Etching parameter adjustment method and apparatus based on plasma simulation

Abstract

This invention provides a method and apparatus for adjusting etching parameters based on plasma simulation, belonging to the semiconductor field. The method includes: acquiring multiple parameter combinations, including radio frequency power, chamber pressure, and chamber temperature; determining a first selection probability corresponding to each parameter combination based on a depth model, and determining a reference subset from the multiple parameter combinations based on the first selection probability; iteratively calculating the reference subset, determining multiple target combinations based on a second selection probability corresponding to each parameter combination during the iteration, calculating etching uniformity based on the power deposition distribution simulation results of the target combinations, and updating the second selection probability and evaluation index based on the etching uniformity; and determining the parameter combination with the highest evaluation index at the end of the iteration as the optimal parameter combination. This invention improves the efficiency of plasma simulation by optimizing the plasma parameter simulation process, thereby solving the technical problem of low efficiency in adjusting etching process parameters in existing technologies.

Description

Technical Field

[0001] This application relates to the semiconductor field, specifically to a method and apparatus for adjusting etching parameters based on plasma simulation. Background Technology

[0002] In wafer etching processes, in-plane uniformity of etching is a key indicator affecting device yield. Good etching uniformity ensures that the electrical performance of chips on the same wafer is similar, thereby directly improving device yield and manufacturing stability.

[0003] In actual production, the quality of etching uniformity is affected by a variety of factors, including the chamber design of the etching equipment, the setting of process parameters, and the plasma state. To achieve ideal uniformity, it is usually necessary to rely on a large number of simulations to adjust the process parameters in the process chamber, which results in long calculation cycles, high time costs, and low efficiency.

[0004] In other words, existing technologies suffer from the problem of low efficiency in adjusting etching process parameters. Summary of the Invention

[0005] This application provides a method and apparatus for adjusting etching parameters based on plasma simulation, which at least partially solves the technical problem of low efficiency in adjusting etching process parameters in the prior art as mentioned in the background.

[0006] The first aspect of this application provides a method for adjusting etching parameters based on plasma simulation, including: Multiple parameter combinations are obtained, including RF power, chamber pressure, and chamber temperature. The first selection probability corresponding to the parameter combination is determined based on the deep model, and a reference subset is determined from multiple parameter combinations based on the first selection probability. The first selection probability is used to characterize the probability that the parameter combination is the optimal parameter combination. The reference subset is iteratively calculated. During the iteration, multiple target combinations are determined according to the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition simulation results of the target combination. The second selection probability and evaluation index are updated according to the etching uniformity. The second selection probability is used to characterize the possibility that the parameter combination is determined as the target combination. The parameter combination with the highest evaluation index at the end of the iteration is determined as the optimal parameter combination.

[0007] In one embodiment of the first aspect, after determining a reference subset from multiple parameter combinations according to a first selection probability, the method further includes: performing plasma simulation on the reference subset to determine the power deposition distribution and electron temperature distribution of each parameter combination in the reference subset; updating the reference subset according to the electron temperature distribution, wherein each parameter combination in the updated reference subset satisfies a preset electron temperature constraint.

[0008] In one embodiment of the first aspect, the iterative calculation of the reference subset includes iterative calculation of the updated reference subset, and the calculation of etching uniformity based on the power deposition simulation results of the target combination includes: obtaining the power deposition distribution of the target combination; determining a first average power deposition density of each sub-region of the wafer and a second average power deposition density of the entire wafer based on the power deposition distribution; calculating a relative deviation rate based on the first average power deposition density and the second average power deposition density, wherein the relative deviation rate is used to describe the power deposition difference of each sub-region relative to the entire wafer; and determining the etching uniformity based on the relative deviation rate.

[0009] In one embodiment of the first aspect, determining etching uniformity based on relative deviation rate includes: determining the deviation level of the corresponding sub-region based on relative deviation rate; determining the weight corresponding to different deviation levels; and determining etching uniformity based on the weighted result of the weight and relative deviation rate.

[0010] In one embodiment of the first aspect, the input to the depth model is a multi-channel feature image, wherein different channels of each pixel in the multi-channel feature image encode radio frequency power, chamber pressure, and chamber temperature, respectively.

[0011] In one embodiment of the first aspect, before determining the first selection probability corresponding to the parameter combination based on the deep model, the method further includes: training the deep model to be trained to convergence using a simulation dataset, wherein each training sample in the simulation dataset includes a reference parameter combination and a reference selection probability label determined after simulation using the reference parameter combination. In one embodiment of the first aspect, the method further includes: obtaining a reference power deposition distribution after simulation of the reference parameter combination; and determining a reference selection probability label based on the reference power deposition distribution.

[0012] In one embodiment of the first aspect, the iterative calculation of the reference subset is performed using an ant colony algorithm, wherein the second selection probability is the state transition probability, and the evaluation index is the pheromone concentration.

[0013] A second aspect of this application provides a computer device including a processor and a memory. The memory stores a computer program, which is loaded and executed by the processor to implement the above-described method for adjusting etching parameters based on plasma simulation.

[0014] A third aspect of this application provides a computer-readable storage medium storing a computer program, the instructions of which are loaded and executed by a processor to implement the above-described method for adjusting etching parameters based on plasma simulation.

