Pith. sign in

REVIEW 4 major objections 4 minor 83 references

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

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read ForgetMeNot: a systems-level model of fluorinated 'forever chemical' emissions in chip manufacturing, claimed accurate to within 5% of real fab emission records.

desk verdict First systems-level tool for fluorinated emissions in computing hardware, with open code and a timely qualitative message, but the 'within 5%' claim is in-sample and needs reframing. read the letter →

arxiv 2505.11236 v1 pith:AHXXQ5CK submitted 2025-05-16 cs.ET cond-mat.mtrl-scics.ARcs.DC

Rohan Basu Roy , Raghavendra Kanakagiri , Yankai Jiang , Devesh Tiwari This is my paper · ORCID
classification cs.ETcond-mat.mtrl-scics.ARcs.DC
keywords sustainablecomputingforeverchemicalsPFASfluorinatedcompoundssemiconductormanufacturingemissionmodelingembodiedcarbondatacenterservers
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that the sustainability conversation about computing hardware is incomplete when it counts carbon dioxide only, because the manufacturing process releases fluorinated compounds—'forever chemicals' with global warming potentials thousands of times that of carbon dioxide. To close the gap it introduces ForgetMeNot, a modeling tool that estimates fluorinated emissions from twelve manufacturing sources by combining fabrication-facility practices (gas recovery factors, wafer yield, usage coefficients) with the specifications of the hardware being made (die area, core and cache counts, TDP, node size, lithography type). The paper's central claim is that ForgetMeNot matches real facility-reported emissions within 5% across four hardware types and three different facilities, and therefore generalizes across fabs and components. If the claim holds, it gives hardware designers and datacenter operators a practical way to see and reduce the fluorinated part of manufacturing emissions rather than treating them as invisible.

What carries the argument

The load-bearing identity is the per-source emission formula $$E_i = N_{\text{Wafers}} \times \text{Usage}_i \times (1-\eta_{\text{Rec}}) \times \text{GWP}_i,$$ summed over twelve sources. $N_{\text{Wafers}}$ is derived geometrically from die area, usable wafer area, wafer diameter, and yield, so that larger dies or lower yields mean more wafers and more emissions. Each source's usage is not a free constant: it is anchored to a reference older-generation component and scaled by node-size exponents ($\alpha$), die area, TDP, package size, and a lithography factor $\phi_{\text{Lith}}$, which is the mechanism that makes EUV lithography less emissive than DUV. The gas recovery factor $\eta_{\text{Rec}}$ (about 0.9 at advanced facilities) appears in every source, which is why the paper treats small improvements in capture as a first-order lever for emission reduction.

What would settle it

A decisive check would be direct process-level metering in one fab for one product line: measuring the mass of each fluorinated gas (NF3, SF6, CF4, C2F6, CHF3, C3F8, C4F8) consumed and captured across etching, chamber cleaning, and lithography, and comparing those source-level quantities to ForgetMeNot's per-source estimates. If the source-level estimates miss the metered values systematically while the facility total still lands within 5%, the model's internal scaling would be falsified even though its headline accuracy survived.

Watch

Extended reading notes

Core claim

ForgetMeNot's central claim is that the fluorinated compound emissions of manufacturing a piece of computing hardware can be predicted from a small set of hardware specifications and fab-level parameters, and that the prediction matches what fabrication facilities actually report. The model decomposes the manufacturing pipeline into twelve emission sources—etching, chamber cleaning, photolithography, heat-transfer fluids, solvent fluids, dielectric fluids, wafer thinning, testing, soldering, vacuum pumps, pulsed-plasma nanocoatings, and packaging—and computes each source's contribution as wafers used times per-wafer fluorinated compound usage times the fraction not captured by gas recovery times the global warming potential of the gases involved. Per-wafer usage is scaled from a reference older-generation component using the ratio of node sizes, die-area terms for cores and cache, TDP, package size, and a lithography factor that makes EUV patterning cheaper than DUV. The paper validates the summed model against public facility-level emission records during the production of four real products—a server CPU fabricated in Oregon, a competing CPU fabricated in Dresden, a DRAM module fabricated in Texas, and a solid-state drive fabricated in Idaho—and reports a maximum difference of 4.7% between modeled and reported total emissions. On that basis it claims ForgetMeNot generalizes across different fabrication facilities and hardware types.

Load-bearing premise

The accuracy claim rests on the assumption that public facility-level emission reports, after subtracting a 10-mile-radius neighborhood baseline, truly measure emissions from the specific hardware products being modeled—even though the same reports were used to fit the model's usage coefficients.

