Abstracts of Magazine FUJITSU 2005-7 (VOL.56, NO.4)

Special Issue : Research and Development

• Low-k Interlayer Dielectrics for 65 nm-Node LSIs

We have developed a novel porous low-k material called Nano-Clustering Silica (NCS) that has a low dielectric constant (k = 2.25) and high film strength (Young’s modulus E = 10 GPa). Conventionally, the pore-creation technique, by which holes with a dielectric constant of 1 are created within a film, is very effective for reducing the dielectric constant, but reduces the film strength. On the other hand, NCS, which was developed using an original technology called nano-clustering, has homogeneously distributed pores of 2.8 nm or less. NCS also has good compatibility with 65 nm-node micro-fabrication processes. Fujitsu has advanced the trial production of multi-level interconnects made of NCS and established an NCS-based multi-layer interconnect formation technology. Our Cu/NCS interconnects are reliable enough to prevent damage to Cu interconnects by the mechanical stresses that occur in wire bonding, packaging, and other processes. This paper describes the problems of conventional porous low-k materials and how they are solved using NCS are described. Then, this paper describes the development status of NCS-based LSI multi-layer interconnects.

• SoC Embedded Memories

The functionality of SoCs for graphics and audio applications is increasing; as a result, the density of embedded memory on these chips is also increasing. Moreover, the increases in chip area and power consumption have become big issues for embedded SRAMs. Extensive research and development of alternative memories have been conducted to address these issues. In this paper, we describe the R&D status of DTMs for high-density, low-power applications; FRAMs for nonvolatile, low-power applications; and MRAMs for nonvolatile, high-speed, and high-density applications.

• High-Density Perpendicular Recording Technology

Although perpendicular magnetic recording is fundamentally an excellent recording method, it is not ready for practical use due to its poor read-write performance and manufacturability of equipment and media. Fujitsu has therefore developed the following elemental technologies that solve these problems: 1) granular materials that improve the recording density by reducing noise from the recording layer; 2) magnetic domain control technology that removes spike noise from the domain structure of the soft under layer (SUL), which helps writing through magnetic interactions with the write head; and 3) multi-layering of the SUL to reduce the wide-area track erasure (WATER) phenomenon, which causes data loss due to large-area erasure of recorded information. By using these technologies, we have obtained excellent performance from perpendicular magnetic recording media in terms of the spike noise and WATER phenomenon. We confirmed that sufficient error rates can be achieved at a linear density exceeding 1000 kBPI. This paper describes the advantages and problems of perpendicular magnetic recording and introduces our elemental technologies that solve these problems. It also describes the performance of our perpendicular magnetic recording media.

• FR-V Single-Chip Multicore Processor

To realize the low power consumption and low-cost equipment needed to encode high-definition broadcasts, Fujitsu has developed a single-chip multicore processor that integrates four 8-way VLIW FR-V processor cores. This new multicore processor is fabricated using a 90 nm, nine metal-layer CMOS process and a 900-pin flip-chip package. The processor cores operate at 500 MHz, the memory interfaces at 250 MHz, and the system bus at 166 MHz. By using a single processor core, MPEG-2 MP@ML video-stream decoding can be performed at 190 MHz. By using four processor cores, MPEG-2 MP@HL video streams can be decoded using just software. Also, this processor only needs about three watts to decode MPEG-2 MP@HL video streams. This paper introduces the hardware and software development environment of this new processor. It also describes the processor's software operation environment and some examples of its application.

• Optics Simulator for Use in Nano-Optics Analysis

Fujitsu has developed an optics simulator for use in nano-optics analysis that performs Finite-Difference Time-Domain (FDTD) calculations. This simulator enables us to make plasmon-enhancement calculations based on Lorentzian and Drude models for metals with a negative-dielectric constant. We confirmed the effectiveness of this simulator by making calculations for analytical models with a negative-dielectric metal. We also found there is good agreement between simulated and experimental results for an optical beam spot profile. This FDTD simulator is a very powerful tool for use in precise nano-optical analysis and design and has an easy-to-use graphical user interface. In this paper, we describe some of the problems encountered when using the FDTD method in optics calculations and their solutions. We also describe how we verified the numerical accuracy of the new simulator.

• Next-Generation Photonic IP Network

Internet data-traffic is rapidly growing as fibre-to-the-home (FTTH) and asymmetric digital subscriber line (ADSL) are becoming the preferred infrastructures. This brings various rich sets of services such as voice, data, and video content to the users. Now that the ubiquitous network society is moving into high gear, future networks must efficiently support various services and transfer data at the requested quality. To satisfy these requirements, next-generation photonic IP networks with high-performance optical/electrical switching and control layers will be fundamental. This paper describes the following technologies that Fujitsu is developing for these next-generation photonic IP networks: a high-performance optical gateway node technology and multi-layer traffic engineering for efficiently accommodating IP traffic by controlling these nodes. This paper also describes a prototype for the next-generation photonic IP networks.

