Why Aluminum Tin Lids Loosen in Transit - and How to Specify One That Does Not

The two rules of thumb, written out, together with the caveat that belongs beside them:

Target application torque (in-lb) ≈ Closure diameter (mm) ÷ 2
Expected removal torque after 24 h ≈ Application torque × 0.5
1 in-lb = 0.113 N·m

For our 41 mm bodies: 41 ÷ 2 = 20.5 in-lb applied, which is 2.32 N·m, settling toward roughly 10 in-lb to remove.
For the 84 mm body: 84 ÷ 2 = 42 in-lb, which is 4.75 N·m, settling toward roughly 21 in-lb.

Those numbers are a starting point and nothing more. Everything below is about why they are a starting point, and what to put in a specification instead.

Closure Torque Reference Chart

Before the detail, the chart. The four bodies we tool, with the rule-of-thumb starting figures in both units, and the column that matters most on the right: the number you have to establish yourself.

Body	Closure diameter	Rule-of-thumb application torque	In N·m	Your measured removal torque at 24 h
10 g tin	41 mm	20.5 in-lb	2.32 N·m	______
15 ml container	41 mm	20.5 in-lb	2.32 N·m	______
200 ml container	82 mm	41 in-lb	4.63 N·m	______
100 g tin	84 mm	42 in-lb	4.75 N·m	______

The right-hand column is empty on purpose. It is the only column in this chart that is evidence rather than arithmetic, and it is the one a buyer has to fill in before a closure specification means anything.

Why the Torque Number You Were Given Has No Standard Behind It

Ask three suppliers for a torque specification and you will get three numbers, all close to half the diameter in millimetres. It looks like consensus. It is not consensus — it is the same undocumented rule circulating.

We went looking for the source. The rule appears on packaging distributors’ technical pages, stated without attribution. The trade body usually credited with the original torque chart, the Closure & Container Manufacturers Association, dissolved on 1 April 2013 and merged into the International Society of Beverage Technologists. ISBT continues to publish finish guidelines — its Threadspecs documents — and states explicitly that they contain no torque values. ASTM publishes methods for measuring torque and publishes no target values at all.

So the honest position is: the rule of thumb is a reasonable place to start and has no authority behind it. That is not a reason to ignore it. It is a reason to convert it into a contractual number backed by your own measurements.

3-Point Torque Register

Point	What to record	Where it comes from	Your SKU
1. Application torque	The torque the capper is set to, in N·m, and the tolerance band	Agreed with the filler; starts from the rule of thumb, ends as a measured setting	______
2. Removal torque at 24 h	Mean and range across 20 pieces, measured 24 hours after capping	ASTM D2063/D2063M-24 (manual) or ASTM D7860-14(2022) (automated)	______
3. Removal torque after transit	Mean and range across 20 pieces measured after a transit simulation	Same methods, run on units that have been through ISTA 3A or ASTM D4169	______

Three numbers, not one. A closure specification containing a single figure has told you what the machine was set to and nothing about what the customer will experience. The 3-Point Torque Register is the minimum record, and it fits on one line of a quality plan.

If your current closure specification contains only an application torque, add points 2 and 3 before your next production run rather than after a complaint.

Why a Lid That Passed the Sample Check Backs Off in a Carton

A hand check on a sample tests the closure at rest. Transit is not rest. A carton on a pallet on a truck spends hours under low-amplitude random vibration, punctuated by shocks, and a threaded closure under vibration does exactly what a threaded fastener under vibration does: it walks.

The mechanism is worth understanding, because it tells you which test to ask for. A threaded closure is held by friction between the thread flanks and by the compressed liner pushing back against them. Vibration momentarily reduces that friction — each shock unloads the joint for a fraction of a second — and any residual unwinding torque stored in the compressed liner is then free to move the closure a few degrees. Repeat that a few hundred thousand times across a truck journey and the closure has turned. Nothing broke, no component was defective, and the sample you approved would still pass a bench check today.

That is also why the two useful methods here are a vibration method and a full journey simulation rather than a static seal test. A static test asks whether the package is sealed. These ask whether it stays sealed while being shaken, which is the question a customer’s parcel actually poses.

This is well-covered ground in packaging test standards, which is convenient, because it means you can specify a transit test by number rather than by argument.

