Title:
|
The life-cycle environmental impacts of etching silicon wafers and (PE)CVD chamber cleaning
|
|
Author(s):
|
|
|
Published by:
|
Publication date:
|
ECN
Solar Energy
|
10-9-2008
|
|
ECN report number:
|
Document type:
|
ECN-M--08-079
|
Conference Paper
|
|
Number of pages:
|
Full text:
|
3
|
Download PDF
|
Presented at: Electronics Goes Green 2008+, Berlin, Germany, 8-10 september 2008.
Abstract:
Fluorinated gases are used by the semiconductor and photovoltaic industry for etching silicon wafers and (PE)CVD chamber cleaning. The desired result is due to F atoms and other reactive species, but the emission of the undecomposed PFC (perfluorinated) gases is unwanted because they have a high global warming effect and high atmospheric life-time. In this study a full life-cycle assessment is used in order to (1) compare the environ-mental impacts of the different technologies and (2) to indicate improvement options.
The steps in the life cycle are the following: synthesis of the compounds, transportation, distribution in the fab (connection of cylinders), use in the process, abatement to destroy the unreacted gases and take-back of cylin-ders. Emissions from each step can be direct (from emission of the fluorinated gases) or indirect (from energy use).
Results, partly based on best guesses, indicate that fugitive emissions of the fluorinated gases during synthesis, downtime of abatement system and cleaning of the not completely empty cylinders dominate the life-cycle global warming effect. This means that the global warming effect of the gas itself determines the effect being the highest for SF6. F2 turns out to be clearly in advantage over the other fluorinated compounds because it has a global warming potential of zero with moderate efforts for synthesis.
Possible improvement options to minimize the use and emission of fluorinated gas are (1) strict procedure for connection of cylinders, (2) complete usage or reliable abatement of the gas from the bottle, (3) the recovery or reliable abatement of unused gases from the process and (4) end-point detection of the process.
Back to List