Optical measurement system and optical measurement method

Abstract

To provide a light measuring system, an information processing device, a light measuring method and an information processing method with which it is possible to measure a specific location easily with good reproducibility in light wave scattering measurement.SOLUTION: A light measuring system comprises an image forming element that forms an image of a prescribed range including a bright point due to light focused on measurement points of a sample to be measured, an image sensor that displays the image of the prescribed range having been formed by the image forming element so as to display measurement points in a prescribed range of the sample, a first drive unit that moves the sample along each of the optical axis of light and an axis orthogonal to the optical axis and thereby moves the measurement points of the sample, and a second drive unit that moves one or both of the image forming element and the image sensor along the optical axis. By adjusting the first and second drive units, it is possible to focus light upon a specific measurement point in the prescribed range the image of which is being formed in the image sensor.SELECTED DRAWING: Figure 1A

Description

Technical Field

[0001] Embodiments of the present invention relate to an optical measurement system, an information processing apparatus, an optical measurement method, and an information processing method.

Background Art

[0002] Light wave scattering measurement (scatterometry) is a technique that irradiates a fine sample with coherent light (interferable light) and estimates its shape and size from the scattering pattern. As an example of the industrial use of light wave scattering measurement, in the manufacturing process of state-of-the-art semiconductors, an apparatus for measuring fine patterns online can be mentioned. When estimating the original shape from the scattering pattern, the vector field theory that can strictly consider diffraction and polarization is used to calculate the scattering pattern, and the correspondence relationship between the shape and the scattering pattern is stored in a database. Based on this database, the original shape is estimated. The vector field theory directly solves Maxwell's equations by numerical calculation, and rigorous coupled wave analysis, time domain difference method, and boundary element method are known.

[0003] Light wave scattering measurement is widely applied to the analysis of periodic structures, and it is possible to measure a three-dimensional shape with a resolution two orders of magnitude higher than that of imaging measurement (see, for example, Non-Patent Documents 1 and 2). Here, imaging measurement assumes a method of directly obtaining the original image by Fraunhofer approximation. In Fraunhofer approximation, optical simulations of scattering and imaging can be performed by Fourier transform. This includes imaging by Fourier transform using hardware such as refractive index lenses and zone plates, and measurement of obtaining the original image by software Fourier transform from the scattering pattern (see, for example, Non-Patent Document 3).

[0004] In optical wave scattering measurements, the CCD (Charge Coupled Device) image sensor required for measuring scattering patterns must have a high dynamic range of three orders of magnitude or more. The frame rate of this CCD image sensor determines the response speed, but due to recent technological advancements, it is possible to perform measurements with a repeat response speed of up to 50 microseconds, enabling high-speed measurements. The frame rate indicates how many times per second it is possible to acquire all pixels of the CCD image sensor. The research group, including the inventors, has developed a method for applying optical wave scattering measurement not only to periodic structures but also to isolated structures, and has shown that the resolution can be made at the same level as that for periodic structures (see, for example, Non-Patent Documents 2, 4, and 5, and Patent Documents 1 and 2). [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Patent No. 6183826 [Patent Document 2] Patent No. 6783461 [Patent Document 3] Japanese Patent Publication No. 61-213653 [Patent Document 4] Special Publication No. 2009-505101 [Patent Document 5] International Publication No. 2020 / 026704 [Non-patent literature]

[0006] [Non-Patent Document 1] Dexin Kong, Daniel Schmidt, Jennifer Church, Chi-Chun Liu, Mary Breton, Cody Murray, Eric Miller, Luciana Meli, John Sporre, Nelson Felix, Ishtiaq Ahsan, Aron J. Cepler, Marjorie Cheng, Roy Koret, Igor Turovets, “Measuring Local CD Uniformity in EUV vias with scatterometry and machine learning”, Metrology, Inspection, and Process Control for Microlithography XXXIV, Proc. of SPIE Vol.11325, p.113251I. [Non-Patent Document 2] Tetsuya Hoshino, Masahiko Shiono, Banerjee Saswatee, Sadao Aoki, Kenji Sakurai, AND Masahide Itoh, High accuracy cross-sectional shape analysis by coherent soft x-ray diffraction, Vol. 59, No.28 / 1 October 2020 / Applied Optics. [Non-Patent Document 3] Goodman, J. W. (2005). Introduction to Fourier optics. Englewood, Colo: Roberts & Co. [Non-Patent Document 4] Tetsuya Hoshino, Toyohiko Yatagai, and Masahide Itoh, “Precise and rapid distance measurements by scatterometry”, 13 February 2012, Vol.20, No.4 / Optics express 3954. [Non-Patent Document 5] TETSUYA HOSHINO, NORIO WATANABE, SADAO AOKI, KENJI SAKURAI, AND MASAHIDE ITOH, “Cross-sectional particle measurement in the resonance domain on the substrate through scatterometry”, Vol. 25, No. 21, 16 Oct 2017, Optics express 26329. [Non-Patent Document 6] Tetsuya Hoshino, Toyohiko Yatagai, and Masahide Itoh, “Scatterometry of Slant Incidence to Isolated Scatterers for High-Density Memory”, Japanese Journal of Applied Physics, 52(2013), 09LA05. [Non-Patent Document 7] Jorg Bischoff, Joachim Bauer, Ulrich Haak, Lutz Hutschenreuther, and Horst Truckenbrodt, “Optical scatterometry of quarter-micron patterns using neural regression”, 526 / SPIE Vol. 3332. [Non-Patent Document 8] D. Kolenov, D. Davidse, J. Le Cam, AND SF Pereira, “Convolutional neural network applied for nanoparticle classification using coherent scatterometry data”, Vol. 59, No.27 / 20 September 2020 / Applied Optics. [Non-Patent Document 9] Tetsuya Hoshino, Saswatee Banerjee, Sadao Aoki, and Masahide Itoh, “Reflectivity Analysis of Isolated Particle on a Substrate on Incoherent Light by RCWA”, Computational Optical Sensing and Imaging 2021. [Non-Patent Document 10] Tetsuya Hoshino, Masahide Itoh, “Cross-sectional shape evaluation of a particle by scatterometry”, Optics Communications, 359, 240-244. [Non-Patent Document 11] Shintaro Naruka, Tetsuya Hoshino, Masahide Ito, Masami Kobayashi, "High-precision 3D measurement of oil droplets containing phosphors," Proceedings of the 82nd Autumn Meeting of the Japan Society of Applied Physics, 22p-P09-11 (2021) [Non-Patent Document 12] Tetsuya Hoshino, Saswatee Banerjee, Sadao Aoki, AND Masahide Itoh, “Reflection analysis of absorbing film with diffractive structures for incoherent light by rigorous coupled-wave analysis”, Vol.60, No.25 / 1 September 2021 / Applied Optics. [Non-Patent Document 13] Tetsuya Hoshino, Saswatee Banerjee, Norio Watanabe, Sadao Aoki, Kenji Sakurai and Masahide Itoh, “Rigorous analysis of reflection spectrum of absorbing film”, International Conference on Optical Technology and Measurement for Industrial Applications (OPTM2019), Yokohama, 2019.5 [Overview of the project] [Problems that the invention aims to solve]

[0007] Optical scattering measurements have not yet become widely used because the measurement points are difficult to identify, they are not suitable for routine sample analysis, require skilled analysis, and it is difficult to ensure objectivity. The object of the present invention is to provide an optical measurement system, an information processing device, an optical measurement method, and an information processing method that can easily and reproducibly measure a specific location in optical wave scattering measurement. [Means for solving the problem]

[0008] One embodiment of the present invention is an optical measurement system for performing light wave scattering measurements, A first light source, a second light source, an objective lens that focuses the light from the first light source onto a measurement point on a sample to be measured, a display means that measures the scattered light scattered from the sample by the objective lens when the light from the first light source is focused onto the measurement point on the sample, and displays the scattering pattern of the sample, and the light from the second light source is focused onto the objective lens Measure The aforementioned The sample includes bright spots caused by light focused at the aforementioned measurement point. Mu range An imaging element that forms an image of the surrounding area, and the image formed by the imaging element Includes bright spots The aforementioned image of the range Convert into electrical signals Image sensor and From the first light source The system comprises a first drive unit that moves the sample along the optical axis of the light and an axis perpendicular to the optical axis, thereby moving the measurement point of the sample, and a second drive unit that moves either or both of the imaging element and the image sensor along the optical axis. The objective lens creates an image of the range including the bright spot from the light transmitted or reflected by the sample from the light irradiated by the second light source, and the imaging element forms the image of the range including the bright spot created by the objective lens onto the image sensor. The first drive unit and the second drive unit but adjustment SoBy doing so, the light is focused on a specific measurement point within the range imaged on the image sensor. Includes bright spots of the range, and the light but is condensed So , The scattering pattern is the angular distribution and wavelength distribution of the scattering intensity of the sample. This is a light measurement system. As the condensing element, a refractive index lens, an elliptical mirror, a Wolter mirror, a Schwarzschild mirror, etc. can be used according to the wavelength. As the imaging element, a refractive index lens, a zone plate, etc. can be used according to the wavelength. The numerical aperture NA of the refractive index lens of the imaging element is preferably 0.1 or more, more preferably 0.2 or more. A higher resolution can be expected when NA is larger. The working distance, which is an index of the distance between the condensing element or imaging element and the sample, is preferably 1 mm or more, more preferably 2 mm or more. Measuring the sample with a larger distance is easier when the sample is tilted. As the image sensor, a CCD or a CMOS (Complementary Metal Oxide Semiconductor) can be used.

[0011] One embodiment of the present invention is that, in the above-described light measurement system, the first driving unit moves the sample by 0.1 mm or more along each of the optical axis of the light and an axis orthogonal to the optical axis, and the second driving unit moves either one or both of the imaging element and the image sensor by 1 mm or more along the optical axis.

[0012] One embodiment of the present invention is In the aforementioned optical measurement system, a receiving unit that receives the angular distribution of the scattering intensity of the sample, and a processing unit that derives the three-dimensional shape of the sample based on the angular distribution of the scattering intensity of the sample received by the receiving unit. The processing unit has learned the relationship between the angular distribution of the scattering intensity of each of a plurality of samples and the three-dimensional shape of each of the plurality of samples through deep learning, and uses the result of this deep learning to derive the three-dimensional shape of the sample from the angular distribution of the scattering intensity of the sample received by the receiving unit. ru.

[0013] One embodiment of the present invention is the above-described Optical measurement systemIn this configuration, the sample is an isolated light scatterer, and the deep learning is performed with a batch size of 48 to 192, a number of iterations of 150 to 500, a number of parameters in each learning layer of 60 to 200, and a number of learning layers of 2 or more.

