Yuanhang
Fan
a,
Zhenyu
Zhang
*a,
Jiaxin
Yu
*b,
Xingqiao
Deng
*c,
Chunjing
Shi
d,
Hongxiu
Zhou
e,
Fanning
Meng
d and
Junyuan
Feng
d
aState Key Laboratory of High-performance Precision Manufacturing, Dalian University of Technology, Dalian 116024, China. E-mail: zzy@dlut.edu.cn
bSchool of Manufacturing Science and Engineering, Southwest University of Science and Technology, Mianyang 621010, China. E-mail: yujiaxin@swust.edu.cn
cSchool of Mechanical and Electrical Engineering, Chengdu University of Technology, Chengdu 610059, China. E-mail: dengxingqiao19@cdut.edu.cn
dSchool of Mechanical Engineering, Hangzhou Dianzi University, Hangzhou 310018, China
eSchool of Energy and Power Engineering, Dalian University of Technology, Dalian 116024, China
First published on 12th December 2023
High-performance devices of quartz glass demand an atomic surface, which induces a challenge for chemical mechanical polishing (CMP) with a high material removal rate (MRR). Moreover, traditional CMP usually employs toxic and corrosive slurries, leading to the pollution of the environment. To overcome these challenges, a novel green photocatalytic CMP is proposed. In the CMP, SiO2@TiO2 core–shell abrasives were developed, and the CMP slurry included the developed abrasives, sodium carbonate, hydrogen peroxide and sorbitol. After photocatalytic CMP, the surface roughness Sa of quartz glass is 0.185 nm, with a scanning area of 50 × 50 μm2, and the MRR is 8.64 μm h−1. To the best of our knowledge, the MRR is the highest on such a big area of atomic surface for quartz glass. X-ray photoelectron spectroscopy reveals that SiO2@TiO2 core–shell abrasives were used as photocatalysts motivated by simulated solar light, generating electrons and holes and producing hydroxyl radicals through hydrogen peroxide. As a result, OH− could combine with Si atoms on the surface of quartz glass, forming Si–OH–Si bonds. Then the formed bonds were removed based on the balance between chemical and mechanical functions. The proposed CMP, developed SiO2@TiO2 abrasives and slurry provide new insights to achieve an atomic surface of quartz glass with a high MRR.
Fortunately, Xiaoguang Guo et al.11 found that the degree of hydroxylation on the surface of quartz glass affects the formation of interfacial chemical bonds, subsequently impacting the removal of surface atoms. When photocatalysts are stimulated by light, they generate electrons and holes, which react with oxidants and OH− to produce hydroxyl radicals. These radicals can promote the formation of a Si–OH bond on surface quartz glass. Wang et al.22,23 employed silica sol as the abrasive and nanoscale TiO2 as the photocatalyst in the process called photogenerated-catalytic CMP on GaN chips. Remarkably, this approach resulted in an almost twofold increase in the material removal rate (MRR) compared to that of conventional CMP techniques. In recent years, the adoption of photocatalytic-assisted chemical mechanical polishing (PCMP) has been widely applied for various materials including diamond, GaN (gallium nitride), and SiC (silicon carbide). Furthermore, materials like (CeO2 + PS), (CeO2 + SiO2), TiO2, and their counterparts have found extensive use as photocatalysts in PCMP.24,25 However, despite its success in various materials, there are still limited reports regarding the application of PCMP in the polishing of quartz glass.
Herein, a green and environmentally friendly polishing solution was developed, composed of Na2CO3, H2O2, tartaric acid, abrasive particles, and deionized water. In particular, spherical SiO2@TiO2 nanoparticles as abrasives were synthesized by a hydrothermal method. The spherical SiO2@TiO2 nanoparticles were then applied in the PCMP experiment of quartz glass. Through single factor variable experiments, the optimal slurry components were projected to inquire into the effects of diverse abrasives under the conditions of photo-catalytic assisted CMP, and the optimal ratio was obtained. In addition, the polishing mechanism was analyzed by XPS, FTIR, and Raman spectroscopy to prove the importance of photocatalysts in the polishing process.
