A gamma-ray based semiconductor substrate conditioning method
By constructing a defect state assessment model and objective optimization function, and adaptively adjusting the gamma-ray irradiation and vacuum annealing processes, the problems of large resistivity dispersion and lattice integrity loss in existing technologies are solved, achieving high-precision control of electrical performance and batch-to-batch consistency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHENGDU GAOTONG ISOTOPE CO LTD
- Filing Date
- 2026-05-08
- Publication Date
- 2026-06-05
AI Technical Summary
In existing gamma-ray irradiation control technologies, the determination of process parameters relies on fixed experience, ignoring the differences in the initial state of the substrate. This results in large resistivity dispersion and inconsistent loss of lattice integrity, making it difficult to stably hit high resistivity targets. Furthermore, the independent processing of irradiation and annealing leads to insufficient repair or over-annealing, making it difficult to guarantee the performance consistency within and between batches.
By constructing a defect state assessment model and objective optimization function, gamma-ray irradiation parameters and vacuum annealing process are adaptively determined. By combining defect concentration prediction and resistivity correlation, deep coupling of irradiation and annealing is achieved, and process parameters are dynamically adjusted to match the initial state of the substrate, thus optimizing the annealing process.
It significantly improves the accuracy and process robustness of synergistic control of irradiation and annealing, stabilizes resistivity within the target value, increases lattice integrity retention to 97%, and greatly reduces performance dispersion among substrates in the same batch.
Smart Images
Figure CN122161353A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of semiconductor processing technology, and in particular to a method for controlling semiconductor substrates based on gamma rays. Background Technology
[0002] Third-generation wide-bandgap semiconductor substrates (such as silicon carbide, gallium nitride, and diamond) have important applications in high-frequency, high-power, and high-temperature devices. Precise control of their electrical properties, such as resistivity and carrier concentration, is crucial for improving device performance and reliability. Currently, common control methods include ion implantation doping, high-temperature annealing modification, and epitaxial growth doping. Ion implantation allows for control over doping concentration and depth, but it easily introduces severe lattice damage into the substrate, requiring subsequent high-temperature annealing for repair; this process is complex and energy-intensive. High-temperature annealing modification requires stringent temperature uniformity and time control precision, easily leading to substrate composition segregation. While epitaxial growth doping can achieve better uniformity, it is expensive, time-consuming, and makes it difficult to adjust the performance of the prepared substrate later. Gamma-ray irradiation, as a non-contact, large-area processing method, has been attempted for controlling the electrical properties of semiconductor substrates in recent years; however, its process parameters typically rely on fixed empirical ranges and lack adaptability to differences in the initial state of the substrate.
[0003] Studies have found that in existing gamma-ray irradiation modulation techniques, process parameters (such as cumulative dose, dose rate, annealing temperature, and time) are often determined based on limited experimental data or general empirical tables, applying similar irradiation conditions to substrates of different material types, initial carrier concentrations, and resistivities. This approach ignores the influence of the substrate's initial state on the irradiation response, resulting in large resistivity dispersion and inconsistent lattice integrity loss after modulation, making it difficult to stably hit high resistivity targets. Furthermore, irradiation introduces various defects, the concentration and distribution of which directly determine the electrical performance after annealing. However, in existing processes, irradiation and annealing are usually treated as independent steps, with annealing parameters (temperature, time, stage) essentially fixed. This often leads to insufficient repair or over-annealing, failing to fully realize the modulation potential of irradiation and making it difficult to guarantee performance consistency within and between batches. Summary of the Invention
[0004] To address the aforementioned problems in the prior art, this invention provides a method for controlling semiconductor substrates based on gamma rays.
[0005] This application provides a method for controlling a semiconductor substrate based on gamma rays, comprising: acquiring detection parameters of the semiconductor substrate to be processed; wherein the detection parameters include at least: substrate thickness and initial resistivity; determining gamma ray irradiation parameters based on the detection parameters of the semiconductor substrate and the target resistivity; wherein the gamma ray irradiation parameters include at least: irradiation distance, irradiation dose, and cumulative irradiation dose; inputting the detection parameters of the semiconductor substrate and the gamma ray irradiation parameters into a pre-constructed defect state assessment model, outputting a defect concentration prediction value, and inputting the defect concentration prediction value into a target optimization function for calculation to determine vacuum annealing process parameters; wherein the vacuum annealing process parameters include annealing temperature and annealing time; the target optimization function is correlated with the target resistivity; controlling a gamma ray radiation source based on the gamma ray irradiation parameters to irradiate the semiconductor substrate; and performing vacuum annealing treatment on the semiconductor substrate based on the vacuum annealing process parameters.
[0006] Optionally, the target optimization function includes a first optimization term and a second optimization term; wherein, the first optimization term is used to calculate the deviation between the defect concentration after vacuum treatment and the target defect concentration, and the second optimization term is used to calculate the deviation between the resistivity after vacuum treatment and the target resistivity; the target optimization function is further constrained by setting constraint terms; wherein, the setting constraint terms are determined by the material type of the semiconductor substrate.
[0007] Optionally, the vacuum annealing process includes M processing stages; M is a positive integer; the objective optimization function is the sum of sub-optimization terms of the M processing stages; wherein, the sub-optimization terms of each stage are constrained by different set constraint terms; the constraint terms of the sub-optimization terms of each stage are determined by the material type of the semiconductor substrate.
[0008] Optionally, after vacuum annealing the semiconductor substrate, the method further includes: acquiring the adjusted performance parameters of the semiconductor substrate; the adjusted performance parameters include current resistivity, current carrier concentration, and lattice integrity; comparing the adjusted performance parameters with target performance parameters and calculating the performance parameter deviation; and, in response to the performance parameter deviation being greater than a set value, adjusting the process parameters of other semiconductor substrates belonging to the same batch as the semiconductor substrate; wherein the process parameters include the adjustment amount of gamma-ray irradiation parameters and the adjustment amount of the vacuum annealing process parameters.
[0009] Optionally, the reverse adjustment of process parameters of other semiconductor substrates belonging to the same batch as the semiconductor substrate includes: constructing a response surface model; wherein the response surface model characterizes the mapping relationship between process parameters and performance; constructing an objective function in combination with the response surface model, and determining the process parameter quantity by solving the objective function; wherein the objective function characterizes the difference between the performance parameter after adjusting the process parameter quantity and the target performance parameter.
