Browse Prior Art Database

Consistent Data Collection Between In-Line, Wafer Final and Module Final Tests

IP.com Disclosure Number: IPCOM000127912D
Original Publication Date: 2005-Sep-14
Included in the Prior Art Database: 2005-Sep-14
Document File: 5 page(s) / 64K

Publishing Venue

IBM

Abstract

Currently, data collection is being done independently between in-line, wafer final and module final test. Therefore process analysis is being done with gross, ie, averaged manner by wafers or lots. Chip level correlation analysis wasn't possible. This invention will describe the chip level analysis data collection technique, starting from the In-Line test. With such technique, we now can perform consistent chip level correlation analysis between In-Line, wafer final and module final tests.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 57% of the total text.

Page 1 of 5

Consistent Data Collection Between In -Line, Wafer Final and Module Final Tests

Page 1
1. Introduction

As of today, wafer final test fails and module final test fails are being analyzed in order to find the root cause in the process parameters by examining in-line measurement data taken from the Kerf, and to improve yield, SPQL and reliability. All the measurements of in-line tests, wafer final tests(WFT) and module final tests(MFT) are sent to a central database. Dataview a powerful data analysis tool, will perform the close correlation analysis between the three different test process data(In-Line, WFT and MFT).

2. Prior Art

In-line data measurements are time consuming and costly since there are many different types of data to measure in the Kerf. Generally such measurements are taken from limited chip sites a wafer and limited wafers a lot. These chip sites are called characterization sites and varies depending upon chip size, and therefore chip counts per wafer. At the WFT, there are also characterization chip sites. Those sites are being data-collected more than other sites. However, such sites are currently decided independently from in-line characterization test sites. Consequently, there is no correlation between in-line and WFT test sites. Therefore current characterization methodology will allow us only to do a gross correlation analysis between wafer to wafer(averaging out all chip measurement to wafer level). At the MFT, the situation is worse. The tester hasn't identified what modules need to be characterized, therefore only arbitrary sampling characterization is performed (Fig.


1.)

1

Page 2 of 5


3. (Key Elements)

This invention relates to a consistently correlated data collection methodology, enabling consistent data collection to allow better correlation analysis for the same chip sites between in-line, WFT and MFT.

3-1. In-line characterization sector will decide the characterization sites based on the process requirements(such as Leff stripe, Vt, etc) of a wafer. The characterization site information will be sent to WFT. The transmission method of the in-line charactericterization sites will be described in the following section.

3-2. At WFT, the tester identifies which chips are characterized(data collected) at in-line test(again, the data transmission methods will be described later) and therefore they collect more data than others on the same chip sites as in-line test.

3-3. At MFT, the tester reads chip IDs(ECID) or just in-line characterization fuse...