength and the reflection wavelength occurs within a few nanometers near the cutoff wavelength. For example, the transition occurs over a width of 10 nanometers or less near the cutoff wavelength, more ideally, 5 nanometers or less, or even 3 nanometers or less near the cutoff wavelength.

[0107] The optical microscope system further comprises a second dichroic filter 18 .

[0108] The second dichroic filter 18 is adapted to reflect a spectral range greater than the cut-off wavelength and to transmit another different spectral range. It is therefore a low-pass dichroic filter.

[0109] Here, the cutoff wavelength is selected to be smaller than the wavelength of the excitation laser beam 160, that is, strictly smaller than 532 nanometers. More precisely, the cutoff wavelength of the second dichroic filter 18 is equal to 495 nanometers. Therefore, the excitation laser beam 160 is reflected by the second dichroic filter.

[0110] The second dichroic filter 18 is also selected to transmit at least part of the spectrum of the white light beam 120. Advantageously, according to one embodiment, the spectrum of the white light beam 120 is completely transmitted by the second dichroic filter 18.

[0111] The second dichroic filter 18 is positioned on the main axis 10 of the optical microscope system between the ring mirror 13 and the image sensor 15. It is also centered relative to the optical axis of the excitation laser beam 160. The surface of the second dichroic filter 18 is angled at 45 degrees to the optical axis of the laser beam to reflect the excitation laser beam 160 along the main axis 10 of the optical microscope system toward the annular opening formed inside the ring mirror 13.

[0112] Then, the excitation laser beam 160 passes through the hole of the ring mirror 13 and propagates toward the sample 200 along the main axis 10 of the optical microscope system.

[0113] A central area 144 of the dark field microscope objective 14 centered on the main axis 10 of the optical microscope system transmits the excitation laser beam 160 of the Raman microspectroscopy modality before focusing the excitation laser beam 160 onto a measurement point on the sample 200 .

[0114] The measuring point is centered on the optical axis of the darkfield microscope objective 14, which coincides with the main axis of the optical microscope system 10. This point coincides, for example, with the geometric center of the illuminated area in a darkfield microscope.

[0115] The extent of the measurement point in the xy plane is defined as a function of the properties of the dark-field microscope objective 14 , in particular its numerical aperture, and the wavelength of the excitation laser beam 160 .

[0116] Typically, the measurement point ranges between a few hundred nanometers and a few micrometers. Here, for example, given the numerical aperture of the darkfield microscope objective 14 and the wavelength of the excitation laser beam 160, the measurement point detects a point on the sample 200 with a diameter of approximately 564 nanometers.

[0117] Through the Raman effect, the sample 200 emits a Raman spectrum associated with inelastic scattered light. The focus here is Stokes radiation, corresponding to radiation with a wavelength greater than or equal to the wavelength of the excitation laser beam 160.

[0118] This light radiation corresponds to the Raman spectrum emitted from the measurement point focused by the excitation laser beam 160. This Raman spectrum includes in particular Raman lines specific to the chemical composition of the measurement point. In order to obtain this spectral information, the scattered light radiation needs to be analyzed.

[0119] Thus, the Raman spectrum in the form of scattered radiation is collected by the central region 144 of the darkfield microscope objective 14, which collimates the scattered radiation into a beam, referred to as the Raman beam 162. The Raman beam 162 is accompanied by a beam of elastically scattered radiation corresponding to Rayleigh scattering. This beam has the same wavelength as the excitation laser beam 160. This beam is hereinafter referred to as the Rayleigh beam.

[0120] The Raman beam 162 and the Rayleigh beam propagate along the main axis 10 of the optical microscope system towards the hole in the ring mirror 13.

[0121] The Raman beam 162, the Rayleigh beam and the excitation laser beam 160 are thus spatially combined and collinear with each other. However, the Raman beam 162 and the Rayleigh beam are counter-propagating relative to the excitation laser beam 160, ie, they propagate in opposite directions relative to each other.

[0122] Therefore, the Raman beam 162 and the excitation laser beam 160 share the same optical path between the sample 200 and the first dichroic filter 17 .

[0123] Therefore, the central opening of the ring mirror 13 allows the Raman beam 162 and the Rayleigh beam collected and collimated by the central region 144 of the dark field microscope objective 14 to pass through.

[0124] The Raman beam 162 and the Rayleigh beam continue to propagate to the second dichroic filter 18. Assuming that the Raman beam 162 corresponds to Stokes light radiation, whose wavelength is greater than or equal to the wavelength of the excitation laser beam 160 and is therefore strictly greater than the cut-off wavelength of the second dichroic filter 18, the Raman beam 162 is therefore reflected, thereby forming an angle of 90 degrees with respect to the main axis 10 of the optical microscope system. The same is true for the Rayleigh beam having the same wavelength as the excitation laser beam 160.

[0125] Therefore, the Raman beam 162 and the Rayleigh beam propagate upward along the optical axis of the excitation laser beam 160 to the first dichroic filter 17 .

[0126] The first dichroic filter 17 transmits the Raman beam 162 toward a Raman spectrometer 19 included in the optical microscope system and reflects the Rayleigh beam. In practice, the first dichroic filter 17 is configured to reflect wavelengths less than or equal to a cutoff wavelength, which is selected to be equal to the cutoff wavelength of the excitation laser beam 160, while transmitting wavelengths greater than the cutoff wavelength.

