Spectrum confocal efficient three-dimensional measurement device and method based on spatial light modulator

Abstract

The invention discloses a spectral confocal efficient three-dimensional measurement device and method based on a spatial light modulator. The device comprises an illumination module and a detection module. The illumination module comprises a white light LED light source, a collimator, an LED condenser, a digital micromirror device DMD and a dispersion objective lens; the detection module comprises an imaging objective lens, a CCD detector and computer processing equipment. According to the invention, traditional point-by-point spectrum confocal measurement is expanded to area array type measurement, so that the measurement speed can be effectively improved; meanwhile, different from a scanning type or fixed pinhole type parallel measurement scheme, the invention uses a pinhole array with controllable diameter and spacing, thereby improving the flexibility and lateral resolution of measurement.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical measurement. Specifically, it relates to a spectral confocal high-efficiency three-dimensional measurement device and method based on a spatial light modulator. Background Art

[0002] Nowadays, the designs of semiconductor devices and optical components are developing towards more precise three-dimensional microstructures. Therefore, it is crucial to achieve efficient and high-precision measurement of the three-dimensional structures of micro-devices. Non-contact measurement represented by optical detection can better meet such detection requirements. Compared with laser triangulation and laser interferometry, which are greatly affected by the environment, scanning confocal microscopes are more stable. They use an optical pinhole to suppress defocused light, thereby achieving higher precision.

[0003] Confocal microscopes use a single-wavelength light source. When the focus is on the surface of the sample, the peak intensity can be detected. A series of optical cross-sections are scanned perpendicular to the optical axis, and finally a depth response curve related to the light intensity is generated. In order to actually measure the entire surface of an object, lateral scanning is also required. Horizontally, representative methods for obtaining parallel measurement points include using a scanning galvanometer, a Nipkow disk, a slit, a microlens array, etc. Restricted by the instability and speed limitations brought by mechanical scanning, and the fixation of static arrays, etc., the overall efficiency and universality are difficult to improve. Therefore, for such devices, it is equivalent to using a zero-dimensional sensor combined with three-dimensional (axial and lateral) scanning to achieve efficient three-dimensional measurement, which increases the measurement time and various measurement uncertainties introduced by mechanical scanning. Overall, improvements need to be made from two aspects: increasing the dimension of sensing and reducing the dimension of scanning. Summary of the Invention

[0004] Aiming at the deficiencies of the above-mentioned prior art, the purpose of the present invention is to provide a spectral confocal high-efficiency three-dimensional measurement device and method based on a spatial light modulator that can achieve high-precision and high-efficiency three-dimensional microstructure measurement. The present invention uses a digital micromirror device (DMD) as a high-speed and high-precision spatial light modulator, extending the traditional point-by-point spectral confocal measurement to area array measurement, which can effectively improve the measurement speed. At the same time, different from the parallel measurement schemes of scanning type or fixed pinhole type, the present invention uses a pinhole array with controllable diameter and pitch, improving the measurement flexibility and lateral resolution.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows.

[0006] A spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator, which includes an illumination module and a detection module; wherein:

[0007] The illumination module includes a white light LED light source, a collimator, an LED condenser lens, a digital micromirror device DMD, and a dispersive objective lens; the detection module includes an imaging objective lens, a CCD detector, and a computer; within the bandwidth of the light source used, the wavelength provided by the dispersive objective lens has a high linearity with the focal shift curve, so that the measurement results have a consistent axial resolution at different wavelengths, and the size of the ON pixels on the digital micromirror device DMD is larger than the lateral resolution on the image plane; the digital micromirror device DMD acts as both an illumination and a detection pinhole array at the same time, the illumination pinhole and the measurement point on the sample surface form a conjugate relationship, and the measurement point on the sample surface and the detection pinhole form a conjugate relationship to achieve the confocal function;

[0008] During operation, the white light LED light source passes through a collimating mirror and an LED condenser lens, and non-parallel light at a certain angle is incident on the surface of the digital micromirror device DMD. The light reflected by each ON pixel on the surface of the digital micromirror device DMD passes through the dispersive objective lens, and light of different wavelengths converges at different depths on the sample surface. Each object point on the sample surface passes through the imaging objective lens in front of the CCD detector again and converges on the corresponding pixel of the CCD detector. The computer analyzes and processes the light intensity at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample.

