Self-adaptive cleaning method for photoinduced microbubble array silicon wafer

Abstract

The invention discloses a self-adaptive cleaning method for a light-induced microbubble array silicon wafer, and relates to the technical field of semiconductors, and the self-adaptive cleaning method for the light-induced microbubble array silicon wafer comprises the following steps: step 1, pretreatment before cleaning; step 2, generating a light-induced microbubble array; step 3, a self-adaptive cleaning process; step 4, pollutant removal and post-treatment; and 5, cleaning is completed. The sensor is adopted to monitor parameters such as turbidity and conductivity of cleaning liquid in real time, monitor the concentration of pollutants on the surface of the silicon wafer in real time and feed back and adjust light source parameters and microbubble density, dynamic optimization of the cleaning strength is achieved, the cleaning period and the agent concentration can be automatically matched according to the silicon wafers with different pollution degrees, and the cleaning efficiency is improved. The consumption of pure water and chemical reagents is reduced, resources are saved, and the applicability is improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and particularly to an adaptive cleaning method for a silicon wafer with a photoinduced microbubble array. Background Art

[0002] In the process of semiconductor silicon wafer manufacturing, the cleaning process has followed an older process for many years. Chemical solutions, ultrasonic and megasonic waves, and physical shaking or vibration are used to clean the metal, particles, and various stains on the silicon wafer surface. However, chemical solutions are likely to leave residues on the silicon wafer surface, and improper power during the use of ultrasonic and megasonic waves, as well as excessive physical shaking and vibration, will cause various physical damages to the silicon wafer. Therefore, many cleaning methods have emerged on the market.

[0003] Regarding the existing technology, there are the following problems:

[0004] The size of the bubbles cannot be controlled, resulting in uncontrollable cleaning accuracy, affecting the cleaning accuracy, poor applicability, no organic phosphate groups or metal ligands are modified on the surface of quantum dots, weak adsorption ability for metal ions, and thus an increased probability of metal contamination. At the same time, there is no circulating filtration system, resulting in the inability to recycle the cleaning solution and quantum dots, causing waste of resources and great limitations. Summary of the Invention

[0005] The present invention provides an adaptive cleaning method for a silicon wafer with a photoinduced microbubble array to solve the problems mentioned in the above background art.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is:

[0007] An adaptive cleaning method for a silicon wafer with a photoinduced microbubble array, the adaptive cleaning method for a silicon wafer with a photoinduced microbubble array includes the following steps:

[0008] Step 1: Pre-treatment before cleaning;

[0009] Step 2: Generation of a photoinduced microbubble array;

[0010] Step 3: Adaptive cleaning process;

[0011] Step 4: Removal of pollutants and post-treatment;

[0012] Step 5: Cleaning completed.

[0013] A further improvement of the technical solution of the present invention lies in that: Step 1 further includes the following steps:

[0014] A1: The loading method uses a bracket made of quartz or polytetrafluoroethylene (PTFE) to fix the silicon wafer by edge clamping or vacuum adsorption, avoiding contact with the surface of the silicon wafer (contact area < 1%), preventing secondary pollution. The loading density is controlled such that the single-slot loading capacity is 10 - 25 pieces according to the silicon wafer size (such as 12 inches), and the spacing between wafers is ≥ 2 cm to ensure uniform circulation of the cleaning liquid and bubbles.

[0015] A2: Use ultrapure deionized water (resistivity ≥ 18.2 MΩ·cm) as the base liquid, filter out particulate impurities through a 0.01 μm membrane, add sodium dodecylbenzenesulfonate (SDBS) or polyoxyethylene ether non-ionic surfactant, with its concentration in the range of 0.1% - 0.5% (mass ratio), and dissolve for 30 minutes by magnetic stirring (rotation speed 300 - 500 rpm) to form a homogeneous solution. At the same time, fine-tune with dilute nitric acid or ammonia according to the type of contamination, with an accuracy of ±0.5.