[0127] The Raman light beam 162 is thus spectrally analyzed by a spectrometer 19. The spectrometer 19 comprises, inter alia, a dispersive element (i.e. an optical element adapted to decompose the light spectrum into a plurality of monochromatic spectral components, such as a prism or a diffraction network) and a light-sensitive matrix sensor 190 (e.g. a CCD sensor or a CMOS sensor).

[0128] The excitation laser beam 160 has been focused on the measurement point through the central area 144 of the dark field microscope objective 14. The spectrometer 19 performs spectral analysis on the Raman beam 162, and the Raman spectrum from the measurement point can be tracked.

[0129] Thus, in embodiments of the presently disclosed dual-modality optical microscopy system, i.e., a system in which a darkfield microscopy modality for imaging is coupled with a Raman microspectroscopy modality, both modalities may be performed simultaneously without the need for mechanical switching from the illumination modality to the other.

[0130] According to a first alternative, in order to implement both modalities simultaneously, non-overlapping wavelength ranges are selected for the spectrum of the white light beam 120 and the wavelength of the excitation laser beam 160. In particular, in order to separate the Raman beam 162 from the Raman microscopy modality from the stray light from the dark field microscopy modality, the spectrum of the white light beam 120 is, for example, selected to be strictly smaller than the wavelength of the excitation laser beam 160. In particular, a low-pass filter placed between the white light source 12 and the ring mirror 13 eliminates unwanted wavelengths. Alternatively, monochromatic radiation having a wavelength strictly smaller than the wavelength of the excitation laser beam 160 is selected as white light. Here, in Figure 1 In the embodiment shown, the wavelengths covered by the laser source 16 and the wavelength range covered by the white light source 12 are considered to be disjoint. Furthermore, the wavelength range covered by the white light source 12 is considered to be strictly lower than the wavelength of the excitation laser beam 160.

[0131] Conversely, according to a second alternative, the two modes can also be implemented in an alternating manner (i.e., one after the other) by controlling the power supply of one or both light sources (i.e., laser light source 16 and white light source 12). Thus, one of the light sources can be selectively switched off and the other switched on rapidly, compared to conventional switching times known in the prior art. In this case, the spectral ranges covered by the two light sources can be freely selected.

[0132] Dynamic Methods

[0133] This optical microscope system is particularly advantageous for implementing the spatial localization and dynamic spectroscopic analysis methods for microparticles 202 described below. It should be noted that in this dynamic method, unlike the static method, in which the entire sample 200 is imaged before performing spectroscopic analysis, only a portion of the sample 200 is imaged, and spectroscopic analysis is performed on the imaged portion. These two steps are repeated until a comprehensive representation of the sample 200 is obtained.

[0134] A first embodiment of the method consists of Figure 2 The first box shown Figure 3 describe.

[0135] The method comprises, in its first embodiment, the following steps, which are detailed in the following description using examples:

[0136] - illuminating 30 an area on the sample 200 using a dark field microscopy modality,

[0137] - Record 31 dark field microscopy images,

[0138] - locating 32 more particles 202 within the image,

[0139] - extracting 33 the coordinates of the points corresponding to these particles 202 to form a list,

[0140] - moving 34 the sample 200 so that the central area 144 of the dark field microscope objective 14 corresponds to a point in the extracted coordinate list,

[0141] - irradiating 35 the sample 200 using the Raman microspectroscopy modality, the excitation laser beam 160 being focused on a measurement point that coincides with the point of the previous step,

[0142] The Raman spectrum emitted from the measurement point is collected 36 by the central area 144 of the dark field microscope objective 14 .

[0143] The movement step 34, the illumination step 35 and the collection step 36 may then be repeated, for example, until all points on the list have been scanned (i.e., have been browsed). A new movement step 37 then moves the sample 200 to restart the method with illumination 30 of a new dark field area that is at least partially different from the previous area.

[0144] The application of the first embodiment of the method is as follows Figure 3 shown.

[0145] In this first embodiment, a mosaic representing the sample 200 is reconstructed using multiple dark field microscope images. Figure 4 Due to the combination of several dark field microscope images covering at least partially non-intersecting areas on the sample 200, the stitching Figure 4 represents an area larger than the image field of view. Here, the images considered are dark field microscope images, covering completely disjoint areas. Ideally, the stitching Figure 4 represents the entire surface of the sample 200 .

[0146] Each of the plurality of dark field microscopy images is referred to herein as a tile 40, ie, a region of the sample 200 extending in the xy plane. Figure 4 The other tiles 40 included in the image are at least partially disjoint. Here, each tile 40 should be completely disjoint from the other tiles. Here, the size of each tile 40 is the same and is defined by the field of view of the image sensor 15, which is included in the illuminated area of ​​the sample 200.

[0147] Therefore, based on the characteristics of the image sensor 10 and the dark field microscope objective 14 , each tile 40 has a rectangular shape of 140 μm×10 5 μm.

[0148] exist Figure 3 In the formation of splicing Figure 4 The tiles 40 are arranged in a rectangular grid, wherein the tiles 40 are adjacent to each other.

[0149] In the first embodiment, splicing Figure 4 The position of each image segment 40 is predetermined upstream of the method execution relative to a coordinate system. This coordinate system is represented, for example, by a translation stage 11, which is integral with the sample holder 20 and, therefore, with the sample 200. Thus, the coordinate system established by the translation stage 11 also locates a point on the sample 200.

