Intelligent control method and device for laser etching for chip processing

Abstract

The present application provides an intelligent control method and device for laser etching for chip processing, which relates to the field of laser etching technology, including: obtaining the laser control parameter space for chip laser etching processing, randomly generating laser control parameters; generating integrated simulated laser etching microgroove images for chip substrates, and identifying and obtaining multiple microgroove bottom roughnesses; performing individual microgroove electrical performance impact analysis, performing continuous microgroove electrical performance impact analysis, and obtaining continuous electrical performance impact parameters; calculating and obtaining the etching score of the laser control parameters, optimizing the laser control parameters, obtaining the optimal laser etching parameters, and controlling the chip laser etching processing. This application can solve the technical problem that the roughness of microgrooves cannot be accurately controlled during the etching process in the prior art, achieve the technical goal of accurately analyzing and optimizing the relationship between laser etching parameters and microgroove electrical performance, and achieve the technical effect of effectively controlling the roughness of the bottom of the microgroove.

Description

Technical Field

[0001] The present application relates to the field of laser etching technology, and in particular to an intelligent control method and device for laser etching for chip processing. Background Art

[0002] In the field of modern semiconductor manufacturing, laser etching technology is widely used in chip micromachining, especially in the manufacture of complex microstructures and microgrooves, laser etching can achieve high-precision processing effects.

[0003] At present, the existing technology lacks a systematic analysis of the impact of different etching parameters on the electrical performance of microgrooves. Most current etching processes focus on geometry and etching depth, but there is a lack of in-depth research and optimization methods for the microscopic surface quality, especially the impact of the roughness of the bottom of the microgroove on electrical performance. Since the roughness of the microgroove will directly affect the contact resistance and signal transmission performance of the chip, if these parameters are not optimized and controlled, the electrical performance of the chip in actual applications may be reduced, thereby affecting the overall performance and reliability of the product.

[0004] In summary, the existing technology has a technical problem that due to the lack of systematic analysis of the impact of different laser etching parameters on the surface roughness of the bottom of the microgroove and its electrical properties, the roughness of the microgroove cannot be accurately controlled during the etching process, further affecting the electrical performance stability and signal transmission efficiency of the chip. Summary of the invention

[0005] The purpose of this application is to provide an intelligent control method and device for laser etching for chip processing, so as to solve the technical problem in the prior art that the roughness of the microgroove bottom surface and its electrical properties cannot be accurately controlled during the etching process due to the lack of systematic analysis of the influence of different laser etching parameters on the surface roughness of the microgroove bottom and its electrical properties, which further affects the electrical performance stability and signal transmission efficiency of the chip.

[0006] In view of the above problems, the present application provides an intelligent control method and device for laser etching for chip processing.

[0007] In a first aspect, the present application provides an intelligent control method for laser etching for chip processing, which is implemented by an intelligent control device for laser etching for chip processing, including: obtaining a laser control parameter space for chip laser etching processing, and randomly generating laser control parameters in the laser control parameter space; generating an integrated simulated laser etching microgroove image for a chip substrate according to the laser control parameters, obtaining a plurality of simulated etching microgroove images, and identifying and obtaining a plurality of microgroove bottom roughnesses; performing individual microgroove electrical performance influence analysis according to the plurality of microgroove bottom roughnesses, respectively, to obtain a plurality of microgroove electrical performance influence parameters, and performing continuous microgroove electrical performance influence analysis in combination with the plurality of microgroove bottom roughnesses, to obtain continuous electrical performance influence parameters; calculating and obtaining an etching score of the laser control parameters based on the plurality of microgroove electrical performance influence parameters and the continuous electrical performance influence parameters, and optimizing the laser control parameters based on the etching score to obtain optimal laser etching parameters, and performing chip laser etching processing control.

[0008] In a second aspect, the present application also provides an intelligent control device for laser etching for chip processing, which is used to execute the intelligent control method for laser etching for chip processing as described in the first aspect, including: a laser control parameter generation module, which is used to obtain the laser control parameter space for chip laser etching processing, and randomly generate laser control parameters in the laser control parameter space; a roughness recognition module, which is used to generate an integrated simulated laser etching microgroove image for the chip substrate according to the laser control parameters, obtain multiple simulated etching microgroove images, and identify and obtain multiple microgroove bottom roughnesses; an electrical performance influence analysis module, which is used to perform individual microgroove electrical performance influence analysis according to the multiple microgroove bottom roughnesses, obtain multiple microgroove electrical performance influence parameters, and perform continuous microgroove electrical performance influence analysis in combination with the multiple microgroove bottom roughnesses to obtain continuous electrical performance influence parameters; a laser etching module, which is used to calculate the etching score of the laser control parameters based on the multiple microgroove electrical performance influence parameters and the continuous electrical performance influence parameters, and optimize the laser control parameters based on the etching score to obtain the optimal laser etching parameters, and perform chip laser etching processing control.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By acquiring the laser control parameter space of the chip laser etching processing, randomly generating laser control parameters in the laser control parameter space; according to the laser control parameters, generating an integrated simulated laser etching microgroove image of the chip substrate, obtaining a plurality of simulated etching microgroove images, and identifying and obtaining a plurality of microgroove bottom roughnesses; performing individual microgroove electrical performance influence analysis respectively according to the plurality of microgroove bottom roughnesses, obtaining a plurality of microgroove electrical performance influence parameters, and performing continuous microgroove electrical performance influence analysis in combination with the plurality of microgroove bottom roughnesses, obtaining a continuous electrical performance influence parameter; based on the plurality of microgroove electrical performance influence parameters and the continuous electrical performance influence parameter, calculating and obtaining an etching score of the laser control parameters, and optimizing the laser control parameters based on the etching score, obtaining an optimal laser etching parameter, and controlling the chip laser etching processing, achieving the technical goal of accurately analyzing and optimizing the relationship between the laser etching parameters and the microgroove electrical performance, and achieving the technical effect of effectively controlling the roughness of the microgroove bottom, thereby improving the electrical performance stability of the chip and the signal transmission efficiency.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 A schematic diagram of the flow of the intelligent control method for laser etching for chip processing in this application;