[0009] In the present invention, a beam splitter is provided between the digital micromirror device DMD and the dispersive objective lens for dividing two beam paths. One path guides the beam to the dispersive objective lens and focuses it at different depths on the sample surface, and the other path sends the reflected light to the detection optical path.

[0010] In the present invention, the ON pixels on the digital micromirror device DMD scan the entire pixel array, converge at different lateral positions on the sample surface through the dispersive objective lens, and form two-dimensional measurement points on the sample surface.

[0011] In the present invention, the computer analyzes and processes the light intensity at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample, which is realized based on the following algorithm:

[0012] It is known that f(λ) is the focal length related to the wavelength λ, f0 is the focal length corresponding to the central wavelength λ0, and k is set as the focal shift factor; f(λ) = f0 + kλ

[0013] Light of different wavelengths converges at different depths z(λ min ) ~ z(λ max ) on the sample surface. z(λ min ), z(λ max ) respectively refer to the depth coordinates of the incident light with the shortest wavelength and the longest wavelength at the focal point on the sample. The difference between z(λ min ) and z(λ max ) is the defocus amount Δz(λ). The defocus amount Δz(λ) is related to the focal shift caused by the wavelength and is expressed as:

[0014] Δz(λ) = Δf(λ i ) = f(λ i ) - f(λ0) = k(λ i - λ0)

[0015] The wavelength - dependent axial light intensity response function I(λ) is as follows:

[0016]

[0017] The M×N point - source array formed by using a digital micromirror device DMD has M and N ON pixels in the x - and y - directions respectively, with a pitch of d. When the incident light is spatially modulated by the digital micromirror device DMD, each ON pixel on the DMD is regarded as an independent point source, and they all have the same source distribution δ(v), centered at v1, v2, …, v M and v1, v2, …, v N Therefore, the input object function is expressed as:

[0018]

[0019] where u and v are the normalized optical coordinates in the longitudinal and transverse directions respectively. Assuming that the lens, light - source function, and detection function all have circular symmetry, where r is the radial coordinate; the intensity distribution on the final image plane obtained using the dot matrix is the result of the superposition of point sources in the object function. The intensity signal detected by the CCD detector is expressed as:

[0020] I(v) = ∫d 2 v S(v)D(v)|h1(u, v)h2(u, v)| 2

[0021] where h1(u, v) and h2(u, v) are the point - spread functions of the dispersive objective lens and the imaging objective lens respectively. The point - spread function is the Fourier transform of the objective - pupil function, and D(v) is the detector sensitivity function.

[0022] In the present invention, the white - light LED light source can be replaced by a mercury lamp.

[0023] The present invention also provides a spectral confocal high - efficiency three - dimensional measurement method based on the above - mentioned device, including the following steps:

[0024] The white light LED light source passes through a collimating mirror and an LED condenser lens, and non-parallel light at a certain angle is obtained and incident on the surface of the digital micromirror device DMD. The light reflected by each ON pixel on the surface of the digital micromirror device DMD passes through a dispersive objective lens, and lights of different wavelengths converge at different depths on the sample surface. Each object point on the sample surface passes through the imaging objective lens in front of the CCD detector again and converges on the corresponding pixel of the CCD detector. The computer analyzes and processes the light intensities at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample.

[0025] In the present invention, based on the multi-focus distribution with a lateral interval of (δx, δy) formed by the two-dimensional point source array generated by modulating the incident light beam by the digital micromirror device DMD on the sample surface, that is, the measurement dot matrix, the CCD detector is used to analyze the position information of each lateral measurement point to obtain the three-dimensional information of the sample.

