A laser slicing technology is proposed as a new high precision slicing method for semiconductor substrates so as to realize slicing thin wafer with small material loss. In this method, appropriate focusing of laser beam can introduce micro cracks inside the material, which are connected by series of laser scanning. Observation in the laser slicing of single crystal sili con wafer under various process conditions revealed mechanism of this method. A crack was propagated along the cleavage face {111} and {110}. The sliced face was created by linkage of the cleavages. In addition, the silicon wafer of 1 -inch was successfully sliced, and uniform sliced surface could be obtained. However, wafer warpage was generated by residual stress.

Aspheric lenses are used in various optical instruments. However, it has complicated surface shape and it is difficult to process. Therefore, in this research, we tried developing a processing method to product the lens directly from the glass using laser. Laser slicing process is performed using an ultrashort pulse laser having transparency to glass. When a continuous process-affected layer is formed inside by the laser, a crack propagates in process-affected layer and then, the glass is peeled off with mirror surface. As a result, three-dimensional processing is realized by forming Process-affected layer three-dimensionally. It was also found that the Process-affected layer can be removed by annealing.

The availability of ultraviolet-excitation polishing for 4H-SiC wafer and the chemical state of the 4H-SiC surface polished using photocatalyst and cathilon were investigated in absence and presence of an ultraviolet (UV). X-ray absorption fine structure (XAFS) spectroscopy conducted the qualitative analysis of the polished 4H-SiC surface. This analysis was used to clarify the oxides that are formed/removed by decomposition of cathilon dye and water during the polishing of 4H-SiC using TiO₂ slurry, cathilon slurry and TiO₂-cathilon slurry (mixed slurry), all of which included diamond particles. In polishing the previously sandblasted and handlapped 4H-SiC, TiO₂ slurry provides large polishing efficiency and large polished surface roughness, and cathilon slurry provides small polishing efficiency and small polished surface roughness under UV irradiation. The large polishing efficiency, small polished surface roughness and large roughness decrease rate were attained using mixed slurry that is suitable to polish 4H-SiC under UV irradiation. The differences of XAFS spectrum are observed in UV absence/presence among three kinds of slurries. The XAFS analysis mainly investigates and indicates that some oxides form on the surface polished using cathilon and mixed slurries in UV absence/presence. Polishing using mixed slurry provides less volumes of oxides inside SiC on the polished surface, particularly.

Tool life depends on not only the physical and chemical properties of tool material, and the combination of tool and work material, and cutting conditions, but also depends on dynamic characteristics of machine tool such as dumping performance and dynamic stiffness. In this paper, tool life was investigated the influence by controlling vibration direction and dynamic stiffness of the tool using parallel leaf spring structure jig. Two types of parallel leaf spring structure jigs were prepared in which dynamic stiffness can be adjusted in thrust force direction and principal force d irection. Then, tool life test was performed continuous cutting and intermittent cutting of the outer circumferential turning in titanium alloy. Cutting speed was 150 m/min, depth of cut was 0.2 mm, and feed rate was 0.1 mm/rev. Lubrication was not applied . As a result, the dynamic stiffness in thrust force direction affects self-excited chatter, and clearance angle varies with the vibration in thrust force direction, so that the progress of the flank wear becomes faster. Dynamic stiffness in principal force direction has little influence on self-excited chatter, and vibration in principal force direction does not occur, so the flank friction speed does not fluctuate and the progress of flank wear does not become fast. Thus, it was suggested that tool life is different depending on the difference in dynamic stiffness in thrust force direction and principal force direction.

Intermittent cutting is prone to vibrations by periodic cutting force, so tool life is affected by dynamic characteristics of the machine tool and the tool. Then, intermittent cutting is a problem of tool damage such as chipping and flaking. The cause is sharp rise of the cutting force and the impact force due to the collision between the tool and the workpiece. In this paper, the leaf spring structure milling tool which is easy to deform in the principal force direction was manufactured to absorb the impact force during intermittent cutting. And the effect on tool life was investigated. Tool life test was performed down cut with dry cut. Cutting speed was 80 m/min, axial depth of cut was 1.0 mm, radial depth of cut was 5.0 mm, and feed rate was 0.115 mm/tooth. As a result, chipping occurred when the flank wear width exceeded 0.2 mm in the conventional tool with high stiffness. On the other hand, chipping did not occur even if leaf spring structure milling tool with low stiffness in principal force direction had flank wear width exceeded 0.2 mm. However, notch wear has increased. The leaf spring structure milling tool has tool life of approximately 2 times longer than the conventional tool. In the leaf spring structure milling tool, the rise time of cutting force is delayed by deforming the leaf spring by cutting force. Therefore, the impact force is absorbed and chipping does not occur. It was suggested that the chipping resistance was improved by the leaf spring structure milling tool and tool life was extended.

This paper describes the evaluation of the inner shape of three-dimensional small glass holes and elucidation of the machining mechanism using statistical analysis based on designed experiment. The glass hole made by rotary ultrasonic machining had a 1mm diameter, and 4 characteristics to evaluate which were expansion of hole size, inner surface roughness, distortion, and direction of distortion. We assumed an elliptical cross-sectional hole shape and distortion, stipulating that an elliptical shape was a more accurate index than a true circle shape to evaluate hole shape and research machining mechanism. Furthermore the replica method used was both accurate and cost efficient enough to evaluate those characteristics. First we did a “Cross Array” experiment to construct models of each characteristic. In this designed experiment, tool factor and mode factor were set in outer array to identify any high order interaction with the machining factors set in inner array. Next we analyzed the correlation of those characteristics and executed analysis of variance. Finally we evaluated hole shape, discussed the machining mechanism and recognized the importance of main effect of the tool and its interactions with feed speed and feed mode. In conclusion, our approach to evaluating three-dimensional hole shape are useful.

In this paper, we propose a system for estimating reading content in a booklet from an indoor surveillance camera. Here, we assume that a reading content can be specified by estimating the following two; which booklet/page, and which region in the page. We implement a reading booklet/page estimation method based on image search, and propose a reading region estimation method based on the slight difference of body pose of the reader. We evaluated the method as a 44 classes classification problem, which consists of eleven pages of booklets and four regions in each pages. We confirmed that the system achieved 25.6% in accuracy of the reading content estimation in our evaluation setting.

Manufacturing precision electronic parts, it is judged by automatic inspection machines whether there are defects in the electronic components or not. Along with the requirement for higher sensitivity of inspection, some defect candidates judged by the inspection machine include many good products, which cause excessive detection of non-defective products. This study aims to develop an effective method classifying defect candidate images into defect images and non-defect images. In this paper, improvement of defect identification performance by dimensional reduction focusing on localization of HOG feature is evaluated, and the results showing defect identification performance are reported.