[0014] One embodiment of the present invention is as described above. Optical measurement system In this context, the three-dimensional shape of the sample includes three-dimensional shapes where the cross-sectional shape of the sample in a plane containing the optical axis is triangular, quadrilateral, or elliptical.

[0015] One embodiment of the present invention is as described above. Optical measurement system In this case, the three-dimensional shape of the sample is such that the sample surface is uneven. ru .

[0016] One embodiment of the present invention is The system comprises a first light source, a second light source, an objective lens for focusing light from the first light source onto a measurement point on a sample to be measured, a display means for measuring scattered light scattered from the sample by the objective lens when light from the first light source is focused onto the measurement point on the sample, and for displaying the scattering pattern of the sample, an imaging element for forming an image of a range including a bright spot caused by light from the second light source focused onto the measurement point on the sample measured by the objective lens, an image sensor for converting the image of the range including the bright spot formed by the imaging element into an electrical signal, a first drive unit for moving the sample along the optical axis of the light from the first light source and an axis perpendicular to the optical axis, thereby moving the measurement point on the sample, and a second drive unit for moving either or both of the imaging element and the image sensor along the optical axis. An optical measurement method performed by an optical measurement system, The objective lens focuses light from a first light source onto a measurement point on the sample to be measured; the display means measures the scattered light scattered from the sample as the light from the first light source is focused by the objective lens onto the measurement point on the sample, and displays the scattering pattern of the sample. The imaging element is Light from the second light source is reflected by the objective lens. Measure The aforementioned Sample The aforementioned Includes bright spots caused by light focused at the measurement point. Mu range The steps of forming an image of the surrounding area and the image sensor, the image formed by the imaging element Includes bright spots The aforementioned image of the range Convert into electrical signals Steps and From the first light source A first drive unit moves the sample along the optical axis of the light and an axis perpendicular to the optical axis, thereby moving the measurement point of the sample, and a second drive unit moves either the imaging element or the image sensor, or both, along the optical axis, thereby adjusting the first drive unit to move the sample along the optical axis, so that the image formed on the image sensor is Includes bright spots The process includes the step of focusing the light onto a specific measurement point within the range. The objective lens creates an image of the range including the bright spot from the light transmitted or reflected by the sample from the light irradiated by the second light source, and the imaging element forms an image of the range including the bright spot created by the objective lens onto the image sensor, and the scattering pattern is the angular distribution and wavelength distribution of the scattering intensity of the sample. This is a method of measurement using light. [Effects of the Invention]

[0018] According to embodiments of the present invention, it is possible to provide an optical measurement system, an information processing device, an optical measurement method, and an information processing method that can easily and reproducibly measure a specific location in optical wave scattering measurement. [Brief explanation of the drawing]

[0019] [Figure 1A] This figure shows an example of an optical measurement system according to this embodiment. [Figure 1B] This figure shows an example of the operation of the optical measurement system according to this embodiment. [Figure 2A] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 2B] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 2C] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 2D] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 2E] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 3A] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 3B] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 3C] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 3D] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 4A] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 4B] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 4C] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 4D] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 5A] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 5B]This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 5C] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 5D] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 6A] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 6B] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 6C] This figure shows an example of measurement results obtained by the optical measurement system according to this embodiment. [Figure 7] This figure shows a comparative example of measurement results obtained using the optical measurement system according to this embodiment. [Figure 8A] This figure shows an example of an information processing device according to this embodiment. [Figure 8B] This figure shows an example of the operation of the information processing device according to this embodiment. [Figure 9] This figure shows an example of analysis using the optical measurement system according to this embodiment. [Figure 10] This figure shows an example of analysis using the optical measurement system according to this embodiment. [Figure 11] This figure shows an example of analysis using the optical measurement system according to this embodiment. [Figure 12A] This figure shows an example of the analysis results obtained by the optical measurement system according to this embodiment. [Figure 12B] This figure shows an example of the analysis results obtained by the optical measurement system according to this embodiment. [Figure 13] This figure shows an example of the analysis results obtained by the optical measurement system according to this embodiment. [Figure 14] This figure shows an example of the analysis results obtained by the optical measurement system according to this embodiment. [Figure 15] This figure shows an example of the analysis results obtained by the optical measurement system according to this embodiment. [Figure 16]This figure shows another example of the optical measurement system according to this embodiment. [Figure 17] This figure shows an example of a sample measured by the optical measurement system according to this embodiment. [Figure 18] This figure shows an example of the measurement results of a sample using the optical measurement system according to this embodiment. [Figure 19] This figure shows another example of the optical measurement system according to this embodiment. [Figure 20] This figure shows an example of measurement results obtained using the optical measurement system according to this embodiment. [Modes for carrying out the invention]

[0020] Next, the optical measurement system, information processing device, optical measurement method, and information processing method of this embodiment will be described with reference to the drawings. The embodiments described below are merely examples, and the embodiments to which the present invention is applied are not limited to the embodiments described below. In all the figures used to illustrate the embodiments, components with the same function are given the same reference numerals, and repeated explanations are omitted. Furthermore, in this application, "based on XX" means "based on at least XX," and includes cases where it is based on another element in addition to XX. Also, "based on XX" is not limited to cases where XX is used directly, but also includes cases where it is based on something that has been calculated or processed from XX. "XX" is any element (for example, any information).

[0021] (Embodiment) (Measurement of scattering patterns) Optical scattering measurements are primarily used to measure periodic structures or the average structure of scattering by a large number of particles. In methods for measuring periodic structures or the average structure of scattering by a large number of particles, it is not necessary to specify the measurement position. Therefore, to measure complex shapes, it is not necessary to clearly define the measurement position of the light focus. Furthermore, while there are examples of performing scattering pattern measurement and imaging on the same sample, these are not for optical scattering measurements of specific microstructures within the sample (see, for example, Patent Documents 3-5).

[0022] In optical wave scattering measurements, which measure the three-dimensional shape of isolated samples fixed to a substrate, imaging measurements are not used. The first reason is that imaging measurements have a three-dimensional resolution two orders of magnitude lower than optical wave scattering measurements. The second reason is that optical wave scattering measurements use a simple optical system, while imaging measurements use a complex optical system and optical axis adjustment is difficult, so there is little advantage in introducing imaging measurements. Through numerous experiments, the inventors concluded that introducing imaging measurement techniques would broaden the applications of optical wave scattering measurements, which measure the three-dimensional shape of isolated samples fixed to a substrate. In particular, imaging measurement is essential when the sample shape is complex, such as that of living cells or wiring in large-scale integrated circuits (ULSIs), rather than simple periodic structures or loosely dispersed microparticles.

[0023] The objective lens used for imaging measurements can be the same as the objective lens used for focusing light in optical wave scattering measurements. A technical problem with imaging measurements in this case is that, for light scattering in isolated systems (structures), the focusing needs to be nearly parallel. Therefore, when focusing light in optical wave scattering measurements, lenses with a large aperture and long focal length were used. With objective lenses with an aperture of 8 mm or less that are effective for imaging measurements, the focal length is generally short, making it difficult to achieve nearly parallel focusing, and it becomes unclear whether a scattering pattern with high contrast can be obtained.

[0024] In optical wave scattering analysis, a technique for measuring specific locations is known in which the irradiation range of the light focused to measure scattered light is extended to the vicinity of the measurement point, thereby projecting the sample and obtaining an enlarged projection image (see, for example, Non-Patent Document 5 and Patent Document 2). Although this method uses a simple optical system, the image may be distorted, similar to what is seen with a fisheye lens, and the field of view may be narrowed. For this reason, it is necessary to mark the light focus position, for example, with a marker (magic ink), and rely on this to identify the focus position. Furthermore, because the sample is temporarily moved away from the focus position for projection, a three-dimensional positional accuracy of less than 1 μm is required to return it to its original position, making it difficult to maintain the same focus conditions, which poses problems in terms of experimental reproducibility.

[0025] In optical wave scattering analysis, another technique for measuring specific locations involves moving the sample on an XY stage and observing the changes in the angular distribution and intensity of reflected light as the focal point is shifted. Ptychography is a representative method, and it can be implemented with a simple optical system without using beam splitters or lenses on the observation side. While this method has the advantage of being able to handle complex shapes, it is not suitable for high-speed measurements because it involves moving the stage and taking measurements at each point along the stage's movement. Therefore, in the optical measurement system according to this embodiment, the measurement location is identified while the focusing position is fixed to some extent. In the optical measurement system according to this embodiment, the focusing position is determined by imaging, and then the focusing position is moved to the measurement location before performing optical wave scattering measurement. With this configuration, the focusing position for optical wave scattering measurement can be easily determined, and high-speed measurement is possible.

[0026] (Analysis of scattering patterns) In optical scattering measurements, the scattering patterns differ completely between periodic structures and isolated systems (structures). In the former, bright spots appear at diffraction angles determined by the period, so the angular distribution is discrete. On the other hand, in the latter, the diffraction angle deviates from the location determined by the period, and strong angular distributions appear outside of the diffraction angle, so the scattering angular distribution is continuous and broad. Optical scattering measurements are also used to measure the size distribution of numerous microparticles from their scattering patterns. When measuring the size distribution from the scattering patterns of numerous microparticles, the particle shape is assumed to be the same, and the size is calculated from the scattering angular distribution. In optical wave scattering measurements, the shape of a light scatterer was first analyzed from the scattering pattern of an isolated system by the inventors' research group (see, for example, Non-Patent Document 5 and Patent Document 1).

[0027] In the optical measurement system according to this embodiment, deep learning is applied to the scattering pattern of an isolated system to analyze and distinguish shapes such as squares (rectangles), ellipses, semi-ellipses, and triangles. In deep learning, the program is optimized using several parameters to determine the optimal values ​​of the parameters. Furthermore, in this embodiment, it is revealed that surface irregularities are an important element for classification. The optical measurement system according to this embodiment allows anyone to easily and quickly perform optical wave scattering measurements on samples with complex shapes. Below, we will describe measurements of visible and ultraviolet light, and ultrashort ultraviolet and soft X-rays. Furthermore, we will describe a deep learning program that can be commonly used as a method for analyzing the scattering patterns obtained from these measurements.

[0028] (Measurement of visible and ultraviolet light) In visible and ultraviolet light, the focusing range is narrowed to approximately the same wavelength as the incident light. Therefore, even if the focusing range is imaged, information about the focusing position cannot be obtained. A separate imaging light source is used to brightly illuminate the area containing the bright spot of focus, thereby imaging a wider area. It is also preferable to use a different CCD image sensor for scattered light and measurement than for light wave scattering.

