100 Å-thick tunnel junction by double impurity doping liquid phase epitaxy for V-band TUNNETT oscillation

The tunnel injection transit time (TUNNETT) diodes with p⁺p⁺n⁺n⁻n⁺ structure were fabricated by liquid phase epitaxy (LPE). About 100 Å tunnel junction (p⁺n⁺) was successfully prepared by the double impurity diffusion of Ge and S during LPE growth. Continuous wave (CW) oscillation was realized at 51.520 GHz in the V-band cavity with the phase noise of −60 dBc/Hz at 1 kHz bandwidth.     

Dynamic properties of an all solid-state frequency-shifted feedbacklaser

Frequency-shifted feedback (FSF) laser exhibits outstanding features in its oscillation spectrum. We analyze build-up dynamic properties of the FSF laser by means of rate equation and analyze steady-state dynamic properties of the FSF laser by means of Wigner-Ville distribution of intracavity electric field. Furthermore, we analyze instantaneous oscillation frequency at peak spectral intensity and oscillation bandwidth of its spectrum using the formula of instantaneous spectral intensity derived from Wigner-Ville distribution of intracavity electric field. These analytical results are in good agreement with the experimental ones which have been observed by a diode-pumped Nd:YVO₄ FSF laser. It becomes clear that the FSF laser supports many frequency components simultaneously even though the gain medium is homogeneously broadened and has a continuously chirped frequency components of comb in which the creation of chirped frequency components are strongly correlated in phase because of a replica of the preceding components. Also, the instantaneous oscillation frequency is closely related to the detuning frequency which depends on the total net gain in the cavity and the gain bandwidth of atomic transition. The oscillation bandwidth is defined as the product of the saturation-broadened bandwidth and the total resonant modes contributing to FSF operation     

Bidirectional finite-element method-of-line beam propagation method(FE-MoL-BPM) for analyzing optical waveguides with discontinuities

A bidirectional finite-element method-of-line beam propagation method (FE-MoL-BPM) is newly proposed for analyzing optical waveguides with discontinuities including transmissions, reflections and radiations. In this approach, the finite-element method (FEM) is introduced to discretize the derivatives of the variable perpendicular to the propagation direction. Since the proposed method is accurate and stable, only a small number of nodal points are required     

Study on the device characteristics of a quasi-SOI power MOSFETfabricated by reversed silicon wafer direct bonding

The device characteristics of a quasi-SOI power MOSFET were investigated to obtain its optimum device structure. The oxide at the original bottom surface of the bulk power MOSFET of the quasi-SOI power MOSFET formed by reversed silicon wafer direct bonding acts as the buried oxide of the conventional SOI power MOSFET. The short channel effect of the quasi-SOI power MOSFET was larger than that in the conventional SOI power MOSFET. It was suppressed by increasing the width of the oxide in the body region, and the parasitic bipolar effect was suppressed by decreasing it. We also propose a new device structure which can suppress the short channel effect and parasitic bipolar effect of a quasi-SOI power MOSFET based on the results of these experiments     

Optical frequency domain ranging by a frequency-shifted feedbacklaser

This paper describes the theoretical and experimental study of a new technique for optical frequency domain ranging (OFDR) by a frequency-shifted feedback (FSF) laser. In conventional OFDR, a frequency chirped single-mode laser is used as a light source to convert a distance into a beat frequency, and a tradeoff exists between measurement range and resolution. The FSF laser output consists of periodically generated chirped frequency components whose chirp rate is faster than 100 PHz/s (P=10¹⁵), By use of the FSF laser, the tradeoff is removed and long-distance high-resolution OFDR is realized In the experiment, a distance of 18.5 km was measured with a resolution of 20 mm     

Effect of Annealing Ambient on the Self-Formation Mechanism of Diffusion Barrier Layers Used in Cu(Ti) Interconnects

Copper (titanium) [Cu(Ti)] films with low titanium (Ti) concentration were found to form thin Ti-rich barrier layers at the film/substrate interfaces after annealing, which is referred to as self-formation of the barrier layers. This Cu(Ti) alloy was one of the best candidates for interconnect materials used in next-generation ultra-large-scale integrated (ULSI) devices that require both very thin barrier layers and low-resistance interconnects. In the present paper, in order to investigate the influences of annealing ambient on resistivity and microstructure of the Cu alloys, the Cu(7.3at.%Ti) films were prepared on the SiO₂ substrates and annealed at 500°C in ultra-high vacuum (UHV) or argon (Ar) with a small amount of impurity oxygen. After annealing the film at 500°C in UHV, the resistivity was not reduced below 16 μΩ-cm. Intermetallic compounds of Cu₄Ti were observed to form in the films and believed to cause the high resistivity. However, after subsequently annealing in Ar, these compounds were found to decompose to form surface TiO _x and interfacial barrier layers, and the resistivity was reduced to 3.0 μΩ-cm. The present experiment suggested that oxygen reactive to titanium during annealing played an important role for both self-formation of the interfacial barrier layers and reduction of the interconnect resistivity.     

Class-Based Packet Scheduling to Improve QoS for IP Video

An important issue for video transmission over IP networks is the preservation of perceived video quality despite packet loss. Packet loss can be detrimental to compressed video. However, reducing packet loss to a very low level is difficult with current techniques. Furthermore, even a very low objective loss probability can still seriously distort perceived video quality. This paper presents a packet scheduling scheme at a router which addresses the loss issues of networked video. Experiments using real video data show that the proposed scheme can significantly improve the visual quality of video and network efficiency. Moreover, it can provide different classes of videos with different levels of loss guarantees, while maintaining service fairness among the same class of videos.     

Application-aware buffer management: new metrics and techniques

An important issue for video transmission over IP networks is preservation of perceived video quality despite packet loss. Packet loss can be detrimental to compressed video. However, reducing packet loss to a very low level is difficult with current loss control techniques. Furthermore, even a very low objective loss probability can still seriously distort perceived video quality. This paper presents two buffer management schemes using video characteristics. They increase maximum loss tolerance for a desired level of video quality, providing better quality at equal loss ratios, when compared to a conventional buffer management scheme. Meanwhile, service fairness and network efficiency are also improved.