[0016] A3: Use glucose or citric acid with a purity ≥ 99.5% to react at 180 - 220 °C for 4 - 6 hours by hydrothermal method, and after cooling, purify through a 0.22 μm filter membrane to make the raw material. The particle size control is adjusted by the reaction temperature (larger particle size is generated at high temperature, and nanoscale particles are generated at low temperature), with an average particle size of 5 - 10 nm. Use tetrabutyl titanate as the precursor and ethanol as the solvent, hydrolyze and polymerize at 60 - 80 °C, and make metal oxide quantum dots with a particle size of 10 - 20 nm through annealing treatment (300 - 500 °C).

[0017] A4: Disperse the quantum dots in ammonium dihydrogen phosphate solution (concentration 0.01 mol / L), stir at 60 °C for 12 hours, graft phosphate groups (-PO3H2) through covalent bonds to enhance the chelating ability for metal ions such as Fe 3+ 、Al 3 + etc., or use ethylenediaminetetraacetic acid (EDTA) as a ligand to form coordination bonds with the metal ions on the surface of the quantum dots for metal ligand modification of the quantum dots, improving the selective adsorption of Cu 2+ . The loading method of the decorated quantum dots is added to the cleaning liquid by ultrasonic dispersion (power 300 W, time 10 minutes), and the concentration is controlled at 5 - 10 mg / L to ensure uniform suspension (sedimentation rate < 5% / hour).

[0018] A further improvement of the technical solution of the present invention is that: the second step further includes the following steps:

[0019] B1: Use an ultraviolet light source or a visible light source. The ultraviolet light source uses a low-pressure mercury lamp (wavelength 254 nm) or UV-LED (365 nm), which is suitable for TiO2 quantum dots. The visible light source uses an LED (400 - 700 nm), matching the light response range of carbon quantum dots (band gap 2.0 - 3.0 eV).

[0020] B2: Control the energy density by adjusting the light source power (0 - 100 W) and the distance from the silicon wafer surface (5 - 15 cm). The density range is 10 - 50 mW / cm 2 ;

[0021] B3: When the quantum dots are excited by light, electron-hole pairs are generated. The holes react with water to form hydroxyl radicals (·OH), and the electrons react with oxygen to form superoxide radicals (·O2 - ), while the radicals decompose organic substances, water molecules are reduced to produce H2 microbubbles, and oxidized to produce O2 microbubbles. The above is the photocatalytic reaction mechanism;

[0022] B4: The size control method for micron-sized (1 - 50 μm) bubbles uses low energy density (10 - 20 mW / cm 2 ) + low pulse frequency (10 - 30 Hz). The bubble generation rate is 10 4 ~10 5 bubbles per second per cm 2 , which is suitable for stripping large particles in the initial washing stage. Nano-sized bubbles (< 100 nm) use high energy density (30 - 50 mW / cm 2 ) + high pulse frequency (50 - 100 Hz), and are generated through the transient cavitation effect. The density reaches 10 9 bubbles / mL, which is suitable for decomposing molecular-level pollutants in the fine washing stage;

[0023] B5: Control the residence time of bubbles in the tank by the cleaning liquid circulation flow rate (5 - 20 L / min): high flow rate (20 L / min) reduces the bubble density, low flow rate (5 L / min) increases the density, and cooperate with the sensor feedback (such as automatically increasing the flow rate to 15 L / min when the turbidity > 1 NTU).

[0024] A further improvement of the technical solution of the present invention is that the step three further includes the following steps:

[0025] C1: An optical scattering type turbidity sensor with a measurement range of 0 - 100 NTU and an accuracy of 0.1 NTU for real-time monitoring of the concentration of particulate pollutants (≥ 0.1 μm particles) in the cleaning liquid, a four-electrode type conductivity sensor with a measurement range of 0 - 200 μS / cm and an accuracy of 0.1 μS / cm for monitoring the metal ion concentration, and an infrared spectrometer for on-line detection of organic pollutants (such as C-H bonds, C=O bonds) on the silicon wafer surface with a resolution of 4 cm -1 , and a scanning speed of 1 time per minute;

[0026] C2: When the turbidity > 0.5 NTU or the conductivity > 1 μS / cm, the system automatically increases the light source energy density by 10% - 20% and increases the quantum dot concentration by 1 - 2 mg / L. If the residual amount of metal ions detected > 1×10 10atoms / cm 2 , initiate the enhanced adsorption mode of surface-modified quantum dots.