[0029] (Measurement of very short ultraviolet and soft X-rays) In optical scattering measurements, the light irradiated onto the sample must maintain coherence. In soft X-rays, one method of focusing incident light that can maintain this state is focusing using an elliptical mirror. The focusing range is about 100 μm, and when the wavelength of the incident light is 10 nm, it becomes 10,000 times the wavelength. Therefore, by imaging the focused range, some information related to the measurement range can be obtained. No other light source is required, and the CCD image sensor used for measurement can be the same as that used for optical scattering measurements. Examples of focusing elements for light scattering include elliptical mirrors, and examples of focusing elements for imaging include zone plates.

[0030] (Analysis of scattering patterns) There are two types of scattering patterns: the intensity distribution with respect to the wavelength of incident light when the measurement angle of scattered light is constant, and the intensity distribution with respect to the scattering angle when the wavelength is constant. Both can be analyzed (see, for example, Non-Patent Document 6 and Patent Document 1). In this embodiment, the intensity distribution with respect to the scattering angle is analyzed. To estimate the shape and size using deep learning, the program parameters are set appropriately. We investigated whether the experimental results could be appropriately reproduced by changing these parameters and training the program.

[0031] When measuring biological samples and rectangular grids, conventional methods make it difficult to determine where light is being focused when the sample shape is complex or when the particles are isolated. After various investigations, it was found that combining imaging and optical scattering measurement allows for the accurate acquisition of the scattering pattern at a specific location. In measurements using visible and ultraviolet light, one important point is to separate the light source for focusing the optical scattering measurement from the light source for imaging. Another important point is to measure focusing and imaging using the same CCD image sensor while maintaining the same optical configuration. By configuring it in this way, the position of the focusing point in the microstructure of the sample can be determined.

[0032] One new finding from this study is that even if the objective lens for focusing light for optical wave scattering measurements and the objective lens for imaging measurements are the same, both can be focused. Since the focusing of light for isolated optical wave scattering measurements and the focusing of light for imaging are not the same, it is not obvious that they can be used together, but rather it is natural to think that there is a difference. Another new finding from this study is that if the measurement points are different within the sample plane, the objective lens needs to be moved along the optical axis, and since the depth of focus cannot be secured even in a narrow field of view, the imaging lens or imaging CCD needs to be moved along the optical axis. In conventional microscopes, measurement is possible within the field of view by adjusting only the distance between the sample and the objective lens. Therefore, in the optical measurement system according to this embodiment, a new way of thinking about optical axis adjustment is required.

[0033] In conventional techniques, imaging and optical scattering measurements are performed to clearly define the measurement target. However, the purpose of performing imaging and optical scattering measurements in conventional techniques is to identify what the flowing sample is, as in flow cytometry (see, for example, Patent Documents 2, 3, and 4), and not to perform imaging measurements to align the focus position of the optical scattering measurement with a specific location on the sample, as in this embodiment. To align the focusing position with the scattering pattern at a specific location on the sample, it is only possible by utilizing the point where the contrast of the scattering pattern is maximized. In conventional optical wave scattering measurements, the focusing position can be aligned with a specific location on the sample by adjusting the focusing range to project a wide area and then gradually bringing the focusing point to the target location (see, for example, Patent Document 2).

[0034] Figure 1A shows an example of the optical measurement system according to this embodiment. The optical measurement system 10 according to this embodiment irradiates light onto a sample SA placed on a support base 101 to acquire information about the sample SA. The optical measurement system 10 comprises an optical measurement device 50, an information processing device 100, and a display 150. (Optical measuring device 50) The optical measuring device 50 comprises a first stage 1, a second stage 2, a third stage 3, a laser light focusing spot shape adjustment means 4, a laser light focusing spot position adjustment means 5, an imaging lens 7, an image sensor 8, a laser light source 12, a light source 14, a light guide LG, a screen SC, and an imaging device 16. In Figure 1, the direction perpendicular to the plane of the paper is defined as the X-axis, the direction vertically upward is defined as the Y-axis, and the direction perpendicular to the X and Y axes is defined as the Z-axis. The direction of the arrows in the figure is positive, and the opposite direction is negative. The laser light source 12, the laser beam focusing spot shape adjustment means 4, and the laser beam focusing spot position adjustment means 5 are arranged in a straight line parallel to the Z-axis. In Figure 1, the laser beam focusing spot shape adjustment means 4 consists of a slit 4a and an iris aperture 4b arranged in that order, starting from the laser light source 12. Also in Figure 1, the laser beam focusing spot position adjustment means 5 consists of a lens 5a, a pinhole 5b, a lens 5c, a polarizing beam splitter 5d, and a lens 5e, all arranged in that order, starting from the laser light source 12. The slit 4a and iris aperture 4b are positioned between the lens 5c and the polarizing beam splitter 5d.

[0035] The imaging lens 7 is fixed on the first stage 1. The position of the first stage 1 is controllable in the Y-axis direction. For example, the first stage 1 can be moved by more than 1 mm along the Y-axis direction by operating the micrometer MMY. The support base 101 is fixed on the second stage 2. The support base 101 can be any shape that allows the sample SA to be placed on it, and the preferred shape varies depending on the form in which the sample SA exists. The second stage 2 is controllable in the X-axis, Y-axis, and Z-axis directions. For example, the second stage 2 can be moved by 1 mm or more along the X-axis, Y-axis, and Z-axis directions, respectively, by operating the micrometers MMX, MMY, and MMZ.

[0036] In conventional methods, the surface of the support base 101 had markers that allowed for the identification of particle positions. Conventionally, the markers were used as guides to correct for shifts in the light collection position when the support base 101 moved. In this embodiment, by forming an image with the image sensor 8 and displaying it on the display 150, the measurement point can be easily identified even without markers. An example of lens 5e is an objective lens, which is fixed on the third stage 3. Lens 5e focuses the laser light from the laser light source 12 onto the measurement point of the sample SA. The third stage 3 is controllable in the Z-axis direction. For example, the third stage 3 can be moved by more than 1 mm along the Z-axis direction by operating the micrometer MMZ. An example of a laser light source 12 is an argon laser. The laser light source 12 is not limited to an argon laser; it may also be synchrotron radiation, X-ray free electron lasers, gas lasers, solid-state lasers, etc. A wide range of wavelengths can be used, including X-rays, ultraviolet / visible rays, infrared rays, and terahertz waves.

[0037] The laser beam focusing spot shape adjustment means 4 has the function of adjusting the shape and size of the focused spot of the laser beam L1 irradiated onto the sample SA using the laser light source 12 so that it is linear and focused only on specific particles. Specifically, a slit 4a, a pupil 4b, etc., are used. The slit 4a is preferably installed between the pupil 4b, which has a variable diameter, and the lens 5c, and is removable. The irradiation range is adjusted by the slit 4a, and the light intensity is adjusted by the pupil 4b. The slit 4a has a rotatable configuration in order to rotate the focused spot. As the pupil 4b, an iris diaphragm or a pinhole of a different diameter can be used, but an iris diaphragm is preferred because its diameter is variable and it does not need to be replaced.

[0038] The laser beam focusing spot position adjustment means 5 has the function of adjusting the focusing spot of the laser beam L1 to match the position of the sample SA. A spatial filter consisting of a lens 5a, a pinhole 5b, and a lens 5c is placed between the laser light source 12 and the laser beam focusing spot shape adjustment means 4 on the optical axis. Light from the laser light source 12 is spread by lens 5a to a width greater than that of the sample S, enters the pinhole 5b, and the light that has passed through the pinhole 5b is parallelized by lens 5c. The light parallelized by lens 5c enters the laser beam focusing spot shape adjustment means 4. By passing the light through the spatial filter, extraneous modes can be removed from the irradiated laser beam L1. For this purpose, the size of the pinhole 5b is preferably about 1 μmφ. A polarizing beam splitter 5d is placed between the laser beam focusing spot shape adjustment means 4 and the lens 5e. The presence of the polarizing beam splitter 5d allows for precise control of the polarization focused on the sample SA, facilitating the calculation of the scattering pattern. Furthermore, the presence of the polarizing beam splitter 5d reduces the distortion of the wavefront of the light focused on the sample SA. In addition, it is preferable that the polarizing beam splitter 5d be rotatable around the optical axis in order to freely change the direction of polarization.

[0039] It is preferable that the first stage 1 to the third stage 3 can be controlled by a computer such as an information processing device 100 so that the position can be feedback-controlled based on the results of the scattering pattern analysis. By configuring it in this way, the range of light irradiated onto the sample SA can be controlled. This makes it easy to compare experimental and calculated values ​​of the scattering pattern. An example of a screen SC is a transmissive type, which uses scattered light from the sample SA to display the scattering pattern of the sample SA. A light-gathering element may be placed between the screen SC and the sample SA to increase the intensity of the measured scattering. The imaging device 16 captures the scattering pattern of the sample SA displayed on the screen SC. An image sensor may be added instead of the screen SC and camera 26. The light source 14 irradiates light onto the sample SA placed on the support base 101. An example of the light source 14 is a halogen lamp. The light source 14 is not limited to a halogen lamp; it may also be synchrotron radiation, a xenon lamp, an infrared lamp, etc., depending on the wavelength range to be measured. The light from the light source 14 is guided to the support base 101 by a light guide LG. Lens 5e creates an image of a predetermined range from the light reflected by the sample SA from the light source 14. An example of the predetermined range of the image is the range that includes the bright spot of light focused at the measurement point of the sample SA. The direction of the light in the predetermined range image created by lens 5e is changed in the -Y direction by the polarizing beam splitter 5d. The imaging lens 7 forms an image of the predetermined range image created by lens 5e onto the image sensor 8. The image sensor 8 converts the image within a predetermined range created by the imaging lens 7 into an electrical signal and extracts it. The image sensor 8 outputs the extracted electrical signal to the information processing device 100.

[0040] [Optical Measurement Method] Figure 1B shows an example of the operation of the optical measurement system according to this embodiment. Referring to Figure 1B, a method for acquiring information about the sample SA as an optical measurement method using the optical measurement system 10 will be described. (Step S1-1) The shape of the focused spot of the laser beam L1 irradiated onto the sample SA is adjusted to be linear by the laser beam focused spot shape adjustment means 4. (Step S2-1) The distance between the light-gathering element, such as lens 5e, and the sample SA, and the distance between the optical axis and the sample SA are adjusted. For example, either or both of the second-stage micrometer MMZ and the third-stage micrometer MMZ are adjusted. (Step S3-1) The laser beam L1 is adjusted to align with the position of the sample SA. For example, the third stage micrometer MMZ is adjusted. (Step S4-1) The support base 101 or the optical axis (X-axis, Y-axis) of the second stage 2 is adjusted so that the laser beam L1 is focused onto the sample SA by moving either or both of the micrometers MMX and MMY, centered around an axis (X-axis, Y-axis) perpendicular to the optical axis (Z-axis) of the laser beam L1.