• Health-Care Technology for Networks

Clearly positioned as part of the social infrastructure, IP networks must offer high availability. Moreover, as our daily lives have recently become more dependent on IP networks, any performance degradation in these networks can significantly impact our lives. For example, an increase in the response time of an Internet shopping site will immediately lead to a decrease in site sales. However, existing technologies such as performance and fault management are currently not able to sufficiently analyze this performance degradation factor. Particularly, in open systems, the problems to be solved are quite complex because the end-to-end services are composed of various network devices and network providers. Fujitsu is therefore developing technologies to solve these problems and meet the demands of next-generation networks. This paper introduces our research activities for next-generation networks. It then introduces network fault analysis and evaluation technologies for solving the current problems. Lastly, it discusses technology migration toward autonomous network operation and management.

• High-Power GaN-HEMT Amplifiers for W-CDMA Wireless Base-Station Applications

The third generation of mobile communication systems is predicted to soon achieve higher transmission rates, and the fourth generation of such systems is expected to be commercialized by around 2010. For up-coming high-capacity base stations of the third generation, the power dissipation and size of transmitter amplifiers need to be reduced. To meet these needs, Fujitsu has developed a gallium nitride based high-electron mobility transistor (GaN-HEMT) capable of high-voltage operation. By combining a high-power amplifier built using this GaN-HEMT with a digital pre-distortion system, we can satisfy the system requirement for high efficiency. We have also fabricated GaN-HEMTs using large-diameter, low-cost conductive substrates that are suitable for mass production. These devices showed sufficient performance for bass-station applications and operated stably under RF stress testing for 1000 h, indicating they are close to being put to practical use. In this paper, we describe our recent progress in developing GaN-HEMT amplifiers.

• High-Performance Techniques for RFID Systems

Radio Frequency Identification (RFID) systems are now recognized as key elements for realizing the ubiquitous network society, and various field tests have been initiated. The main expected applications of RFID systems are traceability and logistics. Traceability is required for assuring food safety and preventing medical mistakes, and logistics is required for supply chain management. Fujitsu has been developing total RFID system components, including tag LSIs, tags, reader/writers, middleware, and application software, to satisfy various user needs. To realize superior traceability systems in distribution and manufacturing processes, tags with rewritable user memory for storing historical data are indispensable, and high-speed data processing techniques are needed to manage numerous tags and their historical data. This paper outlines Fujitsu's RFID technology development and describes a tag LSI with an embedded ferroelectric memory (FRAM) that features high-speed data writing. It also describes high-performance techniques and anti-collision algorithms that enhance the system-level communication speed.

• RF-MEMS Devices for Wireless Communications

Fujitsu is developing novel RF-MEMS devices for wireless communications that are suitable for broadband, globalization, and high-frequency applications. We are developing FBAR technology for the RF band-pass filters of mobile phones. For example, we have developed practical 2 to 5 GHz FBAR filters that are based on our unique FBAR technology. Because of their high Q factor, these filters have a big advantage in high-frequency applications. We are also developing RF-MEMS switches, which are promising because of their low insertion loss and high isolation. One of our RF-MEMS switches has an insertion loss of less than 0.3 dB and an isolation exceeding 30 dB at 2 GHz. Moreover, we have developed a MEMS variable capacitor that has a capacitance change of 4 pF, making it a promising device for tunable circuits. This paper introduces our technology for FBARs, RF-MEMS switches, and MEMS variable capacitors.

• H.264/AVC Video Coding Standard and Fujitsu's Activity

In 2003, the ITU-T and ISO developed a new MPEG video coding standard called H.264/AVC (Advanced Video Coding). H.264/AVC provides more than twice the compression performance of existing video coding standards and will be adopted for one of the next-generation DVD video formats and mobile digital terrestrial broadcasting in Japan. However, due to its aggressive optimization at every coding stage and pursuit of coding efficiency by subdividing coding modes, H.264/AVC introduces increased complexity and requires sophisticated coding control to obtain stable picture quality. In response, Fujitsu has developed a unique coding control algorithm for H.264/AVC to realize stable video quality and a prototype H.264/AVC encoder system for receiving digital terrestrial broadcasts on mobile PCs. This paper introduces H.264/AVC, its coding control algorithm, and the prototype H.264/AVC encoder system.

• Business Expansion of Palm Vein Pattern Authentication Technology

Fujitsu has developed a new contactless palm vein pattern authentication technology that uses vascular patterns as personal identification data. This technology is very secure because the authentication data exists inside the body and is therefore difficult to steal. It is also very accurate―in a test using 140 000 palm profiles of 70 000 Fujitsu Group employees, it had a false acceptance rate of less than 0.00008% and a false rejection rate of 0.01%. The Suruga Bank, Bank of Tokyo-Mitsubishi, Hiroshima Bank, and Bank of Ikeda started customer confirmation using palm vein pattern authentication in July 2004, October 2004, April 2005, and June 2005, respectively. Additionally, in response to the "Act for the Protection of Personal Information," which came into effect in April 2005, an increasing number of financial institutions, including the Nanto Bank, have begun adopting this technology. Fujitsu has also started sales of products for door security, login authentication, and other applications. We will achieve business growth with these solutions by downsizing the palm vein sensor and reducing the authentication time. This paper describes this technology and its expected business expansion.