[0014] Figure 2 This is a schematic diagram of the structure of the intelligent control device for laser etching used in chip processing in this application.

[0015] Description of reference numerals:

[0016] Laser control parameter generation module 11, roughness identification module 12, electrical performance impact analysis module 13, laser etching module 14. DETAILED DESCRIPTION

[0017] This application provides an intelligent control method and device for laser etching for chip processing, which solves the technical problem that the roughness of the microgroove bottom surface and its electrical properties cannot be accurately controlled during the etching process due to the lack of systematic analysis of the effects of different laser etching parameters on the surface roughness of the microgroove bottom surface and its electrical properties, further affecting the electrical performance stability and signal transmission efficiency of the chip. The technical goal of accurately analyzing and optimizing the relationship between laser etching parameters and the electrical properties of the microgroove is achieved, and the technical effect of effectively controlling the roughness of the microgroove bottom surface and thus improving the electrical performance stability and signal transmission efficiency of the chip is achieved.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] For example, please refer to the attached Figure 1 The present application provides an intelligent control method for laser etching for chip processing, which is applied to an intelligent control device for laser etching for chip processing, and specifically includes the following steps:

[0020] Step 1: Obtain a laser control parameter space for laser etching of a chip, and randomly generate laser control parameters in the laser control parameter space.

[0021] Specifically, the laser control parameter space in chip laser etching processing refers to the range of all laser parameters that can be adjusted during laser etching processing, such as laser power, frequency, and pulse width. For example, the laser power is adjusted between 1 and 10 watts, the pulse frequency is between 10 and 1000 Hz, and the pulse width varies from 10 to 100 nanoseconds. Then, a combination is randomly selected in the laser control parameter space to generate the laser control parameters.

[0022] Step 2: According to the laser control parameters, an integrated simulated laser etching microgroove image is generated for the chip substrate to obtain a plurality of simulated etching microgroove images, and a plurality of microgroove bottom roughnesses are identified and obtained.

[0023] Specifically, the bottom morphology of the microgroove produced by laser etching is affected by many factors, such as the uniformity of the chip material. Based on the generative adversarial network, multiple etching microgroove image simulation branches are used, and the laser control parameters are used as input to integrate and simulate the generation of laser etching microgroove images for the chip substrate, that is, the tiny groove structure formed after laser etching is simulated on the chip substrate, and the etching results are predicted by computer simulation technology. For example, the laser power is set to 5 watts, the pulse frequency is 500 Hz, and the scanning speed is 5 mm per second. According to these parameters, multiple simulated images of microgrooves are generated to obtain multiple simulated etching microgroove images, and each simulated etching microgroove image represents a different etching result.

[0024] Then, multiple simulated etching microgroove images are used to observe how different parameter combinations affect the shape, depth and surface characteristics of the microgroove. Next, multiple microgroove bottom roughnesses are obtained by analyzing multiple simulated etching microgroove images, and the roughness of the bottom of the microgroove is calculated. The roughness of the bottom of multiple microgrooves is an indicator to measure the smoothness of the surface. The larger the value, the rougher the surface. For example, if the bottom roughness of a microgroove is 0.5 microns, and the bottom roughness of another microgroove is 0.8 microns, the surface of the microgroove with a bottom roughness of 0.8 microns is rougher. Roughness can affect the electrical performance of the chip, such as affecting the local contact resistance, and affecting the signal transmission integrity of the overall chip.

[0025] Step three: performing individual microgroove electrical performance influence analysis according to the roughness of the bottom of the multiple microgrooves, respectively, to obtain multiple microgroove electrical performance influence parameters, and performing continuous microgroove electrical performance influence analysis in combination with the roughness of the bottom of the multiple microgrooves, to obtain continuous electrical performance influence parameters.

[0026] Specifically, the local electrical performance impact is analyzed based on the multiple roughnesses identified, specifically the change in the contact resistance at the bottom of the local microgroove and the impact on the continuous electrical performance of the entire chip when the roughness of the bottom of the microgroove appears at multiple locations in the microgroove of a chip. Roughness is a parameter that measures the smoothness of the surface at the bottom of the microgroove. For example, when the roughness is high, the contact resistance may increase. The parameter affecting the continuous electrical performance is the signal transmission loss ratio of the chip, for example, the signal transmission loss of the entire microgroove area is 5%.