[0041] (Step S5-1) The scattering pattern of the sample SA is obtained by utilizing the scattered light from the sample SA. For example, the scattering pattern of the sample SA is projected onto a screen SC, and the projected scattering pattern of the sample SA is imaged by the imaging device 16. At least one of the micrometers MMX, MMY, and MMZ of the second stage 2 is adjusted so that the contrast of the scattering pattern of the sample SA is greater than or equal to a predetermined value. By configuring it in this way, information on the shape and size of the sample SA as viewed from each direction can be obtained. (Step S6-1) Light L2 from light source 14 is irradiated onto sample SA via light guide LG. (Step S7-1) Light L2 is adjusted to focus onto the sample SA. (Step 8-1) The Z-direction of the lens 5e and the Y-direction of either or both of the imaging lens 7 and the image sensor 8 are adjusted so that an image within a predetermined range, created from the reflected light of the light source 14 irradiating the sample SA, is formed on the image sensor 8.

[0042] (Example 1 of light wave scattering measurement) Figures 2A to 2E show examples of measurement results obtained by the optical measurement system according to this embodiment. Here, the scattering pattern of rectangular rhodamine B on a glass slide was measured. Figure 2A shows an image from an optical microscope. Figure 2B shows an oblique 52° image from a Scanning Electron Microscope (SEM). Figure 2C shows the incident angle θi and the scattering angle θd. Figure 2D shows an image from an image sensor. Figure 2E shows an example of the scattering pattern of an argon laser with a wavelength of 488 nm. A wide-area image of the sample SA is measured using an imaging lens 7 that has been pre-focused on the substrate sub-surface. A specific position on the sample SA is designated as the focusing position, and with the imaging lens 7 fixed, the distance between the lens 5e and the sample SA is finely adjusted using either or both of the micrometers MMZ of the second stage 2 and the micrometer MMZ of the third stage 3 to maximize the contrast of the scattering pattern. Furthermore, either or both of the imaging lens 7 and the image sensor 8 are finely adjusted in the optical axis direction using the micrometer MMY of the first stage 1.

[0043] Lens 5e is an OLYMPUS PLAN N (X10 0.25 CY ∞ FN22), and imaging lens 7 is a Mitutoyo MT-2. Image sensor 8 is a JAI 2 / 3-inch CCDCV-M4+CL. The first stage 1, second stage 2, and third stage 3 move the positions of either or both of the imaging lens 7 and image sensor 8, the sample SA, and lens 5e, respectively. Slit 4a can reduce (shield) the laser background from the laser light source 12. By setting the diameter of the iris aperture 4b to 1 mmφ or less, the parallelism of the incident light can be ensured when focused by lens 5e. The imaging device 16 captures the scattering pattern projected onto the screen SC. The light source 14 illuminates the sample SA through the light guide LG.

[0044] Sample SA was prepared by spin-coating a rhodamine B ethanol solution onto a glass slide, and then using a focused ion beam to create two parallel rectangular holes measuring 25 μm on each side. The protrusion between the two rectangular holes, approximately 5 μm wide, was used as the measurement grid (Figures 2A to 2C). By determining the focusing position, the scattering pattern could be measured while maintaining the same focusing conditions (Figures 2D, 2E). The screen is NBS Ricoh TP PAPER 901221, A4 size. The imaging device 16 is an Olympus STYLUS XZ-10, and the image was taken in AI mode. In Figure 1, the sample SA is illuminated with incident light from the light source 14 from the side opposite the lens 5e. However, for example, a beam splitter may be used to incident light from the light source 14 onto the sample SA from the side of the lens 5e. A polarizer or phase difference film can be used as appropriate to control the polarization of the laser. In this case as well, light from the light source 14 can be incident onto the sample SA from the side of the lens 5e.

[0045] (Example 2 of light wave scattering measurement) Figures 3A to 3C show examples of measurement results obtained using the optical measurement system according to this embodiment. Here, we show the case where a rat slide specimen is used as the sample. Figure 3A shows an example of a photograph of a rat brain slide specimen, with the laser light source 12 for optical wave scattering measurement turned on and off. Figure 3B shows an example of a transmitted image of a rat brain on the image sensor 8, with the laser light source 12 for optical wave scattering measurement turned on and off. Figure 3C shows an example of the scattering pattern of argon laser light with a wavelength of 488 nm by a rat brain. Figure 3D shows an example of analysis of the angular distribution and cross-sectional shape of the scattering pattern shown in Figure 3C. In Figure 3D, θd is the scattering angle. As shown in Figures 3A to 3C, a scattering pattern with good contrast and a focused image were obtained. From the scattering pattern in Figure 3C, the angular distribution shown in Figure 3D can be obtained, and the cross-sectional shape and size can be calculated from the angular distribution. The method for calculating the cross-sectional shape will be described later. The width in the direction perpendicular to the incident light can be estimated roughly from the spacing of the peaks, which are arranged at approximately equal intervals as indicated by the arrows, without using the exact coupled wave analysis described later (see, for example, Non-Patent Document 10). Specifically, if the average angular spacing is θ and the wavelength is λ, it can be calculated by λ / sin(θ).

[0046] (Example 3 of light wave scattering measurement) Using a sample similar to that used in Example 2 of optical wave scattering measurement, an image of a wide area of ​​the sample SA is measured using an imaging lens 7 that has been pre-focused on the substrate sub-surface. A specific position on the sample SA is designated as the focusing position, and with the imaging lens 7 fixed, the distance between lens 5e and the sample SA is finely adjusted to maximize the contrast of the scattering pattern. Furthermore, either or both of the imaging lens 7 and the image sensor 8 are finely adjusted in the optical axis direction. As a result, the image became significantly blurred.

[0047] (Example 4 of light wave scattering measurement) Figures 4A to 4D show examples of measurement results obtained using the optical measurement system according to this embodiment. Here, we show the case where a slide specimen of a cat's cerebrum was used as the sample. Figure 4A shows an example of a slide specimen of a cat's cerebrum, including laser light indicating measurement points. In Figure 4A, the sample SA is Golgi-stained to identify Purkinje cells. Figure 4B shows the transmission image of the cat's cerebrum on the image sensor 8 and the measurement points. This is shown with the laser light source 12 turned on and off. Figure 4C shows an example of the scattering pattern of the cat's cerebrum. Figure 4D shows an example of analysis of the angular distribution and cross-sectional shape of the scattering pattern shown in Figure 4C. In Figure 4D, θd is the scattering angle. In Figure 4A, sample SA is Golgi-stained to identify Purkinje cells. As shown in Figure 4B, although the pattern is complex, the focal point was aligned to a specific location while observing the image. As shown in Figure 4C, the scattering pattern necessary for analyzing the optical scattering of a specific organ was observed. As shown in Figure 4D, the shape and size can be calculated by analyzing the angular distribution of the scattering pattern.

[0048] (Example 5 of light wave scattering measurement) Figures 5A to 5D show examples of measurement results obtained using the optical measurement system according to this embodiment. Here, we show the case where a human cerebellum slide specimen is used as the sample. Figure 5A shows an example of a slide specimen of human cerebellum myelin sheath stained with Weigert, and includes laser light indicating the measurement points. Figure 5B shows a transmitted image of human cerebellum on image sensor 8 and the measurement points. This is shown with the laser light source 12 turned on and off. Figure 5C shows an example of the scattering pattern of human cerebellum. Figure 5D shows an example of analysis of the angular distribution and cross-sectional shape of the scattering pattern shown in Figure 5C. In Figure 5D, θd is the scattering angle. As shown in Figure 5B, although the pattern was complex, the focal point was aligned to a specific location while observing the image. As shown in Figure 5C, we were able to observe the scattering pattern necessary for analyzing the optical wave scattering of a specific organ. As shown in Figure 5D, the shape and size can be calculated by analyzing the angular distribution of the scattering pattern.

[0049] (Example 6 of light wave scattering measurement) Figures 6A to 6C show examples of measurement results obtained by the optical measurement system according to this embodiment. Here, the case where a resist grid is prepared on a silicon substrate at 20 μm intervals is shown. Figure 6A shows a schematic of the coherent diffraction optical system for ultrashort ultraviolet and soft X-ray experiments, and an example of the sample and imaging element drive unit. Figure 6B shows a dark-field image of an image sensor using a zone plate and its magnified view. The magnified view also shows the intensity projection image. Four gratings are visible at the locations indicated by the arrows. The design value of the grating width is 0.75 μm, and the wavelength is 13.5 nm. Figure 6C shows the angular distribution of diffraction. In Figure 6C, θi is the angle of incidence, which in this example is 43.7°.

[0050] As shown in Figure 6A, after focusing and imaging with a zone plate ZP and moving the sample SA to the desired measurement position, the zone plate ZP is removed and the scattering pattern is measured with an image sensor IS. Both the zone plate ZP and the sample SA were moved on an XYZ stage. The XYZ stage is an XY stage with a linear guide unit fixed to it. This was coupled to a hexagonal tube and moved using micrometers attached to each stage. The zone plate (ZP) is manufactured by Applied Nanotools INC. and consists of a 200nm thick gold layer formed on a 50nm silicon nitride film. The diameter of this device is 750μm, and the outer zone width is 200nm. The focal length is 16.2mm at a wavelength of 10nm. The beam stopper BS is a wire with a diameter of 0.5 mm and is placed at the position where the primary diffracted light of the sample is focused by the zone plate ZP.

[0051] Synchrotron radiation was focused using an elliptical mirror (EM) and incident on the silicon substrate at an angle of 43.7° to the perpendicular. The aperture limit was a 1 mm diameter pinhole (PH). The longer side of the resist grid was positioned perpendicular to the plane formed by the incident and reflected light. The cross-sectional shape of this grid when cut through this plane was measured. At the synchrotron radiation facility, measurements were taken by switching wavelengths using a spectroscopic diffraction grating located upstream of the elliptical mirror. Using light with a wavelength of 13.5 nm, the light was imaged onto the image sensor IS using a zone plate ZP, and it was confirmed that light was hitting at least two gratings (Figure 6B). Furthermore, the focusing element (zone plate ZP) was moved off the optical axis on the stage, and the scattering pattern was measured on the image sensor IS using light with a wavelength of 10 nm (Figure 6C).

[0052] (Comparative example of light wave scattering measurement) (Comparative Example 1) Figure 7 shows a comparative example of measurement results obtained using the optical measurement system according to this embodiment. In this embodiment, the optical system omits the imaging lens 7. To determine the focal point, a simple method is to shift the focal point along the optical axis and project an enlarged image. However, this method distorts the image, narrows the measurement range, and requires marking the sample slide with a marker to determine the measurement position. Figure 7 shows an example of a projection image of particles using laser light (see, for example, Patent Document 2). Rhodamine B particles are shown as an example of particles. In the projection image of particles using laser light in Figure 7, the image is distorted compared to Figure 2D, and the position of the focal point is not clear. In the projection image of particles using laser light in Figure 7, the measurement point is not visible.

