Intelligent modulation and protection system and method of laser and storage medium

Abstract

The invention relates to the technical field of excimer laser annealing equipment, and discloses an intelligent modulation and protection system and method for a laser and a storage medium, and the system comprises a timing unit which is used for monitoring the non-discharge shutdown time in real time; the voltage compensation module is used for presetting a time length and charging voltage attenuation characteristic correlation model; the pre-modulation execution unit is used for determining a first pulse target voltage compensation parameter according to the duration and the correlation model and injecting the first pulse target voltage compensation parameter into a high-voltage power supply charging reference; the current waveform sampling module is used for extracting the dynamic impedance characteristic quantity of the falling edge of the discharge current; the flow field stability judgment unit is used for calculating a characteristic quantity fluctuation variable coefficient; the pulse time sequence scheduling unit is used for adjusting the trigger period according to the variable coefficient to damage the acoustic oscillation coherence superposition state in the cavity, online monitoring of the uniformity of the flow field is achieved by multiplexing discharge loop characteristics, then the first pulse energy step is eliminated, and local hot spot damage, caused by discharge filamentation, of an optical assembly is avoided.

Description

Technical Field

[0001] This invention relates to an intelligent modulation and protection system, method, and storage medium for lasers, belonging to the technical field of excimer laser annealing equipment. Background Technology

[0002] Current excimer laser annealing processes utilize high-energy-density ultraviolet pulses to induce instantaneous phase transitions in amorphous silicon. The laser operates in burst mode on the production line to handle downtime intervals between adjacent substrates. During downtime, the ionization of the gaseous medium within the laser cavity undergoes physical attenuation, and thermal convection tends to stagnate.

[0003] External optical path hardware optimization improves pulse energy distribution, but it struggles to handle the dynamic evolution of the cavity's physical state during process intervals. For example, Chinese invention patent CN110265289B discloses an excimer laser annealing device and a method for preparing polycrystalline silicon thin films. It incorporates a pulse extension module with a beam splitter and transmission components in the optical path, utilizing a double-peak pulse generated by the physical optical path difference to improve crystallization and broaden the process window. However, this approach is a static delay technique; the beam splitting ratio and delay parameters are fixed after hardware assembly and adjustment. It cannot detect the evolution of dielectric ionization during the shutdown quiet period and cannot address the initial burst of the burst. Pulse energy step execution predictive compensation; conventional technology uses feedback adjustment to stabilize pulse energy. Due to the loss of physical memory of the laser cavity when the burst packet is turned on, the energy of the first pulse deviates from the set reference, resulting in an energy step. Feedback adjustment relies on the deviation of the preceding pulse as an adjustment reference. The first pulse lacks a source of deviation, and the adjustment mechanism produces a response delay. Usually, pre-discharge or discarding the first pulse is used to maintain energy stability, which reduces production efficiency. The instantaneous high energy impact of the first pulse increases the cumulative thermal fatigue of optical components and induces the growth of microcracks in the coating. As the equipment service time increases, gas aging and electrode wear cause the preset compensation mapping relationship to drift, resulting in a decrease in energy consistency.

[0004] Therefore, how to utilize the physical state information during laser shutdown to achieve energy pre-prediction compensation and construct collaborative protection logic for optical components and high-voltage power supply has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background section, the technical solution of this invention is as follows:

[0006] A smart modulation and protection system for a laser, used in an excimer laser annealing apparatus, comprising:

[0007] A timing unit is used to monitor the non-discharge shutdown time of the laser in real time;

[0008] The voltage compensation module is used to preset the correlation model between the non-discharge shutdown time and the charging voltage decay characteristics;

[0009] The pre-modulation execution unit, connected to the timing unit and the voltage compensation module, is used to determine the target voltage compensation parameters for the first pulse based on the non-discharge shutdown duration and the correlation model when a burst start command is received, and to inject the target voltage compensation parameters into the charging reference of the high-voltage power supply.

[0010] The current waveform sampling module is used to collect the current falling edge data of the high voltage power supply in real time during each pulse discharge process, and extract the rate of change of the discharge current from the peak point to the preset current ratio point as a dynamic impedance characteristic quantity.

[0011] The flow field stability determination unit is connected to the current waveform sampling module and is used to calculate the fluctuation variation coefficient of the dynamic impedance characteristic between adjacent pulses within the burst packet. The fluctuation variation coefficient is the ratio of the standard deviation to the mean of the dynamic impedance characteristic within the sliding time window.

[0012] The pulse timing scheduling unit, connected to the flow field stability determination unit, is used to change the triggering period of subsequent pulses to disrupt the coherent superposition state of acoustic oscillations in the laser cavity when the fluctuation variation coefficient exceeds the preset instability threshold, thereby suppressing the distortion of laser energy distribution in the spatial dimension and avoiding local hot spot damage to optical components caused by discharge channel filamentation.

[0013] Preferably, it also includes a feedback calibration module, which is connected to the pre-modulation execution unit and is used to collect the start-up delay parameters during the first pulse discharge process. The feedback calibration module generates a voltage gain correction vector for subsequent pulses in the current burst packet based on the deviation of the start-up delay parameters from the standard start-up threshold, and adjusts the output gain of the pre-modulation execution unit to correct the correlation model. The standard start-up threshold is determined by the following calibration conditions: with the laser cavity filled with a reference working gas at a preset pressure and in thermal equilibrium, the average start-up delay time of the laser during continuous discharge at the reference charging voltage is recorded.

