Multi-probe online real-time film thickness measuring method and device

Abstract

The invention discloses a multi-probe on-line real-time film thickness measuring method and a multi-probe on-line real-time film thickness measuring device. In combination with the multi-wavelength analysis capability and multi-channel parallel signal characteristics of the spectrum detection unit, different plane positions are adopted to correspond to different wavelengths or central wavelengths, the transmissivity of each wavelength is obtained, and the thickness of the corresponding position is obtained respectively; the precision and the accuracy of obtaining the thickness by a single wavelength are improved through a multi-mode cooperative signal analysis algorithm, so that the thicknesses of a plurality of different positions of the same thin film are accurately obtained at the same time through one spectrum detection unit, or the thicknesses of a plurality of different thin films are obtained; according to the invention, the problems of large absorption, transparent or non-transparent substrate compatibility, quick response, multi-branch and on-line real-time measurement of the film are successfully solved; the device and the method can be integrated in film growth equipment to carry out rapid and high-precision measurement on the film, so that closed-loop control on film production is realized, the iteration speed is increased, the yield of high-quality film production is greatly increased, and the film production cost is saved.

Description

Technical Field

[0001] This invention relates to optical thin film measurement technology, specifically to a multi-probe online real-time film thickness measurement method and device. Background Technology

[0002] Multi-probe film thickness gauges, as indispensable key testing equipment in precision manufacturing, play a crucial role in high-tech industries such as semiconductors, flat panel displays, and photovoltaics. Optical film thickness measurement technology is a high-precision measurement method based on the principle of optical interference and spectral analysis, integrating multiple advanced technologies such as optical path design, photoelectric detection, spectral analysis, computer modeling, and intelligent algorithms. Due to its non-contact measurement and non-destructive testing characteristics, film thickness gauges based on optical methods (such as laser interferometry, ellipsometric, and spectral reflectance techniques) have become core testing equipment in modern industrial manufacturing. For example, ellipsometers measure film thickness by measuring the change in the polarization state of reflected light from the thin film. However, this instrument is bulky (on the order of meters) and has a large incident angle (typically around 70°). In addition, ellipsometer film thickness measurement is based on complex calculations and fitting, resulting in non-unique results, and is time-consuming (on the order of minutes). Due to its large size, optical path conflicts, and long processing time, ellipsometers are difficult to integrate with thin film growth equipment, making real-time online film thickness measurement impossible. Interferometric thickness gauges calculate film thickness by fitting interference spectra, but the results are not unique and are time-consuming (on the order of tens of seconds). Furthermore, when the film thickness is on the order of tens or hundreds of nanometers, the fitting error increases, resulting in poor accuracy. Existing ellipsometers and thickness gauges cannot achieve millisecond-level rapid film thickness measurement, making it difficult to meet the online real-time measurement requirements of rapid thin film production processes in industry. With the development of processes and the increasing demands on film thickness measurement, thickness gauges are developing towards miniaturization, integration, multifunctionality, intelligence, and speed.

[0003] Multi-probe film thickness measurement technology, based on traditional single-point measurement techniques, addresses the efficiency and accuracy challenges of measuring large-size samples and complex surfaces by integrating multiple optical probes to work synchronously. Furthermore, the multi-probe design allows the system to simultaneously acquire data from different locations on the sample, significantly improving measurement efficiency and providing more comprehensive film thickness distribution information, which is revolutionary for process control and quality assurance in modern manufacturing. However, a single ellipsometer can only measure one thickness, and due to the large size of the equipment, it is difficult to integrate multiple units for multi-thickness measurement. Film thickness gauges based on interferometric methods require the use of all wavelengths to fit the film thickness, thus limiting the measurement to a single thickness and making multi-thickness measurement difficult. Additionally, ellipsometers and film thickness gauges calculate film thickness by measuring the characteristics of reflected light. If the film has high absorption, the fitting error increases or fails to fit, leading to decreased accuracy and precision. Therefore, for the thickness measurement of films with high absorption, existing technologies struggle to meet the requirements of real-time, online, multi-probe, and rapid measurement. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention proposes a multi-probe online real-time film thickness measurement method and device. By measuring the transmittance or transmission spectrum of the thin film, a multi-modal collaborative signal analysis algorithm is used to solve problems such as reflection interference, scattering interference, substrate absorption, weak signal, and channel crosstalk during the measurement process. Combined with the multi-wavelength analysis capability and multi-channel parallel signal characteristics of the spectral detection unit, multiple probes are used to achieve online, rapid, and high-precision multi-thickness measurement of films with absorption.

