Artificial intelligence-based industrial automatic control system and method

Abstract

The application discloses an industrial automatic control system and method based on artificial intelligence, and relates to the technical field of control.The application comprises a data acquisition module, a doping planning module, a doping control module, an early warning terminal and a database.According to the use and lattice structure of a semiconductor and the information of the semiconductor in each doping history, the parameter values of ion liquid and an ion implanter used in doping are acquired.When the semiconductor is controlled according to the acquired data, the doping effect of the semiconductor is analyzed at each collection time.If the doping effect is poor, the electrical property and lattice of each region of the semiconductor are analyzed.If the electrical property of a certain region does not meet the requirements, the doping parameter values are adjusted.If there is a region with damaged lattice, the damage coefficient of the lattice is analyzed, and an annealing operation is performed, so that the effectiveness of the ion liquid configuration and the parameters of the ion implanter is ensured, and the doping efficiency of the semiconductor is ensured.

Description

Technical Field

[0001] This invention relates to the field of control technology, and more specifically to an industrial automatic control system and method based on artificial intelligence. Background Technology

[0002] Semiconductor doping is one of the key processes in semiconductor manufacturing. By introducing ions into semiconductor materials, the electrical properties of semiconductors can be precisely controlled, thereby creating various high-performance semiconductor devices. In advanced processes, in order to achieve nanoscale device structures, the precision and uniformity of doping are extremely important. It is necessary to precisely control the concentration and distribution of dopant ions in order to form high-performance transistors and other devices.

[0003] Traditional industrial automatic control systems and methods configure the ion liquid and set the parameters of the ion implanter based on manual experience before doping. After the semiconductor doping is completed, the doping effect is monitored. If the effect is not good, the doping process is adjusted. Obviously, this industrial automatic control system and method has the following shortcomings: 1. Traditional industrial automatic control systems and methods configure the ion liquid and adjust the parameters of the ion implanter based on manual experience. There is no data support, so the effectiveness of the ion liquid configuration and the effectiveness of the parameters of the ion implanter cannot be guaranteed.

[0004] 2. Traditional industrial automatic control systems and methods detect the doping effect of semiconductors after the doping process is completed. They cannot detect the doping effect during the doping process, and therefore cannot adjust the semiconductor doping process in real time, thus failing to guarantee the doping efficiency of semiconductors. Summary of the Invention

[0005] In view of the above-mentioned technical deficiencies, the purpose of this invention is to provide an industrial automatic control system and method based on artificial intelligence.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: In a first aspect, the present invention provides an industrial automatic control system based on artificial intelligence, including the following modules: a data acquisition module, a doping planning module, a doping control module, an early warning terminal, and a database.

[0007] The data acquisition module is used to obtain the applications and crystal structures of semiconductors from the database.

[0008] The doping planning module includes an ion liquid planning unit and an ion implanter planning unit.

[0009] The ionic liquid planning unit is used to obtain the lattice structure, application and ionic liquid used of the semiconductor during each historical doping from the database, and to configure the ionic liquid based on the application and lattice structure of the semiconductor at that time.

[0010] The ion implanter planning unit is used to obtain the application of semiconductors and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and to analyze the parameter values ​​of the ion implanter.

[0011] The doping control module is used to set each acquisition time according to a preset time interval when doping the semiconductor according to the planned parameters of the ion liquid and ion implanter. At each acquisition time, the doping effect of the semiconductor is analyzed. If the effect is not good, an early warning is issued and adjustments and remedies are made.

[0012] The early warning terminal is used to issue an early warning when the doping effect is poor.

[0013] The database is used to store information and standard lattice diffraction patterns from each historical doping event.

[0014] Secondly, the present invention provides an industrial automatic control method based on artificial intelligence, including the following steps: Step 1, data acquisition: acquiring the application and crystal structure of semiconductors from a database.

[0015] Step 2: Plan the ionic liquid: Obtain the lattice structure, application, and ionic liquid used for each historical doping process from the database, and configure the ionic liquid based on the application and lattice structure of the semiconductor at that time.

[0016] Step 3: Plan the ion implanter: Obtain the application of the semiconductor and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and analyze the parameter values ​​of the ion implanter.