Editorial extensions

If this is right

  • Smaller manufacturing nodes increase fluorinated emissions: the paper reports a 17.5% rise moving from 10 nm to 7 nm and a further 26.9% rise moving from 7 nm to 5 nm for a high-end server CPU made with EUV lithography.
  • Switching from DUV to EUV lithography lowers fluorinated emissions enough that a later CPU generation can have a smaller fluorinated footprint than its predecessor, even as embodied carbon keeps rising.
  • DDR5 memory modules have lower fluorinated compound and embodied carbon footprints per gigabyte than DDR4 modules of the same capacity, across multiple vendors.
  • For storage, per-terabyte fluorinated and carbon footprints fall by roughly 25–40% as capacity increases, and SATA HDDs have substantially lower manufacturing footprints than SSDs.
  • Datacenter operators choosing lowest-emission components can cut manufacturing emissions by roughly 51–55% for general-purpose, compute-optimized, and memory-optimized servers compared with median configurations, while the fastest configurations buy about 8.5% more throughput at about 52% more manufacturing emissions in the general-purpose case.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the model is calibrated on facility-wide reported totals, its twelve-way breakdown by source is an allocation based on process recipes rather than a direct measurement; the aggregate 5% accuracy does not by itself verify the per-source split.
  • The same parameterized structure could be extended to forecast emissions for process changes not yet in production, such as high-NA EUV or new lower-GWP cleaning gases, as long as step-count and reactivity data are available.
  • The dependence of optimal hardware choices on the GWP time horizon is a sensitivity worth probing: the paper notes the lowest-emission storage-optimized configuration under a 500-year horizon becomes tenth-lowest under a 20-year horizon, so procurement rankings may shift with climate-policy timeframes.
  • If fabs ever publish source-level gas metering, the model's exponents and per-source coefficients could be fitted independently instead of through aggregate totals, turning validation from a single number into twelve testable predictions.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. ForgetMeNot introduces a bottom-up model of fluorinated-compound (PFAS/HFC/etc.) emissions from semiconductor manufacturing for computing hardware. The model decomposes emissions into twelve sources (etching, chamber cleaning, photolithography, heat transfer fluids, solvents, dielectrics, wafer thinning, testing, soldering, vacuum pumps, pulsed-plasma nanocoatings, packaging), with usage per source expressed as base coefficients times die-area/step-count scalings that depend on technology node and lithography type (Table 1). Facility-specific base coefficients are calibrated from EPA TRI facility totals after subtracting a 10-mile-radius neighborhood baseline (Sec. 4.3), and the model is then validated against TRI-derived facility emissions for an Intel CPU, an AMD CPU, Samsung DRAM, and a Micron SSD, with a claimed maximum difference of 4.7% (Sec. 4.4, Fig. 9). The paper then uses the model to compare fluorinated-compound footprints across CPU generations, vendors, DRAM generations, and storage types (Sec. 5), demonstrates fab-level reduction levers (Sec. 4.2), and assembles low-emission server configurations that trade performance against manufacturing emissions (Sec. 6). The artifact is open-sourced at Zenodo.

Significance. If the validation claim were established, ForgetMeNot would be a genuinely useful first systems-level accounting of fluorinated emissions for computing hardware, complementing embodied-carbon tools such as ACT. The paper is also commendable for open-sourcing the tool, for decomposing emissions by manufacturing source, and for showing that including fluorinated compounds can change which hardware is ranked as most sustainable. However, the central accuracy claim currently rests on an in-sample comparison: the same TRI facility totals used to fit the usage coefficients are reused as the validation ground truth. With 26 free parameters and only four facility-level totals, agreement within 4.7% is a goodness-of-fit statement rather than an out-of-sample accuracy result. The load-bearing attribution premise, that a TRI facility total minus a neighborhood baseline equals emissions from the specific modeled hardware, is also unverified. These issues make the headline 'within 5%' claim unsupported as presented, although the modeling framework itself is plausible and worth developing.

major comments (4)
  1. [Sec. 4.3 vs Sec. 4.4, Fig. 9] The validation is in-sample. Sec. 4.3 derives facility-specific base usage coefficients (kEtch, kClean, kPhoto, etc.) from TRI facility emissions after baseline subtraction, and Sec. 4.4 then compares ForgetMeNot's output with 'the emissions value from the TRI dataset during the time when the hardware being evaluated was being manufactured.' With approximately 26 parameters and only four facility-level totals, a maximum difference of 4.7% does not establish that the model 'captures emissions within 5% of measured values across facilities.' I ask for an out-of-sample evaluation: for example, calibrate on a subset of facilities or years and validate on held-out facilities or years, or compare against independent, facility-reported fluorinated emissions (e.g., GHGRP electronics-manufacturing data) rather than the same TRI records used for calibration.
  2. [Sec. 4.3, 10-mile baseline; Fig. 9 (left)] The attribution from a facility total to a specific hardware component is not established. A fab typically manufactures many product families and technology nodes simultaneously, and the 10-mile-radius subtraction cannot separate fab process emissions from other local sources or from emissions attributable to other products manufactured in the same facility and time window. Moreover, Fig. 9 (left) labels per-source values as 'Actual Emissions from Fabrication Facility,' but TRI reports facility-level totals, not per-source breakdowns, so the source-level comparison cannot be supported by the cited data. Please provide evidence that the residual TRI value is attributable to the modeled component, or present the validation at the facility level with a clear description of what the TRI residual can and cannot identify.
  3. [Table 1 and Sec. 4.1] Several load-bearing parameters are hand-set or stated as 'typically' values without uncertainties: all alpha exponents, phi_Lith = 0.8, the 1.25x lower-GWP cleaning-step multiplier, and the base coefficients' typical magnitudes. Because these parameters directly determine the emissions computed for new hardware (and hence all conclusions in Secs. 5 and 6), the paper needs a sensitivity analysis: for example, vary alpha_etch, alpha_clean, alpha_photo, alpha_time, and phi_Lith over plausible ranges and report how the 'within 5%' claim and the server-ranking conclusions change. Without such an analysis, the 5% claim carries no error bars and the qualitative conclusions in Sec. 6 cannot be assessed for robustness.
  4. [Sec. 4.4 and Fig. 9 (right)] The claim that ForgetMeNot 'can effectively generalize modeling across different fabrication facilities and hardware types' is supported by only four data points (Intel Oregon, AMD Dresden, Samsung Texas, Micron Idaho), each corresponding to a different hardware type and facility. This is too few points to establish generalization, especially because the validation target is also the calibration target. Please report the number of facilities and hardware types, provide per-facility parameter fits and their uncertainties, and, if possible, include more facilities or a temporal hold-out so that the generalization claim is statistically meaningful.
minor comments (4)
  1. [Fig. 3 and Sec. 2] The sentence in Sec. 2 stating that fluorinated compound emissions have 'consistently been 6 to 7 orders of magnitude higher than embodied carbon emissions' appears inconsistent with Fig. 3, where both quantities are plotted on the same scale in MMTCE; please check the units and the intended comparison.
  2. [Captions, Figs. 3 and 10] The word 'Foorptint' appears in the captions of Fig. 3 and Fig. 10; it should be 'Footprint.'
  3. [Sec. 4.2, Fig. 7] The statement that reducing cores from 32 to 22 while increasing cache from 48 MB to 64 MB 'maintains similar performance (0.1 TFLOPs)' is asserted without a benchmark reference or a model; please provide the basis for the performance equivalence or soften the claim.
  4. [Sec. 4.1, Table 1] The paper would be easier to reproduce if Table 1 (or an appendix) listed the exact numeric values used for each k coefficient, each alpha exponent, phi_Lith, eta_rec, yield, gamma_usable, and the reference-hardware quantities for each of the four validated facilities, rather than only 'typically' values in the text.