[0005] One objective of this invention is to provide a multi-probe online real-time film thickness measurement method.

[0006] The multi-probe online real-time film thickness measurement method of the present invention includes the following steps:

[0007] 1) The thin film to be tested is placed on a substrate, which is placed on the equipment used to produce the thin film, and the absorption coefficient of the thin film is known;

[0008] 2) The light source emits light of multiple wavelengths, which passes through the incident light path and is obliquely incident on different positions on the surface of the thin film to be tested;

[0009] The different locations of the thin film under test are either different locations on the same thin film surface or locations on multiple different thin film surfaces;

[0010] 3) After being transmitted or reflected by the thin film and substrate, light located at different planar positions is collected by its corresponding probe into the acquisition optical path and becomes narrowband light of different single wavelengths or different center wavelengths. Different planar positions correspond to different wavelengths or center wavelengths.

[0011] 4) After passing through the optical acquisition path, the beams of light of different wavelengths are combined and simultaneously transmitted to a spectral detection unit to obtain the spectral signal with thin film and substrate, and the spectral signal is converted into an electrical signal and input into the computer;

[0012] 5) The computer obtains the transmission spectrum and the transmittance of each wavelength based on the spectral signal; the computer then calculates the thickness at the corresponding position based on the transmittance of each wavelength. During the solution process, the accuracy and precision of the thickness calculation for a single wavelength are improved by using a multi-modal collaborative signal analysis algorithm. Thus, the thickness of multiple different positions of the same thin film or the thickness of multiple different thin films can be obtained simultaneously and accurately through a single spectral detection unit.

[0013] Before measurement, the film to be measured is not located on the substrate. First, measure the spectral signal of the substrate without the film according to steps 2) to 4).

[0014] In step 1), either a reflection method or a transmission method is used. When using the reflection method, the substrate is made of a transparent or opaque material, and the light acquisition path and the incident light path are located on the same side of the film under test. When using the transmission method, the substrate is made of a transparent material, and the light acquisition path and the incident light path are located on opposite sides of the film under test.

[0015] In step 2), the light source is a continuous spectrum light source or a combination of multiple narrowband light sources with different center wavelengths. The incident light path uses optical fiber transmission or free space transmission; optical fiber transmission uses one-to-many optical fibers or multiple optical fibers; free space transmission uses a combination of mirrors, beam splitters, and lenses.

[0016] In step 3), the acquisition optical path uses optical fiber or free space transmission, with the corresponding probes being optical fiber probes or lenses or lens groups, respectively; free space transmission uses a combination of mirrors, beam combiners, and lenses. For continuous spectrum light sources, a narrowband filter with a different center wavelength is set on each branch of the acquisition optical path to obtain narrowband spectra with different center wavelengths.

[0017] In step 4), the spectral detection unit is a spectrometer, a spectrophotometer, or a spectroscopic component; the spectroscopic component includes a grating, a CCD, and a lens.

[0018] In step 5), for the reflection mode, the reflection spectrum is obtained by dividing the spectral signal with both the thin film and the substrate by the spectral signal without the thin film and only the substrate. Since the light signal passes through the thin film twice, the square root of the reflection spectrum yields the transmission spectrum. For the transmission mode, since the light signal passes through the thin film once, the transmission spectrum is directly obtained by dividing the spectral signal with both the thin film and the substrate by the spectral signal without the thin film and only the substrate. Multiple thicknesses at different locations of the thin film are simultaneously measured using a single spectral detection unit. To address the problems of weak signal, substrate absorption, channel crosstalk, reflection interference, and scattering interference during thickness measurement, this invention proposes a multi-modal cooperative signal analysis algorithm, including the following steps:

[0019] a) Lock-in amplification algorithm: First, the light intensity of the light source is modulated to an alternating light intensity at a set frequency f. The frequency of the electrical signal of the alternating light is used as a reference signal and transmitted to the computer. The spectral signal received by the spectral detection unit contains the modulated transmission signal and random noise. The computer extracts the signal with the same frequency as the reference signal according to the reference signal. Thus, the lock-in amplification algorithm extracts only the light signal with the same frequency as the reference signal, obtains the amplified transmission spectrum, suppresses noise at other frequencies, and improves the noise suppression ratio (SNR) by several orders of magnitude. This separates the weak absorption signal from the noise, meeting the signal-to-noise ratio requirements for nanometer-scale thickness measurement.

[0020] b) Peak decomposition algorithm: First, wavelet transform is performed to further denoise the spectrum and retain the detailed features of the transmission spectrum; then, peak decomposition is performed on the transmission spectrum. Assuming that the overlapping peaks are composed of multiple Gaussian peaks, the peak position, half width at half maximum (FWHM), and peak area of ​​each peak are decomposed through nonlinear fitting to obtain the transmittance of each probe, thereby eliminating signal crosstalk between multiple optical paths. This allows for the inverse calculation of the film thickness measured by different optical paths, solving the problem of signal overlap and crosstalk in the same spectral detection unit during multi-optical path measurement, thus improving the accuracy and precision of thickness calculation.

[0021] c) Angle Compensation Algorithm: The measured thickness t' of the thin film is calculated using the formula t' = ln[1 / T(λ)] / α(λ), where α(λ) is the absorption coefficient of the thin film at the position corresponding to wavelength λ, and T(λ) is the transmittance at the position corresponding to wavelength λ. For a smooth thin film surface with no scattered light, the incident angle of the incident light is θ, and the refraction angle of the light in the thin film is γ = arcsin(sinθ / n), where n is the refractive index of the thin film. The path length of light through the thin film is corrected to t' = t 光滑 / cosγ, and finally, the true thickness t of the smooth film on the surface is calculated. 光滑 =t'·cosγ; Thus, an angle compensation algorithm is used to correct the relationship between the path length of light through the thin film and the film thickness, and the spectral detection unit does not collect reflected light to eliminate the interference of direct reflected light from the film surface and the substrate surface on the transmitted light;

[0022] d) AFM-T Correlation Algorithm: Most thin film surfaces have undulations and are not ideally smooth; rough film surfaces introduce additional scattered light, affecting transmitted light; the roughness of the film is measured using atomic force microscopy (AFM), and an empirical formula for the relationship between film roughness and transmittance is established to correct the path length of light through the film; first, the roughness R is obtained by scanning the film surface with AFM. a Then, by calibrating the relationship between "roughness and travel deviation" using experimental data, the travel deviation ΔL = η·R is obtained. a η is the calibration coefficient; the true thickness t of the surface rough film is obtained by correcting the calculation based on the AFM-T correlation algorithm. 粗糙 =t 光滑 –ΔL, to obtain the corrected true thickness at different positions corresponding to each wavelength λ.

[0023] In step (a), the light source emits modulated light with a frequency of f, such as square wave, sawtooth, triangular, trapezoidal, sine, cosine, pulse, or sine derivative waveforms. Alternatively, the light source emits continuous light, which is modulated by a modulator to emit modulated light. The modulator is an acousto-optic modulator or a chopper. The frequency of the electrical signal of the light source or modulator is transmitted to the computer as a reference signal.

[0024] Another objective of this invention is to provide a multi-probe online real-time film thickness measurement device.