[0015] A third aspect of this application provides a computer program product storing a computer program that is loaded and executed by a processor to implement the above-described method for adjusting etching parameters based on plasma simulation.

[0016] The technical solutions provided in this application include at least the following beneficial effects.

[0017] This application proposes a method for adjusting etching parameters based on plasma simulation. On one hand, the method determines the first selection probability corresponding to parameter combinations through a depth model, which can be deployed on a GPU with massively parallel computing capabilities. This leverages the GPU's massively parallel computing power to quickly determine a small reference subset from numerous parameter combinations. The parameter combinations within this reference subset are then subjected to more accurate simulation verification, thereby improving the efficiency of etching parameter adjustment. On the other hand, an automated parameter selection method is designed based on the simulation results from the reference subset. After multiple iterations, an optimal parameter combination can be identified, further improving the efficiency of etching parameter adjustment and thus solving the technical problem of low efficiency in etching process parameter adjustment in existing technologies.

[0018] This application also provides a computer device, a computer-readable storage medium, and a computer program product, which can achieve the same technical effects as described above.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0020] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a hardware structure block diagram of a terminal device for an etching parameter adjustment method based on plasma simulation, provided in an embodiment of the present invention. Figure 2 A flowchart illustrating an etching parameter adjustment method based on plasma simulation, provided as an embodiment of the present invention; Figure 3 is a schematic diagram of an etching parameter adjustment method based on plasma simulation provided in an embodiment of the present invention; Figure 4 A schematic diagram of another etching parameter adjustment method based on plasma simulation provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of an etching parameter adjustment method based on plasma simulation provided in an embodiment of the present invention; Figure 6 A flowchart of another etching parameter adjustment method based on plasma simulation provided in an embodiment of the present invention; Figure 7 A flowchart illustrating another method for adjusting etching parameters based on plasma simulation, provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of a computer device for adjusting etching parameters based on plasma simulation, as provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Those skilled in the art should understand that the embodiments described below are only some, not all, of the embodiments disclosed. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] It should be noted that the information (including but not limited to device information, personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the subject or fully authorized by all parties, and the collection, use and processing of related data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0024] To address the problems mentioned in the background, this application provides an etching parameter adjustment method based on plasma simulation. In this method, on one hand, the first selection probability corresponding to the parameter combination determined by the depth model can be deployed on a GPU with large-scale parallel computing capabilities. That is, by utilizing the large-scale parallel computing capabilities of the GPU, a small reference subset can be quickly determined from numerous parameter combinations. Then, the parameter combinations in the reference subset are subjected to more accurate simulation verification, thereby improving the efficiency of etching parameter adjustment. On the other hand, an automated parameter selection method is designed based on the simulation results of the reference subset. After multiple iterations, an optimal parameter combination can be identified, further improving the efficiency of etching parameter adjustment and thus solving the technical problem of low efficiency in etching process parameter adjustment in the prior art.

[0025] The present application will be further described in detail below with reference to several representative embodiments of the present invention.

[0026] This application provides a method for adjusting etching parameters based on plasma simulation. This method can be executed by a computer device, which can be a terminal or a server.

[0027] Taking running on a terminal device as an example, Figure 1 This is a hardware structure block diagram of a terminal device for a distributed training result evaluation method according to an embodiment of this application. For example... Figure 1 As shown, the terminal device may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a central processing unit (CPU) equipped with a neural processing unit (NPU), a microprocessor (MCU) equipped with a neural processing unit (NPU), a graphics processing unit (GPU), etc.) and a memory 104 for storing data are also shown. The terminal device may further include a data transmission module 106 for communicating with other terminal devices or server devices, and an input / output module 108. Those skilled in the art will understand that… Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal device described above. For example, the terminal device may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown, for example, a terminal device may include a central processing unit and a graphics processing unit, wherein the graphics processing unit is used to perform large-scale parallel computing tasks such as image data processing and artificial neural network calculations, and the central processing unit is used to perform other logical computing tasks.

[0028] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to a plasma simulation-based etching parameter adjustment method in this embodiment of the application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104 to implement the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory.

[0029] The data transmission module 106 is used to receive or send data via network transmission or wired transmission. For example, the data transmission module 106 can be a wireless WIFI module, which is used to communicate with other terminal devices or server devices wirelessly.

[0030] Figure 2 This is a flowchart of an etching parameter adjustment method based on plasma simulation provided in an embodiment of this application, as shown below. Figure 2 As shown, the method includes: S21, obtain multiple parameter combinations, including RF power, chamber pressure and chamber temperature; S22, determine the first selection probability corresponding to the parameter combination based on the deep model, and determine the reference subset from multiple parameter combinations according to the first selection probability. The first selection probability is used to characterize the possibility that the parameter combination is the optimal parameter combination. S23 iteratively calculates the reference subset. During the iteration, multiple target combinations are determined according to the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination. The second selection probability and evaluation index are updated according to the etching uniformity. The second selection probability is used to characterize the possibility that the parameter combination is determined as the target combination. S24 determines the parameter combination with the highest evaluation index at the end of the iteration as the optimal parameter combination.

