CN106503374A - The structure of the compound semiconductor device of labyrinth describes method - Google Patents

Abstract

The invention discloses a method for describing the structure of a compound semiconductor device with a complex structure, including: analyzing the geometric features, physical features, and numerical feature parameters of the device structure to generate five typical data types; determining the member assignment syntax rules corresponding to the typical data types; The semiconductor device structure is decomposed into five typical data types and written into a structure description file; a dynamic pointer oriented to the device structure description file is generated, and the structure description file is read according to the description grammar rules; the storage arrangement and data format of the numerical calculation-specific file are determined; The data and the device structure stored by the dynamic pointer are converted into a numerical calculation-specific file expressed by a fixed array of model parameters, positions, and dimensions. The invention makes it possible and convenient to input the structure of compound semiconductor device with complex structure including multiple internal heterojunction interfaces, multiple internal plane doping, multiple quantum confinement regions and multiple nonlocal quantum tunneling regions.

Description

Structural Description Method for Compound Semiconductor Devices with Complicated Structures

technical field

The invention relates to a semiconductor device preparation technology, in particular to a method for describing the structure of a compound semiconductor device with a complex structure.

Background technique

At present, the structure of compound semiconductor devices is becoming more and more complex, such as multi-junction solar cells, high electron mobility transistors, infrared detectors, tunneling resonance diodes, etc., usually contain multiple heterojunctions, and in some compound semiconductor materials Embedded non-local quantum tunneling diodes, or surface doping, or quantum wells, or internal interfaces, etc., which need to be dealt with separately, the number and position of these physical properties before the simulation program is run, and the physical parameters are Unknown, and often there are many parameters that need to be clarified for these physical properties. Taking the non-localized quantum tunneling diode as an example, the parameters to clarify its physical properties include the tunneling direction, the position of the left/right tunneling layer, and the left/right electrons. And hole effective mass, left/right degeneracy parameters, etc. In addition, the more intuitive reflection is that the number of defects in different compound semiconductor material layers is different. In terms of data structure, dynamic pointers are usually used to describe this kind of objects with uncertain numbers and physical characteristics in advance, but a large number of traversal object pointers in numerical analysis software will reduce the calculation speed and also bring about memory utilization. Another direct method is to generate a fixed-length array in advance, and determine whether each physical object exists in turn during calculation. If it exists, traverse the corresponding number of object arrays. This not only limits the number of physical objects, but also The calculation time is wasted on the judgment of this fixed array, and the simulation software that uses this method to input device structure parameters is mostly some with the feature of visual graphical interface, such as wxAMPS and so on. In addition, the input of physical model parameters currently used either requires a fixed format, such as Silvaco, Crosslight, etc., and does not allow changes in parameter data types, arbitrary input positions of structural units, and input of formulaic parameters, which greatly limits the physical model. Flexible input of object parameters, extensibility of device structure files, and optional types.

Contents of the invention

The invention provides a method for describing the structure of a compound semiconductor device with a complex structure, so that the input contains multiple internal heterojunction interfaces, multiple internal plane doping, multiple quantum confinement regions, and multiple nonlocal quantum tunneling regions. The structure of compound semiconductor devices with complex structures becomes possible and convenient, and the maintainability of the structure file reading program is greatly improved.

In order to achieve the above object, the present invention provides a method for describing the structure of a compound semiconductor device with a complex structure, which is characterized in that the method includes:

S1. Generate five typical data types according to the geometric characteristics, physical characteristics, and numerical characteristic parameters of the device structure required for the numerical analysis of compound semiconductor devices;

S2. Determine the member assignment syntax rules corresponding to each typical data type;

S3. Determine the composition rules of the structure description file, decompose the structure of the compound semiconductor device into five typical data types and write the structure description file;

S4. Generate a dynamic pointer oriented to the device structure description file, and read the structure description file according to the description syntax rules;

S5. According to the high-efficiency requirements of the numerical calculation process, determine the storage arrangement and data format of the numerical calculation-specific files;

S6. According to the requirements of S5, convert the device structure stored in the experimental data and dynamic pointer into a numerical calculation exclusive file expressed in a fixed array of model parameters, positions and dimensions, for use in numerical calculation.

