Performance prediction method for hafnium oxide semiconductor material structure

Abstract

The invention provides a performance prediction method for a hafnium oxide semiconductor material structure, and relates to the technical field of semiconductor material research and development. The method comprises the following steps: performing global search in an Hf-O binary system by using crystal structure prediction software, determining the most stable two-dimensional Hf2O4 crystal structure, and further confirming the crystal configuration through structure optimization; calculating an Hf2O4 crystal structure by using density functional theory calculation software, and evaluating the dynamic, thermodynamic and mechanical stability of Hf2O4 by using a calculation result system; the electronic property of the Hf2O4 is obtained through further deep calculation, and the structural stability and semiconductor characteristics of the Hf2O4 are verified. According to the invention, through the first principle and the evolutionary algorithm calculation, the simulation calculation is used to replace the traditional trial-and-error experiment, and the novel semiconductor material with high reliability can be efficiently explored; a theoretical basis is provided for the design of a novel high-performance semiconductor material in advanced chip manufacturing, the research and development period can be expected to be remarkably shortened, and the cost can be reduced.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of semiconductor material research and development, and particularly relates to a hafnium oxide semiconductor material structure performance prediction method. BACKGROUND

[0002] With the semiconductor process technology entering the 3nm and below nodes, the traditional silicon dioxide (SiO2) and hafnium dioxide (HfO2) based gate dielectric faces significant challenges, including high-temperature annealing induced grain boundary defects, mechanical stress induced film cracking, and quantum tunneling effect induced leakage current increase, which seriously affect the performance and reliability of chips. In order to meet the needs of advanced processes for high dielectric constant, wide band gap and high stability, it is urgent to develop new hafnium oxide semiconductor materials. However, the traditional material research and development method relies on a large number of experiments for exploration and testing, and has the problems of long cycle, high cost and low efficiency. At the same time, HfO2 has monoclinic (Monoclinic, m), tetragonal (Tetragonal, t), cubic (Cubic, c), orthorhombic (Orthorhombic, o) and other crystal structures, and all possible crystal structures cannot be completely discovered and determined based on experiments. SUMMARY

[0003] The purpose of the application is to provide a hafnium oxide semiconductor material structure performance prediction method based on evolutionary algorithm and first principle theoretical calculation.

[0004] Technical scheme: A hafnium oxide semiconductor material structure performance prediction method, comprising the following steps:

[0005] S1, using crystal structure prediction software to search the global minimum energy structure of the hafnium-oxygen binary system, and obtaining the minimum energy structure as the most stable structure;

[0006] S2, using crystal structure modeling software to construct the most stable structure unit cell, and using density functional theory calculation software to optimize the structure of the most stable structure unit cell;

[0007] S3, using density functional theory calculation software, based on the most stable structure unit cell, performing simulation calculation, and based on the simulation calculation result of the most stable structure;

[0008] S4, based on the simulation calculation result of the most stable structure, analyzing to obtain the dynamic stability, thermodynamic stability, mechanical stability and electronic properties of the most stable structure.

[0009] Specifically, step S1 comprises: using crystal structure prediction software to search the hafnium-oxygen binary system, obtaining the relationship between the formation energy of each atom in the hafnium-oxygen binary system and the composition, and comparing to obtain the structure with the lowest formation energy as the most stable structure.

[0010] Specifically, the formation energy calculation formula is:

[0011] E f = [E(Hf x O y )- xE(hafnium) / 2-yE(oxygen) / 8] / (x+y)

[0012] In the formula, E f is the formation energy, E(Hf x O y ) is the total energy of Hf x O y , E(hafnium) is the total energy of P6mm-hafnium, E(oxygen) is the total energy of Pm3n-oxygen, x is the number of Hf atoms in Hf x O y , and y is the number of O atoms in Hf x O y .

[0013] Specifically, step S2 comprises: importing the most stable structure data into crystal structure modeling software, establishing a most stable structure unit cell, and optimizing the lattice constant of the most stable structure unit cell by using density functional theory calculation software.

[0014] Specifically, in step S4, the analysis process of dynamic stability comprises: establishing a most stable structure supercell based on the most stable structure unit cell, calculating the phonon spectrum of the most stable structure supercell using the density functional perturbation method, making a phonon spectrum diagram, and if there is no virtual frequency in the phonon spectrum diagram, the most stable structure has dynamic stability.