[0017] Step 4: Controlling Doping: When doping the semiconductor according to the planned parameters of the ion liquid and ion implanter, set the sampling time according to the preset time interval, analyze the doping effect of the semiconductor at each sampling time, and if the effect is not good, issue an early warning and make adjustments and remedies.

[0018] The beneficial effects of this invention are as follows: 1. This invention provides an industrial automatic control system and method based on artificial intelligence. According to the application and crystal structure of the semiconductor, as well as the information of the semiconductor during each historical doping, the system obtains the parameter values ​​of the ion liquid and ion implanter used during doping. When controlling semiconductor doping according to the obtained data, the doping effect of the semiconductor is analyzed at each acquisition time. If the doping effect is not good, the electrical properties and crystal lattice of each region of the semiconductor are analyzed. If the electrical properties of a certain region do not meet the requirements, the values ​​of each doping parameter are adjusted. If there is a region with crystal lattice damage, the crystal lattice damage coefficient is analyzed and annealing is performed. This ensures the effectiveness of the ion liquid configuration and the parameters of the ion implanter, and also ensures the doping efficiency of the semiconductor.

[0019] 2. This invention retrieves the applications and lattice structures of semiconductors from a database for each historical doping event, compares them, obtains each labeled semiconductor, and obtains the band gap and density of states of each labeled semiconductor before doping, as well as the band gap and density of states after doping, and calculates their differences. At the same time, based on the differences in band gap and density of states of each labeled semiconductor, the mixing ratio of each ion in the ion liquid is obtained. Finally, based on the mixing ratio of each ion, the concentration of each ion stock solution, and the required volume of the ion liquid, the ion liquid is prepared, ensuring the effectiveness of the ion liquid preparation.

[0020] 3. This invention obtains the application of semiconductors, the structure and concentration of each ion in the ionic liquid used in each historical doping process from the database, and compares it with the application of semiconductors, the structure and concentration of each ion in the ionic liquid used in the current doping process to obtain the accelerating voltage, electrode spacing and magnetic field strength of the ion implanter, thus ensuring the effectiveness of each parameter of the ion implanter.

[0021] 4. In this invention, during semiconductor doping, qualitative and lattice parameters of each region within the semiconductor are analyzed at each acquisition time. If, at a certain acquisition time, the electrical parameter or lattice parameter of a certain region is 1, it indicates that the doping effect of the semiconductor is poor at that acquisition time. In this case, the electrical parameter of each region is obtained as the lattice parameter. If there is a region with an electrical parameter of 1, the values ​​of each doping parameter are adjusted. If there is a region with a lattice parameter of 1, the regions with a lattice parameter of 1 are obtained, and their lattice damage coefficient is analyzed. At the same time, annealing is performed to ensure the doping efficiency of the semiconductor. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of the system structure connection of the present invention.

[0024] Figure 2 This is a schematic diagram of the implementation steps of the method of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 As shown, the present invention provides an artificial intelligence-based industrial automatic control system, including the following modules: a data acquisition module, a doping planning module, a doping control module, an early warning terminal, and a database.

[0027] The acquisition module is connected to the doping planning module, the doping planning module is connected to the doping control module, the doping control module is connected to the early warning terminal, and the database is connected to the data acquisition module, the doping planning module, and the doping control module.

[0028] The data acquisition module is used to obtain the applications and crystal structures of semiconductors from the database.

[0029] It should be noted that semiconductors are used in applications such as integrated circuits, transistors, light-emitting diodes, and laser diodes.

[0030] It should also be noted that the crystal lattice structures of semiconductors include diamond structure, zincblende structure, and wurtzite structure, among others.

[0031] The doping planning module includes an ion liquid planning unit and an ion implanter planning unit.

[0032] The ionic liquid planning unit is used to obtain the lattice structure, application and ionic liquid used of the semiconductor during each historical doping from the database, and to configure the ionic liquid based on the application and lattice structure of the semiconductor at that time.