[0027] Step 4: Based on the multiple microgroove electrical performance influencing parameters and continuous electrical performance influencing parameters, the etching score of the laser control parameters is calculated, and the laser control parameters are optimized based on the etching score to obtain the optimal laser etching parameters and perform chip laser etching processing control.

[0028] Specifically, based on multiple microgroove electrical performance influencing parameters and continuous electrical performance influencing parameters, the individual parameters and overall parameters of the microgroove electrical performance previously obtained are used as the basis to evaluate the influence of each set of laser control parameters on etching. For example, the contact resistance change caused by the roughness of the microgroove bottom may be 10%, while the signal loss of the overall microgroove area may be 5%. Combined, the comprehensive score of each set of laser control parameters is calculated to obtain the etching score of the laser control parameters, which reflects the quality of the etching effect.

[0029] Next, the laser control parameters are optimized based on the etching score, that is, the laser parameters, such as power, frequency, and scanning speed, are adjusted to optimize the effect of laser etching by improving the score. Among them, the smaller the total electrical performance influencing parameter, the greater the etching score. For example, the interval of the total electrical performance influencing parameter (such as 0-1) can be mapped to the interval of the etching score (0-10), and the total electrical performance influencing parameter 0 corresponds to an etching score of 10, and 1 corresponds to 0. Finally, the optimal laser etching parameters are obtained, that is, the laser control parameter combination with the highest score is found during the optimization process, and the chip laser etching process control is performed to ensure the best etching quality and electrical performance.

[0030] The intelligent control method for laser etching for chip processing is applied to an intelligent control device for laser etching for chip processing, which can achieve the technical goal of accurately analyzing and optimizing the relationship between laser etching parameters and micro-groove electrical performance, and achieve the technical effect of effectively controlling the roughness of the bottom of the micro-groove, thereby improving the stability of the chip electrical performance and the signal transmission efficiency.

[0031] Furthermore, the present application also includes:

[0032] The adjustment ranges of laser power, pulse frequency and scanning speed in chip laser etching processing are obtained, and the laser control parameter space is obtained by combining them; and laser control parameters are randomly generated in the laser control parameter space.

[0033] Specifically, the adjustment range of laser power, pulse frequency and scanning speed in the chip laser etching process is obtained, and the adjustable value range of these parameters in the laser etching process is determined. Laser power is used to control the energy output of the laser beam, the pulse frequency affects the repetition rate of the laser pulse, and the scanning speed determines the speed at which the laser moves on the chip surface. Then, the adjustment range of laser power, pulse frequency and scanning speed is combined to obtain the laser control parameter space.

[0034] Then, randomly generate laser control parameters in the laser control parameter space as parameters used in the etching process. For example, randomly generate a combination of 4 watts of power, 600 Hz of frequency, and 7 mm / s of scanning speed for experiments or optimization of processing conditions.

[0035] By exploring different combinations of processing parameters, the effect of chip laser etching can be optimized.

[0036] Furthermore, the present application also includes:

[0037] Based on the historical data of chip laser etching, a sample laser control parameter set is collected, and the chip microgroove scanning images after etching with different sample laser control parameters are collected to obtain a sample etching microgroove image set; the sample laser control parameter set and the sample etching microgroove image set are used to train an etching microgroove image simulator, and the etching microgroove image simulator includes multiple integrated etching microgroove image simulation branches; the laser control parameters are respectively input into multiple etching microgroove image simulation branches in the etching microgroove image simulator to generate multiple simulated etching microgroove images.

[0038] Specifically, based on the historical data of chip laser etching, as a reference for the current analysis, a sample laser control parameter set is collected, such as historical laser power, historical pulse frequency, and historical scanning speed. Next, the chip microgroove scanning images after etching with different sample laser control parameters in the sample laser control parameter set are collected to obtain a sample etching microgroove image set of the microgrooves on the chip surface after etching corresponding to the sample laser control parameter set. For example, when the laser power is 5 watts, the pulse frequency is 500 Hz, and the scanning speed is 5 mm per second, the obtained scanning image can show the specific shape of the microgroove.

[0039] Then, a sample laser control parameter set and a sample etching microgroove image set are used to train an etching microgroove image simulator, that is, a model for predicting etching images is trained by using known laser control parameters and their corresponding microgroove image data as an etching microgroove image simulator. The integration of multiple etching microgroove image simulation branches in the etching microgroove image simulator means that the etching microgroove image simulator contains multiple independent simulation branches, each of which may focus on predicting a specific type of etching result, or focus on processing certain specific laser parameters, or the training data of multiple branches are different, so the output microgroove images are not exactly the same, thereby improving the prediction ability of the etching microgroove image simulator. Among them, based on the generative adversarial network training, the generator is responsible for generating simulated images, while the discriminator judges the authenticity of the image. Through continuous iterative training, the generator will gradually improve the fidelity of the generated image, and the discriminator will become more and more adept at distinguishing true from false. For example, the microgroove image generated by the generator may be relatively rough at first, but after multiple trainings, it becomes closer and closer to the real scanned image, and details such as the depth and edge morphology of the microgroove will be more refined.

[0040] Next, the laser control parameters are respectively input into multiple etching microgroove image simulation branches in the etching microgroove image simulator. Each branch will generate a corresponding simulated etching microgroove image according to the input parameters. When the training data of multiple branches are different, each laser control parameter may produce different etching results for laser etching.