Circularity Check

1 steps flagged · score 8.0 of 10

The 'within 5%' validation is in-sample: the TRI facility totals used to determine base usage coefficients in Sec. 4.3 are reused as the ground-truth target in Sec. 4.4.

  1. fitted input called prediction [Section 4.3 methodology and Section 4.4 validation (Eq. 1 and Table 1 usage models)]
    "Here, we compare the effectiveness of ForgetMeNot's fluorinated compound emission modeling (Sec. 4.1) by comparing it with the emissions from fabrication facilities (determined via collecting the emissions value from the TRI dataset during the time when the hardware being evaluated was being manufactured, as described in Sec. 4.3)."

    Sec. 4.3 states that the base usage coefficients (kEtch, kClean, kPhoto, etc.) are determined from the TRI dataset: 'we use the Toxics Release Inventory (TRI) dataset ... to determine the base usage coefficients by each source.' Sec. 4.4 then validates against the same TRI facility totals from the hardware's manufacturing period. With no temporal holdout, the reported 'maximum difference of 4.7%' is an in-sample goodness-of-fit check, not an out-of-sample prediction. The model is a sum of terms linear in these coefficients, so matching the facility totals is built into the coefficient-determination step.

full rationale

ForgetMeNot's source equations (Eq. 1 and Table 1) are linear in the base usage coefficients, and Sec. 4.3 says these coefficients are determined from facility TRI emissions after subtracting a 10-mile neighborhood baseline. Sec. 4.4 validates the model by comparing it with the same TRI emissions 'during the time when the hardware being evaluated was being manufactured.' The reported 4.7% maximum difference is therefore a statement about how well the coefficient-determination procedure reproduces its own calibration target; it is not an out-of-sample accuracy result. The paper presents no temporal or facility-level holdout, and the fab total cannot be attributed uniquely to the modeled hardware because a fab makes many products and nodes. The scaling exponents and phi_Lith are hand-set 'typical' values without uncertainty, so the 5% claim has no error bars. For these reasons, the central validation claim reduces by construction to the calibration step. The remaining analyses are model extrapolations whose reliability depends on that unvalidated calibration; no independent external benchmark is provided. I do not see any load-bearing self-citation (the main external references are EPA, IPCC, and published fab studies), so the circularity is specifically the fitted-input-called-prediction pattern.

Assumptions & free parameters 26 free parameters · 4 assumptions · 0 invented entities

The central model depends on many free parameters: 12 base usage coefficients k, 10 scaling exponents alpha, a lithography factor phi_lith, recovery and yield factors, and a 1.25x cleaning multiplier. These are fit or hand-set without uncertainty quantification. The axioms are standard emission-inventory linearity and the assumption that TRI data represent fab emissions. No new physical entities are introduced.

free parameters (26)
  • k_etch = 0.005 g/mm2/step
    Base etching usage coefficient at Intel Oregon DUV fab; fitted to TRI emissions, no error bar.
  • k_clean = 0.003 g/mm2/step
    Base chamber-cleaning usage coefficient; fitted or hand-set.
  • k_photo = 0.0007 g/mm2/step
    Base photolithography usage coefficient.
  • k_HTF = 0.0025 g/hour/W
    Base heat-transfer-fluid usage coefficient.
  • k_solv = 0.001 g/mm2/step
    Base solvent-fluid usage coefficient.
  • k_dielec = 0.01 g/mm2
    Base dielectric-fluid usage coefficient.
  • k_thin = 0.0002 g/mm2
    Base wafer-thinning usage coefficient.
  • k_test = 0.0001 g/mm2/step
    Base testing usage coefficient.
  • k_VPS = 0.4 g/step
    Base vapor-phase-soldering usage coefficient.
  • k_vacuum = 0.02 g/step
    Base vacuum-pump usage coefficient.
  • k_PPNC = 0.0001 g/mm2/step
    Base pulsed-plasma-nanocoating usage coefficient.
  • k_pack = 0.0002 g/mm2
    Base packaging usage coefficient.
  • alpha_etch = 0.5
    Scaling exponent for etch steps vs node size; hand-set.
  • alpha_clean = 0.5
    Scaling exponent for cleaning steps; hand-set.
  • alpha_photo = 1.0
    Scaling exponent for photolithography steps; hand-set linear.
  • alpha_time = 0.5
    Scaling exponent for processing time; hand-set.
  • alpha_solv = 0.5
    Scaling exponent for solvent steps; hand-set.
  • alpha_test = 1.0
    Scaling exponent for testing count; hand-set.
  • alpha_VPS = 1.0
    Scaling exponent for soldering steps vs package size; hand-set.
  • alpha_vacuum = 0.8
    Scaling exponent for vacuum operations; hand-set.
  • alpha_PPNC = 1.0
    Scaling exponent for plasma coating steps; hand-set.
  • phi_lith = 0.8 (EUV); 1.0 (DUV)
    Lithography efficiency factor; chosen from observed EUV step reduction.
  • eta_rec = 0.9
    Gas recovery factor for Intel Oregon fab; assumed high-efficiency capture.
  • yield = 0.8
    Wafer yield factor; assumed for example CPU.
  • gamma_usable = 0.95
    Usable wafer area factor; assumed for example CPU.
  • lower-GWP cleaning step multiplier = 1.25x
    Factor increasing cleaning steps when switching to CHF3 and C4F8; derived from relative gas reactivity, not measured.
assumptions (4)
  • domain assumption Total emissions are the sum over sources of N_wafers * Usage_i * (1 - eta_rec) * GWP_i (Eq. 1).
    This assumes linear additivity and no interactions between sources, and that the only factors are wafer count, usage, recovery, and GWP. This is the standard inventory approach but is a modeling choice.
  • domain assumption The TRI facility-level emissions minus a 10-mile-radius neighborhood average isolate the fab's fluorinated compound emissions.
    Used in Sec. 4.3 to estimate k coefficients and in Sec. 4.4 as validation ground truth; assumes no other nearby emitters and no systematic bias in the baseline.
  • ad hoc to paper Emission source step counts and usage scale as (N_ref / N)^alpha with hand-set exponents alpha, and EUV reduces usage by factor phi_lith=0.8.
    These power laws appear in Table 1 and Sec. 4.1; they are chosen to match observed trends but are not derived from process physics, and their values directly determine the claimed predictions.
  • domain assumption The GWP values and compound-to-source mappings in Table 1 from IPCC and EPA accurately represent fab practices.
    The model's gCO2eq totals depend linearly on these external constants; while standard, the specific apportionment of compounds to sources is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing." pith.science (2026). https://pith.science/paper/AHXXQ5CK

