DEVICE FOR OBJECT RECOGNITION BY HOLOGRAPHY

Description

TECHNICAL FIELD

[0001] The present invention relates to the field of detection and analysis of objects by holographic imaging. It finds particularly advantageous application in the field of detection and analysis of small objects such as particles.

[0002] The invention can thus be used in the following non-limiting fields: air quality control, detection of microbiological species, detection of explosive powder as well as alarm systems such as those identifying smoke particles to detect fires. STATE OF THE ART

[0003] Holographic imaging of an object, based on recording a hologram with an image sensor and then digitally reconstructing the object, is commonly used to analyze a sample in order to identify and characterize particles or microscopic objects present in the sample.

[0004] As illustrated in figure 1, this technique makes it possible to observe a sample by placing it between a light source 10 and an image sensor 40, without having an optical magnification lens between the sample 20 and the image sensor 40. Thus, the sensor 40 collects an image 41 of the light wave transmitted by the sample. This image, also called a hologram, is formed of interference patterns between a light wave 11, called a reference wave, emitted by the source 10 and transmitted by the sample 20, and diffraction waves 13, resulting from the diffraction by the sample 20 of the light wave 11 emitted by the source 10. These interference patterns are sometimes called diffraction patterns, or designated by the English term “diffraction pattern”. If a holographic reconstruction of the sample is carried out using a computer system 60, an image 50 of the digitally reconstructed object 50 can be displayed on a screen 62.

[0005] Document WO2008090330 describes a device for observing biological samples, in this case cells, by lensless imaging. The device allows each cell to be associated with an interference pattern whose morphology allows the cell type to be identified. Lensless imaging then appears to be a simple and inexpensive alternative to a conventional microscope. In addition, its field of observation is significantly larger than that of a microscope. It is therefore clear that the application prospects linked to this technology are significant.

[0006] The document entitled "Practical algorithms for simulation and reconstruction of digital in-line holograms" (Applied Optics, Vol. 54, Issue 9, pp. 2424-2434 (2015) DOI: 10.1364 / AO.54.002424) or the document US2012 / 0218379, describe in particular holographic reconstruction algorithms which, applied to the image formed on the detector, make it possible to reconstruct the image of the sample in different reconstruction planes.

[0007] Thus, holographic imaging makes it possible to observe a sample containing particles, on the basis of elementary figures formed by each particle under the effect of illumination by an incident wave.

[0008] However, known solutions do not have sufficient resolution to identify samples containing dispersed particles with dimensions close to the illumination wavelength. It is therefore not possible to detect, identify or characterize these small objects.

[0009] Known holographic reconstruction solutions also do not allow morphological identification (e.g., external shape and structure) of small, near-wavelength or sub-wavelength particles.

[0010] This problem is illustrated in Figures 2A to 2C .

[0011] There Figure 2Aillustrates a sample 20 comprising objects such as dark particles 21a, 21b. In this sample, the critical dimensions CD are very much greater than the wavelength λ of the radiation emitted by the source. Indeed, a critical dimension CD defined by the distance between the two particles 21a and 21b is such that CD = 2µm (10 -6< meters), which is much greater than the wavelength λ = 450 nm (10 -9< meters) of the incident wave 11. In this example, the support 30 has dimensions X and Y along the x and y axes, such that X = Y = 50µm.

[0012] This sample 20 is placed near a support 30 having a light background. A sensor 40, placed at a distance Z1 = 1.1 cm (10 -2 < meters) from the support 30 carrying the sample 20, collects waves scattered 13 by the particles 21a, 21b and waves not scattered 12 by the particles. The recombination of these waves forms on the sensor 40 the interference pattern, i.e., the hologram 41 illustrated in Figure 2B .

[0013] It is possible to distinguish on this hologram 41 the particles 21a and 21b. If a holographic reconstruction of the sample is carried out, by applying one of the reconstruction methods mentioned in the state of the art above, the result illustrated in Figure 2C . This image, referred to as a reconstructed object, also allows particles 21a and 21b to be identified. It is possible to count the particles. The graph of the Figure 2Cwhich illustrates the light intensity as a function of the position on the y axis allows the identification of particle 51a and its position. Particle 51b can possibly be detected.

[0014] The result is quite different if the critical dimensions of the objects contained in the sample are close to or smaller than the wavelength of the radiation emitted by the source. This problem is illustrated with the Figures 3A to 3C .

[0015] There Figure 3Aillustrates a sample 20 comprising dark particles 21a-21c. In this sample, the critical dimensions CD are smaller than the wavelength λ of the radiation emitted by the source. In this example, a critical dimension is the distance between the two particles 21a and 21b. Here, this distance is equal to 189 nm. Another critical dimension is the width of the particle 21c. In this example, the wavelength λ = 450 nm (10 -9< meters). The support 30 has dimensions X and Y along the x and y axes, such that X = Y = 1000 µm (10 -6< meters). The distance Z1 between the support 30 carrying the sample 20 and the sensor 40 is equal to 1 cm.

[0016] The hologram 41 formed on the sensor 40 is illustrated in Figure 2B . It is not possible to distinguish particles 21a-21c on this hologram 41.

[0017] If we perform a holographic reconstruction of the sample we obtain a digitally reconstructed object illustrated in Figure 3C . This image also does not allow us to identify particles 21a-21c. A fortiori, it is impossible to count them and even less to characterize them. The graph of the Figure 3C which illustrates the light intensity as a function of the position on the y axis also does not allow the particles to be identified. In practice, we see that the resolution limit of such a system is 2λ, i.e., if a critical dimension of the sample 20 is less than or equal to 2λ, then this critical dimension is no longer identifiable.

[0018] It is noted that if the size of the objects had been much greater, without presenting critical dimensions, but the distance between two objects was close to or less than the wavelength, these two objects would not appear as distinct on the hologram or on the reconstructed object.

[0019] There is therefore a need to propose a solution to limit, or even eliminate, the disadvantages of known solutions. In particular, there is a need to enable the detection, or even the counting or identification of small objects, i.e. objects with critical dimensions close to the wavelength or less than it.

[0020] Other objects, features, and advantages of the present invention will become apparent from the following description and accompanying drawings. It is understood that other advantages may be incorporated. SUMMARY

[0021] To achieve this objective, according to one embodiment, a device is provided for detecting at least one object, for example present in a sample, the device comprising: a light source configured to emit incident waves at a wavelength λ, a detection volume intended to receive the object, and to receive the incident waves, an image sensor, positioned to receive at least scattered light waves obtained by scattering the incident waves on the object and a reference wave coming from the source and not having scattered on the object and to generate a holographic image depending at least on the scattered light waves and the reference waves, a computer data processing device configured to, according to one embodiment, digitally reconstruct the object depending at least on the holographic image and the wavelength λ.

[0022] The detection device also comprises a support comprising patterns organized to form at least one diffraction grating. The patterns of the at least one grating are organized periodically according to a pitch P, such that λ / 2≤ P ≤2λ.

[0023] The medium is configured so that the scattered waves collected by the sensor include either: waves from the source which have been scattered by the object without being diffracted by the at least one diffraction grating, and also waves from the source which have both been diffracted by the at least one diffraction grating and been scattered by the object. Reference waves are incident waves that have or have not been diffracted by the diffraction grating. The computing device is configured to digitally reconstruct the object as a function, in addition to the holographic image and the wavelength λ, of a distance Z1 between the support and the image sensor, the distance Z1 preferably being taken along a propagation axis z of the incident wave.

[0024] So the detection device is configured so that the sensor collects: waves from the source which have both been diffracted by the at least one diffraction grating and have both been scattered by the object, and waves from the source which have been scattered by the object without being diffracted by the at least one diffraction grating, waves from the source which have not been scattered by the object.

[0025] Thus, the diffraction grating whose period is close to the incident wavelength λ, makes it possible to increase the interaction between light and the object such as a particle. The image sensor collects more light scattered by the object. To the light scattered by the object and transmitted by the grating is added the light scattered by the object and diffracted by the grating.

[0026] Thus, the diffraction grating carried by the support acts as a spatial filter which increases the resolution of the reconstruction of the object from the image captured by the sensor, that is to say from the hologram.

[0027] The device according to the invention allows a resolution of critical dimensions of the order of the wavelength λ, or even less than the wavelength λ. Typically, the device offers very good resolution for critical dimensions greater than λ / 4.

[0028] The proposed system thus makes it possible to: distinguish, and therefore for example count, particles separated by a distance of the order of wavelength λ or even less than wavelength λ without the risk of considering that it is a single particle of larger size, to identify particles whose size is of the order of wavelength λ, or even less than wavelength λ.

[0029] The proposed device also allows for counting and morphological identification of larger objects, typically larger than 10λ. This is the case, for example, for pollens and molds. In addition, the present invention makes it possible to remove constraints on the position of the object to be analyzed. Indeed, the object to be analyzed does not necessarily have to be in contact with the support.

[0030] According to a second embodiment, which may be an alternative or combined with the first embodiment in which the data processing computer device is configured to digitally reconstruct the object, the data processing computer device is configured to carry out a count of the objects present in the detection volume. For this, filters are applied to the generated hologram in order to single out the different particles.

[0031] Optionally, the device according to the invention may further have at least any one of the following characteristics which may be taken separately or in combination: According to one example, the device does not include a magnifying optic between the support and the image sensor. According to one example, the grating has a pitch P of between 0.75*λ and 1.5*λ. Preferably, the grating has a pitch P equal to λ.

[0032] The invention finds particularly advantageous, but not limited, applications in the monitoring of air quality, the detection of microbiological species, the detection of explosive powder as well as alarm systems such as those identifying smoke particles to detect fires.