[0041] Through historical data collection, sample laser parameters and corresponding microgroove images are collected, and an image simulator is trained to predict the etching effect under different parameters. Then, new parameters are input into the simulator to generate corresponding microgroove images to predict the etching effect.

[0042] Furthermore, the present application also includes:

[0043] The sample laser control parameter set and the sample etching microgroove image set are used as an etching simulation construction data set, and a first etching simulation training data set is randomly selected from the etching simulation construction data set; based on a generative adversarial network, a first etching microgroove image simulation branch is constructed, and the etching microgroove image simulation branch includes a first etching microgroove image generation path and a first etching microgroove image discrimination path; using the first etching simulation training data set, the first etching microgroove image simulation branch is supervised and trained until convergence; continuing to randomly select etching simulation training data sets from the etching simulation construction data set, and training multiple etching microgroove image simulation branches to obtain an etching microgroove image simulator.

[0044] Specifically, the sample laser control parameter set and the sample etching microgroove image set are used as the etching simulation construction data set, that is, the collected laser control parameters and the corresponding microgroove image data are integrated together to form a complete data set for subsequent simulation and training. Next, an etching simulation training data set is randomly selected from the etching simulation construction data set as the first step of the training data set to train the etching simulation model.

[0045] Based on the generative adversarial network, the first etching microgroove image simulation branch is constructed. The generative adversarial network contains a generator and a discriminator, which compete with each other to improve the generation ability of the model. The etching microgroove image simulation branch includes the first etching microgroove image generation path and the first etching microgroove image discrimination path, which means that the generator is responsible for generating simulated etching images, while the discriminator is responsible for judging whether these images match the real images. For example, the generation path tries to generate microgroove images that look similar to the actual etching effect, while the discrimination path determines whether the generated image meets the characteristics of the actual etching image.

[0046] Next, the first etching simulation training data set is used to supervise the first etching microgroove image simulation branch until convergence, which means using the previously randomly selected training data to train the generator and discriminator in the generative adversarial network. Supervised training obtains the real microgroove image corresponding to each laser parameter, and optimizes the network parameters through multiple iterations until the generator can generate an image that is very similar to the real microgroove image.

[0047] Finally, the etching simulation training data set is randomly selected from the etching simulation construction data set, and multiple etching microgroove image simulation branches are trained to obtain an etching microgroove image simulator. The training data is randomly selected from the data set again, and different simulation branches are trained to finally form a complete etching microgroove image simulator. The etching microgroove image simulator consists of multiple branches, each of which can handle different types of laser etching conditions, helping to more accurately predict the etching effects under different parameter combinations.

[0048] By randomly selecting data sets and conducting multiple trainings, we eventually built a simulator that can accurately simulate the etching effects under different laser parameters and generate microgroove images that are close to the actual etching effects.

[0049] Furthermore, the present application also includes:

[0050] According to the microgroove roughness detection data of chip etching, a set of sample etched microgroove images is collected, and the maximum peak-to-valley height difference of the microgroove bottom in the sample etched microgroove image is marked as the sample roughness to obtain the sample microgroove bottom roughness set; the sample etched microgroove image set and the sample microgroove bottom roughness set are used to train a bottom roughness identifier based on a convolutional neural network; the bottom roughness identifier is used to perform microgroove bottom roughness identification on the multiple simulated etching microgroove images to obtain multiple microgroove bottom roughnesses.

[0051] Specifically, according to the roughness detection data of the microgroove etched by the chip, a set of sample etched microgroove images is collected to reflect the morphology of the microgroove under different etching conditions. The roughness detection data is obtained by measuring the smoothness of the bottom surface of the microgroove. For example, the roughness of the etched microgroove is obtained by scanning and detecting the degree of surface unevenness. Then, the maximum peak-to-valley height difference at the bottom of the microgroove in the sample etched microgroove image is marked as the sample roughness, and the roughness set of the bottom of the sample microgroove is obtained, that is, the maximum peak-to-valley height difference at the bottom of the microgroove is extracted from the image as a measure of roughness, and the roughness data of each sample is marked. For example, the peak-to-valley height difference of one sample is 0.5 microns, and that of another sample is 0.8 microns, which constitute a roughness set.

[0052] Next, we used a set of sample etched microgroove images and a set of sample microgroove bottom roughness to train a bottom roughness identifier based on a convolutional neural network to learn the rules of identifying roughness from images. A convolutional neural network is a deep learning model that automatically extracts features from images. Through training with a large amount of sample data, it learns to predict the roughness of microgrooves based on images.

[0053] Subsequently, the bottom roughness identifier is used to identify the bottom roughness of the microgrooves of the multiple simulated etched microgrooves, predict the roughness of the bottom of the microgrooves in the simulated images, and obtain multiple bottom roughness of the microgrooves. For example, the bottom roughness identifier predicts that the roughness of the simulated image is 0.6 microns, and the roughness of another simulated image is 0.7 microns.

[0054] By training the roughness identifier, the roughness of the simulated image is predicted using the identifier, and then the surface quality of the microgroove is judged.

