Metal film plane stress measurement device and method based on picosecond laser ultrasound

Abstract

The application discloses a picosecond laser ultrasonic-based metal film plane stress measuring device and method, and the device comprises a femtosecond laser, an optical isolator, a 1 / 2 wave plate and a first polarization beam splitter; the first polarization beam splitter divides the incident laser into pump light and probe light, the pump light passes through a frequency doubling assembly and an electro-optic modulator, and the probe light passes through an adjustable optical delay line and a polarization control element, and the pump light and the probe light are combined at a dichroic mirror; the method comprises the following steps: changing the optical path difference of the pump light and the probe light; collecting the reflectivity change and averaging multiple times; performing background removal and digital filtering processing to obtain a smooth echo characteristic signal; applying a peak searching algorithm to determine the time delay corresponding to the first echo, and combining the known film thickness to calculate the longitudinal wave speed; calibrating the speed under different known stress conditions; establishing a speed-stress mapping relationship to obtain the in-plane residual stress value of the metal film. The application realizes sub-picosecond time, micron-level space and non-contact film stress detection.

Description

Technical Field

[0001] This invention belongs to the field of metal thin film plane stress measurement technology, specifically relating to a metal thin film plane stress measurement device and method based on picosecond laser ultrasound. Background Technology

[0002] In the manufacturing processes of integrated circuits, microelectromechanical systems (MEMS), and advanced packaging, residual stress inevitably occurs in thin film materials such as metal interconnects, insulating layers, and epitaxial layers. Excessive localized stress can lead to film peeling, void formation, or device performance drift; therefore, accurate measurement of stress is crucial.

[0003] Current mainstream thin-film stress measurement technologies have the following limitations: Firstly, the wafer curvature method: based on the Stoney formula, stress is calculated by measuring the curvature of the entire wafer. However, this method can only provide average stress information at the wafer scale or over a large area, and it is difficult to characterize the local stress in specific micro-regions inside the chip, such as micrometer-level interconnects.

[0004] Secondly, X-ray diffraction: Although it has high precision, the equipment cost is high, the measurement speed is relatively low, and it usually has certain requirements on the crystallinity of the material and the test conditions. It has limitations in applicability for stress measurement of amorphous materials or ultrathin films.

[0005] Picosecond ultrasound, a non-contact characterization technique that uses femtosecond lasers to excite and detect high-frequency sound waves, has been applied in the measurement of parameters such as thin film thickness and sound velocity. However, in existing thickness measurement algorithms, sound velocity is usually treated as a constant, while the physical phenomenon of sound velocity varying with stress (acoustoelastic effect) is often regarded as a disturbing factor and not used for stress inversion.

[0006] To date, there is a lack of publicly available information on industrial solutions for rapid detection of thin film stress with micron-level spatial resolution that directly utilizes the time-of-flight characteristics of picosecond ultrasound technology. Summary of the Invention

[0007] This invention is proposed to solve the problems existing in the prior art, and its purpose is to provide a device and method for measuring planar stress of metal thin films based on picosecond laser ultrasound.

[0008] The technical solution of the present invention is: a metal thin film plane stress measurement device based on picosecond laser ultrasound, comprising a femtosecond laser that emits femtosecond ultrashort pulse laser, an optical isolator arranged in the optical path direction of the femtosecond ultrashort pulse laser, and a half-wave plate through which the laser passes and a first polarization beam splitter arranged sequentially at the output end of the optical isolator. The first polarization beam splitter splits the input laser into pump light and probe light. The pump light, after passing through a frequency doubling component and an electro-optic modulator, and the probe light, after passing through an optical delay line to introduce time delay and polarization control element to adjust polarization state, are combined at a dichroic mirror. The dichroic mirror separates or combines the pump light and probe light.

[0009] Furthermore, the combined beam is focused onto the sample surface by a 10× objective lens.

[0010] Furthermore, the light transmitted by the first polarizing beam splitter is the pump light. The pump light passes through a focusing lens and focuses the beam onto the frequency doubling crystal. The output light after frequency doubling enters the electro-optic modulator.

[0011] Furthermore, a long-wavelength cutoff filter is placed between the electro-optic modulator and the dichroic mirror.

[0012] Furthermore, the light reflected by the first polarizing beam splitter becomes the probe light. After passing through the 3× beam expander, the probe light enters the mechanical delay stage for controllable optical delay.