[0010] Optionally, the detection parameters further include the initial carrier concentration; determining the gamma-ray irradiation parameters based on the detection parameters of the semiconductor substrate and the target resistivity includes: identifying the material type of the semiconductor substrate; wherein the material type includes silicon carbide, gallium nitride, and diamond; obtaining the target resistivity and instantaneous dose rate corresponding to the material type; determining the cumulative radiation dose of the semiconductor substrate based on the material type, the target resistivity, the initial resistivity, and the initial carrier concentration; and determining the irradiation distance based on the instantaneous dose rate and the substrate thickness.
[0011] Optionally, determining the cumulative radiation dose of the semiconductor substrate includes: determining a modulation coefficient based on the material type; generating a correction weighting factor based on the initial carrier concentration; and determining the cumulative radiation dose based on the modulation coefficient, the correction weighting factor, the target resistivity, and the initial resistivity; wherein the cumulative radiation dose is directly proportional to the logarithm of the ratio of the target resistivity to the initial resistivity; the cumulative radiation dose is inversely proportional to the modulation coefficient; and the cumulative radiation dose is inversely proportional to the correction weighting factor.
[0012] Optionally, determining the irradiation distance based on the instantaneous dose rate and the substrate thickness includes: obtaining a reference distance and a reference dose rate measured at the reference distance; obtaining a thickness compensation coefficient corresponding to the material type; and determining the irradiation distance based on the instantaneous dose rate, the reference dose rate, the reference distance, the substrate thickness, and the thickness compensation coefficient.
[0013] Optionally, before obtaining the detection parameters of the semiconductor substrate to be processed, the method further includes: identifying the material type of the semiconductor substrate; wherein the material type includes silicon carbide, gallium nitride, and diamond; determining pretreatment process parameters for the material type of the semiconductor substrate; and performing cleaning, immersion, and drying treatments on the semiconductor substrate based on the pretreatment process parameters.
[0014] Optionally, the detection parameters further include: surface roughness and lattice integrity retention rate; before determining the γ-ray irradiation parameters based on the detection parameters and target resistivity of the semiconductor substrate, the method further includes: determining that the surface roughness of the semiconductor substrate is less than a set upper roughness threshold, and determining that the lattice integrity retention rate of the semiconductor substrate is greater than a set lower integrity threshold.
[0015] The beneficial effects of this invention include: By constructing a closed-loop process for acquiring detection parameters, determining irradiation parameters, assessing defect state, and optimizing annealing parameters, this application deeply couples gamma-ray irradiation and vacuum annealing through a defect concentration prediction model and an objective optimization function. This overcomes the shortcomings of traditional processes where irradiation and annealing are independent, and annealing parameters rely on fixed experience and cannot adapt to fluctuations in defect state after irradiation. This method can adaptively determine irradiation parameters based on the initial thickness and resistivity of the substrate, and dynamically calculate the optimal annealing temperature and time after irradiation based on the defect concentration predicted by the model. This significantly improves the accuracy and process robustness of the coordinated control of irradiation and annealing.
[0016] Furthermore, by introducing a defect state assessment model to quantitatively predict the concentration of intermediate-state defects generated by irradiation, and combining this with an optimization function associated with the target resistivity to determine annealing parameters, this method can perform differentiated annealing repair for substrates with different initial states. This effectively avoids insufficient repair (resistivity not meeting standards) or over-annealing (decreased lattice integrity) caused by missing defect predictions. Compared with existing empirical processes with fixed parameter ranges, this method can stably control the final resistivity within one order of magnitude of the target value, while improving the substrate lattice integrity retention rate to over 97%, and significantly reducing the performance dispersion between substrates in the same batch. Attached Figure Description
[0017] Figure 1 A flowchart illustrating the steps of the first γ-ray-based semiconductor substrate modulation method provided in this embodiment of the invention; Figure 2 A flowchart illustrating the steps of a second γ-ray-based semiconductor substrate modulation method provided in an embodiment of the present invention; Figure 3 The flowchart illustrates the steps of the third γ-ray-based semiconductor substrate modulation method provided in this embodiment of the invention. Detailed Implementation
[0018] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0019] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0020] Studies have found that in existing gamma-ray irradiation modulation techniques, process parameters (such as cumulative dose, dose rate, annealing temperature, and time) are often determined based on limited experimental data or general empirical tables, applying similar irradiation conditions to substrates of different material types, initial carrier concentrations, and resistivities. This approach ignores the influence of the substrate's initial state on the irradiation response, resulting in large resistivity dispersion and inconsistent lattice integrity loss after modulation, making it difficult to stably hit high resistivity targets. Furthermore, irradiation introduces various defects, the concentration and distribution of which directly determine the electrical performance after annealing. However, in existing processes, irradiation and annealing are usually treated as independent steps, with annealing parameters (temperature, time, stage) essentially fixed. This often leads to insufficient repair or over-annealing, failing to fully realize the modulation potential of irradiation and making it difficult to guarantee performance consistency within and between batches.
[0021] In view of the above problems, this application proposes the following embodiments to solve the above technical problems.
[0022] Please see Figure 1 This application provides a method for controlling a semiconductor substrate based on gamma rays, including steps 101 to 105.
[0023] Step 101: Obtain the detection parameters of the semiconductor substrate to be processed.
[0024] The detection parameters include at least the substrate thickness and the initial resistivity.
[0025] The above-mentioned detection parameters can be measured using standard semiconductor characterization equipment such as Hall effect testers and four-probe resistance testers.
[0026] Prior to step 101, the semiconductor substrate to be processed may be cleaned, soaked, and dried.
[0027] In the embodiments of this application, the semiconductor substrate to be processed includes at least third-generation wide-bandgap semiconductor materials such as silicon carbide (SiC), gallium nitride (GaN), and diamond.