[0025] The multi-probe online real-time film thickness measurement device of the present invention includes: a substrate, a light source, an incident optical path, a probe, a collection optical path, a spectral detection unit, and a computer; wherein, the film to be measured is placed on the substrate, the substrate is placed on the equipment for producing the film, and the absorption coefficient of the film is known; the light source emits light including multiple wavelengths, which passes through the incident optical path and is obliquely incident on different positions on the surface of the film to be measured; the different positions on the film to be measured are different positions on the same film surface or positions on multiple different film surfaces; after being transmitted or reflected by the film and the substrate, the light located at different planar positions is collected by its corresponding probe and sent to the collection optical path, and is transformed into different wavelengths. The single-wavelength or narrow-band spectrum is obtained, with different planar positions corresponding to different wavelengths or center wavelengths. After passing through the acquisition optical path, the light beams of different wavelengths are combined and simultaneously transmitted to a spectral detection unit to obtain spectral signals. The spectral signals are then converted into electrical signals and input into a computer. The computer obtains the transmittance of each wavelength and calculates the thickness at the corresponding position based on the transmittance of each wavelength. During the solution process, a multi-modal collaborative signal analysis algorithm is used to improve the accuracy and precision of obtaining the thickness at a single wavelength. Thus, through a single spectral detection unit, the thickness of multiple different positions of the same thin film, or the thickness of multiple different thin films, can be obtained simultaneously and accurately.

[0026] When measuring the transmittance or transmission spectrum of a thin film using a reflection method, the substrate is made of a transparent or opaque material, and the acquisition light path and the incident light path are located on the same side of the thin film under test. When measuring the transmittance or transmission spectrum of a thin film using a transmission method, the substrate is made of a transparent material, and the acquisition light path and the incident light path are located on opposite sides of the thin film under test.

[0027] The light source is a continuous spectrum light source or a combination of multiple narrowband light sources with different center wavelengths. The incident light path uses optical fiber transmission or free space transmission; optical fiber transmission uses one-to-many optical fibers or multiple optical fibers; free space transmission uses a combination of mirrors, beam splitters, and lenses.

[0028] The acquisition optical path employs either fiber optic transmission or free-space transmission, with the corresponding probes being fiber optic probes, lenses, or lens groups, respectively. Free-space transmission utilizes a combination of mirrors, beam combiners, and lenses. For continuous-spectrum light sources, a narrowband filter with a different center wavelength is placed on each branch of the acquisition optical path, thereby obtaining narrowband spectra with different center wavelengths.

[0029] Advantages of this invention:

[0030] This invention combines the multi-wavelength analysis capabilities and multi-channel parallel signal characteristics of a spectral detection unit. It uses different planar positions corresponding to different wavelengths or center wavelengths to obtain the transmittance of each wavelength, and calculates the thickness at the corresponding position based on the transmittance of each wavelength. During the solution process, a multi-modal collaborative signal analysis algorithm improves the accuracy and precision of thickness calculation for a single wavelength. Thus, a single spectral detection unit can simultaneously and accurately obtain the thickness of multiple different positions of the same thin film, or the thickness of multiple different thin films. This invention features a simple device, small relative error, and high accuracy. It achieves simultaneous, high-precision, stable, and rapid measurement of multiple thicknesses of perovskite films using a single spectral detection unit and developed algorithms, successfully solving challenges related to high absorption, compatibility with transparent or non-transparent substrates, rapid response, multiple probes, and online real-time measurement. This invention can be integrated into thin film growth equipment for rapid and high-precision measurement of thin films, enabling closed-loop control of thin film production, accelerating iteration speed, significantly increasing the yield of high-quality thin films, and saving thin film production costs. Attached Figure Description

[0031] Figure 1 This is an overall schematic diagram of the multi-probe online real-time film thickness measurement device of the present invention;

[0032] Figure 2 This is an optical path diagram of an embodiment of the multi-probe online real-time film thickness measurement device of the present invention;

[0033] Figure 3 The transmission spectrum obtained in Embodiment 1 of the multi-probe online real-time film thickness measurement method according to the present invention;

[0034] Figure 4 The curves showing the change of thickness values ​​of three different perovskite films over time, obtained according to Embodiment 1 of the multi-probe online real-time film thickness measurement method of the present invention.