[0014] Preferably, it also includes a waveform adaptive adjustment module for determining the thermal stress load level of the optical component based on the non-discharge shutdown duration; the waveform adaptive adjustment module adjusts the pre-ionization triggering sequence of the initial pulse in the burst packet and the synchronization delay interval between it and the main discharge triggering sequence according to the thermal stress load level, thereby reducing the instantaneous peak power acting on the optical component by widening the rising edge slope of the initial pulse.

[0015] Preferably, it also includes an electrode loss inversion module and a curve correction unit; the electrode loss inversion module is used to acquire the frequency of the residual oscillation waveform generated by the high-voltage power supply at the end of the discharge cycle; the curve correction unit is connected to the voltage compensation module and the electrode loss inversion module, and is used to adjust the calculation slope of the target voltage compensation parameter according to the electrode loss state determined by the offset of the residual oscillation waveform frequency relative to the preset resonance reference.

[0016] Preferably, it also includes a power behavior analysis unit and a thermal load scheduler; the power behavior analysis unit is used to obtain the second derivative of the voltage rise waveform when the high-voltage power supply charges the energy storage component as curvature; the thermal load scheduler is connected to the power behavior analysis unit and the pulse timing scheduling unit, and is used to reduce the transient thermal load of the high-voltage power supply by extending the trigger interval of subsequent pulses when the curvature indicates that the temperature rise of the high-voltage power supply exceeds the heat dissipation threshold of the power device.

[0017] Preferably, the fluctuation variation coefficient C is determined by the following formula: C=σ / μ, where σ is the standard deviation of the dynamic impedance characteristic within the sliding time window, and μ is the mean of the dynamic impedance characteristic within the sliding time window.

[0018] Preferably, the pulse timing scheduling unit introduces non-equal time windows to guide the turbulent structure in the laser cavity to perform thermal equilibrium relaxation, so that the energy density component of the pulse energy fluctuation residual sequence in the burst packet is lower than the preset judgment threshold in the acoustic coherence frequency band. The pre-modulation execution unit is also connected to a voltage physical limiter, which is used to limit the target voltage compensation parameter within the safe discharge range of the laser cavity.

[0019] Preferably, it also includes an accumulated thermal load calculation module, which is used to perform time-dimensional integral calculation on the energy density of the continuously output pulses to determine the equivalent thermal accumulation state of the optical components; the pulse timing scheduling unit adjusts the energy envelope distribution of the laser pulse according to the weight value of the product of the equivalent thermal accumulation state and the fluctuation variation coefficient; the system synchronously performs open-loop voltage pre-compensation for the first pulse and impedance monitoring for the uniformity of the cavity flow field by reusing the discharge circuit current signal of the high-voltage power supply.

[0020] A method for intelligent modulation and protection of a laser includes the following steps:

[0021] Step 101: Monitor the non-discharge shutdown time of the laser in real time;

[0022] Step 102: Preset the correlation model between non-discharge shutdown time and charging voltage decay characteristics;

[0023] Step 103: Upon receiving the burst start command, determine the target voltage compensation parameters for the first pulse based on the non-discharge shutdown duration and the correlation model, and inject the target voltage compensation parameters into the charging reference of the high-voltage power supply.

[0024] Step 104: Real-time acquisition of current falling edge data of high voltage power supply during each pulse discharge process, and extraction of the rate of change of discharge current from peak point to preset current ratio point over time, as a dynamic impedance characteristic quantity.

[0025] Step 105: Calculate the fluctuation variation coefficient of the dynamic impedance characteristic between adjacent pulses within the burst packet. The fluctuation variation coefficient is the ratio of the standard deviation to the mean of the dynamic impedance characteristic within the sliding time window.

[0026] Step 106: When the fluctuation variation coefficient exceeds the preset instability threshold, the triggering period of subsequent pulses is changed to disrupt the coherent superposition state of acoustic oscillations in the laser cavity, thereby suppressing the laser energy distribution distortion in the spatial dimension and avoiding local hot spot damage to optical components caused by discharge channel filamentation.

[0027] A smart modulation and protection storage medium for a laser is provided for implementing a smart modulation and protection method for a laser. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of a smart modulation and protection method for a laser.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. In the intelligent modulation of the laser, the non-discharge shutdown time of the laser is monitored in real time and the degree of cavity plasma reset is mapped. When the burst start command is received, the target voltage compensation parameter of the first pulse is determined according to the duration and the high voltage power supply charging reference is injected, so that the laser output changes from oscillation convergence to instantaneous steady state, eliminating the energy step of the first pulse caused by the loss of cavity physical memory.

[0030] 2. The equivalent thermal accumulation state of the optical component is determined by the integral calculation of the pulse energy density of continuous output. The cold brittleness level of the optical component is identified by combining the non-discharge shutdown time. By adjusting the synchronous delay interval between the pre-ionization triggering sequence and the main discharge triggering sequence to widen the pulse rising edge slope, the instantaneous peak power acting on the optical component is reduced without changing the total pulse energy, thus alleviating the transient thermal stress gradient impact on the optical coating.

[0031] 3. Collect the residual oscillation waveform characteristics generated at the end of the discharge cycle and deconstruct the inherent resonant frequency. Map the physical loss state of the electrode surface in reverse through the resonant frequency offset. Calculate the gain by dynamically adjusting the voltage compensation parameters according to the loss state, so that the compensation logic has the ability to self-heal against hardware aging and ensure that the device maintains consistent energy stability throughout the entire electrode life cycle. Attached Figure Description

[0032] Figure 1 This is a block diagram illustrating the overall architecture and logical interaction of the intelligent modulation and protection system of the present invention.

