Why Your Pantone Color Drifts Between Aluminum Tin Runs

The colour specification, written out as the five fields it actually has:

Colour spec = Target L*a*b* + ΔE00 tolerance + Measurement condition + Substrate build + Viewing condition

A complete line looks like this, and it fits on a drawing:
Lid face: L* 52.1 a* 38.4 b* −12.0, ΔE00 ≤ 2.0, measured M1, on white base coat + gloss overvarnish, assessed under ISO 3664:2025 P1

Compare that with what most artwork packs contain: PANTONE 234 C. Five fields against one, and the one that is present is defined on a substrate you are not printing on. Everything below is about closing that gap.

Colour Control Reference Chart

Before the detail, the chart. Four decoration routes, what each can actually hold, and what it costs you to hold it. The right-hand column is the one to fill in for your own programme.

Decoration route	What controls the colour	Realistic run-to-run consistency	Your requirement
Screen print on formed body or lid	Ink mix, mesh, squeegee pressure, cure temperature	Good within a run; the base coat under it matters as much as the ink	______
Offset / dry offset on flat sheet before forming	Ink film thickness, base coat, overvarnish; the only route with an ISO process-control standard	Best measurable control, but the sheet is then drawn and the image distorts with it	______
Pad print	Pad durometer, ink viscosity, cliché depth	Suited to small marks and curved faces; not the route for a large flood colour	______
Anodised colour (dyed anodic film)	Anodic coating thickness, dye bath concentration and time, sealing	Depth of colour follows coating thickness, so thickness control is colour control	______

One number worth having in mind while you read the rest. The lid face on our 84 mm body is a circle of radius 42 mm, so its printable area is πr² = 55.4 cm². That is a large, flat, uninterrupted field of a single colour at eye level. Large flat fields are exactly where a small colour difference becomes visible, which is why tins are a harder colour problem than most packaging of the same value.

Why a PMS Number Cannot Be Met on Metal

Pantone’s physical guide systems cover coated and uncoated paper, plastics, textiles and metallic inks. The Metallics Guide is metallic ink printed on paper — not ink on metal. There is no official Pantone guide for a metal substrate, and this is not an oversight: a PMS number is defined by a specific ink formulation on a specific paper stock under a defined illuminant, and none of those three conditions survives the move to aluminum.

Pantone is admirably direct about the underlying variability. Its own support material states that slight variations occur between printings and during each printing, that differences between paper lots and mills cause differences in absorption and ink hold-out, that the formulas in its guides are starting points requiring adjustment for pigments, ink bases and stocks, and that guides should be replaced every 12 to 18 months because normal usage and light exposure render colours inaccurate.

That last one deserves a moment. If your brand team has been matching against the same fan deck for four years, part of your “colour drift” is the reference, not the print.

What exists for metal is a different mechanism: PantoneLIVE dependent standards, which Pantone describes as digitally characterised representative examples of PANTONE colours reproduced using a given printing process and typical substrate. Its libraries include aluminum families for can and sheet — bare, with gloss varnish, with matt varnish, with opaque ink, with transparent ink — and equivalents for steel and metallised film. That is the correct reference to name for a tin, and it is one of the two things you can actually contract to.

4-Line Color Contract

Line	What to write	What it removes from the argument	Your value
1. The target	Either a PantoneLIVE dependent standard naming the exact substrate and varnish build, or measured L*a*b* from a signed physical drawdown on production substrate	“It does not look like the Pantone book” — the book was never the target	______
2. The tolerance	A ΔE00 figure, agreed, with the parametric factors stated	“Close enough” as a judgement rather than a measurement	______
3. The measurement condition	M0, M1 or M2 per ISO 13655:2017, plus the instrument aperture	Two labs producing two numbers for one lid	______
4. The viewing condition	ISO 3664:2025, naming the condition (P1 for critical comparison)	Approval meetings held under whatever light the meeting room has	______

Four lines. They cost nothing to write and they are the difference between a colour dispute that has an answer and one that does not. Every section below is a reason one of these four lines exists.