[0150] Therefore, each tile 40 is located by a pair of coordinates along the x-axis and the y-axis in the coordinate system. i ,y i ) locates the geometric center 41 of the tile 40 relative to the sample 200. The geometric center 41 of the tile 40 here represents the intersection of the diagonals of the tile 40 having a rectangular shape. Figure 3 For each tile 40, the geometric center 41 is represented by a crosshair, i.e. a cross and a circle.

[0151] As an alternative, each tile 40 can also be positioned by one of its corners.

[0152] In the first embodiment, a set of coordinate pairs is created for each image block 40 in advance, ie, before the implementation of the presently disclosed method for spatial localization and spectral imaging of the microparticles 202 . Each image block 40 occupies a predetermined spatial range on the sample 200 .

[0153] Here, for illustration purposes, Figure 3 Only the mosaic reconstructed on the sample is shown in Figure 4 part of.

[0154] Figure 3 The splicing shown Figure 4 The portion includes twenty tiles 40, each tile 40 having a size of 140 micrometers x 105 micrometers.

[0155] Typically, the sample 200 considered has a size of about 10 mm in diameter (e.g. 13 mm, 25 mm or 42 mm, a few centimeters at most), and is composed of a mosaic of at most 1000×1000 tiles 40. Figure 4 That is, one million tiles 40, or even more.

[0156] For example, for a sample 200 having a diameter of 27 mm, if the tiles 40 are rectangular and each measures 100 μm x 100 μm, then π*(D / 2)2 tiles need to be acquired, where D corresponds to the diameter of the sample 200. This means that approximately 58,000 tiles need to be acquired to cover the entire surface of the sample 200.

[0157] Small-sized patches 40 , corresponding to images acquired using high-magnification objectives, are preferred, in particular to increase spatial resolution in the plane of the sample 200 .

[0158] To record the splicing Figure 4 For each dark-field microscopy image of a tile 40, the translation stage 11 moves the specimen 200 so as to, for example, bring one of a set of coordinate pairs pre-established for each tile 40 (here representing the geometric center 41 of each tile 40) into coincidence with the optical axis of the dark-field microscopy objective 14. More generally, the problem is to bring the field of view contained in the region of the specimen 200 illuminated by the dark-field microscopy modality into coincidence with the predetermined tile 40.

[0159] Within the framework of a dynamic method for spatial localization and spectral imaging of particles 202, coordinate points 42 associated with potential particles 202 are located within each image segment 40 imaged using dark-field microscopy, and Raman spectra at these coordinate points 42 are recorded using Raman microspectroscopy before proceeding to the acquisition of the next image segment 40. Prior to implementing this process, the sequential order of image segments 40 to be imaged is predefined. For example, the surface of the sample 200 is imaged here using a line-by-line scanning mode, also known as a raster scan, or using a serpentine scan, or any other suitable scanning mode.

[0160] Therefore, according to the first embodiment in which the positions of the tiles 40 are predefined, Figure 4 For each of the n tiles 40 , the steps described below are performed, between each step the sample 200 being moved 37 from one tile to another.

[0161] The application of this method is now described in detail, for example, to obtain Figure 4 The upper left corner of the tile 40, such as Figure 3 shown.

[0162] First, after the translation stage 11 is positioned, the geometric center 41 defined for one of the image segments 40 corresponds to the optical axis of the dark field microscope objective 14 .

[0163] For each tile 40, an area of ​​the sample 200 is illuminated 30 using an annular light cone 124. It should be noted that this light cone is generated by reflecting a portion of the white light beam 120 by the annular mirror 13. The reflected portion, forming a hollow cylinder, is transmitted by the peripheral region 142 of the darkfield microscope objective 14 and focused onto the sample 200, thereby illuminating the area on the sample 200 using only light that is highly inclined relative to the optical axis of the darkfield microscope objective 14.

[0164] Image sensor 15 is then used to record a first darkfield microscopy image 31, which corresponds to one of image segments 40. As described above, light rays deflected by sample 200, and in particular, the particles 202 contained therein, are collected by central region 144 of darkfield microscopy objective 14. These deflected light rays form a first darkfield microscopy image, in which contrast is present only for structures that deflected the light, namely, particles 202. The first darkfield microscopy image is recorded over the image field included in the illuminated region of sample 200. This first darkfield microscopy image corresponds to first image segment 40, here located in the upper left corner of the mosaic.

[0165] This dark field microscopy modality optimizes the contrast on samples 200 (here microparticles 202 of plastic material) or mineral particles that have significant transparency. This makes the next step easier.

[0166] In practice, based on the first dark-field microscopy image, potential particles within the first image are then localized 32. This localization 32 is performed, for example, using an image processing algorithm. In particular, such an algorithm can use shape detection or contour detection. Alternatively, a machine learning algorithm can be used to identify potential particles 202. This machine learning algorithm, such as a shape recognition algorithm, is pre-trained using a library of images to recognize patterns of interest.

[0167] Once potential particles are identified in the first microscopic image, the process continues with extracting 32 the coordinates of points within the dark field image that may correspond to particles 202. This extraction is performed, for example, by the image processing unit 60 using image processing algorithms or machine learning algorithms.

[0168] Therefore, for the corresponding splicing Figure 4 A first dark field microscope image of one of the image blocks 40 is obtained to obtain a point list formed by coordinate points 42. These coordinate points 42 are defined in the coordinate system of the translation stage 11.