In the above S1, the five typical data types mentioned include:

single fixed-value data type;

Geometric features and basic physical object data types involved in partial differential equations that dominate the physical properties of carriers;

Composite data types that contain physical phenomena of more than one primitive data type;

An associative data type that describes the spatial correlation between different physical phenomena;

A control data type describing the operating conditions of the compound semiconductor device being simulated.

In the above S2, the typical data type member assignment syntax rules include:

A single fixed-value data type assignment in a primitive data type takes the form of [type]name:value;

The member assignment of the basic physical object data type in the basic data type takes the form of [type] name (can be default): member 1 = , member 2 = , ...;;

A member assignment of a compound data type takes the form {type} name: {[member1] name: ...; [member2] name: ...; ...};

Member assignments of associated data types take the form [type]name: member1 = , member2 = , ...;;

A member assignment of a control data type takes the form {type} name: {[member1] name: ...; [member2] name: ...; ...};

Member assignments of generic data types take the form {type} name: {[member1] name: ...; [member2] name: ...; ...}.

In the above S3, the compound semiconductor device structure is decomposed into five typical data types in S1 and written into a structure description file.

In the above S4, a dynamic pointer oriented to the device structure description file is generated, and the structure description file is read according to the syntax rules described in S2.

In the above S5, according to the high-efficiency requirements of the numerical calculation process, the position, arrangement and format of various data in the numerical calculation exclusive file are determined.

In the above S6, the device structure stored by the experimental data and dynamic pointers is converted into a numerical calculation exclusive file expressed by an array of model parameters, fixed positions and fixed dimensions according to S5, for use by the numerical calculation program.

The above-mentioned structure description method is applied in the numerical simulation and analysis of the compound semiconductor device structure; the compound semiconductor device structure includes multi-junction solar cells, microwave devices, photodetectors, and lasers.

Compared with the prior art, the method for describing the structure of a compound semiconductor device with a complex structure in the present invention has the advantage that the present invention enables the input to contain multiple internal heterojunction interfaces, multiple internal surface doping, and multiple quantum confinement regions 1. The structure of compound semiconductor devices with complex structures in multiple non-localized quantum tunneling regions becomes possible and convenient, and at the same time, the processing method of the present invention greatly improves the maintainability of the structure file reading program;

The invention can be applied in the numerical simulation and analysis of various compound semiconductor device structures, such as multi-junction solar cells, microwave devices, photodetectors, lasers and the like.

Description of drawings

Fig. 1 is the flow chart of the structure description method of the compound semiconductor device of complex structure of the present invention;

Fig. 2 is five kinds of data type schematic diagrams that the present invention describes device structure;

Fig. 3 is the schematic diagram of five kinds of data types of description device structure that the present invention reads;

Fig. 4 is a schematic diagram of different characteristic data description areas in the numerical calculation exclusive file of the present invention;

5 is a schematic diagram of Embodiment 1 of the method for describing the structure of a compound semiconductor device with a complex structure in the present invention;

FIG. 6 is a schematic diagram of Embodiment 2 of the method for describing the structure of a compound semiconductor device with a complex structure according to the present invention.

detailed description

Specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

As shown in FIG. 1, it is an embodiment of a method for describing the structure of a compound semiconductor device with a complex structure in the present invention. The method specifically includes the following steps:

S1. Five typical data types are generated according to the parameter representations of the geometric features, physical features, and numerical features of the device structure required for the numerical analysis of compound semiconductor devices.

The geometric features describing the compound semiconductor device structure include the data type of the device area (area) and its grid division (grid) data type. The device area includes characteristic parameters such as device geometric size and spatial distribution, and the grid division includes the grid type on a certain area. , grid size, grid number, grid distribution and other characteristic parameters;

The physical characteristics describing the structure of compound semiconductor devices include the physical objects and device working conditions involved in the partial differential equations that dominate the physical characteristics of the carriers.