Add the 4-Line Color Contract to your artwork pack before your next tender, not after your next rejection — a supplier cannot hit a target that was never stated.

Why Two Labs Measured the Same Lid and Disagreed

This one produces more wasted email than any other item in this article, and it is entirely avoidable.

ISO 13655:2017 defines measurement conditions M0, M1, M2 and M3, which differ in the spectral content of the illumination the instrument uses — principally in how much ultraviolet is present. Substrates and coatings containing optical brightening agents fluoresce under UV, so their measured colour changes with the M-condition. Two competent laboratories, two calibrated instruments, one lid, two different numbers, both correct.

Then there is the light the humans use. ISO 3664:2025, the fourth edition published in July 2025, superseded ISO 3664:2009 and specifies viewing conditions based on CIE standard illuminant D50. The 2025 edition adds two new conditions, P3 and P4, for D50 excluding the ultraviolet component — which exists precisely because of the brightener problem above.

2-Condition Measurement Gate

Gate	What to fix in writing	What goes wrong without it	Agreed value
1. Instrument condition	M-condition per ISO 13655:2017, plus aperture size and whether the reading is with or without the overvarnish	Supplier and buyer measure the same part and disagree by more than the tolerance they are arguing about	______
2. Human viewing condition	ISO 3664:2025 condition, named — and the same condition at both ends	A lid approved under warehouse fluorescents and rejected under a showroom spot	______

If you take one operational change from this article, make it this: put the M-condition in the specification. It is four characters and it eliminates a whole category of dispute.

Before the next colour submission, agree the M-condition and the viewing condition with your supplier in writing — it is the cheapest quality improvement available in this process.

Why “ΔE 1.0 Is Just Noticeable” Is Not in Any Standard

Almost every colour brief eventually contains a sentence like “a ΔE of 1.0 is the threshold of perception for a trained observer”. It is repeated with great confidence and it does not appear in the standards.

We went looking. It is not in ISO/CIE 11664-6:2022, the CIEDE2000 colour-difference formula. It is not in ISO/CIE 11664-4:2019, which defines the CIE 1976 L*a*b* colour space. It is not in CIE 015:2018, Colorimetry, 4th edition. It is not in CIE 142:2001, the technical report that introduced CIEDE2000 — a document which does not use the phrase “just noticeable difference” at all, and which instead defines reference conditions: material and viewing-environment characteristics to which the colour-difference model applies.

The honest position is that perceptibility depends on where in colour space you are, how large the samples are, whether they are edge to edge or separated, what surrounds them, and who is looking. CIEDE2000 was calibrated against perceptual datasets; that is a relationship, not an equivalence.

Which leaves a practical question: what tolerance should you set? The answer is that it is a commercial decision, not a scientific one, and it should be set by looking at real drawdowns on real substrate and deciding what you are prepared to ship.

3-Number Tolerance Set

Number	How to arrive at it	Why it is not a lookup	Your value
1. Target	Measured from the signed drawdown on production substrate	The target has to exist on the substrate you are buying	______
2. Tolerance	Set by viewing a set of deliberately off-target drawdowns and choosing the last one you would accept	No standards body publishes a perceptibility threshold you can adopt	______
3. Formula and parametric factors	ΔE00 with the parametric factors stated, per ISO/CIE 11664-6:2022	ΔE*ab and ΔE00 give different numbers for the same pair; a tolerance without a formula is ambiguous	______

One published figure worth borrowing for scale rather than for compliance: Pantone states that its goal when manufacturing its own guide pages is to achieve less than 2 ΔE2000 for all colours on the page. That is a controlled offset press printing on paper — the most favourable case there is. It is a useful reality check on any brief demanding a tighter number on drawn metal.

Set your tolerance by looking at physical drawdowns rather than by copying a figure — and record which drawdown you rejected, because that is the evidence behind the number.

