公开(公告)号：US6040912A

公开(公告)日：2000-03-21

申请号：US163314

申请日：1998-09-30

申请人： Steven J. Zika ,  C. Bradford Hopper

发明人： Steven J. Zika ,  C. Bradford Hopper

IPC分类号： G01N21/956 ,  G01B11/00

CPC分类号： G01N21/95607

摘要： A method and apparatus for detecting random layout structures sensitive to process induced pattern errors in semiconductor device manufacturing applies a first manufacturing process to a first wafer containing semiconductor devices. A second manufacturing process is applied to a second wafer containing semiconductor devices. The second manufacturing process is similar to, but different from the first manufacturing process. The first and second wafers are compared by image subtraction to detect systematic pattern defects in the semiconductor devices of one of the first and second wafers. After differences are detected, the layout is examined to determine whether the difference represents a defect. If so, the design rules of the layout can be changed to accommodate a wider process variation and improve processing yield.

摘要翻译： 用于检测对半导体器件制造中的处理感应图案错误敏感的随机布局结构的方法和装置将第一制造工艺应用于包含半导体器件的第一晶片。 将第二制造工艺应用于包含半导体器件的第二晶片。 第二个制造过程与第一个制造过程类似，但不同。 通过图像相减来比较第一和第二晶片，以检测第一和第二晶片之一的半导体器件中的系统图案缺陷。 检测到差异后，检查布局以确定差异是否表示缺陷。 如果是这样，可以改变布局的设计规则以适应更广泛的工艺变化并提高加工产量。

公开(公告)号：US06947804B1

公开(公告)日：2005-09-20

申请号：US10340124

申请日：2003-01-10

申请人： James M. Pak ,  Steven J. Zika

发明人： James M. Pak ,  Steven J. Zika

IPC分类号： G05B19/418 ,  G06F19/00 ,  H01L21/00

CPC分类号： G05B19/41865 ,  H01L21/67276 ,  Y02P90/20

摘要： A method for detecting sequential processing effects on integrated circuits to be manufactured in a manufacturing process, includes: determining a first random sequence for a plurality of wafers; performing a first process step on the plurality of wafers by a first process tool in accordance with the first random sequence; determining a second random sequence for the plurality of wafers; and performing a second process step on the plurality of wafers by a second process tool in accordance with the second random sequence. The method performs randomization of wafer processing sequences at the process tool itself. By performing randomization of wafer processing sequences at the process tool, the need for separate wafer handlers is eliminated, resulting in significant cost reduction, clean-room space savings, improved yield, improved manufacturing cycle time, and improved signal-detection capabilities.

摘要翻译： 一种用于检测在制造过程中要制造的集成电路的顺序处理效果的方法，包括：确定多个晶片的第一随机序列; 通过第一处理工具根据第一随机序列对多个晶片执行第一处理步骤; 确定所述多个晶片的第二随机序列; 以及根据所述第二随机序列，通过第二处理工具对所述多个晶片执行第二处理步骤。 该方法在处理工具本身执行晶片处理序列的随机化。 通过在处理工具处执行晶片处理序列的随机化，消除了对单独的晶片处理器的需要，导致显着的成本降低，清洁室空间节省，提高的产量，改进的制造周期时间和改进的信号检测能力。

公开(公告)号：US06496596B1

公开(公告)日：2002-12-17

申请号：US09274212

申请日：1999-03-23

申请人： Steven J. Zika ,  Christopher Lee Pike

发明人： Steven J. Zika ,  Christopher Lee Pike

IPC分类号： G06K900

CPC分类号： H01L22/20 ,  G01N21/9501

摘要： A defect categorization method is described. Two or more regions are defined on a die. These regions can be repeated for other dies on the same wafer or corresponding regions on other dies on other wafers. The wafer(s) are scanned for defects. The location of each defect, along with any other desired defect data is stored. The stored defect data is then categorizing by region or regions within a die. As a result of the categorization, optional defect generation predictions can be made or optional yield or kill ratio predictions can be made. Also, specific regions(s) can be rescanned based on the result of the categorization.

摘要翻译： 描述缺陷分类方法。 在模具上限定两个或更多个区域。 这些区域可以在同一晶片上的其他管芯或其他晶片上的其它管芯上的相应区域重复。 扫描晶圆的缺陷。 存储每个缺陷的位置以及任何其它所需的缺陷数据。 存储的缺陷数据然后根据模具内的区域或区域进行分类。 作为分类的结果，可以进行可选的缺陷产生预测，或者可以进行可选的产量或杀死率预测。 此外，可以基于分类的结果来重新扫描特定区域。

公开(公告)号：US5862055A

公开(公告)日：1999-01-19

申请号：US896340

申请日：1997-07-18

申请人： Ming Chun Chen ,  Paul J. Steffan ,  Steven J. Zika

发明人： Ming Chun Chen ,  Paul J. Steffan ,  Steven J. Zika

IPC分类号： G06T7/00 ,  G06F19/00

CPC分类号： G06K9/00 ,  G06T7/0004 ,  G06K2209/19 ,  G06T2207/30148

摘要： A method of determining classification codes for defects occurring in semiconductor manufacturing processes and for storing the information used to determine the classification codes. A wafer is selected from a production lot after the lot is sent through a first manufacturing process. The selected wafer is scanned to determine if there are defects on the wafer. Images of selected defects are examined and a numerical value is assigned to each of N elemental descriptor terms describing each defect. A classification code is determined for each defect based upon the numerical values assigned to the N elemental descriptor terms. The classification code and numerical values assigned to the N elemental descriptor terms are stored in a database. The wafer is sent through each sequential process and classification codes are assigned to additional defects selected after each sequential process. The classification codes and numerical values assigned to the N elemental descriptor terms for the additional selected defects are stored in the database.

摘要翻译： 确定半导体制造过程中出现的缺陷的分类代码并存储用于确定分类代码的信息的方法。 在批次通过第一制造过程发送之后，从生产批次中选择晶片。 扫描所选择的晶片以确定晶片上是否存在缺陷。 检查所选缺陷的图像，并且将数值分配给描述每个缺陷的N个元素描述符术语中的每一个。 基于分配给N个元素描述符项的数值，为每个缺陷确定分类代码。 分配给N个元素描述符词的分类代码和数值被存储在数据库中。 通过每个顺序过程发送晶片，并且将分类代码分配给在每个顺序处理之后选择的附加缺陷。 分配给附加选定缺陷的N个元素描述符项的分类代码和数值存储在数据库中。