@misc{pith2026250511236,
  author       = {Pith},
  title        = {Pith review of: ForgetMeNot: Understanding and Modeling the Impact of Forever Chemicals Toward Sustainable Large-Scale Computing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AHXXQ5CK}},
  note         = {Machine review of arXiv:2505.11236}
}
read the original 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

Figures reproduced from arXiv: 2505.11236 by the authors.

Figure 1
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 𝐶𝑂2-eq manufacturing emissions. do not account for fluorinated compound emissions toward the overall… view at source ↗
Figure 2
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 [41]). 2 Impact of fluorinated compound emissions from semiconductor manufacturing Fluorinated compounds, essential in semiconductor manufacturing, are persistent environmental pol… view at source ↗
Figure 3
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 [4]). While computing systems research has mostly focused on the carbon footprint, the embodied carbon calculations of computing systems only take into account the CO2 emissions from energy production during material procurement, ma… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Various steps of manufacturing contribute toward emissions of fluorinated compounds. Crystal growth for wafer fabrication. Wafer fabrication begins with growing pure silicon crystals via the Czochralski or Float Zone methods, producing high-purity ingots that are slice…
Figure 5
Figure 5. Figure 5: Fluorinated compound emissions vary across emission sources during fabrication and across different nanometer technology node sizes. Chemical mechanical polishing. Chemical Mechanical Polishing (CMP) ensures a flat surface after layering, using abrasive slurries with f…
Figure 6
Figure 6. Figure 6: ForgetMeNot’s models fluorinated compound emission of new hardware to be manufactured by considering fabrication facility-specific practices, fluorinated compound usage for an older reference hardware manufacturing, and specifications of the new hardware. minimal recla…
Figure 7
Figure 7. Figure 7: Fabrication facilities can use ForgetMeNot to form design and manufacturing decisions for controlling fluorinated compound emissions. Emissions from vacuum pumps and pulsed-plasma nanocoatings. Emissions from both vacuum pumps and plasma coatings are influenced by decr…
Figure 8
Figure 8. Figure 8: We develop a methodology to estimate the emission parameters of fabrication facilities using public toxic release datasets. This will help researchers to analyze emissions for manufacturing hardware at different fabrication facilities. To set parameters for the referen…
Figure 9
Figure 9. Figure 9: ForgetMeNot’s modeling closely matches the emissions from each source from fabrication facilities (left). ForgetMeNot’s modeling closely matches the total emissions from fabrication facilities for different hardware (right). CPU manufactured in Dresden (Germany), Samsu…
Figure 10
Figure 10. Figure 10: Emissions vary across used manufacturing technologies, vendors, and hardware generations. 32 64 DRAM Capacity 0 50 100 150 Carbon/ Fluorinated Compound Foorptint p e r G B (g C O 2 e q) Samsung DDR4 Embodied Carbon Fluorinated Compounds 32 64 128 DRAM Capacity 0 50 10…
Figure 11
Figure 11. Figure 11: DDR5 memory modules have lower environmental footprints than DDR4 memory modules, especially at higher DRAM memory capacities. In [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]
Figure 12
Figure 12. Figure 12: Carbon and fluorinated compound footprints per TB reduce with increased storage capacity. HDDs have a significantly lower environmental footprint than SSDs. beyond the 2𝑛𝑑 generation CPU. The drop in fluorinated compound emissions from 2𝑛𝑑 to 3𝑟𝑑 generation CPU is due…
Figure 13
Figure 13. Figure 13: Performance-optimized configurations achieve only limited improvement in throughput but result in significantly higher emissions (summation of embodied carbon and fluorinated compound emissions) compared to emission-optimized configurations. 3300 choices, considering …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

83 extracted references · 78 canonical work pages

  1. [1]

    https://www

    9 Key Statistics on New Semiconductor Fabs Being Built Around the World - Z2Data — z2data.com. https://www. z2data.com/insights/9-statistics-on-new-semiconductor-fabs-being-built

  2. [2]

    https: //www.interface-eu.org/publications/chip-productions-ecological-footprint

    Chip Production’s Ecological Footprint: Mapping Climate and Environmental Impact — interface-eu.org. https: //www.interface-eu.org/publications/chip-productions-ecological-footprint

  3. [3]

    https://eto.tech/dataset-docs/chipexplorer/

    Documentation: Advanced Semiconductor Supply Chain Dataset – Emerging Technology Observatory — eto.tech. https://eto.tech/dataset-docs/chipexplorer/

