Differential centrifugation method for separating α-Al2O3 nanoparticles based on morphology response differences
By eliminating irregularly shaped particles through a multi-round differential centrifugation process, the morphology selectivity and particle size consistency control of α-Al2O3 nanoparticles were achieved, solving the problem of insufficient morphology control in the existing technology and improving the performance stability of ceramic sintering and SiC-CMP.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MICRO-NANO ADVANCED MATERIALS (BEIJING) CO LTD
- Filing Date
- 2025-11-05
- Publication Date
- 2026-05-26
AI Technical Summary
Existing differential centrifugation technology has difficulty effectively controlling the morphology of α-Al2O3 nanoparticles, resulting in uneven grain growth and pore aggregation during sintering, which affects the strength of the ceramic structure and the surface quality consistency of CMP materials.
A multi-round differential centrifugation process based on morphological response differences was adopted, including high-speed centrifugation rejection, ultrasonic redispersion, and medium-speed centrifugation enrichment. Particles were identified by morphological response differences in the sedimentation path, and an operation sequence of 'high-speed rejection → redispersion → medium-speed enrichment' was constructed to achieve synergistic control of morphological selectivity and particle size consistency of α-Al2O3 nanoparticles.
It significantly improves the purity and particle size uniformity of equiaxed particles, enhances the end-use performance and stability of ceramic sintering and SiC-CMP applications, improves particle morphology purity and particle size uniformity, and reduces surface defect density and grain size deviation.
Smart Images

Figure CN121342059B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of nanomaterial classification and purification technology, specifically involving a fine separation method for α-Al2O3 nanoparticles based on the difference in morphological response of particles in the sedimentation path in a centrifugal field as the identification basis, and employing a multi-round differential centrifugation structured process of "high-speed centrifugation rejection - ultrasonic redispersion - medium-speed centrifugation enrichment". It is suitable for application scenarios with high morphological selectivity and particle size consistency requirements (such as high-performance ceramic sintering, SiC-CMP polishing slurry, etc.). Background Technology
[0002] α-Al₂O₃ nanoparticles have wide applications in high-performance ceramics, wafer chemical mechanical polishing (CMP) slurries, and ceramic matrix composites due to their excellent thermal, mechanical, and dielectric properties. Currently, differential centrifugation technology is mainly used to control the particle size distribution of nanoparticles, but its effectiveness in separating and controlling particle morphology, especially equiaxed or near-spherical particles, is still immature.
[0003] Current differential centrifugation processes mostly focus on controlling particle size distribution, while there is a lack of systematic research on particle morphology selectivity. Existing technologies often fail to significantly remove irregularly shaped particles, leading to uneven grain growth and porosity aggregation during sintering due to chaotic particle packing patterns, ultimately resulting in decreased ceramic structural strength and performance fluctuations. Especially in ceramic substrates and CMP material applications requiring high density and high uniformity, equiaxed or near-spherical α-Al2O3 particles have higher value for finished product control.
[0004] The specific limitations of existing technologies are as follows: the current classification mainly relies on particle size differences and commonly uses single-wheel medium-speed differential centrifugation. This is insufficient for removing long strips / eccentric non-equiaxed particles with L / W>1.2, which leads to an increase in the risk of sintered crystal boundary distortion, pore defects, CMP scratches, local wear, etc., making it difficult to guarantee the consistency of final performance and surface quality.
[0005] While multi-stage centrifugation schemes have been developed to enhance particle size accuracy, they still lack morphology screening dimensions. Therefore, there is an urgent need for a new separation technology that simultaneously possesses morphology recognition and particle size consistency control. Summary of the Invention
[0006] To address the problems mentioned in the background art, this invention proposes a differential centrifugation method for α-Al2O3 nanoparticles based on morphological response differences. Unlike the traditional approach of relying on "differences in particle size settling velocity," this method uses the morphological response differences along the settling path as the identification criterion and constructs a multi-round differential centrifugation process of "high-speed rejection → redispersion and reconstruction → medium-speed enrichment." This improves the purity and particle size consistency (30–70 nm, σ / μ ≤ 15%) of particles with L / W ≤ 1.2, achieving synergistic control over the morphological selectivity and particle size consistency of α-Al2O3 nanoparticles. Consequently, it enhances the final performance and stability of applications such as ceramic sintering and SiC-CMP.