[0035] Figure 5 This is an optical path diagram of Embodiment 2 of the multi-probe online real-time film thickness measurement device of the present invention;

[0036] Figure 6 This is an optical path diagram of Embodiment 3 of the multi-probe online real-time film thickness measurement device of the present invention;

[0037] Figure 7 This is an optical path diagram of Embodiment 4 of the multi-probe online real-time film thickness measurement device of the present invention. Detailed Implementation

[0038] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0039] Example 1

[0040] In this embodiment, the transmission spectrum of the thin film is measured by reflection. The acquisition optical path and the incident optical path are located on the same side of the thin film under test. The light source is a continuous spectrum light source. The incident optical path uses optical fiber transmission. The optical fiber transmission uses a one-to-many fiber splitter. The acquisition optical path uses optical fiber transmission. A narrowband filter with a different center wavelength is set on each branch of the acquisition optical path. The spectral detection unit uses a spectrometer.

[0041] like Figure 1 As shown, the multi-probe online real-time film thickness measurement device of this embodiment includes: a substrate, a light source, an incident light path, a collection light path, a spectral detection unit, and a computer; wherein, the film to be measured is placed on the substrate, the substrate is placed on the equipment for producing the film, and the absorption coefficient of the film is known; as shown... Figure 2 As shown, the frequency f of the light source is transmitted to the computer as a reference signal; the continuous spectrum light source emits light with continuous wavelengths, which passes through a N-splitter fiber (N is a natural number ≥ 2) and is obliquely incident on N different positions 1 to N on the surface of the thin film under test; each acquisition fiber has an optical fiber probe at its inlet; after reflection from the thin film and substrate, the light at the N different planar positions is collected by its corresponding N optical fiber probes and fed into the corresponding N acquisition fibers, and then passes through N narrowband filters a1 to a2 with different center wavelengths. N The light is divided into N narrowband beams with different center wavelengths, and different planar positions correspond to different center wavelengths. The beams of different wavelengths are then combined and simultaneously transmitted to a spectral detection unit. The spectral signal is converted into an electrical signal and input into a computer to obtain the transmittance at different wavelengths. A multimodal cooperative signal analysis algorithm is then used to obtain N different thicknesses t1 to t2. N .

[0042] Before measurement, the spectral signal of the substrate without a thin film is obtained by following steps 2) to 4).

[0043] The multimodal cooperative signal analysis algorithm in this embodiment includes the following steps:

[0044] For the reflection method, the reflection spectrum is obtained by dividing the spectral signal with both thin film and substrate by the spectral signal without thin film but only substrate. Since the light signal passes through the thin film twice, the square root of the reflection spectrum is used to obtain the transmission spectrum.

[0045] a) Lock-in amplification algorithm: The light source emits square wave, sine, or cosine modulated light, or the light source emits continuous light that is modulated by a modulator to emit modulated light. The modulator uses an acousto-optic modulator or a chopper. The frequency of the electrical signal of the light source or modulator is transmitted to the computer as a reference signal. The spectral signal received by the spectral detection unit contains the modulated transmission signal and random noise. The computer extracts the spectral signal with the same frequency as the reference signal according to the reference signal. Thus, the lock-in amplification technology extracts only the light signal with the same frequency as the reference signal, i.e., the modulated transmission spectrum, and suppresses noise of other frequencies (such as detector dark current, light source intensity fluctuations, ambient light interference, and other broadband noise), improving the noise suppression ratio (SNR) by several orders of magnitude. This separates the weak absorption signal from the noise, meeting the signal-to-noise ratio requirements for nanometer-level thickness measurement.

[0046] b) Peak decomposition algorithm: First, wavelet transform is performed, and then Daubechies wavelet is used to denoise the spectrum while preserving the detailed features of the transmission spectrum. Then, peak decomposition is performed. Assuming that the overlapping peaks are composed of multiple Gaussian peaks, the peak position, half width at half maximum (FWHM), and peak area of ​​each peak are decomposed by nonlinear fitting of the Levenberg-Marquardt algorithm to obtain the transmittance of each probe. This eliminates signal crosstalk between multiple optical paths and allows for the inverse calculation of the film thickness measured by different optical paths. This solves the problem of signal overlap and crosstalk in the same spectral detection unit during multi-optical path measurement, thereby improving the accuracy and precision of thickness calculation.

