CN117314926A - Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus - Google Patents
Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus Download PDFInfo
- Publication number
- CN117314926A CN117314926A CN202311621051.1A CN202311621051A CN117314926A CN 117314926 A CN117314926 A CN 117314926A CN 202311621051 A CN202311621051 A CN 202311621051A CN 117314926 A CN117314926 A CN 117314926A
- Authority
- CN
- China
- Prior art keywords
- wafer
- layer
- laser
- images
- different depths
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 94
- 238000012545 processing Methods 0.000 title claims abstract description 79
- 230000004048 modification Effects 0.000 title claims abstract description 48
- 238000012986 modification Methods 0.000 title claims abstract description 48
- 238000012423 maintenance Methods 0.000 title claims abstract description 36
- 238000003860 storage Methods 0.000 title claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 62
- 238000012549 training Methods 0.000 claims abstract description 6
- 238000010586 diagram Methods 0.000 claims description 19
- 238000011176 pooling Methods 0.000 claims description 17
- 230000015654 memory Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 235000012431 wafers Nutrition 0.000 description 39
- 239000000463 material Substances 0.000 description 8
- 238000012790 confirmation Methods 0.000 description 7
- 238000005520 cutting process Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 5
- 210000002569 neuron Anatomy 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000032683 aging Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000003754 machining Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 239000004065 semiconductor Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010329 laser etching Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013386 optimize process Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30148—Semiconductor; IC; Wafer
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Business, Economics & Management (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
Abstract
The invention discloses a method, equipment and a storage medium for confirming maintenance of laser modification processing equipment, wherein the method comprises the steps of carrying out laser modification processing on a wafer under the adjusted technological parameters, and obtaining images of different depths of the wafer; taking images of different depths of the wafer as input samples, and taking a corresponding set of process parameters as output samples to construct a sample data set; constructing a parameter prediction model, and training the parameter prediction model by using a sample data set to obtain a target parameter prediction model; in the prediction process, laser modification processing is carried out on the wafer under the adjusted process parameters, images of different depths of the wafer are obtained, and the images of different depths of the wafer are predicted by utilizing a target parameter prediction model to obtain predicted process parameters; and comparing the adjusted and optimized technological parameters with predicted technological parameters to determine whether the laser modification processing equipment needs maintenance or not. The invention does not require a shutdown and a professional serviceman determines whether the device requires maintenance.
Description
Technical Field
The invention belongs to the technical field of prediction of laser modification processing technological parameters, and particularly relates to a confirmation method for maintenance of laser modification processing equipment.
Background
The laser internal modification cutting process is an important processing technology in the chip manufacturing process. To improve the processing efficiency of chips, chips are typically etched on a circular substrate in a dense array, which is called a wafer. The necessary step before the chips become marketable products is to cut the chips on the wafer by knife wheel dicing or laser etching.
The laser internal modification cutting process is a laser processing process which is developed for cutting the wafer, and can achieve the cutting effect with high precision and high quality by focusing laser in the wafer and then adopting a splitting mode to enable the wafer to generate natural cracks.
At present, in the semiconductor integrated production process, laser modification processing equipment for performing a laser internal modification cutting process often needs to work continuously, and long-time operation can cause ageing or part fatigue of the equipment, so that the cutting effect of a wafer can be influenced; meanwhile, the problems of ageing, part fatigue and the like of the equipment are not easy to observe, and the equipment is usually required to be shut down for maintenance and can be discovered through the operation of professional maintenance personnel.
In the academic world, the action mechanism of the laser modification processing on the semiconductor material is not completely clear, the technological parameters of the laser modification processing are various, the influence relationship between the technological parameters and the processing effect is very auxiliary, the processing effect is often displayed in a single picture form, the information amount is not deep, and the action mechanism is difficult to find.
Disclosure of Invention
The invention aims to provide a method, equipment and a storage medium for confirming maintenance of laser modification processing equipment, which are used for solving the problems that equipment detection and maintenance time is long and labor cost is high due to the fact that the traditional processing equipment is required to be stopped for maintenance and professional maintenance personnel work.