[0007] The technical solution adopted by the present invention to achieve the above objectives is as follows.
[0008] A differential centrifugation method for separating α-Al₂O₃ nanoparticles based on morphological response differences includes the following steps: S1 Particle dispersion step: α-Al₂O₃ nanoparticles are dispersed in anhydrous ethanol at a concentration of 0.55 wt%, and the particles are continuously dispersed using an ultrasonic device to form a stable suspension; S2 High-speed centrifugation step: The suspension is centrifuged at 10,000–15,000 rpm, the lower precipitate is discarded, and the supernatant is retained to remove particles with larger diameters or irregular morphologies; S3 Ultrasonic redispersion step: The obtained supernatant is... The liquid is again dispersed using ultrasonication under the same conditions as in step S1 to break up potential agglomerates and reconstruct the particle dispersion state; S4 medium-speed centrifugation step: the dispersed system is centrifuged at a medium speed of 4000-7000 rpm, and the lower precipitate is collected as part of the target particle product; S5 circulation and drying step: the combined process of steps S2-S4 is repeated 2-3 times to achieve an optimized balance between morphological purity, particle size uniformity and process efficiency; all precipitates obtained from the medium-speed centrifugation stage are combined and dried to obtain the target α-Al2O3 nanoparticles.
[0009] Preferably, the ultrasonic parameters in steps S1 and S3 are a frequency of 40kHz, a power of 150W, and a duration of 30 to 60 minutes.
[0010] Preferably, in step S2, the centrifugation speed is 12000rpm±3000rpm and the centrifugation time is 2 to 5 minutes.
[0011] Preferably, the centrifugation time in step S4 is 10 to 20 minutes.
[0012] Preferably, the proportion of particles with L / W ≤ 1.2 in the target α-Al2O3 nanoparticles is not less than 80%, and this proportion can be obtained by statistically analyzing not less than 100 particles using TEM images and ImageJ software.
[0013] Preferably, the target α-Al2O3 nanoparticles have a particle size distribution between 30 and 70 nm, and the standard deviation of the particle size distribution does not exceed 15% of the average particle size. This data is based on the statistical analysis results of 100 particles in TEM images.
[0014] The method of this invention is based on the difference in morphological response of particles during sedimentation in a centrifugal field to construct an identification mechanism, which is fundamentally different from traditional screening methods that rely solely on differences in particle size and sedimentation velocity. The mechanism of action of this invention can be summarized in three stages: structural stability identification → dynamic attitude sieving → enrichment and purification.
[0015] Specifically, due to their geometric symmetry, equiaxed particles have stable sedimentation paths, controllable postures, and uniform stress in a medium-speed centrifugal field, making it easy to form a dense sedimentation layer. Non-equiaxed particles (such as elongated or eccentric particles) exhibit dynamic instability behaviors such as tumbling, turning, and lateral displacement in a centrifugal field, resulting in large fluctuations in sedimentation paths and easy deviations in direction, making it difficult to achieve stable enrichment. The multi-round "high-speed-redispersion-medium-speed" combined process allows large particles or agglomerates to be removed first in the high-speed stage, followed by ultrasonic redispersion to break down weak agglomerates, and finally, precise screening based on sedimentation stability in the medium-speed stage.
[0016] This continuous feedback control mechanism forms a closed-loop path of "structural stability identification - dynamic attitude screening - enrichment and purification", which significantly improves separation efficiency and structural consistency control capabilities, and supports the innovation and reproducibility of this invention from a mechanistic perspective.
[0017] Compared with the prior art, the present invention has the following beneficial effects:
[0018] (1) Morphology control effect: The proportion of equiaxed particles with a length-to-length ratio L / W≤1.2 in the target particles obtained by this method is more than 80%, which is significantly higher than the result obtained by the traditional single-wheel medium-speed centrifugation process. This shows that the present invention can effectively remove long strip or eccentric morphological particles, significantly improve the purity of the target morphology, and provide more consistent raw materials for subsequent applications.