[0028] This step is used to remove organic contaminants, metal ions, particles, and the natural oxide layer from the surface of the semiconductor substrate, resulting in a clean, activated surface. The cleaning process may include ultrasonic cleaning with organic solvents (such as acetone or ethanol), followed by a short immersion in a specific acidic solution (such as diluted hydrofluoric acid solution) to remove oxides, and finally rinsing with high-purity deionized water and drying in a clean environment.
[0029] Step 102: Determine the gamma-ray irradiation parameters based on the detection parameters of the semiconductor substrate and the target resistivity.
[0030] Among them, the gamma-ray irradiation parameters include at least: irradiation distance, irradiation dose, and cumulative irradiation dose.
[0031] That is, after obtaining the detection parameters, the process parameters for gamma-ray irradiation can be automatically determined by combining the target electrical performance requirements (such as target resistivity) or fixed performance requirements.
[0032] In one embodiment, a mapping table of detection parameters of the semiconductor substrate, target resistivity, and gamma-ray irradiation parameters can be preset, and then the gamma-ray irradiation parameters can be determined based on the mapping table.
[0033] The target resistivity can be determined based on the material type of the identified semiconductor substrate, or a uniform target resistivity can be used as needed.
[0034] Step 103: Input the detection parameters and gamma-ray irradiation parameters of the semiconductor substrate into the pre-built defect state assessment model, output the defect concentration prediction value, and input the defect concentration prediction value into the objective optimization function for calculation to determine the vacuum annealing process parameters.
[0035] Among them, the vacuum annealing process parameters include annealing temperature and annealing time; the objective optimization function is related to the objective resistivity.
[0036] In this embodiment, the defect state assessment module is a supervised machine learning model, which can be a deep neural network model or a gradient boosting decision tree. During training, each sample is labeled with the defect concentration actually measured by a precision instrument. The machine learning model can learn complex, nonlinear, and multi-factor coupled relationships from massive historical data, and its prediction accuracy for defect concentration is significantly higher than that of analytical physical models based on simplifying assumptions.
[0037] This application combines high-precision defect prediction and multi-objective optimization to dynamically generate optimal annealing processes for different semiconductor substrates. It should be noted that traditional high-temperature annealing processes use the same temperature and time for most batches of substrates, or are set based on human experience. When introducing vacuum annealing, it was found that differences in initial defect states due to variations in irradiation dose and materials can easily lead to insufficient repair or over-annealing. Based on this, a defect state assessment model is introduced to determine the internal defect state of each substrate after irradiation. Then, combined with the objective optimization function, the optimal temperature and time are determined to ensure a precise match between the annealing process and the irradiation results. This maximizes repair efficiency, effectively eliminates lattice damage with the most suitable annealing conditions, and maximizes lattice integrity retention. It also precisely stabilizes electrical performance, avoiding performance degradation due to over-annealing or performance instability due to under-annealing.
[0038] Step 104: Irradiate the semiconductor substrate by controlling the gamma-ray radiation source based on the gamma-ray irradiation parameters.
[0039] Then, using a calibrated gamma-ray irradiation device (such as a cobalt-60 or cesium-137 radioactive source), the semiconductor substrate is irradiated at room temperature and normal pressure according to the gamma-ray irradiation parameters determined in step 102.
[0040] Step 105: Perform vacuum annealing on the semiconductor substrate based on the vacuum annealing process parameters.
[0041] After irradiation, the semiconductor substrate is subjected to vacuum annealing based on the vacuum annealing process parameters determined in step 104.
[0042] The principle of irradiation vacuum annealing provided in the embodiments of this application is explained below. When a semiconductor substrate is irradiated with gamma rays, the high-energy photons they carry interact with the atoms inside the semiconductor substrate, generating ionization and excitation. This results in the formation of an appropriate amount of intrinsic defects inside the substrate. These defects can effectively capture charge carriers (electrons or holes) in the substrate, significantly reducing the charge carrier concentration and thus increasing the substrate resistivity. At the same time, the energy of gamma rays can promote the migration and rearrangement of impurity atoms inside the substrate, reducing the contribution of impurity energy levels to charge carriers, optimizing charge carrier mobility, and achieving synergistic regulation of electrical performance. After irradiation, vacuum annealing is performed. Vacuum annealing can further repair the slight lattice defects generated during irradiation, ensuring the crystal quality and structural stability of the substrate, while locking in electrical performance parameters to avoid performance degradation during long-term use.
[0043] This application's embodiments construct a closed-loop process for acquiring detection parameters, determining irradiation parameters, assessing defect state, and optimizing annealing parameters. It deeply couples gamma-ray irradiation and vacuum annealing through a defect concentration prediction model and an objective optimization function, overcoming the shortcomings of traditional processes where irradiation and annealing are independent, and annealing parameters rely on fixed experience and cannot adapt to post-irradiation defect state fluctuations. This method can adaptively determine irradiation parameters based on the initial substrate thickness and resistivity, and dynamically calculate the optimal annealing temperature and time after irradiation based on the defect concentration predicted by the model. This significantly improves the accuracy and process robustness of the coordinated control of irradiation and annealing.
[0044] Furthermore, by introducing a defect state assessment model to quantitatively predict the concentration of intermediate-state defects generated by irradiation, and combining this with an optimization function associated with the target resistivity to determine annealing parameters, this method can perform differentiated annealing repair for substrates with different initial states. This effectively avoids insufficient repair (resistivity not meeting standards) or over-annealing (decreased lattice integrity) caused by missing defect predictions. Compared with existing empirical processes with fixed parameter ranges, this method can stably control the final resistivity within one order of magnitude of the target value, while improving the substrate lattice integrity retention rate to over 97%, and significantly reducing the performance dispersion between substrates in the same batch.
[0045] Optionally, the objective optimization function includes a first optimization term and a second optimization term.
[0046] The first optimization term calculates the deviation between the defect concentration after vacuum treatment and the target defect concentration, while the second optimization term calculates the deviation between the resistivity after vacuum treatment and the target resistivity. The target optimization function is also constrained by setting constraint terms, which are determined by the material type of the semiconductor substrate.
[0047] It can be understood that the objective optimization function provided in the embodiments of this application is a mathematical function model of dual objective collaboration and material adaptive constraint.