[0033] In a specific embodiment, the specific process of configuring the ionic liquid is as follows: S11, obtain the application and lattice structure of the semiconductor during each historical doping from the database, and compare it with the application and lattice structure of the semiconductor being doped this time. If the application and structure of the semiconductor during a certain historical doping are the same as the application and lattice structure of the semiconductor being doped this time, then the semiconductor that was doped in that historical period is called the marker semiconductor. Each marker semiconductor is obtained by this method.

[0034] It should be noted that the ions include boron and phosphorus, among others.

[0035] S12. Obtain the band gap and density of states of each labeled semiconductor before doping, and the band gap and density of states after doping from the database. Calculate the band gap difference and density of states difference of each labeled semiconductor, and compare them. Select the labeled semiconductor with the smallest band gap difference and the smallest density of states difference. Obtain the mixing ratio of each ion in the ionic liquid used for this labeled semiconductor from the database, and use it as the mixing ratio of each ion in the ionic liquid used in this doping. Calculate the volume of each ionic stock solution in the ionic liquid according to the required ionic liquid volume, the mixing ratio of each ion, and the concentration of each ionic stock solution, and configure them accordingly.

[0036] It should be noted that the required volume of ionic liquid was set by the staff.

[0037] It should be noted that the required volume of the ionic liquid is... Let L be the volume of the boron and phosphorus original solutions in the ionic liquid, respectively. L and Given that the concentrations of boron and phosphorus in the original solutions are 1 mol / L and 2 mol / L respectively, and the mixing ratio of boron to phosphorus is 1:2, then... Thus obtain , .

[0038] The ion implanter planning unit is used to obtain the application of semiconductors and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and to analyze the parameter values ​​of the ion implanter.

[0039] It should be noted that ions have different ionic structures. Taking boron as an example, its ionic structure includes... Ionic structure and Ionic structure, etc.

[0040] In a specific embodiment, the analysis of the parameters of the ion implanter is carried out as follows: S21, the application of the semiconductor during each historical doping is compared with the application of the semiconductor during the current doping. If the application of the semiconductor during a certain historical doping is the same as the application of the semiconductor during the current doping, then the semiconductor during that historical doping is called a semiconductor sample. Each semiconductor sample is obtained in this way, and the availability coefficient of the ion implanter acceleration voltage of each semiconductor sample is analyzed. The ion implanter acceleration voltage of the semiconductor sample with the highest availability coefficient is obtained and used as the acceleration voltage of the ion implanter during the current doping.

[0041] S22. Obtain the structure and concentration of each ion in the ion solution used for each semiconductor sample, and compare it with the structure and concentration of each ion in the ion solution used during doping. If the structure and concentration of each ion in the ion solution used for a semiconductor sample are the same as those in the ion solution used during doping, then the semiconductor sample is called the control semiconductor. Obtain each control semiconductor in this way, and obtain the ion concentration range of each control semiconductor from the database. Calculate the difference between the maximum and minimum values ​​of the ion concentration range of each control semiconductor, and call it the ion concentration range difference of each control semiconductor. Select the electrode spacing and magnetic field strength of the ion implanter of the control semiconductor with the smallest difference, and use it as the electrode spacing and magnetic field strength of the ion implanter for this ion implantation.

[0042] The above-mentioned analysis of the availability coefficient of the ion implanter accelerating voltage for each semiconductor sample involves the following steps: The doping depth, lattice damage area during doping, and doping area of ​​each semiconductor sample are obtained from the database and normalized. Then, based on the analysis formula: Get the first Availability coefficient of accelerating voltage for semiconductor sample ion implanter In the formula Representing the Lattice damage area when a semiconductor sample is doped Representing the The doping area of ​​a semiconductor sample Representing the Doping depth of a semiconductor sample Represents the target doping depth of a semiconductor. The number representing each semiconductor sample. =1,2,3,..., , Represents the total number of semiconductor samples. and All are positive integers.

[0043] The doping control module is used to set each acquisition time according to a preset time interval when doping the semiconductor according to the planned parameters of the ion liquid and ion implanter. At each acquisition time, the doping effect of the semiconductor is analyzed. If the effect is not good, an early warning is issued and adjustments and remedies are made.