The physical objects involved in the system of partial differential equations governing the physical properties of carriers include:

1. Material energy band parameters reflecting the electronic characteristics of intrinsic compound semiconductor materials; for example: compound semiconductor electron affinity reflecting energy band arrangement, band edge state density of conduction band and valence band, static dielectric constant, energy band Gap, spontaneous emission recombination coefficient, Auger recombination coefficient, avalanche coefficient, etc.

2. Parameters that reflect the spatial distribution of electron/hole mobility in materials; for example, the mobility of electron/hole may be fixed, or may present a certain functional distribution in space with the distribution of doping concentration.

3. Discrete energy-level defects and continuous-distributed energy-level defects in bulk materials; for example: doping that belongs to single-level defects (SLD) to adjust the chemical potential of materials and deep-level and surface doping as recombination centers (delta), band-tailed state (BT), Gaussian state (GS), etc. Some single-level defects (SLD) and surface doping (delta) in these physical phenomena often have spatially non-uniform distributions, such as In MOCVD (metal organic chemical vapor deposition) growth, the concentration of dopant atoms is often set in the form of a+b(1-x)n or c+dxm, and the distribution of diffuse atoms often obeys the residual error function distribution, Gaussian distribution, etc.

4. Discrete energy level defects and continuous energy level defects in surfaces and interfaces; for example: outer surface (Osurf) of device structure, heterojunction interface (Isurf), etc.

5. Additional properties across different material layers; for example: quantum confinement (QC) in bulk materials, nonlocal quantum tunneling (IBQT) between different energy bands of bulk materials, nonlocal quantum tunneling across interfaces (HFQT) Wait.

6. The physical model connecting the optical characteristics of the device structure and the material properties; for example: optical generation rate (OPG), optical reflection and transmittance (OR), etc.

The actual working conditions of compound semiconductor devices include:

1. Temperature conditions

For example: operating temperature range, low temperature and high temperature point, temperature step and other control parameters

2. Bias conditions

For example: applied bias voltage range, bias voltage start and end points, bias voltage application direction, bias voltage application step size and other control parameters

3. Lighting conditions

For example: light wavelength range, starting wavelength, light intensity distribution with wavelength and other control parameters

The numerical characteristics describing the structure of compound semiconductor devices include solution methods and solution accuracy.

The numerical solution method involves control parameters such as direct algorithm or iterative algorithm for linear equations obtained from discrete differential equations, coupling solution or sequential solution between different differential equations, and storage method for linear equations. The solution accuracy involves control parameters such as maximum number of iterations, parameters for judging convergence of iterations, solution error, and solution error information.

As shown in Figure 2, according to the three parameter characteristics of the device structure, it is decomposed into five typical data types, including basic data types, compound data types, associated data types, control data types, and general data types:

The basic data types include both single fixed values and physical objects involved in partial differential equations governing the physical properties of the carriers.

1. Single fixed value data type

It consists of four fixed-value data such as integer, real, boolean, and string, and is used to describe numbers, values, logic, names, etc.

2. Geometric features and basic physical object data types involved in partial differential equations that dominate the physical characteristics of carriers

By device area (area), grid division (grid), single parameter function distribution (fun), discrete energy level defect electronics parameters (dbsld), discrete energy level defect composite parameters (dbsrh), continuous energy level defect electronics parameters (cbsld), continuous level defect composite parameter (cbsrh), interface discrete level defect electronics parameter (dssld), interface discrete level defect composite parameter (dssrh), interface continuous level defect electronics parameter (cssld), interface Continuum level defect composite parameters (cssrh), band tail state (BT), Gaussian state (GS), optical layer property parameters (opl), etc., are used to describe the basic physical phenomena in partial differential equations.

The device area (area) describes the spatial composition and geometric dimensions of the device structure, including members such as device geometric dimensions and spatial distribution,

Grid division (grid) describes the spatial grid generation for a certain sub-area, including members such as grid type, grid size, grid number, and grid distribution on a certain area;

Function distribution (fun) describes the distribution of polynomial, exponential function, Gaussian function, residual error function, Lorenz function and other distributions composed of x, 1-x, including two group members such as function type and function expression. The number of parameters n, the coefficient array P[n], and the index array I[n] are four members.