[0013] Furthermore, the delayed probe light enters the second polarization beam splitter, and one of the output beams from the second polarization beam splitter passes through a quarter-wave plate and enters the dichroic mirror.

[0014] Furthermore, the light beam reflected from the sample surface passes through a 10× objective lens and a dichroic mirror and returns to form the probe light.

[0015] Furthermore, after the probe light passes through a quarter-wave plate and a second polarization beam splitter, it is focused onto the photodetector via a short-wave cutoff filter and a focusing lens.

[0016] A method for measuring planar stress in thin metal films based on picosecond laser ultrasound includes the following steps: A. By changing the optical path difference between the pump light and the probe light, a continuously varying relative time delay is generated; B. The photodetector converts the detection light reflected from the sample into an electrical signal; the lock-in amplifier synchronously detects the electrical signal with the pump light modulation signal as a reference, and outputs a signal corresponding to the change in reflectivity; the computer collects and records the output of the lock-in amplifier to obtain the reflectivity change signal. C. Construct the dynamic curve of transient reflectivity change with time delay; D. Control the mechanical delay stage to perform multiple reciprocating scans to obtain a high signal-to-noise ratio average time-domain dynamic curve; E. The acquired average time-domain dynamic curve is processed to remove background, and the echo characteristic signal is obtained; F. Perform digital filtering on the echo characteristic signal to obtain a smooth characteristic signal; G. Apply the peak-finding algorithm to the smoothed feature signal to obtain the first echo time delay, and calculate the current longitudinal wave velocity of the metal thin film using the known film thickness; H. Repeat step AG for thin films under different known stress conditions to establish the longitudinal wave sound velocity-stress mapping relationship and calibrate the acoustoelastic coefficient; I. Based on the longitudinal wave velocity and the calibrated acoustoelastic coefficient, the in-plane residual stress value of the metal thin film is obtained, and the measurement is completed.

[0017] The beneficial effects of this invention are as follows: This invention innovatively applies picosecond ultrasound technology to the field of thin film stress measurement, providing a method for detecting uniaxial stress in thin films that achieves sub-picosecond time, micrometer-level spatial resolution, non-contact operation, and rapid analysis based on time-domain signals.

[0018] This invention detects the time-of-flight variation of ultrasonic waves propagating along the thickness direction of a thin film, and combines the Poisson effect and acoustoelasticity theory to invert the stress state inside the thin film.

[0019] This invention features high spatial resolution. By utilizing the focusing properties of lasers, the detection spot can be reduced to 1-2 micrometers, thereby enabling the measurement of minute local stresses that cannot be identified by traditional wafer curvature methods.

[0020] This invention taps into the long-neglected sound velocity-stress sensitivity in picosecond ultrasonic technology, expanding the functionality of the device without altering existing thickness measurement hardware, thus achieving multi-purpose functionality.

[0021] This invention only requires analysis of the relative displacement in the time-of-flight (TOF) domain, without the need for complex frequency domain analysis or full tensor inversion, making it suitable for rapid screening in industrial production lines. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the device in this invention; Figure 2 This is a schematic diagram of the heat removal process in this invention; Figure 3 This is a schematic diagram of the SG smoothing filter in this invention; Figure 4 This is a fitting curve of the acoustoelastic coefficient calibration (sound velocity-stress mapping relationship) in this invention; The components include: 1. Femtosecond laser; 2. Optical isolator; 3. First polarization beam splitter; 4. Frequency doubling crystal; 5. Electro-optic modulator; 6. Long-wavelength cutoff filter; 7. Dichroic mirror; 8. 10× objective lens; 9. 3× beam expander; 10. Mechanical delay stage; 11. Second polarization beam splitter; 12. Short-wavelength cutoff filter; 13. Photodetector; 14. Sample; 15. Signal generator; 16. Lock-in amplifier; 17. Computer. Detailed Implementation

[0023] The present invention will now be described in detail with reference to the accompanying drawings and embodiments: like Figures 1 to 4 As shown, the metal thin film plane stress measurement device based on picosecond laser ultrasound includes a femtosecond laser 1 that emits a femtosecond ultrashort pulse laser, an optical isolator 2 arranged in the optical path direction of the femtosecond ultrashort pulse laser, and a half-wave plate through which the laser passes and a first polarization beam splitter 3 arranged sequentially at the output end of the optical isolator 2. The first polarization beam splitter 3 splits the input laser into a pump beam and a probe beam. The pump beam, after passing through a frequency doubling component and an electro-optic modulator, and the probe beam, after being time-delayed by an optical delay line and having its polarization state adjusted by a polarization control element, are combined at a dichroic mirror 7. The dichroic mirror 7 performs color separation or beam combining of the 400 nm pump beam and the 800 nm probe beam.