[0053] (Comparative Example 2) In this embodiment, the optical system omits the imaging lens 7. We considered observing from an oblique angle using a fiberscope with a thin tip that is less likely to interfere with the scattering pattern (USB fiberscope 3 in 1 autofocus endoscope camera 5 million pixel USB endoscope Anykit AKNTC140AF), but it was difficult to focus on the sample and it interfered with the measurement of the scattering pattern.

[0054] (Comparative Example 3) In this embodiment, the optical system omits the imaging lens 7. We attempted observation using a USB-powered handheld digital microscope (made of Per Pet) which is made of plastic and easy to install, but focusing from an oblique angle was difficult, which interfered with the measurement of the scattering pattern.

[0055] (Comparative Example 4) In this embodiment, the optical system omits the imaging lens 7. Observation was performed from an oblique angle using a small optical microscope with the objective lens mounted horizontally, but focusing each time was time-consuming and cumbersome, and scattering pattern measurements could not be performed simultaneously.

[0056] (Comparative Example 5) In this embodiment, the optical system omits the light source 14 and the light guide LG. We investigated a method of widening the focusing range of the laser light and then imaging the reflected light, but it was difficult to pinpoint the focusing position, and the reproducibility of the scattering pattern and operability also deteriorated.

[0057] (Comparative Example 6) In this embodiment, the optical system omits the imaging lens 7. The light source 14 is moved to the position of the laser light source 12, and the light source 14 and the laser light source 12 are used by switching between them. In the area of ​​the screen SC, the screen SC is placed when the laser light source 12 is used, and the image sensor is placed when the light source 14 is used. An example of a pinhole diameter is 10 μm. The laser light from the focusing laser light source 12 and the light from the light source 14, which covers the area including the bright spot created by the focusing of the laser light, were alternately irradiated through the same lens 5e, and the white light image and the projected laser image were observed with an image sensor located on the opposite side of the objective lens SA across the sample SA. However, due to the large image distortion and narrow field of view, it was difficult to pinpoint the focusing position.

[0058] (Comparative Example 7) The optical system is shown in Figure 6A, with the zone plate omitted. Soft X-rays with a wavelength of 10 nm were shone onto the resist grating, and the scattering pattern was measured. The gratings are spaced at 20 μm intervals. Under conditions where there was no imaging data of the grating, it was not possible to determine how many grating sections the light was focused onto, or how the focused light was distributed on the grating.

[0059] In this embodiment, the case in which the image sensor used for optical wave scattering measurement and the image sensor used for imaging measurement are different has been described, but the embodiment is not limited to this example. The image sensor used for optical wave scattering measurement and the image sensor used for imaging measurement may be the same or different. When the image sensor used for optical wave scattering measurement and the image sensor used for imaging measurement are the same, imaging measurement is performed when there is an imaging element (when imaging is performed with the imaging lens 7), and optical wave scattering measurement is performed when there is no imaging element. For this reason, the imaging element is driven with good reproducibility and switched between being placed on the optical axis when performing imaging measurement and being removed from the optical axis when performing optical wave scattering measurement. When the image sensor used for optical wave scattering measurement and the image sensor used for imaging measurement are different, the imaging element may remain on the optical axis.

[0060] In this embodiment, an image sensor may be used instead of the screen SC. In this case, whether the image sensor used for optical wave scattering measurement and the image sensor 8 used for imaging measurement are the same or different, it is important to measure both the scattering pattern and the image without changing the focusing state in the optical wave scattering measurement. This is because adjusting the focusing position in optical wave scattering measurement is difficult, and once it is moved, it is difficult to return it to its original position with an accuracy of about 1 μm. For this reason, one of the imaging elements is precisely moved and adjusted to a position on the optical axis suitable for imaging. Furthermore, the image must be such that the focusing position can be determined.

[0061] The optical system for optical wave scattering measurement includes a focusing element such as a lens 5e for focusing incident light at the measurement point of the sample SA, and a screen SC for measuring scattered light. An image sensor (not shown) may be used instead of the screen SC. When using the screen SC and projecting onto the screen SC, the scattering pattern is captured and analyzed by a camera (lens and image sensor) such as a separately prepared imaging device 16. Furthermore, a focusing element such as an imaging lens 7 and an image sensor 8 are necessary to obtain an image of the same or wider area as the bright spot, including the position of the focused spot. When obtaining an image of reflection, a focusing element such as a lens 5e may be combined with a focusing element such as an imaging lens 7 on the same optical axis. It is preferable that the focusing element such as a lens 5e can maintain the coherence of the incident light when focusing.

[0062] To identify the measurement point of the sample SA, a sufficiently wide imaging range is necessary. For sensitive optical scattering measurements, it is preferable that the full width at half maximum of the light intensity distribution of the focused bright spot is less than 10,000 times the short axis of the light scatterer being measured. When imaging the same area as the bright spot, the incident light for optical scattering measurement and the incident light for imaging may be the same, but it is preferable that the size is 1000 times the short axis of the light scatterer or 1000 times the central wavelength of the incident light for optical scattering measurement. This is because if the size is not large enough, it becomes difficult to see the surrounding area and determine the measurement position.

[0063] The optical measurement system 10 according to this embodiment uses lens 5e for both light wave scattering measurement and imaging measurement. When used for light wave scattering measurement, it is used as a light-gathering element. Lens 5e is a powerful means of focusing light while maintaining coherence in light wave scattering measurement using visible and ultraviolet light. By using lens 5e in common for both light wave scattering measurement and imaging measurement, the effort of focusing can be reduced and the optical system can be simplified. Lens 5e can reduce the diameter of the incident light focus. On the other hand, since it becomes difficult to determine the position of the focus point with lens 5e, it is preferable to prepare another light source to widen the illumination range and form an image over a wide area.

[0064] In this embodiment, the case in which light incident from the light source 14 is irradiated onto the sample SA from the opposite side of the lens 5e has been described, but the embodiment is not limited to this example. For example, a beam splitter may be used to irradiate the sample SA with light from the light source 14 from the side of the lens 5e. To control the polarization of the laser, a polarizer or phase difference film may be used as appropriate to irradiate the sample SA with light from the light source 14 from the side of the lens 5e.

[0065] The optical measurement system 10 according to this embodiment includes a first drive unit that can move either or both of the imaging lens 7 and the image sensor 8, which are used to image light containing image information of a sample SA that has passed through the lens 5e, along the optical axis by 1 mm or more, independently of the lens 5e and the sample SA. For example, the first drive unit may move the micrometer MMY of the first stage. The position of the first drive unit may be feedback controlled by moving the micrometer MMY of the first stage based on the imaging result. The optical measurement system 10 according to this embodiment includes a second drive unit that can move the relative position between the lens 5e and the sample SA by 0.1 mm or more in each of the three axes: the X, Y, and Z axes. For example, the second drive unit may move the second stage micrometers MMX, MMY, and MMZ. The second drive unit may provide feedback control of the position by moving the second stage micrometers MMX, MMY, and MMZ based on either or both of the results of the scattering pattern analysis and / or the imaging results. The optical measurement system 10 according to this embodiment includes a third drive unit that can move the lens 5e in the Z-axis direction by 0.1 mm or more. For example, the third drive unit may move the micrometer MMZ of the third stage. The position of the third drive unit may be feedback controlled by moving the micrometer MMZ of the third stage based on either the result of the scattering pattern analysis or the imaging result, or both. Preferably, either or both of the second drive unit and the third drive unit can be moved with an accuracy of 1 μm or better along the optical axis direction, and more preferably with an accuracy of 30 nm or better. By moving with an accuracy of 1 μm, an analyzable scattering pattern can be obtained. By moving with an accuracy of 30 nm, differences in the scattering pattern due to the focusing position can be eliminated, making highly reproducible experiments possible.

[0066] Preferably, either or both of the imaging lens 7 and the image sensor 8 can be moved 10 mm or more along the optical axis independently of the lens 5e and the sample SA, and more preferably 20 mm or more. By moving either or both of the imaging lens 7 and the image sensor 8 10 mm or more along the optical axis independently of the lens 5e and the sample SA, most biological samples can be accommodated. By moving either or both of the imaging lens 7 and the image sensor 8 20 mm or more along the optical axis independently of the lens 5e and the sample SA, samples for which light scattering measurement is possible can be accommodated.

[0067] In a typical optical microscope, the distance between lens 5e and sample SA may be changed, but the position of either or both of the imaging lens 7 and image sensor 8 is never moved along the optical axis. Experimental studies have shown that when optical wave scattering measurements are performed by changing the target of the light scatterer within sample SA, the image becomes blurred, making it preferable to change the position of either or both of the imaging lens 7 and image sensor 8.

[0068] (Information processing device 100) The information processing device 100 calculates the three-dimensional shape of an isolated light scatterer in optical wave scattering measurement by solving an inverse problem regarding the scattering intensity angular distribution and the shape and size of the light scatterer. Deep learning is used to calculate the three-dimensional shape of the isolated light scatterer. For example, a deep learning program is used with a number of iterations between 150 and 500, a batch size between 48 and 192, a number of parameters in each learning layer between 60 and 200, and a number of learning layers of 2 or more. Here, the inverse problem is the problem of calculating the original shape from the scattering pattern. The length of the short side of the calculated original shape is about the same as or smaller than the focusing diameter.

[0069] Important parameters in deep learning programs are the number of training iterations, batch size, number of training parameters, and number of training layers (see, for example, Non-Patent Documents 7 and 8). The number of iterations is the number of times the learning algorithm learns the entire training dataset. If the number of training layers is 3, training on 3 layers is counted as one iteration. The batch size is the number of samples to process before updating the internal model parameters. The parameter count is the number of parameters used in one layer of deep learning. For the first time, it has been revealed that scattering patterns obtained by optical wave scattering measurements of isolated light scatterers can be analyzed using deep learning. Furthermore, it was found that triangular, quadrilateral, semi-elliptical, and elliptical shapes are suitable for this analysis. In addition, four appropriate parameter values ​​were calculated.

[0070] Deep learning associates scattering patterns with the original shape, but what aspects of the scattering pattern are associated varies. Here, the scattering intensity was calculated by supplementing the individual angles of the calculated scattering pattern with experimental data, and the data lengths were made the same. In addition, zero-order light distributed between approximately 0° and 5° was excluded from the deep learning target. Fully connected deep learning is performed by directly coupling the values ​​of each angle with the shape. The number of iterations is preferably between 150 and 500. Generally, a higher number of iterations leads to better accuracy, but in this embodiment, around 150 to 300 yielded the best results. The batch size is preferably between 48 and 192, and more preferably between 64 and 128. An appropriate batch size improves the accuracy rate. The number of parameters is preferably between 60 and 200. An appropriate number of parameters improves the accuracy rate. The number of layers is preferably two or more, and more preferably three or more. A larger total number of layers tends to improve the accuracy of finding the correct answer, but it also increases computation time and requires more memory.

