The Decaf Chemical War

How Decaf Coffee’s Solvent Residues Became a Dispute Over Safety and Regulation

In April 2024, the U.S. Environmental Protection Agency finalized a rule restricting most uses of methylene chloride regulated under the Toxic Substances Control Act (TSCA), with workplace controls for certain uses that would remain permitted [1]. On September 15, 2026, the Fifth Circuit vacated that rule and its associated risk determination and remanded the matter to EPA [26][27]. The ruling did not itself decide whether methylene chloride may be used to decaffeinate coffee, which is governed separately under federal food law [3]. The court held that EPA’s whole-chemical risk determination violated TSCA’s requirements and that its assumption about workers’ use of protective equipment was unsupported by substantial evidence [26]. In its earlier risk evaluation, EPA had identified neurotoxicity from short-term exposure and liver effects and cancer from long-term exposure; that evaluation covered the conditions of use EPA assessed, not an estimate of risk from drinking decaf [2]. Vacatur of the rule does not show that the chemical is safe. The solvent EPA had sought to restrict for most uses remains separately authorized for coffee decaffeination under a food regulation [3].

Section 3(2)(B)(vi) of TSCA excludes food and food additives from the definition of a chemical substance [24]. Methylene chloride’s use in decaffeinated coffee is addressed separately under Food and Drug Administration (FDA) food regulations [3]. Under those regulations, methylene chloride is authorized in coffee decaffeination up to a maximum residue of 10 parts per million (10 mg/kg) in decaffeinated roasted coffee and soluble coffee extract [3].

The central question is what evidence would justify continuing, restricting, or withdrawing the food authorization. A maximum residue limit sets a legal ceiling, not typical retail residue levels or consumer health risk, so an answer requires separating hazard classifications from measured residues and comparing the standards that govern them.

What Happens Inside the Green Bean

Decaffeination generally occurs before roasting, while coffee beans are still green [14]. Depending on the method, some aromatic compounds can be removed along with the caffeine [17]. Commercial decaffeination uses four main approaches:

Methylene chloride. In the process often called the “European Method,” green beans are treated with dichloromethane (CH₂Cl₂) [3]. Solvent processes can be direct or indirect [17]. In the direct method, the solvent is circulated through moist green beans and recaptured, and steaming then removes residues to trace levels [14]. In the indirect method, the solvent does not contact the beans directly [17].

Water-based diffusion. The Swiss Water Decaffeinated Coffee Company, with a facility in Delta, British Columbia, operates a water-based process that it describes as “chemical-free” [16]. In this type of process, a green-coffee extract passes through activated charcoal, which absorbs the caffeine, and the caffeine-reduced extract is reused to begin later batches [14].

Carbon dioxide (CO₂). A visitor to CR3-Kaffeeveredelung in Bremen, Germany, reported that green beans are soaked in water, then CO₂ under pressure circulates through an extraction vessel and selectively removes caffeine; the CO₂ is cleaned in a separating tank and reused [15].

Ethyl acetate. This organic ester (CH₃COOCH₂CH₃) is produced synthetically or derived from fermented sugarcane, as at Descafecol in Colombia, which describes using sugarcane-derived ethyl acetate [18]. Sugarcane-derived ethyl acetate is sometimes marketed as “natural,” but its origin does not establish that the process is safer or has lower environmental costs. Likewise, calling water or CO₂ methods “chemical-free” is imprecise; the accurate distinction is that they use no added organic extraction solvent.

The Regulatory Divide: United States, Europe, and California

The clearest regulatory contrast is between U.S. and EU residue limits for roasted coffee. The FDA allows up to 10 mg/kg (10 ppm) of methylene chloride in decaffeinated roasted coffee [3], while European Union Directive 2009/32/EC sets a maximum residue limit of 2 mg/kg (2 ppm) in roasted coffee [4]. The U.S. regulation also covers soluble coffee extract [3]. For roasted coffee, the U.S. limit is numerically five times the EU limit, but that difference is not evidence that U.S. consumers receive five times the exposure; actual exposure depends on measured residue concentrations and coffee consumption.