[0169] A coordinate point 42 is defined for each particle 202 within the image field of view, for example, the coordinate point is set at the geometric center of each particle 202. In other words, each particle 202 appearing on the dark field imaging block 40 is associated with a single coordinate point 42.

[0170] The geometric center of each particle 202 is located inside the particle 202 and is set as far away from the edge of the particle 202 as possible.

[0171] Alternatively, a plurality of coordinate points 42 can be extracted from the same particle 202 appearing in the first dark-field microscope image. This large number of points then takes the form of a matrix of points defined on each particle 202. The spacing between the coordinate points 42 extracted on the same particle 202 is defined as a function of the displacement accuracy of the translation stage 11 and the spatial resolution of the Raman microspectroscopy modality. Thus, a fine grid is obtained for each particle 202. This thin grid can then be used to measure a spectrum, here a Raman spectrum at each point of the matrix, thereby achieving a spectrogram to obtain the chemical distribution inside the particle 202. Alternatively, all spectra acquired within a given particle 202 can also be averaged to obtain an average spectrum representative of that particle 202.

[0172] Optionally, the particles 202 from which the coordinate points 42 were extracted may be pre-screened, as described above. For example, screening criteria may be established based on criteria associated with morphological parameters based on the first dark field microscopy image. For example, these morphological parameters may be size or shape criteria.

[0173] Here, Figure 3 A variation is shown in which a single coordinate point 42 is extracted for each particle 202 that appears in a different tile 40. Figure 4 Each coordinate point 42 on a different tile 40 is shown by a dotted cross.

[0174] Therefore, in Figure 3 In the first dark field microscopic image shown in the upper left corner, three coordinate points 42 are extracted: coordinate point 421, coordinate point 422, and coordinate point 423. These three points together constitute the point list of the image block.

[0175] The sample is moved 34 using the translation stage 11 so that the optical axis of the dark field microscope objective 14 coincides with it. In other words, this movement 34 aligns the central area 144 of the dark field microscope objective 14 with a point from the point list established in the previous step. For example, this is the coordinate point 421, as Figure 3 shown.

[0176] The measurement point on the sample 200 is then illuminated 33 using the Raman microspectroscopy modality. For this purpose, the excitation laser beam 160 transmitted through the ring mirror 13 is focused by the central region 144 of the dark field microscope objective 14 .

[0177] Since the measurement point is centered relative to the optical axis of the dark field microscope objective 14 , the measurement point coincides with the point from the coordinate point list. In other words, the excitation laser beam 160 is focused on the coordinate point 421 .

[0178] The white light source 12 is then switched off, for example via its power supply, which is connected to the control unit 6 .

[0179] However, maintaining both illumination modalities simultaneously is advantageous, particularly when moving stage 11. Indeed, the simultaneous illumination of the measurement point by annular light cone 124 and by the focused laser beam makes it possible, in particular, to observe both the illuminated area and the measurement point on sample 200 simultaneously using image sensor 15 in dark-field microscopy mode. A very small portion of excitation laser beam 160 passes through second dichroic filter 18. This very small portion, representing, for example, less than 0.001% of the optical power of excitation laser beam 160, is nevertheless sufficient to be observed on image sensor 15. Thus, the relative position of the measurement point generated by excitation laser beam 160 relative to the dark-field illuminated area on sample 200 can be observed. The measurement point then appears as a small bright spot in the image field.

[0180] Therefore, if disjoint wavelength ranges are chosen such that the spectra of the white light beam 120 and the excitation laser beam 160 are strictly below the wavelength of the excitation laser beam 160 , both illuminations can be maintained simultaneously without risk of interference.

[0181] A step is then performed to collect 36 the Raman spectrum emitted from the measurement point. This spectrum is collected by the central region 144 of the darkfield microscope objective 14 and propagates along the optical path described in detail above to the spectrometer 19 for analysis. As with the illumination of the sample 200 by the excitation laser beam 160, the collection of the Raman spectrum does not require mechanical switching, which makes it possible to prevent misalignment of the optical elements along this optical path.

[0182] Other coordinate points 422, 423, ... extracted from the same dark-field microscopy image are also detected to collect Raman spectra. To this end, the movement 34 of the sample 200 is controlled by the translation stage 11 so that the measurement point successively coincides with each of the other coordinate points 422, 423, ... in the point list of the considered image segment. Thus, within the framework of the illustrated example, the translation stage 11 moves the sample 200 to scan the coordinate point 422 and then the coordinate point 423. The scanning pattern within the image segment 40 is indicated by the arrows.

[0183] A Raman spectrum is collected at each point in the point list by moving 34 the stage 11 and then collecting 36 a spectrum.

[0184] These two steps of collecting 36 and moving 34 are recursively applied to all coordinate points 42 in the list of points extracted from the first dark-field microscope image, ie the list of points extracted from the first image block 40 .

[0185] After scanning all coordinate points 42 in the point list of the first image block 40, a new dark field microscope image is recorded as part of the process. For this purpose, the sample is moved 37 so that one of its geometrical centers, for which a pair of coordinates has been predefined, coincides with the optical axis of the dark field microscope objective.

[0186] Then reset the point list.

