Broadband profilometer system and method for constructing a three-dimensional profile of an object

Abstract

The three-dimensional profilometer includes a broadband radiation source (110, 10, 10', 10”). An interferometric measurement system (100, 90) receives the radiation and includes a first beam splitter (114, 14, 14', 14”) and a second beam splitter (123, 23, 23', 23”), a moving time-delayed induced reflector (116, 16, 16', 16”), and a fixed reflector (118, 18, 18', 18”). The interferometric measurement system (100, 90) uses the first beam splitter (114, 14, 14', 14”) to create a time-delayed optical sample radiation source and an optical reference incident radiation source. A stationary sample container receives the optical sample incident radiation. A reference plane (140, 40, 40') receives the optical reference incident radiation. The detector (170, 70, 70', 70”) receives the interference signal from the reflected or scattered optical sample radiation and the reflected or scattered optical reference radiation. The processor (80, 80', 80”) extracts the optical path difference between the reference plane (140, 40, 40') and the sample (150, 50, 50') and reconstructs the three-dimensional morphology of the sample.

Description

Technical Field

[0001] The present invention provides a non-invasive, non-contact, and label-free device and method for obtaining the three-dimensional contour of an object, in one embodiment of which the object may be a translucent micro-object (e.g., a living cell). Background Technology

[0002] Interferometry is a widely used technique for measuring various properties of a target sample. An interferometer uses the principle of interference, where light or other electromagnetic waves superimpose; the resulting interference is analyzed to extract sample characteristics. In an interferometer, the source light is split into two paths, one pointing towards the sample and the other towards a reference component. Reflected or scattered light is combined and sent to a detector. The interference fringes provide information about the optical path difference between the sample light and the reference light, which can be correlated with properties such as the surface features of the sample.

[0003] While interferometry can be used in a wide range of technical disciplines, its application can be limited when analyzing biological samples. Biological samples, such as translucent living cells, are often immersed in liquids that require measurement / imaging through these liquids. Due to the presence of liquids or multiple layers in the sample, conventional interferometers may be unable to determine the optical path difference between the sample and the reference arm.

[0004] Other techniques can be used to determine the three-dimensional properties of an object. However, techniques such as scanning electron microscopy require time-consuming and expensive sample preparation and are performed in a vacuum environment. Other techniques, such as cell staining, kill the sample and therefore cannot be used to observe real-time cellular functions such as cell division.

[0005] Therefore, there is a need in the art for improved interferometers and methods that can efficiently measure the multi-layer profiles of samples (including biological samples) in a non-destructive manner with minimal sample preparation. This invention addresses this need. Summary of the Invention

[0006] This invention provides three-dimensional contours of multi-layered samples, such as biological samples immersed in liquid. This is achieved by using time-delayed light in the source arm. In this way, vibrations and movements of the objective lens or sample are eliminated, and the living sample and the surrounding fluid are unaffected by motion.

[0007] In one aspect, the present invention provides an optical system for reconstructing the morphology of a sample, wherein in one embodiment the sample may be a translucent biological microsample. The system includes a broadband radiation source and, optionally, a first lens system, wherein the first lens system communicates with the radiation source and outputs a collimated radiation beam. An interferometric measurement system is configured to receive the collimated radiation beam and includes at least a first beam splitter, a moving time-delayed induced reflector, and a fixed reflector. The interferometric measurement system utilizes the first beam splitter to create a time-delayed optical sample incident radiation source and an optical reference incident radiation source. A stationary sample container receives the optical sample incident radiation. A reference plane receives the optical reference incident radiation. A detector receives the interference signal of radiation reflected or scattered from the optical sample and radiation reflected or scattered from the optical reference. A processor extracts the optical path difference between the reference plane and the sample and reconstructs the three-dimensional morphology of the sample. Attached Figure Description

[0008] Figure 1A-1B An optical system according to a first embodiment and an overview of the optical system are schematically depicted;

[0009] Figure 2 It is an optical system according to another embodiment;

[0010] Figure 3 It is an optical system according to another embodiment;

[0011] Figure 4 This is a graph of the signal peak value generated by the system in Figure 1;

[0012] Figure 5 From Figure 4 The description of peak intensity information extracted from the curve.