[0027] 5. A method for self-adaptive cleaning of a photogenerated microbubble array silicon wafer according to claim 4, characterized in that: said step three further includes the following steps:

[0028] C3: By controlling the energy density range of the light source to be 30-50 mW / cm 2 , and the pulse frequency is 30-50 Hz, use microbubbles with a surfactant concentration of 0.3%-0.5% for primary cleaning for 30%-40% of the total cycle time;

[0029] C4: Then control the energy density range of the light source to be between 10-20 mW / cm 2 , the pulse frequency is between 50-100 Hz, use nanobubbles with a quantum dot concentration between 8-10 mg / L, and then perform fine cleaning for 60%-70% of the total cycle time.

[0030] A further improvement of the technical solution of the present invention is that: said step four further includes the following steps:

[0031] D1: The cleaning tank adopts an overflow system design. The specific setting method is to set an annular overflow port at the top of the cleaning tank, with a flow rate of 1-5 L / min. Use buoyancy to discharge the pollutants carried by the bubbles with the cleaning liquid. The overflow liquid enters the recovery tank after being filtered by a 0.01 μm membrane. A circulating filtration system is set in the recovery tank. The specific filtration methods include primary filtration with a 5 μm polypropylene filter element, precision filtration with a 0.1 μm polyvinylidene fluoride (PVDF) membrane, and magnetic separation (if the quantum dots are loaded with Fe3O₄ magnetic cores) or an affinity chromatography column (immobilized phosphate group ligand) quantum dot recovery system. Primary filtration is used to remove large particles, and precision filtration is used to intercept quantum dots and colloids. The quantum dot recovery rate is ≥90%;

[0032] A further improvement of the technical solution of the present invention is that: said step four further includes the following steps:

[0033] D2: Use a rinsing method of spraying + soaking (spraying for the first 2 minutes and soaking for the next 3 minutes), with an ultra-pure water flow rate of 10-15 L / min, a resistivity ≥18.2 MΩ·cm, and a temperature of 25±2 °C to clean the silicon wafer. At the same time, a control system is set for the key points of rinsing. The control method is to stop when the conductivity of the rinsing water is monitored online to be <0.5 μS / cm, which takes 3-5 minutes.

[0034] A further improvement of the technical solution of the present invention is that: said step four further includes the following steps:

[0035] D3: High-purity nitrogen with a purity of 99.999% is used to blow the surface of the silicon wafer at an angle of 45° under a pressure of 0.2 - 0.3 MPa for 1 - 2 minutes to remove the residual water droplets on the surface. For silicon wafers with a process below 7 nm, the CO2 critical drying method (temperature 31.1 °C, pressure 7.38 MPa) is used to avoid surface damage caused by capillary force, and the residual watermark after drying is ≤ 0.01%.

[0036] Due to the adoption of the above technical solution, the technical progress achieved by the present invention compared with the prior art is as follows:

[0037] 1. The present invention provides a method for self-adaptive cleaning of a silicon wafer with a photogenerated microbubble array. By using photoluminescent carbon quantum dots or metal oxide quantum dots as catalysts, reactive oxygen species are generated through photoexcitation to decompose organic pollutants on the surface of the silicon wafer, improving the cleaning efficiency. Organic phosphoric acid groups or metal ligands are modified on the surface of the quantum dots to enhance the adsorption and catalytic ability for specific metal ions, reducing the probability of metal contamination.

