Papers by Keyword: CMP

Abstract: A novel approach for processing SiC wafers has been developed to grind then polish 150 and 200mm SiC wafers without lapping. The purpose of this work was to optimize the processing of SiC wafers sliced from boules to finished epi-ready wafers by grinding and chemical-mechanical polishing (CMP). Diamond vitrified wheels were used for coarse and fine grinding to correct the irregular shape of SiC wafers before reducing surface roughness by CMP. 4H-SiC wafers were sliced by diamond embedded/slurry wire saw and laser split techniques. Incoming wafer condition was seen to affect coarse grinding wheel performance depending on incoming surface roughness and shape. Wheel characteristics, including abrasive size, abrasive concentration, and bond structure, were adjusted to improve grinding efficiency based on incoming conditions. Coarse grinding wheels were able to reduce wafer total thickness variation to 3-5um and average surface roughness to 20-30nm (Ra). Fine grinding wheels were optimized to reduce total thickness variation (TTV) below 2um and surface roughness to 1-2nm Ra and peak-to-valley height of 20-30nm (Rt). Coarse and fine wafering time was less than 30 minutes total to remove 50 microns on both Si and C-face per wafer. Surface damage from grinding was removed after one hour of polishing each wafer by CMP, achieving surface roughness of 0.4nm Ra and 5-7nm Rt. The benefit of optimizing coarse and fine grinding of 150 and 200mm SiC wafers is demonstrated by producing flat wafers, which reduced overall processing time to prepare an epi-ready condition by CMP.

Abstract: We carried out chemical mechanical polishing (CMP) on commercially available 6 inch SiC wafers (epi-ready products) with slurries containing different abrasive types and evaluated the latent scratch density from the mapping measurement of the wafers using mirror projection electron microscope (MPJ). Comparing to the wafer before polishing, the latent scratch density decreased on the wafer polished with MnO₂+KMnO₄, while that increased by polishing with Al₂O₃+KMnO₄. The two-step polishing using first Al₂O₃+KMnO₄ and then SiO₂+H₂O₂ can reduce the latent scratch density to the same level as that with MnO₂+KMnO₄, but long polishing time is required because of the low polishing rate in the process with SiO₂+H₂O₂. We investigated the reason why MnO₂ slurry can suppress the occurrence of latent scratches by a polishing test on a wafer with an SiO₂ film on its (0001)Si surface. The results suggest the oxidation of the SiC surface is rate-determining step for polishing with MnO₂+KMnO₄. It was also found that wafers without an SiO₂ film could not be polished with only MnO₂ abrasives. Thus the mechanical contribution to polishing by MnO₂ abrasives in KMnO₄-based slurry is smaller than the chemical contribution, which can suppress the occurrence of latent scratches. KMnO₄-based slurry containing MnO₂ abrasives performs the CMP process with low latent scratch density in a time shorter than that containing Al₂O₃ or SiO₂ abrasives.

Abstract: In chemical mechanical planarization (CMP) processes, ceria is generally used as the abrasive . After the CMP process, many ceria particles adhere to the wafer surface and must be removed prior to subsequent processing. In this study, the effect of varied viscosity was investigated during the buffing CMP process for ceria particle removal. After contaminating the wafer surface with ceria slurry, the ceria particles were removed through the buffing CMP process. The wafer surface was analyzed through inductively coupled plasma mass spectrometry (ICP-MS) to confirm cleaning efficiency. The ICP-MS data showed that, as buffing CMP solution viscosity increased, cleaning efficiency improved. These results suggest that increasing the viscosity of the buffing CMP solution improves its effectiveness in removing ceria particles.

Abstract: The sapphire substrate have a significant impact on reducing the light scattering loss and in accordance with the crystalline surface and can improve the light emission efficiency is the surface roughness of the sapphire LED is an important factor. When the CMP process after measuring the transmittance, haze, and surface roughness had a transmittance of 84% or more transmittance at all wavelengths, regardless of the thickness it was confirmed that the haze also has a low value, regardless of the thickness. Was confirmed that the surface roughness is greater the thicker the thickness of the sapphire it affects the pressure of the CMP polishing and then take a lot of importing a low surface roughness value of the thickness of the sapphire CMP process the surface roughness value.

Abstract: This paper presents a combined diamond disk that is designed and manufactured to markedly improve the leveling of diamond tips, thereby reducing the amount of diamond grit. First, a small brazed diamond disk was manufactured. Second, 12-or 24-disk brazed diamond disks were mounted onto a substrate with a diameter of 108 mm for completing the combined diamond disk. Four types of disks were fabricated, and their performances were compared with that of a conventional diamond disk. The experimental results showed that the combined diamond disk has more complete diamond crystal shapes and distribution with better leveling, which increases the effectiveness of the working diamond grits. Compared to a conventional diamond disk, the combined diamond disk achieved a higher wafer removal rate and better uniformity while consuming less pad material. The number of diamond grits required was significantly lower. Roughly, 7,600 and 12,000 diamond grits were used for the 24-and 12-disk brazed diamond disks, respectively, in the new disk, whereas 20,000 diamond grits were used in a conventional diamond disk. In the case of the conventional diamond disk, the diamond tips are leveled to more than 50 μm. However, in the case of the combined diamond disk, the diamond tips can be leveled to less than 30 μm because the diamond tips are already leveled. These results contribute to the understanding of conditioning techniques and further improvement of the chemical mechanical polishing process.