[0024] The combined beam is focused onto the surface of sample 14 by a 10× objective lens 8.

[0025] The light transmitted by the first polarizing beam splitter 3 is the pump light. The pump light passes through the focusing lens and focuses the beam onto the frequency doubling crystal 4. The output light after frequency doubling enters the electro-optic modulator 5.

[0026] A long-wavelength cutoff filter 6 is also provided between the electro-optic modulator 5 and the dichroic mirror 7.

[0027] The light reflected by the first polarization beam splitter 3 is the probe light. After passing through the 3× beam expander 9, the probe light enters the mechanical delay stage 10 for controllable optical delay.

[0028] The delayed probe light enters the second polarization beam splitter 11, and one output light from the second polarization beam splitter 11 enters the dichroic mirror 7 through a quarter-wave plate.

[0029] The light beam reflected from the surface of sample 14 returns through the 10× objective lens 8 and the dichroic mirror 7 to form the probe light.

[0030] After passing through a quarter-wave plate and a second polarizing beam splitter 11, the probe light passes through a short-wave cutoff filter 12 and a focusing lens, and is focused onto a photodetector 13.

[0031] Specifically, the femtosecond laser 1 (Ti:Sapphire) emits an ultrashort pulse laser with a pulse width of 20 fs and a center wavelength of 800 nm.

[0032] Specifically, the optical isolator 2 prevents the return light from the downstream optical path from interfering with the mode-locking state of the laser, ensuring stable output.

[0033] Specifically, the half-wave plate adjusts the energy ratio of the pump light and the probe light.

[0034] As one implementation method The 800nm ​​pump light passes through a focusing lens and focuses the beam onto the center of the frequency doubling crystal 4. The laser undergoes nonlinear frequency transformation after passing through the frequency doubling crystal 4, doubling the frequency and changing the wavelength to 400nm.

[0035] The 400nm laser is re-collimated and shaped by a lens, and the residual 800nm ​​fundamental frequency light is filtered out.

[0036] The filtered 400nm laser light passes through an electro-optic modulator 5 (EOM) and is modulated into a ~1MHz intensity modulation signal, which provides a reference frequency for the lock-in amplifier to improve the signal-to-noise ratio.

[0037] The modulated pump light is introduced through the reflector and reaches the beam-combining position; the 400nm pump light is reflected by the dichroic mirror 7, combined with the probe light, and fills the entrance pupil of the 10× objective lens 8, and is finally focused on the surface of the metal thin film sample to excite transient acoustic waves.

[0038] As one implementation method The 800nm ​​probe light passes through 3×beam expanders 9 to reduce the divergence angle during long-distance transmission.

[0039] The 800nm ​​probe light passes through the mechanical delay stage 10, producing an optical time delay that is precisely controllable compared to the pump light.

[0040] The delayed probe light enters the second polarization beam splitter 11. After passing through the second polarization beam splitter 11, the beam passes through a quarter-wave plate (QWP), and the polarization state changes from linear polarization to circular polarization.

[0041] The 800nm ​​probe light, which becomes circularly polarized, passes through the dichroic mirror 7, which has high transmittance for 800nm ​​and is coaxial with the pump light. It is then focused onto the surface of the sample 14 by the 10× objective lens 8.

[0042] As one implementation method The probe light (circularly polarized) passes through the quarter-wave plate again. Because it passes through the wave plate twice, the polarization direction of the probe light is rotated by 90° relative to its incident direction. Therefore, it separates from the incident light path at the second polarization beam splitter 11 from the other output port and enters the probe channel.

[0043] Furthermore, the light passes through the second polarization beam splitter 11, the short-wave cutoff filter 12 (which filters out 400nm stray light), and the focusing lens, and is focused onto the photodetector 13.

