Application torque is what the capping head puts into the cap; removal torque is what it takes to break that cap loose later on the filled, rested pack. The first is a machine setting you adjust, the second is the pack property you accept or reject against, and the two are never equal because the liner relaxes after capping. No figure transfers between packs: a torque window belongs to one bottle, closure, liner and product combination, and every number should be treated as a reference range to be confirmed against actual measurement.
Three torque terms and what each one describes
Most arguments between a buyer and a supplier over a torque figure start because each side had a different measurement in mind. Pin the vocabulary down before any number is exchanged.
| Term | Other names | What it is | Where it is read | What you use it for |
|---|---|---|---|---|
| Application torque | Capping torque, on-torque, sealing torque | Rotational load delivered by the capping head while seating the cap. A machine input. | At the head, with a suitable gauge | Setting and correcting the line. It shifts with head adjustment, clutch condition, belt or gripper wear and how slippery the cap is. |
| Removal torque | Opening torque, break-loose torque, off-torque | Load needed to start the closed cap moving in the opening direction. A pack output. | On a bench instrument, bottle clamped, after the pack has been filled and rested | Acceptance limits and batch-to-batch comparison. It reflects what the seal is doing and what the customer feels. |
| Unscrewing torque | Over-the-thread torque | Sustained load while the cap keeps turning after it has broken loose | From the plateau of a torque trace | Diagnosing a stiff or galled thread, where opening is hard even though the break-loose peak looks right |
On a single unit, application and removal torque sit closest together straight after capping. Over the following hours and days the liner creeps and the removal figure drifts downward. A lone number with no label, no conditioning period and no instrument attached is therefore unusable, whoever supplies it.
Unscrewing torque normally sits well under the break-loose peak. Request it alongside the peak when the thread is long or when a customer has already complained that the cap is hard to turn.
Instruments and test conditions
Choosing the instrument
A continuous-rotation torque tester clamps the bottle, holds the cap in a chuck or jaws, and turns it while logging torque against angle. Its strength is that it returns a curve: the peak is break-loose torque, the plateau behind it is unscrewing torque, and an irregular shape flags a mechanical fault such as a thread jump or a liner that has partly bonded. For removal torque on production samples this is the right tool.
A slip or click gauge, handheld or bench-mounted, uses a calibrated clutch and reports the load at which that clutch slips. It is quick and inexpensive, and it suits verification of the machine setting at the capping head. It gives one value, no curve, and the result depends on the operator, so removal data taken with it on finished packs will not reproduce from one person to the next.
Four conditions that decide whether a reading means anything
- Where the cap is gripped. A plain cap can simply be held by its full skirt. A child-resistant cap whose outer shell spins freely cannot: holding the shell measures the clutch inside the closure, not the seal, and the result is invalid. It must be held on the inner body or at its designated points, and the method written down. A dropper cap is held by the collar only, never by the rubber bulb. Two labs gripping the same pack differently will report different figures, and both will be right.
- Time since capping. The same pack read five minutes after capping, after twenty-four hours and after seven days yields three figures. Fix one interval, state it and stick to it. Twenty-four hours is widely used because it is practical; a liner that takes longer to settle needs a longer interval, otherwise batches cannot be compared.
- Temperature and contents. Warmth softens the liner and the cap skirt and lowers removal torque, while cold pushes some materials the other way. Product contact matters just as much. Oil soaking into a foam liner, or a solvent drawing plasticiser out of a gasket, alters the seal load over weeks. For an aggressive product the honest test is a filled pack held for its intended shelf life, not an empty bottle checked on capping day.
- Calibration. Either instrument needs calibration traceable to a torque standard, and the clamp and chuck need a repeatability check. An uncalibrated gauge on a loose clamp generates a spread wide enough to bury a real fault, and it usually surfaces as a disagreement with a supplier whose readings differ from yours.
How torque produces a seal, and where it stops helping
Nobody buys torque. The goal is a pack that stays sealed, and torque is specified only because it is the most convenient measurable input that predicts the seal.
Turning a threaded cap down makes the thread translate rotation into axial load, and that load squeezes the liner onto the bottle rim. The squeeze closes the microscopic path between rim and liner and keeps contact around the full rim circle. Torque is thus an indirect handle on compression, and anything in the conversion alters the seal at an unchanged torque: thread pitch, thread friction, the cap's coefficient of friction, how the liner compresses, and the geometry of the finish.
