Before a refill system goes to production, the interface is qualified on 20 samples cycled 100 times each: engagement and release force measured with a force gauge must stay inside the specified band with no cracking of the click feature. The pod-to-outer seal is held inverted at 5°C and 40°C with no leakage or gap. The pod sits in the actual formula at 48°C for 30 days. Airless cartridges add −0.08 MPa for 15 min with ≤0.5% loss and an inner-to-outer pull of ≥50 N. The pod must still click after a 76 cm to 1.5 m drop. All of it runs inside the 40–55 day cycle on a stock outer.
Why a refill interface is tested differently
Three things happen to a refill system that never happen to a single-use pack. The interface is cycled: every refill is a release and a re-engagement, and any drift in force means the system is loosening or wearing. The seal is thermally cycled in use: the same jar lives in a fridge at 5°C and a bathroom cabinet at 40°C, and a seal that holds at room temperature can open when warm. And the contact material changes: the customer is no longer touching the decorated outer but the pod, so compatibility has to be proven on the pod resin, not on the glass.
The qualification battery below answers those three, then runs the standard checks from the quality standards page on the assembled set. The figures are the thresholds applied by partner factories and written into the specification before samples are made; product-specific bands are agreed on the drawing.
The refill-specific tests
| Test | Method | Pass threshold | Why it matters |
|---|---|---|---|
| Pod engagement / release | Force gauge; click the pod in and out; 100 cycles on each of 20 samples | Force stays inside the specified band across all cycles; no cracking of the click feature | Drift in force over cycles means the system degrades; consumers notice a loose or stiff pod long before it fails |
| Pod-to-outer seal at temperature | Filled system held inverted at 5°C and at 40°C | No leakage, no gap at either temperature | Fridge storage and bathroom heat are the real service range; a seal fine at 22°C can leak warm |
| Pod formula compatibility | Pod in contact with the actual bulk at 48°C for 30 days, beside ambient and cold controls | No discolouration, odour, swelling, cracking or decoration attack | The pod is the contact material; PP, PCR-PP and mono-PE behave differently with actives, oils and high-water formulas |
| Airless cartridge vacuum leak | −0.08 MPa for 15 min | ≤0.5% loss; inner-to-outer pull ≥50 N; priming ≤8 strokes | A cartridge is an airless bottle without its outer and is tested as one |
| Inner-lid seal (jar cup) | Colour-water test at −0.06 MPa for 5 min | No dye outside the seal area | Confirms the replaceable cup seals against its own lid, not just against the outer |
| Drop with the outer | 76 cm to 1.5 m at 45° depending on format; glass outer in its carton | No functional damage; the pod still clicks after the drop | A refill outer that survives the drop but loses its click is a failed system |
The engagement band is specified as a range, not a single number, because the geometry is a click or twist feature and the customer needs a definite action in both directions: firm enough to hold inverted, light enough to release without tools. The band is set on the drawing when the bore, click or twist geometry is drawn to the pod, and the same band applies to every later resin lot. Interface geometry and neck standards for bottle systems (18/410, 20/410, 24/410 on pump-on-cartridge) are on neck finish and closure standards.
The standard battery, applied to the set
Once the interface passes, the assembled system runs the checks every cosmetic component runs. What changes is that the outer and pod are judged together, as a set.
| Test | Method | Pass threshold |
|---|---|---|
| Pump / dispenser life | Continuous actuation; dose weighed 50 strokes ×3 | ≥1,500 cycles; output ≥95% (aqueous) / ≥90% (non-aqueous) |
| Thermal cycling | Filled; 48°C and −15°C, 24–48 h each stage, 2 h recovery | No cracking, warping, leakage or colour change |
| Decoration adhesion | 3M 600 tape on print and hot stamp; 3M 810 over 2 mm cross-hatch on spray and plating | No lift; pencil hardness ≥2H on sprayed outers |
| Closure torque | Torque meter after application | Opening 0.2–0.7 N·m on standard caps; no thread strip |
| Colour | Against the signed golden sample under D65 | ΔE within the agreed tolerance; outer and pod judged side by side |
| Heavy metals | Third-party laboratory on the pod resin lot | Pb + Cd + Hg + Cr(VI) ≤100 ppm |
| Acceptance sampling | ISO 2859-1 | AQL agreed per defect class; OQC report before shipment, as a set |
Two of these deserve attention on refill outers specifically. Laser engraving on aluminium or bamboo has no ink to wear across years of reuse, so it sidesteps the adhesion question entirely. And on anodised aluminium outers, the colour test depends on a retained control sample, because a second anodising batch rarely matches the first without one — a problem that only shows up when a customer buys a second outer a year later. The finish comparison is on the decoration process comparison; airless thresholds on airless, pump and sprayer specifications.
Where the tests sit in the timeline
Qualification is not a step added after production; it sits inside the sampling stage and gates the golden sample. On a stock outer with a custom pod, the whole cycle is 40–55 days at the factory gate, and the sequence runs:
- Stock outer and stock pod samples — 7–10 days, to confirm the outer and the bore.
- Pod tool — a closure-size part at roughly USD 3,000–10,000; first samples off the tool, then a revision if the click band is missed.
- Interface qualification — 100 cycles on 20 samples, 5°C / 40°C seal, drop with the outer. Runs in parallel with decorated outer samples.
- Compatibility hold — pod in the actual bulk at 48°C for 30 days. This is the long pole and starts as soon as pod samples exist.
- Golden sample signed — outer and pod as a set, colour chip under D65, force band recorded. Production starts only after signature.
- In-process and final QC — force and seal checks at set intervals in process; FQC repeats engagement on the production lot; OQC to AQL as a set, certificates and test results packed with the inspection report.
On a fully custom outer the tool adds 45–60 days and the programme runs 90–110 days; the mould is paid 30/40/30 with ownership written to the brand, per mould cost and ownership. Then the calendar: ×1.3 buffer, +15 days across Chinese New Year. The route table is in sourcing refillable cosmetic packaging from China.
What to ask a supplier
- What is the engagement force band on the drawing, and can I see the 100-cycle results on 20 samples from the pre-production lot?
- At what temperatures was the seal held, and was the system inverted?
- Was compatibility tested on the pod resin with my actual bulk, at 48°C for 30 days, or on the outer?
- Is the pod still clicking after the drop test with the outer in its carton?
- Will FQC repeat the engagement test on the production lot, and will the results ship with the inspection report?
A supplier who cannot answer these with numbers is describing a jar with a loose inner, not a refill system. The wider questions are in how to evaluate a China packaging supplier.
Vella is a Guangzhou-based beauty packaging solutions provider that helps indie and emerging cosmetic brands specify, sample and quality-control refill systems from vetted Chinese partner factories operating to ISO 22716 / GMP, with the force band, seal temperatures and compatibility hold written into the specification before the pod tool is cut. System types are on refillable packaging; the format overview is on the sustainable and refillable solution.
Common questions
Qualifying a refill system?
Get the test plan before the tool.
Send the outer, the pod concept and the formula type. Within 24 hours you get the force band, seal and compatibility plan and the minimum per component.
Send a brief →