A petition filed with the FDA in December 2023 by the Environmental Defense Fund and allied petitioners sought to revoke food-additive approvals for methylene chloride and three other solvents; the agency published notice in January 2024 [7]. The petition invokes the Delaney Clause (21 U.S.C. § 348(c)(3)(A)), which addresses food additives found to induce cancer in humans or animals [8]. The 1987 Public Citizen v. Young ruling concerned color additives and does not, by itself, resolve this petition [9].

The FDA reopened the comment period in a May 28, 2026 notice, with comments closing June 29, 2026 [10]. A trade-press report published after the period closed described the decision as pending [11]. The FDA’s petition tracker listed the petition as under review in its August 19, 2026 update [25]. No later final decision was identified in the sources checked for this revision (October 21, 2026).

At the state level, California Assembly Bill 2066 originally proposed prohibiting methylene chloride in coffee decaffeination from January 1, 2027. A later amended version instead proposed requiring disclosure on product labels. The National Coffee Association opposed the bill and the federal petition, arguing that a ban would defy science and noting that regulators including the FDA and the European Food Safety Authority authorize European Method decaf [13]. The bill failed in May 2024 [12].

The Exposure Question: Hazard Classification versus Residue Measurements

The chemical’s carcinogenic hazard is established. In 2017, the International Agency for Research on Cancer classified dichloromethane as Group 2A, meaning probably carcinogenic to humans, based on sufficient evidence in experimental animals and limited evidence in humans [5].

A hazard classification is not a numerical estimate of cancer risk from a cup of coffee. Workplace and dietary exposures differ in route and dose, so evidence of occupational harm cannot, by itself, quantify the risk from residues in coffee.

Methylene chloride boils at about 39.6°C (103.3°F) at atmospheric pressure [20]. Coffee roasting involves substantially higher bean temperatures, which vary with the roasting profile [19]. The solvent is volatile, but temperature alone cannot establish how much remains; that depends on processing conditions and must be measured.

A 1987 study by FDA analyst D. L. Heikes used gas chromatography to measure chlorinated solvents in 14 ground and 11 instant decaffeinated coffees from nine brands [6]. It reported up to 640 parts per billion (0.64 mg/kg) of methylene chloride in ground decaf. In brews from only three samples of each type, the maxima were 4.8 ppb (4.8 µg/kg) for ground coffee and 1.1 ppb (1.1 µg/kg) for instant coffee [6].

At 4.8 ppb, a 240-gram cup of brewed coffee would contain about 1.15 micrograms of methylene chloride. This is an estimate based on a small sample’s maximum, not a typical current cup. The FDA’s 10 mg/kg ceiling and the EU’s 2 mg/kg limit apply to roasted coffee and dry extract rather than liquid brews, so brewed concentrations and dry-product limits measure different matrices and should not be compared directly. Because the study is historical, current exposure estimates require independent retail surveys that report sample sizes and analytical detection limits.

A prospective analysis of two U.S. cohorts examined decaf consumption and cancer risk [21], but it cannot, by itself, establish the cancer risk from methylene chloride residues. That question requires exposure estimates and evidence specific to the solvent.

Where the Global Supply Chain Leads

Decaffeination takes place in specialized facilities, including those below. The list is a selection, not a map of the industry, and it shows that alternative processes exist, not that their capacity could replace methylene chloride processing. Entries establish only what their sources document; company descriptions are the companies’ own.

• Germany, CR3-Kaffeeveredelung, Bremen: A visitor’s account describes CR3 using both methylene chloride and CO₂ methods [15].
• Colombia, Descafecol: An importer’s description says the plant uses sugarcane-derived ethyl acetate and water from the Nevado del Ruiz [18].
• Canada, Swiss Water, Delta: The company announced a new facility in Delta in 2021 and describes its process as “chemical-free” [16].