Why the Standard Your Printer Cites Is for Paper

Colour specifications for tins routinely cite ISO 12647-2. It is a real and important standard — and its scope runs to sheet-fed and web offset printing on paper and cardboard packaging. Metal is not in it.

There is a metal decorating standard, and most buyers have never heard of it.

3-Standard Scope Check

Standard	Actual scope	Applies to your tin?	Checked
ISO 12647-2:2013 — offset lithographic processes	Colour separations, proofs and production prints for four-colour sheet-fed and web offset; substrates are paper and cardboard packaging	No	______
ISO 12647-9:2021 — metal decoration printing by offset lithography	Requirements for systems producing offset prints for process colour reproduction on metallic substrates printed with a white coating. ISO states it is intended for flat sheet metal and does not cover shaped or pre-formed metals such as cans	Partly — the flat sheet before forming, not the drawn body	______
ISO 12647-5:2015 — screen printing	Display, signage and graphics using flat bed or cylinder equipment	No — container screen printing is outside it	______

Put plainly: screen printing on a formed aluminum tin has no ISO colour standard, and offset metal decorating has one that stops at the flat sheet. That is not a reason for despair; it is the reason the 4-Line Color Contract exists. Where there is no standard to invoke, the contract has to carry the specification.

What does exist for the flat-sheet route is a characterisation dataset: FOGRA60, the printing condition “Metal Printing MPC1”, based on ISO 12647-9:2021 and available since 2022. If your decoration is offset on sheet, naming FOGRA60 is a meaningful thing to write in a specification.

Ask your decorator which of these three standards they are actually working to, and what they do where none applies — the answer tells you whether their colour control is a system or a habit.

Why the Same Ink Looks Different on Your Four Sizes

A brand running several tin sizes will eventually notice that the same specified colour does not look identical across them, even from one production run. This is not a defect. It is geometry and process interacting.

5-Row Decoration Method Matrix

Cause	Effect on colour	What to do about it	Applies to your SKUs?
Curvature of the printed face	A curved body wall reflects differently from a flat lid; the same ink reads darker or lighter by position	Approve the lid and the body separately, with separate targets	______
Draw distortion	Artwork printed on flat sheet then drawn is stretched non-uniformly; ink film thins where the metal thins	Distortion-compensated artwork, and colour measured after forming	______
Base coat	Colour sits on whatever is under it; a white base and a bare metal ground give different results from identical ink	Fix the base coat in the specification, not just the ink	______
Ink film thickness	The single largest lever on a spot colour; more ink is not linearly more colour	Control by density measurement in-run, not by eye	______
Overvarnish	Gloss and matt varnishes shift measured L* and the perceived saturation	Measure and approve with the final varnish present, never before it	______

Row 3 and row 5 are the two that most often explain a “the ink is wrong” complaint where the ink is in fact identical. A colour approved before the overvarnish is a colour approved on a different product.

If you run more than one size, decide now whether you want them matched to each other or each matched to the brand target — those are different instructions and they produce different tooling and approval plans.

Why the Colour Changed and the Ink Did Not

Anodised colour deserves its own note, because its failure mode is completely different and is the one buyers least expect. In dyed anodising there is no ink. The colour is dye absorbed into the pores of the anodic oxide film, so the depth of colour follows the thickness of that film. A batch anodised slightly thinner takes up less dye and comes out lighter, with no change to the dye bath at all.

That gives you a measurable control that has nothing to do with colour instruments: anodic coating thickness, measured by the amplitude-sensitive eddy-current method under ISO 2360:2017 or ASTM B244-09(2021).