[0033] The present invention also relates to a system comprising a detection device according to the invention in which the system is taken from: a fire alarm system, a fire detection system, a system for analyzing the quality of a fluid such as air or water, an anti-pollution alarm system, an explosive powder detection system, a microbiological species detection system, a DNA analysis system, a pollen analysis system, a mold analysis system.

[0034] According to one embodiment, a device is provided for detecting at least one object, for example present in a sample, the device comprising: an image sensor, positioned to receive at least scattered light waves obtained by diffraction on the object of incident waves of wavelength λ emitted by a source and reference waves coming from the source and not having diffracted on the object and to generate a holographic image depending at least on the scattered light waves and the reference waves.

[0035] The device also comprises a support comprising patterns organized to form at least one diffraction grating. The patterns of the at least one grating are organized periodically according to a pitch P, such that λ / 2≤ P ≤2λ.

[0036] The detection device is configured so that the sensor collects at least: waves from the source which have both been diffracted by the at least one diffraction grating and have both been scattered by the object, and waves from the source which have been scattered by the object without being diffracted by the at least one diffraction grating, waves from the source which have not been scattered by the object.

[0037] The device may or may not understand the source. BRIEF DESCRIPTION OF THE FIGURES

[0038] The aims, objects, as well as the characteristics and advantages of the invention will emerge more clearly from the detailed description of an embodiment thereof which is illustrated by the following accompanying drawings in which: There figure 1 schematically represents a holographic imaging detection device according to the state of the art. The Figure 2A to 2C illustrate the results obtained, with a detection device according to the state of the art, to analyze a first sample. The Figures 3A to 3C illustrate the results obtained, with a detection device according to the state of the art, to analyze a second sample. The figure 4 schematically represents an example of a holographic imaging detection device according to an embodiment of the present invention. The Figures 5A to 5C , illustrate the results obtained by analyzing the same sample as for the Figure 2A to 2C, but with a device according to the invention. The figure 6 schematically represents an example of a two-dimensional diffraction grating equipping a detection device according to the present invention. Figure 7A illustrates the results obtained by analyzing a sample with a device according to the state of the art. The Figures 7B to 7D , illustrate the results obtained by analyzing the same sample as for the Figure 7A , but with devices according to the invention having networks whose geometries are identical to that of the figure 6 , but whose periods vary. The Figure 8A schematically represents an example of a one-dimensional diffraction grating equipping a detection device according to the present invention. The Figures 8B and 8C , illustrate the results obtained by analyzing the same sample as for the Figures 7A to 7D , but with the device of the Figure 8A integrating a one-dimensional network. The Figures 9A to 9Crepresent the results obtained by analyzing the same sample with networks presenting identical periods but presenting different recovery rates. The Figures 10A to 10E illustrate a detection device comprising two distinct periodic networks as well as the results obtained with this device. The Figures 11A to 11C illustrate the results obtained with a detection device comprising four distinct periodic gratings. The Figures 12A to 12C illustrate a detection device comprising two interlaced periodic networks as well as the results obtained with this device. The Figure 13A illustrates a holographic image obtained with a device whose image sensor has a certain dimension. The Figures 13B and 13C represent the results obtained by digital reconstruction of the holographic image of the Figure 13A . There Figure 14Aillustrates a holographic image obtained with a device whose image sensor has dimensions different from those of the image sensor of the Figure 13A . THE Figures 14B and 14C represent the results obtained by digital reconstruction of the holographic image of the Figure 14A . There Figure 15A illustrates a holographic image obtained with a device whose image sensor has dimensions different from those of the image sensors of Figures 13A And 14A . THE Figures 15B and 15C represent the results obtained by digital reconstruction of the holographic image of the Figure 15A . THE Figures 16A to 16C are photos, taken by scanning electron microscope, of several objects to be analyzed. The figure 17 illustrates, in a schematic manner, a device according to the state of the art used to analyze the objects of the Figures 16A to 16C . THE Figures 18A to 18Dillustrate a detection device according to the invention comprising two diffraction gratings as well as the results obtained with this device for analyzing the objects of the Figures 16A to 16C . THE Figures 19A to 19C illustrate another detection device according to the invention as well as the results obtained with this device for analyzing a strand of DNA. The Figures 20A to 20C illustrate another detection device according to the invention, comprising a one-dimensional network, as well as the results obtained with this device for analyzing a strand of DNA. The figure 21 illustrates, schematically, a structured support for a detection device according to the invention. The Figures 22A to 22C illustrate particular examples of support comprising diffraction gratings. In these examples, the grating patterns are separated by apertures. The figure 23schematically represents an example of a holographic imaging detection device comprising several diffraction gratings arranged in series. The figure 24 schematically represents a portion of an example of a holographic imaging detection device according to the state of the art. The figure 25 schematically represents a portion of an example of a holographic imaging detection device according to the invention.

[0039] The drawings are given as examples and are not limiting of the invention. They constitute schematic representations of principle intended to facilitate the understanding of the invention and are not necessarily to the scale of practical applications. In particular, the relative dimensions of the objects to be analyzed and of the different patterns of the networks are not representative of reality. DETAILED DESCRIPTION

[0040] Before commencing a detailed review of embodiments of the invention, optional features which may optionally be used in combination or alternatively are set out below:

[0041] In one example, the network extends in an xy plane perpendicular to the propagation axis z.

[0042] The reference wave corresponds to the incident wave which has not diffused on the object.

[0043] A wave collected by the sensor after having been both diffracted by at least one diffraction grating and having been scattered by the object means that this same wave was either first scattered by the object then diffracted by the grating or conversely that it was first scattered by the grating then diffracted by the object.

[0044] According to one example, the support is arranged between the detection volume and the image sensor. Thus, waves are scattered by the object and then diffracted by the grating. According to an alternative embodiment, the detection volume is arranged between the support and the image sensor. Thus, waves are diffracted by the grating and then scattered by the object.

[0045] In one example, the network is a network of parallel straight lines. Thus, it is a one-dimensional network. In one example, the straight lines extend in the plane of the support. This embodiment has the advantage of being particularly simple and inexpensive to produce. In addition, when the lines form ribs and trenches, it is possible to arrange the objects that one wishes to produce inside the trenches. This allows their position to be controlled with greater precision. Furthermore, this allows their position to be maintained during handling or transport.

[0046] According to another example, the grating is a two-dimensional grating, along a first dimension of the grating, taken in the xy plane, the grating has a pitch Pa and along a second dimension of the grating, taken in the xy plane and perpendicular to the first dimension, the grating has a pitch Pb, with Pa and Pb such that λ / 2≤ Pa ≤2λ and λ / 2≤ Pb ≤2λ. According to one example, Pa = Pb, preferably, the two-dimensional grating comprises or is formed of point patterns, the distance between two point patterns corresponding to the pitch P = Pa = Pb of the grating. This embodiment has the advantage that the patterns of the grating occupy a smaller surface area than the patterns of a one-dimensional grating. The support carrying a two-dimensional grating is therefore more transparent than a support carrying a one-dimensional grating. This makes it possible to improve the contrast and the signal-to-noise ratio.

[0047] According to one example, the support comprises patterns organized so as to form at least two periodic diffraction gratings, including a first grating having a pitch Pa, and a second grating having a pitch Pb, with Pa and Pb such that Pa≠Pb, λ / 2≤ Pa ≤2λ and λ / 2≤ Pb ≤2λ. According to one example, the light source is configured to emit at least a first incident wave and a second incident wave, respectively having a wavelength λa and λb, such that λa / 2≤ Pa ≤2λa and λb / 2≤ Pb ≤2λb, preferably 0.75*λa / 2≤ Pa ≤1.5*λa and 0.75*λb / 2≤ Pb ≤1.5*λb and preferably λa=Pa and λb=Pb.

[0048] According to one embodiment, the first and second waves are emitted simultaneously by the source. According to another embodiment, the first and second waves are emitted successively by the source. This embodiment makes it possible to obtain several images, each image corresponding to an analysis wavelength. This thus allows a chemical analysis according to the molecular absorption of the constituent components of the particles. For example, λa belongs to the infrared range and λb belongs to the ultraviolet range.

[0049] According to one example, the at least two gratings are located on separate areas of the support. Thus, the gratings are not interlaced. This embodiment has the advantage of being able to associate with each of the areas of the hologram an image corresponding to a wavelength. The processing is thus simplified.

[0050] Preferably, the gratings are distinct but contiguous. Alternatively, they may be distinct and separated from each other by a gap in a plane parallel to an xy plane perpendicular to a principal propagation axis z of the incident wave arriving at the medium.

[0051] According to one example, the at least two networks are at least partly and preferably entirely interlaced.

[0052] Thus, in projection onto an xy plane, the networks are superimposed.

[0053] This embodiment makes it possible to reduce the size of the device.

[0054] According to one embodiment, the networks are in contact or are contained in one or more identical planes, parallel to the xy plane. The networks may also not be in contact. In this case, they do not have a space between them capable of allowing particles to pass through.

[0055] This embodiment allows each of the networks to be produced during the same manufacturing step, for example during the same lithography step. This eliminates the alignment problems that would occur if each area had to undergo a dedicated lithography step. This embodiment thus makes it possible to address the problem of improving the accuracy of the detector while limiting its manufacturing cost.

[0056] According to another embodiment, the networks are superimposed without being in contact. They are then at a distance from each other, a non-zero distance, taken along the z direction, separating them. This is for example the case if a network is formed by a first face of a layer (for example if this face is structured to form this network) and another network is formed by a second face, opposite the first face of this same layer or another network is formed in the thickness of this layer (for example if this network is formed by the network of atoms of this layer).