[0044] A method for measuring planar stress in thin metal films based on picosecond laser ultrasound includes the following steps: A. By changing the optical path difference between the pump light and the probe light, a continuously varying relative time delay is generated; B. The photodetector 13 converts the detection light reflected from the sample into an electrical signal; the lock-in amplifier 16 synchronously detects the electrical signal with the pump light modulation signal as a reference, and outputs a signal corresponding to the change in reflectivity; the computer 17 collects and records the output of the lock-in amplifier 16 to obtain the reflectivity change signal. C. Construct the dynamic curve of transient reflectivity change with time delay; D. Control the mechanical delay stage to perform multiple reciprocating scans to obtain a high signal-to-noise ratio average time-domain dynamic curve; E. The acquired average time-domain dynamic curve is processed to remove background, and the echo characteristic signal is obtained; F. Perform digital filtering on the echo characteristic signal to obtain a smooth characteristic signal; G. Apply the peak-finding algorithm to the smoothed feature signal to obtain the first echo time delay, and calculate the current longitudinal wave velocity of the metal thin film using the known film thickness; H. Repeat step AG for thin films under different known stress conditions to establish the longitudinal wave sound velocity-stress mapping relationship and calibrate the acoustoelastic coefficient; I. Based on the longitudinal wave velocity and the calibrated acoustoelastic coefficient, the in-plane residual stress value of the metal thin film is obtained, and the measurement is completed.

[0045] Specifically, step A changes the optical path difference between the pump light and the probe light, resulting in a continuously varying relative time delay. The specific process is as follows: Computer 17 controls mechanical delay stage 10 to move continuously at a constant speed, changing the optical path difference between pump light and probe light, thereby generating a continuously changing relative time delay.

[0046] The constant speed of the mechanical delay stage 10 can be, but is not limited to, 0.1 mm / s.

[0047] Specifically, in step B, photodetector 13 converts the detection light reflected from the sample into an electrical signal; lock-in amplifier 16 synchronously detects the electrical signal with the pump light modulation signal as a reference, and outputs a signal corresponding to the change in reflectivity; computer 17 collects and records the output of lock-in amplifier 16 to obtain the reflectivity change signal. The specific process is as follows: First, while the mechanical delay stage 10 is moving, the photodetector 13 converts the detection light reflected by the sample into an electrical signal and inputs it into the lock-in amplifier 16. Then, the lock-in amplifier 16 uses the reference frequency provided by the electro-optic modulator 5 to perform synchronous detection of the signal, and the computer 17 collects and records the output of the lock-in amplifier 16 at a set sampling frequency, extracting the reflectivity change signal that is in sync with the pump light modulation frequency. .

[0048] The sampling rate of the lock-in amplifier 16 can be, but is not limited to, 5 Hz, that is, 5 points are collected per second, and can be set according to the scanning speed and signal-to-noise ratio requirements.

[0049] Specifically, step C constructs the dynamic curve of transient reflectivity change with time delay, and the specific process is as follows: Computer 17 synchronously records the real-time position of mechanical delay stage 10 and the output data of lock-in amplifier 16 to construct the transient reflectivity change with time delay. Dynamic curve.

[0050] The above curve is a composite signal, superimposed with the thermal background signal caused by electron-phonon coupling and thermal diffusion, and the picosecond ultrasonic echo characteristic signal caused by stress wave propagation.

[0051] Specifically, step D involves controlling the mechanical delay stage to perform multiple reciprocating scans to obtain a high signal-to-noise ratio average time-domain dynamic curve. The specific process is as follows: To suppress system random noise and improve the signal-to-noise ratio (SNR), the mechanical delay stage 10 is controlled to perform multiple reciprocating scans, for example, 10-20 times. The computer 17 aligns and superimposes the time-domain curves acquired multiple times, and outputs a high signal-to-noise ratio average time-domain dynamic curve.

[0052] Specifically, step E involves background removal processing of the acquired average time-domain dynamic curve to obtain the echo characteristic signal. The specific process is as follows: First, extract the data. Data segments with a time delay of 20 ps or more were selected for analysis to avoid... The interference of coherent artifacts near the time point and the initial non-equilibrium electronic excited state on the signal; Then, background fitting A double exponential decay function was used to fit the truncated thermal background signal to accurately describe the electronic thermal relaxation process. Finally, signal extraction The echo characteristic signal is extracted by subtracting the fitted double-exponential thermal background curve from the original data. Figure 2The original echo signal shown is accompanied by high-frequency noise. This signal characterizes the reflection or acoustic resonance properties of the acoustic strain pulse along the thickness direction of the metal thin film.