Economics and the Policy Choice Ahead

Switching away from methylene chloride involves equipment and operational trade-offs. CO₂ extraction requires pressurized extraction vessels and separation equipment [15]. Water-based processing requires carbon filtration to remove caffeine so the extract can be reused [14]. I did not find standardized, audited comparisons of energy, water, yield, and cost per kilogram across these methods, so claims about relative costs require independent evidence.

A key commercial question is whether alternative processors have sufficient capacity. A rapid phase-out could create supply pressures if they do not; this is a conditional scenario, not a forecast. The scale of any disruption would depend on available alternative capacity, the time needed to expand it, and how quickly brands could change suppliers.

Research has identified naturally caffeine-free Coffea species, such as Coffea charrieriana [23], and explored gene-editing approaches that could disrupt caffeine biosynthesis [22]. These approaches are research directions, not an existing industrial replacement for decaffeination.

Policymakers have several options, each addressing a different part of the problem:

1. Tightening residue limits: Lowering the FDA’s 10 mg/kg limit toward the EU’s 2 mg/kg standard governs product composition, but does not address workplace exposure or tell consumers how their coffee was made.
2. Mandatory process labeling: Requiring retail packages to disclose the decaffeination method (methylene chloride, ethyl acetate, water, or CO₂) gives consumers process information, but does not change residue levels or workplace exposure.
3. Phased transition and workplace controls: If regulators revoke methylene chloride’s food-additive authorization, multi-year transition timelines and occupational exposure controls would address adjustment time and worker exposure, but they depend on capacity data and do not show whether revocation is warranted.

No option resolves all of these issues. The regulatory choice turns on two unknowns: how much methylene chloride remains in coffee sold today, and how quickly other processes could absorb demand. Until both are measured, neither the legal ceiling nor the existence of alternatives can settle whether continued authorization is justified.