[0057] Thus, according to one embodiment, the two networks can cover the same area. Alternatively, they can overlap on only part of the areas that they respectively cover.

[0058] Preferably, each wave propagates along at least the propagation axis z.

[0059] According to one example, the object has a critical dimension CD less than or equal to λ.

[0060] According to one example, the at least one network occupies on the support an area having, at the wavelength λ, a transmission coefficient T such that T ≥ 0.5, preferably T ≥ 0.75 and preferably T ≥ 0.96. This makes it possible to improve the contrast of the image obtained on the sensor as well as the background noise signal.

[0061] In one example, the network patterns are made of a dielectric material.

[0062] Alternatively, the network patterns are made of a material having a heat conduction coefficient greater than or equal to 5 Wm -< 1< .K -1< , the lattice patterns preferably being made of a metallic material such as tungsten. In one example, the lattice patterns form heating elements when energized with electricity.

[0063] Preferably, the patterns are made of metal.

[0064] Typically, the support patterns can be brought to a temperature of at least 200°C for the desorption of organic or volatile materials. The temperature can go up to 1700°C or even 2400°C for refractory metals such as molybdenum or tungsten. Those skilled in the art may use alloys such as molybdenum disilicide MoSi2 or silicon carbide SiC, in particular due to a lower cost.

[0065] These embodiments allow the support to be heated. This has the advantage of being able to destroy the particles in contact with the support by heating, thus allowing the latter to be cleaned. This embodiment thus provides a solution to the problem of extending the lifespan and sensitivity of the detector.

[0066] These embodiments also make it possible to differentiate the nature of the particles. Indeed, heating the network will degrade certain particles, typically organic particles, without degrading other particles, typically metallic particles. By making successive recordings, without, then with heating, it is then possible to discriminate the analyzed objects according to their sensitivity to heat.

[0067] In one example, the lattice comprises, or is solely formed by, at least one crystal lattice, the units of the lattice being formed, at least in part, by atoms of the crystal lattice.

[0068] According to one example, the support comprises a layer having a face structured so as to form a first network of relief patterns and a second network formed by said crystalline network.

[0069] According to one example, the source emits the incident wave along a main propagation direction z, and the support has a main face extending in a plane P 30 perpendicular to the propagation axis z.

[0070] According to one example, the source emits the incident wave along a main propagation direction z, and the sensor has a plurality of photodetectors extending in a plane P 0 perpendicular to the propagation axis z.

[0071] The source, the support and the image sensor are aligned, preferably along a propagation direction (z) of the incident wave. Thus, the detector is said to be "in-line".

[0072] According to an alternative embodiment, the sensor receives the waves diffused by the support on the one hand and receives, on the other hand, the reference wave, without the latter passing through the support. For example, a reflection device or an interferometer is provided in order to divide the radiation from the source into a wave arriving at the object and a wave arriving directly, or by reflection, at the sensor without passing through the support. In this case, we speak of a transmission hologram, in an “off axis” configuration (i.e. off-axis).

[0073] Reference waves are incident waves that have or have not been diffracted by the diffraction grating.

[0074] According to one example, the object has a critical dimension CD less than or equal to λ.

[0075] According to another example, the object has a critical dimension CD, such that CD≥2*λ and preferably CD≥10*λ. According to one embodiment, CD≥ 10µm 10 -6< meters).

[0076] In one example, the image sensor includes a plurality of photodetectors, typically arranged in an array.

[0077] In one example, the object is opaque. It is placed on a transparent background.

[0078] In one example, the object is mobile in the detection volume.

[0079] In one example, the support includes openings located between the patterns. Preferably, the openings separate the patterns. In one embodiment, the openings extend from a given pattern and to another pattern, adjacent to the given pattern.

[0080] In one example, the support forms a grid whose mesh is formed by the patterns.

[0081] In one example, the patterns are fully supported by the frame. In one embodiment, the frame extends only around the perimeter of the network.

[0082] According to one example, the detection device comprises a plurality of supports arranged in series, spaced from each other along the propagation axis z and each comprising at least one diffraction grating, such that the at least one object present in the detection volume is located between the gratings of two adjacent supports. The gratings thus channel the objects.

[0083] According to an example, the distance between the sample, more particularly the object(s) to be analyzed, and the network is less than Z1. Thus, we can have Z2≠0 and more generally 0 <Z2<Z1.

[0084] According to a non-limiting example, the data processing computing device is configured to obtain the reconstruction of at least one image from a digital hologram generated by the sensor, by performing at least the following steps using a computing device: 1 / definition of a propagation matrix in the image plane as a function of N (number of pixels of the image sensor); λ; the size of the object (typically the width of the matrix assimilated to the object); the distance Z1 between the support (or the object in contact with the matrix) and the image sensor. 2 / multiplication of the digital hologram by the propagation matrix, 3 / obtaining at least one reconstructed image by performing an inverse Fourier transformation on the result of the previous step.

[0085] In the context of the present invention, the term “particle” or its equivalents is defined as a constituent of a physical system considered to be elementary in relation to the properties studied.

[0086] The term "particle" refers in particular to a solid, liquid or wet solid object suspended in the air and whose size is microscopic. For example, a particle is an element of matter whose largest dimension is less than a few millimeters (10 -3< meters), preferably one millimeter, preferably less than 100µm, preferably a few tens of micrometers (10 -6< meters) and preferably less than a micrometer, or even of the order of a nanometer (10 -9< m). These dimensions may correspond to a diameter of the particle or the diameter of a sphere in which the particle is inscribed.

[0087] More generally, particles are larger than 40 Å (10 -10 < m) and are therefore considered optically continuous. Generally speaking, these are objects composed of matter whose dimensions are small compared to the dimensions of the cavity or channel through which the particles circulate.

[0088] The "size" or "diameter" of a particle is the maximum distance between two points on the particle. Typically, a particle is thought of as a spherical object, so its size corresponds to the diameter of the sphere.

[0089] In the present description, a material is considered transparent when it allows at least 50% of light radiation to pass through, preferably at least 75%, preferably at least 90% and advantageously at least 95%. A so-called transparent material corresponds to this definition at least for the wavelength at which the detector incorporating this material is intended to operate.

[0090] In the present description, a material is considered opaque when it blocks at least 50% of light radiation, preferably at least 75%, preferably at least 90% and advantageously at least 95%. A material corresponds to this definition at least for the wavelength at which the detector incorporating this material is intended to operate.

[0091] In the following, the terms "diffraction", "diffusion" or their equivalents refer to the phenomenon by which a propagation medium produces a distribution, in many directions, of the energy of an electromagnetic wave, light for example.

[0092] There figure 4represents an example of a device according to the invention. A light source 10 is configured to emit at least one light wave 11, called an incident light wave, propagating towards a sample 20. This propagation can be carried out directly on the sample 20, for example by propagating along a main propagation axis, designated propagation axis z of the orthogonal reference frame xyz illustrated in figure 4 Alternatively, the incident light wave 11 can reach the sample 20 after reflection on optical devices.

[0093] The light wave 11 is emitted according to a spectral band Δλ, comprising a wavelength λ. This wavelength may be a central wavelength of said spectral band. According to a non-limiting example, λ belongs to the infrared or ultraviolet wavelengths.

[0094] The sample 20 is a sample that one wishes to characterize. It may in particular be a medium comprising one or more objects 21 such as particles. The medium, in which the particles are located, may be a gaseous fluid, such as air. This is the case, for example, if the particles to be detected are polluting particles such as smoke particles, etc. According to another embodiment, the medium may be a liquid in which the particles are bathed. It may, for example, be a bodily fluid, a culture medium, or a liquid taken from the environment or from an industrial process. It may also be a solid medium or one having the consistency of a gel. It may also be an evaporated, fixed, or frozen sample.

[0095] The sample 20 is placed on or near a support 30. The support 30 has at least one face, facing the sample 20 and extending mainly along an XY plane perpendicular to the propagation axis z of the incident wave 11 upon its arrival on the object 20. According to one embodiment, the sample 20 is held on the support 30. Alternatively, it may be at a distance Z2≠0 from the latter, in particular when the sample 20 comprises particles suspended in a gas flow.

[0096] The support 30 has at least one periodic diffraction structure. This periodic structure forms a network 32 of patterns 31. Preferably, the support 30 has a support, also called a background, which is transparent and the patterns 31 of the network are opaque. In this case, the patterns 31 are preferably made of a dielectric material or a metallic material such as tungsten.

[0097] According to an alternative embodiment, the support 30 has an opaque support or background and the patterns 31 of the network are transparent. The patterns 31 of the network 32 are organized periodically according to a pitch P, such that λ / 2≤P≤2*λ, preferably λ / 0.75≤P≤1.5*λ and preferably λ / 0.9≤P≤1.1*λ. The network acts as a spatial filter. The characteristics, functions and advantages of the support 30 carrying at least one network 32 will be detailed later.

[0098] The distance between the light source 11 and the sample 10 is preferably greater than 1 cm. It is preferably between 2 and 30 cm. It can be said that it is a plane or spherical wave, which makes it possible to overcome the distance from the source. For reconstruction from the digital hologram, we are limited to these two cases for the description of the source.

[0099] Preferably, the light source 10, seen by the sample 20, is considered to be point-like. This means that its diameter (or its diagonal) is preferably less than one tenth, better still one hundredth of the distance between the sample 20 and the light source 10. Thus, preferably, the light reaches the sample 20 in the form of plane waves, or waves that can be considered as such. The light source 10 may be a light-emitting diode or a laser diode. It may be associated with a diaphragm, or spatial filter. Alternatively, the light may reach the sample 20 in the form of spherical waves.