[0047] c) Angle Compensation Algorithm: The measured thickness t' of the thin film is calculated using the formula t' = ln[1 / T(λ)] / α(λ), where α(λ) is the absorption coefficient of the thin film at the position corresponding to wavelength λ. For a smooth surface with no scattered light, the incident angle of the incident light is θ, and the refraction angle of the light in the thin film is γ = arcsin(sinθ / n), where n is the refractive index of the thin film. The path length of the light through the thin film is corrected to t' = t 光滑 / cosγ, and finally, the true thickness t of the smooth film on the surface is calculated. 光滑 =t'·cosγ; Thus, an angle compensation algorithm is used to correct the relationship between the path length of light through the thin film and the film thickness, and the spectral detection unit does not collect reflected light to eliminate the interference of direct reflected light from the film surface and the substrate surface on the transmitted light;

[0048] d) AFM-T Correlation Algorithm: Most thin film surfaces have undulations and are not ideally smooth; rough film surfaces introduce additional scattered light, affecting transmitted light; the roughness of the film is measured using atomic force microscopy (AFM), and an empirical formula for the relationship between film roughness and film transmittance T is established to correct the path length of light passing through the film; first, the roughness R is obtained by scanning the film surface with AFM. aThen, by calibrating the relationship between "roughness and travel deviation" using experimental data, the travel deviation ΔL = η·R is obtained. a η is the calibration coefficient; the true thickness t of the surface rough film is obtained by correcting the calculation based on the AFM-T correlation algorithm. 粗糙 =t 光滑 –ΔL, to obtain the corrected true thickness at different positions corresponding to each wavelength λ.

[0049] Experimentally, perovskite films of varying thicknesses were prepared, and their transmission spectra were measured using a UV-Vis spectrophotometer. Based on these transmission spectra, the absorption coefficient α(λ) at wavelength λ was calculated. The transmission spectra of the perovskite were then measured using the aforementioned algorithm, as shown below. Figure 3 As shown, this perovskite film exhibits strong absorption below wavelength λ = 800 nm, resulting in low transmittance. From Figure 3 The transmittance at wavelength λ = 720 nm can be read as T ≈ 39.1%. Previously, the absorption coefficient of the perovskite film at wavelength λ = 720 nm was found to be α ≈ 0.0039. Based on formula (1) and the multimodal cooperative signal analysis algorithm, the thickness of the perovskite film can be calculated to be 238.7 nm, with a relative error of approximately 2% compared to the profilometer measurement, demonstrating high accuracy.

[0050] The thickness curves of three different perovskite films over time were measured, as shown below. Figure 4 As shown. From Figure 4 It can be seen that the measurement of the thicknesses t1 to t3 of three different perovskite films is very stable, with a measurement accuracy better than 1 nm and a resolution better than 0.1 nm. The measurement accuracy and stability can be further improved by selecting high signal-to-noise ratio, high-sensitivity optoelectronic devices, high-stability light sources, high-stability power supplies and circuits, and high-efficiency optical components. Therefore, this invention uses a spectrometer and developed algorithms to achieve simultaneous, high-precision, high-stability, and rapid measurement of multiple thicknesses of perovskite films, successfully solving the challenges of high absorption, compatibility with transparent / non-transparent substrates, rapid response, multiple probes, and online real-time measurement. This invention can be integrated into thin film growth equipment for rapid and high-precision measurement of thin films, enabling closed-loop control of thin film production, accelerating iteration speed, greatly increasing the yield of high-quality thin film production, and saving thin film production costs.

[0051] Example 2

[0052] In this embodiment, as Figure 5As shown, the transmission spectrum of the thin film is measured using a reflection method; the light source includes the first to the Nth narrowband light source with different center wavelengths; the incident optical path uses N incident optical fibers, each connected to a narrowband light source, and the exit of each optical fiber leads to N different positions; the acquisition optical path uses N acquisition optical fibers, each with an optical fiber probe at its inlet, and the N optical fiber probes correspond to N different positions 1 to N; the ends of the N acquisition optical fibers are bundled together to the spectral detection unit; the spectral detection unit uses a spectrometer.