[0031] It should be noted that the RF power, chamber pressure, and chamber temperature mentioned above are the RF power, chamber pressure, and chamber temperature that need to be selected or adjusted in a plasma etching chamber. See the schematic diagram of the plasma etching chamber above. Figure 3 , Figure 3 The plasma etching chamber 301 in this embodiment is a capacitive plasma etching chamber (CCP). Unless otherwise specified, all embodiments in this application refer to it as such. Figure 3 The capacitive plasma etching chamber (etching chamber 301) shown is explained below, and it is understood that... Figure 3 These are merely illustrative examples; the embodiments of this application can also be illustrated with... Figure 3 This can be implemented in different capacitive plasma etching chambers or other types of etching chambers (such as inductive plasma etching chambers, microwave plasma etching chambers), without specific limitations here.

[0032] Figure 3 In the process, the radio frequency power source 302 is electrically connected to the source electrode 303 and the bias electrode 304 to form an electric field. In the etching chamber 301, all surfaces in contact with the plasma form a sheath. The sheath can be regarded as a capacitor with a certain dielectric constant. That is, the applied power is transferred to the plasma through the coupling effect of a capacitor. The energy distribution of the plasma is the power deposition distribution in step S23 above.

[0033] It is worth noting that in the etching chamber 301, the power deposition is often spatially uneven, which affects the etching uniformity. For example, when the power deposition at the center of wafer 304 is higher than that at the edge, it will lead to over-etching at the center of wafer 304, and vice versa. To avoid uneven power deposition as much as possible, it is necessary to determine the power deposition distribution under a certain combination of etching parameters through simulation. In this embodiment, the optimal parameter combination needs to be determined by simulation of the power deposition distribution, which is related to the RF power, chamber pressure, and chamber temperature in step S21 above.

[0034] Specifically, due to the radio frequency electric field in such Figure 3 The etching chamber 301 shown is not completely uniform and has a boundary effect. When the applied power exceeds a certain threshold, the boundary effect will be significantly amplified, exacerbating the non-uniformity of the power deposition distribution. Pressure and temperature change the gas density and plasma diffusion ability, which in turn affect the power deposition distribution. All three factors work together. Therefore, in order to determine the power deposition distribution under a certain combination of etching parameters through simulation, it is necessary to first perform the above step S21 to obtain multiple parameter combinations, including radio frequency power, chamber pressure and chamber temperature.

[0035] It is understood that the multiple parameter combinations in step S21 above can be selected within a pre-defined search space, with one or more parameters selected according to a certain step size. Each combination is considered as one of the aforementioned parameter combinations. For example, the search space for RF power can be set to 50 to 500W, with a step size of 10W; the search space for chamber pressure can be set to 0.5 to 2 Torr, with a step size of 0.1 Torr; and the chamber temperature can be fixed at 300K. Alternatively, the search space for RF power can be set to 10 to 50W, with a step size of 1W; the search space for chamber pressure can be set to 1 to 2 Torr, with a step size of 0.2 Torr; and the search space for chamber temperature can be set to 280 to 400K, with a step size of 10K. These are merely illustrative examples and are not intended to impose specific limitations.

[0036] Furthermore, after obtaining multiple parameter combinations through the above step S21, it is necessary to determine the power deposition distribution corresponding to different parameter combinations. However, since the simulation consumes a lot of computing resources and has a long calculation cycle, and there are hundreds or even thousands of parameter combinations, simulating them one by one is inefficient. Therefore, in this embodiment, an artificial neural network model (deep model) is used to pre-select several sets of parameter combinations that may become the optimal parameter combinations from all parameter combinations. That is, the above step S22 is executed to determine the first selection probability corresponding to the parameter combination based on the deep model, and a reference subset is determined from multiple parameter combinations based on the first selection probability. The first selection probability is used to characterize the possibility of the parameter combination becoming the optimal parameter combination.

[0037] In one possible implementation, the input to the deep model is a multi-channel feature image, where different channels of each pixel in the multi-channel feature image encode radio frequency power, chamber pressure, and chamber temperature, respectively.

[0038] It is understood that the aforementioned deep model is a pre-trained neural network model, such as a convolutional neural network (CNN). To adapt to the extraction of the spatial features of the power deposition distribution, in this embodiment, the parameter combinations are encoded into a multi-channel feature image. Specifically, each parameter combination is encoded into a pixel in the aforementioned multi-channel feature image. For example, radio frequency power, chamber pressure, and chamber temperature correspond to the r, g, and b channels in a pixel, respectively. This image-based input format enables the deep model to capture the coupling relationship between radio frequency power, chamber pressure, and chamber temperature using convolution operations.

[0039] The above parameter combinations can be arranged in a certain pattern in the multi-channel feature image. For example, pixels can be filled from left to right and then from top to bottom. The three channels of pixels in the above multi-channel feature image that are not filled with parameter combinations can all be set to 0. No specific restrictions are made here.

[0040] In one possible implementation, the output of the aforementioned deep model can be a single-channel grayscale image or a sequence. When the output is a single-channel grayscale image, its resolution is the same as the input image; that is, the grayscale value of each pixel in the output image represents the first selection probability of the corresponding parameter combination. It should be noted that the magnitude of the first selection probability characterizes the likelihood that the corresponding parameter combination will become the globally optimal parameter combination in subsequent simulation verification. A larger first selection probability value indicates a higher probability that the parameter combination will become the globally optimal parameter combination.