Too little torque fails in the obvious way. Compression is short, the contact path is incomplete and the pack leaks.
Too much fails less obviously and costs more. An over-compressed liner cold-flows off the rim and can open a gap on the far side of the seal. A plastic skirt can split along a stress line. A thin glass finish can chip, or pick up a hairline that turns into a crack later. Removal torque also climbs until an older user cannot open the pack. The usable window has two real edges.
Why a figure cannot be borrowed from another pack
A request such as "the torque target for a 28 mm cap" has no honest answer until more is known. The obstacles are mechanical, not commercial.
The finish
Neck finishes differ in thread pitch and thread form. For the same input torque a coarse thread generates a different axial load than a fine one, so two finishes sharing a nominal diameter may need clearly different settings to reach equal sealing compression. Thread engagement length and the friction between glass and closure material shift the conversion further. Reusing a number from an earlier project on another finish is a dependable way to make leakers, which is why the finish designation, its T, E and H dimensions and the matching closure have to be settled before torque work starts.
The liner
Liners move the window as much as threads do. Identical threads with different liners do not share one.
- Compressible foam accepts a broad span of compression and so tolerates a fairly wide window, which explains its popularity on general-purpose packs.
- Pulp or coated pulp is stiffer and wants a more tightly controlled load.
- An induction foil seal is not really sealed by compression. A heat-activated layer bonds to the glass, and cap torque only matters because the liner must be pressed to the rim while activation happens.
- PTFE-faced liners, plain cone liners and rubber gaskets for aggressive chemicals each behave in their own way again.
Changing the liner compound cancels an existing window even if the closure drawing has not been touched.
The closure type
A child-resistant cap contains a deliberate slipping or ratcheting element, and a reading taken through it describes the mechanism. A dropper cap has a pipette collar whose shoulder seats on the finish; overload it and the collar deforms, after which the pipette neither seats nor draws as it should. A metal lug lid is held by internal vacuum and lug engagement, so applying thread torque figures to it is a category error; the headspace and product conditions that make that closure work are explained in our guide to lug cap glass jars.
Glass and plastic bottles cannot share a figure directly either. Glass is rigid and passes the load into the finish. Plastic gives, so an equal torque can leave visibly different compression and a different removal reading afterwards, and the closure's thread form and friction against each material differ too.
Measurement and acceptance by cap type
| Cap type | Usual liner | How to measure | What goes wrong outside the window | How to word acceptance | What to confirm with the supplier |
|---|---|---|---|---|---|
| Continuous thread screw cap | Foam, coated pulp, plain cone or induction foil | Removal torque on a continuous-rotation tester, bottle clamped, full skirt gripped, after a fixed conditioning interval at a stated temperature | Too low: weeping or leaking at the rim. Too high: liner crushed or cold-flowed, skirt split, glass finish chipped, cap too stiff for older users. | Instrument, grip point, conditioning interval, temperature and sample plan, followed by a removal torque window given as a reference range confirmed by your own measurement | Finish designation and thread form; the liner grade actually delivered; the seal threshold measured from a low-to-high torque series with an integrity test at every level |
| Press-on or crown closure | Cork, crown liner or plastisol gasket | Thread torque is not applicable. The crimp or interference fit does the sealing, so the test is a release or pull-off force. | A weak crimp lets go under internal pressure. An over-tight one damages the glass lip or cannot be removed by hand. | A dimensional or force requirement on the closing operation. Leave torque out altogether so the specification contains no clause that cannot be measured. | Whether the filler applies the closure or it arrives pre-crimped, and which dimensional check the line is able to run on finished packs |
| Dropper cap with pipette collar | Typically a cone or small gasket inside the collar | Hold the collar, not the rubber bulb, and note the collar design. The collar shoulder seats against the finish, so record the load at which it seats and the load that would deform it. | Over-torque distorts the collar and the pipette stops seating or drawing properly. Under-torque leaves it loose and product seeps round the collar thread. | Removal torque taken at the collar, a seating check, and a functional draw test on the assembled pack as the true acceptance criterion | Collar material and wall thickness; pipette length against fill volume; whether the collar maker rates it for this liner and product |
| Child-resistant cap | Foam, coated pulp or a gasket that also survives the mechanism | The outer shell may spin freely, so measure on the inner body or at the designated grip points, say which, and test the finished pack with the safety mechanism in place. | A wrong grip returns a false reading on a good pack. Genuine failures are a loose seal, a cap that jams on push-and-turn, and opening effort high enough to fail the senior-adult panel. | The torque window together with the opening procedure, with the integrity check and the applicable panel testing run on one and the same assembled pack | Any features the closure needs on the bottle; the direction of grip required; the upper torque limit the closure maker sets before the mechanism is affected |
For a pipette closure a good number is not enough. An over-compressed collar can return a perfectly acceptable torque reading and still fail to deliver a dose, which is why the draw check described with our dropper caps belongs next to the torque reading.