[0053] Specifically, step F involves digitally filtering the echo characteristic signal to obtain a smooth characteristic signal. The specific process is as follows: First, algorithm selection A Savitzky-Golay (SG) smoothing filter is used; Then, parameter optimization. Smoothing can be achieved by adjusting the sliding window size of the filter and the order of the polynomial fitting, or by performing multiple iterations. More specifically, Gaussian-like function fitting In the obtained smooth feature signal, a local time window containing the target echo peak (e.g., the first echo) is selected, and the local echo peak is fitted with a Gaussian function model to further suppress residual noise and improve the extraction accuracy of echo time of flight (TOF). The Gaussian fitting can be solved using methods such as least squares, and the center position of the fitted peak is used as the time delay parameter corresponding to the echo. Finally, the processing results While effectively filtering out the spikes and high-frequency noise superimposed on the echo, it preserves the original shape, height, and time position of the peak to the greatest extent (without phase shift), obtaining a smooth characteristic signal, which provides a high-quality data foundation for subsequent high-precision peak finding.

[0054] Specifically, step G applies a peak-finding algorithm to the smoothed feature signal to calculate the current longitudinal wave velocity of the metal thin film. The specific process is as follows: First, peak identification Search and lock The first significant characteristic peak that appears afterward corresponds to the echo signal of the sound wave after completing its first round trip in the thickness direction of the thin film. Then, the flight time is determined. Accurately extract the time delay value corresponding to the peak of the wave. It is defined as the round-trip time of flight (TOF) of the sound wave within the thin film. Finally, the speed of sound calculation Combined with known metal thin film thickness According to the formula Calculate the current longitudinal wave velocity of the metal thin film.

[0055] Specifically, step H repeats step AG for thin films under different known stress conditions to establish the longitudinal wave sound velocity-stress mapping relationship and calibrate the acoustoelastic coefficient. The specific process is as follows: First, construct the acoustic elasticity equation. Although the measurement focuses on the longitudinal wave velocity along the thickness direction, based on the Poisson effect, in-plane stress leads to Poisson strain along the thickness direction, which in turn alters the longitudinal wave velocity through the third-order elastic constant. A linear coupling model is then established based on this. ; in, , To measure the speed of sound, The zero-stress reference speed of sound The equivalent acoustic elastic coefficient, This refers to in-plane residual stress; Then, determine the reference speed of sound ( ) A standard sample of the same material as the sample to be tested and which has undergone sufficient annealing and stress relief treatment was selected, and its longitudinal wave velocity was measured as the zero-stress reference value. ; Next, the acoustic elasticity coefficient was calibrated. ) In-situ calibration was performed using a precision uniaxial compression loading device. Then, multi-gradient samples Prepare at least three standard thin film samples of the same material with different thicknesses (e.g., 125 nm, 170 nm, 230 nm); Finally, linear fitting Known uniaxial compressive stresses were applied to each sample in stages, and the corresponding changes in sound velocity were recorded. All data were then compiled, and a global linear regression was performed with stress on the x-axis and the relative rate of change of sound velocity on the y-axis. The resulting slope is the high-precision equivalent acoustoelastic coefficient. .

[0056] Within the tested stress range, the relative rate of change of longitudinal wave sound velocity exhibits extremely high linear significance with respect to uniaxial stress. Statistical analysis shows that the R-squared value of the linear regression is... 2 All values ​​were greater than 0.9, and the P-values ​​were all much less than 0.001, which fully verified the high signal-to-noise ratio of the measurement method and the accuracy of the acoustic elasticity model.

[0057] Specifically, step I, based on the longitudinal wave velocity and calibration, obtains the in-plane residual stress value of the metal thin film, completing the measurement. The specific process is as follows: First, calculate the current measured speed of sound. Substitute these parameters into the defined physical model and calibration parameters; Then, calculate the relative rate of change of sound speed. ; Finally, using the formula The in-plane residual stress value of the metal thin film is obtained by inverse solving (positive value represents tensile stress, negative value represents compressive stress), and the measurement is completed.

[0058] This invention innovatively applies picosecond ultrasound technology to the field of thin film stress measurement, providing a method for detecting uniaxial stress in thin films that achieves sub-picosecond temporal resolution, micrometer-level spatial resolution, non-contact operation, and rapid analysis based on time-domain signals.

[0059] This invention detects the time-of-flight variation of ultrasonic waves propagating along the thickness direction of a thin film, and combines the Poisson effect and acoustoelasticity theory to invert the stress state inside the thin film.