[0048] The first optimization term can be expressed as the difference between the predicted residual defect concentration in the substrate after annealing at temperature T and time t and the ideal target defect concentration, which is used to directly quantify the repair effect of the annealing process on irradiation.
[0049] The second optimization term can be expressed as the predicted difference between the annealed resistivity of the substrate and the target resistivity after annealing at temperature T and time t, used to quantify the difference between the electrical performance and the target.
[0050] The constraints are set to be adapted according to different material types. For example, the annealing temperature range for gallium nitride substrates is 400-600℃, and the annealing temperature range for silicon carbide substrates is 300-500℃.
[0051] Ultimately, determining the vacuum annealing process parameters is transformed into solving the aforementioned constrained optimization problem. By constructing an optimization function that includes dual objective terms of defect repair and electrical performance, and applying dynamic constraints based on material properties, this application achieves precise control over the annealing processes of different semiconductors while ensuring process safety.
[0052] Furthermore, defect types can be introduced as labels during the training of the defect status assessment model, enabling the model to output predicted defect concentration values for each type of defect in practical applications.
[0053] In the subsequent construction of the objective optimization function, the first optimization term can be refined into the summation of the differences between the predicted repair values and the target values for all types of defects.
[0054] Furthermore, the aforementioned objective optimization function may also include process energy consumption and time cost optimization terms, which are not limited in this application.
[0055] Optionally, the vacuum annealing process includes M processing stages; M is a positive integer.
[0056] The objective optimization function is the sum of the sub-optimization terms of the M processing stages. Each sub-optimization term can be referred to the description in the previous embodiments, and will not be repeated here.
[0057] In this process, the sub-optimization terms of each stage are constrained by different set constraints; the constraints of the sub-optimization terms of each stage are determined by the material type of the semiconductor substrate.
[0058] It should be noted that the purpose of breaking down the vacuum annealing process into multiple different processing stages is to control different optimization objectives through different processes. For example, in the embodiments of this application, the vacuum annealing process is divided into three processing stages. Taking a gallium nitride substrate as an example, the first stage (low-temperature repair stage, e.g., 400-500℃): the main goal is to repair point defects (such as isolated vacancies and interstitial atoms) with low migration activation energies caused by irradiation. This stage has a low temperature but may take a long time, with the aim of gently eliminating most unstable simple defects without causing drastic lattice reconstruction. The second stage (medium-temperature recombination stage, e.g., 500-580℃): the main goal is to promote the decomposition and recombination of composite defects, as well as to promote the orderly arrangement of beneficial defects (such as specific vacancies). The third stage (high-temperature stabilization stage, e.g., 580-600℃): the main goal is to stabilize the lattice structure and lock in the final electrical properties.
[0059] By decomposing the annealing process into multiple stages with clearly defined objectives and different constraints, and using a collaborative optimization algorithm for global planning, this application can achieve precise management of the entire process of defect repair and performance stabilization.
[0060] Please see Figure 2 Optionally, after vacuum annealing the semiconductor substrate, the method further includes steps 201 to 203.
[0061] Step 201: Obtain the performance parameters of the semiconductor substrate after modulation.
[0062] The adjusted performance parameters include the current resistivity, current carrier concentration, and lattice integrity.
[0063] It should be noted that lattice integrity can be measured by high-resolution X-ray diffraction, which measures the full width at half maximum (FWHM) of characteristic diffraction peaks. By comparing these values with standard samples or theoretical values, the retention rate of lattice quality can be quantitatively assessed.
[0064] Step 202: Compare the adjusted performance parameters with the target performance parameters and calculate the performance parameter deviation.
[0065] Step 203: In response to the performance parameter deviation being greater than the set value, the process parameters of other semiconductor substrates belonging to the same batch as the semiconductor substrate are adjusted in reverse.
[0066] Among them, the process parameters include the adjustment of gamma-ray irradiation parameters (such as cumulative radiation dose, instantaneous dose rate, and irradiation distance) and the adjustment of vacuum annealing process parameters (annealing temperature and annealing time).
[0067] For example, if the current resistivity deviation is low, the cumulative radiation dose may increase. Similarly, if the lattice integrity is low, the annealing temperature may be increased or the annealing time extended.
[0068] In other words, when the deviation of the performance parameters is greater than the set value, it indicates that the adjustment of the current process parameters still deviates from the expectation. Here, the process parameters in the preceding process can be adjusted in reverse by using the final test results.
[0069] By employing the above methods, process deviations can be detected early, and compensatory adjustments can be made immediately within the same batch, avoiding the risk of scrapping the entire batch. This approach significantly improves batch uniformity and yield: ensuring that the processes experienced by all substrates in the same batch are dynamically optimized, thereby guaranteeing that the key performance parameters of the final product are consistently controlled within the target specifications, significantly reducing performance dispersion.
[0070] Optionally, the above steps reverse the process parameter quantities of other semiconductor substrates belonging to the same batch as the semiconductor substrate, including: constructing a response surface model; wherein the response surface model characterizes the mapping relationship between process parameters and performance; constructing an objective function in combination with the response surface model, and determining the process parameter quantities by solving the objective function; wherein the objective function characterizes the difference between the performance parameters after adjusting the process parameter quantities and the target performance parameters.
[0071] It should be noted that the response surface methodology (RSM) is constructed based on historical production data or specially designed process experimental data. Its input variables (independent variables) are key controllable process parameters, such as cumulative radiation dose, annealing temperature, and annealing time. Its output variables (dependent variables) are the final performance parameters, such as resistivity and lattice integrity. Through regression analysis (usually using a second-order polynomial model), this model can fit a continuous mathematical surface, i.e., the response surface, to quantitatively describe the complex nonlinear relationships and interactions between process parameters and performance indicators.
[0072] Then, a specific objective function is constructed based on the above response surface model. The core purpose of this objective function is to quantitatively evaluate the adjustment scheme. The basic form of the objective function can be the difference between the adjusted performance and the target performance. In this embodiment, a regularization term is added to the objective function to balance achieving the target and the adjustment magnitude. The objective function is usually transformed into a constrained least squares optimization problem, which can be solved efficiently using numerical optimization algorithms such as ridge regression and sequential quadratic programming. The final output is the new combination of process parameters (e.g., a 5% increase in cumulative radiation dose, an 8°C decrease in annealing temperature, etc.).