• Quantum Well Infrared Photodetectors

Infrared imaging sensors visualize objects in the dark by detecting their heat radiation. Recently, the importance of remote night-vision sensors in future sensor networks that will be built to secure a safe and convenient society has been recognized. In response, we have developed a Quantum Well Infrared Photodetector (QWIP) that operates using photo-excitation in GaAs-based multi-quantum-well layers. We have realized large-scale QWIP focal plane arrays (QWIP-FPAs), which are the key devices in infrared imaging systems, by hybridizing 2-dimensional arrays of QWIP and CMOS readout circuits. In this paper, we introduce the infrared sensing mechanism of QWIP pixel structures and our elliptical-pattern optical couplers, which are essential components in QWIP-FPAs. We also describe the sensor characteristics of our QWIP and the infrared imaging performance of QWIP-FPAs.

• Key Device for Quantum Communication: Single-Photon Generator

A single-photon generator is a light source that generates an optical pulse consisting of a single photon. Even though laser technology is a branch of quantum electronics/quantum optics (because laser light is generated by a quantum-mechanical process), the laser is quite different from the single-photon generator because laser light is more like a classical wave, for example, a radio wave. On the other hand, with the single-photon generator, both the generation process and the generated optical pulse have quantum characteristics. The single-photon generator is expected to become a key device for quantum information processing and highly secure quantum encryption. However, a single-photon generator that can be put to practical use does not yet exist. Fujitsu together with the University of Tokyo have become the first in the world to generate and transmit a single-photon pulse that has a wavelength suitable for optical fiber transmission, and we expect to build a practical generator within several years. In this paper, we describe our development of single-photon generation technology.

• Content Creation-Support Technology and Its Commercialization

As we enter the ubiquitous broadband era, e-learning, which enables people to study whatever they want, anytime and anywhere, is becoming popular. E-learning is conducted by delivering, through VOD and other means, conventional lecture videos and also videos that are synchronized with slides. However, this synchronized material is usually created manually using authoring tools, which is a very costly process. This paper describes a synchronous extraction technology that solves this problem by automatically synchronizing the frames of a lecture video and slides offline using a character and image recognition technique and a content virtual-editing technology that intensively manages created content. Then, this paper introduces Campusmate/CourseNavig, which is a Fujitsu e-learning solution for universities that incorporates this synchronous extraction technology.

• Visual Text Mining Technology for Patent Mining

In recent years, patents have gained recognition as an important business property. To fully utilize patents in business management and thereby increase income and strengthen competitiveness, companies must understand the position of their own patents and the technological strategies of their competitors. However, in order to do this, companies must spend a lot of time and money searching and analyzing patent information. To make patent mining more efficient, Fujitsu Laboratories has developed a knowledge management tool that mines patent information using visual text mining technologies. This tool is used within the Fujitsu Group and has also been released as a commercial product. In this paper, we give an overview of this tool and describe the practical use of its functions.

• Software Modeling Technology for Effective Use of IT

Business environments are changing rapidly, and IT systems are becoming increasingly complicated. As a result, when coordinating business activities and IT systems, the modeling technology for software development is becoming more important. Fujitsu Laboratories has been studying software modeling technology for effective use of IT and has developed three new technologies: 1) a business-process modeling tool based on UML that reduces the cost of designing business processes by more than 25%, 2) a business process management technology that enables business processes to be monitored and analyzed by collecting data about business events from IT systems without modifying the systems, and 3) a development tool that automatically generates more than 75% of the Java programs of a Web application from a screen transition model described in UML. In this paper, we describe the trends of software modeling technology and these three technologies.

• Advanced Technologies That Support Environment-Friendly Products and Software Services

Coexistence with the environment has become one of the most important issues for humans in the 21st century, and various approaches have been investigated on a global scale. Fujitsu Laboratories Ltd. is researching and developing environmental technologies for sustainable growth. We think that one of the most important activities for R&D in the Fujitsu Group is to research and develop advanced environmental technologies for environment-friendly products (Green Products and Super Green Products) and software services. This paper introduces some of Fujitsu's new green technologies, for example, PC housings made from a corn-based resin, an environmental purification technology that uses a new photocatalyst (Ti apatite), a high-efficiency micro fuel-cell, and a method for quantitatively evaluating the environmental burden and contribution of the products and solutions of the Fujitsu Group.

• Fujitsu Service Robot Equipped with 3-D Vision System

In recent years, robot development has expanded from the industrial field to include new robots for personal use. Robots such as entertainment robots and robots with housekeeping-related functions are now being developed and marketed; and in the future, a very large number of people will ask for robots that provide many other services. To respond to these needs, Fujitsu has developed a service robot that can provide various services in offices, retail facilities, and other public areas in which people work and spend leisure time. This robot can provide services such as greeting and escorting guests onto elevators, operating the elevators, and making security patrols of buildings. This paper describes the concept behind this robot and its system configuration. It also describes the architecture and functions of this robot's 3-D vision system, which is one of its special features, and the autonomous navigation functions that are realized using this vision system.