Before releasing a bulk order, buyers should run six compatibility tests against the actual palette, the actual pan coating, and the actual formula: accelerated aging at elevated temperature, a direct contact and staining test using the pigmented oil phase, a pan adhesion test after formula contact, a filled drop and handling test, a closure and hinge performance test on filled units, and a decoration durability test on every decorated surface that the formula or the consumer will touch. Those six cover the failure modes that reach the consumer — oil migration and plasticization, stress cracking, irreversible staining, pan pop, hinge fatigue, and print or coating breakdown — and each of them has a defensible method, a sample size, and a numeric pass criterion that can be written into a purchase specification.
The list is short, but the discipline behind it is not. Most compatibility failures in eyeshadow packaging are not caused by a bad material or a bad formula in isolation; they are caused by a combination that was never tested together. A cream shadow that performs flawlessly in a glass jar can soften a styrenic compact within eight weeks. A pressed powder that looks inert can leave a permanent halo of carmine red on a white shell after a single summer shipment. A hinge that passes 20,000 empty cycles can seize after the formula's volatile carrier has been partitioning into the polymer for a month. The purpose of a pre-production compatibility programme is to force those combinations to fail in a laboratory, at a time when the failure costs a sample batch rather than a container load.
Formula compatibility is the ability of a filled package to remain within its functional, cosmetic, and safety specifications for the whole of its intended shelf life, under the conditions of storage, transport, and use that the product will realistically meet. It is a property of a system, not of a material. The system comprises the formula, the pan or godet and its coating, the adhesive or retention geometry that holds the pan, the polymer of the shell and lid, the hinge, the closure mechanism, the mirror and its reflective coating, the gasket or liner if one exists, and every decoration — spray coating, metallization, hot stamping, UV varnish, printing, and label adhesive.
That definition matters commercially because each element is usually sourced or specified by a different party. The formula house supplies the bulk. The converter moulds the shell. A second supplier coats the pan. A third prints or metallizes the lid. The brand approves each one separately against its own datasheet, and then assembles them for the first time in the filling hall. Compatibility is the only specification that crosses all of those boundaries, and it is the one most often left implicit in a purchase order.
There is a second reason to treat compatibility as a system property: the failure is usually time-dependent and non-linear. A polymer that gains two percent in mass over four weeks may gain fifteen percent over twelve, because absorption follows a diffusion front that slows as it advances but does not stop. A bond that holds at 23 °C may creep at 45 °C because the adhesive's modulus drops through its softening range. A crack that needs 10,000 cycles to initiate at room temperature may need 900 at 50 °C. Single-point testing at ambient conditions therefore answers the wrong question. The right question is not "does it work today" but "where is the knee in the curve, and does our shelf life sit safely on the flat side of it".
Suppliers frequently describe a resin or a coating as cosmetic grade, food contact compliant, or suitable for cosmetic packaging. These statements are useful starting points and almost meaningless as acceptance criteria. They typically confirm that the material's formulation avoids a short list of restricted substances and that the supplier has run a generic migration screen, usually against a simulant such as ethanol or olive oil at a defined temperature for a defined time. They do not tell a buyer how that material behaves against a specific cream containing 35 percent ester oils, 8 percent volatile silicone, a dispersed organic pigment, and a fragrance with a high ketone content, held at 45 °C in a shipping container for six weeks.
The gap is not dishonesty; it is scope. A resin supplier cannot test every formula in the world, and a formula house cannot test every polymer. The compatibility test exists precisely to close that gap for one formula and one package. A buyer who accepts a generic compliance letter in place of a system test is accepting a probability, not a specification, and the probability is worse than it looks because the highest-risk formulas are exactly the ones that generic simulants under-represent: high-spreadability creams, balms with high oil load, and hybrids that combine volatile carriers with high pigment loading.
In practice, compatibility failures cluster at three interfaces, and a test programme should be organised around them rather than around the bill of materials.