[0187] This new dark field microscopy image corresponds to the mosaic defined according to the selected scanning mode Figure 4 Here, it is, for example, the block to the right of the first block 40 described above, i.e. Figure 3 on the first line and the second colon.

[0188] For the next tile 40 and splicing Figure 4 Repeat the above series of steps for all blocks 40 included in the image until a spectral image representing the entire sample 200 is obtained. Figure 6 shown.

[0189] On this spectral image, visible structures (here corresponding to particles 202) are imaged based on the dark field microscopy modality, and colors (here represented by textures) are associated with each particle 202. These colors indicate the chemical properties, i.e., the composition, of the particle 202. Figure 6 In the illustrated example, the colors correspond, in a non-limiting manner, to particles 202 of a mineral material, more specifically chalk. Furthermore, these colors correspond to pseudo-colors arbitrarily selected based on the chemical composition of the particles. The composition of each particle 202 is determined by analyzing a point Raman spectrum collected using a Raman microspectroscopy modality, specifically based on the presence of Raman lines in the collected spectrum.

[0190] In practice, the Raman spectrum of each particle 202 is identified and assigned to a class by a dedicated algorithm. For example, and not by way of limitation, the chemical composition of the particle 202 can be identified by correlation between the spectrum acquired on the particle 202 and a pre-established basis of spectral data stored in memory. The particle 202 is then identified as corresponding to the known chemical species whose spectrum has the closest correlation.

[0191] A multivariate analysis can also be performed, for example by means of multivariate curve analysis (MCR). This multivariate analysis makes it possible to classify the particles 202 into different classes and then identify these classes manually or automatically.

[0192] According to another alternative, the identification of the chemical composition of the microparticles 202 is performed by machine learning. A model is then trained beforehand on a large amount of labeled data, thus allowing the identification of the spectra of the microparticles 202 to be characterized.

[0193] Thus, for each tile 40 obtained and corresponding to a dark field microscope image, spectral information is extracted relating to the imaged microparticles 202 in the tile 40. This spectral information then enables the chemical properties of these microparticles 202 to be tracked.

[0194] The dynamic acquisition method described in the first embodiment has the following advantages:

[0195] First, in the splicing Figure 4 During the block-by-block reconstruction, the morphological parameters of particles 202 are determined in real time. Therefore, the spatial resolution of dark-field microscopy can be fully utilized during the morphological parameter determination process. This effectively improves the accuracy of the determination. In fact, in the static mode mentioned in the introduction, the resolution of the dark-field microscopy images must be compressed during acquisition to limit the storage space required to store these dark-field images.

[0196] In addition, in splicing Figure 4 During the block-by-block reconstruction, the spectral analysis is also performed in real time. Therefore, the movements performed by the translation stage 11 for positioning each particle are small, in the order of a few microns. Now, the accuracy of the translation stage 11 is usually given in microns / mm, typically 3 microns / mm. Therefore, a movement of the order of a few microns is more accurate than a movement of a few millimeters. Given that the system described here is applied to research objects of micron size, an accuracy of the same order of magnitude, i.e. in the micron range, becomes crucial. Therefore, the method provides a gain in spatial resolution when measuring using the Raman microspectroscopy modality. Within the framework of the static method, as described in the introduction, the sample 200 is imaged over its entire surface, typically an area of ​​1 cm by 1 cm, before returning to scan the point of interest on the particle 202 using the Raman microspectroscopy modality. This leads to potential positioning errors of several microns.

[0197] Furthermore, the dynamic approach described in the first embodiment also limits the inconvenience of changing environmental conditions encountered in the static mode. In fact, the complete mosaic is acquired using dark field microscopy modality. Figure 4 It takes tens of minutes, or even an hour. Therefore, during the switching between dark-field microscopy and Raman microspectroscopy, environmental conditions (such as temperature, humidity, and vibration) may change and produce measurement inaccuracies. The dynamic mode described in this paper through two embodiments allows the two modalities to be linked together in a short time, or even allow the two modalities to be implemented simultaneously. As a result, the system is insensitive to environmental changes that occur on long time scales.

[0198] In addition to these advantages over static methods, the present invention also avoids the idle time corresponding to the mechanical switching time of switching from one modality to another within block 40. In fact, thanks to the system described above, switching from darkfield microscopy modality to Raman microspectroscopy modality is performed without the need for mechanical switching. Therefore, no mechanical switching time is required, which represents a significant time saving.

[0199] In practice, for microscopy applications, a high-precision mechanical switching system is used. Typically, such a mechanical switching system requires several seconds, for example, 6 seconds, to switch from dark field microscopy mode to Raman microspectroscopy mode, and several seconds, for example, 6 seconds, to switch in the opposite direction, which is time wasted on each imaged tile 40. Here, 12 seconds are lost for each imaged tile 40. In the case of a mosaic representing a large sample 200, Figure 4 At a scale of about a few centimeters by a few centimeters, including hundreds of tiles 40, or even millions of tiles 40, a lot of measurement time will be lost.

[0200] Even for more powerful mechanical switching systems, which are more expensive to extend, the mechanical switching time is still hundreds of microseconds. In fact, even in such mechanical switching systems, the mechanical switching time is only a few milliseconds. At the scale of large sample sizes of 200, these switching times are not negligible.

[0201] Therefore, the proposed method for spatial localization and spectral analysis of particles in images implemented by the system 1 saves time in analyzing large-sized samples (i.e., tens of millimeters on a side, or even several centimeters, and including low-contrast particles 202).