3-Layer Build Lock

Layer	What to specify	How it is verified	Your spec
Substrate treatment	Mill finish, brushed, or anodised — and for anodised, the coating thickness with a tolerance	ISO 2360:2017 or ASTM B244-09(2021) eddy current	______
Colour layer	Ink system and film thickness, or dye and sealing method	In-run density for ink; thickness plus lightfastness rating for anodised	______
Overvarnish	Gloss level and coverage, present on every approval sample	Gloss meter reading recorded with the colour reading	______

For a coloured anodised finish, ISO 7599:2018 is the document that specifies decorative and protective anodic coatings, and colour durability is tested under ISO 2135:2024 for accelerated light fastness of coloured anodic coatings, or ISO 6581:2018 for comparative fastness to ultraviolet light and heat. A specification for a coloured anodised tin that names a colour but no minimum light-fastness rating has left out the property most likely to disappoint.

If you are specifying coloured anodising, add a coating-thickness tolerance and a minimum light-fastness rating — without them, “anodised blue” is a description rather than a specification.

Why Your Print Survived the Sample and Not the Shelf

Colour accuracy is one problem. Colour durability is another, and it is tested by different methods, which is worth knowing before a specification asks for the wrong one.

3-Test Print Durability Register

Property	Correct method to cite	The trap	Result
Adhesion	ASTM D3359-23 Method B, or ISO 2409:2020 cross-cut	Method A is for films over 125 µm; a decorative print is far thinner, so Method B is the right one. ISO 2409 is not suitable above 250 µm or for textured coatings	______
Flexibility and impact	ASTM D522/D522M-17(2021) mandrel bend; ASTM D2794-93(2024) rapid deformation	D2794 itself warns that reproducibility between laboratories is poor — use it within one lab, for comparison, not as an absolute acceptance number	______
Light fastness	ISO 16474-2:2013 xenon-arc for paints and varnishes; ASTM D4587-23 for fluorescent UV exposure of paints and related coatings	ISO 4892-2 is the same test written for plastics, and ASTM G154 is an apparatus practice that does not on its own deliver a result. Neither is the right citation for a printed lacquer	______

Add chemical resistance if the tin will meet cosmetics, sanitiser or cleaning products in use: ASTM D1308-20(2025) for the effect of household chemicals on coating systems, and ASTM D5402-19(2024) for solvent rubs — remembering that D5402 does not specify the solvent or the number of rubs, so a specification citing it must state both or it says nothing. Note also that ASTM D1647 for varnish water and alkali resistance was withdrawn in 2004 with no replacement, so a specification quoting it is quoting a document that no longer exists.

Decide which of these three properties actually matters for how your product is used, and test that one properly rather than all three superficially.

A Worked Example, From Brand Colour to Approved Print

A brand with one spot colour, PANTONE 2935 C in its guidelines, ordering our 100 g body screen printed on the lid face, 30,000 pcs, sold in retail under showroom lighting.

6-Step Color Approval Trace

Step	Input	Working	Output
1	PANTONE 2935 C in brand guidelines	Treated as a starting point, not a target — Pantone’s own words	Ink mixed toward it on production substrate
2	Substrate build fixed	White base coat + spot ink + gloss overvarnish, all named	Drawdowns produced on the real build, with varnish
3	Three drawdowns: on-target, slightly light, slightly heavy	Brand team views all three under ISO 3664:2025 P1 and picks the last acceptable one	Target and tolerance both established by looking
4	Chosen drawdown measured	M1 condition per ISO 13655:2017, aperture recorded, varnish present	Target L*a*b* recorded; ΔE00 tolerance agreed at the measured distance to the rejected drawdown
5	4-Line Color Contract completed	Target, tolerance, M-condition, viewing condition all written on the drawing	A specification a second supplier could also work to
6	Production	In-run density control; first-off and last-off measured against the target	Colour approved against a number, with the signed drawdown retained by both parties

Step 3 is the step people skip, and it is the one that makes the rest possible. Producing a deliberately-wrong drawdown feels like a waste of a proof. It is how you find out what your tolerance actually is, rather than adopting a number from an article.

Ask for three drawdowns rather than one on your next colour development — the two you reject are what turn your tolerance into evidence.