The invention solves the technical problems by the following technical scheme: a confirmation method for maintenance of laser modification processing equipment comprises the following steps:
carrying out laser modification processing on the wafer under the adjusted technological parameters, and acquiring images of different depths of the wafer in the laser modification processing process, so as to obtain the images of different depths of the wafer under different groups of industrial parameters;
taking images of different depths of the wafer as input samples, and taking a corresponding set of process parameters as output samples to construct a sample data set;
the method comprises the steps of constructing a parameter prediction model, wherein the parameter prediction model comprises a VGG-16 model and an LSTM model, a feature layer of the VGG-16 model is connected with an input layer of the LSTM model, and a full connection layer is added at the rear end of the LSTM model;
training the parameter prediction model by using the sample data set to obtain a target parameter prediction model;
in the prediction process, the wafer is subjected to laser modification processing under the adjusted process parameters, images of different depths of the wafer are obtained in the laser modification processing process, and the images of different depths of the wafer are predicted by utilizing the target parameter prediction model to obtain predicted process parameters;
and comparing the adjusted and optimized technological parameters with predicted technological parameters to determine whether the laser modification processing equipment needs maintenance or not.
Further, in the laser modification processing process, images of different depths of the wafer are obtained by changing the position of the industrial camera, which specifically includes:
the industrial camera is arranged above the wafer, and for a certain processing position on the wafer, the surface of the processing position is used as a focusing point of the industrial camera;
the position of the industrial camera in the vertical direction is changed, so that the focusing depth of the industrial camera in the wafer is changed, and images corresponding to the focusing depth are acquired by the industrial camera every time the position of the industrial camera is changed, so that images of different depths of the wafer are obtained.
Further, the position of the industrial camera in the vertical direction is changed by using a stepping motor.
Further, before the sample data set is constructed, format conversion and size adjustment processing are further performed on each image, so that the processed image accords with the input mode of the VGG-16 model.
Further, the VGG-16 model comprises 5 convolution layers, 4 pooling layers and 1 feature layer, and the connection sequence is as follows: the device comprises a first convolution layer, a first pooling layer, a second convolution layer, a second pooling layer, a third convolution layer, a third pooling layer, a fourth convolution layer, a fourth pooling layer, a fifth convolution layer and a feature layer; the first convolution layer and the second convolution layer employ 2 3×3 convolution kernels, and the third convolution layer, the fourth convolution layer, and the fifth convolution layer employ 3×3 convolution kernels.
Further, the LSTM model comprises an input layer, an LSTM layer, a full connection layer and an output layer which are sequentially connected.
Further, the confirmation method further includes:
selecting one of a set of process parameters;
modifying the output layer category of the Grad-Cam algorithm into selected technological parameters, modifying the image data path into a storage path of a sample data set, and obtaining a thermodynamic diagram corresponding to the selected technological parameters by utilizing the Grad-Cam algorithm;
repeating the steps of selecting the technological parameters and obtaining the thermodynamic diagrams to obtain the thermodynamic diagrams corresponding to each technological parameter.
Based on the same conception, the invention also provides an electronic device, comprising:
a memory for storing a computer program;
and a processor for implementing the confirmation method of the maintenance of the laser modification processing equipment when executing the computer program.
Based on the same conception, the present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of confirming maintenance of a laser upgrading processing apparatus as described above.
Advantageous effects
Compared with the prior art, the invention has the advantages that:
the parameter prediction model constructed by the invention is a mixed neural network model formed by a VGG-16 model and an LSTM model, the model expands the input of the LSTM model through the processing of the VGG-16 model, a wider input interface is provided, the LSTM model is utilized to uniformly process the image data with sequential relation, the association relation between the image data can be found, and the prediction precision of the technological parameters is ensured; and comparing the process parameters after the real-time optimization with the predicted process parameters, and determining whether the laser modification processing equipment needs maintenance according to the comparison result by taking the predicted process parameters as the process parameters when the equipment is not aged or the parts are not tired, wherein the laser modification processing equipment does not need to be shut down and is detected and determined by a professional maintenance staff, thereby greatly reducing the cost of equipment maintenance determination, improving the equipment maintenance confirmation efficiency and realizing the on-line confirmation of the equipment maintenance.
