ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing Conference: ; ; CCS: Computer systems organization Cloud computing CCS: Social and professional topics Sustainability
Rohan Basu Roy, Raghavendra Kanakagiri, Yankai Jiang, Devesh Tiwari · 2025
Read ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing Conference: ; ; CCS: Computer systems organization Cloud computing CCS: Social and professional topics Sustainability by Rohan Basu Roy, Raghavendra Kanakagiri, Yankai Jiang, Devesh Tiwari (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-11236.txt.
Introduction
Rohan Basu Roy and co-authors build ForgetMeNot, a tool that counts the climate cost of so called forever chemicals in chip making, in a 2025 paper. Fluorinated compounds drive key steps of fabrication and linger in the air for millennia, with warming power thousands of times that of carbon dioxide. Sustainability work in computing had counted carbon alone.
The tool blends each factory's practices with hardware specs to estimate the fluorinated output. Checks against real factory data validated its numbers. Runs across processors, memory, and storage show how design choices move emissions. Data center buyers can also pick lower emission servers for a set performance target.
Catalog record
| CREATOR | Rohan Basu Roy, Raghavendra Kanakagiri, Yankai Jiang, Devesh Tiwari |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 86 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-11236
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing Conference: ; ; CCS: Computer systems organization Cloud computing CCS: Social and professional topics Sustainability Rohan Basu Roy, Raghavendra Kanakagiri, Yankai Jiang, Devesh Tiwari Abstract . 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. 1. Introduction 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. Why should the computer systems community focus on forever chemicals? The usage of fluorinated compounds including PFAS, hydrofluorocarbons, and others – often called forever chemicals due to their longevity and high resilience to degradation in the environment, are integral ingredients in different steps of modern semiconductor manufacturing like wafer fabrication and photolithography to plasma etching, chamber cleaning, and others ( garg2020review ; epaPFASExplained2023 ; chen2019release ; njdohNitrogenTrifluoride2001 ) . Their global warming potential (GWP) vastly exceeds that of carbon dioxide, with magnitudes up to tens of thousands of times higher. Additionally, these compounds persist in the atmosphere for centuries or millennia, making any emissions effectively permanent contributors to global warming ( ghgprotocolGWPValues2024 ) . The total global warming impact of all greenhouse gases, including carbon dioxide ( CO_{2} ) and fluorinated compounds, is expressed in terms of CO_{2} -equivalent ( CO_{2} -eq), which is the amount of CO_{2} with the same global warming effect. The unit for expressing the impact of both carbon dioxide ( CO_{2} ) and fluorinated compounds is grams of CO_{2} -equivalent (g CO_{2} -eq). The sustainability studies in the systems and architecture community ( gupta2021chasing ; gupta2022act ; li2023toward ; souza2023ecovisor ; bashir2023sustainable ) , have primarily focused on the impact of carbon dioxide ( CO_{2} ) toward embodied carbon (the emissions incurred to produce energy for hardware manufacturing, transportation, and end-of-life processing) and operational carbon (the emissions caused to produce energy during the hardware’s runtime). Current sustainability modeling tools (e.g., ACT ( gupta2022act ) ) primarily focus on estimating carbon dioxide ( CO_{2} ) emissions and do not account for fluorinated compound emissions toward the overall manufacturing CO_{2} -eq emissions. However, recent findings by Elgamal et al. ( elgamalenvironmental ) emphasize that fluorinated compounds can significantly affect the manufacturing carbon footprint. This highlights the need for fluorinated compound modeling in addition to existing embodied carbon modeling tools like ACT (primarily models for only CO_{2} ), for a comprehensive assessment of manufacturing CO_{2} -eq emissions. Fig. 1 shows this point. The systems community has explored the operational emissions and embodied carbon (primarily carbon dioxide emissions) part of the manufacturing emissions. The impact of fluorinated compounds in the manufacturing emissions of computing systems (contributing toward the total CO_{2} -eq emissions), remains unexplored. Fluorinated compound emissions do not correlate directly with the power grid’s energy consumption but arise from material usage in specific fabrication processes such as during chamber cleaning, plasma etching, photolithography, usage of heat transfer fluids, and others. Tracking them demands a detailed understanding of fabrication-specific practices, chemical usage profiles, gas recovery factors, and manufacturing steps at varying technology node sizes. Consequently, although these emissions dwarf carbon emissions in terms of atmospheric impact, they remain largely overlooked in systems research ( elgamalenvironmental ) . ForgetMeNot addresses a critical research gap. In this paper, we close this research gap by introducing a modeling tool, ForgetMeNot, that enables the systems community to characterize and mitigate fluorinated-compound emissions in computing hardware manufacturing. To build this tool, we first identify the fluorinated compound emission sources across the various steps of a semiconductor manufacturing pipeline: from wafer fabrication to final packaging. Based on this, we build ForgetMeNot which incorporates fabrication facility-specific chemical usage practices, node-specific lithography details (e.g., DUV v/s EUV), and hardware-specific attributes (e.g., number of cores, die area, TDP, wafer yield, nanometer scale factor, etc.) to estimate the net fluorinated compound releases for new hardware designs. We validate ForgetMeNot against real fab emissions and show that our model captures emissions within 5% of measured values across facilities. ForgetMeNot can be used to make actionable manufacturing decisions, hardware emission analysis, and building sustainable servers. ForgetMeNot enables informed design decisions and provides fabrication facilities with a tool to model emissions based on material usage, helping to minimize forever-chemical emissions without compromising performance or yield targets. For instance, we demonstrate how ForgetMeNot can model strategies to reduce emissions, such as adjusting core counts and cache sizes to minimize die area (and associated fluorinated compound usage), adopting advanced lithography techniques with fewer etch steps, optimizing manufacturing processes (e.g., reducing high-GWP fluorinated compound usage during chamber-clean frequencies with more number of etching steps), and implementing advanced gas-capture and recycling technologies. Since ForgetMeNot models fluorinated compound emissions from each source across the semiconductor manufacturing pipeline, it provides emissions estimates that can be used to explore emission reduction strategies across different manufacturing steps. These capabilities are not supported by the embodied carbon-centric models (account for carbon dioxide emissions), currently used in the computer systems community ( e.g. , ( gupta2022act ) ), as they do not capture fluorinated compound emissions and their distribution across manufacturing steps. For the broader systems community’s benefit, we design a methodology to show how public data sources (e.g., Toxic Release Inventory logs ( epaToolbox ) ) can be used to calibrate ForgetMeNot for specific fabs, enabling consistent carbon-plus-fluorine emission estimates for proposed hardware designs. Using ForgetMeNot, we analyze fluorinated emissions across hardware types, highlighting trade-offs between performance and sustainability, driven by fabrication methods and technology transitions. We show how datacenter operators can use ForgetMeNot to select hardware configurations that minimize emissions while maintaining competitive performance. Furthermore, we show that hardware determined as most sustainable when considering only embodied carbon – using existing modeling tools (e.g., ( gupta2022act ) ) that primarily account for carbon dioxide – differs from the hardware choices determined when accounting for both embodied carbon and fluorinated compounds (via ForgetMeNot). Modeling and including the fluorinated compound emissions in the total CO_{2} -eq emission estimation provides a more accurate representation of sustainable hardware. Contributions. Specifically, this paper makes the following