[0026] Above, most of the existing confocal or spectral confocal measurement techniques are single-point or line-type measurements. The present invention adopts a dot matrix scanning method, and its significant advantage is that it expands the one-time measurement range. Secondly, compared with other types of parallel scanning measurement methods such as the Nipkow disk, microlens array, mechanical galvanometer scanning, etc., the present invention using DMD has the following advantages: (1) Different from the pinhole type or slit type, which requires mechanical scanning to obtain lateral detection, this device has no moving parts, improving the stability of the entire measurement system.

[0027] (2) The micromirrors in the off state in the DMD can directly act as pinholes, replacing the illumination and detection pinhole arrays.

[0028] (3) The intensity and signal-to-noise ratio of the spectral confocal sensor signal largely depend on the pinhole diameter and spacing. When using a fixed microlens array or Nipkow disk, once the system is established, the pinhole pattern cannot be changed, and the tolerance to alignment errors is low. However, the programmable pixel array of DMD can be flexibly adjusted to balance between the signal intensity and the lateral resolution for different application scenarios or measurement objects. Description of the Drawings

[0029] Figure 1 is a schematic optical path diagram of the spectral confocal measurement device of the present invention.

[0030] Figure 2 is a schematic diagram of the multi-foci on the image plane generated by using non-parallel light incident on the DMD of the present invention.

[0031] Figure 3 is a schematic diagram of the M×N point source array formed by using the DMD of the present invention.

[0032] Figure 4 is a schematic diagram of the principle of obtaining three-dimensional information at one time of the present invention.

[0033] Reference numerals in the figure: 1 is a white light LED light source, 2 is a collimator, 3 is an LED condenser lens, 4 is a digital micromirror device DMD, 5 is a beam splitter, 6 is a dispersion objective lens, 7 is the sample surface, 8 is an imaging objective lens, 9 is a CCD detector, 10 is a computer, 13a is the light ray guided by the OFF pixel, 13b is the light ray guided by the ON pixel, and 14 is the light ray convergence situation after passing through the dispersion objective lens. Detailed implementation mode

[0034] The technical solution of the present invention will be further explained below in conjunction with the accompanying drawings.

[0035] As Figure 1 shown, the present invention provides a schematic optical path diagram of a spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator, which includes an illumination module and a detection module, and two beam paths are divided by a cube beam splitter. The solid line and the dotted line indicate two of the paths of the two-dimensional multi-light spot array. The illumination module includes a white light LED light source 1, a collimator 2, an LED condenser lens 3, a programmable spatial light modulator (digital micromirror device DMD 4), and a dispersion objective lens 6; the detection module includes an imaging objective lens 8, a CCD detector 9, and a computer 10. In the present invention, the confocal optical path requires two groups of one-to-one conjugate relationships: the conjugate relationship between the illumination pinhole and the detection point on the sample surface 7, and the conjugate relationship between the detection point and the detection pinhole. The DMD acts as both the illumination and detection pinhole arrays at the same time to achieve the confocal function.

[0036] The white light LED light source 1 and the dispersion objective lens 6 are the core of the spectral confocal technology, and the two jointly generate axial chromatic aberration. High dispersion means artificially introducing axial chromatic aberration into the system, mainly introduced in the focusing objective lens, and also including the residual chromatic aberration of various components of the system, and only spherical aberration is eliminated. Due to the action of axial chromatic aberration in the system, only the wavelength component focused on the object surface will be focused on the plane of the CCD detector 9. At different heights on the sample surface 7, light of different wavelength components is focused and reflected.

[0037] The digital micromirror device DMD 4 is used to generate a measurement dot matrix, and its micromirror array acts as both the illumination and detection pinhole arrays at the same time. The ON pixel guides the light beam to the sample surface 7, and the OFF pixel makes the light beam away from the sample, producing a pinhole filtering effect. The ON pixels quickly scan the field of view according to a pre-designed two-dimensional array pattern.

[0038] The LED condenser lens 3 of the illumination module is used to control the non-parallel light with a certain angle to be incident on a series of micro mirrors of the DMD, so as to generate a multi-focus distribution corresponding to the DMD ON pixels on the sample surface 7.