[0202] Tom Thumb method

[0203] A second embodiment of the disclosed method consists of Figure 4 Box in Figure 5 shown.

[0204] The box Figure 5 The initialization includes the following steps:

[0205] - using dark field microscopy modality to illuminate 50 an area on the sample 200,

[0206] - Recorded 51 dark field microscope images,

[0207] - locating 52 a plurality of particles 202 within the image,

[0208] - extracting 53 the coordinates of the points corresponding to these particles 202 to form a list,

[0209] - moving 54 the sample 200 so that the central area 144 of the dark field microscope objective 14 corresponds to a point in the extracted coordinate list, - illuminating 55 the sample 200 using the Raman microspectroscopy modality, the excitation laser beam 160 being focused on a measurement point that coincides with the point of the previous step,

[0210] The Raman spectrum emitted from the measurement point is collected 56 by the central area 144 of the dark field microscope objective 14 .

[0211] Once initialized, the method according to the second embodiment continues by moving 54 the sample so that the optical axis of the dark field microscope objective 14 corresponds to another point in the extracted coordinate list. Thereafter, the steps of irradiation 50, recording 51, positioning 52 using the dark field microscope modality are performed simultaneously with, before or after the steps of irradiation 55 and collection 56 using the Raman microspectroscopy modality.

[0212] Therefore, the method loops to the moving step 54 , for example, as long as the point list contains coordinate points 42 that have not yet been scanned and measured by Raman microspectroscopy.

[0213] In this second embodiment, a mosaic representing the sample 200 is reconstructed using multiple dark field microscope images. Figure 4 , as described in the first embodiment. Therefore, the splicing Figure 4 represents a larger field of view than the field of view of the dark field microscope image. The mosaic 4 consists of tiles 40, i.e. dark field microscope images, which are arranged to reproduce the analyzed sample 200. In a second embodiment, some of these tiles 40 at least partially overlap. In other words, the mosaic Figure 4 Some of the tiles 40 partially cover the same spatially identical area.

[0214] As with the first embodiment, each patch 40 has a rectangular shape of 140 micrometers by 105 micrometers according to the characteristics of the image sensor 15 and the dark field microscope objective lens 14 for collecting dark field microscope images.

[0215] The splicing obtained according to the second method Figure 4 Part of Figure 5 The rectangle represents Figure 3 The boundary of mosaic part 4 is also Figure 6 Reappearance in.

[0216] In this embodiment, the geometric center 41 of the tiles 40 is not predefined. Likewise, the arrangement of the tiles 40 relative to each other is not defined upstream in the process. In fact, in this second embodiment, the coordinate pairs defining the position of each tile 40 are defined during the process based on information collected during the implementation of the process.

[0217] More specifically, a geometric center 41 is selected based on the coordinate points 42 extracted during this process, thereby acquiring a new dark field microscope image and extracting new coordinate points 42. These new coordinate points 42 potentially become the geometric center 41 of the image block 40, and so on.

[0218] Thus, unlike the first embodiment, moving from one tile 40 to another does not necessarily involve a dedicated movement; Figure 4The tiling is performed when the movements 54 are completed. These movements 54 are intended to scan different coordinate points 42 in the point list, which then form the geometric center 41 of the tile 40.

[0219] With the help of Figure 5 With reference to the example in FIG, the steps previously described for the second embodiment of the method will now be described in detail.

[0220] To start the process, the stage 11 places the sample 200 at an arbitrary position. For example, Figure 5 The geometric center 41 of the first image segment 40 shown in FIG is selected here by the displacement amplitude of the displacement plate 11 along the x-axis and the y-axis at the mid-stroke. This position corresponds to the middle of the displacement stage 11.

[0221] The next step of the method is to illuminate 50 the area of ​​the sample 200 according to the dark field microscopy modality. As described above, the illumination 50 of the area of ​​the sample 200 is generated by the annular light cone 124 (centered on the geometric center 41 of the first image segment 40), which is focused onto the sample 200 through the peripheral area 142 of the dark field microscope objective 14.

[0222] exist Figure 5 In FIG. 4 , the geometric center 41 of each tile 40 is represented by a solid line crosshair, ie a cross surrounded by a circle.

[0223] Illumination 50 of the area using a darkfield microscope modality causes a first darkfield microscope image to be recorded 51. Again, the field of view of the image is contained within the illuminated area of ​​the sample 200. The first darkfield microscope image is obtained by collecting the stray light in the central area 144 of the darkfield microscope objective 14.

[0224] Based on the first dark field microscopy image, the potential particle 202 within the first image is positioned 52. This position 52 should be the same as that described in the first embodiment.

[0225] After the potential particles 202 contained in the image field have been located, for example, by means of a dedicated image processing algorithm, the coordinates of the points corresponding to the potential particles 202 are extracted 53. This extraction 53 is performed by the processing unit using a suitable algorithm and / or by the user, as described above. Similarly to the first embodiment, a single coordinate point 42 is associated with each particle 202 observed in the dark field, although alternative solutions are possible.

[0226] These extracted coordinate points 42 are added to the point list.

[0227] In the first block 40, two coordinate points 42 are extracted, namely coordinate point 421 and coordinate point 422. These coordinate points 42 are Figure 5 Indicated by a dotted line sight.