Drawings
In order to more clearly illustrate the technical solutions of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawing in the description below is only one embodiment of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for confirming maintenance of laser modifying processing equipment in an embodiment of the invention;
FIG. 2 is a schematic view of an industrial camera image acquisition in an embodiment of the present invention;
FIG. 3 is a network architecture diagram of a parametric prediction model in an embodiment of the present invention;
FIG. 4 is a diagram of a VGG-16 model network architecture in an embodiment of the invention;
FIG. 5 is a diagram of an LSTM model network architecture in an embodiment of the invention;
FIG. 6 is a diagram of a thermodynamic diagram obtained by applying Grad-Cam algorithm in an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made more apparent and fully by reference to the accompanying drawings, in which it is shown, however, only some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The technical scheme of the present application is described in detail below with specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
As shown in fig. 1, the method for confirming maintenance of a laser modification processing apparatus according to the present embodiment includes the following steps:
step 1: and acquiring image data.
Since the laser internally modified cut SiC wafer material is transparent, there is typically no obvious machining mark on the surface of the internally modified cut material, and thus it is necessary to extract machining features from the interior of the material. The single wafer has a plurality of cutting points (i.e. processing positions), and usually, the plurality of processing positions of the single wafer all adopt the same set of optimized processing parameters, and the same set of optimized processing parameters are also adopted for the wafers in the same batch.
In order to obtain the processing characteristics of the material, a process personnel is required to perform adjustment and optimization of the process parameters of the laser modification processing equipment, a group of process parameters (i.e. a group of process parameters after adjustment and optimization) are determined, the laser modification processing equipment is used for performing laser modification processing on the wafer under the group of process parameters, and images of different depths of the wafer are obtained in the laser modification processing process.
In this embodiment, in the laser modification processing process, images of different depths of the wafer are obtained by changing the position of the industrial camera, as shown in fig. 2, and the specific operations are as follows: the method comprises the steps that an industrial camera is arranged above a certain processing position of a wafer to be processed, and a lens of the industrial camera is aligned to the processing position and focused on the surface of the processing position; the industrial camera is controlled by the stepping motor to step downwards in the vertical direction, and the industrial camera is used for acquiring images once in each step, so that images of different depths of the wafer can be acquired when the position of the industrial camera moves downwards due to the fact that the focusing depth of the lens is unchanged, and the images of different depths of the wafer can be obtained in the laser modification processing process. For other optimized process parameters, images of different depths of the wafer can be obtained in the laser modification processing process, so that images of different depths of the wafer under different industrial parameters can be obtained.
For example, if the stepping motor is set to step by 0.156um each time, the first image collected by the industrial camera is a cross-sectional image of the wafer at 0.156um (the distance from the upper surface of the wafer is 0.156 um), the second image is a cross-sectional image of the wafer at 0.312um, and so on, multiple images with different depths can be obtained at each processing position of the wafer under a set of process parameters, the number of images is related to the thickness of the wafer, the number of images is generally 200-400 (the number of images in the embodiment is 200), and the obtained image format is tiff.
And storing the images under the same set of process parameters into the same folder, naming each image according to the stepping times, and arranging the images according to the acquisition sequence to obtain the processing characteristic image data under a certain set of process parameters. In this embodiment, the process parameters include laser power, laser frequency, processing speed, and depth of focus. Exemplary, the set of process parameters after tuning are specifically: the method comprises the steps that the laser Power is 800A current, the laser Frequency is 50kHz, the processing Speed is 400mm/s, the focusing Depth is 50um, the folder name is Power800Frequency50Speed400Depth50, 200 images exist in the folder, and each image is named according to the acquisition sequence and 1-200 images.
Step 2: the image data is preprocessed.