[0041] Furthermore, after determining the first selection probability corresponding to each of the multiple parameter combinations based on the deep model, a reference subset can be determined from the multiple parameter combinations based on the first selection probability. For example, the parameter combinations with the highest first selection probability in the top 10% can be determined as the above reference subset, or the parameter combinations with the highest first selection probability in the top 50 can be determined as the above reference subset. No specific limitation is made here.

[0042] Furthermore, parameter combinations within a reference subset with high selection probabilities, as determined by the deep model prior, can be simulated and verified. This means that only a limited number of parameter combinations within this reference subset are subjected to high-cost plasma simulations, eliminating the need to traverse all original parameter combinations. As mentioned earlier, since there are hundreds or even thousands of original parameter combinations, simulating all of them would result in low computational efficiency. The reference subset determined in step S22 represents a portion of the original number of combinations. For example, in the case of 700 parameter combinations, if the reference subset is determined based on the first 10% of the aforementioned first selection probabilities, it means that only 70 parameter combinations need to be simulated, significantly reducing the time cost of simulation. Moreover, the first selection probabilities corresponding to parameter combinations determined by the deep model can be deployed on GPUs with massively parallel computing capabilities. This leverages the massively parallel computing power of GPUs to quickly determine a small reference subset from numerous parameter combinations, allowing for more accurate simulation verification of the parameter combinations within this reference subset, thereby improving the efficiency of etching parameter adjustment.

[0043] It should be noted that after obtaining the reference subset through step S22 above, the method further includes: S1, Perform plasma simulation on the reference subset to determine the power deposition distribution and electron temperature distribution for each parameter combination in the reference subset; S2, update the reference subset according to the electron temperature distribution, wherein each parameter combination in the updated reference subset satisfies the preset electron temperature constraint.

[0044] It should be noted that the above plasma simulation operations can be based on MATLAB, COMSOL, VSim or other simulation platforms that can perform plasma simulations, and no specific limitations are made here.

[0045] To be used on the simulation platform Figure 3 Taking the plasma etching chamber (etching chamber 301) shown as an example for simulation, the above plasma simulation operation requires the definition of relevant simulation parameters, including but not limited to: discharge gap, electrode diameter, electrode area, radio frequency frequency, radio frequency power, chamber pressure, and chamber temperature.

[0046] The discharge gap is the distance between the two electrode plates, which is... Figure 3The spacing between the source electrode 303 and the bias electrode 304 is such that the electrode diameter is [missing information]. Figure 3 The diameters of the source electrode 303 and the bias electrode 304, and the electrode areas are... Figure 3 The area of ​​the source electrode 303 and the bias electrode 304 in the figure, the radio frequency is Figure 3 The frequency of the RF power source 302 is [frequency value missing], and the RF power is [power value missing]. Figure 3 The power of the RF power source 302 and the chamber pressure are... Figure 3 The pressure and temperature inside the etching chamber 301 shown are... Figure 3 The temperature inside the etching chamber 301 shown.

[0047] It is understandable that during the process of adjusting the etching parameters of an etching device, the etching chamber inside the plasma is fixed, that is, the discharge gap, electrode diameter, and electrode area do not change, and the radio frequency is generally fixed at 13.56MHz. However, the radio frequency power, chamber pressure, and chamber temperature can vary. As mentioned above, each set of radio frequency power, chamber pressure, and chamber temperature is a parameter combination. After multiple parameter combinations are screened through the above step S22, the above reference subset is determined, and then plasma simulation is performed on each parameter combination in the reference subset.

[0048] It should be noted that each parameter in the above parameter combination does not necessarily change. As mentioned earlier, the search space for RF power is set to 50 to 500W, and the search is performed in steps of 10W; the search space for chamber pressure is set to 0.5 to 2 Torr, and the search is performed in steps of 0.1 Torr; while the chamber temperature is fixed at 300K. Each parameter combination is encoded into a multi-channel feature image, and each parameter combination corresponds to a pixel in the multi-channel feature image. RF power, chamber pressure, and chamber temperature each correspond to the r, g, and b channels in the pixel. No specific restrictions are placed on the encoding method here. It is only necessary to ensure that the encoded data conforms to the input specifications of the RGB channel. For example, the input data is divided by ten to avoid the value exceeding 255. Then, the first selection probability corresponding to the parameter combination is determined through the aforementioned deep model, and a reference subset is determined from multiple parameter combinations based on the first selection probability. Further, plasma simulation is performed on the reference subset using a simulation platform to determine the power deposition distribution and electron temperature distribution for each parameter combination in the reference subset. The power deposition distribution map and electron temperature distribution map obtained after simulating a parameter combination using the simulation platform are shown below. Figure 4 , Figure 5 .

[0049] Furthermore, the reference subset is updated based on the electron temperature distribution, wherein each parameter combination in the updated reference subset satisfies a preset electron temperature constraint.