[0039] The detection module is used to analyze the light reflected from the surface 7 of the sample. The focused light has a higher intensity than the defocused light, so as to obtain the intensity peak related to the wavelength, associate the height of the sample surface 7 with the light source wavelength, and no axial displacement of the sample is required during the whole process, thus realizing non-mechanical depth sensing.

[0040] More specifically, in the present invention, the white light LED light source 1 passes through the collimating mirror 2 and the LED condenser lens 3 to obtain non-parallel light at a certain angle and incident on the surface of the digital micromirror device DMD 4. The digital micromirror device DMD 4 is a micro-opto-electro-mechanical system, and its chip surface is composed of more than two million rectangular aluminum micro-mirrors, corresponding to individually addressable pixels. Each micromirror is in a static state of 0°, and can rotate ±12° separately around the flexible hinge, corresponding to the ON / OFF states respectively, so as to selectively guide the light.

[0041] When using the digital micromirror device DMD 4 as a spatial light modulator, the object-side telecentric optical path is usually adopted, and only the chief ray parallel to the optical axis participates in imaging. The difference between the digital micromirror device DMD 4 and the microlens array is that the digital micromirror device DMD 4 is composed of a series of mirrors. When the incident light is parallel light, the reflected light of all micromirrors is parallel light, and after being converged by the lens, the result should be a single image point on the image plane. Compared with using multiple sub-intensity distributions within a focus to identify, the multi-focus distribution is more likely to identify different measurement points. Figure 2 Taking the 2×2 ON pixels as an example, these two situations are compared. Therefore, in order to be able to generate different intensities at the lateral position, it is necessary to add a certain angle to the incident light through the LED condenser lens 3.

[0042] The light reflected from each ON pixel on the surface of the digital micromirror device DMD 4 passes through the dispersion objective lens 6. Different from the achromatic lens or ordinary lens, when designing this kind of dispersion objective lens optically, the optimization purpose is to increase the axial chromatic aberration within a certain wavelength bandwidth and reduce the spherical aberration. Usually, the structure of a refractive lens group, a refractive-diffractive hybrid lens or a pure diffractive lens is used, which is beneficial to increasing the chromatic aberration. Lights of different wavelengths converge at different depths (z(λ min )~z(λ max )) on the surface 7 of the sample, z(λ min ), z(λ max) are the coordinates of different focal points in space, corresponding respectively to the depth coordinates of the incident light with the shortest wavelength and the longest wavelength at the sample focal point. The difference, i.e., the defocus amount Δz(λ), is associated with the wavelength. During the design process of the dispersive objective lens, it is also required that the linearity of the wavelength versus focal shift curve be as high as possible to ensure consistent axial resolution at different wavelengths. Each ON pixel horizontally distributed on the digital micromirror device DMD 4 converges at different horizontal positions on the sample surface. Each object point on the sample surface 7 passes through the objective lens again and converges at the corresponding pixel on the CCD detector 9. The computer 10 analyzes and processes the intensities at different points of the CCD detector 9, and based on the Δz(λ)~I(λ) relationship established through pre-calibration, finally obtains the depth information of each measurement point on the sample. In a confocal microscope illuminated by monochromatic light, the axial light intensity is distributed according to the square of the sinc function:

[0043]

[0044] where a is the radius of the rotationally symmetric lens, is the representation of the normalized optical coordinate, f is the focal length related to the wavelength, and f0 is the focal length corresponding to the central wavelength λ0. In color confocal microscopy, I(u) no longer follows the sinc 2 distribution. Let k be the focal shift factor, f(λ) = f0 + kλ. The defocus amount Δz(λ) is related to the focal shift caused by the wavelength and can be expressed as:

[0045] Δz = Δf(λ i ) = f(λ i ) - f(λ0) = k(λ i - λ0)

[0046] f(λ i ), f(λ0) are the focal lengths (length values) corresponding to the wavelengths λ i and λ0 respectively,

[0047] Substituting the above into the formula, the axial light intensity response function I(λ) related to the wavelength can be obtained as:

[0048]