[0100] Preferably, the emission spectral band Δλ of the incident light wave 11 has a width less than 100 nm. By spectral bandwidth is meant a width at half-maximum of said spectral band. In general, a wave is considered to be quasi-monochromatic if Δλ << λ (λ being usually designated λ 0 in this context) with typically Δλ ~ ±5%. Generally, monochromaticity is defined with respect to the wave frequency f 0 of the optical signal: Δf << f 0 .

[0101] The sample 20 and the support 30 are arranged between the light source 10 and an image sensor 40. The latter preferably extends parallel, or substantially parallel to the plane along which the sample 20 extends. The term substantially parallel means that the two elements may not be strictly parallel, an angular tolerance of a few degrees, less than 20° or 10° being allowed. In the example illustrated in figure 4, these planes correspond to the xy plane of the xyz reference frame.

[0102] The image sensor 40 is capable of forming an image along a detection plane P 0 . The detection plane P 0 preferably extends perpendicular to the propagation axis z of the incident light wave 11, that is to say it extends parallel to the xy plane. In the example shown, the image sensor 40 may comprise a matrix of photodetectors, each photodetector comprising a pixel, of the CCD or CMOS type. CMOS are the preferred sensors because the size of the pixels is smaller, which makes it possible to acquire images whose spatial resolution is more favorable.

[0103] The distance Z1 between the support 30, more particularly between the network 32 of the support 30 and the pixel matrix of the image sensor 40 is preferably between 0.01 cm (10 -2 < meters) and 20 cm, and preferably between 0.1 cm and 1 cm. More precisely, Z1 is measured between a front face of the sensor 40 and the rear face of the network 32. The front face of the sensor 40 is turned opposite the rear face of the network 32

[0104] According to one example, the objects 21 to be analyzed, whether static or mobile in a fluid, are located between the network 32 and the image sensor 40. The detection volume intended to comprise the particles is therefore preferably located between the network 32 and the image sensor 40, preferably along the propagation axis z. Thus, if the detection device comprises a channel configured to allow the flow of a fluid carrying the particles, this channel is therefore preferably located between the network 32 and the image sensor 40. Alternatively, the detection device is configured so that the objects 21 are located upstream of the network 32 relative to the incident wave. Thus, the network 32 is located between the object 21 and the image sensor 40 along the propagation axis z.

[0105] Preferably, the device is further configured so that the distance Z2 between the sample 20, more particularly the objects 21 to be analyzed and the network 32 is less than Z1.

[0106] Thus, as indicated above, it is not necessary for the objects 21 to be arranged in contact with the network 32. Thus, we can have Z2≠0 and more generally 0 <Z2<Z1. Cela permet de relâcher les contraintes sur le positionnement des particules 21. Il est donc possible d'analyser un échantillon dans lequel la position des particules 21 n'est pas parfaitement maîtrisée. Tel est, par exemple, le cas de particules comprises dans un flux gazeux. De préférence, si Z2≠0, c'est-à-dire si l'objet 21 n'est pas au contact du réseau 32, alors on prévoit que l'objet 21 soit à une distance Z2 du réseau 32 de sorte que Z2 ≤ 100 * λ. Cela permet d'améliorer la qualité de l'hologramme et de l'objet reconstruit numériquement.

[0107] We can also have all or some of the objects 21 arranged in contact with the network 32 such that Z2 = 0. It is in this scenario that we position ourselves in the different examples below.

[0108] Z1 and Z2 are measured along the propagation axis z.

[0109] The paragraphs below describe several embodiments of the support 30 comprising the diffraction grating 32.

[0110] According to a first embodiment, the support 30 comprises, for example, a transparent layer on which the network 32 of patterns 31 rests. There transparent layer can serve as a network holding layer 32, or even as a handling substrate. Theretransparent layer is, for example, formed of silicon. Preferably, this silicon layer has a low thickness to reduce absorption. Its thickness is typically less than or equal to 100nm. Preferably, this layer is quartz-based or is made of quartz for visible radiation.

[0111] The patterns are, for example, formed from metal or dielectric material such as SixOy or SixNy, with x and y being non-zero integers.

[0112] To produce the support 30, it is possible, for example, to provide a stack comprising at least one transparent layer and one opaque layer covering the transparent layer. The patterns 31 of the network 32 are then formed by removing a portion of the opaque layer. For this, one of the known lithography techniques may be used. It will be noted that it is possible to maintain a continuous layer of the opaque material if, between the patterns 31, the latter has a thickness that is sufficiently thin to allow a sufficient portion of the light waves to pass through. Such would be the case, for example, with silicon. It is then possible to use nanometric printing techniques.

[0113] Alternatively, it is possible to provide for the network 32 already formed on the support 30.

[0114] In another embodiment, the periodic structures do not rest on a layer or support element. They are suspended in a vacuum or air, as illustrated in Figures 22A to 22C .

[0115] The patterns 31 are then separated from each other without continuity of material. Physical openings 37, i.e. vacuum or air, therefore separate the patterns 31. The patterns 31 are held together, for example by their ends or by junction points 38 as illustrated in Figure 22A . A frame 39 can also be provided which holds the patterns 31 as illustrated in Figure 22B . The diffraction grating 32 thus comprises a frame 39 supporting the patterns 31. The support 30 presented is thus in the form of a grid. The patterns 31 form the meshes of the grid. According to one embodiment, the patterns 31 are entirely supported by the frame 39. According to one embodiment, the frame 39 extends only around the periphery of the grating 32.

[0116] Only the openings 37 can be provided between the patterns 31, i.e. an absence of material as illustrated in figure 22A or 22B . It is also possible to provide that the patterns 31 are connected to each other by transparent portions 37' and that openings 37 are also present between the patterns 31. An example of this embodiment is illustrated in Figure 22C .

[0117] All these embodiments comprising openings 37 make it possible to further increase the transparency of the areas between the patterns 31 and ultimately to obtain an even more precise reconstructed image. On the Figures 2A to 2C , one-dimensional patterns are illustrated. Naturally, these embodiments with openings 37 are applicable to any type of pattern.

[0118] We note the absence of magnification optics between the sensor 40 and the sample 20. This does not prevent the possible presence of focusing microlenses at each pixel of the sensor 40, the latter not having a function of magnifying the image acquired by the sensor 40. The microlenses of the CMOS sensors make it possible to collect more optical flow to be concentrated on the active zone of the sensor 40 which only occupies a small part of the image pixel.

[0119] Under the effect of the incident light wave 11, the object 21 of the sample 20 can generate a diffracted wave 13, capable of producing, at the detection plane P 0 , interference, in particular with a wave called a reference wave. This reference wave is an incident wave which may or may not have been diffracted by the diffraction grating. This reference wave is advantageously a part of the incident light wave transmitted 12 by the sample 20, without diffraction by the object 21 of the sample. It can be transmitted through the support 30. It will be noted that with or without a diffraction structure, only a part of the incident wave is transmitted. Furthermore, the sample 20 can absorb a portion of the incident light wave 11. Thus, the light wave transmitted by the support 30, and to which the image sensor 40 is exposed, can comprise: a component not diffracted by the sample 20 and at least part of which is possibly diffracted by the grating 32 of the support 30. This component corresponds to a part of the incident light wave 12 not absorbed by the sample 20. a component resulting from the diffraction by the sample 20 of the incident light wave 11 and at least part of which is also diffracted by the grating 32.

[0120] This resulting light wave 14 can also be referred to as the “exposure light wave”. This exposure light wave 14 generates on the sensor 40 an image 41 also referred to as a holographic image or hologram.

[0121] It should be noted that the detection device illustrated in figure 1is said to be "in line", since the object 20, the support 30 and the sensor 40 are located on the propagation axis z of the incident wave 11. It will be noted that the invention also applies to detection devices which are not in line and which are usually referred to as transmission holograms, in "off axis" configuration (i.e. off-axis). In these detection devices, the component not diffracted by the sample and the component diffracted by the sample take different optical paths, having non-parallel axes. It can be provided that one of these two components reaches the image sensor 40 after at least one reflection on a reflector.

[0122] A computer equipment 60, comprising a data processing device 61 such as a processor, for example a microprocessor, is capable of processing each image acquired by the image sensor 40. In particular, the processor is a microprocessor connected to a programmable memory in which is stored a sequence of instructions for carrying out the image processing operations and calculations for reconstructing the holographic image so as to obtain a digitally reconstructed image 50. The processor can be coupled to a screen 62 and possibly to a user interface 63. The screen 62 allows the display of images acquired by the image sensor 40 or calculated by the processor.

[0123] The image 41 acquired on the image sensor 40, also called a hologram, does not make it possible to obtain a sufficiently precise representation of the observed sample 20. A propagation operator can be applied to each image 41 acquired by the image sensor 40, so as to calculate a quantity representative of the exposure light wave 14 transmitted by the sample 20, and to which the sensor 40 is exposed. Such a method, conventionally designated by the term holographic reconstruction, makes it possible in particular to reconstruct an image of the modulus or the phase of this exposure light wave 14 in a reconstruction plane parallel to the detection plane P 0 , and in particular in the plane P 30 along which the sample 20 mainly extends.

[0124] To perform a holographic reconstruction from the image 41 generated by the sensor, one can refer to the following publications "Practical algorithms for simulation and reconstruction of digital in-line holograms" (Applied Optics, Vol. 54, Issue 9, pp. 2424-2434 (2015) DOI: 10.1364 / AO.54.002424).