2026 Print Sourcing: Metal Decorating Finally Has Its Own Certification

Two developments worth knowing this year, both of which change what you can ask a supplier for.

Metal decorating can now be certified against its own standard. Until recently, metal printers had no individual route to demonstrate compliance with ISO 12647-9 and were audited against paper-based process-standard criteria. Fogra issued its first PSO certification for metal decoration print in September 2024. The matching characterisation dataset, FOGRA60 for printing condition MPC1, has been available since 2022. If your decoration is offset on flat sheet, “certified to ISO 12647-9 / FOGRA60” is now a question with a real answer, and it was not two years ago.

The viewing standard was replaced. ISO 3664:2025, published 11 July 2025, superseded ISO 3664:2009 and added viewing conditions P3 and P4 for D50 excluding the ultraviolet component. If your colour approval procedure names ISO 3664:2009, it names a withdrawn edition — and the new conditions exist precisely to handle the optical-brightener problem that causes so many measurement disputes.

Update the standard editions in your artwork pack this quarter, and ask any offset metal decorator whether they hold a current process-standard certification — the question is now answerable.

References

1. ISO 12647-2:2013, Graphic technology — Process control … Part 2: Offset lithographic processes. Scope covers paper and cardboard packaging. iso.org/standard/57833.html
2. ISO 12647-9:2021, Graphic technology — Process control … Part 9: Metal decoration printing processes using offset lithography. iso.org/standard/74360.html
3. ISO 12647-5:2015, Graphic technology — Process control … Part 5: Screen printing. Scoped to display, signage and graphics. iso.org/standard/60479.html
4. ISO 13655:2017, Graphic technology — Spectral measurement and colorimetric computation for graphic arts images — source of measurement conditions M0 to M3. iso.org/standard/65430.html
5. ISO 3664:2025, Graphic technology and photography — Viewing conditions, 4th edition, 11 July 2025; supersedes ISO 3664:2009. iso.org/standard/83759.html
6. ISO/CIE 11664-6:2022, Colorimetry — Part 6: CIEDE2000 colour-difference formula, 2nd edition. iso.org/standard/82662.html
7. ISO/CIE 11664-4:2019, Colorimetry — Part 4: CIE 1976 L*a*b* colour space. iso.org/standard/74166.html
8. CIE 015:2018, Colorimetry, 4th edition. cie.co.at
9. CIE 142:2001, Improvement to industrial colour-difference evaluation — the technical report introducing CIEDE2000. cie.co.at
10. Fogra, Characterisation data — FOGRA60, printing condition Metal Printing MPC1, based on ISO 12647-9:2021. fogra.org
11. Fogra press release, First Fogra certification for metal decoration print, September 2024. fogra.org
12. Pantone, PantoneLIVE Library Specifications v1.4 — definition of dependent standards and the aluminum and steel substrate families. pantone (PDF)
13. Pantone Support, Normal printing process colour variances explained. support.pantone.com
14. ASTM D3359-23, Standard Test Methods for Rating Adhesion by Tape Test. store.astm.org/d3359-23.html
15. ISO 2409:2020, Paints and varnishes — Cross-cut test. iso.org/standard/76041.html
16. ASTM D522/D522M-17(2021), Standard Test Methods for Mandrel Bend Test of Attached Organic Coatings. store.astm.org/standards/d522
17. ASTM D2794-93(2024), Standard Test Method for Resistance of Organic Coatings to the Effects of Rapid Deformation (Impact). store.astm.org/standards/d2794
18. ISO 16474-2:2013, Paints and varnishes — Methods of exposure to laboratory light sources — Part 2: Xenon-arc lamps. iso.org/standard/56830.html
19. ASTM D4587-23, Standard Practice for Fluorescent UV-Condensation Exposures of Paint and Related Coatings. store.astm.org/standards/d4587
20. ISO 7599:2018, Anodizing of aluminium and its alloys — Method for specifying decorative and protective anodic oxidation coatings on aluminium. iso.org/standard/70156.html
21. ISO 2135:2024, Anodizing of aluminium and its alloys — Accelerated test of light fastness of coloured anodic oxidation coatings using artificial light. iso.org/standard/88294.html
22. ISO 2360:2017, Non-conductive coatings on non-magnetic electrically conductive base metals — Measurement of coating thickness — Amplitude-sensitive eddy-current method. iso.org/standard/69943.html

