We have developed separation by bonding Si islands (SBSI) process for advanced CMOS LSI applications. In this process, the Si islands that become the SOI regions are formed by selective etching of the SiGe layer in the Si/SiGe stacked layers, and those are bonded to the Si substrate with the thermal oxide layers by furnace annealing. The etching selectivity for SiGe/Si and surface roughness after the SiGe etching were found to be improved by decreasing the HNO3 concentration in the etching solution. The thicknesses of the fabricated Si island and the buried oxide layer also became uniform by decreasing the HNO3 concentration. In addition, it was found that the space formed by SiGe etching in the Si/SiGe stacked layers was able to be filled with the thermal oxide layer without furnace annealing.

The role of electric field in metal-induced lateral crystallization (MILC) of amorphous Si (a-Si) under limited Ni-supply condition has been investigated. The nominal lateral-growth rate was increased from 3.6 µm/h (no-electric field) to 23 µm/h at the positive electrode side and reduced to 2.8 µm/h at the negative electrode side in presence of the electric field of 20 V/cm. However, spontaneously nucleated needle-like Si crystals were observed in the enhanced positive electrode side, which have been found to be independent of the MILC. Further investigation under the condition where Ni in the supply region was removed on the way of crystallization revealed that the electric field enhanced crystallization greatly reduced. These results indicate that the electric field does not enhance the MILC growth but enhances the diffusion of Ni in a-Si which takes place prior to the MILC growth.

We propose a sea surveillance system that automatically detects intruding objects in the sea. The difficulty with an automatic system is detecting objects such as moving boats while reducing false positives caused by some waves and reflections in the sea. A false positive is reporting an object which doesn't actually exist, while a false negative is a failure in detecting an intruding object. Firstly, we identify factors of false positives. Secondly, we propose a new surveillance system considering these factors. Our proposed system combines three detecting methods. The first method is detection of Differences between Surveillance images and Flapping Reference images (DSFR). The second method is detection of Contours from Averaging images (CA). The third method is Silhouette object Detection (SD). The combination of DSFR and CA detects various moving objects under normal light conditions, while SD detects objects under backlight conditions. Finally we apply our proposed method to actual situations. Our proposed method detected boats while reducing false positives effectively.

This paper presents a VLSI design methodology for the MAC-level DWT/IDWT processor based on a novel limited-resource scheduling algorithm. The r-split Fully-specified Signal Flow Graph (FSFG) of limited-resource FIR filtering has been developed for the scheduling of the MAC-level DWT/IDWT signal processing. Given a set of architecture constraints and DWT parameters, the scheduling algorithm can generate four scheduling matrices that drive the data path to perform the DWT computation. Because the memory for the inter-octave is considered with the register of FIR filter, the memory size is less than the traditional architecture. Besides, based on the limited-resource scheduling algorithm, an automated DWT processor synthesizer has been developed and generates constrained DWT processors in the form of silicon intelligent property (SIP). The DWT SIP can be embedded into a SOC or mapped to program codes for commercial off-the-shelf (COTS) DSP processors with programmable devices. As a result, it has been successfully proven that a variety of DWT SIPs can be efficiently realized by tuning the parameters and applied for signal processing applications.

A multifunctional Boolean logic circuit composed of single-electron transistors (SETs) was fabricated and its operation demonstrated. The functions of Boolean logic can be changed by the half-period phase shift of the Coulomb-blockade (CB) oscillation of some SETs in the circuit, and an automatic control based on a feedback process is used to attain an exact shift. The amount of charges in the memory node (MN), which is capacitively coupled to the SET, controls the phase of the CB oscillation, and the output signal of the SET controls the amount of charge in the MN during the feedback process. This feedback process automatically adjusts SET output characteristics in such a way that it is used for the multifunctional Boolean logic. We experimentally demonstrated the automatic phase control and examined the speed of the feedback process by SPICE circuit simulation combined with a compact analytical SET model. The simulation revealed that programming time could be of the order of a few ten nanoseconds, thereby promising high-speed switching of the functions of the multifunctional Boolean logic circuit.