[0013] Figure 6 The image shows a red blood cell produced by the optical system in Figure 1.

[0014] Figures 7A-7D Multiple images of various samples that can be observed / examined using the optical system of the present invention are shown. Detailed Implementation

[0015] Please refer to the attached diagram for details. Figure 1A The main components of an optical system 100 according to one embodiment are schematically depicted. In one aspect, the invention compensates for the increase in focal length caused by liquid within the sample to be measured by using a time delay in the sample measurement path of the optical system. The optical system 100 can be used to reconstruct the morphology of a three-dimensional sample; in one embodiment, this can be a translucent micro-sample.

[0016] System 100 includes an electromagnetic radiation source 110, which may be, for example, light, UV light, infrared light, etc. A beam splitter 114 directs the light to a movable reflector 116, which introduces a time delay, and a fixed reflector 118. The delayed light is incident on a sample 150, which in one embodiment may be a stationary sample comprising a liquid. Light from the fixed reflector is incident on a reference plane 140. An interference pattern containing information about both the sample and the reference reflected light enters a detector 170 and can be modeled as follows:

[0017] I = I(x, y) + c * f(x,y,z)f(x,y,z-dz)cos(φ+kdz)

[0018] Figure 1B A more detailed description of an optical system 90 according to an embodiment is shown. System 90 uses a broadband electromagnetic radiation source 10. In one aspect, the broadband electromagnetic radiation source 10 may be a broadband light source that emits continuous spectrum light. This light may be white light, such as from a white LED or halogen light source, or it may use other spectral regions of electromagnetic radiation, such as ultraviolet or infrared light. The light enters a selectively configured lens 12, such as a collimating lens, to produce collimated incident light 13.

[0019] Collimated beam 13 enters a first beam splitter 14, which transmits half the light to a fixed / stationary mirror 18 and reflects the other half to a movable reflector 16. The movable reflector induces a time delay through movement in a direction perpendicular to its surface. In one aspect, the movable reflector 16 may be a piezoelectric mirror 16. The piezoelectric mirror 16, under the direction of a mirror controller 20, introduces a controlled time delay component into the reflected beam 17.

[0020] In contrast, the light 19 reflected by the fixed reflector 18 does not have a time delay. The two beams of light reflected by the fixed reflector and the movable reflector produce a coaxially overlapping coherent correlated image. These reflected beams are then guided by the first beam splitter 14 to the second beam splitter 23 after passing through the optical lens 21.

[0021] In the second beam splitter 23, light is transmitted / reflected into beam 30 (optical sample beam) and beam 35 (optical reference beam). Each of these beams passes through corresponding objectives 32 and 39. The optical sample beam is incident on sample 50 located on the fixed platform / stage 60, while the optical reference beam is incident on reference plane 40. Reference plane 40 may be a partial reflector located at the focal plane of objective 39.

[0022] It should be noted that the light incident on the sample is sufficient to image the entire area of ​​the sample using tomographic techniques, where the light is incident at different thicknesses of the sample. Therefore, the entire xy-plane image at thickness Z will be captured from the light reflected / scattered by the sample. However, despite... Figure 1B The system does not use this feature, but it can also generate a more focused beam to scan the sample surface in a raster-like manner. The sample is placed on a sample platform 60 (which is used for biological samples immersed in liquid), and the sample platform 60 may include a waterproof sample holder, such as a culture plate. The controller 55 is able to precisely position the sample platform 60 to ensure optical alignment.

[0023] Light is reflected / scattered by sample 50 and combined with light reflected / scattered by reference plane 40 in second beam splitter 23 to generate an interference light signal 67, which is guided to detector 70 through selectively positioned lens 65. The detector can be one or more charge-coupled devices (CCDs), a camera, or any other known detector (e.g., a UV or infrared detector when the light source is UV or IR). To create a full three-dimensional image, tomography is used to image the sample across its continuous thickness. As used herein, the term "tomography" is broadly used to refer to any type of imaging performed on multiple cross-sections of the sample in the z-direction (i.e., perpendicular to the incident source) by radiation from an incident electromagnetic source (e.g., light, infrared light, or UV light).