[0033] Figure 2 This is a comparison chart of the charging voltage rise characteristics of the high-voltage power supply under different thermal states of the present invention;

[0034] Figure 3 This is a mapping diagram of the causes of laser output instability and the control and protection logic of the present invention. Detailed Implementation

[0035] The technical solution of the present invention will be described in detail below through specific embodiments. The content disclosed in this part is intended to explain the present invention and is not intended to limit the scope of protection of the present invention.

[0036] This invention provides an intelligent modulation and protection system for a laser, comprising a timing unit, a voltage compensation module, a pre-modulation execution unit, a current waveform sampling module, a flow field stability determination unit, and a pulse timing scheduling unit. The timing unit monitors the non-discharge shutdown duration of the laser in real time. The voltage compensation module presets a correlation model between the non-discharge shutdown duration and the charging voltage attenuation characteristics. The pre-modulation execution unit connects the timing unit and the voltage compensation module and determines the target voltage compensation parameters for the first pulse upon receiving a burst-start command. The current waveform sampling module collects and extracts current falling edge data during the discharge process in real time. The dynamic impedance characteristic quantity is calculated using a flow field stability determination unit connected to a current waveform sampling module. This unit calculates the fluctuation variation coefficient of the dynamic impedance characteristic quantity. A pulse timing scheduling unit, connected to the flow field stability determination unit, changes the trigger period of subsequent pulses when the fluctuation variation coefficient exceeds the instability threshold. In excimer laser annealing, the laser typically operates in burst mode. The downtime between adjacent substrates causes spontaneous decay of the ionization degree of the gaseous medium within the laser cavity. When the next burst is initiated, the first pulse experiences an energy step due to the change in the cavity's physical state. To address this challenge, a timing unit monitors the laser's non-discharge downtime t in real time. off The timing unit consists of a counter with a sampling frequency of not less than 100MHz, used to record the time interval between the end signal of the previous burst packet and the current burst start command; the voltage compensation module stores a preset correlation model, which is used to describe the non-discharge shutdown duration t. off The functional relationship between the charging voltage attenuation and the pre-modulation execution unit; within a sub-microsecond time window after receiving the burst start command, the pre-modulation execution unit determines the current non-discharge shutdown duration t based on the current non-discharge shutdown duration t. offThe target voltage compensation parameter ΔV for the first pulse is calculated using the correlation model, and this parameter ΔV is injected into the charging reference of the high-voltage power supply; when the non-discharge shutdown time t off At 2.0s, the pre-modulation execution unit calculates that the voltage compensation required for the first pulse is 150V. The high-voltage power supply then adds the 150V bias voltage to the standard charging voltage. The system compensates for the energy decay during the shutdown silence period in advance, so that the laser can reach steady-state output at the moment the burst packet is turned on, thus suppressing the energy step of the first pulse.

[0037] During high-frequency burst operation, uneven heating of the gas inside the laser cavity can easily induce coherent superposition of acoustic oscillations, leading to filamentation of the discharge channel and distortion of the laser energy distribution in the spatial dimension, increasing the probability of local hot spot damage to optical components. Based on this, the current waveform sampling module collects the current falling edge data of the high-voltage power supply during each pulse discharge process in real time, and extracts the discharge current from the peak point I... p The current drops to the preset proportional point, i.e., 10%I. p The rate of change of the dynamic impedance characteristic S with time is used as the dynamic impedance characteristic quantity S; the flow field stability determination unit calculates the fluctuation variation coefficient C of the dynamic impedance characteristic quantity S within the sliding time window, which includes 10 consecutive pulse signals; the fluctuation variation coefficient C is determined by the following formula: C=σ / μ, where C is the fluctuation variation coefficient, σ is the standard deviation of the dynamic impedance characteristic quantity S within the sliding time window, and μ is the dynamic impedance characteristic quantity S within the sliding time window. The mean value; the current waveform sampling module extracts the falling edge data of the high-voltage power supply discharge circuit current at a sampling rate of 2GS / s, and extracts the discharge current from the peak point I. p Decreased to 10% p The rate of change yields the dynamic impedance characteristic quantities S and I. p S is the peak current point, and S is the dynamic impedance characteristic quantity. The flow field stability determination unit calculates the standard deviation σ and mean μ of 10 consecutive pulses S within the sliding time window, and obtains the fluctuation variation coefficient C=σ / μ, where σ is the standard deviation, μ is the mean, and C is the fluctuation variation coefficient. When C exceeds the preset instability threshold of 0.05, the pulse timing scheduling unit switches the constant trigger period of 166μs to a non-equal interval sequence of 160μs and 175μs, which disrupts the coherent superposition state of acoustic oscillation in the laser cavity and suppresses the distortion of laser energy distribution.

