Czochralski silicon single crystal growth method and device for cooperatively controlling volume microdefects and silicon wafer

Abstract

The invention provides a czochralski silicon single crystal growth method and device for cooperatively controlling volume microdefects and a silicon wafer. According to the method, a nitrogen-oxygen-vacancy ternary cooperative control strategy is implemented in the crystal pulling process, and the method specifically comprises the steps that the nitrogen doping amount is calculated in the charging stage, the rotating speed of a crucible and an airflow field are controlled in the melting and stabilizing stage to adjust the oxygen content, and a thermal field is shaped through an inverted-cone-shaped guide cylinder in the equal-diameter growth stage. And locking the V / G ratio of a silicon single crystal solid-liquid growth interface in a pure vacancy defect generation region by executing pulling speed locking logic. According to the invention, the nitrogen concentration in the crystal is limited in a low concentration range of 1 * 10 < 13 > to 5 * 10 < 13 > atoms / cm < 3 >, the oxygen concentration is controlled in a range of 11 to 13 ppma, and the high-concentration vacancy of the pure vacancy region is utilized to compensate the insufficient nucleation power under the low-nitrogen condition, so that the nitrogen-related defects are inhibited, the volume micro-defects with high density and uniformly distributed in the radial direction are obtained, and the high-purity nitrogen-containing crystal is obtained. And the gate oxide integrity of the silicon wafer is improved.

Description

Technical Field

[0001] This disclosure relates to the field of semiconductor material preparation technology, and in particular to a method, apparatus and silicon wafer for Czochralski silicon single crystal growth with coordinated control of bulk micro-defects. Background Technology

[0002] In the field of semiconductor integrated circuit manufacturing, silicon single crystals grown using the Czochralski (CZ) method are currently the mainstream substrate material. With the continuous miniaturization of device process nodes (e.g., entering 14nm, 7nm and more advanced nodes), the requirements for defect control within silicon wafers have reached a high standard. Among these, bulk microdefects (BMDs) are one of the most important microscopic defects within silicon wafers, and are mainly composed of oxygen precipitates (SiO₂). x The structure consists of BMD. An appropriate amount of BMD can create an internal gettering (IG) effect, effectively capturing heavy metal impurities introduced during the process, thereby purifying the active region of the device. However, the formation and distribution of BMD are complexly influenced by the thermal history and impurity concentration during crystal growth.

[0003] To enhance the mechanical strength of silicon wafers and promote the formation of bone matrix mating structures (BMDs), existing technologies commonly employ nitrogen doping. Nitrogen atoms can pin dislocations in the silicon lattice and act as heteronuclei, lowering the nucleation barrier for oxygen deposition. Therefore, conventional processes typically incorporate high concentrations of nitrogen (e.g., 1 × 10¹). 4 atoms / cm 3 (The above) is used to ensure a high-density BMD. While this maintains gettering capability, it also introduces side effects. In particular, the increase in nitrogen concentration alters the equilibrium of point defects and may lead to the formation of complex nitrogen-oxygen complexes.

[0004] However, a core technical problem exists in existing operating methods: they cannot maintain the radial uniformity and density of BMD while suppressing nitrogen-related defects (such as NDP). Specifically, when the nitrogen concentration is too high (e.g., exceeding 5 × 10⁻⁶), the problem becomes more difficult to resolve. 13 atoms / cm 3Nitrogen atoms readily aggregate to form stable clusters or combine with vacancies to form defects. These defects lead to nitrogen defect pits (NDPs) on the silicon wafer surface during subsequent high-temperature processing. When NDPs are located at the gate oxide layer of a device, they severely damage the integrity of the oxide layer, causing device leakage and affecting the performance of devices in advanced processes. To solve the NDP problem, the nitrogen concentration must be reduced. However, once the nitrogen concentration is reduced, the nucleation kinetics of BMDs become insufficient, resulting in a significant decrease in BMD density and uneven distribution, typically exhibiting a high density in the center and a low density at the edges, with a fluctuation rate often exceeding 20%. This non-uniformity leads to poor mechanical strength and weak gettering ability in the wafer edge region, failing to meet the requirements for consistent yield across the entire wafer. Therefore, how to obtain high-density and uniform BMDs at low nitrogen concentrations is a pressing technical challenge that needs to be addressed. Summary of the Invention

[0005] The technical problem this application aims to solve is to address the issue in existing nitrogen-doped Czochralski silicon single crystal growth technologies that cannot simultaneously suppress nitrogen-related defects (NDP) and maintain the density and uniformity of bulk microdefects (BMD). This application provides a Czochralski silicon single crystal growth method, apparatus, and silicon wafer based on a ternary synergistic mechanism. This technical solution can achieve the distribution ratio of nitrogen, oxygen, and vacancies through hardware structure control and full-process logic locking, thereby suppressing NDP while ensuring that the silicon wafer has radially uniformly distributed and moderately dense BMD.

