2002•Unpublished venueRequires access

Construction of a "Grand Pareto" for line yield loss, by process loop using limited data sets

J. Li, M. McIntyre

Open publisher page 7 citations

Abstract

Identifying defect yield loss down to a single operation would provide the ultimate understanding of that operation's contribution to yield. Unfortunately, defect detection resources are seldom available in sufficient quantity to allow a yield enhancement engineer to pinpoint yield loss contributions down to a single operation. This paper will explain a methodology to build a by-operational loop defect pareto for an entire manufacturing process, which normally has only limited inspection points in-line due to scan capacity constraints. Establishing a total line pareto is essential in effectively distributing limited engineering resources to improve yields on defect limited devices. Key enables for checking and cross referencing the validity of short loop experiments will be outlined in detail throughout the paper. This paper will illustrate how, with confidence, a "Grand Pareto" for processing can be supported by the available data. In addition, a demonstration on how loss rates partitioned between "fixed" inspection points are then allocated to the intermediate operation will be shown. This methodology is required to get a total understanding of die loss contribution. Finally we will identify follow up activities that had occurred as a result of the initial "Grand Pareto".

About this research paper

What this paper is about

Identifying defect yield loss down to a single operation would provide the ultimate understanding of that operation's contribution to yield. Unfortunately, defect detection resources are seldom available in sufficient quantity to allow a yield enhancement engineer to pinpoint yield loss contributions down to a single operation. This paper will explain a methodology to build a by-operational loop defect pareto for an entire manufacturing process, which normally has only limited inspection points in-line due to scan capacity constraints. Establishing a total line pareto is essential in effectively distributing limited engineering resources to improve yields on defect limited devices. Key enables for checking and cross referencing the validity of short loop experiments will be outlined in detail throughout the paper. This paper will illustrate how, with confidence, a "Grand Pareto" for processing can be supported by the available data. In addition, a demonstration on how loss rates partitioned between "fixed" inspection points are then allocated to the intermediate operation will be shown. This methodology is required to get a total understanding of die loss contribution. Finally we will identify follow up activities that had occurred as a result of the initial "Grand Pareto".

Why it matters

OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Identifying defect yield loss down to a single operation would provide the ultimate understanding of that operation's contribution to yield. Unfortunately, defect detection resources are seldom available in sufficient quantity to allow a yield enhancement engineer to pinpoint yield loss contributions down to a single operation. This paper will explain a methodology to build a by-operational loop defect pareto for an entire manufacturing process, which normally has only limited inspection points in-line due to scan capacity constraints. Establishing a total line pareto is essential in effectively distributing limited engineering resources to improve yields on defect limited devices. Key enables for checking and cross referencing the validity of short loop experiments will be outlined in detail throughout the paper. This paper will illustrate how, with confidence, a "Grand Pareto" for processing can be supported by the available data. In addition, a demonstration on how loss rates partitioned between "fixed" inspection points are then allocated to the intermediate operation will be shown. This methodology is required to get a total understanding of die loss contribution. Finally we will identify follow up activities that had occurred as a result of the initial "Grand Pareto".

Key concepts: Pareto principle, Pareto analysis, Key (lock), Process (computing), Computer science, Yield (engineering), Pareto chart, Line (geometry)

Related papers

Back to paper searchBrowse research topicsOriginal source
Construction of a "Grand Pareto" for line yield loss, by process loop using limited data sets — Research Paper | ScholarLens