[0024] To image the entire thickness of the sample, consecutive images of the xy plane at different incident light depths (z-direction) are captured. Light is allowed to be incident at different thicknesses by repositioning the movable reflector 16 to different positions in the z-direction (i.e., perpendicular to the incident light direction).

[0025] Processor 80 extracts the optical path difference between the reference plane and the sample, and reconstructs the images captured at each thickness; due to the low coherence of light (due to the use of a broadband light source), only relatively thin depths can generate interference signals. All xy images across the entire sample thickness are merged and reconstructed using tomographic images to reconstruct the three-dimensional topography of the sample. The phase and intensity of light in each pixel are determined.

[0026] The light stripes represent the phase difference between the sample and the reference plate; by numerically calculating the location of the signal, the optical path difference between the sample surface and the reference plane will be extracted, the tomographic phase and intensity information can be extracted layer by layer, and finally the three-dimensional cell morphology can be reconstructed.

[0027] A reflection image can be represented by a 3D reflection profile f(x, y, z). An interferometric image containing information about both the sample and reference reflected light can be modeled as follows:

[0028] I=I(x,y)+c*f(x,y,z)f(x,y,z-dz)cos(φ+kdz)

[0029] The first term is the background image (which remains unchanged during scanning), and the second term is the image modulated by low-coherence interference. As the delay is scanned, the image changes with the autocorrelation of the three-dimensional reflectivity.

[0030] If the image is captured with a quarter-wave step delay, f(x, y, z-dz) can be obtained. The quarter-wave step delay is achieved by moving a piezoelectrically controlled mirror.

[0031] I1(x,y)=I+c*f0*f(x,y,z)*cos[φ(x,z)]φ′(z)=0(0°)

[0032] I2(x,y)=Ic*f0*f(x,x,z-λ / 4)*sin[φ(x,y)]=π / 2(90°)

[0033] I3(x, y)=Ic*f0*f(x, y, z-2λ / 4)*cos[φ(x, y)]=π(180°)

[0034] I4(x, y)=I+c*f0*f(x, y, z-3λ / 4)*sin[φ(x, y)]=3π / 2(270°)

[0035] but

[0036] (x, y, z) = tan -1 [(I4(x, x) - I2(x, y)) / (I1(x, y) - I3(x, y))]. Here, it is assumed that f(x, y, z) remains constant over a phase delay of one wavelength.

[0037] The intensity variation caused by the quarter-wavelength phase delay can be represented as stripes on the image. This image was captured by detector 70. Figure 4 The image shows the intensity variation for each pixel.

[0038] For an image in the xy-plane, the irradiance at each pixel is recorded. The best fringe localization position for each pixel is determined by a series of pre-processing, processing, and post-processing steps. During pre-processing, image filtering algorithms can be applied to minimize the effects of dispersion imbalance, phase changes, and system vibrations. Image enhancement algorithms can be applied depending on the nature of the sample. These algorithms benefit from the high resolution in axial scanning and the "over-resolution" in lateral scanning.

[0039] During the processing phase, a fast envelope and peak detection algorithm is applied to determine the surface and potential multilayer characteristics. Then, various methods, including frequency domain analysis, phase change detection, and time domain filtering, are used to analyze the characteristics of individual fringes and pinpoint their locations. This process is accelerated by the Compute Unified Device Architecture (CUDA), thereby reducing processing time. Figure 5 The image shown is a processed image, displaying signal peaks amplified by noise. In a later stage, the point cloud generated by the aforementioned techniques is analyzed using various methods such as clustering and segmentation, where each point is... Figure 5 The signal peak value is shown.

[0040] Clustering algorithms are a form of unsupervised machine learning. They divide data into multiple subgroups based on learned similarity among various data points. Each data point in a cluster is more similar to other data points within that cluster, and data points belonging to different clusters are compared. Segmentation involves grouping points in subsets / segments that share one or more common features. The similarity between data points is highly dependent on the specific samples being analyzed; therefore, learning can take into account the type of sample (e.g., biological samples, inorganic samples, semiconductor devices, etc.).