4-Stage Transit Gate

Stage	Method to cite	What it reproduces	Pass criterion you set
1. Conditioning	Per the chosen ISTA or ASTM procedure	Temperature and humidity at destination; both change gasket compression	______
2. Vibration	ASTM D999-08(2023), repetitive shock and resonance dwell	The mechanism that actually backs a lid off	______
3. Full distribution simulation	ISTA 3A (18-22) for parcel shipments 70 kg or less, or ASTM D4169-23e1 with the distribution cycle that matches your route	The whole journey in sequence, not one hazard in isolation	______
4. Post-transit torque	ASTM D2063/D2063M-24 on 20 units taken from the tested cartons	What the customer’s hand will meet	______

Two notes on choosing between ISTA 3A and ASTM D4169. ISTA 3A is a general-simulation performance test for parcel-delivery shipments and runs fourteen test blocks, including a leak test block for liquids. ASTM D4169 is a practice rather than a single test: you select a numbered distribution cycle, DC-1 through DC-18, that matches your real route, and an assurance level of I, II or III, where level I is the most severe. For a small-parcel e-commerce SKU, ISTA 3A is the usual answer; for a palletised export shipment, D4169 with the export cycle is closer to the truth.

Pick the procedure that matches how the goods actually travel, name it and its version in the purchase order, and require the post-transit torque reading as a deliverable rather than an option.

Why Two Suppliers’ Lids Do Not Fit Each Other’s Tins

This is the section most buyers do not expect, and it is the single most useful thing in this article.

There is no international standard for the rolled thread on a small aluminium tin. Not an ISO one, not a DIN one, not an EN one. We checked the published catalogue of ISO/TC 52, the ISO technical committee for light gauge metal containers, in full. It contains ISO 90-1, -2 and -3 (definitions and dimensions for open-top cans, general use containers and aerosol cans), ISO 5099:2022 on easy-open and peel-off ends, ISO 10653 and ISO 10654 and the ISO/TR 11761 and 11762 technical reports on round open-top cans, ISO/TS 21985:2022 on LPG cartridges, and ISO 24021-1:2022 and -2:2024 on vocabulary and classification. Not one of them covers a screw thread or a threaded finish.

Standards for threaded container finishes do exist — but every one of them is for a different material. DIN 168-1 covers knuckle threads for glass containers. DIN 6063-1 covers buttress threads mainly for plastic containers. ISO 9056 covers pilferproof finishes on glass. The GPI 400/410/415 finish series that US buyers know is a set of voluntary industry drawings for glass and plastic, not an ISO standard. ISBT’s Threadspecs are voluntary and beverage-focused, cover PET and glass only, and state that cap dimensioning falls under the proprietary expertise of cap manufacturers.

The practical consequences follow directly, and they are worth putting in front of a procurement team:

• Thread geometry on an aluminium tin is supplier-proprietary. Two factories quoting “84 mm screw top” are quoting two different threads.
• Lids and bodies are not interchangeable between suppliers, so a second source is a second qualification, not a drop-in.
• There is no dimensional standard to specify against, so the approved sample is your specification — which is exactly the role the CISG gives an approved sample in a contract of sale.

2-Line Thread Compatibility Lock

Line	What to secure in writing	What happens without it	On file
1. Retained reference samples	Sealed, dated body-and-lid pairs from the production tool, held by both parties	No physical reference exists when a later batch fits differently, and the argument has no evidence in it	______
2. Change notification	A written undertaking that any change to the thread form, the lid liner or the tool is notified before shipment	A silent tooling change arrives as a line stoppage	______

If you are qualifying a second supplier, budget a full closure qualification rather than a sample approval — there is no standard that makes the two threads compatible, so nothing but testing will tell you.

Why One Sample Cannot Tell You a Torque Value

Torque is not a property of a design. It is a distribution across a population, produced by the interaction of thread form, liner compression, capper setting, and the head-to-head variation on the line. One unit gives you one point from that distribution and no idea of its width.

The width is the part that matters. A mean removal torque of 10 in-lb with a range of 8–12 is a controlled process. The same mean with a range of 3–19 contains units that leak and units a customer cannot open, and its mean looks identical on a certificate.

20-Piece Removal Torque Trace

Step	What to do	Why 20 and not 3	Result
1	Cap 20 filled units at the agreed setting, from at least two capping heads	Head-to-head variation is a common and invisible cause of spread	______
2	Hold 24 hours at the storage temperature	Liner relaxation and thermal cycling both reduce retained torque	______
3	Measure removal torque per ASTM D2063/D2063M-24, on a meter calibrated per ASTM D3474-23	An uncalibrated meter turns a specification into a number with no traceability	______
4	Record mean, minimum and maximum — not just the mean	The minimum predicts leaks; the maximum predicts consumer complaints	______
5	Repeat on 20 units after the 4-Stage Transit Gate	This is the number your customer actually meets	______

Twenty is the figure we use because it is small enough to run on a bench in an afternoon and large enough for a range to mean something. If your quality team wants a statistically-derived sample size, use theirs; what matters is that it is not one.