[0073] It should be noted that the above method transforms the reverse adjustment from a qualitative rule-based decision-making process into a computational process based on quantitative models and mathematical optimization. The response surface model provides a global and continuous determination of the process characteristics, while the optimization of the objective function ensures that the adjustment scheme is the current optimal solution. This method overcomes the shortcomings of empirical rules in handling multi-parameter coupling and nonlinear effects, making parameter adjustment more accurate and reliable.
[0074] Optionally, the detection parameters also include the initial carrier concentration. See the relevant documentation for details. Figure 3 The above steps determine the gamma-ray irradiation parameters based on the detection parameters and target resistivity of the semiconductor substrate, including steps 301 to 304.
[0075] Step 301: Identify the material type of the semiconductor substrate.
[0076] The materials include silicon carbide, gallium nitride, and diamond.
[0077] The identification process may include reading the identification code on the semiconductor substrate product and performing online analysis through spectral detection.
[0078] Step 302: Obtain the target resistivity and instantaneous dose rate corresponding to the material type.
[0079] The target resistivity can be preset to a typical range of target resistivity for different materials and application scenarios, based on the application requirements of the final device (such as the withstand voltage level of power devices).
[0080] The instantaneous dose rate is primarily determined based on the response characteristics of different materials to gamma rays, safety regulations, and equipment capabilities, setting a recommended dose rate range or fixed value for each material. For example, for GaN, the recommended instantaneous dose rate might be 20-40 kGy / h; for diamond, due to its better radiation resistance, the dose rate can be appropriately increased. The selection of the instantaneous dose rate affects process time and uniformity.
[0081] In one embodiment, a specific value can be randomly determined from a recommended range of instantaneous dose rates based on the material type of the semiconductor substrate.
[0082] Of course, a specific value can also be determined from the recommended instantaneous dose rate range by combining the initial carrier concentration and initial resistivity of the semiconductor substrate.
[0083] Step 303: Determine the cumulative radiation dose of the semiconductor substrate based on the material type, target resistivity, initial resistivity, and initial carrier concentration.
[0084] Step 304: Determine the irradiation distance based on the instantaneous dose rate and substrate thickness.
[0085] Determining the irradiation distance requires considering both the instantaneous dose rate and the substrate thickness. The instantaneous dose rate determines the basic order of magnitude of the irradiation distance to ensure the achievement of the target dose rate, while the substrate thickness is used to compensate for the exponential decay effect of gamma rays within the material. The greater the thickness, the more severe the surface dose loss after ray penetration. If the distance is not appropriately increased and thickness compensation is not introduced, it will lead to a significant difference in the concentration of defects on the front and back surfaces of the substrate, ultimately resulting in uneven resistivity distribution along the depth direction. Combining both factors satisfies the process's precise dose rate requirements while actively optimizing the dose uniformity along the thickness direction, thereby ensuring the consistency of the overall and depth-oriented electrical properties of the substrate after modulation. Therefore, this step combines the instantaneous dose rate and substrate thickness to determine the irradiation distance.
[0086] Of course, in practical applications, different irradiation distance ranges are set for semiconductor substrates of different material types, and then a specific value in the irradiation distance range is determined based on the instantaneous dose rate.
[0087] In summary, by identifying the material type of the semiconductor substrate and obtaining the corresponding target resistivity and instantaneous dose rate, and then combining the initial resistivity, initial carrier concentration, and substrate thickness, the cumulative radiation dose and irradiation distance are determined. This achieves material-adaptive and initial-state-adaptive irradiation parameters, avoiding the problems of large resistivity control dispersion and process window drift caused by using the same empirical parameter range for different materials and different batches of substrates in traditional processes. This significantly improves the process's compatibility with different substrates and the success rate of control.
[0088] Optionally, the above steps for determining the cumulative radiation dose of the semiconductor substrate include: determining a modulation coefficient based on the material type; generating a correction weighting factor based on the initial carrier concentration; and determining the cumulative radiation dose based on the modulation coefficient, the correction weighting factor, the target resistivity, and the initial resistivity. The cumulative radiation dose is directly proportional to the logarithm of the ratio of the target resistivity to the initial resistivity; the cumulative radiation dose is inversely proportional to the modulation coefficient; and the cumulative radiation dose is inversely proportional to the correction weighting factor.
[0089] For the specific formula of cumulative radiation dose, please refer to: ; in, Indicates cumulative radiation dose. Indicates the initial resistivity. Indicates the target resistivity; This represents the control coefficient; it reflects the efficiency of different semiconductor materials in generating effective defect compensation per unit dose under the same radiation irradiation. It is related to factors such as the material's displacement threshold energy, atomic density, and defect recombination rate. For example, silicon carbide has high atomic bond energy and low defect introduction rate, so the control coefficient can range from 0.3 to 0.6; gallium carbide has a medium defect introduction rate, so the control coefficient can range from 0.5 to 0.8; and diamond has strong atomic bond energy, so the control coefficient can range from 0.1 to 0.3. This represents the correction weighting factor; the correction weighting factor is used to correct for the dose consumption caused by the initial carrier concentration. The higher the initial carrier concentration, the more defects are needed to compensate for these free carriers, therefore the correction weighting factor decreases as the initial carrier concentration increases.
[0090] It is evident that by basing the determination of cumulative radiation dose on the theory of physical compensation, and by introducing material control coefficients and initial carrier concentration correction factors, a quantitative formula in which the dose is proportional to the logarithmic change in resistivity is constructed. This allows the irradiation dose to be accurately back-calculated based on the initial resistivity of the substrate, the target resistivity, material properties, and the initial carrier concentration. This significantly reduces dose setting deviations and minimizes the risk of substandard resistivity and excessive lattice damage caused by insufficient or excessive dose.
[0091] Optionally, determining the irradiation distance based on the instantaneous dose rate and substrate thickness includes: obtaining a reference distance and a reference dose rate measured at the reference distance; obtaining a thickness compensation coefficient corresponding to the material type; and determining the irradiation distance based on the instantaneous dose rate, reference dose rate, reference distance, substrate thickness, and thickness compensation coefficient.