[0038] 2. The present invention provides a method for self-adaptive cleaning of a silicon wafer with a photogenerated microbubble array. By using sensors to real-time monitor parameters such as the turbidity and conductivity of the cleaning solution, and real-time monitor the concentration of pollutants on the surface of the silicon wafer, the light source parameters and microbubble density are feedback-regulated to achieve dynamic optimization of the cleaning intensity. It can automatically match the cleaning cycle and reagent concentration for silicon wafers with different pollution degrees, reducing the consumption of pure water and chemical reagents, saving resources, and improving applicability.

[0039] 3. The present invention provides a method for self-adaptive cleaning of a silicon wafer with a photogenerated microbubble array. The lifespan of the cleaning solution is extended through a circulating filtration system, reducing the frequency of reagent replacement; the quantum dots can be recycled and reused, reducing the treatment cost, further reducing the loss of resources, and improving applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic structural diagram of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0041] The following further describes the present invention in detail with reference to the embodiments:

[0042] Embodiment 1

[0043] As Figure 1 shown, the present invention provides a method for self-adaptive cleaning of a silicon wafer with a photogenerated microbubble array. A method for self-adaptive cleaning of a silicon wafer with a photogenerated microbubble array includes the following steps:

[0044] Step 1: Pretreatment before cleaning;

[0045] Step 2: Generation of a photogenerated microbubble array;

[0046] Step 3: Adaptive cleaning process;

[0047] Step 4: Pollutant removal and post-treatment;

[0048] Step 5: Cleaning completed.

[0049] Step 1 also includes the following steps:

[0050] A1: The loading method uses a bracket made of quartz or polytetrafluoroethylene (PTFE), and the silicon wafer is fixed by edge clamping or vacuum adsorption, avoiding contact with the surface of the silicon wafer (contact area < 1%), preventing the introduction of secondary pollution. The loading density is controlled such that the single-slot loading capacity is 10 - 25 pieces according to the silicon wafer size (such as 12 inches), and the wafer spacing is ≥ 2 cm to ensure uniform circulation of the cleaning liquid and bubbles;

[0051] A2: Use ultrapure deionized water (resistivity ≥ 18.2 MΩ·cm) as the base liquid, filter out particulate impurities through a 0.01 μm membrane, add sodium dodecylbenzenesulfonate (SDBS) or polyoxyethylene ether non-ionic surfactant, with its concentration in the range of 0.1% - 0.5% (mass ratio), and dissolve for 30 minutes by magnetic stirring (rotation speed 300 - 500 rpm) to form a homogeneous solution. At the same time, fine-tune with dilute nitric acid or ammonia according to the type of pollution, with an accuracy of ±0.5;

[0052] A3: Use glucose or citric acid with a purity ≥ 99.5% to react at 180 - 220 °C for 4 - 6 hours by hydrothermal method. After cooling, purify through a 0.22 μm filter membrane to make the raw material. The particle size control is adjusted by the reaction temperature (larger particle sizes are generated at high temperatures, and nanoscale particle sizes are generated at low temperatures), with an average particle size of 5 - 10 nm. Use tetrabutyl titanate as the precursor and ethanol as the solvent, hydrolyze and polymerize at 60 - 80 °C, and make metal oxide quantum dots with a particle size of 10 - 20 nm through annealing treatment (300 - 500 °C);

[0053] A4: Disperse the quantum dots in ammonium dihydrogen phosphate solution (concentration 0.01 mol / L), stir at 60 °C for 12 hours, graft phosphate groups (-PO3H2) through covalent bonds to enhance the chelating ability for metal ions such as Fe 3+ 、Al 3 + etc., or use ethylenediaminetetraacetic acid (EDTA) as a ligand to form coordination bonds with the metal ions on the surface of the quantum dots for metal ligand modification of the quantum dots, enhancing the selective adsorption of Cu 2+ . The loading method of the decorated quantum dots is added to the cleaning liquid by ultrasonic dispersion (power 300 W, time 10 minutes), and the concentration is controlled at 5 - 10 mg / L to ensure uniform suspension (settling rate < 5% / hour).