Discrete-level defect electronics parameters (dbsld) describe the physical characteristics of defects obtained through instrumental testing, including six members including defect concentration, energy level position, electron/hole capture interface, and electron/hole degeneracy.

Discrete-level defect complex model parameters (dbsrh) describe the Schokley-Read-Hall model of defects, including defect concentration, electron/hole intrinsic lifetime, and equivalent concentration of defects to conduction band/valence band. The discrete energy level defect compound model parameters are obtained by converting the material discrete energy level defect electronic parameters through physical relations, and are the data finally provided to the device structure file.

Continuous level defect electronics parameter (cbsld) describes the physical characteristics of defects obtained through instrument testing, including five members including defect concentration energy spatial distribution function, electron/hole capture cross section, electron/hole degeneracy, among which defect concentration The energy spatial distribution function is expressed in the form of a parameter function distribution (fun).

Continuous level defect recombination parameter (cbsrh) describes the Schokley-Read-Hall model of defects, including five groups of defect concentration energy spatial distribution function, electron/hole intrinsic lifetime, defect equivalent concentration to conduction band/valence band Members, where the defect concentration energy spatial distribution function is expressed in the form of a parameter function distribution (fun). The continuous energy level defect composite model parameters are obtained by converting the material continuous energy level defect electronic parameters through physical relations, and are the data finally provided to the device structure file.

Interface Discrete Level Defect Electronics Parameters (dssld) describe the physical characteristics of interface defects obtained through instrumental testing, including six members including defect concentration, energy level position, electron/hole capture interface, and electron/hole degeneracy.

Interface discrete energy level defect composite parameter (dssrh) describes the Schokley-Read-Hall model of interface defects, including five members including defect surface density, electron/hole intrinsic lifetime, and equivalent concentration of defects to conduction band/valence band . The interface discrete energy level defect compound model parameters are obtained by converting the interface discrete energy level defect electronic parameters through physical relations, and are the data finally provided to the device structure file.

Interface continuous level defect electronics parameters (cssld) describe the physical characteristics of interface defects obtained through instrumental testing, including five members including defect density energy spatial distribution function, electron/hole capture cross section, and electron/hole degeneracy, among which The energy spatial distribution function of the defect concentration on the street surface is expressed in the form of a parameter function distribution (fun).

The interface continuous level defect recombination parameter (cssrh) describes the Schokley-Read-Hall model of the interface defect, including the interface defect concentration energy spatial distribution function, the electron/hole intrinsic recombination rate, and the equivalent concentration of the defect to the conduction band/valence band and other five team members, in which the energy spatial distribution function of the interface defect concentration is expressed in the form of a parameter function distribution (fun). The continuous energy level defect composite model parameters are obtained by converting the material continuous energy level defect electronic parameters through physical relations, and are the data finally provided to the device structure file.

Band tail state (BT) describes the physical characteristics of conduction band/valence band tailing defect states, including six members including band tail state density, exponential distribution decay factor, electron/hole capture cross section, and electron/hole degeneracy. After the tail state parameters are transformed through the physical relationship, the four members finally provided to the device structure file are the band tail state density, the exponential distribution decay factor, and the electron/hole intrinsic lifetime.

Gaussian state (GS) describes the defect state with a Gaussian distribution of state density in the material band gap, including seven members such as concentration, broadening factor, energy level position, electron/hole capture cross section, and electron/hole degeneracy. The Gaussian state parameters are finally provided to the device structure file after the physical relationship conversion, including five members: concentration, broadening factor, energy level position, and electron/hole intrinsic lifetime.

Optical layer attribute parameters describe the optical properties of the material layer, including three members: material name, material thickness, optical parameter model and optical parameter file name.