[0125] In a very simplified manner, from the image 41 generated by the sensor 40, the wavelength(s) λ, the distance Z1 and the nature of the incident wave (plane or spherical), the following main steps are carried out to digitally reconstruct the image of the object 21: 1 / Reading the digital hologram. This matrix is, for example, named "Holo(x,y )". It describes the hologram. This matrix corresponds a light intensity to each point of the hologram. This matrix is a function of: the number of pixels of the image sensor 40 "NbPix". For the sake of simplification, we can provide that the matrix of the image sensor is square and has a number N of pixels per side (i.e. per row and per column). Also for the sake of simplification, we provide that the object is assimilated to a matrix of the same dimensions as those of the image sensor. This simplification is, however, perfectly reasonable in the absence of an optical device located between the object and the image sensor. the size of the object "sizeObject". This size of the object thus corresponds to the width of the matrix assimilated to the object. The distance Z between the imager and the object.For the sake of simplification, we consider that the object is in contact with the diffraction grating. Thus, Z=Z1. 2 / Definition of the propagation matrix in the image plane. This matrix is a function of: N (number of pixels per side); λ; ObjectSize; Z1). This matrix is, for example, named "Prop(N; lambda; ObjectSize; Z)". We can define Z as an iteration variable to optimize the ideal position for image reconstruction. Indeed, the distance between the particle and the plane of the digital sensor recording the hologram can vary and not be perfectly known. We can then find it by the reconstruction calculation, by adjusting the value of Z. 3 / multiplication of the hologram by the propagation matrix, i.e.: HoloZ(x,y)=Holo(x,y)*prop(N; lambda; objectSize; Z) 4 / We then perform the inverse transformation of the result obtained in the previous step: Rec=FFTinverse 2D(HoloZ). .

[0126] More generally, reference may be made to document WO 2017 / 178723 to precisely characterize an object or a particle from an image 41 produced by the sensor 40.

[0127] According to an optional embodiment, the presence of the network 32 can be taken into account by applying image processing, for example by applying a spectral filter. This can further improve the quality of the reconstruction of the object.

[0128] As indicated in this document, in order to reconstruct an image of the modulus or phase of the exposure light wave 14 in a reconstruction plane parallel to the detection plane P 0 of the image sensor 40, a convolution product of the image 41 acquired by the image sensor 40 is performed by a propagation operator h. It is then possible to reconstruct a complex expression A of the exposure light wave 14 at any point with coordinates (x, y, z) in space, and in particular in a reconstruction plane P z located at a distance | z | from the image sensor 40, this reconstruction plane possibly being the plane of the sample P 30 , in the case where | z | = Z1. The complex expression A is a complex quantity whose argument and modulus are respectively representative of the phase and intensity of the light wave 14 to which the image sensor 40 is exposed.The convolution product of the image 41 by the propagation operator h makes it possible to obtain a complex image A z representing a spatial distribution of the complex expression A in a plane, called the reconstruction plane P z , extending to a coordinate z of the detection plane P 0 . In this example, the detection plane P 0 has the equation z = 0. This complex image corresponds to a complex image of the sample in the reconstruction plane P z . It also represents a two-dimensional spatial distribution of the optical properties of the wave 14 to which the image sensor 40 is exposed.

[0129] The propagation operator has the function of describing the propagation of light between the image sensor 40 and a point with coordinates (x, y, z), located at a distance | z | from the latter. It is then possible to determine the modulus M(x, y, z) and / or the phase φ (x,y,z) of the light wave 14, at this distance | z | , called the reconstruction distance, with: M x y z = abs A x y z φ x y z = arg A x y z

[0130] The operators "abs" and "arg" denote the modulus and the argument respectively. In other words, the complex expression A of the light wave 14 at any point with coordinates (x, y, z) in space is such that: A x y z = M x y z e jφ x y z

[0131] Obtaining a complex image Az of the sample by applying a propagation operator to a hologram is known from the prior art, in particular for particle characterization purposes, as evidenced by the references indicated above and in the section relating to the state of the art.

[0132] By providing on the support 30 a network 32 acting as a spatial filter, the device according to the invention makes it possible to considerably improve the resolution of the reconstruction of the object from the image 41.

[0133] As illustrated in figure 24 , with a detection device according to the prior art ( figure 1) the recording of an online hologram of a light-scattering particle 21 is limited by the interaction between the incident wave 11 and the light-scattering particle 21 on the one hand and, on the other hand, by the distance from the image sensor 40 and its surface (in the xy plane) which records the hologram. The more light the image sensor 40 collects from the particle 21, the more faithful the reconstruction can be to the particle 21 to be analyzed. This is the case for a large image sensor 40 whose distance Z to the particle 21 to be analyzed is small, in order to have the largest possible numerical aperture.

[0134] As illustrated in figure 25, with a detection device according to the invention, the provision of a diffraction grating 32 whose period is close to the incident wavelength λ 11, makes it possible to increase the interaction between the light and the particle 21. For the same recording configuration (distance Z and width of the image sensor 40) the image sensor 40 will collect more light scattered by the particle 21. Indeed, if the particle 21 is arranged between the grating 32 and the image sensor 40, the light diffracted by the grating 32 which will then also be intercepted by the particle 21 will be added to the incident light scattered by the particle 21 and transmitted by the grating 32. The effect is the same if the diffraction grating 32 is located between the particle 21 and the image sensor 40.

[0135] In the optimum case, particle 21 is in contact with the grating (Z2=0), the grating has a period P=λ. Furthermore, the diffraction is at 90° to the incident wave (transversely) which allows particle 21 to collect and interact with a transverse wave.

[0136] For gratings with a period P < λ, the diffraction angles will be smaller and therefore less effective when particle 21 is in contact with the grating. But, in this case, particle 21 can be shifted from the grating to a distance Z2≠0, the optimum of which will be a function of P, λ, the size of particle 21 and Z.

[0137] THE Figures 5A and 5C correspond to the same case as that illustrated on the Figures 3A to 3C . Sample 20 includes a scene 22 showing the same particles 21a-21e as in the Figures 3A to 3CThe wavelength of the source is identical (450 nm) and the detector is identical except that the support 30 has a grating 32. Here, this grating 32 is formed of lines parallel to each other and distributed periodically according to a pitch P = λ = 450 nm. The Figures 5B and 5C show the results of the digital reconstruction of image 41 generated by sensor 40. These results are indisputably better than those of the Figure 3C . In particular, it is possible to identify particles and even distinguish particles 21a and 21b as illustrated by tasks 51a and 51b of the Figure 5B . It should be noted that a still significantly improved result can be obtained by choosing a network whose patterns would be better suited to this scene 22. The identification of an optimal network 32 can be obtained by trying various networks and then comparing the resolutions obtained with each network.

[0138] The paragraphs below describe the effects of the period of the grating patterns on the contrast and resolution of the reconstructed hologram.

[0139] There figure 6 illustrates a support 30 having a network 32 formed of point patterns 31. This network 32 can thus be described as a two-dimensional network. Indeed, the patterns 31 have a first alignment along the x direction and have a second alignment along the y direction. The pitch P of the network is identical along these two directions.

[0140] The observed sample 20 notably presents objects 21a and 21b which appear enlarged in scenes 22a and 22b.

[0141] In this example, the wavelength λ is equal to 450 nm. The critical dimensions CD1, CD2 illustrated in figure 6 , are respectively of the order of 225 and 337 nm.

[0142] THE Figures 7A to 7Dillustrate the results of the reconstruction of the holographic image 41 generated by the sensor as a function of the support 30 used.

[0143] There figure 7A corresponds to the case where the support 30 does not have a network 32. The support thus forms a transparent plate. This embodiment thus corresponds to the solutions of the state of the art. It is clear that it is very difficult to identify the particles and in particular the particle 21a. Indeed, on the digitally reconstructed image 50 the object 51a appears very indistinctly. The contrast C is equal to 0.5. In the equation C appearing on the figures 7A à 7D , L mx and L bg correspond respectively to the maximum received light intensity and the average light intensity. The calculation of the contrast C can be carried out on the following equation C = (L mx -L bg ) / L bg . In this embodiment, the image 50 is a square with a side of 25.58µm.

[0144] There figure 7B corresponds to the case where the support 30 has a network 32 whose pitch P = λ / 2. We see that it is possible to identify all the different particles. In particular, on the image 50 obtained by reconstruction of the hologram the images 51a, 51b of the particles 21a, 21b are very easily identifiable. The contrast C is equal to 4.3.

[0145] There figure 7C corresponds to the case where the support 30 has a network 32 whose pitch P = 2λ. We see that with this network also it is possible to identify the different particles. The contrast C is equal to 1.

[0146] There figure 7D corresponds to the case where the support 30 has a network 32 whose pitch P= λ. The resolution is still clearly superior to that of the figure 7B or of the figure 7C The contrast C is equal to 5.5, which is much higher than that of the embodiments with P = λ / 2 or P = 2λ.

[0147] In these embodiments of the figures 7A à 7D , the size of the images 41 are squares whose side lengths are respectively 25.58 µm, 53.15 µm and 106.3 µm. In the examples described, the same dimensions and the same number of pixels were used for the calculations to describe the object and the hologram. At the sensor 40, the pixel density per line is 925 pixels. The size or pitch of each pixel in the pixel matrix of the sensor 40 are, respectively for the embodiments of the figures 7B à 7D of: 28nm (P = λ / 2), 112.5 nm (P = 2λ) and 56 nm (P = λ).

[0148] It is clear from these figures that a step P equal to or close to λ is very clearly advantageous. A step P greater than 2λ no longer allows a truly satisfactory resolution to be obtained.