[0019] (2) Particle size uniformity performance: The target particles obtained by this method are mainly concentrated in the range of 30~70nm, and the standard deviation of particle size distribution is controlled within 15% of the average value. This invention not only has the selective advantage in morphology screening dimension, but also achieves a high level of control in particle size distribution uniformity, and has dual regulation capability.
[0020] (3) Synergistic Enhancement Mechanism: Compared with the traditional single-round medium-speed centrifugation method, the present invention significantly improves the morphology enrichment efficiency and particle size uniformity through a structured multi-round combination process of "high-speed pre-screening – redispersion – medium-speed enrichment", while effectively reducing the residue of large particles and agglomerates. This method achieves synergistic optimization of morphological purity and particle size consistency, forming a positive feedback mechanism between the two. The "high-speed → redispersion → medium-speed" operation process of the present invention is based on the morphological response difference mechanism of the particle sedimentation path in the centrifugation field, which is fundamentally different from the traditional classification principle that only relies on particle size difference, forming an irreplaceable operation sequence and selective identification path; the structured process of "high-speed → redispersion → medium-speed" is not a simple superposition of parameters, but generates a synergistic effect of morphology-particle size integration in the multi-round centrifugation process, thereby achieving the synchronous improvement of particle size consistency and morphological purity.
[0021] (4) When the α-Al2O3 particles obtained in this invention are used for chemical mechanical polishing (CMP) of SiC wafers, they exhibit a higher polishing removal rate, lower surface roughness and less surface defect density compared with particles obtained by traditional single-stage centrifugation methods.
[0022] (5) When the α-Al2O3 particles obtained by the present invention are used in ceramic sintering, the resulting ceramic material can have higher density, more concentrated grain size distribution and better structural consistency.
[0023] (6) The present invention has equipment versatility. The high-speed and medium-speed centrifugation operations in this method can be completed on a conventional adjustable speed laboratory benchtop or industrial centrifuge. The high-speed section has a speed of 10,000 to 15,000 rpm, and the medium-speed section has a speed of 4,000 to 7,000 rpm. The whole method does not depend on special equipment structure or environmental conditions. Attached Figure Description
[0024] Figure 1 : A schematic diagram of the steps of a preferred embodiment of the present invention.
[0025] Figure 2 TEM morphology image of the particles obtained in Example 1.
[0026] Figure 3 TEM morphology of the particles obtained in Comparative Example 1.
[0027] Figure 4 Comparison of particle size distribution curves between Example 1 and Comparative Example 1. Detailed Implementation
[0028] like Figure 1As shown, a schematic flowchart of a preferred embodiment of the present invention is presented. α-Al₂O₃ nanoparticles are dispersed in anhydrous ethanol at a concentration of 0.55 wt%. The particles are continuously dispersed using ultrasound at a frequency of 40 kHz and a power of 150 W for 30–60 minutes to form a stable suspension. Step S2: High-speed centrifugation. The suspension is centrifuged at 10,000–15,000 rpm for 2–5 minutes. The lower precipitate is discarded, and the supernatant is retained to remove larger or irregularly shaped particles. Step S3: Ultrasonic redispersion. The obtained supernatant is then subjected to the same process as in step S1. Ultrasonic dispersion under the same conditions is used to break up potential agglomerates and reconstruct the particle dispersion state; S4 medium-speed centrifugation step: the dispersed system is centrifuged at 4000-7000 rpm for 10-20 minutes, and the lower precipitate is collected as part of the target particle product, enriching particles with a length-to-short axis ratio (L / W) ≤ 1.2; S5 circulation and drying step: the combined process of steps S2-S4 is repeated 2-3 times to achieve an optimized balance between morphological purity, particle size uniformity and process efficiency, and all precipitates obtained from the medium-speed centrifugation stage are combined and dried to obtain the target α-Al2O3 nanoparticles. Figure 1 The combined process of "high-speed centrifugation removal → ultrasonic redispersion → medium-speed centrifugation enrichment" is demonstrated. This process is repeated 2-3 times in sequence to gradually remove large and irregularly shaped particles and enrich equiaxed particles with stable sedimentation paths. Specific embodiments and control experiments are as follows.