[0006] The technical solution disclosed herein is implemented as follows: In a first aspect, this disclosure provides a method for Czochralski silicon single crystal growth with coordinated control of bulk micro-defects, the method comprising: Nitrogen dopant is incorporated into the silicon melt in the single-crystal furnace, resulting in a nitrogen concentration of 1 × 10⁻⁶ for the grown silicon single crystal. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; The process parameters are adjusted so that the interstitial oxygen concentration of the silicon single crystal is between 11 ppma and 13 ppma. During the constant diameter growth stage, the thermal field distribution of the single crystal furnace is adjusted and a pulling speed locking process is executed to lock the V / G ratio of the solid-liquid growth interface of the silicon single crystal in the pure vacancy defect generation region, where V is the pulling speed and G is the axial temperature gradient.

[0007] In a second aspect, this disclosure provides a Czochralski silicon single crystal growth apparatus for synergistically controlling bulk microdefects, the apparatus being used to perform the method according to the first aspect, the apparatus comprising: Furnace body; A crucible assembly, disposed within the furnace body, is used to contain molten silicon; A heater for heating and melting the molten silicon within the crucible assembly; A pulling mechanism for pulling silicon single crystals from the silicon melt; and A control unit, connected to the heater, the pulling mechanism, and the crucible assembly, is configured to control the amount of nitrogen dopant added, such that the nitrogen concentration of the grown silicon single crystal is 1 × 10⁻⁶. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 Adjust the process parameters so that the interstitial oxygen concentration of the silicon single crystal is 11 ppma to 13 ppma; during the constant diameter growth stage, adjust the thermal field distribution of the single crystal furnace and execute the pulling speed locking process to lock the V / G ratio of the solid-liquid growth interface of the silicon single crystal in the pure vacancy defect generation region.

[0008] Thirdly, this disclosure provides a silicon wafer, which is fabricated by the method according to the first aspect, and the silicon wafer has the following parameter characteristics: Nitrogen concentration range is 1×10 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; The interstitial oxygen concentration is 11 ppma to 13 ppma; The bulk micro-defect density is 1×10 13 pcs / cm 3 Up to 5×10 13 pcs / cm 3 ; The fluctuation rate of the bulk micro-defect density in the radial direction of the silicon wafer is less than 5%.

[0009] This disclosure provides a method, apparatus, and silicon wafer for Czochralski silicon single crystal growth with coordinated control of bulk micro-defects. Firstly, the problem of nitrogen-defect (NDP) defects is suppressed and resolved by limiting the nitrogen concentration to 5 × 10⁻⁶. 13 atoms / cm 3The following improvements reduce the material basis for nitrogen atom aggregation to form large-sized clusters, suppressing potential leakage current in devices and improving the gate oxide integrity of the silicon wafer. Secondly, high density and radially uniform distribution of BMD are achieved. This application utilizes the high concentration of vacancies in the "pure vacancy region" as an auxiliary nucleation factor, effectively compensating for insufficient nucleation kinetics under low nitrogen conditions. Combined with the rectification effect of the inverted conical flow guide tube on the thermal field, the entire radial direction of the crystal is kept within the pure vacancy region, thereby maintaining the radial fluctuation rate of BMD density to less than 5%, improving the mechanical strength and gettering consistency of the wafer. The device in this application, through the combination of pull-speed locking logic and a specific thermal field structure, provides a method applicable to industrial production, capable of controlling the V / G ratio and avoiding crystal quality instability caused by thermal field fluctuations in traditional processes. Attached Figure Description

[0010] Figure 1 This is a cross-sectional schematic diagram of the overall structure of the Czochralski silicon single crystal growth apparatus for controlling micro-defects in the collaborative control body provided in the embodiments of this application.

[0011] Figure 2 This is a partially enlarged schematic diagram showing the positional relationship between the inverted conical guide tube and the crystal, melt, and crucible in the embodiments of this application, illustrating the details of the thermal rectification structure.

[0012] Figure 3 This is a control flow diagram of the control unit executing the speed locking logic in the embodiment of this application.

[0013] Figure 4 This is a schematic flowchart of the Czochralski silicon single crystal growth method for co-controlling micro-defects provided in the embodiments of this application. Detailed Implementation

[0014] The technical solutions in this disclosure will now be clearly and completely described with reference to the accompanying drawings.