[0049] There are two ways to generate a matrix of light points using the digital micromirror device DMD 4. One way is to turn on the ON pixel array at the same time to create a synchronous pinhole array. With this scheme, the acquisition time can be reduced, but the illumination of each point does not increase, and DMD is generally not recommended to load the same image for a long time. Usually, the second method is used. Based on the ultra-high speed and high contrast performance of the digital micromirror device DMD 4, a single ON pixel is designed to perform a zigzag scan across the entire field of view at a certain interval (i.e., OFF pixels). According to the numerical aperture and magnification of the objective lens, the resolution Δα on the image plane can be calculated. Usually, the DMD pixel size is 7.6 - 13.68μm, and the ON pixel size needs to be greater than Δα. Considering the available light source power in the experiment and the sensitivity of the CCD detector 9, multiple ON pixels can be turned on as a single point source to ensure that high-quality confocal images can be detected.

[0050] Figure 3 is a schematic diagram of the M×N point source array formed by using the digital micromirror device DMD 4. Assume that the point source array formed by DMD has M and N ON pixels in the x and y directions respectively, with a spacing of d. When the incident light is spatially modulated by DMD, each ON pixel on DMD can be regarded as an independent point source, all having the same source distribution δ(v), centered at v1, v2, …, v M and v1, v2, …, v N respectively. Therefore, the input object function can be expressed as:

[0051]

[0052] where u and v are the normalized optical coordinates in the longitudinal and transverse (radial) directions respectively, assuming that the lens, light source function, and detection function all have circular symmetry. Among them, r is the radial coordinate. The intensity distribution on the final image plane obtained using the dot matrix is the result of the superposition of point sources in the object function. The detected intensity signal can be expressed as:

[0053] I(v) = ∫d 2 v S(v)D(v)|h1(u, v)h2(u, v)| 2

[0054] where h1(u, v) and h2(u, v) are the point spread functions of the dispersive objective lens and the imaging objective lens respectively. The point spread function is the Fourier transform of the objective lens pupil function, and D(v) is the detector sensitivity function.

[0055] To simplify the calculation, it can be assumed that two identical objective lenses are used, one as the imaging objective lens in front of the CCD and the other as the dispersive objective lens, i.e., h1 = h2 = h;

[0056] The detected intensity signal can be expressed as:

[0057] I(v) = ∫d 2 v S(v)D(v)|h(u, v)h(u, v)| 2

[0058] where it is assumed that the focusing objective lens is the same as the dispersive objective lens, h(u, v) is the objective lens point spread function, which is the Fourier transform of the pupil function, and D(v) is the detector sensitivity function.

[0059] Figure 4 The multi - focal distribution and axial chromatic aberration of the final image plane of the spectral confocal optical path in the present invention are simulated, solving the technical problem of measuring the three - dimensional information of the test microstructure. The one - time three - dimensional measurement includes two aspects: (1) The spectral confocal principle provides the one - dimensional depth discrimination ability δz(λ) without mechanical scanning. (2) The two - dimensional point source array generated by modulating the incident beam by the DMD can form a multi - focal distribution with a lateral interval of (δx, δy) on the sample surface, that is, the measurement lattice. The CCD is used to analyze the position information of each lateral measurement point, and combined with depth decoding, the three - dimensional information of the sample is finally obtained.

Claims

1. A spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator, characterized in that, It includes an illumination module and a detection module; among which: The illumination module includes a white light LED light source, a collimator, an LED condenser lens, a digital micromirror device DMD, and a dispersion objective lens; the detection module includes an imaging objective lens, a CCD detector, and a computer; within the light source bandwidth used, the wavelength provided by the dispersion objective lens and the focal shift curve have high linearity, so that the measurement results have consistent axial resolution at different wavelengths, and the size of the ON pixel on the digital micromirror device DMD is larger than the lateral resolution on the image plane; the digital micromirror device DMD serves as both an illumination and a detection pinhole array at the same time, the illumination pinhole and the measurement point on the sample surface form a conjugate relationship, and the measurement point on the sample surface and the detection pinhole form a conjugate relationship to achieve the confocal function; During operation, the white light LED light source passes through the collimator and the LED condenser lens to obtain non-parallel light at a certain angle and is incident on the surface of the digital micromirror device DMD. The light reflected by each ON pixel on the surface of the digital micromirror device DMD passes through the dispersion objective lens. Within the bandwidth of the composite light, lights of different wavelengths converge at different depths on the sample surface. Each object point on the sample surface passes through the imaging objective lens in front of the CCD detector again and converges on the corresponding pixel on the CCD detector. The computer analyzes and processes the light intensities at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample.

