一种新型的自对准源漏接触技术

提出了一种用于亚微米尺寸以下MOSFET器件的自对准源漏接触技术。这种新型的集成方法改变了传统的工艺步骤,即光刻接触孔,然后向里填充导电材料。在形成栅极、绝缘介质侧墙以及自对准金属硅化物以后,通过沉积金属膜层,并各向异性刻蚀以在绝缘介质侧墙两旁形成一对金属的侧墙结构。这个侧墙连接底部的源漏区和上部的互联区,作为底层的金属接触引出。这个方法不仅减小了刻蚀接触孔的难度,且采用自对准的方法形成金属接触也减小了源漏接触的距离,提高了集成度。这项工艺集成技术尝试应用于0.5μm栅长MOSFET器件结构中,并仿真得到了良好的电学性能。     

Characteristics of the sampling paddle modules of neutron wall from the cosmic ray test

In the construction of a large area neutron detector （neutron wall） that is used to detect neutrons at GeV energies, the performances of all the sampling paddle modules prepared for the neutron wall are investigated with a specially designed test bench. Tested by cosmic rays, an average intrinsic time resolution of 222.5 ps is achieved at the center of the modules. The light attenuation length and the effective speed of the light in the module are also investigated.     

Scalar mesons and glueballs in Dp-Dq hard-wall models

We investigate the light scalar mesons and glueballs in the Dp-Dq hard-wall models, including D3-Dq, D4-Dq, and D6-Dq systems. It is found that only in the D4-D6 and D4-Ds hard-wall models are the predicted masses of the clq scalar meson fo scalar glueball consistent with their experimental or lattice results. This indicates that D4-D6 and D4-Ds hard-wall models are the favorite candidates of the realistic holographic QCD model.     

Electron Acceleration in the Bubble Regime with Dense-Plasma Wall Driven by an Ultraintense Laser Pulse

An optimizing scheme for electron acceleration in the wake bubble with dense-plasma wall driven by an ultraintense laser pulse is presented and investigated by particle-in-cell simulation. The wall has an inner diameter matching the expected lateral bubble size. The bubble shape can be transversely controlled and longitudinally shrunk. The accelerated electrons as a bunch have a high quality because the electrons almost stay close to the bottom of the bubble and are accelerated to much high energy with narrow energy spread.     

异构多核SoC处理器内部存储架构优化

异构多核技术的发展使微处理器的性能有了较大提升,而处理器与外部存储器之间的带宽差异限制了处理器的性能发挥,“存储墙”问题日益严重。针对一种用于高密度计算的异构多核SoC系统,文中提出了一套存储设计方案。该方案通过复用一些长时间闲置的本地空闲存储资源作为二级共享缓存来增加访存带宽,减少访问外部存储频率。分布式高速共享二级缓存结合多路并行访问外部存储的层次化存储结构,缓解了系统处理数据与外部存储器间的速度差异,提高了数据的存取效率,优化了系统的性能。综合资源消耗和计算效率,文中所提设计相比普通二级缓存节约了69.36%的片上SRAM资源,相比无缓存结构提高了41.2%的加速比,整体任务计算时间平均减少了约40.6%。     

激光内送粉变姿态增材制造薄壁墙的试验研究

为了探索在非水平基面上激光增材制造薄壁件的成形规律，拓展激光增材修复技术的广泛应用。基于激光内送粉技术，分别在0°～150°的倾斜基板上进行薄壁件的成形试验，研究了不同倾斜角度下薄壁墙的成形规律，并分析了非水平基面下熔池的受力规律。试验结果表明：随基板倾斜角度的增大，薄壁墙的总体宽度及高度增大，熔池长度变短；随着基板倾斜角度的增大，成形件尾部塌陷越来越严重，在基板角度150°时，产生最大角度为21°的斜坡。该研究结果可以为非水平基面上激光增材制造及修复提供参考价值。     

Ventuz参展BIRTV2014 开启互动创意之旅

正8月25日下午,维图斯(中国)有限公司召开新闻发布会,宣布ventuz将独立参展广电行业盛会——BIRTV2014,并与中国传媒大学合作,成立官方双认证培训中心.首次授课将于9月开始。Ventuz作为国际著名在线包装软件提供商,其主要产品Vennuz是一款实时的图像内容创作、编辑和控制播出软件,它专注于高端视听内容的制作,包括专业展示、视频墙安装、实况现场显示及互动、播出等领域。