A flexible and robust rate-distortion optimization algorithm is presented to select macroblock coding mode for H.264/AVC transmission over wireless channels subject to burst errors. A two-state Markov model is used to describe the burst errors on the packet level. With the feedback information from the receiver and the estimation of the channel errors, the algorithm analyzes the distortion of the reconstructed macroblock at the decoder due to the channel errors and spatial and temporal error propagation. The optimal coding mode is chosen for each macroblock in rate-distortion (R-D)-based framework. Experimental results using the H.264/AVC test model show a significant performance of resilience to the burst errors.

Bit-errors in a subband of a wavelet-based video frame during network transmission affect not only lower-level subbands within the same frame but also the subsequent frames. This is because the video frame is wavelet-transformed image with multi-levels and referenced from later frames. In this paper, we propose a new motion estimation scheme for wavelet-based video called Intra-frame Motion Estimation (IME), in which each subband except the LL subband refers to the 1-level-lower subband in the same orientation within the same frame. This scheme protects video quality by confining the effects of the bit-errors of all subbands, except the LL subband, within a frame. We evaluated the performance of our proposed scheme in a simulated wireless network environment. As a result of tests, it was shown that the proposed IME algorithm performs better than MRME, a motion-compensated video coding scheme for wavelet video, in a heavy motion video sequence, while IME outperforms MRME at a high bit-rate in small motion video sequence.

This paper proposes a new method, EC-MDC that can detect and correct bit errors in the bitstream generated by multiple description coding. The proposed method generates two sub-bitstreams having a few redundancies as much as conventional multiple description coding. If a sub-bitstream at one side has bit error, the bit error can be corrected by using sub-bitstream of the other side. In BER-SNR experiments, reconstruction quality of the proposed method shows about 11dB higher than that of the conventional MDC at BER < 10-3 when a sub-bitstream is corrupted.

The notion of k-wise independent permutations has several applications. From the practical point of view, it often suffices to consider almost (i.e., ε-approximate) k-wise independent permutation families rather than k-wise independent permutation families, however, we know little about how to construct families of ε-approximate k-wise independent permutations of small size. For any n > 0, let Sn be the set of all permutations on {0,1,..., n - 1}. In this paper, we investigate the size of families of ε-approximate k-wise independent permutations and show that (1) for any constant ε 0, if a family Sn of permutations is ε-approximate k-wise independent, then || n(n - 1) (n - k + 1) if ε< 1; || {n(n - 1) (n - k + 1)}/(1 +ε) otherwise; (2) for any constant 0< ε 1, there exists a family Sn of ε-approximate k-wise independent permutations such that || = ; (3) for any constant ε> 0 and any n = pm - 1 with p prime, it is possible to construct a polynomial time samplable family Sn of ε-approximate pairwise independent permutations such that || = O(n(n - 1)/ε); (4) for any constant ε> 0 and any n = pm with p prime, it is possible to construct a polynomial time samplable family Sn of ε-approximate 3-wise independent permutations such that || = O(n(n - 1)(n - 2)/ε). Our results are derived by combinatorial arguments, i.e., probabilistic methods and linear algebra methods.

A silicon (Si) wire waveguiding system fabricated on silicon-on-insulator (SOI) substrates is one of the most promising platforms for highly-integrated, ultra-small optical circuits, or microphotonics devices. The cross-section of the waveguide's core is about 300-nm-square, and the minimum bending radius are a few micrometers. Recently, crucial problems involving propagation losses and in coupling with external circuits have been resolved. Functional devices using silicon wire waveguides are now being tested. In this paper, we describe our recent progress and future prospects on the microphotonics devices based on the silicon-wire waveguiding system.

An amperometric thin-film glucose biosensor based on a plasma-polymerized film using hexamethyldisiloxane as the monomer is presented. The plasma-polymerized film, achieved in plasma in the vapor phase, offers a new alternative for use in the design of the electrode-enzyme interface of biosensors. The film shows promise of high sensor performance; namely, rapid sensor response, low noise, a wide dynamic range, reproducibility, and reduction in the effects of interfering materials such as ascorbic acid. In this study, we examined the usefulness of the hexamethyldisiloxane plasma-polymerized film and investigated how the thickness of the plasma-polymerized film on a platinum electrode affected sensor characteristics: the selectivity for hydrogen peroxide versus interfering agents, the sensor response due to enzymatic reaction, and oxygen depletion.

