A surface integrity reconstruction process for high-transparency backside-inspected silicon nitride ceramic substrates

Abstract

The application discloses a surface integrity reconstruction process for a high-transparency back-detection silicon nitride ceramic substrate, and belongs to the technical field of precise processing of ceramic substrates. The process comprises the following steps: back light transmission detection is performed on the silicon nitride ceramic substrate, an optical abnormal response area is obtained, and a position correspondence relationship is established; near-surface defect structures are identified according to abnormal area morphological characteristics; a near-surface low-density defect layer is removed by quantitative grinding; low-damage fine finishing is performed by using multi-scale gradient abrasives to reconstruct a dense, uniform and low-scattering surface layer; after cleaning and drying, secondary back light transmission detection is performed to evaluate the effect; and the processing parameters are adjusted according to the secondary detection result to form a closed-loop process. Through defect identification, quantitative removal and surface reconstruction, the application effectively reduces the abnormality of dark spots, bright spots and foggy areas in back light detection, improves the back light transmission uniformity of the substrate and the detection pass rate, and is suitable for efficient batch processing of silicon nitride ceramic substrates for power semiconductors.

Description

Technical Field

[0001] This invention belongs to the field of precision machining technology for ceramic substrates, and specifically relates to a surface integrity reconstruction process for silicon nitride ceramic substrates. Background Technology

[0002] Silicon nitride ceramic substrates are widely used in high-power, high-reliability packaging fields such as power semiconductor modules, new energy vehicle electronic control systems, photovoltaic inverters, and energy storage converters due to their high strength, high toughness, good electrical insulation, and excellent thermal conductivity. As power devices develop towards higher voltage and higher current density, higher requirements are placed on the surface integrity of the substrates and the consistency of batch testing.

[0003] Currently, the production process of silicon nitride ceramic substrates typically includes ceramic powder preparation, tape casting, high-temperature sintering, planar grinding, and polishing. During these sintering, leveling, and machining processes, various defect structures are easily generated near the surface of the substrate. These mainly include: micropore-rich layers, weakly bonded grain boundary layers, and residual glass phase segregation layers formed during sintering, as well as powdered layers, microcracks, and scratch-damaged layers introduced by machining. These near-surface defects are difficult to effectively identify using conventional visual inspection or contact surface roughness measurement. However, in backlight transmission testing, due to the refractive index difference and light scattering effect between the defect area and the dense substrate, they can form dark spots, bright spots, hazy transmission areas, or stripe-like anomalies, significantly reducing the substrate's inspection pass rate.

[0004] Traditional processing methods mainly aim to control the flatness, thickness tolerance and surface roughness of the substrate. The setting of grinding and polishing process parameters lacks specific consideration for near-surface defect structures, making it difficult to effectively remove low-density defect layers hidden in the depth range of several micrometers to tens of micrometers below the surface. This leads to repeated backlight detection anomalies, which restricts the quality consistency and yield improvement in mass production. Summary of the Invention

[0005] This invention addresses the technical problems of existing silicon nitride ceramic substrate processing methods, such as difficulty in effectively removing near-surface defects and low backlight inspection pass rates. It provides a surface integrity reconstruction process for high-transmittance backlight inspection silicon nitride ceramic substrates. By establishing a complete closed-loop processing link of defect identification, quantitative removal, surface reconstruction, and effect verification, it achieves targeted elimination of low-density defect layers near the surface and reconstruction of surface structural integrity.

[0006] This invention is achieved through the following technical solution: A surface integrity reconstruction process for a high-transmittance back-light transmission inspection silicon nitride ceramic substrate includes the following steps: performing backlight transmission inspection on the silicon nitride ceramic substrate to obtain optically abnormal response regions on the substrate surface and establishing a correspondence between the optically abnormal response regions and the positions on the substrate surface; identifying near-surface defect structures based on the morphological characteristics of the optically abnormal response regions; performing quantitative grinding on the optically abnormal response regions based on the identification results to remove the near-surface low-density defect layer; using multi-scale gradient abrasives to perform low-damage fine finishing on the ground substrate surface to reconstruct a dense, uniform, and low-scattering surface layer; cleaning and drying the finely finished substrate; performing a second backlight transmission inspection on the dried substrate to evaluate the surface integrity reconstruction effect; and adjusting the grinding removal amount, abrasive particle size, and fine finishing parameters based on the second backlight transmission inspection results to form a closed-loop processing technology with the backlight inspection pass rate as the target.