3. The compound data type mentioned refers to those physical phenomena that contain multiple basic data types. The compound device structure mainly involves the following types:

Material internal defect (BSLD), interface/surface defect (SSLD): including defect name (single fixed value data type), attribute logic feature (single fixed value data type), coordinate space distribution (function distribution), electronic parameters, Composite model parameters and several other basic data types.

The material growth layer (gl) describes the compound data type in which the material components are fixed and other conditions change during the material growth process: including basic and compound data types such as mobility function, material internal defects, and interface/surface defects (composite data types).

The heterogeneous material layer (ml) describes data types with different material compositions or components, and may contain several material growth layers: basic and composite data types including material energy band parameters (single fixed value data type), material growth layers, etc.

4. The associated data type describes the positional relationship between the heterogeneous material layer and the heterojunction interface/surface (gl_to_surface), the positional relationship between the optical layer and the heterogeneous material layer (opl_to_hl), and the nonlocality of the heterojunction interface Five associated data types including quantum tunneling (HFQT), interband nonlocal quantum tunneling (IBQT), and quantum confinement (QC).

Among them, the positional relationship between the heterogeneous material layer and the heterojunction interface/surface includes two members: the name of the material itself and the name of the heterojunction interface/surface, and the positional relationship between the optical layer and the heterogeneous material layer includes the name of the optical layer and the name of the heterojunction interface/surface. The name of the heterogeneous material layer has two members, and the nonlocal quantum tunneling at the heterojunction interface includes five members including the arrangement of the energy bands of the heterojunction, the names of the heterogeneous material layers on both sides involved in the tunneling, and the name of the growth layer. Nonlocal quantum tunneling between bands includes nine members including the type, the name of the material growth layer on both sides involved in the tunneling, the effective mass of electrons/holes on both sides, and the degeneracy of both sides, and the quantum confinement includes the type and energy band of the heterojunction Arrangement, name of heterogeneous material layer on both sides involved in tunneling, name of growth layer, name of extended heterogeneous material layer on both sides of wave function, name of growth layer and other ten members

5. The control data type describes the working conditions of the simulated and analyzed compound semiconductor device, including five control data types such as applied bias voltage, applied light, transient characteristics, ambient temperature, and numerical accuracy.

Among them, the applied bias voltage includes four members such as bias direction, start/stop bias voltage, and bias voltage step size. The external pressure light includes three members such as start/stop wavelength and light intensity function. The transient characteristics include start/stop There are three group members including time and time step distribution function, and the ambient temperature includes three group members including start/stop temperature and temperature step distribution function. Through the combination of the above control data types, many actual test conditions can be generated, such as the combination of external bias voltage and external light can generate light I/V curves, the combination of external bias voltage and transient characteristics can simulate time-resolved characteristics, external voltage bias Combined with transient and temperature, it can simulate the characteristics of deep level spectrum.

The general data type describes the general physical space and typical characteristics of the device, including members such as time scale, space scale, and device structure file description.

S2. Determine the member assignment syntax rules corresponding to each typical data type.

S2.1. The single fixed-value data type assignment in the basic data type takes the form of [type] name: value

For example, the expression that defines the concentration of doping D1 as 1E18cm-3 is:

[real] D1: 1.0d+18;

S2.2. The member assignment of the basic physical object data type in the basic data type takes the form of [type] name (can be default): member 1 = , member 2 = , ...;.

For example, define a deep level point defect D2 at 0.8eV on the valence band, with a concentration of 1E16cm-3, an electron/hole capture cross section of 1E-15cm2, and an electron/hole degeneracy of 2 as:

[dbsld] D1: edc = 0.8, den = 1.0d+16; sgn = 1e-15, sgp = 1e-15, g = 2.0;

For example, define an exponential function as:

[fun] f: type = 2; p = 50 + e^(10.*x^5) – e^(7.16*(1-x)^3);

S2.3. The member assignment of a compound data type takes the form of {type} name: {[member 1] name: ...; [member 2] name: ...; ...}

For example, the point defect deep energy level D1 is defined as

{bsld}D1:

{

[logical] ionized : false;

[fun] spatial distribution: ...;

[dbsld] energy level: ...;

[dbsrh] composite: ...;

…

}

Since the composite data type is consistent with the basic data type, the assignment method of each member is the same as in 2.