[0050] It should be noted that electron temperature is a key parameter affecting plasma activity and ion energy, and its value needs to be maintained within a certain range. Otherwise, even if the power deposition uniformity is good, excessively high electron temperature may cause photoresist damage, or excessively low electron temperature may lead to insufficient etching rate. Therefore, this method introduces a preset electron temperature constraint condition to perform secondary screening of parameter combinations in the reference subset.

[0051] The aforementioned electron temperature constraint condition can be interpreted as requiring all regions in the simulated plasma etching chamber to meet the electron temperature constraint condition, i.e., requiring that... Figure 5 The electron temperature distribution diagram shown satisfies the electron temperature constraint condition.

[0052] In one possible implementation, the above-mentioned electron temperature constraint condition can be established based on the electron mobility-temperature formula, specifically, according to:

[0053] in, The current electron mobility; For reference electron mobility, the corresponding reference temperature is ( The electron mobility at a given temperature; T is the current electron temperature, which can be obtained from, for example... Figure 5 Extracted from the electron temperature distribution diagram shown; For reference temperature, a value of 300K can be used; Temperature coefficient, which can be 1.5.

[0054] Furthermore, we have:

[0055] Therefore, it is easy to obtain:

[0056]

[0057] Can be set The specific scope is then obtained and For example, it can be set At this point, we can obtain For 0.9 The value is 1.1, which is only used as an example and is not a specific limitation. Therefore, it can be solved... and Therefore, electronic temperature confinement conditions are constructed. That is, the requirement is as follows Figure 5 The data at each point in the electron temperature distribution graph shown must meet the following requirements. If the conditions are not met, the parameter combination is removed from the reference subset. After confirming whether the electron temperature distribution of all parameter combinations in the reference subset meets the electron temperature constraint and removing parameter combinations that do not meet the electron temperature constraint from the reference subset, the above-mentioned updating of the reference subset based on the electron temperature distribution is completed.

[0058] Furthermore, after updating the reference subset, step S23 is executed to iteratively calculate the reference subset. During the iteration, multiple target combinations are determined based on the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination, and the second selection probability and evaluation index are updated based on the etching uniformity. The second selection probability is used to characterize the probability that the parameter combination is determined as the target combination.

[0059] It should be noted that the iterative calculation of the reference subset in step S23 above includes the iterative calculation of the updated reference subset. That is, in each iteration, multiple parameter combinations are selected from the updated reference subset as target combinations according to the second selection probability of each parameter combination in the updated reference subset, and the power deposition simulation results of each target combination are obtained as described above. Figure 4 The power deposition distribution diagram shown is, if Figure 4 If the parameter combination corresponding to the power deposition distribution map shown is selected as the above target combination, then... Figure 4 The data in the above are used as the corresponding simulation results of the power deposition distribution, and then, according to Figure 4 The etching uniformity of the target combination is calculated based on the data, and the second selection probability and evaluation index of the target combination are updated after the completion of this iteration based on the etching uniformity.

[0060] The above-mentioned second choice probabilities can be assigned the same value under the initial conditions, for example, all of them can be 0, and they can be changed in each subsequent iteration. No specific restrictions are made here.

[0061] See Figure 6 The calculation of etching uniformity based on the power deposition simulation results of the target combination in step S23 above includes: S231, Obtain the power deposition distribution of the target combination; S232, determine the first average power deposition density of each sub-region of the wafer and the second average power deposition density of the entire wafer based on the power deposition distribution; S233, calculate the relative deviation rate based on the first average power deposition density and the second average power deposition density, wherein the relative deviation rate is used to describe the power deposition difference of each sub-region relative to the whole wafer; S234, determine the etching uniformity based on the relative deviation rate.

[0062] First, perform step S231 above to obtain the power deposition distribution of the target combination. Figure 4 For example, if Figure 4 If the corresponding parameter combination is selected as one of the multiple target combinations in this iteration, Figure 4 The data in the figure represents the power deposition distribution described above.

[0063] Then, step S232 is executed to determine the first average power deposition density of each sub-region of the wafer and the second average power deposition density of the entire wafer based on the power deposition distribution.

[0064] It should be noted that the wafer is housed within the plasma etching chamber, therefore the wafer's diameter is not equal to the long side of the bottom surface of the plasma etching chamber. However, the wafer's center can be considered to have the same (x, y) coordinates as the center of the bottom surface of the plasma etching chamber in a spatial coordinate system, thus it is not necessary to... Figure 4 All data in the data set only needs to be selected based on the coordinates corresponding to the wafer region. The corresponding coordinates can be coordinates of the same or similar regions as the wafer region; no specific limitation is made here.

[0065] The wafer region can be further divided into multiple sub-regions. For example, dividing the wafer into 10×10 grids can be simplified to... Figure 4 The extracted wafer region is divided into ten equal sub-regions. Each sub-region represents ten wafer regions with the same x-coordinate. Then, based on... Figure 4 The data in the figure are used to determine the first average power deposition density of each sub-region of the wafer and the second average power deposition density of the entire wafer. This is only an example, and in practice, more or fewer sub-regions can be divided than in the above implementation.

[0066] It is understandable that the aforementioned first average power deposition density is the same as that in... Figure 4 The average value of one of the 10 regions is determined, which represents the first average power deposition density corresponding to each of the ten wafer sub-regions.