[0015] Specifically, in step S4, the analysis process of thermodynamic stability comprises: establishing a most stable structure supercell based on the most stable structure unit cell, performing ab initio molecular dynamics simulation on the most stable structure supercell, making a structure diagram of the ab initio molecular dynamics simulation and a total temperature fluctuation diagram and a total energy fluctuation diagram of the ab initio molecular dynamics simulation, and if the structure diagram has no distortion, the total temperature fluctuates around the target value in the total temperature fluctuation diagram, and the total energy oscillates in a narrow range in the total energy fluctuation diagram, then the most stable structure has thermodynamic stability.

[0016] Specifically, in step S4, the analysis process of mechanical stability comprises: based on the most stable structure unit cell, calculating the elastic constant of the most stable structure by using the strain energy method, judging the elastic constant by using the Born-Huang criterion, and then calculating the orientation dependence of Young's modulus and the orientation dependence of Poisson's ratio, and if the Young's modulus is greater than a set threshold value and the Poisson's ratio is less than a set threshold value, the most stable structure has mechanical stability.

[0017] Specifically, the orientation-dependent calculation formula of Young's modulus is as follows:

[0018]

[0019] In the formula, E(θ) is Young's modulus, θ is the included angle with the x-axis, C 11 , C 12 , C 22 and C 66 are elastic constants.

[0020] The orientation-dependent calculation formula of Poisson's ratio is as follows:

[0021]

[0022] In the formula, v(θ) is Poisson's ratio.

[0023] Specifically, in step S4, the analysis process of the electronic property includes: calculating the charge distribution between each atom in the most stable structure unit cell, analyzing the atomic bonding; calculating the electron localization function, analyzing the chemical bonding mode; performing energy band calculation to obtain the energy band structure, and analyzing the band gap; performing total state density and projected state density calculation, and analyzing the energy band formation mechanism and the spatial distribution rule of atomic orbital contribution.

[0024] Specifically, the most stable structure is a Hf2O4 two-dimensional crystal structure, which belongs to an orthorhombic system, and the lattice parameters are as follows: In the formula, c represents the vacuum layer.

[0025] Advantages: Compared with the prior art, the significant effect of the present application is that: the present application uses crystal structure prediction software to perform global search in the Hf-O binary system, determines the most stable Hf2O4 crystal structure, and further confirms the crystal configuration through structure optimization; the density functional theory calculation software is used to calculate the Hf2O4 crystal structure, and the dynamics, thermodynamics and mechanical stability of Hf2O4 are systematically evaluated; and further in-depth calculation obtains the electronic property, and verifies the structure stability and semiconductor characteristics of Hf2O4. The present application uses first principle and evolutionary algorithm calculation to replace the traditional trial and error experiment with "calculation simulation driving", which can efficiently explore new semiconductor materials with high reliability; the present application provides a theoretical basis for the design of high-performance and high-reliability semiconductor materials in advanced chip manufacturing, significantly shortens the research and development cycle and reduces the cost, and has important industrial application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a method flowchart of the present application.

[0027] Figure 2 (a) is a schematic diagram of the formation energy of each atom of HfxOy and the composition.

[0028] Figure 2 (b) is a schematic diagram of the structure of Hf2O4.

[0029] Figure 3 is a phonon spectrum of Hf2O4.

[0030] Figure 4 (a) is a schematic diagram of the structure of Hf2O4 at 300 K from AIMD simulation.

[0031] Figure 4 (b) is a schematic diagram of the structure of Hf2O4 at 1000 K from AIMD simulation.

[0032] Figure 5 (a) is a schematic diagram of the total temperature and total energy fluctuations at 300 K from AIMD simulation.

[0033] Figure 5 (b) is a schematic diagram of the total temperature and total energy fluctuations at 1000 K from AIMD simulation.

[0034] Figure 6 (a) is a schematic diagram of the orientation dependence of Young's modulus.

[0035] Figure 6 (b) is a schematic diagram of the orientation dependence of Poisson's ratio.

[0036] Figure 7 (a) is a schematic diagram of the charge transfer of Hf2O4.

[0037] Figure 7 (b) is a schematic diagram of the ELF of Hf2O4.

[0038] Figure 8 is a schematic diagram of the band structure, TDOS and PDOS of Hf2O4. DETAILED DESCRIPTION

[0039] One preferred embodiment of the present application will be further described below with reference to the drawings.

