Radiation defect evolution fluence rate simulation method and system based on molecular dynamics

Abstract

The application provides a kind of irradiation defect evolution fluence rate simulation method and system based on molecular dynamics, belongs to the technical field of simulation.The method comprises: obtaining the number of PKA generated by single incident particle irradiation semiconductor device;The semiconductor device is grid divided, the number of the initial PKA in each grid is obtained, and a simulation system model equal in size to the grid is established;Based on the simulation system model, the molecular dynamics method and the kinetic Monte Carlo method are used to simulate the defect evolution of the grid, and the defect information in the grid under the incident particle calibration fluence rate is obtained;The defect information in all the grids is summarized to obtain the comprehensive defect information of the semiconductor device.The application combines molecular dynamics and kinetic Monte Carlo method, realizes the simulation calculation of the defect evolution process of the whole semiconductor device under different fluence rates, and the calculation logic is clear, simple and easy to operate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of simulation, in particular to a simulation method and system for irradiation defect evolution fluence rate based on molecular dynamics. BACKGROUND

[0002] In the space environment, semiconductor devices are prone to damage by space charged particles. Especially high-energy incident particles interact with device materials, producing primary collision particles and their cascade collision processes, causing device irradiation damage. The cascade damage of primary collision particles is at the microscale, which is difficult to observe experimentally, and there are deficiencies such as long required period, high cost, complex method, and harsh conditions. Molecular dynamics is an important computer simulation method for solving many-body problems at the atomic level. In order to accurately describe the cascade microdynamic evolution process of primary collision particles, molecular dynamics method can be used.

[0003] However, the current simulation calculation of semiconductor device irradiation defects generally uses a relatively rough simulation method. For example, the relevant information of the input irradiation particle flow often only has the particle number and particle energy. However, as can be known from the experimental results, the fluence rate of irradiation particles also has a very important influence on the evolution of semiconductor device irradiation defects, resulting in low simulation accuracy of the existing method. SUMMARY

[0004] The present application solves the problem of how to conveniently and accurately simulate the fluence rate effect in semiconductor device defect evolution simulation.

[0005] To solve the above problems, the present application provides a simulation method for irradiation defect evolution fluence rate based on molecular dynamics, comprising:

[0006] Step S1, obtaining the number of primary collision particles generated by irradiation of a single incident particle on a semiconductor device;

[0007] Step S2, performing grid division on the semiconductor device to obtain the number of primary collision particles ni in each grid.

[0008] Step S3, establishing a simulation system model with the same size as the grid;

[0009] Step S4, based on the simulation system model, using molecular dynamics method and kinetic Monte Carlo method to simulate defect evolution of the grid, and obtaining defect information in the grid under the incident particle calibration fluence rate;

[0010] Step S5, aggregating the defect information in all the grids to obtain comprehensive defect information of the semiconductor device.

[0011] Preferably, the defect evolution simulation of the grid by the molecular dynamics method and the kinetic Monte Carlo method comprises:

[0012] In step S41, the generation process of the primary collision particles in the semiconductor device is simulated by the molecular dynamics method, and a steady-state structure under the molecular dynamics time scale is outputted;

[0013] In step S42, the steady-state structure is simulated by the kinetic Monte Carlo method, and the simulation time is the time of generation of the primary collision particles in adjacent batches;

[0014] The step S41 and the step S42 are repeated.

[0015] Preferably, the time of generation of the primary collision particles in adjacent batches is Δt, and Δt is obtained according to formula (1) and formula (2);

[0016]

[0017]

[0018] wherein, is the number of the primary collision particles generated per second in the grid under the calibrated injection rate of the incident particles, FR is the calibrated injection rate of the incident particles, S is the cross-sectional area of the grid perpendicular to the incident direction of the incident particles, n i is the number of the primary collision particles generated in the i-th grid, and N is the total number of the incident particles.

[0019] Preferably, the obtaining of the number of the primary collision particles generated by a single incident particle irradiating the semiconductor device comprises:

[0020] The number of the primary collision particles generated after different numbers of incident particles irradiating the semiconductor device are obtained, and the number of the primary collision particles generated after the incident particles irradiating the semiconductor device tends to a convergence value as the number of the incident particles increases;

[0021] The sum of the numbers of the primary collision particles generated by all the incident particles is counted;

[0022] The sum of the numbers of the primary collision particles is divided by the total number N of the incident particles to obtain the number of the primary collision particles generated by a single incident particle.

[0023] Preferably, the establishment of the simulation system model comprises: establishing a pure silicon model in LAMMPS, and introducing a doping element into the pure silicon model;

[0024] The doping elements include at least one of C, O, B and P, the doping modes of C and O are interstitial atom doping, the doping rate is 10 12 / cm 3 -10 14 / cm 3 The doping modes of B and P are substitution atom doping, the doping rate is 10 18 / cm 3 -10 20 / cm 3 .

[0025] Preferably, the obtaining of the defect information in the grid at the incident particle calibration fluence rate comprises:

[0026] Obtaining the defect information corresponding to each of the primary collision particles generated in the same batch in the grid,

[0027] Obtaining the average value of the defect information corresponding to each of the primary collision particles, to obtain the defect information in the grid at the incident particle calibration fluence rate.

[0028] Preferably, the defect information includes the type of defects and the number or concentration of each type of defects.

[0029] Preferably, the summarizing of the defect information in all the grids comprises:

[0030] Proportionally sampling the grids containing the same number of primary collision atoms;

[0031] Obtaining the type of defects and the number of each type of defects in the sampled grid, and proportionally summing to obtain the type of defects and the total number of each type of defects in all the grids.

[0032] Preferably, the summarizing of the defect information in all the grids further comprises:

[0033] Dividing the total number of each type of defects by the total volume of the semiconductor device to obtain the type of defects and the concentration of each type of defects in the semiconductor device.

[0034] The advantage of the irradiation defect evolution fluence rate simulation method based on molecular dynamics of the present application over the prior art is that:

[0035] In the process of space irradiation, the parameters of the incident particle flow will have an important influence on the irradiation defect evolution of the semiconductor device, and the existing technology pays more attention to the number and energy of the incident particles. The simulation method of the present application can consider the influence of the fluence of the incident particles on the irradiation defect evolution, thereby solving the problem that the fluence effect cannot be reflected in the defect evolution simulation of the semiconductor device. The simulation method combines molecular dynamics and kinetic Monte Carlo method to realize the simulation calculation of the defect evolution process of the entire semiconductor device under different fluences. The calculation logic is clear, the steps are simple and easy to operate, and the method is in close contact with the actual situation in the entire simulation process, and the simulation result is close to the experimental data.

[0036] The present application also provides a fluence simulation system for irradiation defect evolution based on molecular dynamics, comprising:

[0037] A grid division module is configured to divide the semiconductor device into grids;

[0038] An acquisition module is configured to acquire the number of primary collision particles generated by irradiation of a single incident particle on a semiconductor device;

[0039] The acquisition module is further configured to acquire the number of primary collision particles ni in each grid;

[0040] The acquisition module is further configured to acquire the defect information in the grid under the calibrated fluence of the incident particles;

[0041] A modeling module is configured to establish a simulation system model with the same size as the grid;

[0042] A simulation module is configured to perform molecular dynamics simulation and kinetic Monte Carlo simulation on the simulation system model;

[0043] A statistical module is configured to aggregate the defect information in all the grids to obtain comprehensive defect information of the semiconductor device.

[0044] The fluence simulation system for irradiation defect evolution based on molecular dynamics of the present application has the same advantages as the fluence simulation method for irradiation defect evolution based on molecular dynamics of the prior art, which will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The flowchart of the fluence simulation method for irradiation defect evolution based on molecular dynamics in the embodiments of the present application;

[0046] Figure 2 The defect evolution calculation result graph under different incident particle injection rates in the embodiments of the present application. DETAILED DESCRIPTION

[0047] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, specific embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0048] Please refer to Figure 1 The simulation method of the embodiment of the present application based on molecular dynamics for irradiation defect evolution fluence rate comprises:

[0049] Step S1, obtaining the number of primary collision particles generated by irradiation of a single incident particle on a semiconductor device;

[0050] Step S2, performing grid division on the semiconductor device to obtain the number of primary collision particles ni in each grid.

[0051] Step S3, establishing a simulation system model with the same size as the grid;

[0052] Step S4, based on the simulation system model, using the molecular dynamics method and the kinetic Monte Carlo method to simulate the defect evolution of the grid, and obtaining the defect information in the grid under the incident particle calibration fluence rate;

[0053] Step S5, summarizing the defect information in all grids to obtain the comprehensive defect information of the semiconductor device.

[0054] The simulation method of the embodiment combines molecular dynamics and kinetic Monte Carlo method to realize the simulation calculation of the defect evolution process of the entire semiconductor device under different fluence rates, solves the problem that the defect evolution simulation of the semiconductor device in the prior art cannot reflect the fluence rate effect, the calculation logic of the simulation method is clear, the steps are simple and easy to operate, and the method is in close contact with the actual situation in the entire simulation process, and the simulation result is close to the experimental data.

[0055] In some embodiments, the step S1 of obtaining the number of primary collision particles generated by irradiation of a single incident particle on a semiconductor device comprises:

[0056] Obtaining the number of primary collision particles generated after irradiation of the semiconductor device by different numbers of incident particles, and as the number of incident particles increases, the number of primary collision particles generated after irradiation of the semiconductor device by the incident particles tends to a convergence value;

[0057] Summing up the number of primary collision particles generated by all incident particles;

[0058] Dividing the sum of the number of primary collision particles by the total number N of incident particles to obtain the number of primary collision particles generated by a single incident particle.

[0059] Specifically, for a certain type of incident particles, the number of primary knock-on atoms (PKA) generated by the incident particles at a certain energy is calculated using the Monte Carlo method. To ensure accuracy, the calculation is performed for different numbers of incident particles. The number of incident particles is continuously increased until the number of PKA generated by the incident particles converges to a value, and then normalized by dividing the total number of generated PKA by the total number of incident particles to obtain the number of PKA generated by a single incident particle.

[0060] where the number of PKA generated by a single incident particle refers to the ratio of the total number of PKA generated by irradiation T pka to the total number of incident particles N, which does not change significantly with the increase of incident particles. The convergence value is used because the Monte Carlo method calculates the number of incident particles as a single particle, so a converged normalized value is better for description.

[0061] In some embodiments, the semiconductor is meshed in step S2. To ensure moderate grid density in the semiconductor device and facilitate subsequent calculations, the grid box size is set to 70nm*70nm*70nm. After meshing, the grid boxes without PKA are cleared, and the grid boxes containing PKA are retained. The PKA information contained in each box after the incident particles are incident on the semiconductor device is counted, including the number, energy, position, and incident direction of the PKA.

[0062] In some embodiments, establishing a simulation system model includes: first, establishing a pure silicon model in LAMMPS, and expanding the Si unit cell by a suitable multiple in the x, y, and z directions to form a silicon structure equal in size to the grid size. In this embodiment, since the grid size is 70nm*70nm*70nm, the Si unit cell is expanded by 129 times in the x, y, and z directions to form a 70nm*70nm*70nm Si structure.

[0063] Then, to ensure the fit of the simulation model with the actual, doped elements are introduced into the pure silicon model. Specifically, according to the actual doping method and doping concentration, various existing doped elements such as C, O, B, and P are introduced into the pure Si system. Among them, C and O elements are doped into Si as interstitial atoms, with a doping rate of 10 12 / cm 3 —10 14 / cm 3 , B and P elements are doped into Si as substitutional atoms, with a doping rate of 10 18 / cm 3 —10 20 / cm 3 .

[0064] In some embodiments, the fluence rate effect is simulated in step S4 while keeping the fluence unchanged. The fluence rate is achieved by setting the number of PKA in the ith box as n i , the nominal fluence rate of the incident particles is FR, and the cross-sectional area of the box perpendicular to the incident direction of the incident particles is S. Then, the number of PKA falling in the box per second at the nominal fluence rate is calculated as follows in equation (2):

[0065]

[0066] where N is the total number of incident particles.

[0067] The time interval Δt between the falling of the PKA in adjacent batches is calculated according to the number of PKA falling per second, as follows in equation (1):

[0068]

[0069] That is, according to the number of primary collision particles generated in the grid after irradiation of a single incident particle, the nominal fluence rate of the incident particles, and the area of the grid, the number of primary collision particles generated in each batch at the nominal fluence rate of the incident particles and the total number of primary collision particles can be obtained.

[0070] Further, the structure state when the PKA in the second batch falls is determined according to Δt. This process needs to be closely combined with molecular dynamics and Kinetic Monte Carlo. Generally, the time scale of molecular dynamics is in the order of picoseconds, and Δt here is in the order of seconds. Therefore, the defect evolution simulation of the grid by using molecular dynamics and Kinetic Monte Carlo includes:

[0071] Step S41, the generation process of the primary collision particles in the semiconductor device is simulated by using molecular dynamics, and a steady-state structure in the time scale of molecular dynamics is outputted;

[0072] Step S42, the steady-state structure is simulated by using Kinetic Monte Carlo, and the simulation time is the time of the generation of the primary collision particles in adjacent batches;

[0073] Steps S41 and S42 are repeated.

[0074] ​Specifically, the molecular dynamics method is used to evolve to a steady state under the molecular dynamics time scale, and the evolution time is about 100 ps. Then the structure is evolved under the Kinetic Monte Carlo method to time Δt, and the structure evolved by the Kinetic Monte Carlo method is returned to the molecular dynamics method to calculate the defect evolution process of the second batch of PKAs. If the number of PKA batches increases, the above process is repeated.

[0075] In some embodiments, the step S4 of obtaining the defect information in the grid under the calibration fluence of the incident particles comprises:

[0076] obtaining the defect information corresponding to each primary collision particle generated in the same batch in the grid,

[0077] obtaining the average value of the defect information corresponding to each primary collision particle to obtain the defect information in the grid under the calibration fluence of the incident particles.

[0078] In some embodiments, due to the computing power, the step S5 can not be calculated for all boxes, but only for the boxes containing the same number of PKAs. After statistical calculation by proportional sampling, the defect types and the concentration of each defect of the entire device can be obtained by proportional summation and division by the total volume of the device.

[0079] In an example, the defect information includes the type of defects and the number of each type of defects, and therefore, the step of summarizing the defect information in all grids comprises: proportionally sampling the grids containing the same number of primary knock-on atoms; obtaining the type of defects and the number of each type of defects in the sampled grids, and performing proportional summation to obtain the type of defects and the total number of each type of defects in all grids. Figure 2 As shown in FIG. 6, the number of vacancy defects generated after the incident particles irradiate the semiconductor device under different fluences is shown. It can be seen that the number of defects generated under different fluences is different, and thus, the embodiment realizes the defect evolution simulation calculation of the entire semiconductor device under different fluences in a combination of the Monte Carlo, molecular dynamics and Kinetic Monte Carlo methods. Through the simulation of the fluence effect, the influence of the fluence on the defect evolution of the semiconductor device can be understood.

[0080] In another example, the defect information includes the type of defects and the concentration of each type of defects. Therefore, the step of summarizing the defect information in all grids comprises: proportionally sampling the grids containing the same number of primary knock-on atoms; obtaining the type of defects and the number of each type of defects in the sampled grids, performing proportional summation, and then dividing the total number of each type of defects by the total volume of the semiconductor device to obtain the type of defects and the concentration of each type of defects in the semiconductor device.

[0081] The simulation system comprises a molecular dynamics-based irradiation defect evolution fluence rate simulation system, a grid division module, a modeling module, a simulation module and a statistics module.

[0082] The grid division module is configured to divide the semiconductor device into grids.

[0083] The acquisition module is configured to acquire the number of primary collision particles generated by the single incident particle irradiating the semiconductor device.

[0084] The acquisition module is further configured to acquire the number of primary collision particles ni in each grid.

[0085] The acquisition module is further configured to acquire the defect information in the grid under the incident particle calibration fluence rate.

[0086] The modeling module is configured to establish a simulation system model with the same size as the grid.

[0087] The simulation module is configured to perform molecular dynamics simulation and kinetic Monte Carlo simulation on the simulation system model.

[0088] The statistics module is configured to aggregate the defect information in all grids to obtain comprehensive defect information of the semiconductor device.

[0089] Although the present disclosure is disclosed as above, the protection scope of the present disclosure is not limited to this. Those skilled in the art can make various modifications and changes without departing from the spirit and scope of the present disclosure, and these modifications and changes shall fall within the protection scope of the present disclosure.

Claims

1. A method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics, characterized in that, include: Step S1: Obtain the number of primary collision particles generated by a single incident particle irradiating a semiconductor device; Step S2: Divide the semiconductor device into grids and obtain the number of primary collision particles in each grid. Step S3: Establish a simulation system model with the same size as the mesh; Step S4: Based on the simulation system model, the defect evolution of the mesh is simulated using molecular dynamics and dynamic Monte Carlo methods, and the defect information in the mesh under the incident particle calibration flux rate is obtained. Step S5: Summarize the defect information in all the grids to obtain the comprehensive defect information of the semiconductor device; The method of using molecular dynamics and dynamic Monte Carlo methods to simulate the defect evolution of the mesh includes: Step S41: The generation process of the primary colliding particles in the semiconductor device is simulated using molecular dynamics methods, and the steady-state structure under the molecular dynamics timescale is output. Step S42: The steady-state structure is simulated using the dynamic Monte Carlo method, and the simulation duration is the time it takes for two adjacent batches of primary collision particles to be generated. Repeat steps S41 and S42; The time between the generation of the primary collision particles in the two adjacent batches is , Obtained according to equations (1) and (2); （1）； （2）； in, FR is the number of primary collision particles generated per second in the grid at the calibrated injection rate of the incident particles, and S is the cross-sectional area of ​​the grid perpendicular to the incident direction of the incident particles. The number of primary collision particles generated in the i-th grid, where N is the total number of incident particles.

2. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 1, characterized in that, The method of obtaining the number of primary collision particles generated by a single incident particle irradiating a semiconductor device includes: The number of primary collision particles generated after different numbers of incident particles irradiate the semiconductor device is obtained, and as the number of incident particles increases, the number of primary collision particles generated after the incident particles irradiate the semiconductor device tends to converge to a convergent value. The sum of the number of primary collision particles produced by all the incident particles is counted; The number of primary collision particles generated by a single incident particle is obtained by dividing the sum of the number of primary collision particles by the total number of incident particles.

3. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 1, characterized in that, The establishment of the simulation system model includes: establishing a pure silicon model in LAMMPS and introducing doping elements into the pure silicon model; The doping element includes at least one of C, O, B, and P, wherein the C and O elements are doped via interstitial atom doping with a doping rate of 10. 12 / cm 3 —10 14 / cm 3 The doping method for elements B and P is substitutional atom doping, with a doping rate of 10. 18 / cm 3 —10 20 / cm 3 .

4. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 1, characterized in that, The step of obtaining defect information in the mesh at the incident particle calibration flux rate includes: Obtain the defect information corresponding to each of the primary collision particles generated in the same batch in the mesh. The average value of the defect information corresponding to each of the primary collision particles is obtained to obtain the defect information in the grid under the incident particle calibration flux rate.

5. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 4, characterized in that, The defect information includes the type of defect and the number or concentration of each type of defect.

6. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 5, characterized in that, The process of summarizing the defect information in all the grids includes: Extract meshes containing the same number of the aforementioned primary ejected atoms proportionally; Obtain the type of defect and the number of defects of each type in the extracted mesh, and sum them proportionally to obtain the defect types of all the meshes and the total number of defects of each type.

7. The method for simulating the fluence rate of irradiation defect evolution based on molecular dynamics according to claim 6, characterized in that, The process of summarizing the defect information in all the grids also includes: Dividing the total amount of each type of defect by the total volume of the semiconductor device yields the type of defect and the concentration of each type of defect in the semiconductor device.

8. A fluence rate simulation system for irradiation defect evolution based on molecular dynamics, characterized in that, include: A mesh generation module, which is used to perform mesh generation on semiconductor devices; The acquisition module is used to acquire the number of primary collision particles generated by a single incident particle irradiating a semiconductor device; The acquisition module is also used to acquire the number ni of the primary collision particles in each grid; The acquisition module is also used to acquire defect information in the grid under the incident particle calibration flux rate; A modeling module, which is used to establish a simulation system model of the same size as the mesh; The simulation module is used to perform molecular dynamics simulations and dynamic Monte Carlo simulations on the simulation system model; The defect evolution simulation of the mesh using molecular dynamics and dynamic Monte Carlo methods includes: Step S41: The generation process of the primary colliding particles in the semiconductor device is simulated using molecular dynamics methods, and the steady-state structure under the molecular dynamics timescale is output. Step S42: The steady-state structure is simulated using the dynamic Monte Carlo method, and the simulation duration is the time it takes for two adjacent batches of primary collision particles to be generated. Repeat steps S41 and S42; The time between the generation of the primary collision particles in the two adjacent batches is , Obtained according to equations (1) and (2); （1）； （2）； in, FR is the number of primary collision particles generated per second in the grid at the calibrated injection rate of the incident particles, and S is the cross-sectional area of ​​the grid perpendicular to the incident direction of the incident particles. The number of primary collision particles generated in the i-th grid, where N is the total number of incident particles; The statistics module is used to summarize the defect information in all the grids to obtain the comprehensive defect information of the semiconductor device.

Images