[0015] See Figure 1 , Figure 1 A cross-sectional view of the overall structure of a Czochralski silicon single crystal growth apparatus for co-controlling micro-defects according to an embodiment of this application is shown. The apparatus mainly includes a furnace body 1, a crucible assembly 3, a heater 4, a pulling mechanism 6, a thermal rectification structure, and a control unit 10. The furnace body 1 is typically made of stainless steel and has a double-layer water-cooled wall structure to maintain a low temperature on the outer wall of the furnace body during high-temperature crystal growth. A sealed processing chamber 2 is defined inside the furnace body 1, which can maintain a high vacuum or be filled with an inert gas (such as argon) at a controlled pressure during operation. An exhaust port 18 is provided at the bottom of the furnace body 1 for connecting to a vacuum pump system to discharge waste gas and control the pressure inside the furnace. An argon inlet 17 is provided at the top of the furnace body 1 for introducing high-purity argon to form a top-down protective gas flow.

[0016] At the center of the sealed processing chamber 2, a crucible assembly 3 is disposed. The crucible assembly 3 includes an inner quartz crucible 31 that directly contacts the silicon melt 8 and an outer graphite crucible 32 that supports the quartz crucible 31. The graphite crucible 32 is typically made of high-purity graphite and has a three- or multi-lobed spliced ​​structure to accommodate thermal expansion. The crucible assembly 3 is mounted at the top of a crucible shaft 33. The crucible shaft 33 passes through the bottom of the furnace body 1 and is connected to a rotary lifting drive mechanism (not shown). This drive mechanism is capable of driving the crucible assembly 3 to rotate at a predetermined speed (e.g., 0.1 to 15 rpm) and can also drive the crucible assembly 3 to rise as the level of the silicon melt 8 decreases during crystal growth, thereby maintaining a constant position of the melt level relative to the heater 4.

[0017] A heater 4 is arranged around the crucible assembly 3. The heater 4 is the core heat source of the entire thermal field and is typically made of isotropic high-purity graphite, designed in a tortuous or grid-like structure to provide uniform circumferential heat radiation. The heater 4 is connected to an external DC power supply via heating electrodes, melting the polycrystalline silicon raw material inside the quartz crucible 31 into silicon melt 8 through resistance heating. An insulation cylinder 5 is arranged around the heater 4. The insulation cylinder 5 is typically made of graphite felt or carbon fiber composite material, serving as thermal insulation to prevent heat loss to the cold walls of the furnace body 1, improving energy utilization efficiency, and creating the required radial temperature gradient.

[0018] Above the furnace body 1, a lifting mechanism 6 is installed. The lifting mechanism 6 includes a rotating chamber located at the top of the furnace body, a hoisting mechanism, and a lifting rope 16 that extends through a rotating sealing device into the sealed processing chamber 2. The lifting rope 16 is typically made of high-strength tungsten wire or steel wire rope. The lower end of the lifting rope 16 is connected to a seed crystal chuck for holding the seed crystal 15. The lifting mechanism 6 can control the rotational speed and vertical lifting speed of the seed crystal 15 and the grown silicon single crystal 7. During crystal growth, the seed crystal 15 is first immersed in the silicon melt 8, and then slowly lifted, going through steps such as necking, shoulder formation, and shoulder rotation, finally growing a cylindrical silicon single crystal 7 of the target diameter.

[0019] A key aspect of this application lies in the design of the thermal rectification structure, particularly the introduction of the inverted conical guide tube 9 and its specific geometry. See [link to relevant documentation]. Figure 2 An inverted conical flow guide tube 9 is coaxially positioned between the lifting mechanism 6 and the crucible assembly 3, suspended from the upper structure of the furnace body 1 or the insulation cylinder 5. The inverted conical flow guide tube 9 is positioned around the growth path of the silicon single crystal 7. Its main function is to shape the thermal and airflow distribution within the furnace. The inverted conical flow guide tube 9 has an upper opening 91 and a lower opening 92. The sidewalls of the inverted conical flow guide tube 9 gradually slope inward from the upper opening 91 to the lower opening 92, forming an inverted conical structure. This structure differs from the traditional straight-cylinder type flow guide tube.

[0020] Furthermore, the structural details of the inverted conical guide tube 9 are crucial to achieving the objectives of this application. The inverted conical guide tube 9 comprises an inner graphite shell 94 and an outer graphite shell 95, with a carbon felt insulation layer 96 filling the space between the two shells. This sandwich structure maintains both the mechanical strength of the guide tube and provides thermal insulation. To prevent carbon particles from contaminating the silicon melt 8, the outer surface of the outer graphite shell 95 is typically coated with a dense silicon carbide (SiC) coating. The argon gas flow entering from the argon inlet 17 is guided by the inverted conical guide tube 9 and blows downwards across the surface of the silicon melt 8.