S2.4. The member assignment of the associated data type takes the form of [type] name (can be default): member 1 = , member 2 =, ...;

Define the left and right associated interfaces of the 1D material layer GaInP as:

[gl_to_surface] GaInP: left_surface_type = , left_surface_name = , right_surface_type = , right_surface_name = ;

Define the interband nonlocal quantum tunneling that occurs between the material growth layers d1 and d2 as:

[IBQT] QT1: type = , left_gl_name = , right_gl_name = , mel = , mer = , mhl =, mhr = , lg = , rg = ;

S2.5. The member assignment of the control data type takes the form of {type} name: {[member 1] name: ...; [member 2] name: ...; ...};

For example, define an applied bias voltage as:

{bias}B1 :

{

[logical]direction:;

[real] Vstart:;

[real] Vstop:;

[fun] Vstep: type = , P = ;

}

S2.6. The member assignment of a general data type takes the form of {type} name: {[member 1] name: ...; [member 2] name: ...; ...};

For example, define the general description of the device structure as:

{general}sample2:

{

[string]tscale:ns;

[string]xspace:nm;

[string] date: 2016-8-31

}

S3. Determine the composition rules of the structure description file, decompose the structure of the compound semiconductor device into five typical data types, and write the structure description file.

According to the structural characteristics of compound semiconductor devices and the data characteristics required for numerical simulation analysis, the device structure file is mainly divided into the following four parts:

1. General feature description, using general data types;

2. The description of physical objects such as material layers, interfaces/surfaces, optical layers, and mesh generation adopts basic data types and composite data types;

3. The description of the relationship between each material layer, interface/surface, and optical layer adopts the associated data type;

4. Description of device control characteristics, including working conditions and numerical characteristics, using control data types.

S4. Generate a dynamic pointer oriented to the device structure description file, and read the structure description file according to the description syntax rules;

In order to facilitate the increase and decrease of compound semiconductor device members and the maintainability and scalability of the device structure file, before reading the specific file, first generate pointers to the characteristics of the above four parts, which are sequentially represented in the file by the data structure of the linked list For each characteristic object encountered in , each object in the linked list is read according to the syntax rules in S2.

Since the reading of the file adopts the pointer to dynamically allocate space, the four parts of the file can intersect with each other, and the position is not necessarily fixed. The reading program will automatically allocate space after the data type that has been read according to the data type, and according to the general physical The principle is to preliminarily judge the rationality of the data.

As shown in Figure 3, it is the storage method of the final device structure. Material growth layers 311 , 312 comprising communication connections; heterojunction interfaces 321 , 322 ; material layer surfaces 331 , 332 ; associated data 341 , 342 ; control data 351 , 352 .

S5. According to the high-efficiency requirements of the numerical calculation process, determine the position, arrangement and format of various data in the numerical calculation exclusive file.

In order to improve the calculation efficiency as much as possible, three points need to be done: 1) Various data are stored in a fixed location in the file, and the numerical calculation software will read the relevant data sequentially; 2) The data format of the linked list is converted into a fixed-dimensional array format; 3) Transform experimental data into physical model parameters in a system of partial differential equations. details as follows:

1. According to the requirements of the calculation simulation software, determine the arrangement position of various data in the numerical calculation exclusive file.

The inventive method takes following position arrangement:

1) The data involved in the description of general characteristics;

2) The data involved in the description of physical objects such as material layers, interfaces/surfaces, optical layers, and grid generation;

3) The data involved in the description of the relationship between each material layer, interface/surface, and optical layer;

4) The data involved in the description of device control characteristics.