[0040] Example 1

[0041] Referring to Figure 1 The embodiment shown provides a hafnium oxide semiconductor material structure performance prediction method, including the following steps:

[0042] S1, using the crystal structure prediction software USPEX to search for the global minimum energy structure of the Hf-O binary system, obtaining the minimum energy structure as the most stable structure;

[0043] S2, using the crystal structure modeling software VESTA to construct the most stable structure unit cell, and using the density functional theory calculation software VASP to optimize the structure of the most stable structure unit cell;

[0044] S3, using the density functional theory calculation software VASP, based on the most stable structure unit cell, based on the most stable structure simulation calculation results;

[0045] S4, based on the most stable structure simulation calculation results, the dynamic stability, thermodynamic stability, mechanical stability and electronic properties of the most stable structure are analyzed.

[0046] In this embodiment, the software used includes USPEX, VESTA, VASP, VASPKIT and Origin, which are briefly described below.

[0047] USPEX, which stands for Universal Structure Predictor: Evolutionary Xtallography, is a crystal structure prediction method and program developed by Oganov Laboratory since 2004.

[0048] VESTA, which stands for Visualization for electronic and structural analysis, is a crystal structure visualization software that can assist first-principles calculations. It can be used for crystal structure modeling, viewing structure information, adjusting crystal structure parameters, etc.

[0049] VASP, which stands for Vienna Ab-initio Simulation Package, is a software package for electronic structure calculations and quantum mechanics-molecular dynamics simulations. It is based on density functional theory and uses the projected augmented wave (PAW) method for calculations. The input files include INCAR, POSCAR, KPOINTS, and POTCAR. INCAR is used to control the calculation of properties and how to calculate. POSCAR describes the crystal structure file, giving the material's basis vector, symmetry, and specific atomic coordinates. KPOINTS specifies the grid size and path of K space. POTCAR gives the pseudopotential of each element. The main output files of VASP calculation are CHGCAR (charge density output file), WAVECAR (wave function output file), OSZICAR (energy output file), XDATCAR (atomic motion trajectory file), and ELFCAR (electron localization function file). For output file processing, OSZICAR provides the energy corresponding to different structures. In the INCAR of static calculation, the parameter LORBIT = 11 is added. CHGCAR file can be directly plotted to get the charge density by VESTA software. XDATCAR can get the atomic motion trajectory, and ELF can be obtained by plotting ELF through VESTA software.

[0050] VASPKIT is a software used to adapt VASP, when the user has a POSCAR file, VASPKIT can automatically generate input files POTCAR, INCAR and KPOINTS; and can be used to analyze the input file of VASP, and perform graphical drawing.

[0051] Origin is a data analysis and drawing software, with statistical, peak analysis and curve fitting analysis functions, which can be used to draw two-dimensional and three-dimensional graphs.

[0052] The application of the above software in the method in a typical application scenario is described as follows:

[0053] S1, prediction of the most stable structure: using USPEX software, configuring input file INPUT.txt. Generating initial structure, calling VASP software for structure optimization, extracting the lowest energy structure, i.e. the most stable structure.

[0054] S2, establishment of the most stable structure unit cell: using modeling software VESTA to construct the most stable structure unit cell and optimizing the structure by VASP software, and converting the model data file to VASP software input file POSCAR by VESTA software.

[0055] S3, simulation calculation:

[0056] (a) optimizing the structure of the POSCAR obtained in step S2 by VASP software to generate stable structure data file CONTCAR, and renaming it as POSCAR, which is used as the input file for subsequent dynamic, thermodynamic and mechanical stability and electronic property calculation;

[0057] (b) configuring INCAR required for phonon spectrum calculation by VASP software, constructing mechanical file FORCE_CONSTRAINS, editing band.conf file to obtain band.yaml file, so as to obtain phonon spectrum data file band.dat;

[0058] (c) configuring INCAR required for AIMD simulation by VASP software, selecting ensemble to obtain output files CONTCAR, OSZICAR and XDATCAR;

[0059] (d) configuring INCAR required for elastic constant calculation by VASP software, selecting strain energy method to obtain elastic constants C 11 , C 22 , C 12 and C 66 ;

[0060] (e) The optimized structure of step (a) is further subjected to static calculation by VASP software to generate charge density CHGCAR;

[0061] (f) The required INCAR for ELF calculation is configured by VASP software to perform static calculation to obtain ELFCAR;

[0062] (g) The band calculation INCAR is configured by VASP software to perform band calculation to obtain output files EIGENVAL, OUTCAR and DOSCAR, and band data file band.dat is obtained by processing VASPKIT software;

[0063] (h) The state density calculation INCAR is configured by VASP software to obtain output files DOSCAR and OUTCAR, and TDOS.dat and PDOS_X_Atom.dat are obtained by processing VASPKIT software, wherein X in PDOS_X_Atom.dat is the number of selected Hf atoms and O atoms.