Source List

1. U.S. EPA, “Methylene Chloride; Regulation Under TSCA Section 6(a),” 89 Fed. Reg. 39254 (May 8, 2024). https://www.govinfo.gov/content/pkg/FR-2024-05-08/html/2024-09606.htm
2. U.S. EPA, “Final Risk Evaluation for Methylene Chloride”. https://epa.gov/assessing-and-managing-chemicals-under-tsca/final-risk-evaluation-methylene-chloride
3. 21 C.F.R. § 173.255, Methylene chloride. https://www.law.cornell.edu/cfr/text/21/173.255
4. Directive 2009/32/EC, EUR-Lex. https://eur-lex.europa.eu/eli/dir/2009/32/oj
5. IARC, Dichloromethane, Monographs Vol. 110 (2017), overall evaluation. https://www.ncbi.nlm.nih.gov/books/NBK436246/
6. Heikes, D. L., “Determination of Residual Chlorinated Solvents in Decaffeinated Coffee by Using Purge and Trap Procedure,” J. AOAC Int. 70(1):176–180 (1987). https://doi.org/10.1093/jaoac/70.1.176
7. FDA, “Filing of Food Additive Petition From Environmental Defense Fund, et al.,” 89 Fed. Reg. 1857 (Jan. 11, 2024), Docket FDA-2023-F-5684. https://regulations.justia.com/regulations/fedreg/2024/01/11/2024-00411.html
8. 21 U.S.C. § 348(c)(3)(A). https://www.law.cornell.edu/uscode/text/21/348
9. Public Citizen v. Young, 831 F.2d 1108 (D.C. Cir. 1987). https://openjurist.org/831/f2d/1108
10. FDA, “Food Additive Petition from Environmental Defense Fund, et al.; Reopening of the Comment Period,” 91 Fed. Reg. 31681 (May 28, 2026). https://www.govinfo.gov/content/pkg/FR-2026-05-28/pdf/2026-10615.pdf
11. “FDA Weighs Next Steps on Decaf Solvent Rules Following New Comments,” Daily Coffee News (June 30, 2026). Cited only for what it reported on that date. https://dailycoffeenews.com/2026/06/30/fda-weighs-next-steps-on-decaf-solvent-rules-following-new-comments/
12. California AB 2066, “The California Food Safety Act,” bill history, Digital Democracy. https://calmatters.digitaldemocracy.org/bills/ca_202320240ab2066
13. National Coffee Association, “NCA: Banning European Method decaf would defy science and harm Americans’ health” (Mar. 8, 2024). https://www.ncausa.org/Newsroom/NCA-Banning-European-Method-decaf-would-defy-science-and-harm-Americans-health
14. Clydesdale, F., “How is caffeine removed to produce decaffeinated coffee?” Scientific American. https://www.scientificamerican.com/article/how-is-caffeine-removed-t/
15. “Field trip to CR3 Decaffeination plant,” Square Mile blog. https://squaremileblog.com/2016/03/24/roasters-field-trip-to-cr3-decaffeination-plant-in-bremen-germany/
16. “Swiss Water’s New Decaffeination Flagship,” Swiss Water Decaffeinated Coffee Company (Apr. 5, 2021). https://swisswater.com/blogs/sw/swiss-waters-new-decaffeination-flagship
17. Ball, L., and Burch, E., “How is decaf coffee made? And is it really caffeine-free?” University of Queensland. https://public-health.uq.edu.au/node/11382
18. “Decafecol Ethyl Acetate Decafe from Cauca, Colombia,” InterAmerican Coffee. https://interamericancoffee.com/colombia-descafecol-cauca-decaf/
19. Schwartzberg, H., “Roasting Chemistry and Profile Roasting.” Cited for the general point that bean temperatures vary with the roasting profile, not for a specific temperature range. https://library.sweetmarias.com/wp-content/uploads/2020/08/Roasting-Chemistry-and-Profile-Roasting.pdf
20. Haynes, W. M., ed., CRC Handbook of Chemistry and Physics, 92nd ed.; PubChem CID 6344. https://pubchem.ncbi.nlm.nih.gov/compound/Dichloromethane
21. Zhang, Y., et al., “Decaffeinated coffee consumption and risk of total and site-specific cancer,” Ann. Oncol. (2025). https://doi.org/10.1016/j.annonc.2025.03.018
22. “The biological feasibility and social context of gene-edited, caffeine-free coffee,” Food Sci. Biotechnol. (2022). https://pmc.ncbi.nlm.nih.gov/articles/PMC9133285/
23. Stoffelen, P., et al., “A new caffeine-free coffee from Cameroon,” Bot. J. Linn. Soc. 158(1):67–72 (2008). https://doi.org/10.1111/j.1095-8339.2008.00845.x
24. 15 U.S.C. § 2602(2)(B)(vi). https://www.law.cornell.edu/uscode/text/15/2602
25. FDA, Food Additive Petitions Under Review (petition tracker), entry for FAP 4A4839 (Environmental Defense Fund, et al.), update dated August 19, 2026. https://hfpappexternal.fda.gov/scripts/fdcc/index.cfm?set=FAP-CAP
26. East Fork Enterprises, Inc. v. EPA, No. 24-60227 (consolidated with No. 24-60256) (5th Cir. Sept. 15, 2026) (vacating 89 Fed. Reg. 39254 and the associated unreasonable-risk determination and remanding to EPA). https://www.ca5.uscourts.gov/opinions/pub/24/24-60256-CV0.pdf
27. Arnold & Porter, “Fifth Circuit Vacates EPA’s TSCA Risk Management Rule for Methylene Chloride” (Sept. 16, 2026). Secondary summary confirming the date and holdings. https://www.arnoldporter.com/en/perspectives/blogs/environmental-edge/2026/09/fifth-circuit-vacates-epas-tsca-risk-management-rule-for-methylene-chloride