[0228] The translation stage 11 moves 54 relative to the sample 200 so that the optical axis of the darkfield microscope objective 14 coincides with a coordinate point in the point list. Specifically, a point that has not yet been scanned by the Raman microspectroscopy modality is selected. For example, this is coordinate point 421, which corresponds to coordinate point 421 associated with the first particle 202 identified and located within the image block 40.

[0229] The following steps involve irradiating 55 a measurement point on the sample 200 using the Raman microspectroscopy modality. The laser beam is focused onto the measurement point on the sample 200 via the central region 144 of the darkfield microscope objective 14. The measurement point is here a continuation of the optical axis of the darkfield microscope 14.

[0230] The movement 54 thus causes the illuminated measurement point to coincide with the coordinate point 421 taken from the point list.

[0231] After the illumination 55 and the movement 54 , the Raman spectrum emitted from the measurement point is collected 56 , which is collected by the central region 144 of the dark-field microscope objective 14 .

[0232] Advantageously, given that switching from one modality to another does not require mechanical switching, a new recording 50 of a new image block 40 can be made, the geometrical centre 41 of which coincides with one of the coordinate points 42 from the list L of points.

[0233] exist Figure 5 In the example shown, the geometric center 41 of the new image block 40 coincides with the coordinate point 421. The geometric center 41 of the first image block 40 is represented by a dotted cross, and the point is represented by a solid crosshair.

[0234] In a first alternative of this second embodiment, to obtain the new image segment 40, both illuminations are used simultaneously if the spectral ranges of the laser source 16 and the white light source 12 do not intersect and if the wavelengths of the components of the white light spectrum are lower than the wavelength of the excitation laser beam 160. This is assumed to be the case presented here.

[0235] As described above, the measurement point is illuminated 55 by the excitation laser beam 160 , the emitted Raman spectrum is subsequently collected 36 , and a new dark field microscope image is recorded 51 , the field of view of which is included in the illuminated area of ​​the sample 200 .

[0236] This recording 51 is followed by the localization 52 of the microparticle 202 and the extraction 53 of the coordinates of the point corresponding to the microparticle 202. Here, no point corresponding to the microparticle 202 is detected.

[0237] According to a second alternative, after the movement 54 , illumination 50 , recording 51 , positioning 52 and extraction 53 are first performed according to the dark field imaging modality, and then illumination 55 and collection 56 are performed according to the Raman spectroscopy modality.

[0238] Regardless of the order in which the two modalities are used, the method continues by shifting 54 samples so that the measurement point coincides with another point in the list L of points.

[0239] Here is the coordinate point 422, such as Figure 5 Then, this coordinate point 422 becomes the geometric center 41 of the new image block 40. What is imaged here is the third image block 40.

[0240] The steps of the method are then repeated by irradiating the potential 50, recording 51 a new dark field microscopy image forming the third image block 40, locating 52 and extracting 53 the coordinates of the points corresponding to the potential particles 202, these steps being performed in parallel with the steps of measuring the points by irradiating 55 the excitation laser beam and collecting 56 the Raman spectrum. In particular, the coordinate points 423 and 424 are detected and extracted from the third image block 40. These coordinate points 423 and 424 are added to the list of points to be scanned.

[0241] A new movement 54 of the stage then sets the measuring point at a point in the point list, and the steps of the method are thus repeated.

[0242] Assembly representing the structure and chemical properties of sample 200 Figure 4 In other words, the point list L is continuously updated during the implementation of the method and is not reset between each mosaic 40 .

[0243] With coordinate point 423 as the geometric center 41, a fourth image block 40 corresponding to a fourth dark-field microscopy image is acquired. Coordinate points 425 and 426 are extracted from this fourth image, and stage 11 moves sample 200 to detect coordinate point 424 using Raman microscopy. Coordinate point 424 becomes the geometric center 41 of the fifth image block 40, and the method steps are repeated for this fifth image block 40. Next, coordinate point 425 in point list L is scanned to collect a Raman spectrum, as well as a sixth dark-field microscopy image corresponding to the sixth image block 40. Coordinate point 427 is then extracted from this sixth image.

[0244] This method produces a mosaic corresponding to the entire surface of the representative sample 200 Figure 4 The spectral image of Figure 6 , i.e., it corresponds to the same spectral image as that obtained with the above-described first embodiment.

[0245] The structure of the microparticles 202 was obtained using a dark field microscopy modality that accentuates contrast even on these highly transparent samples 200, while the false colors were associated with the composition determined by Raman spectroscopy.

[0246] Within the framework of this second embodiment, the simultaneity of the two illuminations is exploited to use the movements required to detect the point of interest by means of the Raman microspectroscopy modality, also imaging the sample 200 by means of dark-field microscopy.

[0247] For example, the order in which the coordinate points 42 in the point list L are scanned is selected to minimize the distance the stage moves during the movement. For example, it is advantageous to minimize the Euclidean distance between two consecutively scanned coordinate points 42. Alternatively, the total movement distance required to scan all coordinate points 42 may be minimized.

[0248] Therefore, in the process, especially when a new coordinate point 42 is extracted, the scanning order is reorganized.

[0249] A suitable protocol can also be implemented if there are no coordinate points 42 in the list of points L that have not yet been scanned. For example, it is then decided to move the sample 200 by a sufficient amount to explore a new area of ​​the sample 200.

[0250] This second embodiment has the same advantages as the first embodiment, in particular eliminating the prohibitive mechanical switching time when switching from dark field microscopy mode to Raman microspectroscopy mode.