[0029] Example 1
[0030] Step S1 Raw materials and dispersion operation: α-Al2O3 powder with an initial particle size of 20–200 nm was selected, with equiaxed particles (L / W≤1.2) accounting for about 40%, and dispersed in anhydrous ethanol at a concentration of 0.55 wt%. The ultrasonic conditions were 40 kHz and 150 W, and the dispersion was carried out continuously for 40 min to form a stable suspension.
[0031] Step S2 First high-speed centrifugation pre-screening: The suspension is centrifuged at 12000 rpm for 3 minutes, the lower sediment is discarded, and the supernatant is retained to remove particles with larger diameter or irregular shape.
[0032] Step S3: Redispersion: After ultrasonic redispersion under the same conditions as in step S1, potential agglomerates are broken and the particle dispersion state is reconstructed.
[0033] Step S4 First medium-speed centrifugation enrichment: Centrifuge the dispersed system at 6000 rpm for 15 min and collect the lower precipitate.
[0034] Step S5: Cycling and Drying: Perform two more rounds of the "high-speed-redispersion-medium-speed" process, combine all the precipitates obtained in the medium-speed centrifugation stage, and obtain the target α-Al2O3 nanoparticles after drying.
[0035] Results analysis: such as Figure 2 As shown, TEM images were taken from 100 samples. Approximately 80% of the samples had an L / W ratio of ≤1.2. These were typical equiaxed particles (L / W≈1.1), with particle sizes concentrated between 30 and 70 nm and a standard deviation of <15%.
[0036] Comparative Example 1
[0037] Conditions and process: The raw materials, dispersion system and parameters are the same as in Example 1, except that a single-wheel medium-speed centrifugation (6000 rpm, 20 min) is used, without high-speed pre-screening and redispersion steps.
[0038] Results analysis: such as Figure 3 As shown, TEM images were taken from 100 samples. The proportion of particles with L / W ≤ 1.2 was about 60%, with typical elongated particles (L / W ≈ 1.4). There were many elongated and eccentric particles. The particle size range expanded to 30–120 nm, the standard deviation increased significantly, and agglomeration was obvious.
[0039] Table 1. Particle proportion data for Example 1 and Comparative Example 1 in different L / W ranges.
[0040]
[0041] Data statistics were performed based on ≥100 particles in TEM images, combined with ImageJ for ellipse fitting analysis. The analysis showed that in the particles obtained by the method of the present invention (Example 1), the proportion of equiaxed particles reached 80%, while that of the traditional method (Comparative Example 1) was only about 60%. The data in Table 1 can further support the advantage of the "high-speed-redispersion-medium-speed" process in the present invention for screening target morphology particles.
[0042] Combined Figure 4 The particle size distribution ranges of Example 1 and Comparative Example 1 are shown. The particles obtained by the method of the present invention are mainly distributed in the 30-70 nm range, with the standard deviation controlled within 15% of the average value. The distribution is concentrated and the size consistency is good. In contrast, the particle size distribution range of Comparative Example 1 is wider and the deviation is larger. This figure and Table 1 form synergistic evidence, demonstrating the dual advantages of particle size control and morphology screening of the method of the present invention.
[0043] The separated particles obtained in Example 1 and Comparative Example 1 were applied to SiC wafer CMP: two polishing solutions with a particle concentration of 5 wt% and a pH of 7.0 ± 0.5; other components were the same; polishing conditions: pressure 3 psi, rotation speed 60 rpm, flow rate 150 mL / min, and time 5 min.
[0044] Table 2. Performance comparison of the separated particles obtained in Example 1 and Comparative Example 1 in SiC wafer CMP.