[0021] More importantly, the geometry of the inverted conical guide tube 9 was determined based on thermal field simulation to optimize the axial temperature gradient G at the solid-liquid growth interface. Specifically, the diameter of the lower opening 92 of the inverted conical guide tube 9 (denoted as D) shield The target diameter (denoted as D) of the silicon single crystal to be grown is the same as that of the target diameter of the silicon single crystal to be grown. crystal The ratio of D is set between 1.1 and 1.3. For example, for a silicon single crystal with a diameter of 300 mm (12 inches), D shield The range should be controlled between 330mm and 390mm. If this ratio is too large, the radiation shielding effect of the guide tube on the crystal edge will be weakened, resulting in excessive heat dissipation at the crystal edge and an increased axial temperature gradient G at the edge. edge greater than the central axial temperature gradient G center This will lead to drastic radial fluctuations in the V / G ratio, making it impossible to achieve pure vacancy region control throughout the entire radial direction. If the ratio is too small, accidents such as crystals touching the flow guide tube are likely to occur, and excessively high airflow velocity will cause oscillations on the melt surface. By limiting the ratio to 1.1 to 1.3, the inverted conical flow guide tube 9 can effectively block radiation from the cold furnace wall, playing a "heat preservation" role for the crystal edges, thereby reducing G. edge , making G edge Approaching G center This causes the radial temperature gradient distribution to tend to flatten out. This is the physical basis for achieving a uniform BMD distribution across the entire radial direction.

[0022] Furthermore, a gas guiding gap 14 (also known as a melt gap) is formed between the lower opening 92 of the inverted conical guide tube 9 and the liquid surface of the silicon melt 8. The height of this gap (denoted as H) directly affects the thermal state of the solid-liquid growth interface 13. To maintain the stability of the growth process, this device is equipped with a high-precision liquid level detection sensor 11. The liquid level detection sensor 11 can be an optical sensor based on the laser triangular reflection principle or a vision sensor based on image processing. The liquid level detection sensor 11 is installed outside the observation window of the furnace body 1 to monitor the absolute position of the melt surface in real time. The liquid level detection sensor 11 transmits the signal to the control unit 10. The control unit 10 controls the rising speed of the crucible shaft 33 according to a preset program to compensate for the melt consumption caused by crystal growth, thereby always keeping the height H of the gas guiding gap 14 within a constant range of 20 mm to 30 mm. The constant H value maintains a stable flow rate of argon gas across the melt surface, thereby stabilizing the oxygen volatilization rate and also maintaining constant heat dissipation conditions at the bottom of the crystal.

[0023] The device also includes a diameter detection sensor 12, typically a high-resolution CCD camera or a line-scan camera mounted on the furnace top window. The diameter detection sensor 12 captures images of the crystal growth region in real time, particularly the high-brightness meniscus at the solid-liquid growth interface 13. Using image processing algorithms, the control unit 10 can calculate the current crystal diameter and the changing trend of the solid-liquid growth interface.

[0024] The control unit 10 is the core of the entire device. The control unit 10 is connected to the power supply of the heater 4, the servo motor of the lifting mechanism 6, the drive motor of the crucible assembly 3, and various sensors. The control unit 10 typically consists of a main controller 100, a storage module 101, and a computing module 102. The storage module 101 stores a preset crystal growth control program, which includes the core "pulling speed locking logic" of this application.

[0025] The following is combined Figure 3 The paper details the specific process steps and elaborates on the specific implementation methods for silicon single crystal growth using the aforementioned device, particularly how to achieve ternary synergistic control of nitrogen, oxygen, and vacancies.

[0026] First, the charging and melting steps are performed. High-purity electronic-grade polycrystalline silicon raw material is charged into the quartz crucible 31. To control the nitrogen concentration, a calculated amount of nitrogen dopant needs to be added at this stage. The nitrogen dopant can be a silicon wafer with a silicon nitride film or a pre-prepared high-nitrogen-doped silicon master alloy. The control unit 10 is based on the segregation coefficient of nitrogen in the silicon melt (k≈7×10). -4 ) and the upper limit of nitrogen concentration at the head of the target crystal (5×10) 13 atoms / cm 3A reverse calculation is then performed. Due to the low segregation coefficient of nitrogen, most of it will accumulate in the remaining melt as the crystal grows, resulting in a much higher nitrogen concentration at the tail end of the crystal compared to the head end. To prevent excessive nitrogen concentration at the tail end from causing NDP (nitrogen dihydrogen condensation), the initial addition amount should be conservatively calculated. The goal is to control the nitrogen concentration in the entire constant-diameter section of the crystal to 1 × 10⁻⁶. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 Within the range. Below 1×10 13 atoms / cm 3 This will result in too few nuclei for BMD formation, exceeding 5×10. 13 atoms / cm 3 This will induce NDP defects.