Rough polishing significantly reduced the Sa of the quartz glass and improved the efficiency of the overall polishing process. The rough polishing solution comprised 3 wt% SiO2 abrasives (500 nm), Na2CO3, 0.5 wt% sorbitol, 0.5 wt% H2O2, and DI·H2O. Na2CO3 was employed to regulate the pH of the solution to 9. During the solution preparation, the solution was continuously stirred to ensure uniform distribution. The sample quartz glasses underwent polishing for 60 min under the following experimental conditions: 100 rpm polishing speed, 18 kPa polishing pressure, and 5 mL min−1 flow rate. Following the completion of the rough grinding process, the sample surfaces were thoroughly washed with copious amounts of DI·H2O and promptly dried using compressed air. Subsequently, the surface roughness Sa of the quartz glass was measured to be 1.008 nm for over a 50 × 50 μm2 measurement area.
The precision polishing process was carried out using a single-factor controlled variable method, with variations in the type of abrasive grains. The precision polishing solution comprised 1 wt% mixed abrasives (the ratio of SiO2 to TiO2 is 0:1, 1:1, 3:2, 4:1, and 1:0), Na2CO3, 0.5 wt% sorbitol, 0.5 wt% H2O2, and DI·H2O. Sorbitol was employed as a dispersant in the polishing solution. Na2CO3 was utilized to regulate the pH of the polishing solution to 9. To ensure a homogeneous distribution of the solution, the slurries were continuously stirred throughout the fine polishing preparation process. The sample quartz glasses underwent precision polishing for 15 min under the following experimental conditions: 80 rpm polishing speed, 18 kPa polishing pressure, and 5 mL min−1 flow rate. The experimental conditions of polishing are shown in Table 1.
Polishing solution composition | pH | Polishing pressure (kPa) | Polishing speed (rpm) | Polishing time (min) | Average optical power density (Sun) | Polishing solution flow rate (mL min−1) | |
---|---|---|---|---|---|---|---|
a Where 1 Sun is equivalent to 1000 W m−2 solar constant. What's more, the standard value of the solar constant is approximately 1361 W m−2. | |||||||
Rough polishing | SiO2 abrasive, Na2CO3, sorbitol, H2O2, and DI·H2O | 9 | 18 | 100 | 60 | 0 | 5 |
Precision polishing | Mixed/SiO2@TiO2 abrasives, Na2CO3, sorbitol, H2O2, and DI·H2O | 9 | 18 | 80 | 15 | 15 | 5 |
Throughout the entire precision polishing process, a light source (HXF300-T3, Beijing Zhongjiao Jinyuan Technology Co., Ltd, China) was employed to activate the polishing solution, providing an average optical power density of 15 Sun. Following the completion of the fine grinding process, the sample surfaces were thoroughly washed using a generous amount of DI·H2O and promptly dried using compressed air.
(1) |
In addition, the sample quartz glasses were analyzed using X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, America) after undergoing different treatments. The treatments included immersing the quartz glasses in the polishing solution containing SiO2@TiO2 synthesized abrasives for 7 days, as well as precision polishing using the polishing solution containing SiO2@TiO2 synthesized abrasives. All the sample quartz glasses were processed under photocatalytic conditions. Furthermore, Raman spectroscopy (Horiba LabRAM HR Evolution, Japan) was performed to analyze the sample quartz glasses before and after polishing. Finally, FTIR (Thermo Scientific Nicolet iS20, USA) was employed to assess whether the quartz glasses were contaminated after the precision polishing process.