[0149] THE figures 8A à 8C illustrate the result obtained with the same sample as that observed on the figures 7A à 7D but with a one-dimensional (1D) grating 32. The grating 32 is formed of patterns 31 defined by parallel lines. Each line is, for example, formed by an opaque line. The period of the grating is equal to λ. In this embodiment, the width of a line, taken in a direction perpendicular to the main direction of extension of the line, is approximately λ / 4. The figures 8A et 8C show that the resolution obtained is very good. The images 51a, 51b of the particles 21a, 21b in particular, are very easily identifiable. The contrast C is equal to 6.8. This contrast is better than that obtained in the embodiment of the 2D network 32 having a pitch P = λ, because the patterns are finer and therefore the support 30 blocks less light, thereby improving the contrast.

[0150] It is clear from these examples that any type of periodic pattern 31 (1D or 2D) can be used to improve the reconstruction of the object from the recording of the hologram 41. A periodic grating 32 with a period close to the illumination wavelength λ of the object 20 will be optimal for the digital reconstruction of the object from the recorded image 51 of the hologram.

[0151] THE figures 9A à 9C illustrate the influence of the filling rate of the diffraction grating 32 on the resolution obtained.

[0152] For each of these embodiments, the observed sample 20 comprises the same scene 22 as that of the figures 3A Or 5A. For each of these embodiments, the network 32 is a 1D network, formed of opaque lines, extending in this non-limiting example along the x axis. The pitch P of each network is identical. Only the thickness of each pattern varies, i.e. the width of each line, taken along the y direction. This variation in thickness thus directly impacts the filling rate of the structure 30. The filling rate refers to the percentage of the surface occupied by the opaque patterns on the total surface of the support 30. Conversely, the transmission rate refers to the percentage of the surface not occupied by the opaque patterns on the total surface of the support 30. If the opaque patterns occupy the entire surface of the support 30, then the transmission rate is zero. Conversely, if the support 30 does not include opaque patterns, then the transmission rate is equal to 100%.

[0153] In the example illustrated in figure 9A , the thickness of each line is such that the transmission rate T = 50%. The contrast C = 0.68. The signal-to-noise ratio SBR = 1.68.

[0154] In the example illustrated in figure 9B , the thickness of each line is such that the transmission rate T = 75%. The contrast C = 1.46. The signal-to-noise ratio SBR = 2.54.

[0155] In the example illustrated in figure 9C , the thickness of each line is such that the transmission rate T = 96%. The contrast C = 6.85. The signal-to-noise ratio SBR = 7.85.

[0156] Il It is clear from these figures that the contrast and the signal-to-noise ratio are all the better when the periodic structure 32 is transparent.

[0157] In the following examples, the periodic diffraction structure of the support 30 comprises several gratings 32. The pitch P of each grating 32 is adapted for a given wavelength. Thus, by illuminating the sample 20 with waves whose lengths are different, as many images 41 are obtained as there are exposure wavelengths. The incident exposure waves having different wavelengths can be emitted simultaneously or successively onto the sample 20.

[0158] For example, it is possible to provide that one grating has a pitch adapted to infrared wavelengths and that another grating has a pitch adapted to ultraviolet wavelengths. This then allows a chemical analysis according to the molecular absorption of the constituent components of the particles 21 of the sample 20. The image of the digitally reconstructed object will have a variation in intensity depending on the absorption of light specific to the chemical composition of the particle 21 analyzed. It is also possible to provide a larger number of gratings 32. It is possible, for example, to provide on the same support 30 at least three gratings, including for example: a 32a grating adapted to the wavelengths of the blue color, the pitch Pa of this grating will be, for example, 450 nm, a 32b grating adapted to the wavelengths of the green color, the pitch Pb of this grating will be, for example, 600 nm, a 32c grating adapted to the wavelengths of the red color, the pitch Pc of this grating will be, for example, 750 nm.

[0159] According to a first embodiment, the support 30 comprises several networks distinct from each other. These networks do not overlap. They occupy distinct areas of the support 20. This embodiment will be detailed with reference to figures 10A à 11C . Another embodiment in which the different networks partially or completely overlap will be detailed with reference to the figures 12A à 12C .

[0160] A first example of support comprising separate networks is illustrated with reference to figures 10A à 10D . The support 30 comprises a first network 32a and a second network 32b, as shown in figure 10E .

[0161] The first grating 32a comprises patterns 31a forming parallel lines, extending vertically (x) and distributed periodically at a pitch P1. The second grating 32b comprises patterns 31b forming parallel lines, extending horizontally (y) and distributed periodically at a pitch P2≠P1. In this example, Pb = 2Pa. The exposure wavelengths of the sample 20 are respectively λa = 450nm and λb= 900nm. Z1 = 1cm. The sample 20 comprises in particular the particle 21a located on or opposite the grating 32a and the particle 21b located on or opposite the grating 32b. These particles 21a, 21b are here rectangles whose dimensions, in projection on a plane parallel to the xy plane, are equal to 1.53µm * 0.27µm.

[0162] By exposing the sample 20 simultaneously or successively with incident waves of wavelengths λa and λb, two images or holograms 41 are obtained. The figures 10A et 10B illustrate the reconstructed object 50 by exposure at wavelength λa. The contrast obtained is 5.9. The figures 10C et 10D illustrate the reconstructed object 50 by exposure at wavelength λb. The contrast obtained is 5.9. We find on the digitally reconstructed images 50, and illustrated on the figures 10A And 10C , zones 32a' and 32b' which correspond to the 2 networks 32a, 32b of the support 30.

[0163] A second example of a support 30 comprising separate networks is illustrated in figures 11A à 11C . The support 30 comprises four networks 32a, 32b, 32c, 32d. Here, each network occupies a quarter of the surface of the support 30.

[0164] The analyzed scene 22 includes particles 21a-21d which are distributed in relation to the networks of the support 30.

[0165] In this non-limiting example, the support 30 comprises two 2D networks and two 1D networks. More precisely, in this example: network 32a is a 2D network, in the form of a matrix, whose pitch Pa = λa = 900nm, network 32b is a 2D network in the form of a matrix, whose pitch Pb = λb= 450nm, network 32c is a 1D network, formed of horizontal lines (y axis) distributed periodically according to a pitch Pc = λc = 450nm, network 32d is a 1D network, formed of vertical lines (x axis) distributed periodically according to a pitch Pd = λd = 450nm.

[0166] In this example, as in the previous examples, the critical dimensions of the sample 20 are between 280 and 336 nanometers. The distance Z1 between the support 32 and the image sensor 40 is equal to 1 cm.

[0167] There figure 11B corresponds to image 50 the digitally reconstructed object when the source 10 emits an incident wave of wavelengths λ = λb= λc = λd = 450nm. The figure 11C corresponds to the digitally reconstructed object when source 10 emits an incident wave of wavelengths λ = λa= 900nm.

[0168] As indicated above, the networks may, according to another embodiment, overlap completely or partially. Such an example will now be described with reference to Figures 12A to 12D.

[0169] In this example, two gratings are fully interlaced. They occupy the same area of the support 30. In this example, these are two point pattern gratings. These two gratings are interlaced. A first grating has a pitch Pa = λa = 450nm and a second grating has a pitch Pb = λb= 900nm. In this example, Z1 = 1cm.

[0170] There figure 12B corresponds to the digitally reconstructed object when the source 10 emits an incident wave of wavelengths λ = λa = 450nm. The figure 12C corresponds to the digitally reconstructed object when source 10 emits an incident wave of wavelength λ = λb= 900nm. Images 51a-51d of particles 21a-21d are very easily identifiable.

[0171] THE figures 13A à 15C illustrate the influence of the size of the pixel matrix of the sensor 40 collecting the exposure wave 14 propagating from the sensor 40. Indeed, the larger the size of the pixel matrix (and therefore of the sensor), the more light and information it captures on the object.

[0172] In these examples, Z1 = 1cm and λ = 450nm. The support 30 is identical to that of the figures 12A à 12C . Thus, the support 30 comprises two interlaced networks 32, formed of point patterns. Typically, the pitches of each of the networks are 450 and 900nm.

[0173] Sample 20 is identical for each of these figures. The figures 13A , 14A And 15Aillustrate the images 41 generated by the sensor 40, i.e. the holograms.

[0174] In each of these examples, the sensor 40 is a square pixel matrix, the length of one side of which is L and having a number of pixels per line equal to Nbpix.

[0175] In the first example, illustrated in figure 13A , L = 112.52µm and Nbpix = 1000. The digitally reconstructed image 51 is illustrated in figure 13B The contrast obtained is equal to 7.6. The light intensity as a function of the position of the pixels is illustrated in figure 13C .

[0176] In the second example, illustrated in figure 14A , L = 56.26µm and Nbpix = 500. The digitally reconstructed image 51 is illustrated in figure 14B The contrast obtained is equal to 7.1. The light intensity as a function of the position of the pixels is illustrated in figure 14C .

[0177] In the third example, illustrated in figure 15A , L = 28.13µm and Nbpix = 300. The digitally reconstructed image 50 is illustrated in figure 15B The contrast obtained is equal to 3.5. The light intensity as a function of the position of the pixels is illustrated in figure 15C .

[0178] Il It is clear from these examples that the object field, respectively image, defined by the set of points of the object crossed by the incident light rays, impacts by its lateral dimension the resolution and the contrast of the image reconstructed from the hologram 41. The larger the sensor 40, the higher the contrast and the signal to noise ratio.

[0179] According to one embodiment, the patterns 31 of the network 32 are heat-conducting elements or heat-producing elements. In the first case, the patterns may be metallic. They may be made entirely or partially of tungsten. Their temperature rises and releases heat when they are in conduction with a heat source. In the second case, it may be provided, for example, that the patterns are resistive elements which generate heat when they are supplied with electricity.