[0060] This invention features high time resolution. It utilizes a femtosecond ultrashort pulse laser pump-probe time-delay scanning method to accurately measure the ultrasonic echo time of flight (TOF) in the thickness direction of thin films, achieving sub-picosecond time resolution. This enables the resolution of minute changes in sound velocity and time of flight caused by stress variations, making it suitable for rapid inversion of residual stress within the thin film plane.

[0061] This invention features high spatial resolution. By utilizing the focusing properties of lasers, the detection spot can be reduced to 1-2 micrometers, thereby enabling the measurement of minute local stresses that cannot be identified by traditional wafer curvature methods.

[0062] This invention taps into the long-neglected sound velocity-stress sensitivity in picosecond ultrasonic technology, expanding the functionality of the device without altering existing thickness measurement hardware, thus achieving multi-purpose functionality.

[0063] This invention only requires analysis of the relative displacement in the time-of-flight (TOF) domain, without the need for complex frequency domain analysis or full tensor inversion, making it suitable for rapid screening in industrial production lines.

Claims

1. A metal thin film plane stress measurement device based on picosecond laser ultrasound, comprising a femtosecond laser (1) emitting femtosecond ultrashort pulse laser, characterized in that: An optical isolator (2) is provided in the optical path direction of the femtosecond ultrashort pulse laser. A half-wave plate through which the laser passes and a first polarization beam splitter (3) are arranged sequentially at the output end of the optical isolator (2). The first polarization beam splitter (3) splits the incident laser into pump light and probe light. The pump light, which passes through the frequency doubling component and the electro-optic modulator, and the probe light, which has been time-delayed by the optical delay line and adjusted by the polarization control element, are combined at the dichroic mirror (7). The dichroic mirror (7) separates or combines the pump light and the probe light.

2. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 1, characterized in that: The combined beam is focused onto the surface of the sample (14) by the 10× objective lens (8).

3. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 2, characterized in that: The light transmitted by the first polarization beam splitter (3) is the pump light. The pump light passes through the focusing lens and focuses the beam onto the frequency doubling crystal (4). The output light after frequency doubling enters the electro-optic modulator (5).

4. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 3, characterized in that: A long-wavelength cutoff filter (6) is also provided between the electro-optic modulator (5) and the dichroic mirror (7).

5. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 2, characterized in that: The light reflected by the first polarization beam splitter (3) is the probe light. After passing through the 3× beam expander (9), the probe light enters the mechanical delay stage (10) for controllable optical delay.

6. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 5, characterized in that: The delayed probe light is injected into the second polarization beam splitter (11), and one output light of the second polarization beam splitter (11) enters the dichroic mirror (7) through the quarter-wave plate.

7. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 6, characterized in that: The light beam reflected from the surface of the sample (14) returns through the 10× objective lens (8) and the dichroic mirror (7) to form the probe light.

8. The metal thin film plane stress measurement device based on picosecond laser ultrasound according to claim 7, characterized in that: After passing through a quarter-wave plate and a second polarization beam splitter (11), the probe light passes through a short-wave cutoff filter (12) and a focusing lens, and is focused onto a photodetector (13).

9. The method for measuring planar stress of a metal thin film based on picosecond laser ultrasound according to claim 1, characterized in that: Includes the following steps: A. By changing the optical path difference between the pump light and the probe light, a continuously varying relative time delay is generated; B. The photodetector (13) converts the detection light reflected by the sample into an electrical signal; the lock-in amplifier (16) synchronously detects the electrical signal with the pump light modulation signal as a reference and outputs a signal corresponding to the change in reflectivity; the computer (17) collects and records the output of the lock-in amplifier (16) to obtain the reflectivity change signal. C. Construct the dynamic curve of transient reflectivity change with time delay; D. Control the mechanical delay stage to perform multiple reciprocating scans to obtain a high signal-to-noise ratio average time-domain dynamic curve; E. The acquired average time-domain dynamic curve is processed to remove background, and the echo characteristic signal is obtained; F. Perform digital filtering on the echo characteristic signal to obtain a smooth characteristic signal; G. Apply the peak-finding algorithm to the smoothed feature signal to obtain the first echo time delay, and calculate the current longitudinal wave velocity of the metal thin film using the known film thickness; H. Repeat step AG for thin films under different known stress conditions to establish the longitudinal wave sound velocity-stress mapping relationship and calibrate the acoustoelastic coefficient; I. Based on the longitudinal wave velocity and the calibrated acoustoelastic coefficient, the in-plane residual stress value of the metal thin film is obtained, and the measurement is completed.

Images