[0064] S4. Result processing and analysis:

[0065] (a) The formation energy of all structures in step S1 is analyzed, and the structure with the lowest energy is the most stable structure, and the structure diagram of the most stable structure is made using VESTA software;

[0066] (b) According to the band data file band.dat obtained in step S3(b), the phonon spectrum diagram is made using Origin software to judge the dynamic stability of the most stable structure;

[0067] (c) According to CONTCAR, OSZICAR and XDATCAR in step S3(c), combined with VESTA software and Origin software, the structure diagram after ab initio molecular dynamics AIMD simulation and the total temperature fluctuation diagram and total energy fluctuation diagram of AIMD simulation are made to judge the thermodynamic stability of the most stable structure;

[0068] (d) According to the elastic constants C 11 , C 22 , C 12 and C 66 in step S3(d), combined with Origin software, the Young's modulus and Poisson's ratio orientation dependence diagram is made to judge the mechanical stability of the most stable structure;

[0069] (e) According to the CHGCAR files obtained in two states in step S3(e), the specific charge transfer data is analyzed by processing VASPKIT software and combining VESTA software;

[0070] (f) According to the ELF CAR in step S3(f), the VESTA software is combined to make the electronic local function (ELF) diagram;

[0071] (g) According to the band data file band.dat obtained in step S3(g), the Origin software is combined to make the band structure diagram;

[0072] (h) According to the TDOS.dat and PDOS_X_Atom.dat obtained in S3(h), the Origin software is combined to make the TDOS diagram and the PDOS diagram.

[0073] The specific structure and performance prediction process of the hafnium oxide semiconductor material are described below, and in the present application, the hafnium oxide semiconductor material mentioned is a two-dimensional material.

[0074] Prediction of the most stable structure: using the USPEX software to perform a global search on the Hf-O binary system, the Hf x O y The relationship between the formation energy of each atom and the composition is shown in FIG. Figure 2 (a). From Figure 2 (a), it can be seen that the formation energy of the Hf2O4 two-dimensional crystal structure is lower than that of Hf2O2 and other structures, and it is the structure with the lowest formation energy, that is, the most stable structure. Figure 2 (b) is a structure diagram of Hf2O4, and it can be seen that the Hf / O ratio of Hf2O4 is 2:4, and it belongs to the orthorhombic system, and the lattice parameters are: In the present application, c represents a vacuum layer, and each unit cell contains 2 Hf atoms and 4 O atoms.

[0075] The calculation formula of the formation energy is:

[0076] E f =[E(Hf x O y )-xE(hafnium) / 2-yE(oxygen) / 8] / (x+y)

[0077] In the formula, E f is the formation energy, E(Hf x O y ) is the total energy of Hf x O y , E(hafnium) is the total energy of P6mm-hafnium, E(oxygen) is the total energy of Pm3n-oxygen, x is the number of Hf atoms in Hf x O y , and y is the number of O atoms in Hf x O y .

[0078] Structure modeling: Based on the most stable structure data obtained by USPEX, import VESTA software to establish Hf2O4 unit cell structure, and optimize the structure by VASP software, use VESTA software to convert the unit cell into the required POSCAR file, first use VASP software to optimize the lattice constant of the unit cell structure, and get the initial stable lattice constant. Structure optimization parameters: ISIF = 2, NSW = 200, the total energy and eigenvalue change in two consecutive iterations is less than 10 -6 eV, the atomic structure optimization stops, and the accurate structure file POSCAR is obtained, and a 5*6*1 supercell model is established, which contains 60 Hf atoms and 120 O atoms.