Finding the window by test
Since torque only controls the seal indirectly, the window has to be found by experiment. Cap filled samples in groups across a deliberate spread of application torque, running from plainly too low to plainly too high, then put every group through a seal integrity test. Vacuum decay is the usual method for rigid packs, with dye penetration or bubble emission as alternatives; flexible and semi-rigid packs call for other methods.
What comes out is a threshold, the application torque above which all samples pass. Production does not run at that threshold. The window is placed above it with a margin for ordinary variation, because a filling line is not a lab and torque wanders from unit to unit and from hour to hour.
The same data settles leakage complaints. Take the retention samples for the batch and measure them after the standard conditioning interval. If they sit within the window and pass the integrity check, the seal was made to specification and the cause is somewhere in handling or distribution. If they sit under it, the capping head drifted and the batch record will show when. Without a measured window, each side is only offering an opinion.
What the filling line has to control
A window on paper turns into a production fact only if the line can hold it. Three things have to be in place before it is worth putting into a purchase order.
A head that holds its setting. On a simple friction head, delivered torque depends on the clutch setting, head speed relative to the bottle, and the belt or gripper that stops the bottle spinning. Pads wear, belts stretch and clutches drift as they warm, so the first hour of a shift and the eighth are not automatically alike. Verifying the head against a gauge at intervals, and logging it in the batch record, is the least that will do.
A consistent finish. Thread dimensions vary inside the moulding tolerance, from mould to mould and from run to run. A finish at the small end of tolerance engages differently from one at the large end, and an identical head setting produces a different removal torque. Treat the finish as an inspected characteristic on the bottle, not a nominal note on the drawing, and treat a new bottle mould as a reason to requalify the window.
A named liner grade. Liners come in grades with a defined thickness and density, and substitutions are frequent when a batch runs short. A different density compresses differently, so the threshold moves. Unless the grade is written into the specification, the development window may describe a liner that is never supplied again. With foil seals, the activation parameters of the sealing equipment sit inside this control loop as well, since an under-activated seal leaks however well the cap went on.
The line must also check its own output. A workable minimum is a torque check on a fixed sample size when a run starts, at intervals through it and after every stoppage, with readings logged against the batch. Base the sampling plan on a recognised scheme such as ISO 2859-1, choose an AQL that reflects what a leak would cost, and take the acceptance limits from the measured window, not from a figure remembered from some other product.
When the pack is a standard threaded Boston round bought with its cap and nothing beyond the standard closure is needed, much of this work can be skipped; the finish and cap pairings that are stocked together are shown under Boston round bottles with caps.
How removal torque changes with age and transport
Capping day gives the highest removal torque a pack will ever show. Later readings are lower, and that decline is physics, not a defect.
Stress relaxation in the liner
A compressible liner is viscoelastic. Kept under load it slowly deforms, spreads the compression, and pushes back with less axial force. The seal holds for as long as the remaining compression still closes the rim path, so a pack that passes integrity on day one may fail a month later if its window was set too near the threshold. How fast this happens depends on liner material, temperature and product, and an oil migrating into the liner speeds it up markedly. Shelf-life claims should therefore rest on filled packs tested at intervals, not on empty bottles tested once.
Backing off under vibration
The alternating motion of transport can walk a cap round in the opening direction a little at a time, and removal torque drops with it. The extent depends on closure mass, thread form, liner-to-rim friction, and how severe and how long the vibration is. Use a recognised distribution simulation, such as an ISTA procedure or an ASTM test method suited to the pack, and take the reading after the vibration. A comfortable margin before dispatch does not prevent a loose cap on arrival if it was never tested against a vibration cycle.