[0041] Density-based clustering algorithms can be chosen for clustering. Density-based algorithms can be used when many clustered points (e.g., cellular features such as cell nuclei) are surrounded by featureless regions (e.g., protoplasmic regions primarily based on fluid). Density-based spatial clustering of applications with noise (DBSCAN) can be used to effectively reduce noise from the data. Other algorithms based on the DBSCAN method can also be used. In the post-processing stage, the point cloud is processed, and the clustering and segmentation information are processed together to present an accurate 3D topography. Various tools are provided for surface inspection, roughness measurement, and visualization. An example of the generated 3D topography is... Figure 6 Images of red blood cells.

[0042] Figure 2 Another embodiment of the optical system is depicted in the diagram. Figure 2 In the embodiments, with Figure 1B Components that are substantially similar are indicated by the same reference numerals. Figure 2 In this embodiment, radiation from radiation source 10′ enters collimating lens 12′ and then beam splitter 14′. Reflected / scattered light from the sample and reference planes is combined and penetrates telecentric lens 72 before entering detector 70′.

[0043] Figure 3 An embodiment of the optical system is depicted. Figure 3 In the embodiments, with Figure 1B Components that are substantially similar are indicated by the same reference numerals. Figure 3 In this system, radiation is incident on the opposite side of the sample. Furthermore, polarizers 74 and 76 are used to allow polarized light to enter detector 70". The use of polarized light enhances contrast, thereby improving the quality of the generated 3D image. When the sample is a biological sample in a liquid such as water, the use of polarized light reduces unwanted reflections from the water surface.

[0044] Industry applicability

[0045] This invention can be used to observe three-dimensional cellular structures, such as the cell nucleus or other organelles. Typically, structures can be observed if different proteins have different refractive indices. Such biological samples should be translucent or transparent to obtain three-dimensional imaging. Only a small amount of sample preparation is required; the sample may be a live specimen for observing cellular activity (such as cell division). Applications include pathology, cancer detection and research, and observing topological changes in cell samples.

[0046] Because this technology does not require extensive sample preparation and is non-destructive, it can be used for non-destructive assessment of structures, including crack detection and other defects. It can also serve as a final inspection tool in fields such as semiconductor device manufacturing.

[0047] Figures 7A-7D Multiple images of various samples that can be observed / inspected for quality control or other purposes are shown. Figure 7A It is an image of the ink on a banknote, which can show whether there are any printing defects during the printing process. Figure 7B It is an image of human hair. Figure 7C This is an image illustrating the line sensor of the present invention used for equipment inspection. Figure 7D The image shows an inspection of the spray coating using the present invention, which can be used to determine coating uniformity. Although the images are reproduced in black and white, it should be understood that the system can be used to generate color images.

[0048] It will be apparent to those skilled in the art that many modifications can be made beyond those already described without departing from the inventive concept of this document. Therefore, the subject matter of this invention is not limited except in the spirit of this disclosure. Furthermore, in interpreting this disclosure, all terms should be interpreted in the broadest possible sense, consistent with the context. In particular, the terms “includes,” “including,” “comprises,” and “comprising” should be interpreted as referring to an element, component, or step in a non-exclusive manner, indicating that the referenced element, component, or step may be present, used, or combined with other elements, components, or steps not explicitly referenced.

Claims

1. An optical system for reconstructing a three-dimensional topography of a sample, characterized in that, The system includes: Broadband radiation source; An interferometric measurement system configured to receive supplied radiation from the broadband radiation source, wherein the interferometric measurement system comprises: A first beam splitter, a moving time-delay-inducing reflector, and a fixed reflector, wherein the radiation is directed to the first beam splitter, which transmits half of the radiation to the fixed reflector to generate optical reference incident radiation, and reflects the other half of the radiation to the moving time-delay-inducing reflector to generate time-delayed optical sample incident radiation; and The second beam splitter receives the optical reference incident radiation and the time-delayed optical sample incident radiation; A stationary sample container, on which a sample is placed, is used to receive the time-delayed optical sample incident radiation reflected by a second beam splitter, thereby producing reflected or scattered optical sample radiation, wherein the second beam splitter further receives the reflected or scattered optical sample radiation from the stationary sample container. A reference plane is used to receive the optical reference incident radiation passing through the second beam splitter, thereby generating reflected or scattered optical reference radiation, wherein the second beam splitter further receives the reflected or scattered optical reference radiation from the reference plane, and a first optical path located between the second beam splitter and the stationary sample container and a second optical path located between the second beam splitter and the reference plane are spatially separated. The detector is configured to receive an interference signal generated by combining the reflected or scattered optical sample radiation with the reflected or scattered optical reference radiation through the second beam splitter; A processor is configured to extract the optical path difference between the reference plane and the sample, and generate the three-dimensional morphology of the sample. The processor generates an xy-plane image for the z-direction thickness of each sample, generates and records the irradiance of each pixel in the xy-plane image, and generates the fringe positioning positions of the pixels determined by a series of preprocessing, processing, and post-processing stages. In the preprocessing stage, an image filtering algorithm is applied to minimize dispersion imbalance, phase change, and system vibration. In the processing stage, a fast envelope and peak detection algorithm is applied to determine surface and multilayer characteristics. The generated point cloud is also analyzed by clustering and segmentation machine learning to group points with common features, and each xy-plane image is denoised. In the z-direction, multiple generated xy-plane images are combined to create the three-dimensional morphology of the sample.