[0055] Furthermore, the present application also includes:

[0056] According to the electrical performance test data of the etched microgrooves, a set of sample microgroove bottom roughness is collected, and the variation amplitude of the microgroove bottom contact resistance under different sample microgroove bottom roughness is marked as the sample microgroove electrical performance influencing parameter, and the sample microgroove electrical performance influencing parameter set is obtained; the sample microgroove bottom roughness set and the sample microgroove electrical performance influencing parameter set are used to construct a microgroove electrical performance influencing analyzer, and the individual microgroove electrical performance influencing analysis is performed on the multiple microgroove bottom roughnesses, and multiple microgroove electrical performance influencing parameters are obtained; multiple sample microgroove bottom roughness sets of laser etching of multiple sample chips are collected, and the signal loss ratio of signal transmission of the etched microgrooves under different sample microgroove bottom roughness sets is tested, and the signal loss ratio is marked as the sample continuous electrical performance influencing parameter, and the sample continuous electrical performance influencing parameter set is obtained; the multiple sample microgroove bottom roughness sets and the sample continuous electrical performance influencing parameter set are used to construct a continuous microgroove electrical performance influencing analyzer, and the continuous microgroove electrical performance influencing analysis is performed on the multiple microgroove bottom roughnesses, and the continuous electrical performance influencing parameters are obtained.

[0057] Specifically, according to the electrical performance test data of the etched microgroove, the sample microgroove bottom roughness set is collected, that is, the electrical performance of the actual test microgroove is obtained by the electrical contact performance of the test microgroove, especially the corresponding roughness data is obtained, indicating the influence of the smoothness of the microgroove bottom surface on its electrical performance. Then, the microgroove bottom contact resistance variation range under different sample microgroove bottom roughness is marked as the sample microgroove electrical performance influencing parameter, which means recording the contact resistance variation range corresponding to each roughness. For example, when the microgroove roughness is 0.5 micron, the variation range of the contact resistance is 10%, and when the roughness is 0.8 micron, the contact resistance variation range may be 20%. The variation range is marked as the electrical performance influencing parameter, and finally forms the sample microgroove electrical performance influencing parameter set.

[0058] A microgroove electrical performance impact analyzer is constructed using a set of sample microgroove bottom roughness and a set of sample microgroove electrical performance impact parameters. The roughness data and electrical performance parameters are used to predict or evaluate the impact of microgroove roughness on electrical performance. Next, individual microgroove electrical performance impact analysis is performed on multiple microgroove bottom roughnesses, and the electrical performance impact of each microgroove with different roughness is analyzed separately to generate multiple microgroove electrical performance impact parameters. For example, a microgroove with a roughness of 0.6 microns may have an impact parameter of 15% on contact resistance.

[0059] Collecting a set of roughness of the bottom of multiple sample microgrooves etched by laser from multiple sample chips, and testing the signal loss ratio of the signal transmission of the etched microgrooves under different sets of roughness of the bottom of the sample microgrooves, refers to obtaining different roughness data from multiple laser-etched chips and testing the effect of roughness on signal transmission. The signal loss ratio indicates the attenuation caused by the roughness of the microgroove surface during signal transmission. For example, when the roughness is 0.5 microns, the signal loss ratio is 5%, while when the roughness is 0.8 microns, the loss ratio may reach 10%. The loss ratio is marked as a parameter affecting the continuous electrical performance of the sample to form a set of parameters affecting the continuous electrical performance of the sample.

[0060] A continuous microgroove electrical performance impact analyzer is constructed using a set of multiple sample microgroove bottom roughness and a set of sample continuous electrical performance impact parameters to analyze the impact of microgroove roughness on continuous electrical performance. Next, a continuous microgroove electrical performance impact analysis is performed on multiple microgroove bottom roughnesses to obtain continuous electrical performance impact parameters, that is, each microgroove of each roughness is analyzed to obtain its impact parameters on continuous electrical performance (such as signal transmission).

[0061] The roughness of the bottom of the microgroove and its corresponding electrical performance parameters are obtained through electrical performance testing, the impact of individual and continuous microgrooves on electrical performance is evaluated, and the electrical performance in the laser etching process is optimized.

[0062] Furthermore, the present application also includes:

[0063] Perform weighted calculation on the multiple microgroove electrical performance influencing parameters and the continuous electrical performance influencing parameters to obtain the total chip electrical performance influencing parameters; input the total chip electrical performance influencing parameters into a laser etching score classifier, and classify them to obtain an etching score, wherein the laser etching score classifier is constructed by a mapping relationship between multiple sample total chip electrical performance influencing parameters and multiple sample etching scores, and the sample total chip electrical performance influencing parameters and the sample etching scores are negatively correlated; continue to randomly generate and optimize the laser etching control parameters in the laser control parameter space until convergence, and output the laser control parameters with the largest etching score as the optimal laser etching parameters.

[0064] Specifically, multiple microgrooves' electrical performance influencing parameters and continuous electrical performance influencing parameters are weighted to calculate the comprehensive impact on the electrical performance of the entire chip. The weighting process can assign different weights according to importance. For example, if the electrical performance impact of an individual microgroove is more important, it is given a higher weight, while the impact of continuous electrical performance is given a lower weight. Finally, a total parameter reflecting the electrical performance of the entire chip is obtained.