The format of the image collected by the industrial camera is tiff, the pixel size is 2592 multiplied by 1944, and in order to enable the image data to meet the input mode requirement of a follow-up parameter prediction model, format conversion and size adjustment processing are carried out on each image, so that the processed image meets the input mode of a VGG-16 model. In this embodiment, the python program is used to traverse each image in the folder, and convert each image into jpg format, and the pixel size is 224-224. The preprocessing does not have excessive loss, and the processing speed of the parameter prediction model can be improved.
Step 3: and (3) constructing a sample data set.
And (3) constructing a sample data set based on the preprocessed images in the step (2), wherein in the sample data set, images with different depths of the wafer are used as input samples, and a group of process parameters corresponding to the input samples are used as output samples.
Step 4: and constructing a parameter prediction model.
The parameter prediction model is a hybrid neural network of a VGG-16 model and an LSTM model, a feature layer of the VGG-16 model is connected with an input layer of the LSTM model, and a full connection layer is added at the rear end of the LSTM model, as shown in figure 3. As shown in FIG. 4, the VGG-16 model is a convolutional neural network model proposed by Simonyan and Zisselman of Vsiual Geometry Grop group in document Very Deep Convolutional Networks for Large Scale Image Recognition (literature details: simonyan K, zisselman A. Very deep convolutional networks for large-scale image recognition [ J ]. ArXiv preprint arXiv:1409.1556, 2014.), the VGG-16 model includes 5 convolutional layers, 4 pooled layers and 1 feature layer, connected in the order: the method comprises the steps of a first convolution layer 1, a first pooling layer 2, a second convolution layer 3, a second pooling layer 4, a third convolution layer 5, a third pooling layer 6, a fourth convolution layer 7, a fourth pooling layer 8, a fifth convolution layer 9 and a feature layer 10; wherein the first convolution layer 1 and the second convolution layer 3 adopt 2 convolution kernels of 3×3, and the third convolution layer 5, the fourth convolution layer 7 and the fifth convolution layer 9 adopt 3 convolution kernels of 3×3; the first convolution layer 1 and the second convolution layer 3 are 2-channels, and the third convolution layer 5, the fourth convolution layer 7, and the fifth convolution layer 9 are 3-channels. A pooling layer is arranged between every two convolution layers, and the pooling layer is used for averaging pixel values. The input of the VGG-16 model is a single image in a sample data set, after being processed by a plurality of convolution layers and pooling layers, a 7 multiplied by 512 tensor matrix is output at a characteristic layer, and then output vectors are obtained through one-dimensional flattening (tensor flattening 11).
In the VGG-16 model, image data is input in a single form, and a single one-dimensional output vector with a length of L is obtained after inference and flattening, so that i one-dimensional output vectors (i=200 in the embodiment) can be obtained, and the i one-dimensional output vectors are overlapped in parallel to obtain a matrix in an [ i, L ] form.
As shown in fig. 5, the LSTM model (Long Short-Term Memory) includes an input layer, an LSTM layer, a full-connection layer and an output layer which are sequentially connected, the input format is a matrix of 200× 25088, the matrix is connected with 64 neurons of the LSTM layer by a full-connection mode, the neurons of the LSTM layer have a Memory and forgetting function, the last input result parameters are synchronized into the current process in each training process, the 64 neurons of the LSTM layer are connected with the full-connection layer of 100 neurons of the next layer by the full-connection mode, the full-connection layer is finally directly connected with the output layer, and the output layer has 4 neurons which respectively represent 4 process parameters, namely, the LSTM model finally outputs predicted process parameters.
Compared with the VGG-16 model, the parameter prediction model provides a wider input interface, the subsequent LSTM model can extract the sequence relation in the image data, the relation between the image data can be mined, and the prediction precision of the technological parameters is ensured.
Step 5: training of a parameter prediction model.
And training the parameter prediction model by using the sample data set to obtain a target parameter prediction model, and predicting the technological parameter by using the target parameter prediction model to obtain predicted technological parameters.
Step 6: and (5) predicting process parameters.