[0054] Step 2 also includes the following steps:

[0055] B1: An ultraviolet light source or a visible light source is adopted. The ultraviolet light source uses a low-pressure mercury lamp (wavelength 254 nm) or a UV-LED (365 nm), which is applicable to TiO2 quantum dots. The visible light source uses an LED (400 - 700 nm), which matches the light response range of carbon quantum dots (band gap 2.0 - 3.0 eV).

[0056] B2: The energy density is controlled by adjusting the light source power (0 - 100 W) and the distance from the surface of the silicon wafer (5 - 15 cm). The density range is 10 - 50 mW / cm 2 ;

[0057] B3: When the quantum dots are excited by light, electron-hole pairs are generated. The holes react with water to generate hydroxyl radicals (·OH), and the electrons react with oxygen to generate superoxide radicals (·O2 - ). While the radicals decompose organic substances, water molecules are reduced to produce H2 microbubbles and oxidized to produce O2 microbubbles. The above is the photocatalytic reaction mechanism;

[0058] B4: The method for regulating the size of microscale (1 - 50 μm) bubbles adopts a low energy density (10 - 20 mW / cm 2 ) + a low pulse frequency (10 - 30 Hz), and the bubble generation rate is 10 4 ~10 5 bubbles per second per cm 2 , which is applicable to stripping large particles in the initial washing stage. Nanoscale bubbles (< 100 nm) adopt a high energy density (30 - 50 mW / cm 2 ) + a high pulse frequency (50 - 100 Hz), and are generated through the transient cavitation effect, with a density reaching 10 9 bubbles / mL, which is applicable to decomposing molecular-level pollutants in the fine washing stage;

[0059] B5: The residence time of the bubbles in the tank is controlled by the circulating flow rate of the cleaning liquid (5 - 20 L / min): a high flow rate (20 L / min) reduces the bubble density, and a low flow rate (5 L / min) increases the density, in coordination with the feedback of the sensor (such as automatically increasing the flow rate to 15 L / min when the turbidity > 1 NTU).

[0060] Step three further includes the following steps:

[0061] C1: An optical scattering type turbidity sensor with a measurement range of 0 - 100 NTU and an accuracy of 0.1 NTU for real-time monitoring of the concentration of particulate pollutants (≥ 0.1 μm particles) in the cleaning liquid, a four-electrode type conductivity sensor with a measurement range of 0 - 200 μS / cm and an accuracy of 0.1 μS / cm for monitoring the concentration of metal ions, and an infrared spectrometer for on-line detection of organic pollutants (such as C-H bonds, C=O bonds) on the surface of the silicon wafer with a resolution of 4 cm -1 , and a scanning speed of 1 time per minute;

[0062] C2: When the turbidity > 0.5 NTU or the conductivity > 1 μS / cm, the system automatically increases the light source energy density by 10% - 20% and increases the quantum dot concentration by 1 - 2 mg / L. If the residual amount of metal ions detected > 1×10 10 atoms / cm 2 , the enhanced adsorption mode of surface-modified quantum dots is activated.

[0063] Step three also includes the following steps:

[0064] C3: By controlling the energy density range of the light source to be 30 - 50 mW / cm 2 , and the pulse frequency to be 30 - 50 Hz, micron-sized bubbles with a surfactant concentration of 0.3% - 0.5% are used for primary washing for 30% - 40% of the total cycle time;

[0065] C4: Then control the energy density range of the light source to be between 10 - 20 mW / cm 2 , the pulse frequency to be between 50 - 100 Hz, and nanometer-sized bubbles with a quantum dot concentration of 8 - 10 mg / L are used, and then fine washing is carried out for 60% - 70% of the total cycle time.