[0065] Next, the total chip electrical performance influencing parameters are input into the laser etching score classifier, and the etching effect score is given according to the value of the electrical performance parameters to obtain the etching score. The laser etching score classifier is constructed by the mapping relationship between multiple sample total chip electrical performance influencing parameters and multiple sample etching scores, which includes the total electrical performance parameters of multiple sample chips and their corresponding etching scores. The size of the sample total chip electrical performance influencing parameters and the sample etching score is negatively correlated, which means that the greater the electrical performance impact, the lower the etching score, and vice versa. For example, the total electrical performance parameter of one sample is 50, and the score may be 80; while the total electrical performance parameter of another sample is 80, and the score may be 60.

[0066] Continue to randomly generate and optimize the laser etching control parameters in the laser control parameter space, that is, continue to randomly generate new parameter combinations and optimize them to find the optimal parameter combination. The optimization process is based on the previously generated score feedback, and continuously improves the etching effect by adjusting parameters such as laser power, frequency, and scanning speed until convergence, and outputs the laser control parameters with the largest etching score as the optimal laser etching parameters. That is, during the optimization process, when the score cannot be further improved, the calculation will be stopped, and the laser control parameters with the highest current score will be output as the optimal parameter combination finally selected. For example, the optimal parameter combination finally selected may be a laser power of 6 watts, a frequency of 600 Hz, a scanning speed of 4 mm per second, and an etching score of 95 points.

[0067] By selecting the optimal laser etching solution, we can ensure that the chip electrical performance and etching quality are optimal.

[0068] In summary, the intelligent control method for laser etching for chip processing provided in this application has the following technical effects:

[0069] By acquiring the laser control parameter space of the chip laser etching processing, randomly generating laser control parameters in the laser control parameter space; according to the laser control parameters, generating an integrated simulated laser etching microgroove image of the chip substrate, obtaining a plurality of simulated etching microgroove images, and identifying and obtaining a plurality of microgroove bottom roughnesses; performing individual microgroove electrical performance influence analysis respectively according to the plurality of microgroove bottom roughnesses, obtaining a plurality of microgroove electrical performance influence parameters, and performing continuous microgroove electrical performance influence analysis in combination with the plurality of microgroove bottom roughnesses, obtaining a continuous electrical performance influence parameter; based on the plurality of microgroove electrical performance influence parameters and the continuous electrical performance influence parameter, calculating and obtaining an etching score of the laser control parameters, and optimizing the laser control parameters based on the etching score, obtaining an optimal laser etching parameter, and controlling the chip laser etching processing, achieving the technical goal of accurately analyzing and optimizing the relationship between the laser etching parameters and the microgroove electrical performance, and achieving the technical effect of effectively controlling the roughness of the microgroove bottom, thereby improving the electrical performance stability of the chip and the signal transmission efficiency.

[0070] Embodiment 2: Based on the intelligent control method for laser etching for chip processing in the previous embodiment, the present application also provides an intelligent control device for laser etching for chip processing, as shown in the attached Figure 2 ,include:

[0071] The laser control parameter generation module 11 is used to obtain the laser control parameter space for chip laser etching processing and randomly generate laser control parameters in the laser control parameter space; the roughness recognition module 12 is used to generate an integrated simulated laser etching microgroove image for the chip substrate according to the laser control parameters, obtain multiple simulated etching microgroove images, and identify and obtain multiple microgroove bottom roughnesses; the electrical performance influence analysis module 13 is used to perform individual microgroove electrical performance influence analysis according to the multiple microgroove bottom roughnesses, obtain multiple microgroove electrical performance influence parameters, and perform continuous microgroove electrical performance influence analysis in combination with the multiple microgroove bottom roughnesses to obtain continuous electrical performance influence parameters; the laser etching module 14 is used to calculate the etching score of the laser control parameters based on the multiple microgroove electrical performance influence parameters and the continuous electrical performance influence parameters, and optimize the laser control parameters based on the etching score to obtain the optimal laser etching parameters and perform chip laser etching processing control.

[0072] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0073] The adjustment ranges of laser power, pulse frequency and scanning speed in chip laser etching processing are obtained, and the laser control parameter space is obtained by combining them; and laser control parameters are randomly generated in the laser control parameter space.

[0074] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0075] Based on the historical data of chip laser etching, a sample laser control parameter set is collected, and the chip microgroove scanning images after etching with different sample laser control parameters are collected to obtain a sample etching microgroove image set; the sample laser control parameter set and the sample etching microgroove image set are used to train an etching microgroove image simulator, and the etching microgroove image simulator includes multiple integrated etching microgroove image simulation branches; the laser control parameters are respectively input into multiple etching microgroove image simulation branches in the etching microgroove image simulator to generate multiple simulated etching microgroove images.

[0076] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0077] The sample laser control parameter set and the sample etching microgroove image set are used as an etching simulation construction data set, and a first etching simulation training data set is randomly selected from the etching simulation construction data set; based on a generative adversarial network, a first etching microgroove image simulation branch is constructed, and the etching microgroove image simulation branch includes a first etching microgroove image generation path and a first etching microgroove image discrimination path; using the first etching simulation training data set, the first etching microgroove image simulation branch is supervised and trained until convergence; continuing to randomly select etching simulation training data sets from the etching simulation construction data set, and training multiple etching microgroove image simulation branches to obtain an etching microgroove image simulator.