  4. [4]

    https://www.epa.gov/ghgreporting/ghgrp-electronics- manufacturing

    GHGRP Electronics Manufacturing | US EPA — epa.gov. https://www.epa.gov/ghgreporting/ghgrp-electronics- manufacturing

    work page

  5. [5]

    https://github.com/quarterwave0/LithographySimulator

    GitHub - quarterwave0/LithographySimulator: A Python-based tool for modeling optical photolithography — github.com. https://github.com/quarterwave0/LithographySimulator

    work page

  6. [6]

    https: //github.com/toUpperCase78/intel-processors

    GitHub - toUpperCase78/intel-processors: Datasets for All Processors Maufactured By Intel — github.com. https: //github.com/toUpperCase78/intel-processors

    work page

  7. [7]

    https://chemsec.org/app/uploads/2023/04/Check-your-Tech_230420.pdf

    A guide to pfas in electronics. https://chemsec.org/app/uploads/2023/04/Check-your-Tech_230420.pdf

    work page 2023

  8. [8]

    https://ark.intel.com/content/www/us/en/ark.html

    Intel product specifications — ark.intel.com. https://ark.intel.com/content/www/us/en/ark.html

    work page

Show all 83 references
  1. [9]

    https://echa.europa.eu/hot-topics/perfluoroalkyl- chemicals-pfas

    Per- and polyfluoroalkyl substances (PFAS) - ECHA — echa.europa.eu. https://echa.europa.eu/hot-topics/perfluoroalkyl- chemicals-pfas

  2. [10]

    https: //www.techinsights.com/sustainability/pfas-semiconductor-manufacturing-major-sustainability-challenge

    PFAS in Semiconductor Manufacturing: A Major Sustainability Challenge | TechInsights — techinsights.com. https: //www.techinsights.com/sustainability/pfas-semiconductor-manufacturing-major-sustainability-challenge

  3. [11]

    https://www.ipcc-nggip.iges.or.jp/public/gp/bgp/ 3_6_PFC_HFC_NF3_SF6_Semiconductor_Manufacturing.pdf

    Pfc, hfc, nf3 and sf6 emissions from semiconductor manufacturing. https://www.ipcc-nggip.iges.or.jp/public/gp/bgp/ 3_6_PFC_HFC_NF3_SF6_Semiconductor_Manufacturing.pdf

  4. [12]

    https://www.epa

    Risk Management for Per- and Polyfluoroalkyl Substances (PFAS) under TSCA | US EPA — epa.gov. https://www.epa. gov/assessing-and-managing-chemicals-under-tsca/risk-management-and-polyfluoroalkyl-substances-pfas

  5. [13]

    https://www.mckinsey.com/ industries/industrials-and-electronics/our-insights/semiconductor-fabs-construction-challenges-in-the-united- states

    Semiconductor fabs: Construction challenges in the United States — mckinsey.com. https://www.mckinsey.com/ industries/industrials-and-electronics/our-insights/semiconductor-fabs-construction-challenges-in-the-united- states

  6. [14]

    https://www.epa.gov/eps-partnership/semiconductor-industry

    Semiconductor Industry | US EPA — epa.gov. https://www.epa.gov/eps-partnership/semiconductor-industry

  7. [15]

    https://www.epa.gov/toxics-release-inventory-tri-program/tri-toolbox

    TRI Toolbox | US EPA — epa.gov. https://www.epa.gov/toxics-release-inventory-tri-program/tri-toolbox

  8. [16]

    https://archive.ics.uci.edu/dataset/179/secom

    UCI Machine Learning Repository — archive.ics.uci.edu. https://archive.ics.uci.edu/dataset/179/secom

  9. [17]

    https://www.semiconductors.org/wp-content/uploads/2023/06/FINAL-Plasma-Etch-and-Deposition-White- Paper.pdf

    White paper: Pfas-containing fluorochemicals used in semiconductor manufacturing plasma-enabled etch and depo- sition. https://www.semiconductors.org/wp-content/uploads/2023/06/FINAL-Plasma-Etch-and-Deposition-White- Paper.pdf

    2023

  10. [18]

    Carbon explorer: A holistic framework for designing carbon aware datacenters

    Bilge Acun, Benjamin Lee, Fiodar Kazhamiaka, Kiwan Maeng, Udit Gupta, Manoj Chakkaravarthy, David Brooks, and Carole-Jean Wu. Carbon explorer: A holistic framework for designing carbon aware datacenters. In Proceedings of the This work has been accepted at ACM SIGMETRICS 2025,...

    2025

  11. [19]

    Fluorinated gases for semiconductor manufacture: process advances in chemical vapor deposition chamber cleaning

    Charles C Allgood. Fluorinated gases for semiconductor manufacture: process advances in chemical vapor deposition chamber cleaning. Journal of fluorine chemistry , 122(1):105–112, 2003

    2003

  12. [20]

    Open-source carbon footprint estimator: Development and university declination

    Clément Auger, Benoit Hilloulin, Benjamin Boisserie, Maël Thomas, Quentin Guignard, and Emmanuel Rozière. Open-source carbon footprint estimator: Development and university declination. Sustainability, 13(8):4315, 2021

    2021

  13. [21]

    The sunk carbon fallacy: Rethinking carbon footprint metrics for effective carbon-aware scheduling

    Noman Bashir, Varun Gohil, Anagha Belavadi Subramanya, Mohammad Shahrad, David Irwin, Elsa Olivetti, and Christina Delimitrou. The sunk carbon fallacy: Rethinking carbon footprint metrics for effective carbon-aware scheduling. In Proceedings of the 2024 ACM Symposium on Cloud ...