How the Numbers in This Article Were Calculated

The printable lid area is calculated as πr² from the outside diameter published on our own product page, and is the full circle before any print margin is deducted; the actual printable field is smaller by whatever margin the process requires.

The example colour specification line is illustrative. Its L*a*b* values are not a measurement of any product and must not be used as a target; they are there to show the shape of a complete specification, which is the point of that section.

No numeric colour tolerance is quoted from ISO 12647-2 or ISO 12647-9 in this article. Both standards are paid documents and their tolerance tables are not publicly readable, so rather than repeat a figure circulating second-hand we have deliberately left the number out and described how to set your own. The only ΔE figure quoted from a named source is Pantone’s own manufacturing target for its guide pages, which describes ink on paper and is not transferable to metal.

The statement that no standards body defines a ΔE of 1.0 as a just-noticeable difference is based on checking the scopes and published abstracts of ISO/CIE 11664-6:2022, ISO/CIE 11664-4:2019, CIE 015:2018 and CIE 142:2001 in September 2026. Standard designations, editions and status were checked against ISO, CIE, ASTM and Fogra’s own catalogue entries at the same time.

Frequently Asked Questions

Can you match a Pantone color exactly on an aluminum tin?
No, and no supplier who says yes is describing what will happen. A PMS number is defined for ink on paper under a defined illuminant. On metal the achievable deliverable is a measured colour against a target with a stated tolerance, established from a drawdown on your actual substrate build.

What is an acceptable ΔE for printed packaging?
There is no published threshold to adopt. Set it by viewing physical drawdowns and choosing the last one you would accept, then record that distance as your tolerance with the formula and parametric factors stated. For scale, Pantone’s own target on its paper guides is under 2 ΔE2000.

Why does my tin look different from my Pantone book?
Three reasons, usually together: the book is ink on paper and the tin is ink on metal; the tin has a base coat and an overvarnish that the book does not; and the book itself fades. Pantone recommends replacing guides every 12 to 18 months.

Does ISO 12647-2 apply to printing on metal?
No. Its scope covers offset lithography on paper and cardboard packaging. The metal decorating standard is ISO 12647-9:2021, which is written for flat sheet metal and does not cover shaped or pre-formed cans.

Why do two color readings of the same lid disagree?
Almost always the ISO 13655 measurement condition — M0, M1 and M2 illuminate the sample with different amounts of ultraviolet, and any optical brightener in the coating fluoresces differently under each. Both readings can be correct. Fix the M-condition in the specification.

Should color be measured before or after the overvarnish?
After, always. Gloss and matt varnishes change both the measured L* and the perceived saturation, so a colour approved before varnishing was approved on a different surface from the one you will ship.

Why is my anodised color lighter in one batch?
Usually anodic coating thickness rather than dye. Dye is absorbed into the pores of the oxide film, so a thinner film takes up less dye. Specify the coating thickness with a tolerance and verify it by eddy current under ISO 2360:2017 or ASTM B244-09(2021).

Getting a Colour Approved

Send your brand colour reference, the decoration method you have in mind and the substrate finish you want. We will produce drawdowns on the real build including the overvarnish, measure them under a stated condition, and give you back a complete four-line colour specification you can hold us — or anyone else — to. If your target is not achievable on the finish you have chosen, we will tell you before tooling rather than after the first run.

Send your artwork for a colour check — 5,000 pcs minimum per SKU, 15–20 days production after artwork approval, samples available before bulk. See also our screw-top body specifications and our aluminum tins for cosmetics.