[0007] Furthermore, the near-surface defect structure includes one or more of the following: a low-density layer formed during sintering, a weakly bonded grain boundary layer, a residual glass phase enrichment layer, a micropore enrichment layer, a powdered layer, or a processing-induced damage layer formed during machining.

[0008] Furthermore, the amount of material removed from one side in the quantitative grinding process is determined based on one or more factors, including backlight transmission abnormality intensity, substrate surface roughness, and substrate thickness allowance.

[0009] Preferably, the removal amount per side is 1 μm to 80 μm; more preferably, the removal amount per side is 5 μm to 25 μm.

[0010] Furthermore, the multi-scale gradient abrasive is a diamond abrasive with a progressively smaller particle size.

[0011] Furthermore, the low-damage finishing process includes one or more combinations of precision grinding, precision polishing, chemical mechanical polishing, magnetorheological polishing, or ultrasonic-assisted polishing.

[0012] Furthermore, the backlight transmission detection uses at least one of visible light, near-infrared light, or a line light source as the detection light source.

[0013] Furthermore, the evaluation indicators for the backlight transmission detection include one or more of the following: transmission grayscale uniformity, number of local dark spots, number of bright spots, and area of ​​stripe anomalies.

[0014] Furthermore, the cleaning process includes one or more combinations of ultrasonic cleaning, pure water rinsing, weak alkali cleaning, anhydrous ethanol cleaning, and hot air drying.

[0015] A silicon nitride ceramic substrate prepared using any of the above processes has a reduced scattering anomaly region and an improved detection pass rate under backlight transmission detection.

[0016] This invention establishes a correspondence between near-surface defect structures and backlight transmission anomalies, enabling precise defect localization and targeted removal. Through the synergistic effect of quantitative grinding and multi-scale gradient abrasive finishing, the defect layer is effectively removed while avoiding over-processing and secondary damage, reconstructing a dense, continuous, and low-scattering surface integrity structure. Secondary backlight detection verification and closed-loop parameter optimization ensure the stability and consistency of batch processing results. This process not only improves the backlight transmission uniformity and detection pass rate of the substrate but also provides a good surface foundation for subsequent metallization, electroplating, and packaging processes.

[0017] Beneficial effects: This invention clarifies the correspondence mechanism between backlight detection anomalies and near-surface loose structure, and achieves targeted processing of abnormal areas such as dark spots, bright spots, and hazy areas through defect localization, avoiding the blindness of traditional processing methods.

[0018] This invention effectively reduces the density of light scattering centers on the surface and subsurface by quantitatively removing low-density defect layers near the surface and combining this with multi-scale gradient abrasive finishing, thereby significantly improving the uniformity of backlight transmission and the pass rate of substrate detection.

[0019] The reconstructed surface layer of this invention is dense and continuous, eliminating weak bonding grain boundaries and microcracks, which is beneficial to improving the bonding reliability between subsequent metallization coatings, electroplating layers, and packaging materials and the substrate.

[0020] The process of this invention can be implemented using existing grinding and polishing production lines without the need for additional special equipment. It has a wide process window and is easy to implement in mass production and with automated control. Attached Figure Description

[0021] Figure 1 This is a schematic flowchart of a surface integrity reconstruction process for a high-transparency back-inspection silicon nitride ceramic substrate according to the present invention. Figure 2 This is a schematic diagram illustrating the identification of abnormal areas in backlight transmission detection in an embodiment of the present invention; Figure 3 This is a comparison chart of the secondary backlight detection results in an embodiment of the present invention. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to embodiments, but the implementation of the present invention is not limited thereto.