[0092] For the specific formula for calculating irradiation distance, please refer to: You can refer to the following formula: ; in, This represents the irradiation distance to be calculated. Indicates the reference distance. This indicates the reference dose rate measured at the reference distance. This indicates the determined instantaneous dose rate. Indicates substrate thickness. This represents the thickness compensation coefficient, which is related to the material attenuation coefficient.
[0093] It should be noted that the above calculation method adds a compensation term proportional to the substrate thickness to the traditional geometric distance calculation. This ensures that the irradiation distance setting not only considers dose rate control but also optimizes the dose uniformity along the depth direction within the substrate.
[0094] It is evident that by introducing a compensation term proportional to the substrate thickness into the determination of the irradiation distance, the irradiation distance can be dynamically adjusted according to the substrate thickness, effectively compensating for the dose attenuation of γ-rays along the depth direction, optimizing the dose uniformity of the upper and lower surfaces of the substrate, and is especially suitable for thick or high atomic number substrates (such as diamond and thick silicon carbide), significantly improving the depth consistency and batch uniformity of the resistivity of the substrate after regulation.
[0095] Optionally, the above steps of cleaning, immersing, and drying the semiconductor substrate to be processed may specifically include: identifying the material type of the semiconductor substrate to be processed; wherein the material type includes silicon carbide, gallium nitride, and diamond; determining the preprocessing process parameters based on the material type of the semiconductor substrate to be processed; and cleaning, immersing, and drying the semiconductor substrate based on the preprocessing process parameters.
[0096] That is, different pretreatment process parameters are adapted to different semiconductor materials during the pretreatment process.
[0097] The complete process flow for different semiconductor substrates is described below.
[0098] Example 1: The semiconductor substrate is gallium nitride, and its complete process includes: When the substrate is identified as gallium nitride (in this example, it can be a single-crystal gallium nitride substrate, and the crystal type can be wurtzite), the gallium nitride substrate is sequentially cleaned and dried to remove oil, impurities and oxide layer from the substrate surface, and a clean gallium nitride substrate to be processed is obtained.
[0099] Specifically, first clean with acetone and ethanol using ultrasonic cleaning for 15-20 minutes, then soak in 8%-10% hydrofluoric acid solution for 3-4 minutes, and finally rinse with deionized water; the drying temperature is 120-150℃, and the drying time is 1.5-2 hours.
[0100] Then, the detection parameters of the gallium nitride substrate can be obtained; typically, the selected gallium nitride substrate has a thickness of 100-800 μm and an initial resistivity of 1×10³-1×10⁻⁶. 7 Ω·cm, initial carrier concentration 1×10¹ 5 -1×10¹ 8 cm⁻³.
[0101] Then, based on the detection parameters of the gallium nitride substrate, the gamma-ray irradiation parameters were determined. Typically, the instantaneous dose rate of gamma rays is stable within the range of 20-40 kGy / h; the gamma-ray irradiation source is a cobalt-60 gamma-ray source or a cesium-137 gamma-ray source, the energy of the gamma rays is 0.662-1.33 MeV, the irradiation distance is 60-120 cm, and the cumulative irradiation dose is controlled at 1 × 10⁻⁶ kGy / h. 7 Gy-1×10 9 Within the Gy range.
[0102] Then, the gallium nitride substrate to be treated was placed in the gamma-ray irradiation area and irradiated with gamma rays at room temperature and normal pressure.
[0103] After irradiation, the gallium nitride substrate is removed and subjected to vacuum annealing. The process parameters for vacuum annealing can be related to the material type of the gallium nitride substrate, the cumulative radiation dose, and the instantaneous dose rate. For example, the vacuum degree for vacuum annealing is 1×10⁻³-5×10⁻ 4 The annealing temperature is 400-600℃, the holding time is 2-5h, and after cooling to room temperature, a high-performance gallium nitride substrate with stable electrical properties is obtained. Experiments show that the resistivity of the high-performance gallium nitride substrate is 1×10¹¹-1×10¹³ Ω·cm, the resistivity uniformity error is ≤±4%, the lattice integrity retention rate is ≥97%, the surface roughness is ≤0.6nm, and the carrier mobility retention rate is ≥90%.
[0104] Through the above examples, precise control of the electrical properties of gallium nitride (GaN) substrates can be achieved based on the ionization effect and lattice modification of gamma rays: When a GaN substrate is irradiated with gamma rays, the high-energy photons they carry interact with Ga and N atoms inside the substrate, generating ionization and excitation. This results in the formation of appropriate intrinsic defects (such as Ga vacancies, N vacancies, interstitial atoms, and recombination defects) within the substrate. These defects can effectively trap charge carriers (electrons or holes) in the substrate, significantly reducing the charge carrier concentration and thus increasing the substrate resistivity. Simultaneously, the energy of gamma rays can promote the migration and rearrangement of impurity atoms inside the substrate, reducing the contribution of impurity energy levels to charge carriers, optimizing charge carrier mobility, and achieving synergistic control of electrical properties. Vacuum annealing can further repair minor lattice defects generated during irradiation, ensuring the crystal quality and structural stability of the substrate, while locking in electrical performance parameters to prevent performance degradation during long-term use.
[0105] Example 2: The semiconductor substrate is silicon carbide, and its complete process includes: When the substrate is identified as a silicon carbide substrate (the crystal form of the silicon carbide substrate can be 4H-SiC or 6H-SiC), the silicon carbide substrate is sequentially cleaned and dried to remove oil, impurities and oxide layer from the substrate surface, resulting in a clean silicon carbide substrate to be processed.
[0106] Specifically, first clean with acetone and ethanol using ultrasonic cleaning for 15-20 minutes, then soak in a 10%-15% hydrofluoric acid solution for 3-5 minutes, and finally rinse thoroughly with deionized water; the drying temperature is 120-150℃, and the drying time is 1.5-2 hours.
[0107] Then, the detection parameters of the silicon carbide substrate can be obtained; typically, the selected silicon carbide substrate has a thickness of 200-1000 μm and an initial resistivity of 1×10⁻⁶. 4 -1×10 8 Ω·cm.