SEM, TEM, XRD, and FTIR were employed to analyze the surface appearance, core–shell structure, and the crystal structure of the composite SiO2@TiO2 abrasive particles. Fig. 3 presents the SEM and TEM images and diameter distribution plots of the SiO2 and SiO2@TiO2 abrasive synthesized by the Stöber method. The SEM images (Fig. 3(a) and (d)) and TEM images (Fig. 3(b) and (e)) depict the surface appearance and internal structure of the SiO2 and SiO2@TiO2 abrasive. Twenty randomly selected particles were measured to determine the diameter distribution, resulting in average diameters of 301 ± 34 nm and 387 ± 16 nm for the SiO2 and SiO2@TiO2 abrasive particles, respectively, as illustrated in Fig. 3(c) and (f). Analysis of the SEM and TEM images and diameter distribution plots reveals that the synthesized SiO2 abrasives reveal a smooth spherical morphology and a uniform particle size distribution, making them well-suited for serving as the core component in the composite SiO2@TiO2 abrasive. Additionally, the TEM images demonstrate the distinctive core–shell structure of the composite SiO2@TiO2 abrasive particles.
Fig. 3 The SEM and TEM images and diameter distributions of the synthesized abrasive particles: (a)–(c) SiO2; (d)–(f) SiO2@TiO2. |
The XRD pattern of the composite SiO2@TiO2 abrasive particles is illustrated in Fig. 4(a). An evident characteristic peak is observed at approximately 23° in the SiO2 spectrum, indicating the high purity of the synthesized SiO2. Furthermore, distinct diffraction peaks are observed in the spectrum at 2θ values of 25.4°, 37.9°, 48°, 54.0°, 62.1°, and 62.7°, corresponding to the crystal planes (101), (004), (200), (105), (211), and (204) of the anatase TiO2 phase, respectively.30 According to the latest literature,31 anatase TiO2 has higher photocatalytic activity than other forms which has been verified. The FTIR image of SiO2@TiO2 is illustrated in Fig. 4(b). SiO2@TiO2 abrasive particles exhibit absorption bands at 1082 cm−1, 798 cm−1, 779 cm−1, and 695 cm−1. The absorption band at 1082 cm−1 primarily corresponds to the vibration of the silicon–oxygen (Si–O) bond, while the absorption band at 798 cm−1 represents the bending vibration of the Si–O bond. The absorption bands at 779 cm−1 and 695 cm−1 are associated with the vibration of the titanium–oxygen (Ti–O) bond. The XRD and FTIR characterization studies demonstrate the high purity of the composite SiO2@TiO2 abrasive particles.
The XPS full-spectrum and fine-spectrum images of SiO2 and SiO2@TiO2 composite particles are shown in Fig. 5. In the full-spectrum scan, characteristic peaks of SiO2 can be observed. The peaks at 533 eV, 285 eV, and 103 eV correspond to O 1s, C 1s, and Si 2p, respectively, as shown in Fig. 5(a). As for the SiO2@TiO2 composite particles, the peaks at 532 eV and 284 eV correspond to O 1s and C 1s,32 while the peaks at 458 eV and 103 eV correspond to Ti 2p and Si 2p.33Fig. 5(b)–(d) depict the high-resolution scans of the Ti, Si, and O elements, respectively. In Fig. 5(b), the Ti 2p spectrum of SiO2@TiO2 reveals two peaks at binding energies of 458.55 eV and 464.55 eV, corresponding to the Ti 2p3/2 and Ti 2p1/2 of Ti atoms in TiO2, respectively. In Fig. 5(c), the high-resolution scan of the O element in SiO2 demonstrates that the O 1s peak is located at 532.33 eV. The high-resolution scan of the O element in SiO2@TiO2 exhibits three peaks at 529.89 eV, 530.98 eV, and 532.71 eV, which are assigned to the O atoms in Ti–O–Ti bonds, –OH groups, and Si–O–Si bonds in SiO2, respectively.34 Additionally, as shown in Fig. 5(c), the O 1s peak in the SiO2–TiO2 spectrum is slightly shifted to 530.98 eV, compared to 532.22 eV of SiO2. This shift may happen because of the difference in the chemical environment which was produced by the bonding of surface hydroxyl groups. In Fig. 5(d), the high-resolution scan of the Si element in SiO2 reveals that the main peak of Si 2p is located at 102.85 eV. Furthermore, the Si 2p spectrum of SiO2@TiO2 exhibits two peaks at binding energies of 103.27 eV and 103.80 eV, corresponding to the Si 2p3/2 and Si 2p1/2 of Si atoms in SiO2, respectively. The peak at 103.27 eV represents the Si atom in the Si–O–Ti bond.35 These results indicate a strong bond between SiO2 and TiO2, supporting the successful preparation of a large amount of composite abrasive SiO2@TiO2 for precision polishing in the experiment.