    2024

  14. [22]

    Sustainable computing-without the hot air

    Noman Bashir, David Irwin, Prashant Shenoy, and Abel Souza. Sustainable computing-without the hot air. ACM SIGENERGY Energy Informatics Review , 3(3):47–52, 2023

    2023

  15. [23]

    An overview of semiconductor industry efforts to reduce pfas use and emissions in plasma processes

    Laurie S Beu and Melissa A Gresham. An overview of semiconductor industry efforts to reduce pfas use and emissions in plasma processes. Advanced Etch Technology and Process Integration for Nanopatterning XIII , 12958:162–166, 2024

    2024

  16. [24]

    Life-cycle assessment of semiconductors

    Sarah B Boyd. Life-cycle assessment of semiconductors . Springer Science & Business Media, 2011

    2011

  17. [25]

    The road to net-zero emissions in ic manufacturing

    Olga Bubnova. The road to net-zero emissions in ic manufacturing. Nature Reviews Electrical Engineering, 1(4):214–215, 2024

    2024

  18. [26]

    Sf6 and alternatives in electrical switchgear and related equipment

    Karsten Burges, Kristina Warncke, and Barbara Gschrey. Sf6 and alternatives in electrical switchgear and related equipment. European Commission Briefing Paper, March 2020

    2020

  19. [27]

    Release of perfluoroalkyl substances from melting glacier of the tibetan plateau: Insights into the impact of global warming on the cycling of emerging pollutants

    Mengke Chen, Chuanfei Wang, Xiaoping Wang, Jianjie Fu, Ping Gong, Juping Yan, Zhengliang Yu, Fangping Yan, and Javed Nawab. Release of perfluoroalkyl substances from melting glacier of the tibetan plateau: Insights into the impact of global warming on the cycling of emerging p...

    2019

  20. [28]

    Sf6 usage and emission trends in the tft-lcd industry

    Ju-Hsiu Cheng, Scott Bartos, Wei Lee, and Shou-Nan Li. Sf6 usage and emission trends in the tft-lcd industry. International Journal of Greenhouse Gas Control , 17:106–110, 09 2013

    2013

  21. [29]

    Sustainability and the US EPA

    National Research Council, Policy, Global Affairs, Technology for Sustainability Program, and Committee on Incorpo- rating Sustainability in the US Environmental Protection Agency. Sustainability and the US EPA . National Academies Press, 2011

    2011

  22. [30]

    Pfc emission reduction in the semiconductor industry

    Michael Czerniak. Pfc emission reduction in the semiconductor industry. In Light Metals 2018 , pages 1495–1498. Springer, 2018

    2018

  23. [31]

    Pfas in semiconductor photolithography: a mass balance model

    Ralph R Dammel and David E Speed. Pfas in semiconductor photolithography: a mass balance model. In Advances in Patterning Materials and Processes XLI , volume 12957, pages 40–62. SPIE, 2024

    2024

  24. [32]

    Building an aws ec2 carbon emissions dataset

    Benjamin Davy. Building an aws ec2 carbon emissions dataset. Medium, September, 2021

    2021

  25. [33]

    The linpack benchmark: past, present and future

    Jack J Dongarra, Piotr Luszczek, and Antoine Petitet. The linpack benchmark: past, present and future. Concurrency and Computation: practice and experience , 15(9):803–820, 2003

    2003

  26. [34]

    Kaya for computer architects: Toward sustainable computer systems

    Lieven Eeckhout. Kaya for computer architects: Toward sustainable computer systems. IEEE Micro, 43(1):9–18, 2023

    2023

  27. [35]

    Focal: A first-order carbon model to assess processor sustainability

    Lieven Eeckhout. Focal: A first-order carbon model to assess processor sustainability. In ACM ASPLOS, pages 401–415, 2024

    2024

  28. [36]

    The environmental impact of forever chemicals in computing systems

    Mariam Elgamal, Abdulrahman Mahmoud, Gu-Yeon Wei, David Brooks, and Gage Hills. The environmental impact of forever chemicals in computing systems

  29. [37]

    Franck, Alise Chachereau, and Juriy Pachin

    Christian M. Franck, Alise Chachereau, and Juriy Pachin. Sf6-free gas-insulated switchgear: Current status and future trends. IEEE Electrical Insulation Magazine , 37(1):7–16, 2021

    2021

  30. [38]

    A review on the sources, occurrence and health risks of per-/poly-fluoroalkyl substances (pfas) arising from the manufacture and disposal of electric and electronic products

    Shafali Garg, Pankaj Kumar, Vandana Mishra, Rosanne Guijt, Prabhjot Singh, Ludovic F Dumée, and Radhey Shyam Sharma. A review on the sources, occurrence and health risks of per-/poly-fluoroalkyl substances (pfas) arising from the manufacture and disposal of electric and electr...

    2020

  31. [39]

    Ultra-fast aerial image simulation algorithm using wavelength scaling and fast fourier transformation to speed up calculation by more than three orders of magnitude

    Tsai-Sheng Gau, Po-Hsiung Chen, Burn J Lin, Fu-Hsiang Ko, Chun-Kung Chen, and Anthony Yen. Ultra-fast aerial image simulation algorithm using wavelength scaling and fast fourier transformation to speed up calculation by more than three orders of magnitude. Journal of Micro/Nan...

    2023

  32. [40]

    Semiconductor manufacturing handbook

    Hwaiyu Geng. Semiconductor manufacturing handbook. McGraw-Hill, Inc., 2005

    2005

  33. [41]

    Global Warming Potential Values (August 2024), 2024

    GHG Protocol. Global Warming Potential Values (August 2024), 2024

    2024

  34. [42]

    Flexcooldc: Datacenter cooling flexibility for harmonizing water, energy, carbon, and cost trade-offs

    Wedan Emmanuel Gnibga, Andrew A Chien, Anne Blavette, and Anne Cécile Orgerie. Flexcooldc: Datacenter cooling flexibility for harmonizing water, energy, carbon, and cost trade-offs. In Proceedings of the 15th ACM International Conference on Future and Sustainable Energy System...

    2024

  35. [43]

    Act: Designing sustainable computer systems with an architectural carbon modeling tool

    Udit Gupta, Mariam Elgamal, Gage Hills, Gu-Yeon Wei, Hsien-Hsin S Lee, David Brooks, and Carole-Jean Wu. Act: Designing sustainable computer systems with an architectural carbon modeling tool. In Proceedings of the 49th Annual International Symposium on Computer Architecture ,...