[0044] In a specific embodiment, the doping control module performs the following process: The semiconductor is divided into regions according to a preset area threshold. Simultaneously, a preset time interval is set for each acquisition moment. At each acquisition moment, electrical parameters and lattice diffraction patterns within each region are acquired, and the electrical and lattice indices of each region at each acquisition moment are obtained. The doping effect of the semiconductor at each acquisition moment is analyzed. When the doping effect of the semiconductor is poor at a certain acquisition moment, the electrical and lattice indices of each region at that acquisition moment are obtained. If the electrical index of a certain region is 1, the values ​​of each doping parameter are adjusted. If the lattice index of one or more regions is 1, an annealing operation is performed on each region with a lattice index of 1.

[0045] It should be noted that the preset area threshold is a critical value used to evaluate whether the area division is reasonable, and it is set by the staff.

[0046] It should also be noted that the preset time interval is a threshold used to determine whether the data collection time setting is standard, and it is set by the staff.

[0047] The electrical parameters, including carrier concentration and resistivity, are collected using sensors such as Hall effect sensors and four-probe testers.

[0048] It also needs to be explained that an electron diffractometer is used to collect lattice diffraction patterns.

[0049] The above-mentioned analysis of the doping effect of semiconductors at each acquisition time is specifically carried out as follows: S31, obtain the values ​​of each electrical parameter of each region at each acquisition time, perform normalization processing, and input them into the electrical analysis model, output the electrical index of each region at each acquisition time. If the electrical index of a certain region at a certain acquisition time is 1, it means that the electrical performance of that region does not meet the requirements. If the electrical index of a certain region at a certain acquisition time is 0, it means that the electrical performance of that region meets the requirements.

[0050] It should be noted that the electrical analysis model is as follows: In the formula Representing the The first region The electrical parameter at the th Values ​​at each acquisition time, Representing the The electrical parameter at the th Standard values ​​at each data collection time Representing the The region in the first Electrical parameters at each acquisition time, This represents a preset threshold for the difference in electrical parameters. The numbers representing each region =1,2,3,..., , Represents the total number of regions. The numbers representing the various electrical parameters, =1,2,3,..., , Represents the total number of electrical parameters. The number representing each data collection moment. =1,2,3,..., , This represents the total number of data collection moments. , , , , and All are positive integers.

[0051] It should also be noted that the standard values ​​of each electrical parameter at each acquisition time were set by the staff.

[0052] The preset electrical parameter difference threshold is a critical value used to evaluate whether each electrical parameter meets the requirements, and it is set by the staff.

[0053] S32. Obtain the lattice diffraction pattern of each region at each acquisition time and compare it with the standard lattice diffraction pattern. If the lattice diffraction pattern of a certain region at a certain acquisition time is different from the standard lattice diffraction pattern, it means that the lattice index of that region at that acquisition time is 1, and the lattice in that region is damaged. If the lattice diffraction pattern of a certain region at a certain acquisition time is the same as the standard lattice diffraction pattern, it means that the lattice index of that region at that acquisition time is 0, and the lattice in that region is not damaged.

[0054] It should be noted that the standard lattice diffraction pattern was obtained from a database.

[0055] S33. Obtain the electrical and lattice indices of each region at each acquisition time. When the electrical index or lattice index of one or more regions is 1 at a certain acquisition time, it indicates that the doping effect of the semiconductor at that acquisition time is poor. When the electrical and lattice indices of each region are 0 at a certain acquisition time, it indicates that the doping effect of the semiconductor at that acquisition time is good.

[0056] The specific process of adjusting each doping parameter described above is as follows: S41. When the electrical index of a certain region is 1, a virtual test bench is constructed, and a virtual semiconductor and ion implanter are generated. The electrode spacing range and magnetic field strength range are set. From the ion mixing ratio set, the accelerating voltage set, the electrode spacing range, and the magnetic field strength range, several ion mixing ratios, several accelerating voltages, several electrode spacings, and several magnetic field strengths are randomly selected and randomly combined to obtain several doping combinations. Each doping combination includes each ion mixing ratio, accelerating voltage, electrode spacing, and magnetic field strength.