[0023] Example 1: This embodiment provides a surface integrity reconstruction process for a high-transparency back-scanning silicon nitride ceramic substrate, specifically including the following steps: (1) First backlight transmission detection. The silicon nitride ceramic substrate, after sintering and initial grinding, is placed on a backlight transmission detection device and used as a light source for transmission scanning to obtain a full-area transmission optical image of the substrate. Based on the grayscale distribution in the transmission image, dark areas and hazy transmission areas are identified, the position coordinates of abnormal areas are recorded, and the positional correspondence between abnormal areas and the substrate surface is established; from Figure 2 The image clearly shows abnormal areas such as dark spots, bright spots, and hazy areas. The "bright spots" referred to in this application are... Figure 2 The dark black spots in it.

[0024] (2) Defect identification. Based on the morphological characteristics of dark areas showing a significant decrease in local light intensity and hazy areas showing a diffuse decrease in gray level in the backlight detection image, it was determined that there is a microporous enrichment layer and a loose grain boundary structure in the near-surface layer of the substrate.

[0025] (3) Quantitative grinding process. A diamond grinding disc is used to quantitatively grind the surface of the abnormal area identified in step (2). The amount of material removed on one side is controlled between 5 μm and 25 μm until the abnormal layer shown by the backlight detection is completely removed.

[0026] (4) Low-damage fine finishing. The surface of the substrate after grinding is finely finished in multiple stages using diamond abrasive with progressively smaller particle size. Coarser abrasive is used to remove the grinding scratch layer, and finer abrasive is used to reduce the surface roughness, ultimately obtaining a dense, uniform, and low-scattering finely finished surface layer.

[0027] (5) Cleaning and drying. The finely trimmed substrate is placed in the middle for ultrasonic cleaning, then rinsed with pure water, and finally dried with hot air.

[0028] (6) Secondary backlight transmission test. The dried substrate is subjected to secondary backlight transmission test under the same test conditions as in step (1) to evaluate the uniformity of transmission grayscale and count the number of local dark spots and bright spots.

[0029] In this embodiment, after the above-described process, from Figure 3 Secondary backlight testing results showed that the dark areas were significantly reduced, the hazy transmission area basically disappeared, and the uniformity of transmission grayscale was significantly improved.

[0030] Example 2: This embodiment provides a surface integrity reconstruction process for a high-transparency back-inspection silicon nitride ceramic substrate, suitable for double-sided processing of silicon nitride ceramic substrates with large thickness allowances, and specifically includes the following steps: (1) Double-sided backlight transmission detection was performed on the silicon nitride ceramic substrate with a large thickness margin after sintering. The distribution of optical abnormal response areas on the two main surfaces of the substrate was obtained, and the abnormal position and abnormal intensity level of each surface were recorded.

[0031] (2) Based on the difference in the degree of abnormality of the two surfaces, determine the asymmetric grinding removal scheme: set the single-sided removal amount of the stronger abnormal surface to a larger value, and set the single-sided removal amount of the weaker abnormal surface to a smaller value.

[0032] (3) Using a double-sided grinding equipment, the two surfaces are quantitatively ground according to the removal scheme determined in step (2).

[0033] (4) After grinding, the two surfaces are subjected to multi-scale gradient abrasive low-damage finishing.

[0034] (5) After cleaning and drying, a second backlight transmission test is performed.

[0035] In this embodiment, after asymmetric quantitative grinding and fine finishing, the backlight transmission uniformity of both surfaces of the substrate is significantly improved, the number of bright spots is greatly reduced, and the overall test pass rate meets the batch consistency requirements.

[0036] Example 3: This embodiment provides a closed-loop feedback control method for the surface integrity reconstruction process of a high-transparency back-inspection silicon nitride ceramic substrate, specifically including the following steps: (1) Taking the batch of silicon nitride ceramic substrate as a unit, set the initial grinding removal amount, abrasive particle size combination and finishing parameters as the initial processing window.