[0045]
[0046] Equiaxed particles (L / W≤1.2) exhibit higher distribution stability in the polishing slurry, forming uniform contact and force field at the polishing interface, significantly reducing scratches and localized stress concentration on the surface. Uniformly shaped particles significantly improve the removal rate and reduce defect density. As shown in Table 2, the particles prepared using the method of this invention can increase the polishing removal rate of SiC wafers by approximately 20%, reduce surface roughness (Ra) by approximately 17%, and reduce defect density by approximately 33%. This indicates that the α-Al₂O₃ particles obtained by this method can achieve higher polishing efficiency and surface uniformity in a neutral pH CMP system, exhibiting stable and repeatable performance. This meets the stringent process requirements for third-generation semiconductors such as silicon carbide and is feasible for direct industrialization.
[0047] The separated particles obtained in Example 1 and Comparative Example 1 were applied to ceramic sintering.
[0048] Table 3. Performance comparison of the separated particles obtained in Example 1 and Comparative Example 1 in SiC wafer CMP.
[0049]
[0050] Morphology and particle size consistency directly affect particle packing density and grain growth uniformity. Equiaxed particle systems can form a uniform and dense grain boundary structure during sintering, inhibiting abnormal grain growth. Table 3 shows that ceramic samples prepared using the particles of this invention exhibit a 2.2% increase in density, a 50% decrease in grain size standard deviation, and a more concentrated grain distribution. The particles obtained by this invention can significantly improve sintering density and inhibit grain coarsening, indicating that particle structure consistency directly translates into improved macroscopic performance. This performance improvement helps to enhance the mechanical properties and thermal stability of electronic ceramics, heat sink substrates, and high-strength structural ceramics, meeting the requirements of high-performance ceramic industrialization and satisfying the material uniformity requirements of 5G devices, power modules, and precision equipment manufacturing.
[0051] Furthermore, in functional ceramic composite materials, slurries, and coating systems, the morphology and particle size uniformity determine their dispersion uniformity and interfacial bonding strength. The particles of this invention exhibit excellent dispersibility and morphological stability, making them suitable as highly uniform fillers in high thermal conductivity composite materials, wear-resistant coatings, and precision spraying systems, improving the isotropy and long-term stability of the materials. This application requires no changes to existing equipment; simply replacing the powder raw materials can improve product stability and yield, demonstrating high engineering applicability and market promotion value.
[0052] Therefore, the α-Al2O3 nanoparticles obtained by the method of this invention exhibit significant process advantages and structural performance improvements in key fields such as CMP polishing fluids, high-performance ceramic materials, and functional composite systems.
[0053] Example 2
[0054] Conditions and procedures: Basically the same as in Example 1, except that the high-speed centrifugation conditions in step S2 are adjusted to 13000 rpm and 3 min; and the medium-speed centrifugation conditions in step S4 are adjusted to 5000 rpm and 15 min, and the cycle is repeated for two rounds.
[0055] Results analysis: The proportion of particles with L / W ≤ 1.2 is close to 85%; the particle size is concentrated in the range of 30–70 nm, and the standard deviation is further narrowed. This indicates that the method still has stable selectivity under different combinations of rotation speed parameters, demonstrating strong process window tolerance and suitability for fine-tuning industrial parameters.
[0056] Example 3
[0057] Conditions and procedures: Basically the same as in Example 1, except that the raw materials were replaced with three batches of α-Al2O3 powder with equiaxed particles (L / W≤1.2) accounting for 39%, 42%, and 37% respectively.
[0058] Results analysis: After separation, the target α-Al2O3 nanoparticles obtained from the three batches of powder had L / W≤1.2 ratios of 87%, 85%, and 88%, respectively. The results from different batches of raw materials were basically consistent, all of which could stably obtain an equiaxed particle ratio and concentrated particle size distribution of ≥85%, indicating that the method has good repeatability and batch-to-batch stability and can be directly applied to mass production.
[0059] In addition, the method described in this invention has good versatility and industrial adaptability in terms of equipment selection and process implementation. It can be directly applied to existing laboratories and industrial production systems without the need for special equipment or special conditions. The specific details are as follows.
