ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing

TL;DR

ForgetMeNot addresses the overlooked role of fluorinated forever chemicals in computing hardware manufacturing by integrating fab specific practices, reference hardware, and new hardware specifications to estimate per source fluorinated emissions across the fabrication pipeline. It formalizes emissions with $E_{ ext{Total}} = \sum_i N_{ ext{Wafers}} \times \text{Usage}_i \times (1 - \eta_{ ext{Rec}}) \times \text{GWP}_i$ and derives $N_{ ext{Wafers}}$ from die area and fabrication yields, enabling detailed source level analysis. The tool is validated against real fab data with errors below $5\%$ and is shown to shift optimal hardware choices when fluorinated emissions are included alongside embodied carbon, guiding emission aware design, manufacturing and datacenter procurement. The practical impact is a framework and open source resource that helps fabs and cloud providers reduce emissions through design, process, and hardware selection, while highlighting the need for standardized data reporting in fluorinated gas usage.

Abstract

Fluorinated compounds, often referred to as forever chemicals, are critical in various steps of semiconductor fabrication like lithography, etching, chamber cleaning, and others. Forever chemical emissions can exhibit global warming potentials thousands of times greater than carbon dioxide and persist in the atmosphere for millennia. Despite their severe impact, most sustainability works in computer systems have focused on carbon emissions alone. We address this gap by introducing ForgetMeNot, a modeling tool that quantifies fluorinated compound emissions by integrating fabrication facility-specific practices and hardware specifications, and validate its accuracy using real-world emission data from fabrication facilities. We show how ForgetMeNot can enable fabrication facilities to optimize design and material usage decisions for emission reduction and provide researchers with a methodology to calibrate emission estimates for hardware designs. When ForgetMeNot is applied to analyze emissions for manufacturing CPUs, DRAM, and storage, it illustrates how hardware generations, lithography techniques, and capacities impact fluorinated compound emissions. Finally, we demonstrate how datacenter operators can assemble low-emission servers while balancing performance demands. By factoring in fluorinated emissions into manufacturing decisions, ForgetMeNot paves the way for building more sustainable systems.

Figures (13)

• Figure 1: Systems community has primarily focused on embodied and operational carbon emissions. Fluorinated compound emissions remain unexplored. ForgetMeNot models the fluorinated compound emissions from various sources across the entire semiconductor manufacturing pipeline to provide insights and strategies to reduce overall $CO_2$-eq manufacturing emissions.
• Figure 2: The global warming potential depends on the radiative efficiency and atmospheric lifespan. Fluorinated compounds have tens of thousands of times more impact on global warming than carbon (obtained from IPCC Global Warming Potential values ghgprotocolGWPValues2024).
• Figure 3: The fluorinated compound footprint is significantly higher than the embodied carbon footprint for semiconductor manufacturing (obtained from US EPA's GHGRP electronics manufacturing data epaGHGRPElectronics).
• Figure 4: Various steps of manufacturing contribute toward emissions of fluorinated compounds.
• Figure 5: Fluorinated compound emissions vary across emission sources during fabrication and across different nanometer technology node sizes.