A test plan that covers only the first interface will pass a palette that fogs its own mirror, lifts its own logo, and drops its pans into the consumer's lap. All three must be sampled, and they must be sampled on the same aged units so that the interactions between them are captured rather than averaged out.
The three formula families differ less in what they contain than in how mobile those constituents are. Mobility — the ability of a molecule to diffuse into a polymer at a given temperature — drives every compatibility outcome. Pressed powders are low-mobility systems with high pigment load and minimal liquid phase. Creams are high-mobility systems with a continuous oil phase. Hybrids are high-mobility systems that lose their carrier during use, which means their attack profile changes over the life of the product. The same polymer can be adequate for one and unacceptable for another, and the same formula can be acceptable in one geometry and unacceptable in another because wall thickness, moulded-in stress, and gate location all change the outcome.
A pressed eyeshadow typically contains 3 to 10 percent binder and oil phase, with the balance made up of talc, mica, sericite, silica, bismuth oxychloride, and pigments. On paper this is the mildest of the three families, and in practice pressed powders rarely plasticize a shell. Their failure modes are different and often underestimated.
The first is staining. A pressed powder carries a high concentration of colourant in intimate contact with the shell through loose dust, through the pan rim, and through the gap between the pan and the cavity wall. Organic lakes and dyes — particularly reds, and particularly those carried on an aluminium lake substrate — can migrate into a white or light-coloured styrenic shell and leave a permanent pink or grey shadow around the pan well. The migration is driven by humidity and temperature and is often invisible at 23 °C. It appears after a container crossing the equator.
The second is dust ingress. Fine powder migrates into the hinge, the closure magnet pocket, and the mirror recess. Once there, it acts as an abrasive, increases hinge torque, and produces a visible film on the mirror that the consumer reads as a quality defect even though it wipes off. Powder with a high talc fraction is also strongly triboelectric; it clings to the shell by static and resists brushing, which is why cleanliness specifications need a defined wipe and a defined inspection light rather than a visual glance.
The third is pan adhesion under thermal cycling. A pressed powder is compacted under high pressure, and the compact can shrink or relax after pressing. If the pan is shallow and the press density is high, the cake can lift at the edge and generate a fine ring of loose powder that migrates to the hinge. The shell is not attacked, but the package still fails the consumer's cleanliness expectation.
A cream eyeshadow or a balm typically contains 25 to 60 percent of an oil phase: esters such as isopropyl myristate, isononyl isononanoate, and caprylic/capric triglyceride; natural oils; waxes; petrolatum; silicone fluids; and thickeners. It is poured warm and sets in the pan, and from the moment it sets it begins to partition into everything it touches.
The dominant mechanism is absorption followed by plasticization. Small ester molecules enter the amorphous regions of a polymer, increase the free volume between chains, and lower the glass transition temperature. The visible consequences follow in a predictable order. First the surface loses gloss and takes on a slightly greasy sheen. Then the polymer swells, typically by 0.5 to 3 percent in linear dimension, which is enough to bind a hinge, lift a lid seam, or make a snap closure feel tight and then loose. Then the modulus drops, so a lid that resisted thumb pressure at week zero deflects visibly at week eight. Finally, in severe cases, the shell deforms enough that the lid no longer sits flush and the closure loses its interference fit.
The severity depends on the polymer's solubility parameter relative to the oil. Styrenic polymers — general-purpose polystyrene, SAN, and ABS — are attacked by many esters and by most fragrance solvents, and ABS in particular has poor resistance to long-chain esters because the butadiene phase provides an accessible diffusion path. Polypropylene is more resistant when it is highly crystalline but is attacked by non-polar oils and by some silicones, and its resistance falls sharply at elevated temperature. PETG offers good resistance to many oils but is sensitive to stress cracking in the presence of certain solvents and to hydrolysis at high humidity. Acrylic polymers are optically excellent and chemically poor against esters and ketones. There is no universal answer, which is why the test must be run rather than reasoned.