[0038] When the fluctuation variation coefficient C exceeds the preset instability threshold of 0.05, the pulse timing scheduling unit intervenes in the triggering time of subsequent pulses. By introducing non-equidistant time windows, the pulse timing scheduling unit adjusts the originally constant 166μs pulse interval into a non-equidistant sequence, thereby disrupting the coherent superposition state of acoustic oscillations within the laser cavity. This allows the turbulent structure within the cavity to undergo thermal equilibrium relaxation, avoiding localized hot spot damage to optical components caused by discharge channel filamentation. To address model mismatch caused by electrode loss or gas component drift within the laser cavity, the system also includes a feedback calibration module, connected to the pre-modulation execution unit, used to acquire the ignition delay during the first pulse discharge process. The ignition delay parameter τ is the time difference between the trigger command issuance and the formation of an effective plasma current in the cavity. The feedback calibration module calculates the deviation of the ignition delay parameter τ from the standard ignition threshold and generates a voltage gain correction vector for subsequent pulses within the current burst packet. This vector is used to adjust the output gain of the pre-modulation execution unit and correct the correlation model. By using the first pulse as a probe to sense the instantaneous ionization activity of the cavity, the system achieves online adaptive compensation for environmental aging and hardware wear. Furthermore, to address the stress impact problem easily caused by the optical components being in a state of cooling and contraction after long-term shutdown, the waveform adaptive adjustment module adjusts the waveform based on the non-discharge shutdown duration t. off The thermal stress load level of the optical components is determined; the waveform adaptive adjustment module adjusts the synchronization delay interval Δt between the pre-ionization triggering sequence and the main discharge triggering sequence of the initial pulse in the burst packet according to the thermal stress load level; with the non-discharge shutdown time t off With the increase in [energy / energy], the system actively increases the synchronization delay interval Δt, reducing the instantaneous peak power acting on the optical components by widening the rising edge slope of the initial pulse. This timing-dimensional waveform shaping process smooths the thermal stress gradient without changing the total pulse energy, extending the service life of the optical components. The waveform adaptive adjustment module adjusts the non-discharge shutdown time t according to [the specific parameters]. off The thermal stress bearing capacity of optical components is classified into three states: thermal saturation, transitional stress, and cold brittleness. off This refers to the non-discharge shutdown duration; when t off When the time interval is 0s to 0.5s, the synchronization delay interval Δt is maintained at a baseline value of 80ns, where Δt is the synchronization delay interval; when t off From 0.5s to 5.0s, the slope changes with t at a rate of 6ns / s. off Increase the width Δt; when t off When the time exceeds 5.0s, it is determined to be in a cold brittle state. The Δt is clamped at 110ns. The timing interval between the pre-ionization pulse and the main discharge pulse is adjusted to reduce the power density of the rising edge of the initial pulse, thereby avoiding the transient thermal stress gradient impact on the optical components after long-term shutdown.

[0039] In terms of system hardware protection, the electrode loss inversion module acquires the frequency f of the residual oscillation waveform generated by the high-voltage power supply at the end of the discharge cycle. r The curve correction unit is based on the residual oscillation waveform frequency f. r The calculation slope of the target voltage compensation parameter ΔV is adjusted based on the electrode loss state determined by the offset from the preset resonant reference. Simultaneously, the power behavior analysis unit acquires the curvature of the voltage rise waveform when the high-voltage power supply charges the energy storage capacitor, i.e., the second derivative κ of this waveform. When the value of the second derivative κ indicates that the driving capability of the high-voltage power supply's power devices decreases due to temperature rise, the thermal load scheduler connects to the pulse timing scheduling unit to reduce the transient thermal load of the high-voltage power supply by extending the subsequent pulse trigger interval, achieving flexible dimensionality reduction protection for the high-voltage power supply components. During the initialization phase, the discharge electrode gap d is adjusted, and the residual oscillation frequency f after the high-voltage circuit discharge is turned off is recorded. r The frequency offset Δf was established by calibrating in 0.1mm steps within the range of d from 1.0mm to 2.5mm. r Linear mapping table with equivalent electrode loss depth δ, where d is the discharge electrode gap, f r The residual oscillation frequency is Δf r δ represents the frequency offset, and δ represents the equivalent loss depth; the curve correction unit acquires f in real time. r And calculate the resonant reference frequency f. base Deviation value, f base Using the reference frequency, the gain is calculated by extracting the target voltage compensation parameter ΔV from the linear mapping table. ΔV is the target voltage compensation parameter. The voltage compensation logic migrates with the hardware loss state to maintain the energy stability of the laser throughout its entire life cycle.

[0040] In this embodiment of the invention, the acoustic coherence frequency band is determined based on the coupling relationship between the physical dimensions of the laser cavity and the sound velocity of the working gas. The system is based on the lateral span of the laser cavity discharge region. And the speed of sound v of the working gas under the current operating conditions s Calculate the first or multiple natural resonant frequencies f n =n⋅v s / (2L), In this embodiment, the acoustic coherence frequency band is preset to a bandwidth range of ±15% centered on the inherent resonant frequency; Regarding the spectral analysis of the energy density component, the system takes the energy residual sequence of N consecutive pulses within the burst packet as input, and the sampling length N is preferably 256 or 512 pulse points to meet the baseline resolution requirements. Before performing the discrete Fourier transform, the system applies a Hanning window to the sequence for weighting processing to suppress spectral leakage and improve the main lobe identification accuracy; The integration method adopts the power spectral density (PSD) estimation method, and the pulse timing scheduling unit performs the following... The integral values ​​of the PSD curve within the acoustically coherent frequency band are accumulated to calculate the energy density contribution of that specific frequency band. When the ratio of this contribution to the total power of the entire frequency band exceeds a preset judgment threshold, the cavity is in an acoustic positive feedback state. To achieve precise destruction, the pulse timing scheduling unit introduces a non-equal interval sequence, and its discrete characteristic frequency is designed to actively avoid the aforementioned acoustically coherent frequency band. This engineering configuration not only guides the thermal equilibrium relaxation of turbulence, but also eliminates the inducement of acoustic coherent superposition from the dimension of energy spectrum distribution, thereby solving the problems of discharge channel filamentation and hot spot damage to optical components.