During prediction, laser modification processing is carried out on the wafer under the adjusted process parameters, images of different depths of the wafer are obtained in the laser modification processing process, and the images of different depths of the wafer are predicted by utilizing a target parameter prediction model to obtain predicted process parameters.
Step 7: whether to maintain a confirmation.
And comparing the adjusted and optimized technological parameters with predicted technological parameters to determine whether the laser modification processing equipment needs maintenance or not. In order to determine whether the laser modification processing equipment needs maintenance or not, comparing the technological parameters after the current adjustment with the technological parameters after the adjustment when the equipment is not aged and/or parts are not fatigued, and if the difference between the technological parameters is large, indicating that the equipment has aging and/or parts are fatigued, and stopping maintenance is needed; if the difference is small, the equipment is not aged and/or parts are fatigued, and the equipment can continue to operate.
In this embodiment, the predicted process parameters are used as the process parameters after the equipment is not aged and/or the parts are not fatigued, so the image data acquisition in the step 1 needs to be performed under the conditions that the laser modification processing equipment is not aged and the parts are not fatigued (i.e. the laser modification processing equipment is normal), and the process parameters predicted by the target parameter prediction model are the process parameters of the laser modification processing equipment under the normal condition.
The difference between the adjusted and predicted process parameters can be determined according to a set threshold, if the difference is larger than the set threshold, the difference is larger, otherwise, the difference is smaller. The set threshold may be empirically set.
Step 8: thermodynamic diagrams of different process parameters are acquired.
The action mechanism of the laser modification processing on the semiconductor material is not completely clear, in order to study the influence of the technological parameters on the processing characteristics, thermodynamic diagrams corresponding to different technological parameters are obtained, as shown in fig. 6, the specific process is as follows:
step 8.1: selecting one of a set of process parameters;
step 8.2: modifying the output layer category of the Grad-Cam algorithm (namely Gradient-weighted Class Activation Mapping, gradient weighted class activation heat map) into selected technological parameters, modifying the image data path into a storage path of a sample data set, and obtaining a thermodynamic diagram corresponding to the selected technological parameters by utilizing the Grad-Cam algorithm;
step 8.3: repeating the steps 8.1 and 8.2 to obtain a thermodynamic diagram corresponding to each technological parameter;
step 8.4: the thermodynamic diagram is visually displayed on a graphical interface through a drawing function of OpenCV.
The technological parameters include laser power, laser frequency, processing speed and focusing depth, and by taking laser power as an example, the output layer category of Grad-Cam algorithm is modified to laser power, the image data path is modified to the storage path of the sample data set, and the thermodynamic diagram of the laser power is obtained by utilizing Grad-Cam algorithm. In this embodiment, the number of images under a set of process parameters is 200, and the number of thermodynamic diagrams of laser power that can be obtained is also 200, and the color (i.e. the marking area) of each thermodynamic diagram is different, which indicates that the influence of the process parameter of laser power on different depths inside the material is different. And similarly, the influence of other process parameters on different depths inside the material can be obtained.
The gradient of the technological parameter relative to the image data can be calculated by using the Grad-Cam algorithm, the gradient shows the change condition of a certain technological parameter score relative to the image data, the influence degree of the technological parameter on different depths is reflected, the basis of model prediction can be understood visually by using the Grad-Cam algorithm, and the interpretability and the credibility of the model are improved.
The embodiment of the invention also provides electronic equipment, which comprises: a processor and a memory storing a computer program, the processor being configured to implement a method of validating maintenance of a laser modifying machining apparatus as described above when the computer program is executed.
Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes according to programs and/or data stored in a Read Only Memory (ROM) or programs and/or data loaded from a storage portion into a Random Access Memory (RAM). The processor may be a multi-core processor or may include a plurality of processors. In some embodiments, the processor may comprise a general-purpose main processor and one or more special coprocessors, such as, for example, a Central Processing Unit (CPU), a Graphics Processor (GPU), a neural Network Processor (NPU), a Digital Signal Processor (DSP), and so forth. In the RAM, various programs and data required for the operation of the electronic device are also stored. The processor, ROM and RAM are connected to each other by a bus. An input/output (I/O) interface is also connected to the bus.