[0037] (2) Process the batch of substrates according to the initial processing window, and perform secondary backlight transmission detection on the processed substrates to obtain the transmission gray uniformity, number of dark spots, number of bright spots and stripe abnormal area of ​​each substrate.

[0038] (3) Compare the detection index obtained in step (2) with the preset pass threshold: when the detection index is better than the pass threshold, record the current processing parameters as the effective process window; when the detection index is worse than the pass threshold, adjust the grinding removal amount, such as increasing or decreasing, the abrasive particle size combination, such as refining or coarsening, and the finishing parameters, such as extending or shortening the finishing time, according to the deviation direction and degree of the detection index, to form the corrected processing parameters.

[0039] (4) Process the subsequent batches of substrates with the corrected processing parameters, and continue to verify and iteratively optimize them with the results of the secondary backlight detection.

[0040] This embodiment establishes a closed-loop feedback mechanism between backlight detection results and processing parameters, thereby achieving self-optimization of process parameters for different batches and specifications of silicon nitride ceramic substrates, effectively ensuring the quality consistency of mass production.

[0041] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make many possible variations and modifications to the technical solutions of the present invention, or modify them into equivalent embodiments, without departing from the scope of the technical solutions of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

Claims

1. A surface integrity reconstruction process for a high-transparency back-scanning silicon nitride ceramic substrate, characterized in that, Includes the following steps: Step 1: Perform backlight transmission detection on the silicon nitride ceramic substrate to obtain the optical abnormal response area on the surface of the substrate, and establish the correspondence between the optical abnormal response area and the position on the substrate surface. Step 2: Identify near-surface defect structures based on the morphological characteristics of the optical anomaly response region; Step 3: Based on the near-surface defect structure identification results, perform quantitative grinding on the optical anomaly response region to remove the low-density defect layer near the surface; Step 4: Use multi-scale gradient abrasive to perform low-damage fine finishing on the surface of the polished substrate, reconstructing a dense, uniform, and low-scattering surface layer; Step 5: Clean and dry the finely trimmed substrate; Step 6: Perform a second backlight transmission test on the dried substrate to evaluate the surface integrity reconstruction effect; Step 7: Based on the secondary backlight transmission detection results, adjust the grinding removal amount, abrasive particle size, and finishing parameters to form a closed-loop processing technology with the backlight detection pass rate as the target.

2. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The near-surface defect structure includes one or more of the following: a low-density layer formed during sintering, a weakly bonded grain boundary layer, a residual glass phase enrichment layer, a micropore enrichment layer, a powdered layer, or a processing-induced damage layer formed during machining.

3. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The amount of material removed from one side during the quantitative grinding process is determined based on one or more of the following factors: abnormal backlight transmission intensity, substrate surface roughness, and substrate thickness allowance.

4. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 3, characterized in that, The single-sided removal amount is from 1 μm to 80 μm.

5. The surface integrity reconstruction process for a high-transparency back-scanning silicon nitride ceramic substrate according to claim 4, characterized in that, The single-sided removal amount is 5 μm to 25 μm.

6. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The multi-scale gradient abrasive is a diamond abrasive with a progressively smaller particle size.

7. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The low-damage finishing process includes one or more combinations of precision grinding, precision polishing, chemical mechanical polishing, magnetorheological polishing, or ultrasonic-assisted polishing.

8. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The evaluation indicators for backlight transmission detection include one or more of the following: transmission grayscale uniformity, number of local dark spots, number of bright spots, and area of ​​stripe anomalies.

9. The surface integrity reconstruction process for a high-transparency back-scan silicon nitride ceramic substrate according to claim 1, characterized in that, The cleaning process includes one or more combinations of ultrasonic cleaning, pure water rinsing, weak alkali cleaning, anhydrous ethanol cleaning, and hot air drying.

10. A silicon nitride ceramic substrate prepared by the process described in any one of claims 1 to 9, characterized in that, The silicon nitride ceramic substrate exhibits reduced scattering anomaly areas and improved detection throughput under backlight transmission detection.

Images