[0071] Figure 8A shows an example of an information processing device according to this embodiment. The information processing device 100 is implemented by a device such as a personal computer, server, smartphone, tablet computer, or industrial computer. The information processing device 100 includes, for example, an input unit 102, a receiving unit 104, a processing unit 106, an output unit 108, and a storage unit 110. The input unit 102 receives information. For example, the input unit 102 may have an operating unit such as a keyboard and a mouse. In this case, the input unit 102 receives information corresponding to the operations performed by the user on the operating unit. As another example, the input unit 102 may receive information from an external device. This external device may be, for example, a portable storage medium. The scattering intensity angular distribution is input to the input unit 102.

[0072] The reception unit 104 acquires the scattering intensity angle distribution input to the input unit 102 and accepts the acquired scattering intensity angle distribution. The processing unit 106 includes a learning model 107. The learning model 107 is obtained by machine learning the relationship between the angular distribution of scattering intensity of the sample SA and information that identifies the three-dimensional shape of the sample SA, based on multiple combinations of the angular distribution of scattering intensity of the sample SA obtained by the optical measurement system 10 measuring the sample SA by light wave scattering, and the three-dimensional shape of the sample SA. Examples of the three-dimensional shape of the sample SA are triangles, squares, semi-ellipses, and ellipses. The learning model 107 may be updated by having a human determine the three-dimensional shape of the sample SA, and by using machine learning to study the relationship between the angular distribution of the scattering intensity of the sample SA and the three-dimensional shape of the sample SA.

[0073] The processing unit 106 acquires the angular distribution of the scattering intensity of the sample SA received by the receiving unit 104. Based on the acquired angular distribution of the scattering intensity of the sample SA and the learning model 107, the processing unit 106 determines the three-dimensional shape of the sample SA. The output unit 108 obtains the result of determining the three-dimensional shape of the sample SA from the processing unit 106 and outputs the obtained result of determining the three-dimensional shape of the sample SA. For example, the output unit 108 may output the result of determining the three-dimensional shape of the sample SA by displaying it on the display 150, or by outputting it as sound. The storage unit 110 is implemented using HDD (Hard Disk Drive), flash memory, RAM (Random Access Memory), ROM (Read Only Memory), etc.

[0074] The receiving unit 104, the processing unit 106, and the output unit 108 are realized, for example, by a hardware processor such as a CPU (Central Processing Unit) executing a computer program (software) stored in the storage unit 110. Furthermore, some or all of these functional components may be implemented by hardware (including circuitry) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), and GPU (Graphics Processing Unit), or by the cooperation of software and hardware. Computer programs may be stored in advance in storage devices such as HDDs or flash memory, or they may be stored in removable storage media such as DVDs or CD-ROMs and installed when the storage media is inserted into a drive device.

[0075] (Operation of information processing device) Figure 8B shows an example of the operation of the information processing device according to this embodiment. An example of the operation of the information processing device according to this embodiment will be described. Here, as an example, in the information processing device 100, the processing unit 106 includes a learning model 107 obtained by machine learning the relationship between the angular distribution of scattering intensity of the sample SA and information that identifies the three-dimensional shape of the sample SA, based on multiple combinations of the angular distribution of scattering intensity of the sample SA obtained by the optical measuring device 50 measuring the sample SA by optical wave scattering, and the three-dimensional shape of the sample SA. (Step S1-2) The scattering intensity angular distribution is input to the input unit 102. (Step S2-2) The reception unit 104 acquires the scattering intensity angle distribution input to the input unit 102 and accepts the acquired scattering intensity angle distribution. (Step S3-2) The processing unit 106 acquires the angular distribution of the scattering intensity of the sample SA received by the receiving unit 104. Based on the acquired angular distribution of the scattering intensity of the sample SA and the learning model 107, the processing unit 106 determines the three-dimensional shape of the sample SA. (Step S4-2) The output unit 108 obtains the determination result of the three-dimensional shape of the sample SA from the processing unit 106 and outputs the obtained determination result of the three-dimensional shape of the sample SA.

[0076] In this embodiment, we have described a case in which the optical measuring device 50 has a learning model 107 obtained by machine learning the relationship between the angular distribution of scattering intensity of the sample SA and information that identifies the three-dimensional shape of the sample SA, based on multiple combinations of the angular distribution of scattering intensity of the sample SA obtained by optical wave scattering measurement of the sample SA and the three-dimensional shape of the sample SA. However, the embodiment is not limited to this example. For example, the optical measuring device 50 is not the only option; a general optical measuring device 50 may also be used to obtain a learning model 107 by machine learning the relationship between the angular distribution of scattering intensity of the sample SA and information that identifies the three-dimensional shape of the sample SA, based on multiple combinations of the angular distribution of scattering intensity of the sample SA obtained by optical wave scattering measurement of the sample SA and the three-dimensional shape of the sample SA. The scattering intensity angular distribution input to the input unit 102 is not limited to the scattering intensity angular distribution of the sample SA obtained by the optical measuring device 50 performing optical wave scattering measurements on the sample SA; an angular distribution of the scattering intensity of the sample SA obtained by a general optical measuring device performing optical wave scattering measurements on the sample SA may also be input.

[0077] In this embodiment, deep learning was used as a method for analyzing the cross-section of the sample SA to be measured in a plane containing the optical axis in an optical measurement system 10 that performs optical wave scattering measurement. Deep learning-based shape discrimination was performed using the following classifications of model shapes for deep learning: triangular, quadrilateral, semi-elliptical, and triangular, quadrilateral, and semi-elliptical shapes with surface irregularities. However, for triangular, quadrilateral, and semi-elliptical shapes with surface irregularities, differences in the height of the irregularities were included in the classification for deep learning-based shape discrimination. For example, if the model shape is rectangular, the irregularities can be designed by placing rectangles with a width of 1 wavelength and a height of 0.2 wavelengths on top of it at intervals of 2 wavelengths. By changing the height of 0.2 wavelengths to 0 wavelengths, 0.4 wavelengths, or 0.6 wavelengths, each can be divided into different categories. In this case, three types are added to the triangular, square, and semi-elliptical shapes, resulting in a total of six categories. The scattering patterns of these categories are learned, and the shape is determined from the scattering patterns obtained in experiments.

[0078] In this embodiment, during imaging measurement, incident light 1 propagates in the order of sample SA, imaging element (imaging lens 7), and image sensor 8, while during optical wave scattering measurement, incident light 2 propagates in the order of light-gathering element (imaging lens 7), sample SA, and image sensor (referred to as image sensor A) positioned at the screen SC. In other words, both the imaging range and the irradiation range of the incident light for optical wave scattering measurement can be measured with the same image sensor.

[0079] Image sensor 8 displays an image, while image sensor A displays a scattering pattern. For image formation, if the light scattering measurement's focusing range is 1000 wavelengths or more, incident light 1 and incident light 2 may be the same. Image sensor 8 and image sensor A may also be the same, but in this case, it is preferable to separate the measurement of image formation and the scattering pattern with a time difference. If the incident light is visible or ultraviolet light, image sensor A may be a combination of screen SC (transmissive screen) and imaging device 16 (camera) instead of an image sensor. Furthermore, it is preferable that the imaging element (imaging lens 7) and lens 5e are refractive index lenses. When the incident light is extremely short ultraviolet or X-ray, it is preferable that the imaging element (imaging lens 7) is a zone plate and the light-gathering element (lens 5e) is an elliptical mirror.

[0080] (Example of scattering pattern analysis) (Analysis example 1) Figures 9 to 11 show an example of analysis using the optical measurement system according to this embodiment. Figure 9 shows the focus and sample. Figure 10 shows rectangular, elliptical, semi-elliptical, and triangular models. Figure 11 shows the diffraction patterns. In Figure 11, (1) are the simulated diffraction patterns of four shapes with a width of 40 wavelengths and an aspect ratio of 0.3, and (2) are the simulated and experimental diffraction patterns of the semi-elliptical shape. As shown in Figure 9, sample SA was prepared by spraying oil droplets onto a glass slide substrate, then adding water, and covering it with a coverslip (see, for example, Non-Patent Document 11). The oil droplets contained a phosphor. The scattering patterns were analyzed. The number of samples was 7.

[0081] Referring to Figures 9 to 11, we will explain how to distinguish between elliptical, semi-elliptical, square, and triangular cross-sectional shapes in a plane containing the optical axis in order to calculate the size-aspect ratio. As training data, 6072 scattering angle distributions were calculated using exact coupled-wave analysis of isolated systems by varying the width, aspect ratio, extinction coefficient, and polarization of the incident light of the oil droplet. Triangular shapes were defined with an aspect ratio of 0.1 to 0.3, and other shapes with an aspect ratio of 0.1 to 1. This is because characteristic peaks in triangular shapes fall outside the measurement angle range when the aspect ratio is 0.4 or greater, making them difficult to distinguish from other shapes. The definitions of polarization TE and TM are described in Non-Patent Document 5. For each angle for which calculations were performed, samples were extracted from experimental data. The angle distribution of each sample and its corresponding shape were fully connected for deep learning. An example of a programming language is Python, and the framework is PyTorch. A ReLU (Rectified Linear Unit) layer was used as the activation function layer, and a softmax layer was used as the output layer.

[0082] Figures 12 to 14 show examples of analysis results obtained by the optical measurement system according to this embodiment. Figures 12A and 12B show the results of the accuracy evaluation of the test data. Figure 13 shows the results of the shape estimation. Figure 14 shows the number of correct answers as a result of the shape evaluation. As shown in Figures 12 to 14, the number of iterations, batch size, and number of parameters were optimized. The number of hierarchies was set to 3, and the number of parameters was set to 85 in all cases. When the batch size was 64 and the number of iterations was 150, the accuracy (precision), which is determined by how many correct answers were obtained out of 6072 when validated with the same test data as the training data, was 96%. In shape estimation, all 7 oil droplets were determined to be semi-elliptical. The novelty score, which is an indicator of the probability that the shape is different from a semi-ellipse, was less than 0.01 for all of them. A smaller novelty score indicates higher reliability. After setting the conditions for the deep learning program, the time required for learning the correspondence between shape and scattering pattern was 2 minutes and 40 seconds, and the time for shape analysis was 6 seconds, for a total of 2 minutes and 46 seconds, which was a reduction in time.