[0053] Example 3

[0054] In this embodiment, as Figure 6 As shown, the transmission spectrum of the thin film is measured using a reflection method. The light source is a continuous spectrum light source, coupled to the incident light path through a beam splitting system. The incident light path adopts free space transmission, which uses a combination of reflectors, beam splitters, and lenses. The light source emits continuous wavelength light, which is reflected by the first reflector and then split into N beams by the first to Nth beam splitters, and incident on N different positions 1 to N. Each position corresponds to a probe, which is a lens group. The probe collects the signal light through the corresponding lens group and transmits it to the acquisition light path through the first to Nth beam splitters. The acquisition light path adopts a combination of reflectors, beam combiners, beam splitters, and lenses. A narrowband filter with a different center wavelength is set on each acquisition branch. The first to N-1 beam combiners are set at the end of the first to N-1 acquisition branches, and a second reflector is set at the end of the Nth acquisition branch. The beams are combined into one beam, collected by a lens, and sent to the spectral detection unit, thereby obtaining narrowband spectra with different center wavelengths.

[0055] Example 4

[0056] In this embodiment, as Figure 7 As shown, the transmission spectrum of the thin film is measured by reflection. The light source includes the first to the Nth narrowband light source with different center wavelengths. The incident light path adopts free space transmission. Each narrowband light source passes through a corresponding incident branch, and each incident branch uses a combination of mirrors. Each position corresponds to a probe, which is a lens. The light reflected from N different positions is collected by the corresponding probe and transmitted to the acquisition light path. The light is combined into a single path by the first to the N-1th beam combiner and the mirror and transmitted to the spectral detection unit.

[0057] Finally, it should be noted that the purpose of disclosing the embodiments is to help further understand the present invention. However, those skilled in the art will understand that various substitutions and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the content disclosed in the embodiments, and the scope of protection of the present invention is defined by the claims.

Claims

1. A method for online real-time measurement of multiple film thicknesses, characterized in that, The measurement method includes the following steps: 1) The thin film to be tested is placed on a substrate, which is placed on the equipment used to produce the thin film, and the absorption coefficient of the thin film is known; 2) The light source emits light of multiple wavelengths, which passes through the incident light path and is obliquely incident on different positions on the surface of the thin film to be tested; the different positions on the thin film to be tested are different positions on the same thin film surface or positions on multiple different thin film surfaces; 3) After being transmitted or reflected by the thin film and substrate, light located at different planar positions is collected by its corresponding probe into the acquisition optical path and becomes narrowband light of different single wavelengths or different center wavelengths. Different planar positions correspond to different wavelengths or center wavelengths. 4) After passing through the optical acquisition path, the beams of light of different wavelengths are combined and simultaneously transmitted to a spectral detection unit to obtain the spectral signal with thin film and substrate, and the spectral signal is converted into an electrical signal and input into the computer; 5) The computer obtains the transmission spectrum and the transmittance of each wavelength based on the spectral signal; the computer then calculates the thickness at the corresponding position based on the transmittance of each wavelength. During the solution process, the accuracy and precision of the thickness calculation for a single wavelength are improved by using a multi-modal collaborative signal analysis algorithm. Thus, the thickness of multiple different positions of the same thin film or the thickness of multiple different thin films can be obtained simultaneously and accurately through a single spectral detection unit.

2. The measurement method as described in claim 1, characterized in that, In step 1), either a reflection method or a transmission method is used. When using the reflection method, the substrate is made of a transparent or opaque material, and the light acquisition path and the incident light path are located on the same side of the film under test. When using the transmission method, the substrate is made of a transparent material, and the light acquisition path and the incident light path are located on opposite sides of the film under test.

3. The measurement method as described in claim 1, characterized in that, In step 2), the light source is a continuous spectrum light source or a combination of multiple narrowband light sources; the incident light path adopts optical fiber transmission or free space transmission.