The 3D surface profilometer image of the polished surfaces is presented in Fig. 6(g)–(l). To objectively assess the surface quality of quartz glasses after the precision polishing process under photocatalytic conditions, the surface roughness Sa was measured by randomly selecting 5 points on the sample surface within a 50 μm × 50 μm scanning range. The average roughness values of the samples processed with mixed abrasives were 0.632 ± 0.025 nm, 0.372 ± 0.035 nm, 0.279 ± 0.035 nm, 0.483 ± 0.034 nm, and 0.524 ± 0.015 nm, respectively. Additionally, the average MRR values of the five sample groups were 91.617 ± 12.910 nm min−1, 123.839 ± 13.333 nm min−1, 159.091 ± 12.606 nm min−1, 132.424 ± 10.303 nm min−1, and 125.607 ± 7.576 nm min−1, respectively. It can be observed that the polishing effect is the best and the polishing efficiency is the highest when the mixing ratio of commercial SiO2 (300 nm) and TiO2 (100 nm) is 3:2. Fig. 7 shows the Ra and MRR of the quartz glasses polished with mixed commercial abrasives.
An optical microscope, a 3D optical surface profiler and an AFM were used to observe the surface of quartz glasses after the precision polishing process used the core–shell SiO2@TiO2 abrasives. The 3D surface profilometer image of the polished surface is shown in Fig. 6(l). Within a scanning range of 50 μm × 50 μm, five random points on the sample surface were selected to measure the Sa, which achieved a minimum value of 0.185 nm, indicating the smoothest and most uniform surface. Furthermore, the average MRR of the samples was 144.639 ± 17.445 nm min−1. The surface morphology after polishing is depicted in Fig. 6(f), demonstrating a smooth quartz glass surface without scratches or pits. Fig. 8(a) and (b) present the 2D and 3D morphological images of the AFM measurements on the surface of quartz glass polished with SiO2@TiO2 composite abrasives within a scanning range of 1 × 1 μm2. The obtained Ra value was 0.148 nm and Rq value was 0.192 nm. What's more, the Rq represents the root mean square average of the longitudinal coordinates of the roughness profile. It can be observed that the surface roughness is the lowest when using the composite SiO2@TiO2 abrasives, but the MRR is slightly lower compared to the MRR of the commercially mixed abrasives with a ratio of 3:2. This indicates that the synthesized core–shell SiO2@TiO2 abrasive particles exhibit superior surface flatness processing performance compared to the commercially mixed abrasives. Moreover, the core–shell SiO2@TiO2 abrasive particles offer the highest polishing accuracy and show great potential in the field of ultra-precision manufacturing.
Fig. 8 AFM 2D and 3D morphology patterns of the quartz glass surface after precision polishing using SiO2@TiO2 composite abrasive grains. |
The differences in mechanical friction behavior among single-abrasives, mixed-abrasives, and composite core–shell abrasives can be explained from a physical perspective by considering variations in hardness. The quartz glass is composed of 99.9% SiO2.36 In the case of single-abrasives, TiO2 exhibits a relatively low hardness of 6–6.5 when used for abrasives. Consequently, TiO2 is unable to rapidly remove the reaction layer, resulting in relatively poor mechanical polishing efficiency. During the polishing process, TiO2 may become embedded in the reaction layer, resulting in the formation of numerous surface pits. SiO2 possess a higher hardness of 7 compared to TiO2, thereby exhibiting stronger mechanical removal capabilities. SiO2 abrasives can effectively remove the reaction layer and promote chemical reactions on the newly exposed surface. When SiO2 and TiO2 are mixed in different proportions as abrasive particles for polishing, under photocatalytic conditions, the combined use of these two abrasives can yield a smoother surface that surpasses the performance achievable with single-grain abrasives alone.