The above-described processor is used in combination with a memory to execute a program stored in the memory, which when executed by a computer is capable of implementing the methods, steps or functions described in the above-described embodiments.
Although not shown, embodiments of the present invention also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of validating maintenance of a laser upgrading processing apparatus as described above.
Storage media in embodiments of the invention include both permanent and non-permanent, removable and non-removable items that may be used to implement information storage by any method or technology. Examples of storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, read only compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape storage or other magnetic storage devices, or any other non-transmission medium which can be used to store information that can be accessed by a computing device.
The foregoing disclosure is merely illustrative of specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art will readily recognize that changes and modifications are possible within the scope of the present invention.
Claims (9)
1. A method for confirming maintenance of laser-modified processing equipment, the method comprising the steps of:
carrying out laser modification processing on the wafer under the adjusted technological parameters, and acquiring images of different depths of the wafer in the laser modification processing process, so as to obtain the images of different depths of the wafer under different groups of industrial parameters;
taking images of different depths of the wafer as input samples, and taking a corresponding set of process parameters as output samples to construct a sample data set;
the method comprises the steps of constructing a parameter prediction model, wherein the parameter prediction model comprises a VGG-16 model and an LSTM model, a feature layer of the VGG-16 model is connected with an input layer of the LSTM model, and a full connection layer is added at the rear end of the LSTM model;
training the parameter prediction model by using the sample data set to obtain a target parameter prediction model;
in the prediction process, the wafer is subjected to laser modification processing under the adjusted process parameters, images of different depths of the wafer are obtained in the laser modification processing process, and the images of different depths of the wafer are predicted by utilizing the target parameter prediction model to obtain predicted process parameters;
and comparing the adjusted and optimized technological parameters with predicted technological parameters to determine whether the laser modification processing equipment needs maintenance or not.
2. The method for confirming maintenance of a laser upgrading apparatus according to claim 1, wherein the step of obtaining images of different depths of the wafer by changing a position of an industrial camera during the laser upgrading process comprises:
the industrial camera is arranged above the wafer, and for a certain processing position on the wafer, the surface of the processing position is used as a focusing point of the industrial camera;
the position of the industrial camera in the vertical direction is changed, so that the focusing depth of the industrial camera in the wafer is changed, and images corresponding to the focusing depth are acquired by the industrial camera every time the position of the industrial camera is changed, so that images of different depths of the wafer are obtained.
3. The method for confirming maintenance of a laser upgrading apparatus according to claim 2, wherein the position of the industrial camera in the vertical direction is changed by a stepping motor.
4. The method of claim 1, wherein prior to constructing the sample data set, performing format conversion and size adjustment processing on each image to conform the processed image to the VGG-16 model input mode.
5. The method for confirming maintenance of a laser modifying apparatus according to claim 1, wherein the VGG-16 model comprises 5 convolution layers, 4 pooling layers and 1 feature layer, and the connection order is: the device comprises a first convolution layer, a first pooling layer, a second convolution layer, a second pooling layer, a third convolution layer, a third pooling layer, a fourth convolution layer, a fourth pooling layer, a fifth convolution layer and a feature layer; the first convolution layer and the second convolution layer employ 2 3×3 convolution kernels, and the third convolution layer, the fourth convolution layer, and the fifth convolution layer employ 3×3 convolution kernels.
6. The method of claim 1, wherein the LSTM model includes an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence.
7. The method for confirming maintenance of a laser-modified processing apparatus according to any one of claims 1 to 6, characterized in that the method further comprises:
selecting one of a set of process parameters;
modifying the output layer category of the Grad-Cam algorithm into selected technological parameters, modifying the image data path into a storage path of a sample data set, and obtaining a thermodynamic diagram corresponding to the selected technological parameters by utilizing the Grad-Cam algorithm;
repeating the steps of selecting the technological parameters and obtaining the thermodynamic diagrams to obtain the thermodynamic diagrams corresponding to each technological parameter.
