Almost every affordable earring, chain, watch case, and jeans rivet has nickel in it somewhere. Nickel makes cheap metal harder, brighter, and slower to corrode, which is why it sits under so much plating and inside so much steel. European law has restricted it for three decades, and the restriction rests on a distinction shoppers almost never hear stated: it limits what comes off the metal, not what is in it.
A limit on what comes off, not what is in
The rule began as European Parliament and Council Directive 94/27/EC of 30 June 1994, which amended Directive 76/769/EEC for the twelfth time and set a release rate of 0.5 micrograms per square centimetre per week. It survives today as entry 27 of Annex XVII to REACH, Regulation (EC) No 1907/2006, and the figure has not moved since it was written.
Release rather than composition is the whole design. A ring built largely from a nickel-bearing alloy can comply without difficulty, provided the surface hands over less than the limit across a week. A piece containing very little nickel can fail if the small amount it has sits on the outside in a form that dissolves readily. What a seller says about how much nickel is inside tells you nothing on its own.
Two numbers for two kinds of contact
There are two limits, and the stricter one covers piercings. The 1994 directive handled post assemblies with a composition rule instead of a release rule: a post inserted into a piercing during healing had to be homogeneous, with nickel under 0.05 percent of total mass. That excluded most stainless steel outright, because nickel is part of what gives steel its corrosion resistance in the first place.
Commission Directive 2004/96/EC of 27 September 2004 replaced the composition test with a release limit of 0.2 micrograms per square centimetre per week, on the stated reasoning that a migration limit matched the actual risk better than a content limit did. The wider 0.5 limit applies to articles intended for direct and prolonged skin contact, and the legislation names examples rather than leaving the boundary to argument: earrings; necklaces, bracelets, chains, anklets, and finger rings; wrist-watch cases, watch straps, and tighteners; and rivet buttons, tighteners, rivets, zippers, and metal marks where these are used in garments. The Irish transposition of the 2004 amendment reproduces that list and both limits verbatim, which is an easier way to read the operative wording than working through a consolidated annex.
One week in artificial sweat, then two years of wear
The reference test method is EN 1811. The article goes into an artificial sweat solution for a week under controlled conditions, and the nickel that has migrated into the solution is then measured, usually by atomic absorption spectroscopy or mass spectrometry. The answer is expressed per square centimetre of the surface that touches skin, which is why a broad cuff and a thin wire can carry the same rating while handing over very different total amounts.
Nothing in that sequence involves rubbing, flexing, or a year of friction inside a bag. It is one immersion, one week, one sample of one design.
Which is why coated articles face a second test. Where a nickel-free coating is doing the protective work, that coating has to hold release under the limit for at least two years of normal use, and EN 12472 exists to simulate wear and corrosion before the release measurement runs. The Danish Environmental Protection Agency names both standards and the two-year requirement in its summary of the nickel rules.
Two years of normal use is a definition somebody had to write down, not a property of your week. A chain worn daily through a hot summer is not receiving the same treatment as one worn twice a month, and the moment a thin plate stops covering is the moment the piece starts behaving like whatever sits underneath. Nothing grandfathers an old plated earring into its original test result.
The threshold a laboratory compares against
The comparison is also less tidy than a single figure implies. An earlier version of EN 1811 applied a flat adjustment factor of 0.1 to measured results to absorb testing variability. The 2011 revision dropped that factor and worked with measurement uncertainty instead, after an inter-laboratory comparison in 2008 put the combined uncertainty at 46 percent. The side effect was a band of results around the limit that the standard could only report as no decision.
The amendment EN 1811:2011+A1:2015, published on 25 August 2015, closed that band by folding the uncertainty into the conformity call. Under it an article held to the 0.5 limit is judged non-compliant only above a measured 0.88, and a post assembly held to 0.2 only above 0.35. A testing house set out both decision thresholds when the amendment appeared. CEN issued a further revision, EN 1811:2023, in 2023. The legal limit and the number a laboratory has to beat are two different quantities, and only the first one is ever quoted at a shopper.
Nickel free is nobody's legal term
None of the vocabulary on a product page is anchored to any of this. The restriction defines a release rate and a test method; it does not define a phrase a seller may print.
| What you read | What it is anchored to |
|---|---|
| Complies with REACH entry 27 | Release under 0.5, or 0.2 for a piercing post, measured to EN 1811 |
| Nickel free | The seller's own wording, with no defined threshold behind it |
| Hypoallergenic | A marketing term carrying no release figure at all |
| Surgical steel | A family of steels, most of which contain a good deal of nickel |
| Titanium or niobium post | A named base metal, which is a stronger claim than an adjective |
Surgical stainless steel is the honest confusion in that list. The common grades contain nickel, often at ten percent or more, and still tend to pass the release test because a passivated surface holds onto it. High content and low release sit together without contradiction, which is what the 2004 amendment recognised.
Buying an earring and a strap against the rule
The useful move in a shop is to ask about base metal and finish rather than about nickel, because those are the answers a seller can actually give.
- For anything going through a piercing, buy on a named base metal such as titanium or niobium, since the stricter limit applies there and a plating layer is doing no work inside a healing channel.
- Treat a plated piece as protected for a defined stretch of wear rather than forever, and retire it when the colour breaks at the clasp or the back of a stud.
- On a watch, remember that the case back and the buckle are the parts in prolonged contact, so a steel case with a plated buckle is two different questions.
- Ask whether the seller can point to a test report referencing EN 1811, and treat a shrug as information rather than as a refusal.
One check before a piece joins daily rotation
Give any new metal piece a short trial before it becomes something you put on without thinking. Wear it for a full day, take it off, and look at the contact points under a lamp: the inside of a ring, the back of a stud, the underside of a buckle. Discolouration on the metal or on you is the surface telling you something the packaging did not.
Then adopt the habits that slow release down. Take metal off before a workout or a shower, since sweat and hot water are the conditions the test is imitating. Dry a strap and a case back rather than leaving them damp. Store pieces so clasps and edges are not scraping each other, because every scratch is a shortcut through the coating the two-year clock assumed would hold.