When two different types of abrasive particles are combined, their roles in the polishing process undergo a transformation. SiO2@TiO2 composite abrasive particles, which are a physical combination of two types of abrasive particles, exhibit distinct functionality during the polishing process. According to ref. 37, the quartz glass undergoes hydrolysis to generate ˙OH radicals, which react with the surface to form silanol groups. Under photocatalytic conditions, the outer shell of nano-sized TiO2 acts as a highly active catalyst, demonstrating the highest catalytic activity on fused quartz substrates.38 When TiO2 is used as a catalyst, based on the band theory, photoexcited electron–hole pairs and the corresponding photocatalytic reduction reactions occur when the energy of the incident photons equals or exceeds the band gap. This leads to the excitation of TiO2 from the valence band to the conduction band, resulting in the generation of free electrons.39,40 Simultaneously, positively charged holes are formed in the valence band, leading to the generation of more ˙OH radicals,36 as shown in eqn (2) and (3).
TiO2 + hν → h+ + e− | (2) |
H2O + h+ → ˙OH + H2 | (3) |
Hydroxylation under alkaline conditions forms the basis for enhancing the surface quality of quartz glass. Meanwhile, H2O2 in the polishing solution can promote the dehydroxylation reaction of silanol structures,41,42 further oxidizing Si–OH to form Si–O–Si bonds. Furthermore, the positively charged holes generated in the valence band of TiO2 can undergo oxidation reactions with H2O2 molecules and hydroxide ions (OH−) adsorbed on the surface of catalyst particles in the polishing solution, resulting in the production of more hydroxyl radicals ˙OH with strong oxidative properties. The generated ˙OH radicals react with the surface of quartz glass, leading to the formation of additional silanol structures. This process can be represented by using eqn (4) and (5). Under photocatalytic conditions, and based on eqn (2) and (5), it can be inferred that TiO2, acting as a catalyst, enhances the chemical hydroxylation reaction, thereby resulting in the formation of a thicker reaction layer on the sample.
H2O2 + hν + OH− → ˙OH + H2O | (4) |
˙OH + SiO2 + H2O → Si(OH)40 | (5) |
To investigate the removal mechanism of SiO2@TiO2 composite abrasive particles, XPS full-spectrum scans and Raman spectra were obtained for quartz glass samples soaked in the polishing solution for 7 days and subjected to precision polishing, as shown in the Fig. 9. The Raman spectra were recorded and are presented in Fig. 9(b). The quartz glass surfaces before and after polishing exhibited clear absorption peaks at 460 cm−1, 605 cm−1,36 794 cm−1, and 1050 cm−1. Among them, the band at 460 cm−1 corresponds to the bending vibration peak of the silicon–oxygen bond, while the bands at 605 cm−1 and 1070 cm−1 correspond to the stretching vibration peaks of the silicon–oxygen bond. The band at 794 cm−1 corresponds to the rotational vibration peak of the silicon-oxygen tetrahedron. Additionally, the surface of the quartz glass sample soaked in the polishing solution for 7 days exhibited a notable absorption peak at 2950 cm−1, corresponding to the stretching vibration peak of the O–H group.