[0108] Then, based on the detection parameters of the silicon carbide substrate, the gamma-ray irradiation parameters were determined. Typically, the instantaneous dose rate of gamma rays is stable within the range of 20-40 kGy / h; the gamma-ray irradiation source is a cobalt-60 gamma-ray source or a cesium-137 gamma-ray source, the energy of the gamma rays is 1.17-1.33 MeV, the irradiation distance is 50-100 cm, and the cumulative irradiation dose is controlled at 1 × 10⁻⁶ kGy / h. 5 Gy-1×10 7 Within the Gy range.
[0109] Then, the silicon carbide substrate to be treated was placed in the γ-ray irradiation area and subjected to γ-ray irradiation at room temperature and normal pressure.
[0110] After irradiation, the silicon carbide substrate is removed and subjected to vacuum annealing. The process parameters for vacuum annealing can be related to the material type of the silicon carbide substrate, the cumulative radiation dose, and the instantaneous dose rate. For example, the vacuum degree for vacuum annealing is 1×10⁻³-5×10⁻⁻⁻⁵. 4 Pa, annealing temperature of 300-500℃, holding time of 2-4h, and cooling to room temperature yield a high-performance silicon carbide substrate with a resistivity of 1×10¹²Ω·cm.
[0111] Through the above examples, the resistivity of silicon carbide substrates can be precisely controlled based on the ionization effect and lattice modification of gamma rays: When a silicon carbide substrate is irradiated with gamma rays, the high-energy photons they carry interact with the atoms inside the substrate, producing ionization and excitation, resulting in a small number of intrinsic defects (such as vacancies and interstitial atoms) inside the substrate. These defects can effectively trap charge carriers (electrons or holes) in the substrate, reducing the charge carrier concentration. At the same time, the energy of gamma rays can promote the migration and rearrangement of impurity atoms inside the substrate, reducing the contribution of impurity energy levels to charge carriers, thereby significantly improving the resistivity of the substrate. Subsequent vacuum annealing can further repair the slight lattice defects generated during irradiation, ensuring the crystal quality and structural stability of the substrate, while locking the resistivity to the level of 1×10¹²Ω·cm.
[0112] Example 3: The semiconductor substrate is a diamond substrate, and its complete process includes: When the substrate is identified as a diamond substrate (in this example, it can be a single-crystal diamond substrate with a cubic crystal system), the single-crystal diamond substrate is sequentially cleaned and dried to remove oil, impurities and oxide layer from the substrate surface, resulting in a clean diamond substrate to be processed.
[0113] Specifically, first clean with acetone and ethanol using ultrasonic cleaning for 15-20 minutes, then soak in a 5%-8% (volume fraction) mixed solution of hydrofluoric acid and nitric acid (volume ratio 1:1) for 2-3 minutes, and finally rinse with deionized water; the drying temperature is 100-130℃, and the drying time is 1.5-2 hours.
[0114] Then, the detection parameters of the diamond substrate can be obtained; typically, the selected diamond substrate has a thickness of 50-500 μm and an initial resistivity of 1×10⁻⁶. 4 -1×10 8 Ω·cm, initial carrier concentration 1×10¹ 4 -1×10¹ 7 cm⁻³.
[0115] Then, based on the detection parameters of the diamond substrate, the gamma-ray irradiation parameters were determined. Typically, the instantaneous dose rate of gamma rays is stable within the range of 40-60 kGy / h; the gamma-ray irradiation source is a cobalt-60 gamma-ray source or a cesium-137 gamma-ray source, the energy of the gamma rays is 0.662-1.33 MeV, the irradiation distance is 70-130 cm, and the cumulative irradiation dose is controlled at 1 × 10⁻⁶ kGy / h. 7 Gy-1×10 9 Within the Gy range.
[0116] Then, the diamond substrate to be treated was placed in the gamma-ray irradiation area and irradiated with gamma rays at room temperature and normal pressure.
[0117] After irradiation, the diamond substrate is removed and subjected to vacuum annealing. The process parameters for vacuum annealing can be related to the material type of the diamond substrate, the cumulative radiation dose, and the instantaneous dose rate. For example, the vacuum degree for vacuum annealing is 1×10⁻³-5×10⁻⁻⁻⁵. 4 Pa, annealing temperature of 500-700℃, holding time of 3-6h, and cooling to room temperature, yield a high-performance diamond substrate with stable electrical properties, locking the resistivity to the level of 1×10¹²Ω·cm.
[0118] Through the above examples, the precise control of the electrical properties of diamond substrates can be achieved based on the ionization effect and lattice modification of gamma rays: When gamma rays irradiate a diamond substrate, the high-energy photons they carry interact with the C atoms inside the substrate, producing ionization and excitation, resulting in an appropriate amount of intrinsic defects (such as C vacancies, interstitial C atoms, and recombination defects) within the substrate. These defects can effectively capture charge carriers (electrons or holes) in the substrate, significantly reducing the charge carrier concentration and thus increasing the substrate resistivity. At the same time, the energy of gamma rays can promote the migration and rearrangement of impurity atoms inside the substrate, reducing the contribution of impurity energy levels to charge carriers, optimizing charge carrier mobility, and achieving synergistic control of electrical properties. Subsequent vacuum annealing can further repair the slight lattice defects generated during irradiation, ensuring the crystal quality and structural stability of the substrate, while locking in electrical performance parameters to avoid performance degradation during long-term use.
[0119] Optionally, the detection parameters also include: surface roughness and lattice integrity retention rate; before determining the gamma-ray irradiation parameters based on the detection parameters of the semiconductor substrate, the method further includes: determining that the surface roughness of the semiconductor substrate is less than a set upper roughness threshold, and determining that the lattice integrity retention rate of the semiconductor substrate is greater than a set lower integrity threshold.
[0120] The surface roughness of a semiconductor substrate can be measured using an optical profilometer and can be characterized as arithmetic mean roughness or root mean square roughness. It reflects the microscopic smoothness of the semiconductor substrate surface.
[0121] Lattice integrity can be assessed using X-ray diffraction techniques.
[0122] The upper limit threshold for roughness can be set to 0.8 nm; the lower limit threshold for integrity can be set to 90%, and this application does not limit this.