2. The optical system of claim 1, wherein The reconstruction includes extracting the phase and intensity of the tomographic scan of the generated xy-plane image of the sample.

3. The optical system of claim 1, wherein The movable time-delay induced reflector includes a piezoelectric reflector.

4. The optical system of claim 1, wherein It also includes a first polarizer and a second polarizer, the first polarizer being positioned to generate a polarized optical sample incident radiation source and a polarized optical reference incident radiation source, and the second polarizer being positioned to guide polarized radiation to the detector.

5. The optical system of claim 1, wherein It also includes a collimating lens positioned between the broadband radiation source and the beam splitter.

6. A method for reconstructing a three-dimensional profile of a sample, characterized in that, The method includes: Broadband radiation is directed to an interferometric measurement system, which includes a first beam splitter, a second beam splitter, a movable time-delay induced reflector, and a fixed reflector. The broadband radiation is directed to the first beam splitter, which transmits half of the broadband radiation to the fixed reflector to generate optical reference incident radiation. The first beam splitter reflects the other half of the broadband radiation to the movable time-delay induced reflector to generate time-delayed optical sample incident radiation. The optical reference incident radiation and the time-delayed optical sample incident radiation are received by the second beam splitter. By using the second beam splitter, the time-delayed optical sample incident radiation is directed onto the sample on the stationary sample container, and the optical reference incident radiation is directed onto the reference plane. The second beam splitter reflects the time-delayed optical sample incident radiation onto the sample on the stationary sample container. The second beam splitter allows the optical reference incident radiation to pass through it to reach the reference plane, and the first optical path located between the second beam splitter and the stationary sample container and the second optical path located between the second beam splitter and the reference plane are spatially separated. In the second beam splitter, the optical sample radiation reflected or scattered from the sample on the stationary sample container is combined with the optical reference radiation reflected or scattered from the reference plane, and the combined radiation is guided to the detector. The optical path difference is extracted from the radiation fringes of the merged radiation, wherein the radiation fringes indicate the phase difference between the sample and the reference plane; and The three-dimensional morphology of the sample is generated based on the optical path difference by sequentially performing tomographic imaging on the sample. Generating the three-dimensional morphology includes generating an xy-plane image for the z-direction thickness of each sample, generating and recording the irradiance of each pixel in the xy-plane image, and generating the fringe location of the pixel determined through a series of preprocessing, processing, and post-processing stages. In the preprocessing stage, an image filtering algorithm is applied to minimize dispersion imbalance, phase change, and system vibration. In the processing stage, a fast envelope and peak detection algorithm is applied to determine surface and multilayer characteristics. Furthermore, the generated point cloud is analyzed using clustering and segmentation machine learning to group points with common features, and each xy-plane image is denoised. Multiple generated xy-plane images are then combined in the z-direction to create the three-dimensional morphology of the sample.

7. The method according to claim 6, characterized in that, The time-delayed induced reflector is repositioned to different locations to perform sequential imaging of the sample, such that the time-delayed optical sample incident radiation is incident on different thicknesses of the sample.

8. The method according to claim 6, characterized in that, The sample is immersed in a liquid.

Images