[0079] Dynamical stability calculation: Select 5*6*1 supercell model, use density functional perturbation method (DFPT) to calculate phonon spectrum. First, prepare the optimized structure POSCAR and POTCAR, use the command phonopy-d--dim='5 6 1' to expand the supercell to 5*6*1, then a group of files containing POSCAR-00* and SPOSCAR will be generated, rename the original unit cell POSCAR to POSCAR-unitcell; Copy SPOSCAR to POSCAR and put it in a new folder phonon, also put POSCAR-unitcell in this folder, then copy POTCAR to phonon. Modify INCAR parameters, IBRION = 8, POTIM = 1, NSW = 1, modify KPOINTS file, construct mechanical file FORCE_CONSTRAINS required for phonon spectrum, edit band.conf file, generate band.yaml file, finally get phonon spectrum data file band.dat. Finally, use Origin software to make phonon spectrum graph.

[0080] Thermodynamic stability calculation: Select 5*6*1 supercell model, modify INCAR parameters, for 300K, set TEBEG = 300, TEEND = 300, IBRION = 0, POTIM = 1, NSW = 10000. For 1000K, set TEBEG = 1000, TEEND = 1000, other parameters remain unchanged. After calculation, get CONTCAR, OSZICAR and XDATCAR files, use VESTA software to make structure graph after AIMD simulation, get energy and temperature fluctuation data by VASPKIT processing, import Origin software to make total temperature fluctuation graph and total energy fluctuation graph of AIMD simulation.

[0081] Mechanical stability calculation: select the optimized unit cell model, modify the INCAR parameters, wherein ISIF = 3, IBRION = 6, NFREE = 4, POTIM = 0.015, and after the calculation, view the elastic stiffness matrix in OUTCAR, and use VASPKIT software to convert the stiffness matrix into mechanical parameters such as Young's modulus and Poisson's ratio, to obtain the elastic constant C 11 , C 22 , C 12 and C 66 , make the orientation dependence graph of Young's modulus and the orientation dependence graph of Poisson's ratio.

[0082] The calculation formula of the orientation dependence of Young's modulus is:

[0083]

[0084] In the formula: E(θ) is Young's modulus, angle θ is the angle with the x-axis, C 11 , C 12 , C 22 and C 66 are elastic constants;

[0085] The calculation formula of the orientation dependence of Poisson's ratio is:

[0086]

[0087] In the formula: v(θ) is Poisson's ratio.

[0088] Self-consistent calculation: rename the unit cell CONTCAR file calculated in the structure optimization step as the POSCAR file, and perform static calculation through VASP software, set ISTART = 0, ICHARG = 2, ISPIN = 2, LORBIT = 11, LCHARG =.True., NSW = 0, IBRION =-1 in INCAR, modify the KPOINTS file, and finally output the CHGCAR file.

[0089] Charge transfer calculation: perform static calculation on the unit cell before structure optimization, also get CHGCAR file, then first drag the optimized unit cell CHGCAR file into VESTA, then subtract the CHGCAR file of the unit cell before optimization, and make a charge transfer graph.

[0090] ELF calculation: select the optimized unit cell structure, modify the INCAR parameters, wherein LELF =.TRUE., get ELFCAR, and make an ELF graph.

[0091] Band calculation: copy all files in the self-consistent calculation to the band folder, modify the INCAR parameters, where ISTART = 1, ICHGCAR = 11, NSW = 0. Change the KPOINTS file, get the EIGENVAL, OUTCAR and DOSCAR files, process them through the VASPKIT software to get the band data file band.dat, and use Origin software to make the band structure diagram.

[0092] Density of states calculation: copy all files in the self-consistent calculation to the dos folder, modify the INCAR parameters, where ISTART = 1, ICHGCAR = 11, NSW = 0. Change the KPOINTS file, get DOSCAR and OUTCAR, process them through the VASPKIT software to get TDOS.dat and PDOS_X_Atom.dat, X is the number of selected Hf atoms and O atoms, make the total density of states TDOS and the projected density of states PDOS diagram.

[0093] Through the above steps, a series of test result graphs of Hf2O4 are obtained, which are analyzed and explained as follows.

[0094] Please refer to Figure 3 It can be seen that Hf2O4 has 18 phonon branches (3 acoustic branches and 15 optical branches); there is no imaginary frequency in the entire Brillouin zone, which proves the dynamic stability of Hf2O4 material.