Abstract: The growth of graphene on 3C-SiC/Si heterostructure is a promising approach, which provides low production cost, high scalability and easiness of nanoelectromechanical system fabrication. However, the quality of graphene is still insufficient for device applications due to mediocre morphological and structural quality of the 3C-SiC epilayers compared to bulk SiC crystals and to excessive Si out-diffusion from the Si substrate. Here, we propose a solution of inserting a 4H-AlN layer between 3C-SiC and Si, which allows us to polish the 3C-SiC film without worrying about enhancement of the Si out-diffusion despite the thinning after the polishing. With this insertion, a considerable quality improvement is achieved in our graphene on silicon.

Abstract: Single crystal diamond is widely used in high-tech fields for its remarkable performance on mechanics, calorifics, optics, acoustics, etc. High-quality diamond surface with small roughness and low scathe are required in these applications. However, the extreme hardness and high chemical inertness of diamond result in severe processing difficulties. Chemical mechanical polishing (CMP) is a promising processing method which can obtain super-smooth and low-damage diamond surface. Oxidant is a key issue for CMP of single crystal diamond. In this study, five different oxidants were used to polish diamond samples. The results indicated that Fenton reagent was an appropriate CMP oxidant and a super-smooth diamond surface of Ra 2.4 nm was achieved by using Fenton reagent in CMP.

Abstract: In commercial CMP tools, slurry is applied near the pad center. As the pad rotates, more than 95% of the fresh slurry flows directly off the surface due to bow wave formation and inertial forces without ever entering the pad-wafer interface, resulting in low slurry utilization [1]. Furthermore, some slurry that manages to go under the wafer stays on the pad, mixes with fresh slurry and re-enters the pad-wafer interface. This used slurry contains reaction products, foam and pad debris (due to pad conditioning) that cause wafer-level defects [2]. Such defect-causing by-products keep recirculating on the pad during polishing and accumulate near the retaining ring over time. Also, since large amounts of DI water are used between wafer polishes to rinse off the debris and reaction products, appreciable amounts of water may stay on the pad and inside the grooves. When fresh slurry is introduced to polish the next wafer, it mixes with the residual water and is diluted, resulting in lower material removal. As such, the current slurry application method does not provide efficient slurry utilization and leaves significant room for improving defect levels. Moreover, the constant sweeping of the conditioner arm during in-situ conditioning results in uneven slurry distribution and introduces additional challenges when it comes to carrier multi-zone pressure control for reduced within-wafer removal rate non-uniformity.

Abstract: Surface with nanometer accuracy is required to manufacturing process of integrated circuit (IC) devices. One of the most promising techniques for surface planarization is chemical mechanical polishing (CMP). CMP is a high efficiency process, both due to the varied chemical and physical properties of the surface materials. Conventional CMP uses the slurry, which is composed of abrasive particles suspended in a chemical solution. The abrasive particle is effective for surface planarization by a chemical mechanism with little mechanical abrasion. So chemical reactivity appears to be an essential factor of CMP process. In our conventional study, fullerenols have been proposed as suitable abrasives for copper CMP. The chemical reactivity of fullerenol is suggested by the fact that high removal rate (150 nm/min) and surface flatness (0.6 nm RMS) have been confirmed using fullerenol slurry. In this study, we analyzed the chemical reactivity between fullerenol molecule and copper surface, which is important to understand the material removal mechanism. Using the intrinsic Raman spectroscopic signal of interaction between fullerenol and copper with surface plasmon resonance (SPR), the chemical reactivity over a period of the reaction process was analyzed. Raman spectroscopy is commonly used in chemical analysis, since vibrational information is specific to the chemical bonds and symmetry of molecules. Therefore, it provides a fingerprint by which the molecule can be identified. However, spontaneous Raman spectroscopic signal is typically very weak, and as a result the main difficulty of Raman spectroscopy is separating the weak inelastically scattered light from the intense Rayleigh scattered laser light. Then Raman spectroscopic signal in this study is further enhanced by the SPR, also known as surface enhanced Raman scattering (SERS). The increase in intensity of the Raman spectroscopic signal for adsorbates on copper surface occurs because of an enhancement in the electric field provided by the surface. This technique enables high-sensitive analysis in the near-surface region. The signature of copper-oxygen bond was measured by Raman spectroscopy for fullerenol/copper system by in-process SERS analysis. It is thought to be caused by the hydroxyls of fullerenol molecule adsorbed on the copper surface. This result suggests that fullerenol molecules absorbing onto the copper surface affect the high efficient material removal.

Abstract: During the process of chemical mechanical planarization (CMP) of copper, benzotriazole (BTA) is the most commonly used inhibitor in the slurry. Though the corrosion inhibition mechanism has been studied widely, the mechanism of BTA layer on copper surface in CMP slurries should be further investigated. In this paper, the adsorption mechanisms of BTA were studied by static corrosion tests. Besides, the surface composition was measured by XPS. Combining with CMP experiments, the material removal mechanism of copper CMP depending on pH values was investigated. It was found that the formation of passive film, consisting of Cu-BTA complex, adsorption of BTA and copper oxides, played a dominant role under acidic conditions. While the surface film composed of adsorption layer of BTA and copper oxides under alkaline conditions. The inhibition mechanism of BTA varied with pH values, resulted in corresponding changes of material removal rate and coefficients of friction.