Run the trace once on the current production tool before you change anything — without a baseline distribution, every later measurement is an anecdote.

Why Your Capping Head Strips a Thread That a Hand Test Passes

A hand test applies torque slowly, coaxially, and with a human who stops when it feels tight. A capping head applies torque fast, at whatever alignment the puck presents, and stops when its clutch says so. These are different loads, and a thread that survives the first can strip under the second.

The complication that makes this hard to debug: on an automatic capper you cannot measure application torque directly. The torque is delivered inside a spinning chuck. What is measurable afterwards is removal torque, which is used as the proxy — and the relationship between application and removal torque is specific to each package. That relationship has to be established by correlation testing, per package, not assumed from a chart.

5-Row Capping Line Handoff Sheet

Row	What the filler needs from the tin supplier	What the tin supplier needs from the filler	Exchanged
1	Thread form reference samples, dated, from the production tool	Capper make, model and head count	______
2	Lid liner material and thickness	Chuck type and gripping method	______
3	Dimensional tolerance on the finish, body and lid	Line speed in units per minute	______
4	Recommended starting application torque	Measured removal torque from a trial run	______
5	Confirmation of any change to any of the above	Fill temperature, if the product is filled warm	______

Row 5 on the left and row 5 on the right are the two that get skipped and the two that cause the most expensive surprises. A warm fill changes the headspace pressure as the product cools, which changes retained torque; a silent liner change does the same thing for a different reason.

Send this sheet to your filler and your tin supplier at the same time, before the first trial — it costs an email and removes the most common cause of a stopped line.

Why “It Doesn’t Leak” Is Three Different Tests

“Leak test” is used for at least three different procedures with different sensitivities, different sample states and different purposes. Specifying the wrong one produces a certificate that does not answer your question.

3-Test Seal Verification Matrix

Method	What it actually tests	When it is the right choice	Chosen?
ASTM F2096-11(2019) — internal pressurisation bubble test	Gross leaks in a package by pressurising it and looking for bubbles; the standard cites gross leaks down to about 250 µm at 81% probability of detection	A filled or sealed package where you want a straightforward gross-leak check	______
ASTM D3078-02(2021)e1 — bubble emission under vacuum	Gross leaks in flexible packages containing headspace gas; the standard warns small leaks may not be detected	Pouches and sachets, not rigid tins	______
ASTM D4991-25 — vacuum leakage of empty rigid containers	Establishes the point at which leakage commences in an empty rigid container, framed around UN and ICAO dangerous-goods air transport	Regulatory qualification for hazardous liquids by air — not a routine filled-tin seal check	______

The distinction that catches people: D4991 tests empty containers for a dangerous-goods purpose. A supplier who certifies “leak tested to ASTM D4991” on a filled balm tin has cited a real standard for the wrong situation. If what you want to know is whether a filled tin holds, F2096 is the closer method, and the ISTA 3A leak test block is the closer answer if the question is about transit.

Write the method designation and the acceptance criterion into the specification — “leak tested” on its own is not a specification and cannot be audited.

Why a Child-Resistant Claim Is a Panel Test, Not a Lid Design

Buyers in balm, topical and botanical categories increasingly ask whether a screw-top tin can be child-resistant. The answer is that child resistance is not a feature of a closure — it is the result of a human panel test that a specific package either passes or does not.

Under 16 CFR 1700.20, the US procedure, the child panel is 50 children aged 42 to 51 months, tested in pairs across two five-minute periods, with a sequential protocol that can extend to 200 children. The senior-adult panel is 100 adults aged 50 to 70, 70% female, who must open and resecure the package. There is a further younger-adult panel of 100. ISO 8317:2015 is the international equivalent for reclosable packages and states explicitly that it is for type approval, not for quality assurance. ISO 13127:2012 provides mechanical test methods as a surrogate for the human panels.