[0095] Please refer to Figure 4 (a) and Figure 4 (b), in this embodiment, 10 ps (time step 1 fs) AIMD simulation of Hf2O4 was carried out at 300 K and 1000 K, respectively. The long simulation time and large supercell (in each direction ) ensure the reliability of the results. Hf2O4 maintains the structure at 300 K without obvious distortion, and still maintains the structure at 1000 K without obvious distortion, indicating that Hf2O4 has high thermodynamic stability.

[0096] Please refer to Figure 5 (a) and Figure 5 (b), the green line is the total temperature fluctuation curve, and the orange line is the total energy fluctuation curve. It can be seen that Hf2O4 reaches thermal equilibrium in a short time at 300 K and 1000 K. At 300 K, the total temperature fluctuates around the target value, and the total energy oscillates in a narrow range. At 1000 K, the total temperature also fluctuates around the target value, and the total energy still shows oscillation in a narrow range. This indicates that Hf2O4 is in a stable state during the collection stage, which proves its high thermodynamic stability.

[0097] The elastic constants of Hf2O4 are calculated as follows: C 11 = 150.15 N / m, C 22 = 225.23 N / m, C 12 = 38.13 N / m, C 66 = 46.90 N / m. It is found that the mechanical stability of Hf2O4 satisfies the Born-Huang criterion: C 11 C 22 -C 12 2 > 0 and C 66 > 0. The Young's modulus and Poisson's ratio of Hf2O4 are calculated using the elastic constants, which are shown in Figure 6 (a) and Figure 6 (b), respectively, and it is proved that Hf2O4 has mechanical anisotropy. Furthermore, it is found that E x = 143.7 N / m, E y = 215.53 N / m (the Young's modulus in x and y directions), v x = 0.17 and v y = 0.25 (the Poisson's ratio in x and y directions). The combination of high Young's modulus and low Poisson's ratio represents high stiffness, which embodies the high mechanical stability of Hf2O4.

[0098] Please refer to Figure 7 (a) and Figure 7 (b) (isosurface: 0.70 a.u.). It is found that the charge lost by a single Hf atom is 2.29 e in Figure 7 (a), and the charges gained by the O1 and O2 atoms around the Hf atom are 1.22 e and 1.07 e, respectively. This indicates that there is significant charge transfer in the direction of Hf→O in Hf2O4, resulting in the Hf atom being positively charged and the O atom being negatively charged. The charge difference between the two oxygen atoms reveals their non-equivalent chemical environments. From Figure 7 (b), it can be found that the electrons around the O atom are concentrated and highly localized, while the Hf-O and Hf atoms are highly delocalized, which confirms the dominant interaction mode of ionic bonding and again confirms the high structural stability of Hf2O4.

[0099] Please refer to Figure 8 , it is found that the Hf-O and Hf atoms are highly delocalized, which confirms the dominant interaction mode of ionic bonding and again confirms the high structural stability of Hf2O4. Figure 8As can be seen from the figure, Hf2O4 has a large band gap, which indicates that it has insulating properties and can effectively suppress leakage current, meeting the core requirements of high-k dielectric materials. According to the TDOS and PDOS results, it can be seen that Hf atoms and O atoms contribute to the entire energy range. However, the situation is different, and the heavier Hf atoms mainly act in the conduction band, while the lighter O atoms mainly act in the valence band. This is due to the strong ionicity of Hf2O4 caused by the large electronegativity. Charge transfer leads to the formation of valence bands by filled, low-energy O atom 2p orbitals, while empty, high-energy Hf atom 5d orbitals form conduction bands.

[0100] The present application innovatively combines evolutionary algorithms with first-principle calculations, uses USPEX software to perform global structure search on the Hf-O binary system, efficiently locks the most stable orthorhombic phase Hf2O4 (with the lowest formation energy), and constructs a multi-scale calculation framework: verifies the dynamic stability (no imaginary frequency) through phonon spectrum, confirms that the structure has no distortion at high temperatures of 300K / 1000K (thermodynamic stability) through AIMD simulation, meets the Born-Huang criterion and reveals mechanical anisotropy through elastic constant calculation, reveals mechanical stability through high Young's modulus and low Poisson's ratio. Electronic property prediction shows that it has wide-bandgap semiconductor characteristics and strong ionic bond dominated charge distribution, which can effectively suppress quantum tunneling leakage current. This method replaces the traditional trial-and-error experiment with "computational simulation driven" to provide theoretical support for the design of advanced semiconductor materials, effectively shortening the research and development cycle and cost.

