Challenge: Experiment data management

Vast amounts of development data from diverse sources are employed during semiconductor based process development efforts (see description in “Abundance of data” section).

An overview of the diverse sources has been presented in the “Semiconductor Manufacturing Technology – Instructor Manual” by Michael Quirk and Julian Serda and is depicted below. Industry experts have said that engineers spend up to 80% of their time collecting and arranging data during the experimental development phase and only 20% of their time actually evaluating it. Therefore the management of experiment data is a major time and cost contributor for the development of semiconductor based process development efforts.

Solution:

This challenge can be mitigated by automizing the majority of these efforts, making the data collection and development documentation less of an engineer burden. The “Loading job” functionalities of XperiDesk, part of the XperiShare module, elevate and automize the structuring of raw development data. This allows engineers to focus on knowledge creation and decision taking. The loading functions can be used to auto-collect metrology and other results from diverse sources and integrate them into the knowledge network of XperiLink. Wizards guide engineers through the definition of schedulable jobs, regularly checking for new or updated files in diverse locations. The jobs can be configured to load found, new or updated files into the system and to establish relations to the existing information. Missing related entities like wafers or lots can be created on the fly and be related at the same time. This allows XperiDesk to auto-load the majority of experimental results and therefore removes the burdens on engineers to manually collect and arrange the data.