4-Question Child-Resistance Gate

Question	Why it decides the answer	Your position
1. Is CR required by law for this product in this market, or is it a customer preference?	A legal requirement fixes the protocol; a preference lets you choose a design that is merely difficult	______
2. Which protocol — 16 CFR 1700.20, ISO 8317:2015, or both?	They are different tests with different panels; passing one is not passing the other	______
3. Who commissions and pays for the panel test, and who owns the certificate?	Panel testing is a significant cost and the certificate is package-specific, not lid-specific	______
4. Does the senior-adult accessibility requirement conflict with your torque target?	A closure tight enough to resist a child must still be openable by a 70-year-old — these two pull in opposite directions	______

Question 4 is the one that surprises people. Child resistance and accessibility are tested on the same package, and the torque window that satisfies both is narrower than the window that satisfies either alone.

If a customer has asked for child resistance, establish which protocol applies before quoting anything — the test regime, not the lid, is what you are buying.

A Worked Example, From Product to Approved Closure

A 100 g solid balm, filled warm at 55 °C into our 84 mm body, shipped as single parcels by courier to consumers in the US and the EU.

6-Step Closure Qualification Trace

Step	Input	Working	Output
1	Closure diameter 84 mm	84 ÷ 2	Starting application torque 42 in-lb = 4.75 N·m
2	Warm fill at 55 °C	Headspace contracts on cooling; retained torque typically rises then relaxes	Measure at 24 h after full cool-down, not at the end of the line
3	20-Piece Removal Torque Trace, two heads	Record mean, min, max per ASTM D2063/D2063M-24	Baseline distribution, not a single figure
4	Shipping mode: single parcel	Choose ISTA 3A (18-22) over ASTM D4169 export cycle	Transit procedure named in the PO
5	4-Stage Transit Gate	Condition, vibrate, full simulation, then re-measure 20 units	Post-transit removal torque distribution
6	3-Point Torque Register complete	Application, 24 h removal, post-transit removal all recorded with ranges	Closure approved against evidence, with a retained reference sample pair

Step 2 is the one that is specific to this product and gets missed on almost every warm-filled SKU. Measuring at the end of the line measures a package that has not finished changing.

Tell us the fill temperature when you send the enquiry, not after the first trial — it changes which torque reading is the meaningful one.

2026 Closure Specifications: The Method Most of Them Cite No Longer Exists

If your closure specification, or your supplier’s certificate, cites ASTM D3198 for application and removal torque, it is citing a withdrawn document. D3198-97(2007), Standard Test Method for Application and Removal Torque of Threaded or Lug-Style Closures, was withdrawn in 2016, and ASTM lists no replacement. Its scope described exactly this problem — rating capping machine performance and assessing loosening in storage and shipment — which is why it is still quoted so widely.

The live methods, and the correct citations for a 2026 specification, are:

• ASTM D2063/D2063M-24 — torque retention for packages with continuous thread closures, using non-automated (manual) torque testing equipment.
• ASTM D7860-14(2022) — the same measurement using automated torque testing equipment, covering child-resistant and non-child-resistant packages.
• ASTM D3474-23 — calibration and use of the torque meters themselves.

Note also the vocabulary shift, because it affects what you search for and what a lab quotes you: ASTM’s term is torque retention, not torque decay and not stress relaxation. A request for “torque decay testing” will get you a phone call; a request for D2063 torque retention will get you a quotation.

Search your existing specifications for “D3198” this quarter and replace it — a certificate against a withdrawn method is difficult to defend in a claim.

References

1. ASTM D3198-97(2007), Standard Test Method for Application and Removal Torque of Threaded or Lug-Style Closures. Withdrawn 2016, no replacement. store.astm.org/d3198-97r07.html
2. ASTM D2063/D2063M-24, Standard Test Methods for Measurement of Torque Retention for Packages with Continuous Thread Closures Using Non-Automated (Manual) Torque Testing Equipment. store.astm.org/d2063_d2063m-24.html
3. ASTM D7860-14(2022), Standard Test Methods for Measurement of Torque Retention for Child Resistant and Non-Child Resistant Packages with Continuous Thread Closures Using Automated Torque Testing Equipment. store.astm.org/d7860-14r22.html
4. ASTM D3474-23, Standard Practice for Calibration and Use of Torque Meters Used in Packaging Applications. store.astm.org/standards/d3474
5. ASTM D999-08(2023), Standard Test Methods for Vibration Testing of Shipping Containers. store.astm.org/standards/d999
6. ASTM D4169-23e1, Standard Practice for Performance Testing of Shipping Containers and Systems. store.astm.org/standards/d4169
7. ISTA 3A (18-22), Packaged-Products for Parcel Delivery System Shipment 70 kg (150 lb) or Less, International Safe Transit Association. ista.org/docs/3Aoverview.pdf
8. ASTM F2096-11(2019), Standard Test Method for Detecting Gross Leaks in Packaging by Internal Pressurization (Bubble Test). store.astm.org/standards/f2096
9. ASTM D3078-02(2021)e1, Standard Test Method for Determination of Leaks in Flexible Packaging by Bubble Emission. store.astm.org/standards/d3078
10. ASTM D4991-25, Standard Test Method for Leakage Testing of Empty Rigid Containers by Vacuum Method. store.astm.org/d4991-25.html
11. ISO 8317:2015, Child-resistant packaging — Requirements and testing procedures for reclosable packages, edition 3, confirmed 2021. iso.org/standard/61650.html
12. ISO 13127:2012, Packaging — Child resistant packaging — Mechanical test methods for reclosable child resistant packaging systems. iso.org/standard/53019.html
13. 16 CFR 1700.20, Testing procedure for special packaging, US Consumer Product Safety Commission. ecfr.gov
14. ISO/TC 52, Light gauge metal containers — published catalogue, checked in full for a threaded-finish standard. iso.org/committee/48944
15. ISBT, Threadspecs — voluntary beverage finish guidelines; states that torque values are not included. isbt.com/resources/isbt-threadspecs