[0041] In specific implementation scenarios, the pulse timing scheduling unit performs closed-loop dynamic modulation of the energy envelope, and the system pre-calibrates the maximum heat load threshold H of the optical components through experiments. limit and the flow field instability threshold C th This is used to normalize real-time parameters. The specific mathematical expression for the product weight value W is: W = (H / H limit )⋅(C / C th H is the equivalent thermal accumulation state obtained by integrating the current pulse sequence over time, and C is the real-time monitored fluctuation variation coefficient. Regarding the weight normalization: the system performs a weighted product of the normalized H ratio and the C ratio, limiting the value of W to the range of [0,1]. When W approaches 1, it indicates that the optical component is in a critical damage state under the dual coupling effect of thermal accumulation and flow field instability. The corresponding envelope adjustment method is as follows: the pulse timing scheduling unit dynamically reshapes the energy distribution shape of the burst packet according to the value of W, specifically by adjusting the output gain G of the pre-modulation execution unit. The adjustment logic follows the formula G. adj =G base ⋅(1-η⋅W), where η is the envelope attenuation coefficient; when W is determined to exceed the preset protection limit, the system automatically switches the energy envelope of the burst packet from a flat-top distribution to an exponential attenuation distribution with a steep front and a gentle back, that is, maintaining the pulse intensity of the initial segment of the envelope to ensure the process ignition, and simultaneously reducing the energy injection of the pulse in the later segment of the burst packet proportionally, thereby achieving thermal unloading and physical protection of the optical components by actively reducing the power density at the end of the envelope.

[0042] Example 1: In a specific excimer laser annealing production scenario, when the system faces a 2.0s non-discharge downtime t caused by the switching of adjacent glass substrates... off At that time, the timing unit records the time parameter and transmits it to the pre-modulation execution unit. Because this duration causes a decrease in the ionization degree of the gaseous medium inside the cavity, to prevent a drop in the energy of the first pulse, the pre-modulation execution unit, based on the pre-set correlation model in the voltage compensation module, injects a target voltage compensation parameter of ΔV = 150V into the high-voltage power supply within a nanosecond-level time window after receiving the burst start command. This compensation mechanism works in conjunction with the flow field stability determination unit to maintain the energy density of the first pulse, providing the current waveform sampling module with electrical raw data that meets the signal-to-noise ratio standard, enabling the system to start monitoring the cavity from the first pulse of the burst packet. Online monitoring of the flow field state; when the first pulse discharge occurs, the current waveform sampling module extracts the falling edge waveform of the discharge current at a sampling rate of 2GS / s. The flow field stability determination unit calculates that the fluctuation variation coefficient C within the sliding time window is 0.06. This value exceeds the preset instability threshold of 0.05, indicating that the coherent superposition of acoustic oscillations in the cavity induces the filamentation trend of the discharge channel. The fluctuation variation coefficient C is calculated according to the following formula: C=σ / μ, where C is the fluctuation variation coefficient, σ is the standard deviation of the dynamic impedance characteristic S within the sliding time window, and μ is the mean value of the dynamic impedance characteristic S within the sliding time window.

[0043] The pulse timing scheduling unit then intervenes, actively disrupting the formation conditions of the standing wave field by adjusting the trigger period of subsequent pulses from a constant 166μs to a non-equally spaced sequence consisting of 175μs and 160μs, thus restoring the heat and pressure distribution within the laser cavity to a uniform state. Simultaneously, the waveform adaptive adjustment module increases the synchronization delay interval Δt between the pre-ionization trigger sequence and the main discharge trigger sequence to 110ns, thereby reducing the transient stress gradient acting on the coating of the optical components by decreasing the power rise rate at the pulse leading edge.

[0044] Example 2: On the excimer laser annealing equipment platform, to verify the supporting role of the intelligent modulation mechanism in output consistency and flow field stability, Gaussian white noise with a signal-to-noise ratio of 20dB was actively superimposed on the experimental signal source to simulate high-frequency electromagnetic interference in the industrial environment. The experimental group used a broadband current transformer and an analog-to-digital converter with a sampling frequency of not less than 2GS / s to collect the original waveform of the discharge circuit. The sampling accuracy of the timing unit was set to 10ns. The selection of this parameter is based on the trade-off between the microscopic capture resolution of the downtime and the data throughput rate of the back-end controller. It is used to capture the attenuation window of the dielectric ionization degree without causing communication blockage. The correlation model adopts the form of a piecewise linear lookup table, and its input is the non-discharge downtime t. off The output is the voltage compensation amount for the high-voltage power supply charging reference.

[0045] The experiment examined the response accuracy of the pre-modulated execution unit by setting a shutdown duration gradient, namely a lower limit of 0.5s, a median of 2.0s, and an upper limit of 5.0s, and simultaneously monitored the intermediate variables generated by the flow field stability determination unit, as the non-discharge shutdown duration t increased. off As t increases, the target voltage compensation parameter ΔV output by the voltage compensation module exhibits a nonlinear growth law corresponding to the spontaneous decay trend of the dielectric ionization degree. off After more than 5.0s, as the physical state of the gaseous medium in the laser cavity tends to a quasi-static equilibrium state, the gain slope of the target voltage compensation parameter ΔV tends to flatten out, indicating that the defined parameter window covers the main evolution range of physical memory loss. The experiment uses the existing conventional feedback adjustment logic as the control group and the scheme including voltage pre-compensation and flow field dynamic monitoring as the experimental group. The specific test results are shown in Table 1.