[0078] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0079] According to the microgroove roughness detection data of chip etching, a set of sample etched microgroove images is collected, and the maximum peak-to-valley height difference of the microgroove bottom in the sample etched microgroove image is marked as the sample roughness to obtain the sample microgroove bottom roughness set; the sample etched microgroove image set and the sample microgroove bottom roughness set are used to train a bottom roughness identifier based on a convolutional neural network; the bottom roughness identifier is used to perform microgroove bottom roughness identification on the multiple simulated etching microgroove images to obtain multiple microgroove bottom roughnesses.

[0080] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0081] According to the electrical performance test data of the etched microgrooves, a set of sample microgroove bottom roughness is collected, and the variation amplitude of the microgroove bottom contact resistance under different sample microgroove bottom roughness is marked as the sample microgroove electrical performance influencing parameter, and the sample microgroove electrical performance influencing parameter set is obtained; the sample microgroove bottom roughness set and the sample microgroove electrical performance influencing parameter set are used to construct a microgroove electrical performance influencing analyzer, and the individual microgroove electrical performance influencing analysis is performed on the multiple microgroove bottom roughnesses, and multiple microgroove electrical performance influencing parameters are obtained; multiple sample microgroove bottom roughness sets of laser etching of multiple sample chips are collected, and the signal loss ratio of signal transmission of the etched microgrooves under different sample microgroove bottom roughness sets is tested, and the signal loss ratio is marked as the sample continuous electrical performance influencing parameter, and the sample continuous electrical performance influencing parameter set is obtained; the multiple sample microgroove bottom roughness sets and the sample continuous electrical performance influencing parameter set are used to construct a continuous microgroove electrical performance influencing analyzer, and the continuous microgroove electrical performance influencing analysis is performed on the multiple microgroove bottom roughnesses, and the continuous electrical performance influencing parameters are obtained.

[0082] Furthermore, the intelligent control device for laser etching for chip processing is also used for:

[0083] Perform weighted calculation on the multiple microgroove electrical performance influencing parameters and the continuous electrical performance influencing parameters to obtain the total chip electrical performance influencing parameters; input the total chip electrical performance influencing parameters into a laser etching score classifier, and classify them to obtain an etching score, wherein the laser etching score classifier is constructed by a mapping relationship between multiple sample total chip electrical performance influencing parameters and multiple sample etching scores, and the sample total chip electrical performance influencing parameters and the sample etching scores are negatively correlated; continue to randomly generate and optimize the laser etching control parameters in the laser control parameter space until convergence, and output the laser control parameters with the largest etching score as the optimal laser etching parameters.

[0084] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The intelligent control method and specific examples of laser etching for chip processing in the aforementioned embodiment 1 are also applicable to the intelligent control device for laser etching for chip processing in this embodiment. Through the aforementioned detailed description of the intelligent control method for laser etching for chip processing, those skilled in the art can clearly understand the intelligent control device for laser etching for chip processing in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0085] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0086] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent control method for laser etching for chip processing, characterized in that: The method comprises: Acquire a laser control parameter space for laser etching of a chip, and randomly generate laser control parameters in the laser control parameter space; According to the laser control parameters, an integrated simulated laser etching microgroove image is generated for the chip substrate to obtain a plurality of simulated etching microgroove images, and a plurality of microgroove bottom roughnesses are identified and obtained; According to the roughness of the bottom of the plurality of microgrooves, respectively, individual microgrooves are analyzed for electrical performance influence, and a plurality of microgrooves electrical performance influence parameters are obtained; and in combination with the roughness of the bottom of the plurality of microgrooves, continuous microgrooves are analyzed for electrical performance influence, and continuous electrical performance influence parameters are obtained; Based on the multiple microgroove electrical performance influencing parameters and continuous electrical performance influencing parameters, an etching score of the laser control parameters is calculated, and the laser control parameters are optimized based on the etching score to obtain the optimal laser etching parameters and perform chip laser etching processing control; Among them, according to the roughness of the bottom of the multiple micro-grooves, the electrical performance influence analysis of the individual micro-grooves is performed respectively to obtain multiple micro-groove electrical performance influence parameters, and combined with the roughness of the bottom of the multiple micro-grooves, the electrical performance influence analysis of the continuous micro-grooves is performed to obtain the continuous electrical performance influence parameters, including: According to the electrical performance test data of the etched microgrooves, a set of roughness values ​​of the bottom of the sample microgrooves is collected, and the variation amplitude of the contact resistance of the bottom of the microgrooves under different roughness values ​​of the bottom of the sample microgrooves is marked as a parameter affecting the electrical performance of the sample microgrooves, so as to obtain a set of parameters affecting the electrical performance of the sample microgrooves; Using the sample microgroove bottom roughness set and the sample microgroove electrical performance influencing parameter set, a microgroove electrical performance influencing analyzer is constructed, and individual microgroove electrical performance influencing analysis is performed on the plurality of microgroove bottom roughnesses to obtain a plurality of microgroove electrical performance influencing parameters; Collecting a plurality of sample microgroove bottom roughness sets of laser-etched multiple sample chips, and testing the signal loss ratio of signal transmission by etching microgrooves under different sample microgroove bottom roughness sets, marking them as sample continuous electrical performance influencing parameters, and obtaining a sample continuous electrical performance influencing parameter set; A continuous microgroove electrical performance impact analyzer is constructed by using the plurality of sample microgroove bottom roughness sets and the sample continuous electrical performance impact parameter set, and a continuous microgroove electrical performance impact analysis is performed on the plurality of microgroove bottom roughnesses to obtain continuous electrical performance impact parameters.