How the Numbers in This Article Were Calculated

The application torque figures in the reference chart are the trade rule of thumb — closure diameter in millimetres divided by two, read as inch-pounds — applied to the outside diameters published on our own product pages, and converted at 1 in-lb = 0.113 N·m. They are starting points for a trial, not recommended settings and not values taken from any standard. We could find no standards body or trade association that publishes this rule or any target torque value; the association most often credited with the original chart dissolved in 2013.

The “removal torque settles near 50% of application torque after 24 hours” figure is likewise trade practice, quoted by packaging distributors without attribution to any standard. The relationship between application and removal torque is specific to each package and must be established by correlation testing on your own filled units.

The statement that no international standard exists for the rolled thread on a small aluminium tin is based on a review of the full published catalogue of ISO/TC 52 (light gauge metal containers) in September 2026, together with a check of the threaded-finish standards that do exist for glass and plastic. Standard designations, titles and status — including the withdrawal of ASTM D3198 — were checked against the issuing bodies’ own catalogue entries at the same time.

Frequently Asked Questions

Should I just tighten the lid more?
No, and this is the instinct that causes the most damage. Over-torquing deforms the liner, can strip a rolled thread in aluminium, and produces packages an end customer cannot open — while the units that leak are usually at the low end of a wide distribution, not the whole batch. Narrow the range with the 20-Piece Removal Torque Trace instead of raising the mean.

What torque should I use on an 84 mm aluminum tin?
Start at about 42 in-lb (4.75 N·m) from the diameter rule of thumb, then measure removal torque on 20 filled units at 24 hours and again after transit simulation, and let those numbers set the contractual figure. No standards body publishes a target value for this.

Can I use another supplier’s lids on your tins?
Almost certainly not. There is no ISO or DIN standard for the thread on a small aluminium tin, so every manufacturer’s thread form is its own. Treat a second source as a full closure qualification, not a drop-in substitution.

Why does the lid feel tighter after shipping?
Two effects run in opposite directions. Liner relaxation and vibration reduce retained torque; a warm fill that cools, or a drop in ambient temperature or pressure, can increase it. Which one dominates is specific to your product and fill temperature, which is why the measurement is taken after full cool-down rather than at the end of the line.

Do I need a gasket or liner in the lid?
For anything that flows, or anything travelling by air, a liner is doing the sealing and the thread is only holding it in compression. For a dry solid it may be unnecessary. Specify the liner material and thickness in writing either way — a silent liner change is one of the most common causes of a torque shift between batches.

Is ASTM D3198 still the right test to specify?
No. It was withdrawn in 2016 with no replacement. Cite ASTM D2063/D2063M-24 for manual measurement or ASTM D7860-14(2022) for automated equipment, and use ASTM’s term — torque retention.

How many units should I test?
Twenty, from at least two capping heads, recording mean, minimum and maximum. One unit tells you a value exists; twenty tell you the spread, and the spread is what predicts both leaks and complaints.

Getting a Closure Qualified

Send your fill temperature, your capper make and head count, your shipping mode and your destination markets. We will give you a starting application torque for the body you are considering, supply reference sample pairs from the production tool, and tell you which transit procedure matches your route. If your product needs a certified child-resistant closure, we will tell you what that changes before you commit to a size.

Send your filling details for a closure 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 guide to choosing an aluminum tin supplier.