[0046] Table 1: Comparison of Laser Performance Test Data

[0047] Referring to Table 1, the experimental group maintained the initial pulse energy deviation within 0.88% under different downtime gradients. Compared to the control group's 16.45% energy jump at 6.0s, the pre-modulation execution unit compensated for the start-up delay caused by insufficient dielectric ionization by pre-biasing the charging voltage, eliminating the energy drop at the moment of sudden start-up. When the fluctuation variation coefficient C exceeded the instability threshold of 0.05, the pulse timing scheduling unit intervened to generate a non-equal interval trigger sequence, reducing the instantaneous thermal shock energy density of the optical components from 158.2 mJ / cm³. 2 It decreased to 121.5 mJ / cm 2 The fluctuation variation coefficient C is calculated using the following formula: C = σ / μ, where C is the fluctuation variation coefficient, σ is the standard deviation of the dynamic impedance characteristic S within the sliding time window, and μ is the mean of the dynamic impedance characteristic S within the sliding time window.

[0048] Example 3: This example combines Figures 1 to 3 A description of an intelligent modulation and protection system, method, and storage medium for a laser, such as... Figure 1As shown, it includes a main control computer system as the logic operation and scheduling center, a high-speed signal acquisition and processing unit to perform FPGA / DSP real-time analysis, a high-voltage power supply component responsible for energy injection and modulation execution, and an excimer laser physical device as the opto-mechatronic core. The main control computer system has a timing unit for monitoring non-discharge shutdown time, a voltage compensation module for storing associated model data, and a pulse timing scheduling unit for adjusting the trigger period and disrupting coherent superposition. The high-speed signal acquisition and processing unit has a current waveform sampling module to collect the dynamic impedance characteristics of the falling edge and a flow field stability determination unit to calculate the fluctuation variation coefficient. The high-voltage power supply component includes a pre-modulation execution unit to inject target voltage compensation parameters and a charging reference control loop to receive voltage regulation commands. The excimer laser physical device includes a laser cavity and discharge electrodes to generate plasma and an acoustic oscillation source, optical components to avoid local hot spot damage, and a current / voltage transformer to collect physical signals. The units interact and achieve closed-loop control through voltage compensation commands, high-voltage pulse discharge, loop current signals, and instability state feedback paths.

[0049] like Figure 2 As shown in the figure, the horizontal axis represents time in μs with a scale range of 0 to 10, and the vertical axis represents voltage percentage with a scale range of 0 to 100. The solid line representing the voltage rise waveform under normal conditions and the dashed line representing the voltage rise waveform under temperature rise conditions both exhibit non-linear growth trends. The voltage rise rate under temperature rise conditions is lower than under normal conditions. A horizontal dashed line is included in the figure as a heat dissipation threshold, used to indicate the key judgment boundary in power supply behavior analysis. Figure 3 As shown, this fishbone diagram is result-oriented, focusing on laser output instability and damage. Its upper factor branches include physical state causes and signal monitoring characteristics. The physical state causes include non-discharge shutdown time, intracavity acoustic oscillation, and discharge filamentation caused by coherent superposition. The signal monitoring characteristics include charging voltage decay characteristics, current falling edge data, and dynamic impedance characteristics. The lower factor branches include algorithm judgment logic and regulation and protection actions. The algorithm judgment logic involves instability threshold comparison, fluctuation variation coefficient calculation, and correlation model preset. The regulation and protection actions involve disrupting the coherent superposition state, pulse timing scheduling, and first pulse target voltage compensation.

[0050] Example 4: In the system commissioning scenario of an excimer laser annealing equipment, for a newly connected laser cavity, the calibration unit determines the non-discharge shutdown time t by controlling the laser to execute a burst cycle sequence. offThe functional response characteristics between the target voltage compensation parameter ΔV; in this sequence, the shutdown silence interval between adjacent burst packets is set with nonlinear step gradients (or logarithmic gradients) of 0.1s, 0.5s, 1.0s, 2.0s, 5.0s, and 10.0s. The system records the measured energy value E of the first pulse at each gradient point without pre-control. first and its relative steady-state average energy E avg The deviation rate; the pre-modulation execution unit calculates the compensation voltage gain required to bring the first pulse energy back to the set reference, and establishes a linear lookup table for the correlation model based on this; this procedure transforms the attenuation characteristics of the cavity physical state into quantifiable energy bias parameters, and establishes the initial logic reference for the pre-modulation action.

[0051] To determine the quantization boundary for instability assessment of the intracavity flow field, the system performs a baseline calibration process based on statistical distribution during steady-state laser operation. The current waveform sampling module continuously acquires the falling edge waveform data of 1000 pulses of discharge current and analyzes the corresponding dynamic impedance characteristic S. The flow field stability assessment unit verifies the distribution characteristics of this batch of dynamic impedance characteristic S, extracting its mean fluctuation μ0 and standard deviation σ0. The instability threshold T... h Determine T according to the following formula: h =k⋅σ0 / μ0, where T h The instability threshold is given by k, the confidence factor is given by σ0, the standard deviation of the dynamic impedance characteristic S under steady-state conditions is given by μ0, and the mean fluctuation of the dynamic impedance characteristic S under steady-state conditions is given by μ0. In the specific application of this embodiment, the instability threshold T is obtained by setting k to 3. h The value is 0.0515. This calibration method eliminates background noise interference in flow field monitoring and provides a definite physical criterion for subsequent pulse timing scheduling. Under high-load continuous operation, the thermal load scheduler initiates a real-time thermal protection process for the high-voltage power supply. The power behavior analysis unit periodically acquires the current waveform when the high-voltage power supply charges the energy storage capacitor and calculates the second derivative κ of the voltage rise waveform. When the value of the second derivative κ shows a monotonically decreasing trend within 5 consecutive sampling periods, it indicates that the high-voltage power supply power device has a decrease in driving transconductance due to temperature rise. The pulse timing scheduling unit performs timing compensation action according to the feedback signal, that is, gradually increases the pulse trigger interval in 5μs increments, thereby increasing the heat dissipation window of the high-voltage power supply until the second derivative κ returns to the normal operating range.