The XPS full-spectrum scans of the quartz glass samples, which were soaked in the polishing solution for 7 days and then subjected to precision polishing, have been recorded and are presented in Fig. 9(c) and (d). Following a 7-day immersion in the polishing solution, the surface of the quartz glass sample displays a prominent peak at 1071 eV in Na 1s,32 along with main peaks at 532 eV and 284 eV, corresponding to O 1s and C 1s, respectively. The main peak at 458 eV corresponds to Ti 2p, with a lower peak value, signifying that TiO2 plays a role in polishing as both a catalyst and an abrasive, without involvement in other chemical reactions that result in new substances. After precision polishing, the main peak corresponding to Na 1s nearly disappeared, suggesting the successful removal of the sodium-containing substance that had formed on the surface of the quartz glass due to the presence of SiO2@TiO2 composite abrasives during soaking in the polishing solution. In the XPS full-spectrum scan of the quartz glass sample after fine polishing, the main peaks at 154 eV and 103 eV correspond to Si 2s and Si 2p, respectively. It is known that SiO2 is susceptible to corrosion under alkaline conditions, resulting in the formation of soluble SiO32− ions.43 After precision polishing, there is a notable increase in the Si content on the surface of the silica sample. The Si 1s XPS spectra of quartz glass samples, both before immersion in the polishing slurry and after precision polishing, are presented in Fig. 9(c). In the polished quartz glass sample, the peaks at 102.77 eV and 101.21 eV were identified as corresponding to SiO2 and Na2SiO3, respectively. However, after the polishing process, only a single main peak at 102.83 eV, attributed to Si from SiO2, was observed in the fitted curve. The O 1s XPS spectra of quartz glass, both before immersion in the polishing slurry and after polishing, are presented in Fig. 9(d). The peaks at 531.03 eV, 531.95 eV, and 532.89 eV, observed in the XPS spectra of quartz glass immersed in the polishing slurry, are likely associated with Na2CO3, Na2SiO3, and SiO2, respectively.28 After polishing, only a single main peak at 532.01 eV, attributed to SiO2, was observed in the fitted XPS curve. This observation suggests the successful removal of Na2CO3 and Na2SiO3 species that were generated on the surface of quartz glass after immersion in the polishing slurry.19 Based on the XPS analysis, we can infer that the following reaction occurs during the PCMP process,19,32 leading to the oxidation of Si–OH to Si–O–Si:
Na2CO3 + H2O → 2Na+ + HCO3− + CO2 | (6) |
HCO3− + H2O → H2CO3 + OH− | (7) |
SiO2 + OH− → (SiO3)2− + H2O | (8) |
To demonstrate the environmentally friendly nature of the polishing solution, FTIR characterization was performed on the polished quartz glass samples, as illustrated in Fig. 10. The polished quartz glass surface exhibited absorption bands at 1140 cm−1, 970 cm−1, and 775 cm−1. The bands observed at 1140 cm−1 and 970 cm−1 correspond to the asymmetric stretching vibrations of silicon–oxygen bonds, while the band at 970 cm−1 corresponds to the asymmetric bending vibration of silicon–oxygen bonds. The characterization results indicate the absence of impurities on the surface of the polished quartz glass after polishing, demonstrating the feasibility of visible light-assisted catalytic chemical mechanical polishing in the field of ultra-precision machining.
Based on the aforementioned analysis, a material removal mechanism for polishing quartz glass under photocatalytic conditions using SiO2@TiO2 composite abrasives is proposed, as illustrated in Fig. 11. Initially, driven by external load and kinetic energy, the SiO2@TiO2 abrasives exhibit rapid motion and experience fast friction with the quartz glass surface, leading to impact-induced protrusion collapse and generation of frictional heat. This mechanical effect accelerates the molecular movement of chemical components in the polishing solution, initiating a series of chemical reactions. First, the original fused quartz glass surface reacts with the aqueous solution to form silanol. Subsequently, the silanol reacts with OH− generated in the alkaline environment provided by the Na2CO3 aqueous solution, resulting in the formation of SiO32−. Simultaneously, under the photocatalytic conditions, the TiO2 shell of the SiO2@TiO2 composite abrasives is excited by using a xenon lamp light source, leading to the generation of additional hydroxyl radicals on the fused quartz glass surface, thereby accelerating the reaction rate and forming a thicker chemical reaction layer. Moreover, H2O2 facilitates the dehydrogenation reaction of silanol, further generating Si–O–Si. During the PCMP process, the reaction-generated silicates undergo elastic deformation due to the well-ordered structure of SiO2@TiO2 nanocomposite abrasives, and they are precisely removed under the compression of the polishing pad and quartz glass. Finally, the resultants are discharged along with the flowing polishing solution. After several iterations of this PCMP cycle, the surface precision of the quartz glass eventually reaches the sub-nanometer level.
Fig. 11 The material removal mechanism for polishing quartz glasses using SiO2@TiO2 composite abrasives under photocatalytic conditions. |
This journal is © The Royal Society of Chemistry 2024 |