    2022

  36. [44]

    Chasing carbon: The elusive environmental footprint of computing

    Udit Gupta, Young Geun Kim, Sylvia Lee, Jordan Tse, Hsien-Hsin S Lee, Gu-Yeon Wei, David Brooks, and Carole-Jean Wu. Chasing carbon: The elusive environmental footprint of computing. In 2021 IEEE International Symposium on This work has been accepted at ACM SIGMETRICS 2025, St...

    2021

  37. [45]

    Going green for less green: Optimizing the cost of reducing cloud carbon emissions

    Walid A Hanafy, Qianlin Liang, Noman Bashir, Abel Souza, David Irwin, and Prashant Shenoy. Going green for less green: Optimizing the cost of reducing cloud carbon emissions. In ACM ASPLOS, pages 479–496, 2024

    2024

  38. [46]

    A guide to global warming potentials (gwps)

    LD Danny Harvey. A guide to global warming potentials (gwps). Energy Policy, 21(1):24–34, 1993

    1993

  39. [47]

    Environment, safety, health and sustainability, 2023

    IEEE International Roadmap for Devices and Systems. Environment, safety, health and sustainability, 2023

    2023

  40. [48]

    Perfluorocompounds emission reduction by the semiconductor industry

    Francesca Illuzzi and Harry Thewissen. Perfluorocompounds emission reduction by the semiconductor industry. Journal of Integrative Environmental Sciences , 7(S1):201–210, 2010

    2010

  41. [49]

    Climate Change 2001: The Scientific Basis, 2001

    Intergovernmental Panel on Climate Change (IPCC). Climate Change 2001: The Scientific Basis, 2001. Archived from the original on 31 January 2016

    2001

  42. [50]

    Perfluorocarbon emissions reduction programme: Results of the 2018 Anode Effect Survey

    International Aluminium Institute. Perfluorocarbon emissions reduction programme: Results of the 2018 Anode Effect Survey. Technical Report, 2019

    2018

  43. [51]

    Ecolife: Carbon-aware serverless function scheduling for sustainable computing

    Yankai Jiang, Rohan Basu Roy, Baolin Li, and Devesh Tiwari. Ecolife: Carbon-aware serverless function scheduling for sustainable computing. arXiv preprint arXiv:2409.02085, 2024

    2024 arXiv

  44. [52]

    The semiconductor supply chain: Assessing national competitive- ness

    Saif M Khan, Alexander Mann, and Dahlia Peterson. The semiconductor supply chain: Assessing national competitive- ness. Center for Security and Emerging Technology , 8(8):1–98, 2021

    2021

  45. [53]

    D. Koch. Sf6 properties, and use in mv and hv switchgear. Cahier Technique Schneider Electric, (188), 2003

    2003

  46. [54]

    Carbon connect: An ecosystem for sustainable computing

    Benjamin C Lee, David Brooks, Arthur van Benthem, Udit Gupta, Gage Hills, Vincent Liu, Benjamin Pierce, Christopher Stewart, Emma Strubell, Gu-Yeon Wei, et al. Carbon connect: An ecosystem for sustainable computing. arXiv preprint arXiv:2405.13858, 2024

    2024 arXiv

  47. [55]

    Toward sustainable hpc: Carbon footprint estimation and environmental implications of hpc systems

    Baolin Li, Rohan Basu Roy, Daniel Wang, Siddharth Samsi, Vijay Gadepally, and Devesh Tiwari. Toward sustainable hpc: Carbon footprint estimation and environmental implications of hpc systems. In Proceedings of the International Conference for High Performance Computing, Networ...

    2023

  48. [56]

    thirsty

    Pengfei Li, Jianyi Yang, Mohammad A Islam, and Shaolei Ren. Making ai less" thirsty": Uncovering and addressing the secret water footprint of ai models. arXiv preprint arXiv:2304.03271, 2023

    2023 arXiv

  49. [57]

    Digital & environment: How to evaluate server manufacturing footprint, beyond greenhouse gas emissions?, 2021

    Romain Lorenzini. Digital & environment: How to evaluate server manufacturing footprint, beyond greenhouse gas emissions?, 2021

    2021

  50. [58]

    Demonstrating gwp*: a means of reporting warming-equivalent emissions that captures the contrasting impacts of short-and long-lived climate pollutants

    John Lynch, Michelle Cain, Raymond Pierrehumbert, and Myles Allen. Demonstrating gwp*: a means of reporting warming-equivalent emissions that captures the contrasting impacts of short-and long-lived climate pollutants. Environmental Research Letters, 15(4):044023, 2020

    2020

  51. [59]

    Untangling carbon-free energy attribution and carbon intensity estimation for carbon-aware computing

    Diptyaroop Maji, Noman Bashir, David Irwin, Prashant Shenoy, and Ramesh K Sitaraman. Untangling carbon-free energy attribution and carbon intensity estimation for carbon-aware computing. In Proceedings of the 15th ACM International Conference on Future and Sustainable Energy S...