[0180] In these two embodiments, it is then possible to make the network 32 heatable. This makes it possible to destroy certain particles which have an evaporation or melting temperature lower than the temperature generated by the network 32. This is the case for organic particles. The network heating 32 thus makes it possible to eliminate these fusible particles without degrading other particles which are not fusible or are fusible at a higher temperature.

[0181] Typically, this allows organic particles to be removed without degrading metal particles. It is then possible to differentiate these particles by making successive recordings of holograms. Some holograms are recorded without the application of heat by the grating 32, which makes it possible to identify organic particles and potentially other particles such as metal particles. Other holograms are recorded with the application of heat by the grating 32, which makes it possible to identify only non-organic particles.

[0182] Another advantage of heating the diffracted structure is that it is possible to degrade and eliminate particles that would otherwise stick to the grating 32. This thus makes it possible to avoid the accumulation of particles on the grating 32 by cleaning it. This feature thus addresses the problem of extending the lifetime of the device and maintaining its sensitivity over a long period of time.

[0183] Typically, the device is configured so that the network 32 reaches a temperature of several hundred degrees, typically at least 150°C and preferably at least 300°C, preferably at least 500°C and preferably at least 1000°C. If the materials allow it, it is possible to go beyond 1500°C, or even beyond 1700°C, or even to or beyond 2400°C.

[0184] For example, a quartz support with a MoSi2 metal lattice can reach 1800°C in air, without damaging the electrodes. Furthermore, the temperature can reach 1700°C or even 2400°C for refractory metals such as molybdenum or tungsten. Those skilled in the art can use alloys such as molybdenum disilicide MoSi2 or silicon carbide SiC, particularly due to their lower cost.

[0185] The examples described above clearly show that the invention is particularly effective for counting and analyzing very small particles, i.e. particles whose size is less than or of the order of the wavelength.

[0186] Furthermore, the invention also proves to be very useful for carrying out the morphological identification of larger objects. This is the case for pollen or mold. Indeed, with the holographic imaging solutions of the state of the art, it is possible to count objects such as pollen. However, it is not possible to distinguish between different types of pollen. For example, known solutions do not allow ragweed pollen to be distinguished from birch pollen. figures 16A et 16B are photos obtained using a scanning electron microscope (SEM) for each of these pollens respectively. These pollens have roughly the same size: 25 µm. Known solutions do not allow these pollens to be differentiated by holographic imaging. Therefore, in practice, a sample is usually taken, followed by a SEM photograph, and then each of the particles is counted from this photo. This process is naturally long and tedious. Furthermore, it makes it impossible to count these pollens in real time or continuously. It is therefore not possible to analyze a flow potentially containing these pollens.

[0187] On the contrary, the detection device described above makes it possible to digitally reconstruct an image with a resolution good enough to identify the nature of each of these pollens.

[0188] Furthermore, this device makes it possible to identify objects, sometimes of large size, which are difficult to recognize using conventional techniques. This is the case, for example, of conidiophore molds, which are several hundred micrometers in size, an example of which is illustrated in figure 16C .

[0189] There figure 17 illustrates a holographic imaging device, similar to prior art solutions, used to analyze a sample comprising ragweed pollen 21a, birch pollen 21b, and conidiophore mold 21c. In this example, Z1=1mm, λ=450nm. The imager of the sensor 40 is a square with sides measuring 1000µm.

[0190] In this example, the support 30 does not have a diffracted grating. The object 51 digitally reconstructed from the hologram 41 does not allow each of these particles to be recognized. This digitally reconstructed object 51 does not in fact allow a resolution higher than that of the hologram 41 itself to be obtained.

[0191] A completely different result is obtained by analyzing this same sample 20 using a device according to the invention. This device comprises a support 30 having two interlaced networks. Each network has point patterns, for example dots of circular, polygonal, rectangular or square section. The pitch Pa of a first network is equal to 500 nm, measured in a vertical direction (x). The pitch Pb of a second network is equal to 1000 nm, measured in a horizontal direction (y). The support 30 and the particles 21a-21c are illustrated in figure 18A , there figure 18B being an enlargement of the portion 22a of the sample. The pixel matrix of the sensor 40 forms a square with sides of 250 µm and has 1000 pixels per line.

[0192] By exposing the sample with an incident wave of wavelength λa=500nm, a 50a digital reconstruction of the hologram is obtained. This result is illustrated in figure 18C . By exposing the sample with an incident wave of wavelength λb=1000nm, a 50b digital reconstruction of the hologram is obtained. This result is illustrated in figure 18D . On each of these digitally reconstructed images 50a, 50b, the resolution is largely sufficient to recognize the nature of each of these particles 21a-21c.

[0193] This example clearly shows that the device according to the invention proves to be very effective for the identification and morphological analysis of objects whose size is greater than the exposure wavelength.

[0194] The invention also finds particularly advantageous application in the counting and identification of very small particles such as macromolecules like DNA.

[0195] To produce a detector whose diffraction grating 32 is adapted to this particle size, it may be provided that this grating is formed by the atoms of a lattice. Preferably, it is a crystalline lattice, advantageously monocrystalline. This crystalline lattice is for example a lattice of silicon atoms. A crystalline layer of silicon forms a face-centered cubic (FCC) lattice. The period of a CCC lattice of silicon is equal to the length l of a side of the cube and is equal to 3.11 angstroms (3.11 10 -10 < m). Such a lattice is illustrated in figure 19B .

[0196] According to an advantageous embodiment, the network 32 is formed by a thin layer of monocrystalline silicon forming a membrane.

[0197] To form a silicon membrane <100> , we can for example refer to the following publications: US Patent 8,501,026, entitled "Method for Making a Planar Membrane", issued on August 6, 2013, and listing as inventors Christophe Constancias et al. "Fabrication of large area ultrathin silicon membrane: Application for high efficiency extreme ultraviolet diffraction gratings", Constancias et al. J. Vac. Sci. Technol. B 281, Jan / Feb 2010. Fabrication of Buckling Free Ultrathin Silicon Membranes by Direct Bonding with Thermal Difference Florian Delachat, Christophe Constancias, Frank Fournel, Christophe Morales, Boris Le Drogoff, Mohamed Chaker, and Joelle Margot, ACS Nano, Just Accepted Manuscript DOI: 10.1021 / acsnano.5b00234 • Publication Date (Web): 19 Mar 2015.

[0198] There figure 19A represents the support 30 carrying such a network 32. An enlarged portion 22a illustrates in detail the patterns 31 of the network 32 as well as a portion of an object to be analyzed. This object is here a DNA strand 21. In this symbolic representation, the DNA strand has a radius of 2 nanometers and a pitch of 3.4 nanometers.

[0199] The DNA helix 21 is exposed to an incident wave whose wavelength is 3.11 angstroms, that is to say approximately equal to the pitch of the grating 32. The recombination of the incident wave 12 coming from the source 10 and propagating through the structure 30 with the diffracted wave 13 by the DNA helix 21 and also propagating through the structure 30 forms a holographic image 41 on the sensor 40. The result 50 of the digital reconstruction of this hologram 41 is illustrated in figure 19C .

[0200] We can see that, in this image 50, it is possible to precisely analyze the DNA helix 51. In particular, we can measure the radius and the pitch of this helix.

[0201] THE figures 20A à 20C illustrate an embodiment of a device comprising a structured network, particularly well suited to the analysis of small objects extending mainly in a single direction. This is the case, for example, of a macromolecule such as DNA. In this example, the DNA strand forms a helix whose length is equal to 253 nanometers, the radius is equal to 2 nanometers, the pitch is equal to 3.4 nanometers.

[0202] To do this, a network of parallel lines can be structured so that the object to be analyzed is placed between two parallel lines. This allows the position of the object to be precisely controlled. Furthermore, this allows this position to be controlled when handling or transporting the device containing the sample to be analyzed.

[0203] To achieve such a network, one can, for example, use nanometric lithography techniques such as electron beam lithography (Ebeam), lithography by self-organization of block copolymers, a technique usually referred to by the acronym DSA (Directed Self-Assembly). To date, the smallest period for achieving periodic networks in ring lithography is 18 nanometers or even 10 nanometers.

[0204] For networks with slightly larger sizes, we can use classic lithography techniques and, in particular, nanoimprinting techniques which allow us to obtain high resolutions, of the order of a few tens of nanometers.

[0205] In the example illustrated in figure 20A and of which an enlarged portion 22a is illustrated in figure 20B , the diffraction grating 32 of the support 30 has parallel line patterns extending in the y direction.

[0206] In this non-limiting example, the pitch P of the network 32 is 18 nanometers. The support 30 has a square shape with a side length of 280 nanometers. The distance Z1 between the support 30 and the image sensor 40 is 10 millimeters. The image sensor 40 comprises 1000 pixels per line.

[0207] With this device, by emitting an incident wave whose wavelength λ is close to the pitch of the grating 32, i.e. 18 nanometers in this case and with an energy of 124eV, a hologram is obtained on the image sensor 40. The object 50 digitally reconstructed from this hologram is illustrated in figure 20C . In this figure, we can clearly see DNA helix 51. We can easily measure its length. However, in order to precisely measure the radius and pitch of this helix 51, we would need a source capable of emitting a wave with a shorter wavelength to approach the critical dimension, here the radius of DNA helix 51.

[0208] As a non-limiting example, we can use systems whose sources of a few keV are metal anodes bombarded by electrons. Anodes made of transition metals such as Co, Cu, Ga, Mo, Ag produce X-rays of 6 to 22 keV. We can also use synchrotron sources which can produce any type of X-ray.