[0067] Similarly, it can be based on Figure 4 The data in the calculation of the second average power deposition density of the entire wafer should be noted. It should be understood that the second average power deposition density can be the same as the data used for calculating the second average power deposition density of the wafer. Figure 4 The average value of the data from each sampling point corresponding to the mid-wafer region can also be the average value of all the aforementioned first average power deposition densities, i.e. , The first average power deposition density mentioned above, The second average power deposition density is not specifically defined here.

[0068] Further, step S233 is performed to calculate the relative deviation rate based on the first average power deposition density and the second average power deposition density, wherein the relative deviation rate is used to describe the power deposition difference of each sub-region relative to the whole wafer.

[0069] That is, for each sub-region, the relative deviation rate Defined as:

[0070] Then, step S234 is performed to determine the etching uniformity based on the relative deviation rate. (See also...) Figure 7 In one possible implementation, determining the etching uniformity based on the relative deviation rate includes: S2341, Determine the deviation level of the corresponding sub-region based on the relative deviation rate; S2342, determine the weights corresponding to different deviation levels; S2343, the etching uniformity is determined based on the weighted result of the weight and the relative deviation rate.

[0071] For example, the deviation level can be set to three levels: high, medium, and low. The deviation level of the corresponding sub-region can be determined based on the relative deviation rate. For example, the relative deviation rate can be... Sub-regions with a deviation rate greater than 5% are set to a high deviation level; the relative deviation rate is set to... Sub-regions with deviations greater than 2% and less than 5% are classified as medium deviation levels; the relative deviation rate is... Sub-regions with less than 2% deviation are set to a low deviation level.

[0072] Then, the weights corresponding to different deviation levels are determined. For example, the weight of high deviation level is determined to be 0.6, the weight of medium deviation level is determined to be 0.3, and the weight of low deviation level is determined to be 0.1. No specific limit is made here.

[0073] Furthermore, the etching uniformity is determined based on the weighted result of the weights and the relative deviation rate. : .

[0074] in, Let be the weight of the nth sub-region. This can be understood as... The more sub-regions that belong to the high deviation level, the higher the etching uniformity. The smaller the value, the better.

[0075] Furthermore, in step S23, after determining the etching uniformity, the etching uniformity updates the second selection probability and evaluation index.

[0076] In one possible implementation, the iterative computation of the reference subset is performed using an ant colony algorithm, where the second selection probability is the state transition probability and the evaluation index is the pheromone concentration.

[0077] Specifically, each parameter combination in the reference subset is regarded as a path in the ant colony algorithm. For example, if the number of ants in the ant colony algorithm is defined as 10, and the number of parameter combinations in the reference subset is 70, each of the 10 ants independently selects a parameter combination as the target combination according to the second selection probability in each iteration. After each round of iteration, the second selection probability and evaluation index of the selected parameter combination are updated. After the ant colony algorithm finishes iteration, the parameter combination with the highest pheromone concentration (evaluation index) is selected as the optimal parameter combination.

[0078] Define the state transition probability (second choice probability). :

[0079] in, The pheromone concentration for a chosen path (parameter combination) in the next iteration; The value can be 2 to indicate pheromone importance. The value for the heuristic function importance can be 3. The heuristic function takes the etching uniformity of a path (parameter combination) selected in the next iteration. , The concentration of pheromones on paths not traversed (parameter combinations not selected). Etching uniformity of paths not traversed (a parameter combination not selected before)

[0080] Understandably, as mentioned above, The more sub-regions that belong to the high deviation level, the higher the etching uniformity. The smaller the value of , the smaller its state transition probability (second choice probability) becomes, that is, the smaller the probability of being selected as the target combination.

[0081] It should be noted that the pheromone concentration (evaluation index) can be initially assigned the same value or different values. For example, all values ​​can be initially 0, or the value can be assigned based on the etching uniformity. For example, the etching uniformity can be used as the initial pheromone concentration. No specific restrictions are made here.

[0082] After one iteration of the ant colony algorithm is completed, the second selection probability and evaluation metric are updated. For the evaluation metric (pheromone concentration), we have:

[0083] Among them, z is the pheromone evaporation rate (0 < z < 1), which is used to prevent the algorithm from falling into local optimality. is the total pheromone left on a target combination in this iteration, that is, it represents the pheromones left by each ant that selects this target combination in this round of iteration The sum of, for can be defined as:

[0084] Among them, L is an adjustable coefficient, and U is the aforementioned etching uniformity.

[0085] Furthermore, according to the updated evaluation index (pheromone concentration), update the second selection probability of the corresponding parameter combination, and perform the next round of iteration according to the second selection probability. In this way, until the iteration ends, the condition for the iteration to end can be to set a certain number of rounds or all parameter combinations in the updated reference subset have been selected. No specific limitation is made here.

[0086] Finally, execute the above step S24, and determine the parameter combination with the highest evaluation index at the end of the iteration as the optimal parameter combination.