[0066] Step four also includes the following steps:

[0067] D1: The cleaning tank adopts an overflow system design. The specific setting method is to set an annular overflow port at the top of the cleaning tank with a flow rate of 1 - 5 L / min. The pollutants carried by the bubbles are discharged with the cleaning liquid using buoyancy. The overflow liquid enters the recovery tank after being filtered by a 0.01 μm membrane. A circulating filtration system is set in the recovery tank. The specific filtration methods include primary filtration with a 5 μm polypropylene filter element, precision filtration with a 0.1 μm polyvinylidene fluoride (PVDF) membrane, and magnetic separation (if the quantum dots are loaded with Fe3O4 magnetic cores) or an affinity chromatography column (immobilized phosphate group ligand) quantum dot recovery system. The primary filtration is used to remove large particles, and the precision filtration is used to intercept quantum dots and colloids. The quantum dot recovery rate ≥ 90%;

[0068] Step four also includes the following steps:

[0069] D2: The silicon wafer is cleaned by a rinsing method using a spray + immersion combination (spray for the first 2 minutes and immerse for the next 3 minutes), with an ultra-pure water flow rate of 10 - 15 L / min, a resistivity ≥ 18.2 MΩ·cm, and a temperature of 25 ± 2 °C. At the same time, a control system is set for the key points of rinsing. The control method is to stop when the conductivity of the rinsing water monitored online < 0.5 μS / cm, which takes 3 - 5 minutes.

[0070] Step four also includes the following steps:

[0071] D3: High-purity nitrogen with a purity of 99.999% is used to blow the surface of the silicon wafer at an angle of 45° with a pressure of 0.2 - 0.3 MPa for 1 - 2 minutes to remove the residual water droplets on the surface. For silicon wafers with a process below 7 nm, the CO2 critical drying method (temperature 31.1°C, pressure 7.38 MPa) is adopted to avoid surface damage caused by capillary force, and the residual watermark after drying is ≤ 0.01%.

[0072] In this embodiment, by using photoluminescent carbon quantum dots or metal oxide quantum dots as catalysts, reactive oxygen species are generated by photoexcitation to decompose organic pollutants on the surface of the silicon wafer. The surface of the quantum dots is modified with organic phosphoric acid groups or metal ligands to enhance the adsorption and catalytic ability for specific metal ions, reduce metal pollution, and use sensors to monitor parameters such as the turbidity and conductivity of the cleaning solution in real time, monitor the concentration of pollutants on the surface of the silicon wafer in real time through sensors, feedback and adjust the light source parameters and microbubble density to achieve dynamic optimization of the cleaning intensity, control the bubble size, combine with surfactants to enhance the impurity adsorption efficiency, and achieve directional migration of impurities through an overflow or circulation system, with strong applicability.

[0073] For silicon wafers with different degrees of contamination, the cleaning cycle and reagent concentration are automatically matched to reduce the consumption of pure water and chemical reagents.

[0074] The present invention has been generally and exhaustively described above, but based on the present invention, some modifications or improvements can be made, which are obvious to those of ordinary skill in the technical field. Therefore, modifications or improvements that do not depart from the spirit of the present invention are within the protection scope of the present invention.

Claims

1. An adaptive cleaning method for a silicon wafer with a photo-induced microbubble array, characterized in that: The described self - adaptive cleaning method for a silicon wafer with a photo - induced micro - bubble array includes the following steps: Step 1: Pre - treatment before cleaning; Step 2: Generation of the photo - induced micro - bubble array; Step 3: Self - adaptive cleaning process; Step 4: Pollutant removal and post - treatment; Step 5: Cleaning completed.