[0083] (Analysis example 2) Figure 15 shows an example of the analysis results obtained by the optical measurement system according to this embodiment. Figure 15 shows an example of a rectangular, elliptical, semi-elliptical, and triangular model with uneven surfaces. As in Example 6 of the previously mentioned optical wave scattering measurement, a rectangular resist was imaged at a wavelength of 13.5 nm and optical scattering was measured at 10 nm. The scattering pattern was measured with a CCD camera. The angular distribution of diffracted light intensity was analyzed using deep learning to identify whether the shape was rectangular, semi-elliptical, elliptical, or triangular. For the size, the width and height were calculated using the angle of the peaks or troughs in the angular distribution of scattering intensity as an indicator. The same deep learning program and parameters as described above were used.

[0084] A rectangular lattice with a cross-section of 750 nm in width and 188 nm in height was fabricated on a silicon substrate and used as the sample. Synchrotron radiation with a wavelength of 10 nm was incident at an incident angle of 43.7°, and the angular distribution of the scattered light was measured. As shown in Figure 15, we evaluated the experimental results for a total of seven types of rectangles, varying the surface roughness height (d0) to 0 nm, 2 nm, 4 nm, and 6 nm (four types), and adding three more types: triangular, elliptical, and semi-elliptical. For each type, 1024 training data points were calculated and used as training and test data. The evaluation results showed that the rectangle with a surface roughness height of 4 nm had a novelty score of 7.5 × 10⁻⁵, indicating high confidence. After setting the conditions for the deep learning program, the time required for learning the correspondence between shape and scattering pattern was 2 minutes, and the time for shape analysis was 6 seconds, for a total of 2 minutes and 6 seconds, indicating a reduction in analysis time.

[0085] (Analysis example 3) Figure 16 shows another example of the optical measurement system according to this embodiment. By applying two methods, light wave scattering measurement and imaging measurement, to the same sample, the reliability of the absorption spectrum can be improved. In the optical measurement system according to this embodiment, the light source 14, screen SC, and imaging device 16 are replaced with a switching unit 20. The switching unit 20 includes a spectrometer 21, an optical fiber FB attached to the spectrometer 21, a fourth stage 22, an achromatic lens 23a, a lens 23b, and a light source 24. The fourth stage 22 is equipped with a micrometer MMX that moves the tip of the optical fiber FB in the X-axis direction and a micrometer MMY that moves it in the Y-axis direction. With this configuration, the transmission absorption spectrum can be measured for the same location as the part for which the 3D shape measurement was performed. Therefore, the distortion of the absorption spectrum can be interpreted from the 3D shape, which contributes to the reliability of the experimental data.

[0086] As shown in Figure 16, the experimental system further comprises a light source 25, a light guide LG, a lens 26a, an objective lens 26b, a pinhole 26c, an achromatic lens 26d, and a polarizing beam splitter 5f, in addition to the optical measurement system 10 of Figure 1. An example of the diameter of the pinhole 26c is 25 μmφ. An example of the focal length of the achromatic lens 26d is 45 mm. A polarizing beam splitter 5f is placed between the slit 4a and the iris aperture 4b. Light from the light source 25 is guided in the -Y direction by the light guide LG, spread by the lens 26a and the objective lens 26b, and incident on the pinhole 26c. The light that has passed through the pinhole 26c is parallelized by the achromatic lens 26d. The light parallelized by the achromatic lens 26d is incident on the polarizing beam splitter 5f. The polarizing beam splitter 5d changes the direction of the incident light to the +Z direction. Light from a light source 25, such as a halogen lamp, is collected, and the transmission absorption spectrum of the collected area is measured. The spectrometer 21 is an Ocean Optics Maya2000Pro. A light source 24, such as an LED flashlight, is used as the light source for imaging, and is turned on only when measuring the image. The first stage 1, second stage 2, third stage 3, and fourth stage 22 are stages used to move the imaging lens 7, the sample SA, the lens 5e, and the tip of the optical fiber FB, respectively. The diameter of the pinhole 26c and the focal lengths of the objective lens 5e and achromatic lens 26d are set using the magnification formula for the two combined lenses so that the light-gathering diameter is between 5 μm and 20 μm. The optical axis of lens 5e is aligned by collimating white light from a light source 25 such as a halogen lamp, and then adjusting the distance between the objective lens 5e and the sample SA while viewing the image from the image sensor 8 so that the light-gathering spot on the sample SA is minimized. Furthermore, the position of either the imaging lens 7 or the image sensor 8, or both, is adjusted so that the sample image is in focus.

[0087] Figure 17 shows an example of a sample measured by the optical measurement system according to this embodiment. Figure 17 shows an example of measurement points for the absorption spectrum. Sample SA is a rhodamine B film of dye spin-coated onto a glass slide, with two 25 μm square holes made using a focused ion beam, and a rectangular grid created by spacing the holes 5 μm apart (see, for example, Non-Patent Document 12). Figure 17 shows the imaging result observed by the image sensor 8 regarding the positional relationship between the focus point and the 25 μm square holes.

[0088] Figure 18 shows an example of the measurement results of a sample using the optical measurement system according to this embodiment. Figure 18 shows the absorption spectra of each measurement point. In Figure 18, the horizontal axis represents wavelength (nm), and the vertical axis represents transmittance (%). As shown in Figure 18, there is a large difference in the absorption maximum wavelength of the absorption spectrum between the lattice portion and the flat film portion. This difference is thought to be influenced by the three-dimensional shape of the rectangular lattice. Accurate information on the cross-sectional shape of the rectangular lattice is necessary for this interpretation (see, for example, Non-Patent Document 9).

[0089] (Comparative Example 1) As shown in Analysis Example 1, sample SA was prepared by spraying oil droplets onto a glass slide substrate, then adding water, and covering it with a coverslip. The scattering patterns were then analyzed. They were classified into four types: square, triangular, elliptical, and semi-elliptical. After calculating the scattering patterns, they were assigned to a classification by human judgment. All seven samples were determined to be elliptical. It took an expert 10 minutes to make the judgment. For a novice, it took 1 hour just for the expert's explanation.

[0090] (Comparative Example 2) Measurements similar to those in Analysis Example 2 were performed, and the analysis was conducted using the same deep learning program. For a resist grating with a rectangular cross-section, imaging was performed at a wavelength of 13.5 nm, and light scattering measurements were taken at 10 nm. The scattering pattern was analyzed using deep learning, and the parameters were optimized. The scattering pattern was measured using imaging device 16. The angular distribution of diffracted light intensity was analyzed using deep learning to identify whether it was rectangular, semi-elliptical, elliptical, or triangular. For the size, the width and height were calculated using the angle of the peaks or troughs in the angular distribution of scattering intensity as an indicator. A rectangular grating with a cross-section of 750 nm width and 188 nm height was created on a silicon substrate and used as the sample. Synchrotron radiation with a wavelength of 10 nm was incident at an incident angle of 43.7°, and the angular distribution of the scattered light was measured. Three types were added to the rectangular shape: triangular, elliptical, and semi-elliptical, for a total of four types, and the experimental results were evaluated using deep learning. Surface irregularities of the rectangular shape were not considered. For each type, 1024 training data points were calculated. The evaluation results showed that the rectangle novelty score was greater than 1, indicating that it was not rectangular in shape.

[0091] (Comparative Example 3) Figure 19 shows another example of the optical measurement system according to this embodiment. Figure 19 shows an optical measurement system that measures arrangement by projecting white light. As shown in Figure 19, in analysis example 3, the optical system has an image sensor 8 placed in the switching unit 20, and the imaging lens 7, polarizing beam splitter 5d, and first stage 1 are omitted. To determine the focal point, a simple method is to shift the focal point along the optical axis and project an enlarged image; however, this method distorts the image, narrows the measurement range, and makes it difficult to determine the accurate measurement point.

[0092] Figure 20 shows an example of measurement results obtained by the optical measurement system according to this embodiment. As an example of measurement results, Figure 20 shows two 25 μm square holes and a rectangular grid at their center, projected onto them using white light (see, for example, Non-Patent Document 13). The method for creating the grid was explained in Analysis Example 3, so the explanation is omitted here. Compared to Figure 17, the image in Figure 20 is distorted, and the position of the focal point is not clear. Therefore, the measurement point is unknown. (An example of the best mode for carrying out the invention)

[0093] According to the optical measurement system 10 of this embodiment, the optical measurement system 10 performs light wave scattering measurement. The optical measurement system 10 includes an imaging element (imaging lens 7) that forms an image of a predetermined range including a bright spot due to light focused on the measurement point of the sample SA to be measured; an image sensor 8 that displays the measurement point of the sample SA by displaying the image of the predetermined range formed by the imaging element; a first drive unit that moves the measurement point of the sample SA by moving the sample SA along the optical axis and an axis perpendicular to the optical axis; and a second drive unit that moves either or both of the imaging element and the image sensor 8 along the optical axis. By adjusting the first drive unit and the second drive unit, light can be focused on a specific measurement point in a predetermined range that is imaged on the image sensor 8.

[0094] With this configuration, the measurement point (measurement position) within the sample SA can be determined by imaging measurement. The optical measurement system 10 includes a first drive unit for moving the measurement point and a second drive unit for moving either or both of the imaging element and the image sensor 8. Therefore, based on the obtained image (image within a predetermined range), light can be focused to a specific measurement point within the same sample. For this reason, in optical wave scattering measurement, a specific location can be measured easily and reproducibly. A micrometer, stepping motor, or piezo motor can be used as the driving means for each drive unit. With a micrometer, an accuracy of 1 μm is possible, and with a piezo motor, an accuracy of 30 nm is possible. An example of a piezo motor is the Picomotor manufactured by NEWPORT, which can be automatically controlled from the information processing device 100. Furthermore, for the automatically controlled motor, a closed-loop motor with quantitative measurement of travel distance is preferable to an open-loop motor with no quantitative measurement of travel distance.

[0095] The optical measurement system 10 includes a first light source as a laser light source 12, a second light source as a light source 14, and an objective lens 5e that focuses the light from the first light source onto the measurement point of the sample SA. The objective lens creates an image within a predetermined range from the transmitted or reflected light from the sample SA irradiated by the second light source. The imaging element forms an image of the predetermined range created by the objective lens.

[0096] By configuring it in this way, the objective lens for focusing light and the objective lens for imaging can be the same, and the light source for focusing light and the light source for imaging can be different. By precisely adjusting the position on the optical axis of either one or both of the imaging elements (e.g., imaging lens 7) and the image sensor 8 to focus, one of the imaging elements (e.g., lens 5e) can serve as both the focusing element for imaging measurement and the light wave scattering measurement.

[0097] The optical measurement system 10 includes a display means, which is a screen SC, that displays the scattering pattern of the sample SA by utilizing the scattered light scattered from the sample SA when light from the first light source is focused by the objective lens onto the measurement point of the sample SA. The imaging element forms an image of a predetermined range created by the objective lens from the reflected light from the sample SA of the light irradiated by the second light source onto the image sensor 8. By configuring it in this way, the scattered light from the sample SA can be used to display the scattering pattern of the sample SA on the screen SC.