2. The spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator according to claim 1, wherein A beam splitter is arranged between the digital micromirror device DMD and the dispersion objective lens to divide the two beam paths. One path guides the beam to the dispersion objective lens and focuses it at different depths on the sample surface, and the other path sends the reflected light to the detection optical path.

3. The spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator according to claim 1, wherein, During operation, the ON pixels on the digital micromirror device DMD scan the entire pixel array and converge at different lateral positions on the sample surface through the dispersion objective lens to form two-dimensional measurement points on the sample surface.

4. The spectral confocal high-efficiency three-dimensional measurement device based on a spatial light modulator according to claim 1, wherein The computer analyzes and processes the light intensities at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample, which is realized based on the following algorithm: It is known that f(λ) is the focal length related to the wavelength λ, f0 is the focal length corresponding to the central wavelength λ0, and k is set as the focal shift factor; f(λ) = f0 + kλ Light of different wavelengths converges at different depths z(λ min ) ~ z(λ max ) on the sample surface. z(λ min ) and z(λ max ) respectively refer to the depth coordinates of the incident light with the shortest wavelength and the longest wavelength at the focal point of the sample. The difference between z(λ min ) and z(λ max ) is the defocus amount Δz(λ). The defocus amount Δz(λ) is related to the focal shift caused by the wavelength and is expressed as: Δz(λ) = Δf(λ i ) = f(λ i ) - f(λ0) = k(λ i - λ0) The axial light intensity response function I(λ) related to the wavelength is: The M×N point source array formed by using a digital micromirror device DMD has M and N ON pixels in the x and y directions respectively, with a pitch of d. When the incident light is spatially modulated by the digital micromirror device DMD, each ON pixel on the DMD is regarded as an independent point source, all having the same source distribution δ(v), centered at v1, v2, …, v M and v1, v2, …, v N respectively. Therefore, the input object function is expressed as: where u and v are the normalized optical coordinates in the longitudinal and transverse directions, respectively. It is assumed that the lens, the source function, and the detection function all have circular symmetry. Among them, r is the radial coordinate; the intensity distribution on the final image plane obtained using the dot matrix is the result of the superposition of point sources in the object function. The intensity signal detected by the CCD detector is expressed as: Among them, h1(u, v) and h2(u, v) are the point spread functions of the dispersion objective lens and the imaging objective lens respectively. The point spread function is the Fourier transform of the objective pupil function, and D(v) is the detector sensitivity function.

5. A spectral confocal high-efficiency three-dimensional measurement method based on the device described in claim 1, characterized in that, It includes the following steps: The white light LED light source passes through the collimator and the LED condenser lens to obtain non-parallel light at a certain angle and is incident on the surface of the digital micromirror device DMD. The light reflected by each ON pixel on the surface of the digital micromirror device DMD passes through the dispersion objective lens. Lights of different wavelengths converge at different depths on the sample surface. Each object point on the sample surface passes through the imaging objective lens in front of the CCD detector again and converges on the corresponding pixel on the CCD detector. The computer analyzes and processes the light intensities at different points of the CCD detector, and finally obtains the depth information of each measurement point on the sample.

6. The spectral confocal high-efficiency three-dimensional measurement method based on a spatial light modulator according to claim 5, characterized in that A multi-focus distribution with a lateral spacing of (δx, δy), that is, a measurement lattice, is formed on the sample surface by a two-dimensional point source array generated by modulating an incident beam based on a digital micromirror device (DMD). The position information of each lateral measurement point is analyzed using a CCD detector, and three-dimensional information of the sample is obtained by combining depth decoding.

Images