2. Each physical object exists in the form of an array

When reading the device structure, the heterogeneous material layer, growth material layer, various defects in the material layer, associated characteristics, and control characteristics stored in the linked list are all converted into a fixed array form. For example, the expression of the heterogeneous material layer is as follows:

Intrinsic parameters of heterogeneous material layers

…

Growth material layer number

Intrinsic parameters of growth material layer 1

…

number of defect types

Defect Type Parameters

Defect Type Parameters

…

Intrinsic parameters of growth material layer 2

…

number of defect types

Defect Type Parameters

Defect Type Parameters

…

3. Defect parameters are unified into a composite model

If the material/interface/surface defect description is in the form of electronic parameters, use physical relationships to convert to composite models.

4. Each associative array is indexed by the object array number

When reading the device structure, each physical object indexed by name has been stored in a fixed-dimensional array, so the name index of each associated array is converted into an object array number.

As shown in Fig. 4, the different characteristic descriptions for the compound semiconductor device permutations in the entire numerical calculation-specific file.

S6. According to the requirements of S5, convert the device structure stored by the experimental data and dynamic pointers into a numerical calculation exclusive file expressed by an array of model parameters, fixed position arrangement and fixed dimension, for use by the numerical calculation program.

As shown in FIG. 5 , it is Embodiment 1 of the structure description method of a compound semiconductor device with a complex structure in the present invention. Taking a GaInP/GaAs/InGaAs triple-junction solar cell with a large lattice mismatch as an example, it includes a p-InGaAs contact layer 51 , InAlGaAs back field 52, InGaAs active layer 53, GaInP window layer 54, InAlGaAs n++ doped layer 55, InAlGaAs p++ doped layer 56, p-InAlGaAs buffer layer 57, AlGaAs back field 58, GaAs active layer 59, AlInP A window layer 510 , a GaInP n++ doped layer 511 , an AlGaAs p++ doped layer 512 , an AlGaInP back field 513 , an AlGaInP active layer 514 , an AlGaInP window layer 515 , and a GaAs cap layer 516 . The structure is grown on an n-type GaAs substrate using low-pressure metal-organic chemical vapor deposition equipment. The structure contains two interband nonlocal quantum tunnels, which respectively occur between the InAlGaAs n++ doped layer 55 and the InAlGaAs p++ doped layer 56 and between the GaInP n++ doped layer 511 and the AlGaAs p++ doped layer 512, containing Six non-localized quantum tunnels spanning the heterojunction interface, respectively existing in the InAlGaAs back field 52 and the InGaAs active layer 53, the InGaAs active layer 53 and the GaInP window layer 54, and the AlGaAs back field 58 and the GaAs active layer 59 , between the GaAs active layer 59 and the AlInP window layer 510, between the AlGaInP back field 513 and the AlGaInP active layer 514, between the active layer 514 and the AlInP window layer 515, and so on. According to the method of the present invention, the inter-band non-local quantum tunneling associated data linked list containing two headers and the cross-heterojunction non-local quantum tunneling associated data linked list containing 6 headers are respectively generated, using the method of the present invention Because it no longer cares about the description positions of various physical characteristics, and various data can be added or subtracted at will, the complexity of the device structure description file is greatly reduced, and the readability and maintainability are improved.

As shown in FIG. 6, it is the second embodiment of the method for describing the structure of a compound semiconductor device with a complex structure in the present invention. Taking a 2D electron mobility transistor as an example, it includes a source electrode 61, a gate electrode 62, a drain electrode 63, and an upper barrier layer. 64, a channel layer 65, a lower barrier layer 66 and a substrate layer 67. The structure is grown on a semi-insulating GaAs substrate using low-pressure metal-organic chemical vapor deposition equipment. Wherein the metal-semiconductor contact type of the source electrode 61 and the drain electrode 63 is an ohmic type, and the metal-semiconductor contact type of the gate electrode 62 is a Schottky type, and the upper barrier layer 64 contains plane doping. Nonlocal quantum tunneling occurs between the gate electrode 62 and the external load, including two nonlocal quantum tunneling across the heterojunction interface, which occur respectively in the upper barrier layer 64 and the channel layer 65, and the channel layer Between 65 and the lower barrier layer 66, there is a quantum confinement region, the well region is in the channel layer 65, and the barrier layers are the upper barrier layer 64 and the lower barrier layer 66. In addition to generating various physical object array bits, the method of the present invention also generates linked list associated data types containing one nonlocal quantum tunneling across the Schottky junction interface and two nonlocal quantum tunneling across the heterojunction interface, generating A linked list associated data type with a quantum-limited area is introduced, which greatly improves the ability to deal with different types of associated physical objects.