[0027] After loading, the furnace body 1 is sealed, evacuated, and filled with argon gas. The heater 4 is started to completely melt the raw materials. During melting, the control unit 10 monitors the power output to prevent the quartz crucible from overheating and deforming. After melting, the melt stabilization stage begins. At this time, the control unit 10 adjusts the crucible rotation speed. To meet the low to medium oxygen concentration requirements of this application, the crucible rotation speed is set in a low range, for example, 5 rpm to 8 rpm. The lower crucible rotation speed reduces the scouring effect of melt thermal convection on the quartz crucible wall, thereby reducing the oxygen dissolution rate. Combined with the efficient airflow from the inverted conical guide tube 9 carrying away surface-volatile silicon oxides, the interstitial oxygen (O2) entering the crystal can ultimately be removed. i The oxygen concentration is controlled within a narrow window of 11 ppma to 13 ppma. This oxygen concentration range provides sufficient precipitate feedstock while avoiding excessively large oxidation-induced stacking fault (OSF) rings or BMD sizes caused by excessive oxygen content.

[0028] Next, crystal pulling, necking, and shoulder growth are performed. Seed crystal 15 is immersed in the melt for crystal pulling to suppress dislocations caused by thermal shock. Then, a narrow neck is formed by rapid pulling. Next, the pulling speed and temperature are reduced to gradually increase the crystal diameter, forming a shoulder.

[0029] Once the crystal diameter reaches the target diameter (e.g., 300 mm), it enters the critical constant-diameter growth stage. During this stage, the control unit 10 activates the "pulling speed lock logic." Traditional automatic diameter control typically responds quickly to diameter changes by drastically adjusting the pulling speed V, but this leads to drastic fluctuations in the V / G ratio, causing the crystal to alternately grow in vacancy-dominated and interstitial atom-dominated regions, resulting in stacking faults. This application employs a different control strategy.

[0030] The control unit 10 first determines the critical V / G value (ξ) under the current thermal field configuration (especially when the inverted conical guide tube 9 is present) based on thermal field simulation data or previous experimental data.crit This value marks the boundary between vacancy-type crystals and interstitial atom-type crystals. To obtain the "pure vacancy (Pv) region," the target V / G value must be slightly greater than ξ. crit However, it cannot be too large to avoid entering the void (COP) generation zone. The control unit 10 sets a target lifting speed V. target (For example, 0.65 mm / min), this velocity corresponds to the center position of the pure vacancy region.

[0031] During the growth process, the computing module 102 analyzes the data transmitted back by the diameter detection sensor 12 in real time. When it detects that the crystal diameter is deviating from the target value (e.g., the diameter is increasing), the control unit 10 does not prioritize increasing the pulling speed, but instead prioritizes adjusting the power of the heater 4. By fine-tuning the heating power (e.g., increasing the power by 0.1%), the melt temperature is slightly increased, thereby reducing the crystallization rate and causing the diameter to shrink. During this process, the pulling speed V is "locked" at V0. target Near the same location, the variation in lifting speed V is limited to a small range (e.g., ±0.05 mm / min). This "temperature control priority, lifting speed lock" strategy ensures the stability of the V / G ratio over time.

[0032] Meanwhile, due to the special design of the inverted conical guide tube 9, the temperature gradient G at the solid-liquid growth interface 13 is distributed very uniformly in the radial direction. Therefore, when the pulling speed V is locked, the V / G value of the entire radial section (from the center to the edge) falls into the pure vacancy region. This creates a fully radial environment rich in vacancy.

[0033] In this environment, the ternary synergistic mechanism begins to function. Although the nitrogen concentration is controlled at a low level (less than 5 × 10⁻⁶), 13 atoms / cm 3 While individual oxygen vacancies alone are insufficient to provide a large number of heterogeneous nucleation sites, the high concentration of vacancies provided by pure vacancy regions serves as a compensating mechanism. According to thermodynamic principles, the formation of oxygen precipitates is accompanied by approximately two times volume expansion, which generates enormous lattice strain energy, hindering precipitate growth. The presence of vacancies can absorb this volume expansion, releasing strain energy and thus significantly reducing the critical nucleation radius and nucleation barrier of oxygen precipitates. In other words, vacancies help lower the nucleation barrier of oxygen precipitates. Therefore, this application utilizes a high concentration of vacancies to promote the formation of high-density and stable BMD cores with an appropriate amount of oxygen (11-13 ppma) and trace amounts of nitrogen.

[0034] Under current technological conditions, especially in silicon single crystals grown using high nitrogen doping and conventional thermal field configurations, the radial distribution of defects within such silicon single crystals is typically non-uniform. Specifically, at the crystal edges, due to the larger axial temperature gradient, the V / G ratio is lower, often falling into the interstitial atom-dominated region, making it difficult for boundary defects (BMDs) to form or resulting in low BMD density at the edges. In the crystal center, due to the excessively high nitrogen concentration, nitrogen-related defects (NDPs) are easily formed. Therefore, the defect distribution of the entire wafer exhibits a non-uniform state with NDPs present in the center and a lack of BMDs at the edges. In contrast, silicon single crystals grown using the apparatus and method provided in this application can effectively overcome the aforementioned problem of non-uniform defect distribution. Because the nitrogen concentration in the crystal is controlled at a low level, the formation of NDP defects is suppressed. At the same time, through the rectification effect of the thermal field by the inverted conical guide tube, combined with the precise control of the pulling speed by the control unit, the entire solid-liquid growth interface is in a pure vacancy region in the radial direction. This results in a uniform distribution of BMD across the entire radial cross-section of the silicon single crystal, with no significant edge attenuation, thus achieving a uniform distribution of defect density.