A three-point schedule
One test is not enough. Measure removal torque:
- at the standard conditioning interval;
- after a defined ageing period at the intended storage temperature;
- after distribution simulation on the filled pack.
Readings that stay within the window at all three points give the pack a defensible basis. A window backed only by the first reading describes a moment, not a product.
Reading torque results to find the cause
Provided the readings are taken the same way every time, their pattern points to the source of a problem.
| Pattern | Likely cause | What to do |
|---|---|---|
| Uniformly low across the sample | Head setting or liner, not the bottle. An integrity failure in that batch puts the cause close to the seal. | Check the head against a gauge and confirm the liner grade. |
| Wide scatter, no pattern | Measurement error first, process second: uncalibrated gauge, inconsistent grip point, unstable clamp, or uneven intervals. | Rule out the measurement before chasing the line. |
| One low unit in a tight sample | A finish defect on that bottle, a foreign body under the liner, or a missing or folded liner. | Inspect the bottle together with its reading. Keep the torque bench beside the samples and leave closures on until each value is recorded. |
| Inside the window at first, below it after ageing | Liner and product, not the capping head. | Review the liner grade or seal compatibility. The usual remedy is another liner compound or a higher initial setting. |
| Abnormally high, stiff opening, field complaint | The window came from someone else's pack: a different finish, liner or product, carried over without repeating the conversion. | Trace where the figure originated, then lower the head setting to a window measured on your own pack. |
Two responses deserve caution. Adding torque on the line to offset an ageing seal defeats itself: the pack moves toward the top edge of the window, where liners crush and caps become hard to open, and the relaxation remains. Likewise, lowering a head setting to match a number of unknown origin is no fix at all.
Wording the requirement for incoming inspection
A bare number will be read differently by everyone who handles it. A usable clause has four parts.
- The measurement. Application or removal torque; for removal, whether the figure is the break-loose peak or the sustained unscrewing load.
- The method. Instrument type, grip point on the closure, clamping arrangement and rotation speed.
- The conditions. Conditioning interval after capping, ambient temperature and, where they apply, the ageing period and the product in the pack.
- The limit. A range, noted as a reference range to be confirmed against actual measurement on the specific bottle, closure, liner and product combination. That note is no legal hedge. It accurately states how the number was arrived at.
Then define what happens outside the window. One out-of-range unit in a sample does not automatically condemn a batch, but the reaction must be decided beforehand: resample, check the capping head, check the finish, or hold the batch for review. Where no response is written down, the matter gets settled by negotiation afterwards, which is the costliest route.
On the filling side, the incoming inspection SOP can be short:
- sample size and frequency taken from a recognised sampling scheme;
- a recorded gauge calibration;
- a written grip-point procedure for each closure type in use;
- a filled retention sample kept from every batch;
- a periodic comparison of current readings with those retention samples.
The last item converts isolated numbers into a trend, and a trend reveals a drifting capping head before it turns out a batch of leakers.
Questions buyers ask about cap torque testing
Is there a standard torque value for a 28 mm cap?
No. Anyone quoting one without asking about the finish, the liner and the product is guessing, because nominal diameter says nothing about thread form, material friction or liner compression. Confirm any suggested range by measuring your own pack.
How soon after capping should removal torque be read?
At one fixed interval, used every time. Twenty-four hours is a common compromise between practicality and a stable reading. Extend it where the liner settles slowly or the product migrates into it, and record the interval, temperature and instrument in the specification.
Why is the cap looser after shipping than it was at the filler?
Liner relaxation lowers the axial load with time, and transport vibration lets the cap rotate gradually toward open. Together they mean the post-distribution reading is under the one taken at the plant, so judge the pack on the post-vibration figure.
Does more torque always give a better seal?
No. Beyond the upper edge of the window the liner cold-flows off the rim, skirts split, thin glass finishes chip and older users struggle to open the pack. Find the seal threshold experimentally and run above it with margin; do not simply tighten until the leak stops.
What should we send to get a torque test plan for our pack?
Three items: closure type, liner type and contents. Add the finish designation, fill volume, closing equipment and destination market if any are already fixed. From those we can tell whether thread torque is even the right measurement, what liner behaviour to expect, how the pack should be conditioned before reading, what the seal threshold test should look like on your line, and how to word the resulting window so a supplier and an inspector read it alike.