Although the content of the present invention has been described in detail through the above preferred embodiments, it should be understood that the above description should not be considered as limiting the present invention. Various modifications and alterations to the present invention will become apparent to those skilled in the art upon reading the above disclosure. Therefore, the protection scope of the present invention should be defined by the appended claims.

Claims (8)

1. A method for describing the structure of a compound semiconductor device with a complex structure, characterized in that the method comprises: S1. Generate five typical data types according to the geometric characteristics, physical characteristics, and numerical characteristic parameters of the device structure required for the numerical analysis of compound semiconductor devices; S2. Determine the member assignment syntax rules corresponding to each typical data type; S3. Determine the composition rules of the structure description file, decompose the structure of the compound semiconductor device into five typical data types and write the structure description file; S4. Generate a dynamic pointer oriented to the device structure description file, and read the structure description file according to the description syntax rules; S5. According to the high-efficiency requirements of the numerical calculation process, determine the storage arrangement and data format of the numerical calculation-specific files; S6. According to the requirements of S5, convert the device structure stored in the experimental data and dynamic pointer into a numerical calculation exclusive file expressed in a fixed array of model parameters, positions and dimensions, for use in numerical calculation. 2. The structure description method of the compound semiconductor device of complex structure as claimed in claim 1, is characterized in that, in described S1, described five kinds of typical data types comprise: single fixed-value data type; Geometric features and basic physical object data types involved in partial differential equations that dominate the physical properties of carriers; Composite data types that contain physical phenomena of more than one primitive data type; An associative data type that describes the spatial correlation between different physical phenomena; A control data type describing the operating conditions of the compound semiconductor device being simulated. 3. The structure description method of the compound semiconductor device of complex structure as claimed in claim 1, is characterized in that, in described S2, described typical data type member assignment grammatical rule comprises: A single fixed-value data type assignment in a primitive data type takes the form of [type]name:value; The member assignment of the basic physical object data type in the basic data type takes the form of [type] name (can be default): member 1 = , member 2 = , ...;; A member assignment of a compound data type takes the form {type} name: {[member1] name: ...; [member2] name: ...; ...}; Member assignments of associated data types take the form [type]name: member1 = , member2 = , ...;; A member assignment of a control data type takes the form {type} name: {[member1] name: ...; [member2] name: ...; ...}; Member assignments of generic data types take the form {type} name: {[member1] name: ...; [member2] name: ...; ...}. 4. The structure description method of the compound semiconductor device of complex structure as claimed in claim 1 or 2, is characterized in that, in described S3, five kinds of typical data types in S1 are decomposed into compound semiconductor device structure and write structure description file . 5. The structure description method of the compound semiconductor device of complex structure as claimed in claim 1 or 4, it is characterized in that, in said S4, generate the dynamic pointer facing device structure description file, read structure description according to S2 description grammar rules document. 6. The method for describing the structure of a compound semiconductor device with a complex structure as claimed in claim 1, wherein in said S5, according to the high-efficiency requirements of the numerical calculation process, the position, arrangement and location of various data in the numerical calculation exclusive file are determined. Format. 7. The structure description method of the compound semiconductor device of complex structure as claimed in claim 6, is characterized in that, in described S6, the device structure that will store with experimental data and dynamic pointer is converted into according to S5 with model parameter, position arrangement Numerical calculation-specific files for fixed and fixed-dimensional array expressions, for use by numerical calculation programs. 8. The structural description method of the compound semiconductor device of complex structure as claimed in claim 1, is characterized in that, described structural description method is used in the numerical simulation and the analysis of compound semiconductor device structure; Compound semiconductor device structure comprises multi-junction solar Batteries, microwave devices, photodetectors, lasers.