[0123] It should be noted that a semiconductor substrate is deemed suitable for precise performance control only if it simultaneously meets both of the aforementioned threshold conditions, allowing it to proceed to the step of determining gamma-ray irradiation parameters. If either condition is not met, the semiconductor substrate will be rejected or transferred to a rework process. This mechanism ensures that only semiconductor substrates meeting the surface state and crystal quality requirements can proceed to the subsequent gamma-ray irradiation stage. This fundamentally avoids problems such as inconsistent performance control results and process window drift caused by uneven initial material quality.
[0124] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0125] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0126] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A method for controlling semiconductor substrates based on gamma rays, characterized in that, include: The detection parameters of the semiconductor substrate to be processed are obtained; wherein, the detection parameters include at least: substrate thickness and initial resistivity; Based on the detection parameters and target resistivity of the semiconductor substrate, gamma-ray irradiation parameters are determined; wherein, the gamma-ray irradiation parameters include at least: irradiation distance, irradiation dose, and cumulative irradiation dose; The detection parameters of the semiconductor substrate and the gamma-ray irradiation parameters are input into a pre-constructed defect state assessment model, which outputs a predicted defect concentration. The predicted defect concentration is then input into a target optimization function for calculation to determine the vacuum annealing process parameters. The vacuum annealing process parameters include annealing temperature and annealing time. The target optimization function is associated with the target resistivity. The semiconductor substrate is irradiated by controlling the gamma-ray radiation source based on the gamma-ray irradiation parameters. The semiconductor substrate is subjected to vacuum annealing based on the vacuum annealing process parameters.
2. The semiconductor substrate modulation method based on gamma rays according to claim 1, characterized in that, The objective optimization function includes a first optimization term and a second optimization term; Wherein, the first optimization term is used to calculate the deviation between the defect concentration after vacuum treatment and the target defect concentration, and the second optimization term is used to calculate the deviation between the resistivity after vacuum treatment and the target resistivity; The objective optimization function is further constrained by setting constraint terms; The setting constraint is determined by the material type of the semiconductor substrate.
3. The semiconductor substrate modulation method based on gamma rays according to claim 2, characterized in that, The vacuum annealing process includes M processing stages; M is a positive integer. The objective optimization function is the sum of the sub-optimization terms of M processing stages; In this process, the sub-optimization terms of each stage are constrained by different set constraints; the constraints of the sub-optimization terms of each stage are determined by the material type of the semiconductor substrate.
4. The semiconductor substrate modulation method based on gamma rays according to claim 3, characterized in that, After vacuum annealing the semiconductor substrate, the method further includes: The performance parameters of the semiconductor substrate after modulation are obtained; the performance parameters after modulation include the current resistivity, the current carrier concentration, and the lattice integrity. The adjusted performance parameters are compared with the target performance parameters, and the deviation of the performance parameters is calculated. In response to the performance parameter deviation being greater than a set value, the process parameters of other semiconductor substrates belonging to the same batch as the semiconductor substrate are adjusted in reverse. The process parameters include the adjustment amount of the gamma-ray irradiation parameters and the adjustment amount of the vacuum annealing process parameters.
5. The semiconductor substrate modulation method based on gamma rays according to claim 4, characterized in that, The reverse adjustment of process parameters for other semiconductor substrates belonging to the same batch as the semiconductor substrate includes: Construct a response surface model; wherein the response surface model characterizes the mapping relationship between process parameters and performance; An objective function is constructed by combining the response surface model, and the process parameter quantity is determined by solving the objective function; wherein, the objective function represents the difference between the performance parameter after adjusting the process parameter quantity and the target performance parameter.
6. The semiconductor substrate modulation method based on gamma rays according to claim 1, characterized in that, The detection parameters also include the initial carrier concentration; determining the gamma-ray irradiation parameters based on the detection parameters of the semiconductor substrate and the target resistivity includes: Identify the material type of the semiconductor substrate; wherein the material type includes silicon carbide, gallium nitride, and diamond; Obtain the target resistivity and instantaneous dose rate corresponding to the material type; The cumulative radiation dose of the semiconductor substrate is determined based on the material type, the target resistivity, the initial resistivity, and the initial carrier concentration. The irradiation distance is determined based on the instantaneous dose rate and the substrate thickness.
7. The semiconductor substrate modulation method based on gamma rays according to claim 6, characterized in that, Determining the cumulative radiation dose of the semiconductor substrate includes: Determine the control coefficient based on the material type; Based on the initial carrier concentration, a corrected weighting factor is generated; The cumulative radiation dose is determined based on the control coefficient, the correction weighting factor, the target resistivity, and the initial resistivity. The cumulative radiation dose is directly proportional to the logarithm of the ratio of the target resistivity to the initial resistivity; the cumulative radiation dose is inversely proportional to the control coefficient; and the cumulative radiation dose is inversely proportional to the correction weighting factor.
8. The semiconductor substrate modulation method based on gamma rays according to claim 6, characterized in that, Determining the irradiation distance based on the instantaneous dose rate and the substrate thickness includes: Obtain a reference distance and a reference dose rate measured at the reference distance; Obtain the thickness compensation coefficient corresponding to the material type; The irradiation distance is determined based on the instantaneous dose rate, the reference dose rate, the reference distance, the substrate thickness, and the thickness compensation coefficient.
9. The semiconductor substrate modulation method based on gamma rays according to claim 1, characterized in that, Before obtaining the detection parameters of the semiconductor substrate to be processed, the method further includes: Identify the material type of the semiconductor substrate; wherein the material type includes silicon carbide, gallium nitride, and diamond; Determine the pretreatment process parameters based on the material type of the semiconductor substrate; Based on the pretreatment process parameters, the semiconductor substrate is cleaned, soaked, and dried.
10. The semiconductor substrate modulation method based on gamma rays according to claim 1, characterized in that, The detection parameters also include: surface roughness and lattice integrity retention rate; Before determining the gamma-ray irradiation parameters based on the detection parameters and target resistivity of the semiconductor substrate, the method further includes: The surface roughness of the semiconductor substrate is determined to be less than a set upper roughness threshold, and the lattice integrity retention rate of the semiconductor substrate is determined to be greater than a set lower integrity threshold.