Thermal stability of stacked high-κ dielectrics, especially ZrO2, HfO2 and ZrSiO4 /SiO2 layered structures, on silicon has been investigated in terms of ultrahigh vacuum (UHV), 1 Torr N2 and helium (He) gas annealing with controlled oxygen partial pressure (PO2) at 920. Comparison of 2 nm and 20 nm ZrO2 films under UHV annealing revealed that the trigger of silicidation is the contact of ZrO2, SiO and Si accompanying disappearance of interfacial SiO2 layer due to SiO desorption. In the contact position, a small amount of SiO gas can easily change ZrO2 to ZrSi2. This reaction model is also applicable to the silicidation of HfO2 and ZrSiO4, at not only stacked high-κ film/Si substrate interface, but also at gate poly-Si/high-κ film interface. Moreover, comparison of UHV, N2 and He annealing with controlled PO2 revealed that the optimal PO2 ranges in He at which the thermal stability of layered structure can be achieved are wider than those in UHV and N2. This result suggests that He gas physically may obstruct SiO creation due to the quenching of atomic vibration at degradation-prone sites in the SiO2 /Si interface, thus reducing probability of bond breaking process, which is the first step of silicidation.

This paper presents a novel technique for reconstructing an outdoor sculpture from an uncalibrated image sequence acquired around it using a hand-held camera. The technique introduced here uses only the silhouettes of the sculpture for both motion estimation and model reconstruction, and no corner detection nor matching is necessary. This is very important as most sculptures are composed of smooth textureless surfaces, and hence their silhouettes are very often the only information available from their images. Besides, as opposed to previous works, the proposed technique does not require the camera motion to be perfectly circular (e.g., turntable sequence). It employs an image rectification step before the motion estimation step to obtain a rough estimate of the camera motion which is only approximately circular. A refinement process is then applied to obtain the true general motion of the camera. This allows the technique to handle large outdoor sculptures which cannot be rotated on a turntable, making it much more practical and flexible.

Using various rare earth sesquioxides as additives, silicon nitride (Si3N4) samples were sintered at 1700 for 4 h by millimeter-wave heating performed in an applicator fed by a 28 GHz Gyrotron source under a nitrogen pressure of 0.1 MPa. A comparative study of densification, grain growth behavior and mechanical properties of silicon nitride fabricated by millimeter-wave and conventional sintering was carried out. Bulk densities were measured by Archimedes' technique. Except for the Eu2O3 containing sample, all samples were densified to relative densities of above 97.0%. Microstructure of the specimens was analyzed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). To investigate quantitatively the effect of millimeter-wave heating on grain growth, image analysis was carried out for grains in the specimens. Fracture toughness was determined by the indentation-fracture method (IF method) in accordance with Japan Industrial Standards (JIS). Fully dense millimeter-wave sintered silicon nitride presenting a bimodal microstructure exhibited higher values of fracture toughness than materials processed by conventional heating techniques. Results indicate that millimeter-wave sintering is more effective in enhancing the grain growth and in producing the bimodal microstructure than conventional heating. It was also confirmed that localized runaway in temperature, depending upon the sintering additives, can occur under millimeter-wave heating.

This paper proposes a 10 µm thick oxide layer structure, which can be used as a substrate for RF circuits. The structure has been fabricated by anodic reaction and complex oxidation, which is a combined process of low temperature thermal oxidation (500, for 1 hr at H2O/O2) and a rapid thermal oxidation (RTO) process (1050, for 1 min). The electrical characteristics of oxidized porous silicon layer (OPSL) were almost the same as those of standard thermal silicon dioxide. The leakage current through the OPSL of 10 µm was about 100-500 pA in the range of 0 V to 50 V. The average value of breakdown field was about 3.9 MV/cm. From the X-ray photo-electron spectroscopy (XPS) analysis, surface and internal oxide films of OPSL, prepared by complex process, were confirmed to be completely oxidized. Also the role of RTO was important for the densification of the porous silicon layer (PSL), oxidized at a lower temperature. For the RF test of Si substrate, with thick silicon dioxide layer, we have fabricated high performance passive devices such as coplanar waveguide (CPW) on OPSL substrate. The insertion loss of CPW on OPSL prepared by complex oxidation process was -0.39 dB at 4 GHz and similar to that of CPW on OPSL prepared at a temperature of 1050 (1 hr at H2O/O2). Also the return loss of CPW on OPSL prepared by complex oxidation process was -23 dB at 10 GHz which is similar to that of CPW on OPSL prepared by high temperature oxidation.