[0251] Alternative Embodiments

[0252] The invention is not in any way limited to the embodiments described and shown, but a person skilled in the art will know how to apply any variant according to the invention.

Claims

1. A method for performing spatial image localization and spectral analysis of particles (202) in a sample (200), the method utilizing an optical microscope system comprising the following steps: A) illuminating (30, 50) an area of ​​a sample (200) by an annular light cone originating from a white light beam (120) partially reflected by an annular mirror (13) and focused by a peripheral area (142) of a darkfield microscope objective (14), B) recording (31, 51) a first dark field microscope image over the field of view encompassed by the illuminated region of the sample (200), the first dark field microscope image being collected by a central region (144) of the dark field microscope objective (14), C) locating (32, 52) the particle (202) in said first dark field image, D) extracting the coordinates of the points corresponding to the particles (202) in the dark field image recorded by (33, 53) to form a point list, E) moving (34, 54) the sample (200) so that the optical axis of the dark field microscope objective (14) coincides with a point in the point list, F) irradiating (35, 55) a measurement point on the sample (200) by means of an excitation laser beam (160) which passes through a ring mirror (13) and is focused by a central region (144) of the dark field microscope objective (14), said measurement point coinciding with a point in the point list, G) collecting (36, 56) a Raman spectrum emitted from the measurement point, the Raman spectrum being collected by a central area (144) of the dark field microscope objective (14).

2. The method according to claim 1, wherein Steps A) and F) are performed simultaneously, illuminating an area on the sample (200) through a peripheral area (142) of the dark field microscope objective (14), and illuminating a measurement point within the area through a central area (144) of the dark field microscope objective (14).

3. The method according to any one of claims 1 to 2, wherein A field of view larger than the image field of view is reconstructed by a mosaic image (4) composed of a plurality of the dark field microscopic images, wherein the plurality of dark field microscopic images correspond to a plurality of image blocks (40) of the mosaic image (4).

4. The method according to claim 3, wherein the positioning of each tile (40) of the mosaic (4) is predefined upstream of the process.

5. A method according to claim 4, wherein steps E), F) and G) are repeated until all points of the extracted point list have been scanned on a given tile (40), and then a step (36) of moving the sample (200) is performed to focus the field of view of the dark field microscope objective (14) on another tile (40) of the mosaic (4), and the process is restarted from step A) to record another dark field microscope image.

6. The method according to claim 5, wherein: For each tile (40) of the mosaic (4), the point list comprises a determined number of points.

7. The method according to claim 3, wherein the positioning of at least one tile (40) of the mosaic (4) is established as a function of the coordinates of the points in the point list.

8. The method according to claim 7, wherein: After moving (54) step E) so that the optical axis of the dark field microscope objective (14) coincides with a point on the point list, steps A), B), C), D), F) and G) are performed to record a new dark field microscope image corresponding to a new tile (40) of the mosaic (4) and a Raman spectrum emitted from the measurement point, which steps are followed by another movement (54) of the sample (200) so that the optical axis of the dark field microscope objective (14) coincides with another point in the point list.

9. Method according to claim 8, wherein additional points extracted from the new dark field microscope image recorded after the movement (54) of step E) are added to the list of points to be scanned.

10. A system (1) for spatial localization in image and spectral analysis of microparticles, comprising an optical microscope, a laser source (16), a white light source (12), an image sensor (15), a spectrometer (19) and a control unit (6), wherein the control unit includes an image processing unit (60), the optical microscope comprising a sample holder (20) mounted on a translation stage (11), the sample holder (20) being adapted to accommodate a sample (200), characterized in that The system comprises: an optical system comprising a ring mirror (13) and a dark field microscope objective (14), said optical system being arranged between said laser source (16), said white light source (12) and said sample holder (20), said dark field microscope objective (14) having a central region (144) and a peripheral region (142), - a ring mirror (13) configured to reflect a portion of the white light beam (120) emitted by the white light source toward the peripheral region (142) of the dark field microscope objective (14), and to transmit the excitation laser beam (160) emitted by the laser source (16) toward the central region (144) of the dark field microscope objective (14), - the dark field microscope objective (14) is adapted to transmit the portion of the white light beam to the sample holder (20) via its peripheral region (142) and to transmit the excitation laser beam (160) to the sample holder (20) via its central region (144), - the darkfield microscope objective (14) is adapted to focus a portion of the light beam to illuminate a first area of ​​the sample (200) on the sample holder (20) through an annular light cone (124), and the darkfield microscope objective (14) is further adapted to focus the excitation laser beam (160) to a measurement point contained in the first area, - the central area (144) of the dark field microscope objective (14) is adapted to collect a dark field microscope image over the image field comprised by the first illumination area, the dark field microscope image being recorded by the image sensor (15), - the image processing unit (60) is adapted to locate a potential plurality of microparticles (202) and extract coordinates of points associated with the plurality of microparticles (202) from the dark field microscope image to create a point list, - the translation stage (11) is configured to move the sample holder (200) so that the measurement point coincides with a point in the point list, or to set the image field of the dark field microscope objective (14) to another area that is at least partially different from the first area; - the central area (144) of the dark field microscope objective (14) is also adapted to collect a Raman spectrum generated from the measurement point, said Raman spectrum being recorded by a Raman spectrometer (19).

Images