    2024

  52. [60]

    Semiconductor: Supply chain deep dive assessment

    Margaret Mann and Vicky Putsche. Semiconductor: Supply chain deep dive assessment. Technical report, USDOE Office of Policy (OP), Washington DC (United States), 2022

    2022

  53. [61]

    Electricity Maps Live 24/7, 2025

    Electricity Maps. Electricity Maps Live 24/7, 2025

    2025

  54. [62]

    Semiconductor lithography: principles, practices, and materials

    Wayne M Moreau. Semiconductor lithography: principles, practices, and materials . Springer Science & Business Media, 2012

    2012

  55. [63]

    Right to Know Hazardous Substance Fact Sheet: Nitrogen Trifluoride, March 2001

    New Jersey Department of Health. Right to Know Hazardous Substance Fact Sheet: Nitrogen Trifluoride, March 2001. Accessed: 2024-11-10

    2001

  56. [64]

    Dry etching technology for semiconductors

    Kazuo Nojiri. Dry etching technology for semiconductors . Springer, 2015

    2015

  57. [65]

    2019 refinement to the 2006 ipcc guidelines for national greenhouse gas inventories, 2019

    Intergovernmental Panel on Climate Change (IPCC). 2019 refinement to the 2006 ipcc guidelines for national greenhouse gas inventories, 2019

    2019

  58. [66]

    Achieving sustainability in the semiconductor industry: The impact of simulation and ai

    Wojciech T Osowiecki, Martyn J Coogans, Saravanapriyan Sriraman, Rakesh Ranjan, Yu Lu, and David M Fried. Achieving sustainability in the semiconductor industry: The impact of simulation and ai. IEEE Transactions on Semiconductor Manufacturing, 2024

    2024

  59. [67]

    GHG emissions of semiconductor manufacturing in 2021

    Maxime Pelcat. GHG emissions of semiconductor manufacturing in 2021 . PhD thesis, Univ Rennes, INSA Rennes, CNRS, IETR–UMR 6164, F-35000 Rennes, 2023

    2021

  60. [68]

    Semiconductor manufacturing technology, volume 1

    Michael Quirk and Julian Serda. Semiconductor manufacturing technology, volume 1. Prentice Hall Upper Saddle River, NJ, 2001

    2001

  61. [69]

    The chip manufacturing industry: Environmental impacts and eco-efficiency analysis

    Marcello Ruberti. The chip manufacturing industry: Environmental impacts and eco-efficiency analysis. Science of the Total Environment, 858:159873, 2023

    2023

  62. [70]

    The carbon footprint of pecvd chamber cleaning using fluorinated gases

    Martin Schottler and Mariska de Wild-Scholten. The carbon footprint of pecvd chamber cleaning using fluorinated gases. In Proc. 23rd European Photovoltaic Solar Energy Conference, Valencia, Spain , pages 1–5, 2008

    2008

  63. [71]

    Ecovisor: A virtual energy system for carbon-efficient applications

    Abel Souza, Noman Bashir, Jorge Murillo, Walid Hanafy, Qianlin Liang, David Irwin, and Prashant Shenoy. Ecovisor: A virtual energy system for carbon-efficient applications. In Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languag...

    2023

  64. [72]

    Carbon footprint of a dell rack server

    Markus Stutz, Scott O’Connell, and John Pflueger. Carbon footprint of a dell rack server. In 2012 Electronics Goes Green 2012+, pages 1–5. IEEE, 2012

    2012

  65. [73]

    Sapatnekar, and Vidya A

    Chetan Choppali Sudarshan, Nikhil Matkar, Sarma Vrudhula, Sachin S. Sapatnekar, and Vidya A. Chhabria. ECO-CHIP: Estimation of Carbon Footprint of Chiplet-based Architectures for Sustainable VLSI . In 2024 IEEE International Symposium on High-Performance Computer Architecture ...

    2024

  66. [74]

    On the limitations of carbon- aware temporal and spatial workload shifting in the cloud

    Thanathorn Sukprasert, Abel Souza, Noman Bashir, David Irwin, and Prashant Shenoy. On the limitations of carbon- aware temporal and spatial workload shifting in the cloud. In Proceedings of the Nineteenth European Conference on Computer Systems, pages 924–941, 2024

    2024

  67. [75]

    Environmental Protection Agency

    U.S. Environmental Protection Agency. PFAS Explained, October 2023

    2023

  68. [76]

    U.S. EPA. Global mitigation of non-CO2 greenhouse gases: 2010-2030. Technical Report, 2013

    2010

  69. [77]

    FabSim, a semiconductor factory (fab) simulator — fabsim.com

    Holger Vogt. FabSim, a semiconductor factory (fab) simulator — fabsim.com. http://www.fabsim.com/

  70. [78]

    A new method to determine the tool count of a semiconductor factory using fabsim

    Holger Vogt. A new method to determine the tool count of a semiconductor factory using fabsim. In Proceedings of the 2004 Winter Simulation Conference, 2004. , volume 2, pages 1925–1929. IEEE, 2004

    2004

  71. [79]

    Designing cloud servers for lower carbon

    Jaylen Wang, Daniel S Berger, Fiodar Kazhamiaka, Celine Irvene, Chaojie Zhang, Esha Choukse, Kali Frost, Rodrigo Fonseca, Brijesh Warrier, Chetan Bansal, et al. Designing cloud servers for lower carbon. In ACM/IEEE ISCA, pages 452–470, 2024

    2024

  72. [80]

    Designing cloud servers for lower carbon

    Jaylen Wang, Daniel S Berger, Fiodar Kazhamiaka, Celine Irvene, Chaojie Zhang, Esha Choukse, Kali Frost, Rodrigo Fonseca, Brijesh Warrier, Chetan Bansal, Jonathan Stern, Ricardo Bianchini, and Akshitha Sriraman. Designing cloud servers for lower carbon. In International Sympos...

    2024

  73. [81]

    List of semiconductor fabrication plants

    Wikipedia contributors. List of semiconductor fabrication plants. https://en.wikipedia.org/wiki/List_of_semiconductor_ fabrication_plants

  74. [82]

    Extreme ultraviolet lithography and three dimensional integrated circuit—a review

    Banqiu Wu and Ajay Kumar. Extreme ultraviolet lithography and three dimensional integrated circuit—a review. Applied Physics Reviews, 1(1), 2014

    2014

  75. [83]

    Euv lithography

    Stefan Wurm, Winfried Kaiser, Udo Dinger, Stephan Müllender, Bruno La Fontaine, Obert R Wood, and Mark Neisser. Euv lithography. In Microlithography, pages 163–246. CRC Press, 2020. This work has been accepted at ACM SIGMETRICS 2025, Stony Brook, New York, USA

    2020

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.