2. The adaptive cleaning method of a photo-induced microbubble array silicon wafer according to claim 1, characterized in that: The said Step 1 further includes the following steps: A1: The loading method uses a bracket made of quartz or polytetrafluoroethylene (PTFE) to fix the silicon wafer in an edge - clamping or vacuum - adsorption manner, avoiding contact with the silicon wafer surface (contact area < 1%), preventing the introduction of secondary pollution. The loading density is controlled such that the single - slot loading capacity is 10 - 25 pieces according to the silicon wafer size (such as 12 inches), and the wafer spacing is ≥2 cm to ensure uniform circulation of the cleaning liquid and bubbles; A2: Use ultrapure deionized water (resistivity ≥18.2 MΩ·cm) as the base liquid, filter out particulate impurities through a 0.01 - μm membrane, add sodium dodecylbenzenesulfonate (SDBS) or polyoxyethylene ether non - ionic surfactant, with its concentration in the range of 0.1% - 0.5% (mass ratio), and dissolve it for 30 minutes by magnetic stirring (rotation speed 300 - 500 rpm) to form a homogeneous solution. At the same time, fine - tune it with dilute nitric acid or ammonia according to the pollution type, with an accuracy of ±0.5; A3: Use glucose or citric acid with a purity ≥99.5% to react at 180 - 220°C for 4 - 6 hours by hydrothermal method, and after cooling, purify it through a 0.22 - μm filter membrane to make the raw material. The particle size is controlled by adjusting the reaction temperature (larger particle size is generated at high temperature, and nanoscale particle size is generated at low temperature), with an average particle size of 5 - 10 nm. Use tetrabutyl titanate as the precursor and ethanol as the solvent, hydrolyze and polymerize at 60 - 80°C, and make metal - oxide quantum dots with a particle size of 10 - 20 nm after annealing treatment (300 - 500°C); A4: Disperse quantum dots in ammonium dihydrogen phosphate solution (concentration 0.01 mol / L), stir at 60 °C for 12 hours, graft phosphate groups (-PO3H2) through covalent bonds to enhance the chelating ability for metal ions such as Fe 3+ and Al 3 +, or use ethylenediaminetetraacetic acid (EDTA) as a ligand to form coordination bonds with metal ions on the surface of quantum dots for metal-ligand modification of quantum dots, improving the selective adsorption of Cu 2+ . The loading method of the decorated quantum dots is to add the cleaning solution by ultrasonic dispersion (power 300 W, time 10 minutes), and the concentration is controlled at 5 - 10 mg / L to ensure uniform suspension (sedimentation rate < 5% / hour).

3. The adaptive cleaning method of a photoinduced microbubble array silicon wafer according to claim 1, wherein: The said Step 2 further includes the following steps: B1: Use an ultraviolet light source or a visible light source. For the ultraviolet light source, use a low - pressure mercury lamp (wavelength 254 nm) or UV - LED (365 nm), which is suitable for TiO2 quantum dots. For the visible light source, use LED (400 - 700 nm), which matches the light - response range of carbon quantum dots (band gap 2.0 - 3.0 eV); B2: Controlling the energy density by adjusting the light source power (0 - 100 W) and the distance from the silicon wafer surface (5 - 15 cm), with the density range being 10 - 50 mW / cm 2 ; B3: When quantum dots are excited by light, electron-hole pairs are generated. The holes react with water to form hydroxyl radicals (·OH), and the electrons react with oxygen to form superoxide radicals (·O2 - ). While the radicals decompose organic substances, water molecules are reduced to produce H2 microbubbles and oxidized to produce O2 microbubbles. The above is the photocatalytic reaction mechanism; B4: The method for regulating the size of micro-scale (1 - 50 μm) bubbles uses a low energy density (10 - 20 mW / cm 2 ) + a low pulse frequency (10 - 30 Hz), and the bubble generation rate is 10 4 - 10 5 bubbles per second per cm 2 , which is suitable for stripping large particles in the initial washing stage. Nano-scale bubbles (< 100 nm) are generated by a high energy density (30 - 50 mW / cm 2 ) + a high pulse frequency (50 - 100 Hz) through the transient cavitation effect, and the density reaches 10 9 bubbles / mL, which is suitable for decomposing molecular-level pollutants in the fine washing stage; B5: Control the residence time of bubbles in the tank by the cleaning - liquid circulation flow rate (5 - 20 L / min): High flow rate (20 L / min) reduces the bubble density, and low flow rate (5 L / min) increases the density, in cooperation with sensor feedback (such as automatically increasing the flow rate to 15 L / min when the turbidity > 1 NTU).