[0209] There figure 21 illustrates, in a simplified manner, a support 30 comprising a layer 35 of which a portion 36 is sufficiently thin (along the z axis) to allow an incident wave 11 to pass. This portion 36 is structured so as to form a diffraction grating 32a. This grating 32a comprises trenches 33a, separated from each other by raised patterns 31a forming ribs. The trenches 33a and the ribs are arranged alternately and extend parallel to each other in the y direction. The bottom 34 of the trenches 33a can be used to receive an object such as a strand of DNA 21. This support 30 is exposed to an incident wave 11.

[0210] This embodiment allows the object to be confined and held in place during transport handling. Furthermore, the grating 32a has a diffraction function in addition to its function of holding the DNA strand. More specifically in this case, the grating 32a will have a small-angle diffraction function.

[0211] According to a particularly advantageous embodiment, layer 35 is made of monocrystalline silicon. Portion 36 can then form a silicon membrane which forms a periodic network of atoms.

[0212] Thus, the support 30 comprises a network 32b, in this example a CfC network formed from silicon atoms, in addition to the network 32a formed by the structuring of the layer 35. The support 30 thus comprises two networks 32a, 32b superimposed on the propagation axis z of the light.

[0213] This structure thus makes it possible to obtain very good resolution, in particular for critical dimensions of small sizes, thanks to the 32b network formed by the atoms of the layer, and another 32a network which makes it possible to control the position of the object to be analyzed.

[0214] Naturally, this embodiment is applicable to any material other than silicon and which would make it possible to structure the surface of layer 35, and possibly to create a periodic network of atoms in its thickness.

[0215] According to another embodiment, the network 32a can also be placed on a silicon membrane. For example, it can be a silica (silicon oxide) network produced by electron beam lithography (Ebeam) on a fox ® or HSQ hydrogen silsesquioxane resin, a sol-gel silica which transforms into silica after exposure to an electron beam by Ebeam electron beam lithography, development and annealing.

[0216] In the examples illustrated above, the detection device comprises a single support 30. The support 30 possibly comprises several superimposed diffraction gratings. However, these diffraction gratings are in contact with each other or at least they do not allow the passage of particles between the different gratings.

[0217] The paragraphs below describe an embodiment with several supports each comprising at least one diffraction grating, such that the diffraction gratings are spaced from each other, so as to allow the passage of particles between two adjacent gratings.

[0218] As illustrated in figure 23 , it is possible to provide for arranging several supports 30a-30d comprising diffraction gratings 32a-32d along the axis of propagation of the rays. In the illustrated embodiment, of the line transmission type, these different supports 30a-30d, and therefore their gratings 32a-32d, are aligned along the propagation axis z. Thus, each grating is located at a distance Z1 from the image sensor 40. In the non-limiting example illustrated in figure 23 , the detection device comprises four supports 30a-30d, located respectively at distances Z1a, Z1b, Z1c and Z1d from the image sensor 40.

[0219] These different supports 30a-30d are spaced from each other. The distances Dab, Dbc and Dcd are measured along the propagation axis z and are illustrated in figure 23 .

[0220] Furthermore, the detection device is configured so that two networks channel a detection volume that may include particles. Thus, adjacent networks 32 define between them a portion of the detection volume forming a channel 70. The detection device thus has several channels 70. This is particularly advantageous when a flow of fluid transporting particles is channeled by these different channels 70. The arrows F1 illustrate the direction of flow of the fluid transporting the particles 21a-21d.

[0221] The presence of this plurality of networks 32a-32d is thus particularly advantageous, since for a large detection volume, each particle 21a-21c is in the immediate vicinity of one of the networks. The distance Z2 between each particle and one of the networks 32a-32d is thus smaller. It can therefore be approximated more easily.

[0222] The digital reconstruction of the image can then be carried out by varying the Z1 distance taken into account in the calculation. This makes it possible to restore the particles in position and morphology.

[0223] It will be noted that all the characteristics, technical effects and advantages mentioned in correspondence with embodiments comprising a single diffraction grating are perfectly applicable and combinable with embodiments comprising several diffraction gratings in series.

[0224] Naturally, this embodiment with several diffraction gratings can have at least two diffraction gratings, three diffraction gratings, or even four or more diffraction gratings.

[0225] Furthermore, this embodiment with several supports each comprising at least one diffraction grating is not limiting to a type of diffraction grating and any periodic structure can be envisaged. For example, it can be provided that each support has several superimposed, juxtaposed or interlaced diffraction gratings. The diffraction gratings of the same support do not allow particles to pass between them. On the other hand, the particles can pass between the diffraction gratings belonging to two adjacent supports 30.

[0226] In view of the above description, it appears clearly that the proposed device offers a particularly effective solution for improving the counting, identification and analysis of objects of micrometric size but also of objects of nanometric size.

[0227] The invention is not limited to the embodiments previously described and extends to all embodiments covered by the claims.

Claims

1. Device for detecting at least one object (21), the device including: - a light source (10) configured to emit incident waves (11) at a wavelength λ, - a detection volume intended to receive the object (21), and to receive the incident waves (11), - an image sensor (40) positioned to receive at least scattered waves (13) obtained by scattering incident waves (11) on the object (21) and reference waves (12) corresponding to incident waves (11) not scattered on the object (21), the sensor being configured to generate a holographic image (41) based at least on the scattered waves (13) and on the reference waves (12), - a computer data processing device (61) configured to digitally reconstruct the object based at least on the holographic image (41) and the wavelength λ, characterised in that the device also comprises at least one support (30) comprising patterns (31) organised to form at least one periodic diffraction grating (32) having a pitch P, such that λ / 2≤ P ≤2λ, so that the scattered waves (13) collected by the sensor (40) comprise: - waves from the source (10) which were scattered by the object (21) without being diffracted by the at least one diffraction grating (32), and - waves from the source (10) which were both diffracted by the at least one diffraction grating (32) and scattered by the object (21), the reference waves (12) being incident waves (11) having or not been diffracted by the diffraction grating (32), and wherein the computer device (61) is configured to digitally reconstruct the object based on at least the holographic image (41), of wavelength λ and with a distance Z1 between the support (30) and the image sensor (40), the distance Z1 preferably been taken along a propagation axis z of the incident wave (11).

2. Device according to the preceding claim, wherein the grating (32) has a pitch P lying between 0.75*λ and 1.5*λ.

3. Device according to the preceding claim, wherein the grating (32) has a pitch P equal to A.

4. Device according to any one of the preceding claims, wherein the grating (32) is a two-dimensional grating, according to a first dimension of the grating (32) the grating has a pitch Pa and according to a second dimension of the grating (32), taken in a plane perpendicular to the first dimension, the grating has a pitch Pb, with Pa and Pb such that λ / 2≤ Pa ≤2λ and λ / 2≤ Pb ≤2λ.

5. Device according to any one of claims 1 to 3, wherein the support (30) comprises patterns (31a, 31b) organised so as to form at least two periodic diffraction gratings (32a, 32b), including a first grating (32a) having a pitch Pa, and a second grating (32b) having a pitch Pb, with Pa and Pb such that Pa≠Pb, λ / 2≤ Pa ≤2λ and λ / 2≤ Pb ≤2λ.

6. Device according to the preceding claim, wherein the light source (10) is configured to emit at least a first incident wave (11a) and a second incident wave (11b), respectively having a wavelength λa and Ab, such that λa / 2≤ Pa ≤2λa and λb / 2≤ Pb ≤2λb, preferably 0.75*λa / 2≤ Pa ≤1.5*λa and 0.75*λb / 2≤ Pb ≤1.5*λb and preferably λa=Pa and Ab=Pb.

7. Device according to either one of the preceding two claims, wherein the at least two gratings (32a, 32b) are located on distinct areas (30a, 30b) of the support.

8. Device according to any one of the preceding three claims, wherein the at least two gratings (32a, 32b) are at least partly and preferably entirely superimposed in projection on a plane perpendicular to said direction of propagation z or wherein the at least two gratings (32a, 32b) are at least partly and preferably entirely interlaced.

9. Device according to any one of the preceding claims, wherein the patterns (31) of the grating (32) are made of a dielectric material.

10. Device according to any one of claims 1 to 8, wherein the patterns (31) of the grating (32) are made of a material having a heat conduction coefficient greater than or equal to 5 Wm-1.K-1, the patterns (31) of the grating (32) preferably being made of a metallic material such as tungsten, and preferably wherein the patterns (31) of the grating (32) form heating elements when supplied with electricity.

11. Device according to any one of claims 1 to 8, wherein the grating (32) comprises, or is only formed by, at least one crystal lattice, the patterns (32) of the grating being formed, at least in part, by atoms of the crystal lattice, and wherein, preferably, the support (30) comprises a layer having a face structured so as to form a first grating (32a) of relief patterns (31a) and a second grating (32b) formed by said crystal lattice.

12. Device according to any one of the preceding claims, wherein the object has a critical dimension CD, with: - either CD is less than or equal to λ, - or CD is such that CD≥2*λ and preferably CD≥10*λ.

13. Device according to any one of the preceding claims, wherein the support (30) comprises apertures (37) located between the patterns (31).

14. Device according to any one of the preceding claims comprising a plurality of supports (30a-30d) disposed in series, spaced from each other along the propagation axis z and each comprising at least one diffraction grating (32a-32d), so that the at least one object (21) present in the detection volume is located between the gratings (32a-32d) of two adjacent supports (30).

15. System comprising at least one device according to any one of the preceding claims, wherein the system is taken from: - a fire alarm system, - a fire detection system, - a system for analysing the quality of a fluid such as air or water, - an anti-pollution alarm system, - a system for detecting powder of explosives, - a system for detecting microbiological species, - a DNA analysis system, - a pollen analysis system, - a mould analysis system.