[0057] It should be noted that the electrode spacing and magnetic field strength of the ion implanter during this doping process are obtained and used as the median of the electrode spacing range and magnetic field strength range. Preset electrode spacing thresholds and preset magnetic field strength thresholds are set, and the electrode spacing and magnetic field strength are added to or subtracted from the preset electrode spacing thresholds and preset magnetic field strength thresholds to obtain the maximum and minimum values ​​of the electrode spacing range and magnetic field strength range.

[0058] It should also be noted that the preset electrode spacing threshold is a critical value used to evaluate whether the electrode spacing range setting is reasonable. It is obtained by the deep learning model based on the historical electrode spacing.

[0059] The preset magnetic field strength threshold is a critical value used to evaluate whether the magnetic field strength range setting is reasonable. It is obtained by deep learning based on historical magnetic field strengths.

[0060] S42. Apply each doping combination to a virtual test bench, and simultaneously obtain the doping effect of each doping combination. Select the doping combinations with good doping effects and refer to them as usable doping combinations. In the formula Representative in application number When the number of available doping combinations is , the th The first region A value for an electrical parameter. Representing the Standard values ​​for each electrical parameter Representing the The doping coefficients of a number of available doping combinations The numbers representing each region =1,2,3,..., , Represents the total number of regions. The numbers representing the various electrical parameters, =1,2,3,..., , Represents the total number of electrical parameters. The number representing each available doping combination. =1,2,3,..., , This represents the total number of available doping combinations. , , , , and All are positive integers. The doping combination with the highest doping coefficient is selected. The mixing ratio of each ion, the accelerating voltage, the electrode spacing and the magnetic field strength in the doping combination are used as the adjusted mixing ratio of each ion, the accelerating voltage, the electrode spacing and the magnetic field strength, and then adjusted.

[0061] It should also be noted that the standard values ​​of each electrical parameter are obtained from the database.

[0062] The above-mentioned annealing operation for each region with a lattice index of 1 is specifically performed as follows: Each region with a lattice index of 1 is obtained and referred to as an annealing region. The lattice diffraction pattern of each annealing region is obtained, and the lattice damage coefficient of each annealing region is analyzed. Simultaneously, the lattice damage coefficients from historical annealing events in the database are obtained and compared. If the lattice damage coefficient from a historical annealing event is the same as the lattice damage coefficient of a certain annealing region, then the annealing parameters from that historical annealing event are obtained from the database and used as the annealing parameters for that annealing region. A set of annealing parameters for each annealing region is obtained in this way, where each element represents the value of each annealing parameter. The average value of each annealing parameter in the set of each annealing region is calculated and used as the standard value of each annealing parameter for each region. Simultaneously, the annealing device is controlled to perform annealing operations on each region according to the annealing parameters of each region.

[0063] The above-mentioned analysis of the lattice damage coefficient in each annealing region is specifically carried out as follows: the width and position of the diffraction peaks in the lattice diffraction pattern of each annealing region are obtained, as well as the width and position of the diffraction peaks in the standard lattice diffraction pattern. The width difference and displacement of the diffraction peaks in the lattice diffraction pattern of each annealing region are calculated. Based on the width difference and displacement of the diffraction peaks in the lattice diffraction pattern of each annealing region, the lattice damage coefficient in each annealing region is calculated.

[0064] It should be noted that the lattice damage coefficient for each annealing region is obtained by dividing the difference in the width of the diffraction peaks in the lattice diffraction pattern of each annealing region by the width of the diffraction peaks in the standard lattice diffraction pattern, and then adding the resulting value to the displacement. It is assumed that the width of the diffraction peaks in the standard lattice diffraction pattern is... The positions of the diffraction peaks are The width of the diffraction peak in the lattice diffraction pattern of a certain annealed region is The positions of the diffraction peaks are The difference in the width of the diffraction peaks in the lattice diffraction pattern of the annealed region is then... The displacement is Then the lattice damage coefficient of the annealed region is .

[0065] The early warning terminal is used to issue an early warning when the doping effect is poor.

[0066] The database is used to store information from each historical doping event, standard values ​​of each electrical parameter, and standard lattice diffraction patterns.

[0067] It should be noted that the information for each historical doping event includes the application of the semiconductor, the lattice structure of the semiconductor, the mixing ratio of each ion in the ionic liquid used, the structure of each ion in the ionic liquid, the concentration of each ion, the band gap, the density of states, the doping depth, the lattice damage area during doping, and the doping area.

[0068] Please see Figure 2 As shown, the present invention provides an industrial automatic control method based on artificial intelligence, including the following steps: Step 1, data acquisition: acquiring the application and crystal structure of semiconductors from a database.

[0069] Step 2: Plan the ionic liquid: Obtain the lattice structure, application, and ionic liquid used for each historical doping process from the database, and configure the ionic liquid based on the application and lattice structure of the semiconductor at that time.

[0070] Step 3: Plan the ion implanter: Obtain the application of the semiconductor and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and analyze the parameter values ​​of the ion implanter.

[0071] Step 4: Controlling Doping: When doping the semiconductor according to the planned parameters of the ion liquid and ion implanter, set the sampling time according to the preset time interval, analyze the doping effect of the semiconductor at each sampling time, and if the effect is not good, issue an early warning and make adjustments and remedies.

[0072] This invention, based on the semiconductor's application and crystal structure, as well as information from previous doping processes, acquires the parameter values ​​of the ion liquid and ion implanter used during doping. When controlling semiconductor doping according to the acquired data, the doping effect is analyzed at each acquisition time. If the doping effect is unsatisfactory, the electrical properties and lattice of each region of the semiconductor are analyzed. If the electrical properties of a certain region do not meet the requirements, the values ​​of each doping parameter are adjusted. If there are regions with lattice damage, the lattice damage coefficient is analyzed, and annealing is performed. This ensures the effectiveness of the ion liquid configuration and the parameters of the ion implanter, and also ensures the doping efficiency of the semiconductor.

[0073] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. An industrial automatic control system based on artificial intelligence, characterized in that, Includes the following modules: The data acquisition module is used to retrieve the applications and crystal structures of semiconductors from the database; The doping planning module includes an ion liquid planning unit and an ion implanter planning unit: The ion liquid planning unit is used to obtain the lattice structure, application and ion liquid used of the semiconductor during each historical doping from the database, and to configure the ion liquid based on the application and lattice structure of the semiconductor at that time. The ion implanter planning unit is used to obtain the application of semiconductors and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and to analyze the parameter values ​​of the ion implanter. The doping control module is used to set the acquisition time according to the preset time interval when the semiconductor is doped according to the planned parameters of the ion liquid and ion implanter. At each acquisition time, the doping effect of the semiconductor is analyzed. If the effect is not good, an early warning is issued and adjustments and remedies are made. The early warning terminal is used to issue an early warning when the doping effect is poor. The database is used to store information from each historical doping event, standard values ​​of each electrical parameter, and standard lattice diffraction patterns.

2. The artificial intelligence-based industrial automatic control system according to claim 1, characterized in that, The specific process for preparing the ionic liquid is as follows: S11. Obtain the application and crystal structure of the semiconductor during each historical doping from the database, and compare them with the application and crystal structure of the semiconductor doped in this instance. If the application and structure of the semiconductor during a certain historical doping are the same as those of the semiconductor doped in this instance, then the semiconductor doped in that historical instance is called the marker semiconductor. Obtain each marker semiconductor in this way. S12. Obtain the band gap and density of states of each labeled semiconductor before doping, and the band gap and density of states after doping from the database. Calculate the band gap difference and density of states difference of each labeled semiconductor, and compare them. Select the labeled semiconductor with the smallest band gap difference and the smallest density of states difference. Obtain the mixing ratio of each ion in the ionic liquid used for this labeled semiconductor from the database, and use it as the mixing ratio of each ion in the ionic liquid used in this doping. Calculate the volume of each ionic stock solution in the ionic liquid according to the required ionic liquid volume, the mixing ratio of each ion, and the concentration of each ionic stock solution, and configure them accordingly.

3. The artificial intelligence-based industrial automatic control system according to claim 1, characterized in that, The specific process for analyzing the parameters of the ion implanter is as follows: S21. Compare the applications of semiconductors during each historical doping with the applications of semiconductors used in this doping. If the application of a semiconductor during a historical doping is the same as that of a semiconductor used in this doping, then the semiconductor used in that historical doping is called a semiconductor sample. Obtain each semiconductor sample in this way and analyze the availability coefficient of the ion implanter acceleration voltage of each semiconductor sample. Obtain the ion implanter acceleration voltage of the semiconductor sample with the highest availability coefficient and use it as the acceleration voltage of the ion implanter in this doping. S22. Obtain the structure and concentration of each ion in the ion solution used for each semiconductor sample, and compare it with the structure and concentration of each ion in the ion solution used during doping. If the structure and concentration of each ion in the ion solution used for a semiconductor sample are the same as those in the ion solution used during doping, then the semiconductor sample is called the control semiconductor. Obtain each control semiconductor in this way, and obtain the ion concentration range of each control semiconductor from the database. Calculate the difference between the maximum and minimum values ​​of the ion concentration range of each control semiconductor, and call it the ion concentration range difference of each control semiconductor. Select the electrode spacing and magnetic field strength of the ion implanter of the control semiconductor with the smallest difference, and use it as the electrode spacing and magnetic field strength of the ion implanter for this ion implantation.

4. The artificial intelligence-based industrial automatic control system according to claim 3, characterized in that, The specific process for analyzing the availability coefficient of the accelerating voltage of the ion implanter for each semiconductor sample is as follows: The doping depth, lattice damage area, and doping area of ​​each semiconductor sample are obtained from the database and normalized. Then, based on the analysis formula: Get the first Availability coefficient of accelerating voltage for semiconductor sample ion implanter In the formula Representing the Lattice damage area when a semiconductor sample is doped Representing the The doping area of ​​a semiconductor sample Representing the Doping depth of a semiconductor sample Represents the target doping depth of a semiconductor. The number representing each semiconductor sample. =1,2,3,..., , Represents the total number of semiconductor samples. and All are positive integers.

5. The artificial intelligence-based industrial automatic control system according to claim 1, characterized in that, The doping control module operates as follows: The semiconductor is divided into regions according to a preset area threshold. At the same time, the acquisition time is set according to a preset time interval. At each acquisition time, the electrical parameters and lattice diffraction patterns of each region are acquired, and the electrical and lattice indices of each region at each acquisition time are obtained. The doping effect of the semiconductor at each acquisition time is analyzed. When the doping effect of the semiconductor at a certain acquisition time is poor, the electrical and lattice indices of each region at that acquisition time are obtained. If the electrical index of a certain region is 1, the values ​​of each doping parameter are adjusted. If the lattice index of one or more regions is 1, the regions with the lattice index of 1 are annealed.

6. The artificial intelligence-based industrial automatic control system according to claim 5, characterized in that, The analysis of the doping effect of the semiconductor at each acquisition time is carried out in the following specific process: S31. Obtain the values ​​of each electrical parameter of each region at each acquisition time, perform normalization processing, and input them into the electrical analysis model. Output the electrical index of each region at each acquisition time. If the electrical index of a certain region at a certain acquisition time is 1, it means that the electrical performance of that region does not meet the requirements. If the electrical index of a certain region at a certain acquisition time is 0, it means that the electrical performance of that region meets the requirements. S32. Obtain the lattice diffraction pattern of each region at each acquisition time and compare it with the standard lattice diffraction pattern. If the lattice diffraction pattern of a certain region at a certain acquisition time is different from the standard lattice diffraction pattern, it means that the lattice index of that region at that acquisition time is 1, and the lattice in that region is damaged. If the lattice diffraction pattern of a certain region at a certain acquisition time is the same as the standard lattice diffraction pattern, it means that the lattice index of that region at that acquisition time is 0, and the lattice in that region is not damaged. S33. Obtain the electrical and lattice indices of each region at each acquisition time. When the electrical index or lattice index of one or more regions is 1 at a certain acquisition time, it indicates that the doping effect of the semiconductor at that acquisition time is poor. When the electrical and lattice indices of each region are 0 at a certain acquisition time, it indicates that the doping effect of the semiconductor at that acquisition time is good.

7. The artificial intelligence-based industrial automatic control system according to claim 5, characterized in that, The specific process for adjusting the values ​​of each doping parameter is as follows: S41. When the electrical properties of a certain region are 1, construct a virtual test bench and generate a virtual semiconductor and ion implanter. Set the electrode spacing range and magnetic field strength range. From the set of ion mixing ratios, the set of accelerating voltages, the electrode spacing range, and the magnetic field strength range, randomly select several ion mixing ratios, several accelerating voltages, several electrode spacings, and several magnetic field strengths, and randomly combine them to obtain several doping combinations. Each doping combination includes various ion mixing ratios, accelerating voltages, electrode spacings, and magnetic field strengths. S42. Apply each doping combination to a virtual test bench, and simultaneously obtain the doping effect of each doping combination. Select the doping combinations with good doping effects and refer to them as usable doping combinations. In the formula Representative in application number When the number of available doping combinations is , the th The first region A value for an electrical parameter. Representing the Standard values ​​for each electrical parameter Representing the The doping coefficients of a number of available doping combinations The numbers representing each region =1,2,3,..., , Represents the total number of regions. The numbers representing the various electrical parameters, =1,2,3,..., , Represents the total number of electrical parameters. The number representing each available doping combination. =1,2,3,..., , This represents the total number of available doping combinations. , , , , and All are positive integers. The doping combination with the highest doping coefficient is selected. The mixing ratio of each ion, the accelerating voltage, the electrode spacing and the magnetic field strength in the doping combination are used as the adjusted mixing ratio of each ion, the accelerating voltage, the electrode spacing and the magnetic field strength, and then adjusted.

8. The artificial intelligence-based industrial automatic control system according to claim 5, characterized in that, The annealing operation for each region with a lattice index of 1 is performed as follows: Regions with a lattice index of 1 are identified and designated as annealing regions. Lattice diffraction patterns are obtained for each annealing region, and the lattice damage coefficient within each region is analyzed. Simultaneously, the lattice damage coefficients from historical annealing events are retrieved from the database and compared. If the lattice damage coefficient from a historical annealing event is the same as that of a particular annealing region, the annealing parameters from that historical annealing event are retrieved from the database and used as the annealing parameters for that annealing region. This method is used to obtain a set of annealing parameters for each annealing region, where each element represents the value of the annealing parameter. The average value of the annealing parameters in the set for each annealing region is calculated and used as the standard value for the annealing parameters of each region. Simultaneously, the annealing apparatus is controlled to perform annealing operations on each region according to the annealing parameters of each region.

9. The artificial intelligence-based industrial automatic control system according to claim 8, characterized in that, The analysis of the lattice damage coefficient in each annealing region is performed as follows: Obtain the width and position of the diffraction peaks in the lattice diffraction patterns of each annealing region, as well as the width and position of the diffraction peaks in the standard lattice diffraction pattern. Calculate the width difference and displacement of the diffraction peaks in the lattice diffraction patterns of each annealing region. Based on the width difference and displacement of the diffraction peaks in the lattice diffraction patterns of each annealing region, calculate the lattice damage coefficient in each annealing region.

10. An industrial automatic control method for implementing the artificial intelligence-based industrial automatic control system according to any one of claims 1-9, characterized in that, include: Step 1: Obtain data: Retrieve the applications and crystal structures of semiconductors from the database; Step 2: Plan the ionic liquid: Obtain the lattice structure, application and ionic liquid used for each historical doping process from the database, and configure the ionic liquid based on the application and lattice structure of the semiconductor at that time. Step 3: Plan the ion implanter: Obtain the application of the semiconductor and the ion structure of each ion in the ion liquid used in each historical doping process from the database, and analyze the parameter values ​​of the ion implanter. Step 4: Controlling Doping: When doping the semiconductor according to the doping planning module, set the acquisition time according to the preset time interval, analyze the doping effect of the semiconductor at each acquisition time, and if the effect is not good, issue an early warning and make adjustments and remedies.

Images