2. The intelligent control method for laser etching for chip processing according to claim 1, characterized in that: Acquiring a laser control parameter space for laser etching of a chip, and randomly generating laser control parameters in the laser control parameter space, including: Obtain the adjustment range of laser power, pulse frequency and scanning speed in chip laser etching processing, and combine them to obtain the laser control parameter space; The laser control parameters are randomly generated in the laser control parameter space.

3. The intelligent control method for laser etching for chip processing according to claim 1, characterized in that: According to the laser control parameters, the chip substrate is subjected to integrated simulation of laser etching microgroove image generation to obtain a plurality of simulated etching microgroove images, including: Based on the historical data of chip laser etching, a sample laser control parameter set is collected, and the scanning images of chip microgrooves after etching with different sample laser control parameters are collected to obtain a sample etching microgroove image set; Using the sample laser control parameter set and the sample etching microgroove image set, training an etching microgroove image simulator, wherein the etching microgroove image simulator includes an integrated plurality of etching microgroove image simulation branches; The laser control parameters are respectively input into a plurality of etching microgroove image simulation branches in the etching microgroove image simulator to generate a plurality of simulated etching microgroove images.

4. The intelligent control method for laser etching for chip processing according to claim 3, characterized in that: Using the sample laser control parameter set and the sample etching microgroove image set, training an etching microgroove image simulator includes: The sample laser control parameter set and the sample etching microgroove image set are used as an etching simulation construction data set, and a first etching simulation training data set is randomly selected from the etching simulation construction data set; Based on a generative adversarial network, a first etching microgroove image simulation branch is constructed, wherein the etching microgroove image simulation branch includes a first etching microgroove image generation path and a first etching microgroove image discrimination path; Using the first etching simulation training data set, supervised training is performed on the first etching microgroove image simulation branch until convergence; Continue to randomly select etching simulation training data sets in the etching simulation construction data set, and train multiple etching microgroove image simulation branches to obtain an etching microgroove image simulator.

5. The intelligent control method for laser etching for chip processing according to claim 1, characterized in that: Identify and obtain multiple micro-groove bottom roughness, including: According to the microgroove roughness detection data of chip etching, a sample etching microgroove image set is collected, and the maximum peak-to-valley height difference of the microgroove bottom in the sample etching microgroove image is marked as the sample roughness to obtain the sample microgroove bottom roughness set; Using the sample etching microgroove image set and the sample microgroove bottom roughness set, a bottom roughness identifier is trained based on a convolutional neural network; The bottom roughness identifier is used to identify the bottom roughness of the microgrooves of the plurality of simulated etched microgrooves to obtain the bottom roughness of the plurality of microgrooves.

6. The intelligent control method for laser etching for chip processing according to claim 1, characterized in that: Based on the multiple microgroove electrical performance influencing parameters and continuous electrical performance influencing parameters, an etching score of the laser control parameter is calculated, and the laser control parameter is optimized based on the etching score, including: Performing weighted calculation on the plurality of micro-groove electrical performance influencing parameters and continuous electrical performance influencing parameters to obtain total chip electrical performance influencing parameters; Input the total chip electrical performance influencing parameters into a laser etching score classifier, and classify to obtain an etching score, wherein the laser etching score classifier is constructed by a mapping relationship between a plurality of sample total chip electrical performance influencing parameters and a plurality of sample etching scores, and the sample total chip electrical performance influencing parameters and the sample etching score are negatively correlated; The laser etching control parameters are continuously randomly generated and optimized in the laser control parameter space until convergence, and the laser control parameters with the largest etching score are output as the optimal laser etching parameters.

7. An intelligent control device for laser etching for chip processing, characterized in that: The steps for implementing the intelligent control method for laser etching for chip processing according to any one of claims 1 to 6 include: A laser control parameter generation module, used to obtain a laser control parameter space for laser etching processing of a chip, and randomly generate laser control parameters in the laser control parameter space; A roughness recognition module is used to generate an integrated simulated laser-etched microgroove image for the chip substrate according to the laser control parameters, obtain a plurality of simulated etched microgroove images, and recognize and obtain the roughness of the bottom of a plurality of microgrooves; An electrical performance impact analysis module is used to perform individual microgroove electrical performance impact analysis according to the roughness of the bottom of the multiple microgrooves, respectively, to obtain multiple microgroove electrical performance impact parameters, and to perform continuous microgroove electrical performance impact analysis in combination with the roughness of the bottom of the multiple microgrooves, to obtain continuous electrical performance impact parameters; The laser etching module is used to calculate the etching score of the laser control parameters based on the multiple micro-groove electrical performance influencing parameters and the continuous electrical performance influencing parameters, and optimize the laser control parameters based on the etching score to obtain the optimal laser etching parameters and perform chip laser etching processing control.

Images