[0087] In a possible implementation manner, before determining the first selection probability corresponding to the parameter combination based on the depth model, the above method further includes: Training the to-be-trained depth model with a simulation data set until convergence, where each training sample in the simulation data set includes a reference parameter combination and a reference selection probability label determined after simulating with the reference parameter combination; The above method further includes: obtaining the reference power deposition distribution after simulating the reference parameter combination; determining the reference selection probability label according to the reference power deposition distribution.

[0088] Understandably, before using the aforementioned deep model to predict the first choice probability, the model must be sufficiently trained to learn the mapping relationship between parameter combinations and the probability of becoming the optimal solution. The training set for training the deep model consists of several reference parameter combinations. For each set of reference parameter combinations, a complete plasma simulation is performed to obtain output results such as power deposition distribution and electron temperature distribution. Based on these simulation results, the etching uniformity (e.g., the sum of weighted uniformities) can be calculated in the same way as in step S23 above. The specific method has been explained in detail above and will not be repeated here. Then, the etching uniformity is mapped to a scalar value as a reference choice probability label for that parameter combination. The mapping method can use normalization, for example, linearly transforming the uniformity value from its actual range to the interval between 0 and 1; the higher the uniformity, the closer the label value is to 1. Each set of reference parameter combinations and its corresponding reference choice probability label constitutes a training sample. Multiple such sample sets form the aforementioned simulation dataset, which is then used to train the deep model to be trained until convergence.

[0089] Through the above implementation method, firstly, multiple parameter combinations are obtained, including RF power, chamber pressure, and chamber temperature; then, a first selection probability corresponding to the parameter combination is determined based on the depth model, and a reference subset is determined from the multiple parameter combinations according to the first selection probability; further, the reference subset is iteratively calculated, and multiple target combinations are determined according to the second selection probability corresponding to the parameter combination during the iteration. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination, and the second selection probability and evaluation index are updated according to the etching uniformity; finally, the parameter combination with the highest evaluation index at the end of the iteration is determined as the optimal parameter combination. In this process, on the one hand, the first selection probability corresponding to the parameter combination determined by the deep model can be deployed on a GPU with large-scale parallel computing capabilities. That is, by utilizing the large-scale parallel computing capabilities of the GPU, a small reference subset can be quickly determined from a large number of parameter combinations. Then, the parameter combinations in the reference subset can be more accurately simulated and verified, thereby improving the efficiency of etching parameter adjustment. On the other hand, an automated parameter selection method is designed based on the simulation results of the reference subset. After multiple iterations, an optimal parameter combination can be identified, further improving the efficiency of etching parameter adjustment. This solves the technical problem of low efficiency in etching process parameter adjustment in the existing technology.

[0090] According to another aspect of the embodiments of this application, a computer device for implementing the above-described etching parameter adjustment method based on plasma simulation is also provided. This computer device may be... Figure 1 The terminal device or server shown. This embodiment uses a computer as an example for illustration. Figure 8As shown, the computer device includes a memory 802 and a processor 804. The memory 802 stores a computer program, and the processor 804 is configured to execute the steps in any of the above method embodiments via the computer program.

[0091] Optionally, in this embodiment, the computer device described above may be located in at least one of a plurality of network devices in a computer network.

[0092] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program: S1, obtain multiple parameter combinations, including RF power, chamber pressure and chamber temperature; S2, determine the first selection probability corresponding to the parameter combination based on the deep model, and determine the reference subset from multiple parameter combinations according to the first selection probability. The first selection probability is used to characterize the possibility that the parameter combination is the optimal parameter combination. S3, iterative calculation of the reference subset. During the iteration, multiple target combinations are determined according to the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination. The second selection probability and evaluation index are updated according to the etching uniformity. The second selection probability is used to characterize the possibility that the parameter combination is determined as the target combination. S4. The parameter combination with the highest evaluation index at the end of the iteration is determined as the optimal parameter combination.

[0093] Alternatively, as those skilled in the art will understand, Figure 8 The structure shown is for illustrative purposes only. The computer device can also be a smartphone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile internet device (MID), a PAD, and other terminal devices. Figure 8 This does not limit the structure of the aforementioned computer equipment. For example, the computer equipment may also include components that are more... Figure 8 The more or fewer components shown (such as network interfaces, etc.), or having the same Figure 8 The different configurations shown.

[0094] The memory 802 can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods described in the embodiments of this application. The processor 804 executes various functional applications and data processing by running the software programs and modules stored in the memory 802, thereby implementing the methods provided in the various optional implementations described above. The memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 802 may further include memory remotely located relative to the processor 804, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0095] Optionally, the transmission device 806 described above is used to receive or send data via a network. Specific examples of the network described above may include wired networks and wireless networks. In one example, the transmission device 806 includes a Network Interface Controller (NIC), which can be connected to other network devices and a router via a network cable to communicate with the Internet or a local area network. In one example, the transmission device 806 is a Wi-Fi module, used for wireless communication with the Internet.

[0096] In addition, the aforementioned computer device also includes: a display 808 for displaying pages; and a connection bus 810 for connecting various module components in the aforementioned electronic device.

[0097] In other embodiments, the aforementioned terminal device or server can be a node in a distributed system, wherein the distributed system can be a blockchain system, which is a distributed system formed by connecting multiple nodes through network communication. The nodes can form a point-to-point network, and any form of computing device, such as a server, terminal, or other electronic device, can become a node in the blockchain system by joining this point-to-point network.