A low energy plasma based on an electron discharge was investigated for the pre-epi clean of silicon wafers and for plasma enhanced homo and hetero epitaxial growth of Si and SiGe layers. VS were produced in a short, completely dry process sequence consisting of LEPC and LEPECVD only. The wafer/epilayer interface obtained in this process sequence was suitable to grow high quality VS with low surface roughness and dislocation densities. Based on this process and its implementation in a 200/300 mm single wafer cluster tool, a high volume and economical production of VS seems possible.

Hetero-interface properties of SiO2/4H-SiC on (0001), (11-20), and (03-38) crystal orientations are presented. Epitaxial growth on new crystal orientations, (11-20) and (03-38), is described by comparing with the growth on (0001). Using thermal oxidation with wet oxygen, metal-oxide-SiC (MOS) structure was fabricated. From high-frequency capacitance-voltage characteristics measured at 300 K and 100 K, the interface properties were characterized semi-quantitatively. The interface state density was precisely determined using the conductance method for the MOS structure at 300 K. The new crystal orientations have the lower interface state density near the conduction band edge than (0001). From the characteristics of inversion-type planar MOSFETs, higher channel mobilities were obtained on (03-38) and (11-20) than on (0001). The cause of the difference in the channel mobility is speculated by the difference bond configuration of the three crystal orientations.

A new type of photodetector called "photosensitive floating field emitter, (PFFE)" has been proposed. The PFFE device combines an n-type cone-shaped triode field emitter with a-Si p-i-n photodiode film. However, a PFFE cannot detect two-dimensional distributions of light intensity. In this paper, we propose a novel structure to overcome the above this problem of the PFFE. The device was fabricated on a silicon-on-sapphire substrate to permit irradiation from the backside. p-n photodiodes were constructed within a field emitters, the n+ region being separated by p+ regions to permit detection of two- dimensional light distributions. The emission current of the PFFE/SOS was found to be proportional to the illumination intensity, but the quantum efficiency was only about 2%. This quantum efficiency is lower than that expected. Under irradiation, the emission current increased, but the gate-leakage current increased. This gate-leakage current was several orders of magnitude larger than the emission current. Almost photo-generated electrons lost in the gate electrode.

Implementation issues on generalized multi-protocol label switching (GMPLS) -based photonic switching networks are experimentally analyzed. A resilient control plane architecture using in-fiber and out-of-fiber control channels is proposed to resolve issues of establishing the control plane in out-of-band networks. The resilient control plane is demonstrated in a photonic cross-connect (PXC) -based GMPLS network involving a 1,000 km transmission line. Fast signaling for provisioning and restoration operation is accomplished by implementing in-fiber control channels as primary, and the out-of-fiber control channels effectively operate as secondary and restore messaging of the control information between neighbors. The control channel protection is initiated by the link management protocol (LMP). Using the test bed, optical layer routing operation is investigated to assess the effects on the signal quality of wavelength paths, and transparent routing of the wavelength paths over one-hop and two-hops route is demonstrated within 1 dB difference regarding the Q factor. Stable operation of loss of light (LOL) -triggered restoration is demonstrated by setting the optical level threshold 5 dB higher than the amplified spontaneous emission (ASE) noise level.

This paper proposes an enhanced method for multiple description coding (MDC) using whitening transform. The MDC using correlating transform is an error resilient coding technique that explicitly adds correlation between two descriptions to enable the estimation of one set from the other when one set is dropped in channel. This paper proposes a method to overcome practical problems that decoder must know statistics of original image in the conventional correlating transform method. The MDC using whitening transform does not need additional statistical information to reconstruct a image because the coefficients whitening transformed have uni-variance statistics. Our experimental results show that the proposed method achieves a good trade-off between the coding efficiency and the reconstruction quality. We obtain that PSNR of image reconstructed from two descriptions is about 0.93 dB higher at the 1.0BPP and PSNR from only one description is about 1.88 dB higher than conventional method at the same rate of 'Lena' image.