[0060] (1) Equipment type and compatibility
[0061] The high-speed and medium-speed centrifugation operations required by this method can be performed on conventional adjustable-speed benchtop centrifuges or vertical industrial centrifuges. The centrifuge equipment used does not depend on specific configurations, rotor types, or refrigeration systems, nor does it require ultra-high pressure or special material components. Therefore, it can be implemented directly on common laboratory equipment platforms, facilitating rapid deployment by research institutions and manufacturing enterprises.
[0062] (2) Critical operating speed range
[0063] High-speed operation: 10,000–15,000 rpm, used for rapid removal of large particles and agglomerates; medium-speed operation: 4,000–7,000 rpm, used for enriching equiaxed particles. This range covers a wide range and can meet the requirements of most mainstream small and medium-sized centrifuges on the market. Depending on the batch size and initial particle distribution, parameters can be fine-tuned within this range to achieve optimal separation results.
[0064] (3) Potential for scale-up and continuous implementation
[0065] The "high-speed → redispersion → medium-speed" process structure of this invention is clear and the steps are cyclical, possessing good potential for large-scale production. Through an automated control system, continuous operation of multiple cycles of centrifugation and redispersion can be achieved, allowing the method to smoothly transition from laboratory scale to pilot-scale and industrial production line stages.
[0066] (4) Equipment selection recommendations and compatibility
[0067] For the laboratory stage: a 15,000 rpm adjustable speed centrifuge (such as a high-speed refrigerated centrifuge or a benchtop microcentrifuge) is recommended; for the pilot and industrial stages: vertical centrifuges with diverse rotor configurations or continuous feed centrifuges can be used; for ultrasonic redispersion equipment: only a standard 40kHz, 150W ultrasonic disperser is required, without the need for a special transducer system. All of the above equipment are standardized models and can be directly obtained from the commercial supply chain; no special customization is required.
[0068] Therefore, this invention also has the following advantages in terms of equipment use: no reliance on dedicated equipment – it can be implemented directly using standard laboratory and industrial equipment; wide range of operating parameters – adaptable to various equipment specifications and particle types; cyclical and easily scaled-up process – with the potential for continuous and automated implementation; low implementation cost – no additional hardware investment required, which is conducive to promotion and application.
[0069] In summary, this method not only possesses various verifiable advantages in terms of technical effectiveness, but also aligns with industrialization in terms of equipment and engineering adaptation, enabling its rapid integration into existing material preparation and refining processes, and thus demonstrating high industrial practical value.
Claims
1. A differential centrifugation method for separating α-Al₂O₃ microparticles based on morphological response differences, characterized in that, Includes the following steps: S1 Particle Dispersion Step: α-Al2O3 nanoparticles are dispersed in anhydrous ethanol at a concentration of 0.55wt%, and the particles are continuously dispersed using an ultrasonic device to form a suspension. S2 High-speed centrifugation step: The suspension is centrifuged at 10,000 to 15,000 rpm for 2 to 5 minutes, the lower precipitate is discarded and the supernatant is retained; S3 Ultrasonic redispersion step: The obtained supernatant is dispersed again by ultrasound under the same conditions as in step S1. S4 medium-speed centrifugation step: Centrifuge the dispersed system at a medium speed of 4000-7000 rpm and collect the lower precipitate as part of the target particle product; S5 Cycling and Drying Steps: Repeat the combined process of steps S2-S4 2-3 times, combine all the precipitates obtained in the medium-speed centrifugation stage, and obtain the target α-Al2O3 nanoparticles after drying.
2. The separation method according to claim 1, characterized in that, In steps S1 and S3, the ultrasonic parameters are 40kHz frequency, 150W power, and 30-60 minutes.
3. The separation method according to claim 1, characterized in that, In step S2, the centrifugation speed is 12000 rpm ± 3000 rpm.
4. The separation method according to claim 1, characterized in that, The centrifugation time in step S4 is 10 to 20 minutes.
5. The separation method according to claim 1, characterized in that, The proportion of particles with L / W ≤ 1.2 in the target α-Al2O3 nanoparticles is not less than 80%.
6. The separation method according to claim 5, characterized in that, The target α-Al2O3 nanoparticles have a particle size distribution between 30 and 70 nm, and the standard deviation of the particle size distribution does not exceed 15% of the average particle size.