8. An electronic device, the device comprising:
a memory for storing a computer program;
a processor for implementing the method for confirming maintenance of the laser modification processing apparatus according to any one of claims 1 to 7 when executing the computer program.
9. A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program when executed by a processor implements the method for confirming maintenance of the laser upgrading apparatus according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311621051.1A CN117314926B (en) | 2023-11-30 | 2023-11-30 | Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311621051.1A CN117314926B (en) | 2023-11-30 | 2023-11-30 | Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117314926A true CN117314926A (en) | 2023-12-29 |
| CN117314926B CN117314926B (en) | 2024-01-30 |
Family
ID=89288836
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311621051.1A Active CN117314926B (en) | 2023-11-30 | 2023-11-30 | Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117314926B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004212067A (en) * | 2002-12-26 | 2004-07-29 | Olympus Corp | Defect inspecting apparatus and defect inspection method |
| US20090063378A1 (en) * | 2007-08-31 | 2009-03-05 | Kla-Tencor Technologies Corporation | Apparatus and methods for predicting a semiconductor parameter across an area of a wafer |
| CN108875113A (en) * | 2017-09-07 | 2018-11-23 | 湖南大学 | Fusion penetration prediction and control method and the system of the cross sectional plates that thicken laser welding |
| CN109491216A (en) * | 2018-12-20 | 2019-03-19 | 上海集成电路研发中心有限公司 | A method of optimization lithography process parameters |
| CN109902558A (en) * | 2019-01-15 | 2019-06-18 | 安徽理工大学 | A kind of human health deep learning prediction technique based on CNN-LSTM |
| CN111863645A (en) * | 2019-04-12 | 2020-10-30 | 长鑫存储技术有限公司 | Machining process parameter determination method and device, storage medium and electronic equipment |
| CN113887328A (en) * | 2021-09-10 | 2022-01-04 | 天津理工大学 | Method for extracting space-time characteristics of photonic crystal space transmission spectrum in parallel by ECA-CNN fusion dual-channel RNN |
| CN114925596A (en) * | 2022-04-20 | 2022-08-19 | 永得利科技（无锡）有限公司 | Method for optimizing casting platform finish milling parameters based on wafer test equipment |
| CN115482201A (en) * | 2022-08-24 | 2022-12-16 | 首都师范大学 | Bi-LSTM neural network-based dual-model defect detection method, training method and device |
| CN115760837A (en) * | 2022-12-06 | 2023-03-07 | 泉州装备制造研究所 | Crystal quality evaluation method and system based on deep neural network |
| US20230153640A1 (en) * | 2020-03-31 | 2023-05-18 | Siemens Aktiengesellschaft | Prediction Model Learning Method, Apparatus and System for an Industrial System |
| CN116822342A (en) * | 2023-06-13 | 2023-09-29 | 华中科技大学 | Process identification and performance prediction method for zirconium alloy laser cutting |
| CN117001188A (en) * | 2023-07-26 | 2023-11-07 | 北京中科镭特电子有限公司 | Wafer laser dicing method, dicing system, process apparatus, and readable storage medium |
-
2023
- 2023-11-30 CN CN202311621051.1A patent/CN117314926B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004212067A (en) * | 2002-12-26 | 2004-07-29 | Olympus Corp | Defect inspecting apparatus and defect inspection method |
| US20090063378A1 (en) * | 2007-08-31 | 2009-03-05 | Kla-Tencor Technologies Corporation | Apparatus and methods for predicting a semiconductor parameter across an area of a wafer |
| CN108875113A (en) * | 2017-09-07 | 2018-11-23 | 湖南大学 | Fusion penetration prediction and control method and the system of the cross sectional plates that thicken laser welding |
| CN109491216A (en) * | 2018-12-20 | 2019-03-19 | 上海集成电路研发中心有限公司 | A method of optimization lithography process parameters |
| CN109902558A (en) * | 2019-01-15 | 2019-06-18 | 安徽理工大学 | A kind of human health deep learning prediction technique based on CNN-LSTM |
| CN111863645A (en) * | 2019-04-12 | 2020-10-30 | 长鑫存储技术有限公司 | Machining process parameter determination method and device, storage medium and electronic equipment |
