Industries / Consumer goods / Candles and home fragrance
Packaging and labels
This is the work of choosing and testing the jars, tins, lids, boxes and warning labels that candles and diffusers are sold in, and meeting each market's labelling and packaging rules.
6
Milestones achieved, from 2000 to 2026.
3
To watch for: announced, not yet confirmed outside.
3
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
2000
US candle warning label standard
ASTM published F2058-00, "Standard Specification for Cautionary Labeling for Candles Burned in a Home", which set nationally recognised warning information for candles as sold. Its 2007 revision, renamed Candle Fire Safety Labeling, is the version the CPSC cited in 2014.
Credit
ASTM International, subcommittee F15.45
2002
Glass jar standard for candles
ASTM F2179, first issued as F2179-02, set minimum requirements for soda-lime glass candle containers: annealing (slow cooling of the glass to remove stress), thermal shock tests and scratch tests, to keep leftover stress in the glass at a "commercially acceptable level". The 2026 edition limits the standard to glass made to be filled with candle wax and excludes votive holders and hurricane glasses.
Credit
ASTM International
2007
European pictogram safety labels
EN 15494, published in the UK as BS EN 15494 on 30 November 2007, set product safety labels for candles across Europe, using a general warning sign plus extra symbols or text. The 2019 edition replaced it, fixing a minimum symbol height of 5 mm, stating which safety information is required for which type of candle, and adding options for small labels.
Credit
European Committee for Standardization (CEN), published nationally by BSI and others
2023
Millions of cracking jars recalled
Target recalled about 4.9 million Threshold glass-jar candles on 18 May 2023 after 137 reports of jars breaking or cracking during use and six injuries, including cuts and severe burns, then about 2.2 million more on 10 August 2023 after 19 further reports; both lots were made in Vietnam. In February 2023 about 1.21 million Mainstays three-wick candles sold at Walmart were recalled after 12 reports of flames burning too close to the side of the jar and the glass cracking. The 4.9 million is the largest unit count among the 187 records the CPSC recall database returns for "candle" (count made for this page).
LessonA glass standard does not stop failures that come from the wick, flame and fill working together.
Credit
CPSC with Target Corporation and Star Soap Star Candle Prayer Candle
2024
Open jars not CLP packaging
The European Chemicals Agency (ECHA) issued guidance in August 2024 that candles are chemical mixtures under the EU's CLP rules on classifying, labelling and packaging chemicals. A scented candle with a skin-sensitising fragrance above a set level must name it with the statement EUH208 ("Contains (name). May produce an allergic reaction."), and a candle classed as hazardous needs a notification to poison centres and a UFI, a unique formula code, on its label. ECHA also said an open glass container, a tealight's aluminium cup or an openwork paper box is not adequate packaging for a hazardous mixture, so the label must go on outer packaging.
Credit
European Chemicals Agency
2026
Candle jars outside packaging law
The EU Packaging and Packaging Waste Regulation (Regulation (EU) 2025/40) has applied since 12 August 2026. The Commission's guidance of 5 June 2026 says tealight cups, grave light containers and "other candle containers such as filled glasses and ceramic bowls" are not packaging, so the regulation's rules on recycling, recycled content, reuse and labelling reach a candle's box, inserts and shipping carton, not the jar. The regulation also caps empty space in e-commerce packaging at 50% from 2030.
Credit
European Commission
Watch for
Diptyque refills its classic candle
Diptyque relaunched its Classic Candle in a glass jar about 10% lighter, and says ten core scents will be sold as wax refills from autumn 2026. It says keeping the jar and buying a refill cuts the product's carbon footprint by about 22% (as reported by Homes & Gardens) or 24% (Moodie Davitt Report); neither report gives a method.
Achieved whenrefills are on sale and an independent party confirms the fit and the footprint figure.
Claimed by
Diptyque, April 2026
CLP label rules rewritten
Regulation (EU) 2025/2439, the "stop the clock" law in force from 23 December 2025, pushed the revised CLP rules on label format, advertising and online sales to 2028. A provisional deal between the European Parliament and Council on 17 June 2026 would keep minimum font sizes only for products sold to the public: an x-height, the height of lower-case letters, of at least 1.2 mm, or 0.9 mm for packs of 125 ml or less.
Achieved whenthe final text is published in the Official Journal and applies to scented candles sold in shops and online.