4. The measurement method as described in claim 1, characterized in that, In step 5), the multimodal cooperative signal analysis algorithm includes the following steps: a) Lock-in amplification algorithm: First, the light intensity of the light source is modulated to an alternating light intensity at a set frequency f. The frequency of the electrical signal of the alternating light is used as a reference signal and transmitted to the computer. The spectral signal received by the spectral detection unit contains the modulated transmission signal and random noise. The computer extracts the signal with the same frequency as the reference signal according to the reference signal. Thus, the lock-in amplification algorithm extracts only the light signal with the same frequency as the reference signal to obtain the amplified transmission spectrum. b) Peak decomposition algorithm: First, wavelet transform is performed to further denoise the spectrum and retain the detailed features of the transmission spectrum; then, peak decomposition is performed on the transmission spectrum. Assuming that the overlapping peaks are composed of multiple Gaussian peaks, the peak position, half width at half maximum and peak area of ​​each peak are decomposed by nonlinear fitting to obtain the transmittance of each corresponding probe. c) Angle Compensation Algorithm: The measured thickness t' of the thin film is calculated using the formula t' = ln[1 / T(λ)] / α(λ), where α(λ) is the absorption coefficient of the thin film at the position corresponding to wavelength λ, and T(λ) is the transmittance at the position corresponding to wavelength λ. For a smooth thin film surface with no scattered light, the incident angle of the incident light is θ, and the refraction angle of the light in the thin film is γ = arcsin(sinθ / n), where n is the refractive index of the thin film. The angle compensation algorithm is used to correct the relationship between the path length of light through the thin film and the film thickness. The corrected path length of light through the thin film is t' = t 光滑 / cosγ, and finally, the true thickness t of the smooth film on the surface is calculated. 光滑 =t'·cosγ; d) AFM-T Correlation Algorithm: The roughness of the thin film is measured using an atomic force microscope (AFM), and an empirical formula for the relationship between the film roughness and transmittance is established to correct the path length of light passing through the film. First, the roughness R of the thin film surface is obtained by scanning with AFM. a Then, by calibrating the relationship between "roughness and travel deviation" using experimental data, the travel deviation ΔL = η·R is obtained. a η is the calibration coefficient; the true thickness t of the surface rough film is obtained by correcting the calculation based on the AFM-T correlation algorithm. 粗糙 =t 光滑 –ΔL, to obtain the corrected true thickness at different positions corresponding to each wavelength λ.

5. A multi-film thickness online real-time measurement device, characterized in that, The online real-time measurement device for multiple film thicknesses includes: a substrate, a light source, an incident light path, a collection light path, a spectral detection unit, and a computer. The film to be measured is placed on a substrate, which is placed on equipment used to produce the film. The absorption coefficient of the film is known. The light source emits light of multiple wavelengths, which passes through the incident light path and is obliquely incident on different positions on the surface of the film to be measured. These different positions on the film to be measured can be different locations on the same film surface or locations on multiple different film surfaces. After transmission or reflection by the film and substrate, the light located at different planar positions is collected by its corresponding probe into the collection light path, and is then converted into different single wavelengths or narrow wavelengths with different center wavelengths. The light carries different wavelengths or center wavelengths corresponding to different planar positions. After beaming together, the light of different wavelengths is simultaneously transmitted to a spectral detection unit to obtain spectral signals of the thin film and substrate. The spectral signals are then converted into electrical signals and input into a computer. The computer obtains the transmission spectrum and the transmittance of each wavelength based on the spectral signals. The computer then calculates the thickness of the corresponding position based on the transmittance of each wavelength. During the solution process, a multi-modal collaborative signal analysis algorithm is used to improve the accuracy and precision of the thickness calculation for a single wavelength. Thus, a single spectral detection unit can simultaneously and accurately obtain the thickness of the same thin film at multiple different positions, or obtain the thickness of multiple different thin films.

6. The online real-time measurement device for multiple film thicknesses as described in claim 5, characterized in that, When using a reflection method, the substrate is made of a light-transmitting or opaque material; when using a transmission method, the substrate is made of a light-transmitting material.

7. The online real-time measurement device for multiple film thicknesses as described in claim 5, characterized in that, The light source is a continuous spectrum light source or a combination of multiple narrowband light sources.

8. The online real-time measurement device for multiple film thicknesses as described in claim 5, characterized in that, The incident optical path and the acquisition optical path adopt optical fiber transmission or free space transmission.

9. The online real-time measurement device for multiple film thicknesses as described in claim 5, characterized in that, When the light source is a continuous spectrum light source, a narrowband filter with a different center wavelength is set on each branch of the acquisition optical path to obtain narrowband spectra with different center wavelengths.

Images