Claims

1. A method of performance prediction of a hafnium oxide semiconductor material structure, characterized by, The method comprises the following steps: S1, using crystal structure prediction software to search the global minimum energy structure of the hafnium-oxygen binary system, and obtaining the minimum energy structure as the most stable structure; S2, using crystal structure modeling software to construct the most stable structure unit cell, and using density functional theory calculation software to optimize the structure of the most stable structure unit cell; S3, using density functional theory calculation software, based on the most stable structure unit cell, performing simulation calculation, and obtaining the most stable structure simulation calculation result based on the simulation calculation; S4, based on the most stable structure simulation calculation result, analyzing the dynamic stability, thermodynamic stability, mechanical stability and electronic properties of the most stable structure.

2. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 1, characterized in that The step S1 comprises: using crystal structure prediction software to search the hafnium-oxygen binary system, obtaining the formation energy of each atom in the hafnium-oxygen binary system and the relationship with the composition, and comparing to obtain the structure with the lowest formation energy as the most stable structure.

3. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 2, characterized in that: The formula for calculating the formation energy is: E f = [E(Hf x O y )- xE(hafnium) / 2 - yE(oxygen) / 8] / (x+y) wherein: E f is the total energy of P6mm-hafnium, E(oxygen) is the total energy of Pm3n-oxygen, x is the number of Hf atoms in Hf x O y , and y is the number of O atoms in Hf x O y . x is the total energy of P6mm-hafnium, E(oxygen) is the total energy of Pm3n-oxygen, x is the number of Hf atoms in Hf y O x , and y is the number of O atoms in Hf y O .

4. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 1, characterized in that The step S2 comprises: importing the most stable structure data into the crystal structure modeling software, establishing the most stable structure unit cell, and optimizing the lattice constant of the most stable structure unit cell by using the density functional theory calculation software.

5. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 1, characterized in that In step S4, the analysis process of the dynamic stability comprises: establishing a most stable structure supercell based on the most stable structure unit cell, calculating the phonon spectrum of the most stable structure supercell by using the density functional perturbation method, making a phonon spectrum diagram, and if there is no virtual frequency in the phonon spectrum diagram, the most stable structure has dynamic stability.

6. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 1, characterized in that In step S4, the analysis process of the thermodynamic stability comprises: establishing a most stable structure supercell based on the most stable structure unit cell, performing ab initio molecular dynamics simulation on the most stable structure supercell, making a structure diagram of the ab initio molecular dynamics simulation and a total temperature fluctuation diagram and a total energy fluctuation diagram of the ab initio molecular dynamics simulation, and if the structure diagram has no distortion, the total temperature fluctuates around the target value in the total temperature fluctuation diagram, and the total energy oscillates in a narrow range in the total energy fluctuation diagram, the most stable structure has thermodynamic stability.

7. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 1, characterized in that In step S4, the analysis process of the mechanical stability comprises: based on the most stable structure unit cell, calculating the elastic constant of the most stable structure by using the strain energy method, judging the elastic constant by using the Born-Huang criterion, and then calculating the orientation dependence of Young's modulus and the orientation dependence of Poisson's ratio, and if the Young's modulus is greater than a set threshold value and the Poisson's ratio is less than a set threshold value, the most stable structure has mechanical stability.

8. The method of performance prediction of hafnium oxide semiconductor material structures according to claim 7, characterized in that The formula for calculating the orientation dependence of Young's modulus is: where E(0) is the Young's modulus, 0 is the angle with the x-axis, C 11 , C 12 , C 22 and C 66 are elastic constants; The formula for calculating the orientation dependence of Poisson's ratio is: In the formula, v(θ) is the Poisson's ratio.

9. The method of performance prediction of hafnium oxide semiconductor material structures of claim 1, wherein, In step S4, the analysis process of the electronic properties comprises: calculating the charge distribution between atoms in the most stable structure unit cell, analyzing the atomic bonding; calculating the electron localization function, analyzing the chemical bonding mode; performing energy band calculation to obtain the energy band structure and analyze the band gap; performing total state density and projected state density calculation to analyze the energy band formation mechanism and the spatial distribution rule of atomic orbital contribution.

10. The method of performance prediction of hafnium oxide semiconductor material structures of claim 1, wherein, In step S1, the most stable structure is a Hf2O4 two-dimensional crystal structure, which belongs to an orthorhombic system, and the lattice parameters are: In the formula, c represents a vacuum layer.

Images