Claimed by
European Parliament and Council, June 2026
Fracture atlas for candle jars
Forensic engineers report lab tests that caused flashover, wick drift, secondary ignition and fire outside the jar, and compare the glass fracture patterns with real cases, in a paper that also reviews proposed changes to ASTM F2179.
Achieved whenthe full results are published and used by ASTM or by other investigators.
Claimed by
David Riegner and Gregory S. Chojecki, SEA, Ltd., Columbus, Ohio, at the ASM IMAT conference, October 2026
Open
01
Audit EU candle listing labels
A public, repeatable audit of scented candles and reed diffusers sold online in the EU that checks listing photos and text for the CLP label elements (hazard pictogram, EUH208 allergen statement, UFI code, supplier address and local language) and the EN 15494 fire safety warnings, with every finding linked to a dated screenshot.
The briefCredit · unclaimed02
Test refill fit and savings
A published bench test of candle refill inserts that records whether each refill fits the jars it is sold for, including older versions of the jar, how much packaging it uses compared with a full candle, and how it burns.
The briefCredit · unclaimed03
Log jar wall temperatures
An open dataset of glass wall and base temperatures through full, label-following burns of common shop-bought container candles, with wick position and flame height recorded, so that jar makers and investigators have reference curves.
The briefCredit · unclaimed
Open 01 · the brief
Audit EU candle listing labels
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public, repeatable audit of scented candles and reed diffusers sold online in the EU that checks listing photos and text for the CLP label elements (hazard pictogram, EUH208 allergen statement, UFI code, supplier address and local language) and the EN 15494 fire safety warnings, with every finding linked to a dated screenshot.
Why it is still open
The closest published check, an Irish Health and Safety Authority project reported in 2024, found 60% of 99 scented products failed the hazard labelling rules (78% of those from outside the EU), but it covered sprays, mists, air fresheners and diffuser oils rather than candles and published no product-level data. The CLP rules for online sales are deferred to 2028, so a baseline taken now would show the starting point.
Built with what exists
Playwright to capture listing pages and images; Tesseract to read text from label photos; a small low-cost AI model such as Meta's Muse Ember to pull label elements into a table (see the Note "The jobs agents repeat"); Label Studio for a second check by a person; a checklist drawn from the ECHA candle guidance and EN 15494.
First step
Write the checklist and capture 30 candle listings from one marketplace.
Done looks like
At least 300 listings from at least five EU marketplaces in at least three languages; the share of listings showing each label element; at least 90% agreement between two human checkers on a sample of 50; screenshots and checklist published.
Open 02 · the brief
Test refill fit and savings
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A published bench test of candle refill inserts that records whether each refill fits the jars it is sold for, including older versions of the jar, how much packaging it uses compared with a full candle, and how it burns.
Why it is still open
Brand claims have no public test. Diptyque's 22 to 24% footprint claims come with no method, and Element Brooklyn says it "invented the candle refill" and gives no fit data. Searches found seller pages and no independent test of fit or packaging weight.
Built with what exists
A 0.1 g kitchen scale; digital calipers; a camera on a fixed stand; openLCA for a simple footprint estimate; ordinary burning, attended, in a ventilated room on a heat-proof tray, following the label instructions.
First step
Buy three refills and their matching full candles, then weigh and photograph every packaging part.
Done looks like
At least 10 refill products from at least four brands, each tried in three jars; a fit rate with gaps measured in millimetres, grams of packaging per refill against a full candle, and burn photos, all published as a spreadsheet.
Open 03 · the brief
Log jar wall temperatures
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open dataset of glass wall and base temperatures through full, label-following burns of common shop-bought container candles, with wick position and flame height recorded, so that jar makers and investigators have reference curves.
Why it is still open
CPSC staff note that consumer reports of jars cracking or shattering often have "limited information" and no pictures. The 2026 SEA, Ltd. study gives no measured data in its abstract, and searches found no published temperature dataset for commercial candle jars.
Built with what exists
K-type thermocouples on MAX31855 boards logged by an Arduino or Raspberry Pi; a FLIR ONE thermal camera; a ruler grid for flame height; a ventilated room, heat-proof tray and fire extinguisher. Burns stop at the end point the label states, and no candle is pushed past it.
First step
Fit one thermocouple to one jar and log a single four-hour burn.
Done looks like
30 candles from at least 10 brands, each with sensors at three heights, with full logs and photos released on Zenodo and peak glass temperature reported against the remaining wax depth.