[0035] After the constant diameter growth is completed, the final growth is performed, and then the silicon single crystal 7 is lifted to the upper region inside the insulation cylinder 5 for cooling. The control unit 10 controls the cooling rate, allowing the crystal to remain in the temperature range of 1100℃ to 800℃ for an appropriate time. This temperature range is the critical interval for BMD core growth. Since a uniform and high-density core has already been formed during the growth stage, the controlled cooling at this stage allows these cores to grow into precipitates of moderate size (e.g., 50-100nm), which have both gettering properties and do not induce secondary dislocations or slip.

[0036] Furthermore, to further optimize control, the control unit 10 can also integrate an artificial intelligence-based correlation data model. This model, based on extensive historical production data, establishes a mathematical mapping relationship between vacancy concentration, nitrogen concentration, oxygen concentration, and the final BMD density. During the growth process, the control unit 10 can dynamically fine-tune the target casting speed V based on the measured nitrogen concentration (calculated through segregation) and oxygen concentration (estimated through a melt hydrodynamic model). target This is to maintain a good pure vacancy region state at different crystal length positions, thereby improving the quality consistency along the entire length.

[0037] In summary, the apparatus provided in this specific embodiment successfully implements a ternary synergistic strategy of nitrogen-oxygen-vacancy through hardware support of the inverted conical guide tube 9 and software control of the pull-speed locking logic. This strategy solves the balance problem between NDP suppression and BMD density maintenance in traditional processes, providing a more reliable solution for manufacturing high-quality silicon wafers that meet the requirements of 14nm and below process nodes. The core of this strategy lies in "compensating for the impact of low nitrogen concentration on nucleation by adjusting vacancy concentration," that is, using physical means (vacancy control) to make up for the deficiencies brought about by chemical means (nitrogen reduction).

[0038] Accordingly, see Figure 4 This application also provides a method for Czochralski silicon single crystal growth with coordinated control of bulk micro-defects. This method is performed using the apparatus described in the foregoing embodiments and includes the following steps S401, S402, and S403: S401: Nitrogen dopant is incorporated into the silicon melt in the single-crystal furnace so that the nitrogen concentration of the grown silicon single crystal is 1×10⁻⁶. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; S402: Adjust the process parameters so that the interstitial oxygen concentration of the silicon single crystal is 11 ppma to 13 ppma; S403: During the constant diameter growth stage, the thermal field distribution of the single crystal furnace is adjusted and a pulling speed locking process is executed to lock the V / G ratio of the solid-liquid growth interface of the silicon single crystal in the pure vacancy defect generation region, where V is the pulling speed and G is the axial temperature gradient.

[0039] To further verify the technical effects of this application, a series of embodiments and comparative examples are described in detail below.

[0040] In Example 1, a Czochralski silicon single crystal growth process with a diameter of 300 mm was used. The diameter ratio of the lower opening of the flow guide tube was set to 1.2. Nitrogen dopant was added to the initial charge to achieve a nitrogen concentration of 3.1 × 10⁻⁶ at the head of the crystal for constant diameter growth. 13 atoms / cm 3 By controlling the crucible rotation speed at 6 rpm, the interstitial oxygen concentration was 12.1 ppma. During the growth process, the pulling speed was locked at 0.65 mm / min, with fluctuations controlled within 0.03 mm / min. After a standard heat treatment at 800℃ for 4 h followed by 1000℃ for 16 h, the BMD density inside the silicon wafer was measured at the center to be 3.5 × 10⁻⁶. 9 pcs / cm 3 The value was measured at an edge with a radius of 145 mm as 3.3 × 10⁻⁶. 9pcs / cm 3 The radial variability of BMD density within the silicon wafer is only 3%. Detection using laser scattering tomography and a surface defect scanner shows an NDP density of less than 0.1 particles / cm². 2 The GOI yield reached 99.8%. This indicates that the synergistic mechanism of this application maintains a high density of gettering capability while suppressing defects.