[0052] Example 5: In the initial deployment scenario of the excimer laser annealing equipment, a controlled voltage decay characteristic scanning program is executed based on the extracted reference parameters of the correlation model. During the program startup phase, the laser is in the initial standard state with fresh working gas. The system generates a non-discharge shutdown duration gradient pulse consisting of 0.1s to 10.0s by controlling the timing unit, and records the measured value of the first pulse energy corresponding to each gradient point in the open-loop state without voltage premodulation. The premodulation execution unit uses the acquired data to perform least squares curve fitting to establish the reference voltage compensation amount V. base Non-discharge shutdown duration t off The exponential function characteristics between V base V is determined by the following formula: base =A⋅(1-e -B⋅toff ), where V base The reference voltage compensation is expressed in V, where A is the saturation compensation constant, B is the attenuation coefficient, and t is the reference voltage compensation value. off The non-discharge shutdown duration is defined as e, which is the base of the natural logarithm. This is to clarify that the compensation logic follows the exponential evolution of the ionization degree of the physical medium, rather than a nonlinear increase or decrease relationship.

[0053] The system writes the fitted parameters A and B into memory as a reference for calculating the target voltage compensation parameter ΔV in the production process. This is used to reconstruct the logic baseline after changing the working gas or electrode. When the laser enters continuous production mode and the start-up delay parameter τ of the first pulse deviates from the standard threshold, the feedback calibration module initiates a gain correction procedure to offset the model drift caused by hardware losses. During the execution of this procedure, the feedback calibration module calculates the deviation of the start-up delay parameter τ from the standard value and uses a proportional gain factor to convert this time deviation into a gain correction vector in the voltage dimension. The feedback calibration module limits the step resolution of this correction vector to within 0.1% of the standard discharge voltage to suppress energy overshoot during the closed-loop regulation process. The feedback calibration module superimposes the gain correction vector into the output gain of the pre-modulation execution unit to adjust the charging reference of subsequent pulses in the current burst packet until the start-up delay parameter τ of the first pulse converges back to the preset pulse dead zone.

[0054] Example 6: In the system integration and field verification scenario of excimer laser annealing equipment, for the overflow handling of the timing unit and the quantization calibration of the voltage physical limiter, the system executes a pre-fault-tolerant procedure, that is, determines the clock frequency f of the internal counter of the timing unit. clk The frequency is 100MHz, and the non-discharge shutdown duration t is defined according to the memory bit width. off The measurement boundary is 60.0s; if the non-discharge shutdown duration exceeds this measurement boundary, the system will... offThe forced clamp is positioned at 60.0s, and the target voltage compensation parameter ΔV of the first pulse is calculated based on the saturation point gain in the voltage compensation module; the system detects the breakdown voltage V of the laser cavity by gradually increasing the charging setpoint of the high-voltage power supply. break The voltage physical limiter sets the upper limit of the target voltage compensation parameter ΔV to V based on the detection results. break It achieves 95% protection, thus providing physical overvoltage protection under boundary conditions where the logic algorithm malfunctions.

[0055] To ensure the sampling fidelity of the current waveform sampling module under different impedance environments, the system executes a signal channel correction procedure based on a calibration source. A calibration signal with a rise-edge accuracy of 1.0 ns is injected into the broadband current transformer to calculate the time-domain delay deviation at the analog-to-digital converter output. The flow field stability determination unit adjusts the starting point of the sliding time window based on this delay deviation, ensuring that the extraction of the dynamic impedance characteristic quantity S is always locked within the falling edge interval of the energy concentration in the discharge current waveform. Simultaneously, the system measures the inherent parasitic inductance L of the discharge circuit under no-load operation of the laser. p The induced voltage component generated by the parasitic inductance is then separated from the measured current falling edge data, and the dynamic impedance component caused by the change in plasma flow field is extracted, where the parasitic inductance L... p The measurement was obtained using a sinusoidal excitation signal with a frequency of 10MHz; in the service life monitoring scenario, the system uses a high-voltage sampling circuit to extract the resonant frequency f at the discharge turn-off moment. r The curve correction unit calculates f r The exponential weighting coefficients are adjusted based on the deviation from the initial baseline to compensate for energy response drift caused by electrode aging.