| US20230153640A1 (en) * | 2020-03-31 | 2023-05-18 | Siemens Aktiengesellschaft | Prediction Model Learning Method, Apparatus and System for an Industrial System |
| CN113887328A (en) * | 2021-09-10 | 2022-01-04 | 天津理工大学 | Method for extracting space-time characteristics of photonic crystal space transmission spectrum in parallel by ECA-CNN fusion dual-channel RNN |
| CN114925596A (en) * | 2022-04-20 | 2022-08-19 | 永得利科技（无锡）有限公司 | Method for optimizing casting platform finish milling parameters based on wafer test equipment |
| CN115482201A (en) * | 2022-08-24 | 2022-12-16 | 首都师范大学 | Bi-LSTM neural network-based dual-model defect detection method, training method and device |
| CN115760837A (en) * | 2022-12-06 | 2023-03-07 | 泉州装备制造研究所 | Crystal quality evaluation method and system based on deep neural network |
| CN116822342A (en) * | 2023-06-13 | 2023-09-29 | 华中科技大学 | Process identification and performance prediction method for zirconium alloy laser cutting |
| CN117001188A (en) * | 2023-07-26 | 2023-11-07 | 北京中科镭特电子有限公司 | Wafer laser dicing method, dicing system, process apparatus, and readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| XIAOYAN CHEN 等: "A Light-Weighted CNN Model for Wafer Structural Defect Detection", IEEE ACCESS, vol. 8, pages 24006 - 24018, XP011770568, DOI: 10.1109/ACCESS.2020.2970461 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117314926B (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10504006B2 (en) | Classification, search and retrieval of semiconductor processing metrology images using deep learning/convolutional neural networks | |
| US11379967B2 (en) | Methods and systems for inspection of semiconductor structures with automatically generated defect features | |
| KR102606140B1 (en) | Pattern grouping method based on machine learning | |
| US11686689B2 (en) | Automatic optimization of an examination recipe | |
| JP2010164487A (en) | Defect inspecting apparatus and defect inspecting method | |
| CN115496975B (en) | Auxiliary weighted data fusion method, device, equipment and storage medium | |
| CN115423796A (en) | Chip defect detection method and system based on TensorRT accelerated reasoning | |
| KR20210127069A (en) | Method of controlling performance of fusion model neural network | |
| CN116168033A (en) | Wafer lattice dislocation image detection method and system based on deep learning | |
| CN117314926B (en) | Method, apparatus and storage medium for confirming maintenance of laser modification processing apparatus | |
| JP2005099020A (en) | Method for analyzing structure of material using convergent beam electron diffraction | |
| CN115587989B (en) | Workpiece CT image defect detection segmentation method and system | |
| KR20220062328A (en) | Wafer inspection method and system | |
| CN116205918A (en) | Multi-mode fusion semiconductor detection method, device and medium based on graph convolution | |
| US11961221B2 (en) | Defect examination on a semiconductor specimen | |
| TW202340868A (en) | Method and system of defect detection for inspection sample based on machine learning model | |
| CN115423793A (en) | Method, device and equipment for detecting wafer dislocation defects and storage medium | |
| Cheng et al. | AtomGAN: unsupervised deep learning for fast and accurate defect detection of 2D materials at the atomic scale | |
| US20210303757A1 (en) | Semiconductor fabrication process parameter determination using a generative adversarial network | |
| CN114926488A (en) | Workpiece positioning method based on generalized Hough model and improved pyramid search acceleration | |
| US11631179B2 (en) | Segmentation of an image of a semiconductor specimen | |
| US12007335B2 (en) | Automatic optimization of an examination recipe | |
| TW202409550A (en) | Defect detection of a semiconductor specimen | |
| US20240095903A1 (en) | Image augmentation for machine learning based defect examination | |
| WO2023112302A1 (en) | Training data creation assistance device and training data creation assistance method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |