Method for simulating halogen plasma etching silicon or silicon dioxide through molecular dynamics

Abstract

The invention discloses a method for simulating halogen plasma etching silicon or silicon dioxide through molecular dynamics. The method comprises the steps of modeling through a silicon or silicon dioxide substrate material, performing LAMMPS software initialization setting, performing plasma etching circulation, and performing data output and analysis processing. According to the method, a plasma-surface interaction mechanism is simulated on the atomic scale by using a molecular dynamics method, the property of a surface modification layer is analyzed, the simulated etching conditions are independent and controllable, plasma etching with different energies, incident angles and incident doses can be simulated, and the influence of a single variable on an etching result can be accurately analyzed; the method can accurately and quickly predict the effect of etching silicon or silicon dioxide by the halogen plasma under different incidence conditions, avoids high cost, complexity and uncertainty caused by experiments, and greatly shortens the development process of the process of etching silicon or silicon dioxide by the halogen plasma.

Description

Technical Field

[0001] The invention relates to a method for etching silicon or silicon dioxide using halogen plasma, and in particular to a method for etching silicon or silicon dioxide using halogen plasma using molecular dynamics simulation. Background Art

[0002] Plasma etching is a material removal process widely used in semiconductor manufacturing. It emits a plasma stream of a suitable gas mixture in pulsed form to the sample. The plasma source (called etching species) can be charged ions or neutral radicals, which can remove substances from the surface of the material by physical sputtering or chemical reaction to generate volatile products.

[0003] Among them, in the process of halogen plasma etching of silicon or silicon dioxide, the properties of the substrate surface modification layer greatly affect the final etching effect, but the plasma-surface interaction mechanism at the atomic scale is unclear, quantitative experimental characterization of the surface etching process is difficult, and the details of the surface reaction are difficult to accurately describe; and the halogen plasma etching of silicon or silicon dioxide has high experimental costs, long time periods, high complexity of experimental conditions, uncontrollable ion incidence conditions, and difficult to predict etching results (etching yield, product composition, etc.).

[0004] Molecular dynamics is a set of molecular simulation methods. This method mainly relies on computers to simulate the motion of molecular and atomic systems. It is a multi-body simulation method. By simulating the motion state of molecules and atoms within a certain period of time, the behavior of the system evolving over time can be examined from a dynamic point of view. Usually, the trajectories of molecules and atoms are obtained by numerically solving Newton's equations of motion, and the potential energy (or its first-order partial derivative with respect to the coordinates, i.e., force) can usually be given by the intermolecular interaction potential energy function. Samples are drawn from a system composed of different states of the molecular system to calculate the configuration integral of the system, and the thermodynamic quantities and other macroscopic properties of the system are further calculated based on the results of the configuration integral.

[0005] Existing molecular dynamics simulation technology has been applied to other fields such as magnetron sputtering thin film deposition and C atom bombardment of graphene, but there is a lack of molecular dynamics methods for simulating halogen plasma etching of silicon or silicon dioxide. The main technical obstacle is that it is difficult to extract the key steps in the complex halogen plasma etching of silicon or silicon dioxide process and abstract them into molecular dynamics simulation models to ensure the accuracy of etching simulation results while taking into account simulation efficiency. Therefore, in order to solve the problems of high cost, long cycle, complex conditions, and difficult result prediction in actual etching experiments, it is urgent to research and develop new methods. Summary of the invention

[0006] Purpose of the invention: The purpose of the present invention is to provide an independently controllable and rapidly predictable method for molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide.

[0007] Technical solution: The method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention comprises the following steps:

[0008] (1) Modeling of silicon or silicon dioxide substrate materials: Use the modeling software Materials Studio to build the initial structural model of the substrate material according to the lattice structure of silicon or silicon dioxide, determine the geometric parameters and save them into LAMMPS, and use the fix command to fix the bottom two layers of silicon or silicon dioxide atoms;

[0009] (2) LAMMPS software initialization settings: set the LAMMPS simulation environment parameters, force field, energy minimization, initial temperature, relaxation, plasma generation area and screen output respectively, and group the silicon or silicon dioxide substrate, plasma etching material and etching products for etching process identification;

[0010] (3) Plasma etching cycle: set the number of etching cycles, create etching ions and set the radius and density according to the incident ion type, set the initial velocity according to the ion energy, set the incident angle, and set the atoms to maintain charge neutrality; perform a single ion bombardment cycle: first, except for the fixed atoms at the bottom, all atoms evolve under the nve ensemble for 0.5-2ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs, then the Langevin or Berendsen thermostat algorithm is applied to the entire system to cool the temperature to the initial temperature of 300K, and finally, the substances not bound to the material (all sputtered or desorbed substances) are removed from the system as etching products, and the remaining atoms form the input model for the next ion bombardment;

[0011] (4) Data output and analysis: Output the macroscopic measurement parameters to be measured, visualize the output using OVITO software, and analyze the properties of the surface modification layer of the silicon or silicon dioxide substrate, as well as the silicon etching yield and product composition.

[0012] Wherein, in step (1), the geometric parameters, including the number of atoms, surface area, number of stacking layers and shape structure, are set according to simulation requirements.

[0013] Wherein, in step (2), the settings of the LAMMPS simulation environment parameters are specifically as follows: the units of the simulation system are set to metal, the boundary conditions are set to ppf (x periodic boundary conditions, y periodic boundary conditions, z fixed boundary conditions), the atom type (atom_style) is set to sphere, the neighbor list (neighbor) is defined as default, the time step (timestep) is set to 0.2fs when the ion energy is less than 200eV, and the time step is set to 0.1fs when the ion energy is greater than 200eV and less than 500eV;

[0014] Among them, in step (2), the setting of the force field is the setting of the potential function, which is specifically: setting the corresponding potential function according to the atom type in the simulation system, first using the pair_style command to set the force field type to the SW potential function, and then using the pair_coeff command to set the corresponding atom type to potential function mapping.

[0015] Wherein, in step (2), the energy minimization setting is specifically: using the min_style command to determine that the energy minimization algorithm is the conjugate gradient algorithm, and using the minimize command to set relevant parameters to minimize the system energy and eliminate unreasonable structure of the model;

[0016] In step (2), the initial temperature is set as follows: the temperature of the system is calculated based on the velocity of the atoms, and a random initial velocity is applied to the atoms of the material substrate according to a Gaussian distribution using the velocity command, and the initial temperature of the system is set to 300K;

[0017] Wherein, in step (2), the relaxation setting is specifically: in the NVT ensemble, the system is relaxed for 1-10ps to obtain an equilibrium substrate structure;

[0018] Wherein, in step (2), the plasma generation region is specifically set as follows: a portion slightly larger than the potential function cutoff distance above the substrate surface; Set the plasma generation area, and the substrate surface height is defined as the z coordinate of the highest atom in the system before each etching cycle starts;

[0019] Wherein, in step (2), the screen output is set as follows: using the thermo command to set the screen to print system information every 1000 time steps, and the thermo_style command to set the output physical quantities (temperature, potential energy, kinetic energy, pressure, current time step number, total number of atoms in the system) so as to monitor the system information in real time during the simulation process;

[0020] Wherein, in step (2), the grouping is specifically: using the group command to group the silicon or silicon dioxide substrate, plasma etching material, etching products, etc. for etching process identification.

[0021] Wherein, in step (3), the number of etching cycles is set, and the plasma etching cycle consists of halogen ions cyclically bombarding the surface of a silicon or silicon dioxide substrate; before the etching cycle starts, the loop command is used to set the number of cycles, and the number of cycles is preferably 800-2000 times. The number of cycles should be large enough to ensure that the system reaches a stable state and increase the statistical data of the macro measurement parameters.

[0022] In step (3), the etching ions are created: first, an incident ion is created in the etching material generation area, and its horizontal position is random. The radius and density are set according to the type of incident ion through the set command (the atomic diameter of F is Density is 12.40826g / cm 3 ), set the initial velocity according to the ion energy (225.36m / s for 50eV, 318.7m / s for 100eV, and 390.33m / s for 150eV), and set the incident angle (perpendicular to the substrate surface), etc.; Charge neutrality assumption: Due to the electron transfer of surface atoms, it is assumed that the incident ions are charge neutralized before hitting the surface, so all atoms remain charge neutral in the simulation. (Specifically: The diameter of the F atom is Density is 12.40826g / cm 3 , the energy of 50eV corresponds to a velocity of 225.36m / s, the energy of 100eV corresponds to a velocity of 318.7m / s, and the energy of 150eV corresponds to a velocity of 390.33m / s; the diameter of the Cl atom is Density is 8.99186g / cm 3 , the energy of 50eV corresponds to a velocity of 164.9766m / s, the energy of 100eV corresponds to a velocity of 233.3121m / s, and the energy of 150eV corresponds to a velocity of 285.7479m / s)

[0023] Among them, in step (3), a single ion bombardment cycle can be divided into three stages: first, except for the fixed atoms at the bottom, all atoms evolve in the nve ensemble for 0.5-2ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs. Then, the Langevin or Berendsen thermostat algorithm is applied to the entire system, and the temperature damping is set to 100 times the time step, and the temperature is cooled to the initial 300K for 0.5-2.0ps. Finally, all sputtered or desorbed substances are removed from the system as etching products using the delete_atoms command, and the remaining atoms form the input model for the next ion bombardment.

[0024] Among them, in step (3), in order to prevent the silicon or silicon dioxide substrate from being broken down during the etching process, the create_atoms command is used every 200 time steps to add new 4-8 layers of silicon or silicon dioxide atoms to the bottom of the substrate and the fix command is used to set the forces in the x, y, and z directions to 0 to re-fix the bottom two layers of atoms.

[0025] Among them, in step (4), the dump and dump_modify commands are used to output the macroscopic measurements to be measured, including atomic type, velocity, coordinates, force, diameter and other calculated parameters at regular time intervals (100-5000 steps), and the silicon etching output is the average number of silicon atoms removed in a single ion bombardment.

[0026] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) The molecular dynamics method is used to simulate the plasma-surface interaction mechanism at the atomic scale and analyze the properties of the surface modification layer, thereby deepening the understanding of the mechanism of halogen plasma etching of silicon or silicon dioxide;

[0027] (2) The simulated etching conditions are independently controllable. Halogen plasma etching with different energies, incident angles, and incident doses can be simulated, and the influence of a single variable on the etching results can be accurately analyzed.

[0028] (3) It can accurately and quickly predict the effect of halogen plasma etching of silicon or silicon dioxide under different incident conditions, avoiding the high cost, complexity and uncertainty brought by experiments, and greatly shortening the development process of halogen plasma etching of silicon or silicon dioxide. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of the method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention;

[0030] Figure 2 is a front view of an initial model of fluorine plasma etching silicon in Example 1 of the present invention;

[0031] Figure 3 is a front view of a steady-state model of fluorine plasma etching silicon in Example 1 of the present invention;

[0032] Figure 4 is a graph showing the relationship between the fluorine plasma energy and the properties of the substrate surface modification layer in Example 1 of the present invention;

[0033] Figure 5 is a graph showing the relationship between the chlorine plasma energy and the properties of the substrate surface modification layer in Example 1 of the present invention;

[0034] Figure 6is a graph showing the relationship between the fluorine and chlorine plasma energies and the silicon yield in Example 1 of the present invention;

[0035] Figure 7 is a graph showing the relationship between fluorine plasma energy and product components in Example 1 of the present invention;

[0036] Figure 8 is a graph showing the relationship between chlorine plasma energy and product components in Example 1 of the present invention;

[0037] Fig. 9 It is a front view of the initial model of fluorine reactive ion etching of silicon in Example 2 of the present invention;

[0038] Fig.10 It is a front view of a steady-state model of fluorine reactive ion etching of silicon in Example 2 of the present invention;

[0039] Fig.11 is a graph showing the relationship between the fluorine plasma energy and the properties of the substrate surface modification layer in Example 2 of the present invention;

[0040] Fig.12 is a graph showing the relationship between fluorine plasma energy and silicon yield in Example 2 of the present invention;

[0041] Fig.13 is a graph showing the relationship between fluorine plasma energy and product components in Example 2 of the present invention;

[0042] Fig.14 is a front view of an initial model of a silicon dioxide substrate in Example 3 of the present invention;

[0043] Fig.15 is a front view of a steady-state model of fluorine plasma etching silicon dioxide in Example 3 of the present invention;

[0044] Fig.16 is a graph showing the relationship between the fluorine plasma energy and the properties of the surface modified layer of the silicon dioxide substrate in Example 3 of the present invention;

[0045] Fig.17 is a graph showing the relationship between the chlorine plasma energy and the properties of the surface modified layer of the silicon dioxide substrate in Example 3 of the present invention;

[0046] Fig.18 3 is a graph showing the relationship between the fluorine and chlorine plasma energies and the silicon yield in Example 3 of the present invention. DETAILED DESCRIPTION

[0047] The technical scheme of the present invention is further described below in conjunction with the examples. The test materials used in the examples can all be purchased through conventional channels.

[0048] Example 1

[0049] The method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention comprises the following steps:

[0050] (1) Modeling of silicon or silicon dioxide substrate materials:

[0051] The initial model of the silicon substrate material was constructed using Materials Studio modeling software according to the diamond lattice structure. The lattice constant is The geometric parameters such as the number of atoms, surface area, number of stacking layers, shape structure, etc. are determined according to the simulation needs and saved as a data file that can be recognized by the LAMMPS software. Then, the read_data command is used to read it into LAMMPS, and the fix command is used to fix the bottom two layers of silicon atoms.

[0052] like Figure 2 As shown, in this embodiment, the initial model substrate is a substrate with a side length of The top area of ​​a cube Corresponding to the silicon (100) crystal plane, there are a total of 17 layers of silicon, with 32 silicon atoms in each layer (denoted as 1ML), for a total of 544 silicon atoms.

[0053] (2) LAMMPS software initialization settings:

[0054] LAMMPS simulation environment parameter settings: the unit of the simulation system is set to metal, the boundary conditions in the X and Y directions (horizontal direction) are set to periodic boundary conditions, the Z direction is set to fixed boundary conditions, the atom type is set to sphere, the neighbor list is selected as default, and the time step is set to 0.2fs.

[0055] Force field (potential function) setting: Use the pair_style command to set the Stillinger-Weber (SW) potential function according to the silicon and fluorine or chlorine atom system in the simulation system, and then use the pair_coeff command to set the corresponding SW potential function file and silicon, fluorine or chlorine atom mapping;

[0056] Specifically, the SW potential function is as follows:

[0057]

[0058] Among them, Ф represents the total potential energy of the system, v2 and v3 are the two-body potential term and the three-body potential term respectively. The total potential energy is represented by the sum of the two-body potential term and the three-body potential term. The two-body potential term is of Lennard-Jones type, with an exponential cutoff function to limit its scope of action; the three-body potential term provides that when the configuration of three adjacent atoms deviates significantly from the diamond lattice, that is, when the three-body angle ≠ 109.5°, energy is lost rapidly.

[0059] Energy minimization: Use the min_style command to determine that the energy minimization algorithm is the conjugate gradient algorithm, and the minimize command sets the relevant parameters. By minimizing energy, the unreasonable structure of the model is eliminated;

[0060] Temperature initialization: The temperature of the system is calculated based on the velocity of the atoms, so temperature initialization is to initialize the velocity of the atoms. Use the velocity command to apply a random initial velocity to each atom of the substrate according to the Gaussian distribution. In this example, the initial temperature of the system is set to 300K.

[0061] Relaxation: Under the NVT ensemble, the system was relaxed for 5 ps with a damping time of 0.02 ps (100 times the time step) to obtain the equilibrium substrate structure.

[0062] Set up the ion generation region: Set up the plasma generation region at a height slightly greater than the potential function cutoff distance above the substrate surface. The substrate surface height is defined as the z coordinate of the highest atom in the system before each etching cycle begins.

[0063] Specifically, Figure 2 As shown, in this embodiment, the initial model substrate An ion generation region is set at a height of 1000 m, and in subsequent etching cycles, the substrate surface height and the ion generation region are updated before each ion injection cycle begins as the substrate surface height changes;

[0064] Grouping: Use the group command to group silicon substrates, plasma etching materials, etching products, etc. for etching process identification;

[0065] Screen output settings: Use the thermo command to set the screen to print system information every 1000 steps, and the thermo_style command to set the output physical quantities, including system temperature, potential energy, kinetic energy, and number of atoms, so as to monitor system information in real time during the simulation;

[0066] (3) Plasma etching cycle:

[0067] Set the number of cycles: The plasma etching cycle consists of halogen ions bombarding the silicon surface cyclically. Before the etching cycle starts, use the loop command to set the number of cycles to 1000 (about 31ML);

[0068] Create etching ions: First, use create_atom to create a fluorine ion in the etching material generation area. The horizontal position is generated by a random number (random command), and the diameter is set to Density is 12.40826g / cm 3 (If it is a chloride ion, the diameter is Density is 8.99186g / cm 3 ), the ion energy was set to 50 eV, and the incident angle was set to be perpendicular to the substrate surface downward (v z <0);

[0069] Charge neutrality assumption: Due to the electron transfer of surface atoms, it is assumed that the incident ions are charge neutralized before hitting the surface, so all atoms remain charge neutral in the simulation. Specifically, the potential function does not consider the effect of charge and the atom type is set to sphere, which does not include atomic charge information.

[0070] A single ion bombardment can be divided into three stages: first, except for the two fixed atoms at the bottom, all atoms evolve in the nve ensemble for 1 ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs. Then, the Berendsen thermostat algorithm is applied to the entire system to cool the temperature to the initial temperature of 300 K, with a heat bath damping of 0.02 ps (100 times the time step). Finally, all atoms that have no interaction with the substrate and have a velocity away from the surface (v z >0) are removed from the system and marked as etching products, and the remaining atoms form the input model for the next ion bombardment;

[0071] To prevent the silicon substrate from being broken down during the etching process, two silicon atomic layers were added to the bottom of the substrate after every 200 ion etching cycles and fixed using the fix command, and then the entire system was allowed to relax for 2ps under the nvt ensemble;

[0072] (4) Data output and analysis processing

[0073] Etching steady state: specific, such as Figure 3 As shown, after about 320 (10ML) ion bombardments, the system reaches a stable state, and then begins to count the macroscopic measurement parameters of the etching system, including the etching yield and product components of silicon.

[0074] Data output: Use dump and dump_modify commands to output the macro measurement parameters to be measured every 1000 steps;

[0075] Data analysis and processing: Use OVITO software to visualize the output files, observe whether the evolution of the etching system is reasonable, and analyze the properties of the silicon substrate surface modification layer, silicon etching yield, and product components, where the silicon etching yield is defined as the number of silicon atoms removed in a single ion bombardment;

[0076] Specifically, Figure 4 , Figure 5As shown, according to the data in the molecular dynamics simulation output file, the properties of the surface modification layer of the silicon substrate (the surface structure of the silicon substrate is destroyed by plasma bombardment, the Si-Si bond is destroyed, and Si-F bond or Si-Cl bond is formed) under different halogen (fluorine, chlorine) ion incident energies (50-150eV) in the steady state are obtained, including the depth of the modification layer (the distance between the deepest halogen atom in the substrate and the highest atom on the substrate surface) and the proportion of halogen atoms in the substrate;

[0077] like Figure 6 As shown, according to the data in the molecular dynamics simulation output file, the relationship between the yield of silicon etching by fluorine and chlorine plasma (the average number of silicon atoms sputtered by a single ion bombardment) and the incident ion energy (50-150eV) in the steady state is obtained;

[0078] like Figure 7 , Figure 8 As shown, based on the data in the molecular dynamics simulation output file, a relationship diagram of the composition (proportion of each product) of the products (only the products containing one Si are counted) of fluorine and chlorine plasma etching of silicon in the steady state as a function of the incident ion energy (50-150 eV) is obtained.

[0079] Example 2

[0080] Molecular dynamics simulation of silicon substrate during fluorine plasma reactive ion etching (RIE):

[0081] The method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention is specifically for molecular dynamics simulation of fluorine plasma reactive ion etching of silicon substrate, and comprises the following steps:

[0082] (1) Modeling of silicon or silicon dioxide substrate materials:

[0083] The initial model of the silicon substrate material was constructed using Materials Studio modeling software according to the diamond lattice structure. The lattice constant is The geometric parameters such as the number of atoms, surface area, number of stacking layers, shape structure, etc. are determined according to the simulation needs and saved as a data file that can be recognized by the LAMMPS software. Then, the read_data command is used to read it into LAMMPS, and the fix command is used to fix the bottom two layers of silicon atoms.

[0084] like Figure 2 As shown, in this embodiment, the initial model substrate is a substrate with a side length of The top area of ​​a cube Corresponding to the silicon (100) crystal plane, there are a total of 17 layers of silicon, with 32 silicon atoms in each layer (denoted as 1ML), for a total of 544 silicon atoms.

[0085] (2) LAMMPS software initialization settings:

[0086] LAMMPS simulation environment parameter settings: the unit of the simulation system is set to metal, the boundary conditions in the XY direction (horizontal direction) are set to periodic boundary conditions, the Z direction is set to fixed boundary conditions, the atom type is set to sphere, the neighbor list is selected as default, and the time step is set to 0.2fs.

[0087] Force field (potential function) setting: Use the pair_style command to set the Stillinger-Weber (SW) potential function according to the silicon and fluorine atom systems in the simulation system, and then use the pair_coeff command to set the corresponding SW potential function file and silicon and fluorine atom mapping;

[0088] Specifically, the SW potential function is as follows:

[0089]

[0090] Among them, Ф represents the total potential energy of the system, v2 and v3 are the two-body potential term and the three-body potential term respectively. The total potential energy is represented by the sum of the two-body potential term and the three-body potential term. The two-body potential term is of Lennard-Jones type, with an exponential cutoff function to limit its scope of action; the three-body potential term provides that when the configuration of three adjacent atoms deviates significantly from the diamond lattice, that is, when the three-body angle ≠ 109.5°, energy is lost rapidly.

[0091] Energy minimization: Use the min_style command to determine that the energy minimization algorithm is the conjugate gradient algorithm, and the minimize command sets the relevant parameters. By minimizing energy, the unreasonable structure of the model is eliminated;

[0092] Temperature initialization: The temperature of the system is calculated based on the velocity of the atoms, so temperature initialization is to initialize the velocity of the atoms. Use the velocity command to apply a random initial velocity to each atom of the substrate according to the Gaussian distribution. In this example, the initial temperature of the system is set to 300K.

[0093] Relaxation: Under the NVT ensemble, the system was relaxed for 5 ps with a damping time of 0.02 ps (100 times the time step) to obtain the equilibrium substrate structure.

[0094] Set up the ion generation region: Set up the plasma generation region at a height slightly greater than the potential function cutoff distance above the substrate surface. The substrate surface height is defined as the z coordinate of the highest atom in the system before each etching cycle begins.

[0095] Specifically, Figure 2 As shown, in this embodiment, the initial model substrate An ion generation region is set at a height of 1000 m, and in subsequent etching cycles, the substrate surface height and the ion generation region are updated before each ion injection cycle begins as the substrate surface height changes;

[0096] Grouping: Use the group command to group silicon substrates, plasma etching materials, fluorine free radicals, etching products, etc. for etching process identification;

[0097] Screen output settings: Use the thermo command to set the screen to print system information every 1000 steps, and the thermo_style command to set the output physical quantities, including system temperature, potential energy, kinetic energy, and number of atoms, so as to monitor system information in real time during the simulation;

[0098] (3) Plasma etching cycle:

[0099] Set the number of cycles: The plasma etching cycle consists of halogen ions bombarding the silicon surface cyclically. Before the etching cycle starts, use the loop command to set the number of cycles to 1000 (about 31ML);

[0100] Creating etching materials: In the fluorine plasma reactive ion etching process, creating etching materials is divided into two parts. The first part is to create fluorine free radicals, that is, use create_atom to randomly create 10-30 0.1-1.0eV fluorine atoms in the top surface area of ​​the substrate, and set the diameter to Density is 12.40826g / cm 3 , the incident angle is set to be perpendicular to the substrate surface downward (v z <0); the second part is to create high-energy etching fluoride ions, that is, use create_atom to create a fluoride ion in the etching material generation area, the horizontal position is generated by a random number (random command), and the diameter is set to Density is 12.40826g / cm 3 The ion energy was set to 50 eV and the incident angle was set to be perpendicular to the substrate surface downward (v z <0); Fig. 9 As shown, low-energy fluorine radicals are attached to the substrate surface. The newly generated high-energy fluoride ions are located at the bottom.

[0101] Charge neutrality assumption: Due to the electron transfer of surface atoms, it is assumed that the incident ions are charge neutralized before hitting the surface, so all atoms remain charge neutral in the simulation. Specifically, the potential function does not consider the effect of charge and the atom type is set to sphere, which does not include atomic charge information.

[0102] A single ion bombardment can be divided into three stages: first, except for the two fixed atoms at the bottom, all atoms evolve in the nve ensemble for 1 ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs. Then, the Berendsen thermostat algorithm is applied to the entire system to cool the temperature to the initial temperature of 300 K, with a heat bath damping of 0.02 ps (100 times the time step). Finally, all atoms that have no interaction with the substrate and have a velocity away from the surface (v z >0) are removed from the system and marked as etching products, and the remaining atoms form the input model for the next ion bombardment;

[0103] To prevent the silicon substrate from being broken down during the etching process, two silicon atomic layers were added to the bottom of the substrate after every 200 ion etching cycles and fixed using the fix command, and then the entire system was allowed to relax for 2ps under the nvt ensemble;

[0104] (4) Data output and analysis processing

[0105] Etching steady state: specific, such as Fig.10 As shown, after about 320 (10ML) ion bombardments, the system reaches a stable state, and then begins to count the macroscopic measurement parameters of the etching system, including the etching yield and product components of silicon.

[0106] Data output: Use dump and dump_modify commands to output the macro measurement parameters to be measured every 1000 steps;

[0107] Data analysis and processing: Use OVITO software to visualize the output files, observe whether the evolution of the etching system is reasonable, and analyze the properties of the silicon substrate surface modification layer, silicon etching yield, and product components, where the silicon etching yield is defined as the number of silicon atoms removed in a single ion bombardment;

[0108] Specifically, Fig.11 As shown, based on the data in the molecular dynamics simulation output file, the properties of the surface modification layer of the silicon substrate (the surface structure of the silicon substrate is destroyed by plasma bombardment, the Si-Si bond is destroyed, and the Si-F bond is formed) with the fluorine ion incident energy (50-150eV) in the steady state are obtained, including the depth of the modification layer (the distance between the deepest fluorine atom in the substrate and the highest atom on the substrate surface) and the proportion of fluorine atoms in the substrate;

[0109] like Fig.12 As shown, a relationship diagram of the yield of silicon in fluorine plasma reactive ion etching (the average number of silicon atoms sputtered by a single ion bombardment) under steady state conditions as a function of incident ion energy (50-150 eV) is obtained based on the data in the molecular dynamics simulation output file;

[0110] like Fig.13As shown, based on the data in the molecular dynamics simulation output file, a relationship diagram of the composition (proportion of each product) of the product of fluorine plasma reactive ion etching of silicon (only the products containing one Si are counted) under steady state changes with the incident ion energy (50-150eV) is obtained.

[0111] Example 3

[0112] Molecular dynamics simulation of halogen plasma etching of silicon dioxide substrate:

[0113] The method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention is applied to molecular dynamics simulation of halogen plasma etching of silicon dioxide substrate, comprising the following steps:

[0114] (1) Modeling of silicon or silicon dioxide substrate materials:

[0115] The initial model of the silicon dioxide substrate material was constructed using Materials Studio modeling software according to the silicon dioxide crystal structure. The lattice constant of the ideal β-quartz is The geometric parameters such as the number of atoms, surface area, number of stacking layers, shape structure, etc. are determined according to the simulation needs and saved as a data file that can be recognized by the LAMMPS software. Then, the read_data command is used to read it into LAMMPS, and the fix command is used to fix the bottom two layers of atoms.

[0116] like Figure 2 As shown, in this embodiment, the top surface of the initial model substrate is a The square with the top area Depth There are 12 monolayers, each with 32 silicon atoms and 64 oxygen atoms, for a total of 384 silicon atoms and 768 oxygen atoms. To facilitate the counting of the incident ion dose, we take every 32 incident ions as 1 ML.

[0117] (2) LAMMPS software initialization settings:

[0118] LAMMPS simulation environment parameter settings: the unit of the simulation system is set to metal, the boundary conditions in the XY direction (horizontal direction) are set to periodic boundary conditions, the Z direction is set to fixed boundary conditions, the atom type is set to sphere, the neighbor list is selected as default, and the time step is set to 0.2fs.

[0119] Force field (potential function) setting: Use the pair_style command to set the Stillinger-Weber (SW) potential function according to the silicon, fluorine or chlorine, and oxygen atom systems in the simulation system, and then use the pair_coeff command to set the corresponding SW potential function file and silicon, fluorine or chlorine, and oxygen atom mapping;

[0120] Specifically, the SW potential function is as follows:

[0121]

[0122] Among them, Ф represents the total potential energy of the system, v2 and v3 are the two-body potential term and the three-body potential term respectively. The total potential energy is represented by the sum of the two-body potential term and the three-body potential term. The two-body potential term is of Lennard-Jones type, with an exponential cutoff function to limit its scope of action; the three-body potential term provides that when the configuration of three adjacent atoms deviates significantly from the diamond lattice, that is, when the three-body angle ≠ 109.5°, energy is lost rapidly.

[0123] Energy minimization: Use the min_style command to determine that the energy minimization algorithm is the conjugate gradient algorithm, and the minimize command sets the relevant parameters. By minimizing energy, the unreasonable structure of the model is eliminated;

[0124] Temperature initialization: The temperature of the system is calculated based on the velocity of the atoms, so temperature initialization is to initialize the velocity of the atoms. Use the velocity command to apply a random initial velocity to each atom of the substrate according to the Gaussian distribution. In this example, the initial temperature of the system is set to 300K.

[0125] Relaxation: Under the NVT ensemble, the system was relaxed for 5 ps with a damping time of 0.02 ps (100 times the time step) to obtain the equilibrium substrate structure.

[0126] Set up the ion generation region: Set up the plasma generation region at a height slightly greater than the potential function cutoff distance above the substrate surface. The substrate surface height is defined as the z coordinate of the highest atom in the system before each etching cycle begins.

[0127] Specifically, Fig.14 As shown, in this embodiment, the initial model substrate An ion generation region is set at a height of 1000 m, and in subsequent etching cycles, the substrate surface height and the ion generation region are updated before each ion injection cycle begins as the substrate surface height changes;

[0128] Grouping: Use the group command to group silicon dioxide substrates, plasma etching materials, etching products, etc. for etching process identification;

[0129] Screen output settings: Use the thermo command to set the screen to print system information every 1000 steps, and the thermo_style command to set the output physical quantities, including system temperature, potential energy, kinetic energy, and number of atoms, so as to monitor system information in real time during the simulation;

[0130] (3) Plasma etching cycle:

[0131] Set the number of cycles: The plasma etching cycle consists of halogen ions bombarding the silicon dioxide surface in a cycle. Before the etching cycle starts, use the loop command to set the number of cycles to 1000 (about 31ML);

[0132] Create etching ions: First, use create_atom to create a fluorine ion in the etching material generation area. The horizontal position is generated by a random number (random command), and the diameter is set to Density is 12.40826g / cm 3 (If it is a chloride ion, the diameter is Density is 8.99186g / cm 3 ), the ion energy was set to 50 eV, and the incident angle was set to be perpendicular to the substrate surface downward (v z <0);

[0133] Charge neutrality assumption: Due to the electron transfer of surface atoms, it is assumed that the incident ions are charge neutralized before hitting the surface, so all atoms remain charge neutral in the simulation. Specifically, the potential function does not consider the effect of charge and the atom type is set to sphere, which does not include atomic charge information.

[0134] A single ion bombardment can be divided into three stages: first, except for the two fixed atoms at the bottom, all atoms are evolved in the nve ensemble for 1 ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs. Then, the Berendsen thermostat algorithm is applied to the entire system to cool the temperature to 300 K, with a heat bath damping of 0.02 ps (100 times the time step). Finally, all atoms that have no interaction with the substrate and have a velocity away from the surface (v z >0) are removed from the system and marked as etching products, and the remaining atoms form the input model for the next ion bombardment;

[0135] To prevent the silicon dioxide substrate from being broken down during the etching process, two silicon dioxide atomic layers were added to the bottom of the substrate after every 200 ion etching cycles and fixed using the fix command, and then the entire system was allowed to relax for 2 ps under the nvt ensemble;

[0136] (4) Data output and analysis processing

[0137] Etching steady state: specific, such as Fig.15 As shown, after about 320 (10ML) ion bombardments, the system reaches a stable state, and then begins to count the macroscopic measurement parameters of the etching system, including the etching yield of silicon.

[0138] Data output: Use dump and dump_modify commands to output the macro measurement parameters to be measured every 1000 steps;

[0139] Data analysis and processing: Use OVITO software to visualize the output files, observe whether the evolution of the etching system is reasonable, and analyze the properties of the surface modification layer of the silicon dioxide substrate and the silicon etching yield, where the silicon etching yield is defined as the average number of silicon atoms removed in a single ion bombardment;

[0140] Specifically, Fig.16 , Fig.17 As shown, according to the data in the molecular dynamics simulation output file, the properties of the surface modification layer of the silicon dioxide substrate (the surface structure of the silicon dioxide substrate is destroyed by plasma bombardment, and the Si-O bond is destroyed) under different halogen (fluorine, chlorine) ion incident energies (50-150eV) in the steady state are obtained, including the depth of the modification layer (the distance between the deepest halogen atom in the substrate and the highest atom on the substrate surface) and the proportion of halogen atoms in the substrate;

[0141] like Fig.18 As shown, based on the data in the molecular dynamics simulation output file, a relationship diagram of the silicon yield (the average number of silicon atoms sputtered by a single ion bombardment) of fluorine and chlorine plasma etching of silicon dioxide substrates under steady state changes with the incident ion energy (50-150eV) was obtained.

[0142] Therefore, the method of molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide of the present invention simulates the plasma-surface interaction mechanism at the atomic scale by the molecular dynamics method, analyzes the properties of the surface modification layer, thereby deepening the understanding of the mechanism of halogen plasma etching of silicon or silicon dioxide, accurately and quickly predicts the effect of halogen plasma etching of silicon or silicon dioxide under different incident conditions, and avoids the high cost, complexity and uncertainty brought by the experiment; and with the continuous development of computer CPU, GPU and other hardware, the efficiency of computer simulation calculation is continuously improved, which further greatly shortens the development process of halogen plasma etching of silicon or silicon dioxide process, and has great application prospects.

Claims

1. A method for molecular dynamics simulation of halogen plasma etching of silicon or silicon dioxide, characterized in that: The following steps are involved: (1) Modeling of silicon or silicon dioxide substrate materials: using modeling software to construct an initial structural model of silicon or silicon dioxide substrate materials according to the lattice structure of silicon or silicon dioxide; (2) LAMMPS software initialization settings: set the LAMMPS simulation environment parameters, force field, energy minimization, initial temperature, relaxation, plasma generation area and screen output respectively, and group the silicon or silicon dioxide substrate, plasma etching material and etching products for etching process identification; (3) Plasma etching cycle: set the number of etching cycles, create etching ions and set the radius and density according to the incident ion type, set the initial velocity according to the ion energy, set the incident angle, and set the atomic charge to maintain charge neutrality; perform a single ion bombardment cycle: first, except for the fixed atoms at the bottom, all atoms evolve under the nve ensemble for 0.5-2ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs, then the thermostat algorithm is applied to the entire system to cool the temperature to the initial temperature, and finally the substances not bound to the material are deleted from the system, and the remaining atoms form the input model for the next ion bombardment; (4) Data output and analysis: The macroscopic measurement parameters to be measured are output and visualized using OVITO software, and the properties of the surface modification layer of the silicon or silicon dioxide substrate, as well as the etching yield and product composition of silicon are analyzed.

2. The method according to claim 1, characterized in that: In step (1), the geometric parameters, including the number of atoms, surface area, number of stacking layers and shape structure, are set according to simulation requirements.

3. The method according to claim 1, characterized in that In step (2), the LAMMPS simulation environment parameters are set as follows: the unit of the simulation system is set to metal, the boundary conditions in the x and y directions are periodic boundary conditions, the z direction is fixed boundary conditions, the atom type is sphere, the neighbor list is defined as default, the time step is 0.2fs when the ion energy is less than 200eV, the time step is 0.1fs when the ion energy is greater than 200eV and less than 1000eV, and the time step is 0.05fs when the ion energy is greater than 1000eV and less than 2000eV; The setting of the force field is the setting of the potential function, which is specifically: setting the corresponding potential function according to the atom type in the simulation system, first using the pair_style command to set the force field type to the SW potential function, and then using the pair_coeff command to set the corresponding atom type to the potential function mapping.

4. The method according to claim 1, characterized in that: In step (2), the energy minimization is specifically set as follows: using the min_style command to determine that the energy minimization algorithm is the conjugate gradient algorithm, and using the minimize command to set relevant parameters; The setting of the initial temperature is specifically as follows: the temperature of the system is calculated based on the velocity of the atoms, and the velocity command is used to apply the initial velocity to the atoms of the material substrate according to the Gaussian distribution, and the initial temperature of the system is set to 300K; The relaxation setting is specifically as follows: under the nvt ensemble, the system is relaxed for 5-10ps at 300K to obtain an equilibrium substrate structure.

5. The method according to claim 1, characterized in that In step (2), the plasma generation region is specifically configured as follows: The plasma generation region is set at , and the substrate surface height is defined as the z coordinate of the highest atom in the system before each etching cycle starts; The screen output settings are as follows: use the thermo command to set the time step interval for printing system information on the screen, and use the thermo_style command to set the output physical quantities including system potential energy, kinetic energy, temperature, pressure, current time step number, and total number of system atoms; The grouping setting is specifically: using the group command to group the silicon or silicon dioxide substrate, plasma etching material, and etching products for etching process identification.

6. The method according to claim 1, characterized in that In step (3), the number of etching cycles is set. The molecular dynamics etching simulation consists of cyclic bombardment of the silicon or silicon dioxide surface by halogen ions. Before the etching cycle starts, the number of cycles is set using the loop command.

7. The method according to claim 1, characterized in that In step (3), create etching ions: first create an incident ion in the etching material generation area, and use the set command to set the radius and density according to the incident ion type, set the initial velocity according to the ion energy, and set the incident angle perpendicular to the substrate surface.

8. The method according to claim 1, characterized in that In step (3), a single ion bombardment cycle can be divided into three stages: first, except for the fixed atoms at the bottom, all atoms evolve in the nve ensemble for 0.5-2ps so that the kinetic energy of the incident ions is fully released to the substrate material and a subsequent collision cascade occurs. Then, the Langevin or Berendsen thermostat algorithm is applied to the entire system, and the temperature damping is set to 100 times the time step, and the temperature is cooled to the initial temperature of 300K for 0.5-2.0ps. Finally, all sputtered or desorbed substances are removed from the system as etching products using the delete_atoms command, and the remaining atoms form the input model for the next ion bombardment.

9. The method according to claim 1, characterized in that: In step (3), during the etching process, use the create_atoms command every 200 time steps to add new 4-8 layers of silicon or silicon dioxide atoms to the bottom of the substrate and use the fix command to set the forces in the x, y, and z directions to 0 to re-fix the bottom two layers of atoms.

10. The method according to claim 1, characterized in that In step (4), the dump and dump_modify commands are used to output the macroscopic measurement parameters to be measured, including atom type, velocity, coordinates, force, diameter and other calculated quantities, at regular time intervals. The silicon etching yield is the average number of silicon atoms removed in a single ion bombardment.

Images