[0056] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart modulation and protection system for a laser, used in an excimer laser annealing apparatus, characterized in that, include: A timing unit is used to monitor the non-discharge shutdown time of the laser in real time; The voltage compensation module is used to preset the correlation model between the non-discharge shutdown time and the charging voltage decay characteristics; The pre-modulation execution unit, connected to the timing unit and the voltage compensation module, is used to determine the target voltage compensation parameters for the first pulse based on the non-discharge shutdown duration and the correlation model when a burst start command is received, and to inject the target voltage compensation parameters into the charging reference of the high-voltage power supply. The current waveform sampling module is used to collect the current falling edge data of the high voltage power supply in real time during each pulse discharge process, and extract the rate of change of the discharge current from the peak point to the preset current ratio point as a dynamic impedance characteristic quantity. The flow field stability determination unit is connected to the current waveform sampling module and is used to calculate the fluctuation variation coefficient of the dynamic impedance characteristic between adjacent pulses within the burst packet. The fluctuation variation coefficient is the ratio of the standard deviation to the mean of the dynamic impedance characteristic within the sliding time window. The pulse timing scheduling unit, connected to the flow field stability determination unit, is used to change the triggering period of subsequent pulses to disrupt the coherent superposition state of acoustic oscillations in the laser cavity when the fluctuation variation coefficient exceeds the preset instability threshold, thereby suppressing the distortion of laser energy distribution in the spatial dimension and avoiding local hot spot damage to optical components caused by discharge channel filamentation.

2. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, It also includes a feedback calibration module, which is connected to the pre-modulation execution unit and is used to collect the ignition delay parameters during the first pulse discharge process. The feedback calibration module generates a voltage gain correction vector for subsequent pulses in the current burst packet based on the deviation of the ignition delay parameters from the standard ignition threshold, and adjusts the output gain of the pre-modulation execution unit to correct the correlation model.

3. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, It also includes a waveform adaptive adjustment module, which is used to determine the thermal stress load level of the optical components based on the non-discharge shutdown duration; The waveform adaptive adjustment module adjusts the pre-ionization triggering sequence of the initial pulse in the bursting package and the synchronous delay interval between it and the main discharge triggering sequence according to the thermal stress bearing level. By widening the rising edge slope of the initial pulse, the instantaneous peak power acting on the optical components is reduced.

4. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, It also includes an electrode loss inversion module and a curve correction unit; The electrode loss inversion module is used to acquire the frequency of the residual oscillation waveform generated by the high-voltage power supply at the end of the discharge cycle. The curve correction unit connects the voltage compensation module and the electrode loss inversion module. It is used to adjust the calculation slope of the target voltage compensation parameters based on the electrode loss state determined by the offset of the residual oscillation waveform frequency relative to the preset resonance reference.

5. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, It also includes a power behavior analysis unit and a thermal load scheduler; the power behavior analysis unit is used to obtain the second derivative of the voltage rise waveform when the high-voltage power supply charges the energy storage component as curvature; the thermal load scheduler is connected to the power behavior analysis unit and the pulse timing scheduling unit, and is used to reduce the transient thermal load of the high-voltage power supply by extending the trigger interval of subsequent pulses when the curvature indicates that the temperature rise of the high-voltage power supply exceeds the heat dissipation threshold of the power device.

6. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, The fluctuation variation coefficient C is determined by the following formula: C=σ / μ, where σ is the standard deviation of the dynamic impedance characteristic within the sliding time window, and μ is the mean of the dynamic impedance characteristic within the sliding time window.

7. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, The pulse timing scheduling unit introduces non-equal time windows to guide the turbulent structure in the laser cavity to perform thermal equilibrium relaxation, so that the energy density component of the pulse energy fluctuation residual sequence in the burst packet is lower than the preset judgment threshold in the acoustic coherence frequency band. The pre-modulation execution unit is also connected to a voltage physical limiter, which is used to limit the target voltage compensation parameter within the safe discharge range of the laser cavity.

8. The intelligent modulation and protection system for a laser according to claim 1, characterized in that, It also includes a cumulative thermal load calculation module, which is used to perform time-dimensional integration on the energy density of the continuously output pulses to determine the equivalent thermal stacking state of the optical components. The pulse timing scheduling unit adjusts the energy envelope distribution of the laser pulse based on the product weight value of the equivalent thermal accumulation state and the fluctuation variation coefficient. The system synchronously performs open-loop voltage pre-compensation for the first pulse and impedance monitoring for the uniformity of the flow field in the cavity by reusing the discharge circuit current signal of the high-voltage power supply.

9. A method for intelligent modulation and protection of a laser, used to implement the intelligent modulation and protection system for the laser as described in claim 1, characterized in that, Includes the following steps: Step 101: Monitor the non-discharge shutdown time of the laser in real time; Step 102: Preset the correlation model between non-discharge shutdown time and charging voltage decay characteristics; Step 103: Upon receiving the burst start command, determine the target voltage compensation parameters for the first pulse based on the non-discharge shutdown duration and the correlation model, and inject the target voltage compensation parameters into the charging reference of the high-voltage power supply. Step 104: Real-time acquisition of current falling edge data of high voltage power supply during each pulse discharge process, and extraction of the rate of change of discharge current from peak point to preset current ratio point as a dynamic impedance characteristic quantity. Step 105: Calculate the fluctuation variation coefficient of the dynamic impedance characteristic between adjacent pulses within the burst packet. The fluctuation variation coefficient is the ratio of the standard deviation to the mean of the dynamic impedance characteristic within the sliding time window. Step 106: When the fluctuation variation coefficient exceeds the preset instability threshold, the triggering period of subsequent pulses is changed to disrupt the coherent superposition state of acoustic oscillations in the laser cavity, thereby suppressing the laser energy distribution distortion in the spatial dimension and avoiding local hot spot damage to optical components caused by discharge channel filamentation.

10. A smart modulation and protection storage medium for a laser, characterized in that, A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the intelligent modulation and protection method for a laser as described in claim 9.

Images