[0098] In the optical measurement system 10, the first drive unit moves the sample SA by 0.1 mm or more along the optical axis and the axis perpendicular to the optical axis. The second drive unit moves either the imaging element or the image sensor 8, or both, by 1 mm or more along the optical axis.

[0099] With this configuration, the sample SA can be moved along both the optical axis and the axis perpendicular to the optical axis, allowing the measurement point (measurement position) within the sample SA to be moved by imaging measurement. Since either or both of the imaging element and the image sensor 8 can be moved along the optical axis, the obtained image (image within a predetermined range) can be focused to a specific measurement point within the same sample.

[0100] The information processing device 100 according to this embodiment includes a receiving unit 104 that receives the angular distribution of the scattering intensity of a sample SA acquired by the optical measurement system 10, and a processing unit 106 that derives the three-dimensional shape of the sample SA based on the angular distribution of the scattering intensity of the sample SA received by the receiving unit 104. The processing unit 106 performs deep learning on the relationship between the angular distribution of the scattering intensity of each of the multiple samples and the three-dimensional shape of each of the multiple samples, and uses the results of this deep learning to derive the three-dimensional shape of the sample SA based on the angular distribution of the scattering intensity of the sample SA received by the receiving unit 104.

[0101] By configuring the system in this way, it is possible to perform deep learning on the relationship between the angular distribution of the scattering intensity of each of the multiple samples and the three-dimensional shape of each of the multiple samples. Therefore, using the results of deep learning, it is possible to derive (discriminate) the three-dimensional shape of the received sample SA based on the angular distribution of its scattering intensity.

[0102] In the information processing device 100, the sample SA is an isolated light scatterer, and deep learning is performed with a batch size of 48 to 192, a number of iterations of 150 to 500, a number of parameters in each learning layer of 60 to 200, and a number of learning layers of 2 or more. By configuring it in this way, it is possible to optimize important parameters in the deep learning program, thereby improving the accuracy of the response.

[0103] In the information processing device 100, the three-dimensional shape of the sample includes a three-dimensional shape in which the cross-sectional shape of the sample in a plane containing the optical axis is triangular, quadrilateral, or elliptical. By configuring it in this way, the information processing device 100 can use the results of deep learning to derive (determine) whether the shape of the cross-section of the sample SA in a plane including the optical axis is triangular, square, or elliptical, based on the angular distribution of the scattering intensity of the received sample SA.

[0104] In the information processing device 100, the three-dimensional shape of the sample is such that the sample surface is uneven and includes a three-dimensional shape in which at least some of the unevenness has different heights. By configuring it in this way, the information processing device 100 can use the results of deep learning to derive (discriminate) a three-dimensional shape of the received sample SA from the angular distribution of scattering intensity of the sample SA, where the sample surface is uneven and at least some of the unevenness has different heights.

[0105] According to the optical measurement system 10 of this embodiment, it is possible to measure the light wave scattering of complex shapes by easily selecting a specific position. Furthermore, analysis can be performed at high speed on the order of milliseconds. The specific solutions are as follows: (1) Compared to imaging measurements, light wave scattering measurements can be expected to improve the resolution by two orders of magnitude in three-dimensional measurements. The problem of unclear measurement positions can be solved by combining it with imaging measurements. (2) By performing analysis of light wave scattering measurement results using deep learning, not only can the analysis results be made more objective, but automated measurement can also be performed. The shape can be determined immediately during measurement.

[0106] As a result, three-dimensional micro-shapes can be measured with high resolution and accuracy. In particular, multiple micro-shapes of different widths and shapes can be measured at high speed. To achieve this, the system incorporates a mechanism that forms an image across a wide field of view and moves the focus of the optical wave scattering measurement to a specific measurement point within that field of view. In addition, performing the analysis of optical wave scattering measurements of isolated systems using deep learning also offers the advantage of high-speed analysis. As a result, the 3D shape of a specific area of ​​interest can be measured with high accuracy and speed for complex shapes. The ability to measure three-dimensional shapes at the wavelength level enables the quantification of absorption spectra. Absorption spectra are distorted by diffraction of fine structures because scattering intensity and scattering angle distribution differ depending on the wavelength. If these wavelength dependencies can be estimated through simulation, the distortion can be corrected.

[0107] According to the optical measurement system of this embodiment, it is possible to easily realize this by adding an optical system that focuses white light only on characteristic areas and a spectrometer compatible with optical fibers to the optical system. It corrects the absorption spectrum for isolated shapes (see, for example, Non-Patent Document 9). The inventors realized that if this could be done for the same sample, distortion could be corrected. However, it was not clear what kind of optical system would be capable of performing absorption spectrum measurement, light wave scattering measurement, and imaging in a single system. To summarize, the above points are as follows: 1. The range of applications for optical wave scattering measurements, which was previously limited to periodic or single structures, will expand to include a wide range of applications such as micro-semiconductor circuits and biological neural circuits. 2. Automatic measurement provides objectivity to the results, and since the results are available during measurement, it is possible to make decisions during the measurement process and proceed with the next measurement accordingly. 3. Since the absorption spectrum can be quantified based on the measured three-dimensional structure, it becomes possible to calculate quantitative absorption spectra for complex structures, which was not possible with conventional methods.

[0108] Although embodiments of the present invention have been described above, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, modifications, and combinations can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. The aforementioned information processing device 100 has an internal computer. The processes of each of the aforementioned devices are stored in program form on a computer-readable recording medium, and the above processes are performed when the computer reads and executes this program. Here, a computer-readable recording medium refers to a magnetic disk, magneto-optical disk, CD-ROM, DVD-ROM, semiconductor memory, etc. Alternatively, this computer program may be distributed to a computer via a communication line, and the computer that receives the distribution may execute the program. Furthermore, the above program may be intended to implement some of the functions described above. Furthermore, the aforementioned functions may be implemented in combination with programs already recorded in the computer system, such as so-called differential files (differential programs). [Explanation of symbols]

[0109] 10…Optical measurement system, 1…First stage, 2…Second stage, 3…Third stage, 4…Laser light focusing spot shape adjustment means, 5…Laser light focusing spot position adjustment means, 7…Imaging lens, 8…Image sensor, 12…Laser light source, 14…Light source, 16…Imaging device, 100…Information processing device, 102…Input unit, 104…Reception unit, 106…Processing unit, 107…Learning model, 108…Output unit, 110…Storage unit, 150…Display

Claims

1. A first light source, The second light source and An objective lens that focuses the light from the first light source onto the measurement point of the sample to be measured, A display means for measuring scattered light scattered from a sample by focusing the light from the first light source onto the measurement point of the sample with the objective lens, and displaying the scattering pattern of the sample, An imaging element that forms an image of a range including a bright spot caused by light focused from the second light source onto the measurement point of the sample measured by the objective lens, An image sensor that converts the image, including the bright spot formed by the imaging element, into an electrical signal, A first drive unit moves the measurement point of the sample by moving the sample along the optical axis of the light from the first light source and along an axis perpendicular to the optical axis, A second drive unit moves either the imaging element or the image sensor, or both, along the optical axis. Equipped with, The objective lens creates an image of the range including the bright spot from the light transmitted or reflected by the sample from the light irradiated by the second light source. The imaging element forms an image on the image sensor of the range including the bright spot created by the objective lens. By adjusting the first drive unit and the second drive unit, the light is focused to a specific measurement point within the range including the bright spot that is imaged on the image sensor. The scattering pattern is the angular distribution and wavelength distribution of the scattering intensity of the sample, in this optical measurement system.

2. The first drive unit moves the sample by 0.1 mm or more along the optical axis of the light and the axis perpendicular to the optical axis. The optical measurement system according to claim 1, wherein the second drive unit moves either the imaging element or the image sensor, or both, by 1 mm or more along the optical axis.

3. A receiving unit that receives the angular distribution of the scattering intensity of the aforementioned sample, A processing unit that derives the three-dimensional shape of the sample based on the angular distribution of the scattering intensity of the sample received by the receiving unit. Equipped with, The optical measurement system according to claim 1, wherein the processing unit performs deep learning on the relationship between the angular distribution of the scattering intensity of each of the multiple samples and the three-dimensional shape of each of the multiple samples, and uses the results of the deep learning to derive the three-dimensional shape of the sample based on the angular distribution of the scattering intensity of the sample received by the receiving unit.

4. The aforementioned sample is an isolated light scatterer, The optical measurement system according to claim 3, wherein the deep learning is performed with a batch size of 48 to 192, a number of iterations of 150 to 500, a number of parameters in each learning layer of 60 to 200, and a number of learning layers of 2 or more.

5. The optical measurement system according to claim 3 or 4, wherein the three-dimensional shape of the sample includes a three-dimensional shape in which the cross-sectional shape of the sample in a plane containing the optical axis is triangular, square, or elliptical.

6. The optical measurement system according to any one of claims 3 to 5, wherein the three-dimensional shape of the sample is such that the sample surface is uneven.

7. A first light source, The second light source and An objective lens that focuses the light from the first light source onto the measurement point of the sample to be measured, A display means for measuring scattered light scattered from a sample by focusing the light from the first light source onto the measurement point of the sample with the objective lens, and displaying the scattering pattern of the sample, An imaging element that forms an image of a range including a bright spot caused by light focused from the second light source onto the measurement point of the sample measured by the objective lens, An image sensor that converts the image, including the bright spot formed by the imaging element, into an electrical signal, A first drive unit moves the measurement point of the sample by moving the sample along the optical axis of the light from the first light source and along an axis perpendicular to the optical axis, A second drive unit moves either the imaging element or the image sensor, or both, along the optical axis. An optical measurement method performed by an optical measurement system comprising: The objective lens focuses the light from the first light source onto the measurement point of the sample to be measured. The display means measures scattered light scattered from the sample by focusing the light from the first light source onto the measurement point of the sample with the objective lens, and displays the scattering pattern of the sample. The imaging element forms an image of a range including a bright spot caused by light from the second light source focused on the measurement point of the sample measured by the objective lens, The image sensor performs the steps of converting the image, including the bright spot formed by the imaging element, into an electrical signal. A first drive unit moves the sample along the optical axis of the light from the first light source and along an axis perpendicular to the optical axis, thereby moving the measurement point of the sample; and a second drive unit moves either or both of the imaging element and the image sensor along the optical axis, thereby focusing the light onto a specific measurement point within a range including the bright spot imaged on the image sensor. It has, The objective lens creates an image of the range including the bright spot from the light transmitted or reflected by the sample from the light irradiated by the second light source. The imaging element forms an image on the image sensor of the range including the bright spot created by the objective lens. A method for optical measurement in which the scattering pattern is the angular distribution and wavelength distribution of the scattering intensity of the sample.

Images