Search

Using the pre-warped substrate to pack IGBT modules can increase the contact area between the substrate and the radiator after the reflow soldering. There are many factors affecting the shape of the substrate’s back after soldering, and determining the effect of each factor by experiments would cost significantly. In this paper, a FEA model of the IGBT module’s reflow soldering process is established by ANSYS to predict the back surface of the pre-warped AlSiC substrate after soldering. Initially, in order to improve the accuracy of the FEA model, the material properties of AlSiC were determined at different temperatures. Subsequently, based on the result of DSC, the solder’s phase transition was simulated in the model. The calculated results showed that the proposed FEA model could predict the shape feature of the AlSiC pre-warped substrate’s back surface after the reflow soldering, and the prediction error of substrate’s residual warpage was 46.05%. The model provides a reliable reference for the design of pre-warped substrate in IGBT module, reducing the test cost and shortening the design period substantially.

Fine finishing of tungsten alloy is required to improve the surface quality of molds and precision instruments. Nevertheless, it is difficult to obtain high-quality surfaces as a result of grain boundary steps attributed to differences in properties of two-phase microstructures. This paper presents a theoretical and experimental investigation on chemical mechanical polishing of W–Ni–Fe alloy. The mechanism of the boundary step generation is illustrated and a model of grain boundary step formation is proposed. The mechanism reveals the effects of mechanical and chemical actions in both surface roughness and material removal. The model was verified by the experiments and the results show that appropriately balancing the mechanical and chemical effects restrains the generation of boundary steps and leads to a fine surface quality with a high removal rate by citric acid-based slurry.

Ultrasonic assisted grinding (UAG) has been considered as a prominent processing method of the reaction bonded silicon carbide (RBSiC). To improve the knowledge of UAG process, both conventional grinding (CG) and UAG were used to process the RBSiC for in-depth investigation. Grinding forces, surface topographies, and subsurface damages during CG and UAG were compared. Furtherly, the ground surface was analyzed on aspects of both topographical characteristics and material removal mechanism. The results indicated that the removal of material is mainly achieved by the intersections of cracks initiated from both big SiC particles and mixture area of silicon matrix and small SiC grains. The crack propagation during UAG was more intensified due to the ultrasonic impact, which results in higher efficiency of machining RBSiC.

The slurry flow beneath the wafer in chemical mechanical polishing (CMP), involving the chemical reaction and the lubrication, is critical to the planarity and surface quality of a large-sized silicon wafer. In order to analyse the effects of pad roughness and some important operating parameters on the slurry flow with the suspended abrasives between the wafer and the pad, a complicated three-dimensional model based on the micropolar fluid theory, Brinkman equations and Darcy's law is developed. The effects of pad roughness and vital parameters on the slurry flow between the pad and the wafer are well discussed.

A thin subsurface damage layer is often formed in the ground layer of a silicon wafer. Compressive stresses exist in the damage layer of the wafer and cause the wafer to deflect when it is unloaded from the vacuum chuck. The residual stresses play an important role in evaluating the machining quality of a wafer. A concise yet accurate stress-deflection relationship is desired so that the residual stresses can be calculated from the wafer deflection. The theoretical equation based on minimizing the total strain energy is often used. However, the anisotropic effects of silicon are neglected and would incur errors when it is applied on silicon wafers. This study establishes an empirical equation via a finite element (FE) model. Both equations are verified in experiments. It is found that FE model is more accurate than the theoretical equation.

Flatness measurements of thin and large substrates are greatly affected by gravity. In this work, a new method was proposed for accurately measuring the flatness error of large and thin silicon wafers using a floating technique. In this method, a silicon wafer was immersed in a liquid that had a density slightly lower than that of silicon and was supported by three ball supporters in the liquid. The wafer shape was measured using a laser triangulation sensor. The buoyancy force on the wafer inside the liquid could cancel out most of the effect of gravity. The residual effect of gravity was further removed by subtracting the residual deflection calculated by finite-element modeling from the measured result. The method was validated by comparing the flatness values of a thick wafer measured using this method and via a commercially available interferometer. The deformations of wafers ground by #600 and #2000 diamond wheels were measured and the measurement error was less than 3% of the total deformation.

The influences of the polishing slurry composition, such as the pH value, the abrasive size and its concentration, the dispersant and the oxidants, the rotational velocity of the polishing platen and the carrier and the polishing pressure, on the material removal rate of SiC crystal substrate (0001) Si and a (0001) C surface have been studied based on the alumina abrasive in chemical mechanical polishing (CMP). The results proposed by our research here will provide a reference for developing the slurry, optimizing the process parameters, and investigating the material removal mechanism in the CMP of SiC crystal substrate.

(1 1 1), (1 1 0) Cd0.96Zn0.04Te and (1 1 1) Cd0.9Zn0.1Te semiconductor wafers grown by the modified vertical Bridgman method with dimensions of 10 mm × 10 mm × 2.5 mm were lapped with a 2–5 µm polygonal Al2O3 powder solution, and then chemically mechanically polished by an acid solution having nanoparticles with a diameter of around 5 nm, corresponding to the surface roughnesses Ra of 2.135 nm, 1.968 nm and 1.856 nm. The hardness and elastic modulus of (1 1 1), (1 1 0) Cd0.96Zn0.04Te and (1 1 1) Cd0.9Zn0.1Te single crystals are 1.21 GPa, 42.5 GPa; 1.02 GPa, 44.0 GPa; and 1.19 GPa, 43.4 GPa, respectively. After nanocutting is performed by the Berkovich nanoindenter, the surface roughness Ra of the (1 1 1) Cd0.9Zn0.1Te single crystal attains a 0.215 nm ultra-smooth surface. The hardness and elastic modulus of three kinds of CdZnTe single crystals decrease with the increase of indentation load. When the nanoindenter departs the surface of the crystals, the adherence effects are obvious for the three kinds of single crystals. This is attributed to the plastic sticking behavior of CdZnTe material at a nanoscale level. When the indentation load of the three kinds of CdZnTe single crystals is in the range of 4000–12 000 µN, the adhered CdZnTe material on the nanoindenter falls onto the surface and accumulates around the nanoindentation.