[0041] In Example 2, to verify the rationality of the upper limit of nitrogen concentration, the initial nitrogen doping amount was adjusted to a nitrogen concentration of 4.9 × 10⁻⁶ to achieve uniform crystal growth. 13 atoms / cm 3 The remaining thermal field and process parameters remained consistent with those of Example 1. Test results showed that, since the nitrogen concentration was still within the limit of 5 × 10⁻⁶ as defined in this application, the thermal field performance was good. 13 Within the threshold, the NDP density increased slightly but remained at 0.1 cells / cm². 2 Below. Due to the enhanced nitrogen-assisted effect, the central value of BMD density increases to 4.2 × 10⁻⁶. 9 pcs / cm 3 The edge value is 4.1 × 10 9 pcs / cm 3 The radial volatility was 2.4%. The GOI yield remained at a high level of 99.5%. This demonstrates that limiting the nitrogen concentration upper limit to 5 × 10⁻⁶ is effective. 13 atoms / cm 3 It can effectively balance NDP suppression and BMD density enhancement.

[0042] In Comparative Example 1, the effect of the thermal rectification structure on uniformity was investigated. A conventional straight-cylinder guide tube was used in this experiment. The lower opening diameter ratio of this straight-cylinder guide tube was 1.5. The nitrogen and oxygen concentrations were set exactly the same as in Example 1. Because the straight-cylinder guide tube could not effectively block edge heat dissipation, an axial temperature gradient G occurred at the edge. edge The density is higher than that at the center. At the same average pulling speed, the V / G ratio decreases at the edges, falling into the interstitial atom-dominated region. The results show that although the BMD density is normal in the wafer center region, the BMD density in the edge region drops sharply to 1×10⁻⁶ due to the lack of vacancy-assisted nucleation. 8 pcs / cm 3 Below this, the radial variability exceeds 90%. This uneven distribution makes the wafer edges susceptible to metal contamination in subsequent device processes and results in insufficient mechanical strength, demonstrating the irreplaceable role of the inverted conical guide tube and specific diameter ratio in achieving uniform distribution.

[0043] In Comparative Example 2, the effect of high nitrogen concentration on defects was investigated. The thermal field structure of Example 1 was used, but the nitrogen concentration was increased to 1.5 × 10⁻⁶. 14 atoms / cm 3Although the BMD density further increased to 8.5×10 9 pcs / cm 3 The distribution was uniform, but a high density of NDP defects was detected on the wafer surface, reaching 65 defects / cm. 2 This high-density NDP resulted in premature breakdown of a large number of capacitors during the gate oxide integrity test, causing the GOI yield to drop to 62%. This result strongly supports the necessity of reducing nitrogen concentration in this application, and the effectiveness of replacing the traditional high-nitrogen process with the strategy of "compensating for the impact of low nitrogen concentration on nucleation by adjusting vacancy concentration".

[0044] In Comparative Example 3, the limitations of lower nitrogen concentrations were explored. The nitrogen concentration was reduced to 8.2 × 10⁻⁶. 12 atoms / cm 3 Even with the thermal field and pull-rate locking logic of this application, the BMD density after heat treatment is only 2 × 10⁻⁶ due to the lack of basic heterogeneous nucleation cores. 7 pcs / cm 3 Left and right, it cannot form an effective internal impurity absorption capacity. This proves that 1×10 13 atoms / cm 3 It is the necessary lower limit to ensure the density of micro-defects in the bulk.

[0045] Based on this, the silicon wafer obtained in this application exhibits unique physical characteristics. Due to the synergistic control of nitrogen-oxygen-vacancy throughout the entire process, the vacancies within the silicon wafer are mainly in the form of N2-V. m The composite exists in the form of a microscopic structure that locks in vacancies, preventing their aggregation into harmful large-sized void defects, while simultaneously releasing the strain energy from oxygen precipitation in subsequent processes. The resulting product exhibits stability in a 14nm process. Furthermore, the control unit 10 can integrate an AI-based correlation data model. This model, based on extensive historical data, can dynamically fine-tune the target pulling speed according to the oxygen and nitrogen segregation at different crystal length positions to maintain a good pure vacancy region state throughout the entire crystal length, ensuring that the produced wafer possesses the aforementioned parameters.

[0046] Accordingly, this application also provides a silicon wafer, which is fabricated by the method according to the foregoing embodiments, and the silicon wafer has the following parameter characteristics: Nitrogen concentration range is 1×10 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; The interstitial oxygen concentration is 11 ppma to 13 ppma; The bulk micro-defect density is 1×10 13 pcs / cm3 Up to 5×10 13 pcs / cm 3 ; The fluctuation rate of the bulk micro-defect density in the radial direction of the silicon wafer is less than 5%.

