Electrostatic clamping with an EUVL mask chuck: Particle issues

Particle and defect issues related to electrostatic chucking with an ultra-planar, pin-structured mask chuck for EUVL application were addressed. By mapping particles/defects on the backside of 8 inch Si-wafers before and after chucking, particle transport from the chuck to the wafer was studied at application relevant electrostatic forces. Particles were detected by analysis of stray light intensities on the wafer side. Investigations were performed under ambient conditions at high chucking voltages on a bipolar chuck. Particle transport from the chuck surface to the wafer surface was mapped and found to concentrate at the pin sites. Successively lower particle counts with increasing number of chucked wafers were observed, indicating a “cleaning effect” on the chuck’s surface induced by the electrostatic chucking procedure. No influence of electric field direction (polarity) on particle count was discernable.     

影视后期中局部调色技术的实现

介绍了在After Effects中常用的三种局部调色方法.     

193nm与x射线光刻技术比较

分别从光源、照明系统、掩模、光刻胶、光刻机、光刻工艺、成本等七个方面择１９３ｎｍ与ｘ射线光刻技术进行了对比分析，介绍了目前它们的一些进展情况，并对它们的应用前景进行了简要分析。     

用于MEMS加工的双面深度光刻机对准系统设计

比较了光刻机中几种常用的对准方式,并分析了其优缺点,在此基础上对双面深度光刻机底面对准系统的设计原理、结构及对准过程做了详细阐述,并进行了对准系统的精度分析,最后介绍了为系统实现所采取的其他技术措施。实践证明,该对准系统应用于深紫外双面深度光刻机中,系统运行稳定、可靠,完全能满足系统的精度要求。     

基于 Mask RCNN 的桥小脑角区脑膜瘤与听神经瘤分类定位研究

由于人体桥小脑角区的脑膜瘤与听神经瘤在影像学的表现以及发病位置极其相似,所以临床诊断极易发生误诊．针对此问题,本文应用掩膜区域卷积神经网络（Mask RCNN）对两类肿瘤进行分类定位研究．首先采集89名脑膜瘤与218名听神经瘤患者的T₁WI-SE序列的磁共振图像,对其进行预处理,再结合改进的特征金字塔网络（FPN）算法进行网络训练．本文对比了三种不同的Mask RCNN主干网络对两者分类定位的效果．结果表明,结合改进的FPN算法和ResNet101作为主干网络的Mask RCNN分类定位模型能够有效实现对两类肿瘤的分类定位,精确率为0.918 2、召回率为0.856 9、特异性为0.876 2、均值平均精度（mAP）为0.90．     

EUV光刻技术的发展

介绍了半导体产业在制造极限的技术路线选择方面,下一代光刻技术—极紫外光刻设备面临的挑战及其开发和技术应用的进展现状,指出了在未来的22nm技术节点极紫外光刻进入量产的可能性。     

On stable refinable function vectors with arbitrary support

Refinable function vectors with arbitrary support are considered. In particular, necessary conditions for stability are given and a characterization of the symbol associated with a stable refinable function vector in terms of the transfer operator is provided: this is a generalization of Gundy’s theorem to the vector case. The proof adapts the tools provided in [S. Saliani, On stability and orthogonality of refinable functions, Appl. Comput. Harmon. Anal. 21 (2006) 254–261]. Though complications arise from noncommuting matrix products, the fundamental ideas are the same.     

A ternary 4-point approximating subdivision scheme

In the implementation of subdivision scheme, three of the most important issues are smoothness, size of support, and approximation order. Our objective is to introduce an improved ternary 4-point approximating subdivision scheme derived from cubic polynomial interpolation, which has smaller support and higher smoothness, comparing to binary 4-point and 6-point schemes, ternary 3-point and 4-point schemes (see Table 2). The method is easily generalized to ternary (2n + 2)-point approximating subdivision schemes. We choose a ternary scheme because a way to get smaller support is to raise arity. And we use polynomial reproduction to get higher approximation order easily.     

Costas环单片集成电路

介绍了多功能单片集成Ｃｏｓｔａｓ环ＫＤ８２的电路、版图和工艺设计方法。     

EOS卫星上基于MODIS云产品的AIRS像素云检测

联合使用搭载在地球观测系统(EOS)Aqua卫星上的中分辨力成像光谱仪MODIS和大气红外探测器AIRS,能够提高对全球云量分布情况的检测。利用几何扫描特性的空间匹配算法,MODIS云分类掩膜和云相态掩膜产品,结合业务上的云检测算法,实现了用空间匹配的MODIS数据对AIRS像素云特性的检测。结果表明,用MODIS 1 km分辨力产品可以鉴别出AIRS像素中不同的云类型、云层信息(低云、中云或者高云)和云相态信息(水云、冰云或者混合相态信息)。