4. The self-adaptive cleaning method of a photogenerated microbubble array silicon wafer according to claim 1, characterized in that: The said Step 3 further includes the following steps: C1: An optical scattering type turbidity sensor with a measuring range of 0 - 100 NTU and an accuracy of 0.1 NTU for real-time monitoring of the concentration of particulate pollutants (particles ≥ 0.1 μm) in the cleaning liquid, a four-electrode conductivity sensor with a measuring range of 0 - 200 μS / cm and an accuracy of 0.1 μS / cm for monitoring the metal ion concentration, and an infrared spectrometer for on-line detection of organic pollutants (such as C-H bonds and C=O bonds) on the silicon wafer surface with a resolution of 4 cm -1 , with a scanning speed of 1 time per minute; C2: When the turbidity > 0.5 NTU or the conductivity > 1 μS / cm, the system automatically increases the light source energy density by 10% - 20% and increases the quantum dot concentration by 1 - 2 mg / L. If the residual amount of metal ions detected > 1×10 10 atoms / cm 2 , start the enhanced adsorption mode of surface-modified quantum dots.

5. The self - adaptive cleaning method of a photo - induced micro - bubble array silicon wafer according to claim 4, wherein: The said Step 3 further includes the following steps: C3: By controlling the energy density range of the light source to be 30 - 50 mW / cm 2 , and with a pulse frequency of 30 - 50 Hz, perform primary washing for 30% - 40% of the total cycle time using microbubbles with a surfactant concentration of 0.3% - 0.5%; C4: Then control the energy density range of the light source to be between 10 and 20 mW / cm 2 2, with the pulse frequency between 50 and 100 Hz, use nanobubbles with a quantum dot concentration between 8 and 10 mg / L, and then perform fine washing for 60% - 70% of the total cycle duration.

6. The self - adaptive cleaning method of a photogenerated micro - bubble array silicon wafer according to claim 1, wherein: The said Step 4 further includes the following steps: D1: The cleaning tank is designed with an overflow system. The specific setting method is to set an annular overflow port at the top of the cleaning tank, with a flow rate of 1 - 5 L / min. The pollutants carried by the bubbles are discharged with the cleaning liquid by buoyancy. The overflow liquid enters the recovery tank after being filtered by a 0.01 μm membrane. A circulating filtration system is set in the recovery tank. The specific filtration methods include primary filtration with a 5 μm polypropylene filter element, precision filtration with a 0.1 μm polyvinylidene fluoride (PVDF) membrane, and magnetic separation (if the quantum dots are loaded with Fe3O4 magnetic cores) or an affinity chromatography column (immobilized phosphate group ligand) for the quantum dot recovery system. The primary filtration is used to remove large particles, and the precision filtration is used to intercept quantum dots and colloids. The quantum dot recovery rate is ≥90%.

7. A self - adaptive cleaning method for a photogenerated micro - bubble array silicon wafer according to claim 1, characterized in that: The fourth step further includes the following steps: D2: The silicon wafers are cleaned by a rinsing method of spraying + soaking combination (spraying for the first 2 minutes and soaking for the next 3 minutes), with an ultrapure water flow rate of 10 - 15 L / min, a resistivity ≥18.2 MΩ·cm, and a temperature of 25 ± 2°C. At the same time, a control system is set for the key points of rinsing. The control method is to stop when the conductivity of the rinsing water is monitored online and <0.5 μS / cm, which takes 3 - 5 minutes.

8. An adaptive cleaning method for a photogenerated microbubble array silicon wafer according to claim 1, characterized in that: The fourth step further includes the following steps: D3: High-purity nitrogen with a purity of 99.999% is used to blow the surface of the silicon wafers at an angle of 45° with a pressure of 0.2 - 0.3 MPa for 1 - 2 minutes to remove the residual water droplets on the surface. For silicon wafers with a process below 7 nm, the CO2 critical drying method (temperature 31.1°C, pressure 7.38 MPa) is used to avoid surface damage caused by capillary force. The residual watermark after drying is ≤0.01%.

Images