[0047] It should be noted that although the above embodiments are illustrated using a 300mm (12-inch) silicon single crystal as an example, the apparatus and method of this application are also applicable to the growth of Czochralski silicon single crystals with diameters of 200mm (8 inches), 450mm (18 inches), and others. For crystals of different diameters, the geometric dimensions of the corresponding components only need to be adjusted according to the proportional relationship described in the claims (e.g., a guide tube aperture ratio of 1.1-1.3). Similarly, although specific numerical ranges are mentioned in the text (e.g., rotation speed 5-8 rpm, pulling speed 0.65mm / min), these values ​​are only preferred examples. In actual operation, they can be adjusted within a reasonable range according to the characteristics of the specific thermal field. As long as the core logic of V / G falling within the pure vacancy region and the nitrogen and oxygen concentration being within the limited range is met, it falls within the protection scope of this application.

[0048] It should be noted that the technical solutions described in this disclosure can be combined arbitrarily as long as they do not conflict.

[0049] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for Czochralski silicon single crystal growth with coordinated control of bulk micro-defects, characterized in that, The method includes: Nitrogen dopant is incorporated into the silicon melt in the single-crystal furnace, resulting in a nitrogen concentration of 1 × 10⁻⁶ for the grown silicon single crystal. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; The process parameters are adjusted so that the interstitial oxygen concentration of the silicon single crystal is between 11 ppma and 13 ppma. During the constant diameter growth stage, the thermal field distribution of the single crystal furnace is adjusted and a pulling speed locking process is executed to lock the V / G ratio of the solid-liquid growth interface of the silicon single crystal in the pure vacancy defect generation region, where V is the pulling speed and G is the axial temperature gradient.

2. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 1, characterized in that, The pulling speed locking process includes: Obtain the critical V / G value under the current thermal field structure, where the critical V / G value corresponds to the boundary between the vacancy-dominated region and the interstitial atom-dominated region; Set a target lifting speed such that the V / G value corresponding to the target lifting speed is greater than the critical V / G value and less than the lower limit V / G value of the void-type defect generation zone; When fluctuations in crystal diameter are detected, the power of the heater is adjusted to correct the diameter, maintaining the change in the target lifting speed within a preset threshold.

3. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 2, characterized in that, The variation range of the target lifting speed is limited to ±0.05 mm / min.

4. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 1, characterized in that, The adjustment of the thermal field distribution of the single crystal furnace includes: setting an inverted conical guide tube around the silicon single crystal growth path, wherein the ratio of the diameter of the lower opening of the inverted conical guide tube to the target diameter of the silicon single crystal is 1.1 to 1.

3.

5. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 4, characterized in that, During the constant diameter growth stage, the liquid surface position of the silicon melt is monitored in real time and the crucible lifting speed is adjusted to keep the vertical distance between the lower opening of the inverted conical guide tube and the liquid surface of the silicon melt constant within the range of 20mm to 30mm.

6. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 1, characterized in that, The rotational speed of the crucible assembly containing the molten silicon is controlled between 5 rpm and 8 rpm to adjust the interstitial oxygen concentration.

7. The Czochralski method for co-controlling micro-defects in silicon single crystal growth according to claim 1, characterized in that, The method further includes controlling the silicon single crystal through a temperature range of 1100°C to 800°C according to a preset cooling process in order to control the growth size of oxygen precipitates.

8. The method for Czochralski silicon single crystal growth with coordinated control of micro-defects according to claim 1, characterized in that, The nitrogen dopant is a silicon wafer with a silicon nitride film or a highly nitrogen-doped silicon master alloy.

9. A Czochralski silicon single crystal growth apparatus for synergistic control of bulk micro-defects, characterized in that, The apparatus is used to perform the method according to any one of claims 1 to 8, the apparatus comprising: Furnace body; A crucible assembly, disposed within the furnace body, is used to contain molten silicon; A heater for heating and melting the molten silicon within the crucible assembly; A pulling mechanism for pulling silicon single crystals from the silicon melt; and A control unit, connected to the heater, the pulling mechanism, and the crucible assembly, is configured to control the amount of nitrogen dopant added, such that the nitrogen concentration of the grown silicon single crystal is 1 × 10⁻⁶. 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 Adjust the process parameters so that the interstitial oxygen concentration of the silicon single crystal is 11 ppma to 13 ppma; during the constant diameter growth stage, adjust the thermal field distribution of the single crystal furnace and execute the pulling speed locking process to lock the V / G ratio of the solid-liquid growth interface of the silicon single crystal in the pure vacancy defect generation region.

10. A silicon wafer, characterized in that, The silicon wafer is prepared by the method according to any one of claims 1 to 8, and the silicon wafer has the following parameter characteristics: Nitrogen concentration range is 1×10 13 atoms / cm 3 Up to 5×10 13 atoms / cm 3 ; The interstitial oxygen concentration is 11 ppma to 13 ppma; The bulk micro-defect density is 1×10 13 pcs / cm 3 Up to 5×10 13 pcs / cm 3 ; The fluctuation rate of the bulk micro-defect density in the radial direction of the silicon wafer is less than 5%.

Images