[0098] According to another aspect of this application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the above-described method for adjusting etching parameters based on plasma simulation is loaded and executed by a processor.

[0099] Optionally, the computer-readable storage medium may include: ROM (Read-Only Memory), RAM (Random-Access Memory), SSD (Solid State Drives), or optical disc, etc. The random access memory may include ReRAM (Resistance Random Access Memory) and DRAM (Dynamic Random Access Memory).

[0100] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps: S1, obtain multiple parameter combinations, including RF power, chamber pressure and chamber temperature; S2, determine the first selection probability corresponding to the parameter combination based on the deep model, and determine the reference subset from multiple parameter combinations according to the first selection probability. The first selection probability is used to characterize the possibility that the parameter combination is the optimal parameter combination. S3, iterative calculation of the reference subset. During the iteration, multiple target combinations are determined according to the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination. The second selection probability and evaluation index are updated according to the etching uniformity. The second selection probability is used to characterize the possibility that the parameter combination is determined as the target combination. S4. The parameter combination with the highest evaluation index at the end of the iteration is determined as the optimal parameter combination.

[0101] According to another aspect of this application, a computer program product is provided, including a computer program that, when loaded and executed by a processor, implements the above-described method for adjusting etching parameters based on plasma simulation.

[0102] It should be noted that, in the embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program with a predetermined function, which works together with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0103] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0104] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0105] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

[0106] The above-described preferred embodiments of the present invention are provided as examples, but it will be apparent to those skilled in the art that such embodiments are provided merely by way of example. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and intent of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of the invention. The appended claims are intended to define the scope of protection of the invention and therefore cover the modular compositions, equivalents, or alternatives within the scope of these claims.

Claims

1. A method for adjusting etching parameters based on plasma simulation, characterized in that, include: Multiple parameter combinations are obtained, including radio frequency power, chamber pressure, and chamber temperature; The first selection probability corresponding to the parameter combination is determined based on the deep model, and a reference subset is determined from multiple parameter combinations based on the first selection probability. The first selection probability is used to characterize the probability that the parameter combination is the optimal parameter combination. The reference subset is iteratively calculated. During the iteration, multiple target combinations are determined according to the second selection probability corresponding to the parameter combination. The etching uniformity is calculated based on the power deposition distribution simulation results of the target combination. The second selection probability and evaluation index are updated according to the etching uniformity. The second selection probability is used to characterize the probability that the parameter combination is determined as the target combination. The parameter combination with the highest evaluation index at the end of the iteration is determined as the optimal parameter combination.

2. The method according to claim 1, characterized in that, After determining the reference subset from the plurality of parameter combinations based on the first selection probability, the method further includes: Perform plasma simulation on the reference subset to determine the power deposition distribution and electron temperature distribution for each parameter combination in the reference subset; The reference subset is updated based on the electron temperature distribution, wherein each parameter combination in the updated reference subset satisfies a preset electron temperature constraint.

3. The method according to claim 2, characterized in that, The iterative calculation of the reference subset includes iterative calculation of the updated reference subset, and the calculation of etching uniformity based on the power deposition simulation results of the target combination includes: Obtain the power deposition distribution of the target combination; The first average power deposition density of each sub-region of the wafer and the second average power deposition density of the entire wafer are determined based on the power deposition distribution. The relative deviation rate is calculated based on the first average power deposition density and the second average power deposition density, wherein the relative deviation rate is used to describe the power deposition difference of each sub-region relative to the whole wafer; The etching uniformity is determined based on the relative deviation rate.

4. The method according to claim 3, characterized in that, Determining the etching uniformity based on the relative deviation rate includes: The deviation level of the corresponding sub-region is determined based on the relative deviation rate; Determine the weights corresponding to different deviation levels; The etching uniformity is determined based on the weighted result of the weights and the relative deviation rate.

5. The method according to claim 1, characterized in that, The input to the depth model is a multi-channel feature image, wherein different channels of each pixel in the multi-channel feature image encode the radio frequency power, the chamber pressure, and the chamber temperature, respectively.

6. The method according to claim 5, characterized in that, Before determining the first selection probability corresponding to the parameter combination based on the deep model, the method further includes: The deep model to be trained is trained to convergence using a simulation dataset, wherein each training sample in the simulation dataset includes a combination of reference parameters and a reference selection probability label determined after simulation using the combination of reference parameters. The method further includes: obtaining the reference power deposition distribution after simulation of the reference parameter combination; and determining the reference selection probability label based on the reference power deposition distribution.

7. The method according to claim 1, characterized in that, The iterative calculation of the reference subset is performed using the ant colony algorithm, wherein the second selection probability is the state transition probability, and the evaluation index is the pheromone concentration.

8. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing a computer program, which is loaded and executed by the processor to implement the etching parameter adjustment method based on plasma simulation as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which is loaded and executed by a processor to implement the etching parameter adjustment method based on plasma simulation as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, It includes a computer program, which is loaded and executed by a processor to implement the etching parameter adjustment method based on plasma simulation as described in any one of claims 1 to 7.

Images