Creams also drive a mechanism that buyers rarely specify: wicking through seams. A shell moulded in two parts with a mechanical joint, an ultrasonic weld, or a snap fit presents a capillary path. A low-viscosity oil phase at 45 °C will travel along that path by capillary action and appear as a bead of oil on the outside of the case. The defect is not a leak in the conventional sense — there is no pressure, no liquid head — but it is visually identical to one from the consumer's point of view, and it stains the carton and the outer box.
Cream-to-powder formulas are the most difficult family to qualify because their chemistry changes during the first minutes of use. They are poured or pressed as a semi-solid containing volatile carriers — cyclopentasiloxane, isododecane, or a volatile hydrocarbon — which flash off after application and leave a powder-like film. In the package, before opening, the volatile fraction is trapped in the headspace and saturated against the shell, the liner, the mirror, and the print.
Volatile silicones are aggressive toward many plastics. Cyclopentasiloxane and dimethicone have low surface tension and low viscosity, and they wet and penetrate styrenic and olefinic surfaces readily. Their effect is often not swelling but stress cracking: they lower the surface energy enough to initiate craze lines at points of moulded-in stress, and the craze grows into a crack under the cyclic load of opening and closing. A hybrid palette can look perfect after filling and show hairline cracks radiating from the hinge boss or the snap bead after three weeks at 50 °C.
The second hybrid-specific risk is shrinkage stress at the interface. As the volatile carrier leaves the formula during use, the remaining mass contracts. If the formula has bonded to the pan wall — which is desirable for adhesion — the contraction pulls on the bond line and can generate a clean separation at the pan edge. The result is a cake that rattles in its pan after a month of use, a defect that no incoming inspection of empty components will ever detect.
The third is odour and taint. Volatile carriers and fragrance components absorbed into a polymer are released slowly over the life of the package, and a case that smelled neutral at filling can develop a stale, solvent-like note after storage. For a product positioned as fragrance-free or clean, this is a claim-level failure, not a cosmetic one.
Staining deserves separate treatment because it is the only common compatibility failure that cannot be corrected after the fact. Swelling can recover when the migrant diffuses back out. A crack can be re-moulded. A stain is a chromophore that has diffused into the polymer matrix or reacted with it, and no wiping, polishing, or surface treatment will remove it.
Three mechanisms operate. The first is simple dissolution and diffusion of a dye or an oil-soluble pigment into the amorphous phase; this is dominant in styrenics and in polyolefins with low crystallinity. The second is adsorption onto pigment particles that have themselves migrated into surface micro-cracks, which produces a grey or brown haze rather than a sharp colour. The third is chemical reaction — oxidation of a phenolic antioxidant, reaction of an amine with an acidic dye, or sulphur staining of a silver mirror — which produces yellowing or browning that is often misdiagnosed as "yellowing of the plastic" rather than as a formula interaction.
The practical consequence is that staining must be tested with the real pigment system at the real concentration, on the real colour of shell, at the highest temperature the product will meet. Testing a white shell with a clear oil is not a staining test. Testing a nude shade on a black shell will not reveal the red halo that a berry shade produces on a cream shell. The test has to use the worst-case shade and the worst-case shell colour, and the worst case is usually the combination the brand intends to launch last.
The pan is the most overlooked component in a compatibility programme. Most eyeshadow pans are tinplate with an internal lacquer, aluminium with or without an anodized layer, or a moulded plastic godet. Each presents a different surface to the formula, and each fails differently.
Tinplate pans rely on an organic lacquer for corrosion resistance and for adhesion. A cream with a free fatty acid content, a low pH, or a high water activity can attack that lacquer at the cut edge or at a coating pinhole, producing a ring of corrosion that lifts the lacquer and releases the pressed cake. The same lacquer can be softened by ester oils, which turns a hard, adherent film into a tacky layer that flows under the compaction pressure of a pressed powder and allows the cake to shift.
Aluminium pans without a coating form a thin native oxide that adheres well to many pressed powders but is amphoteric and corrodes in acidic or alkaline formulas. Anodized aluminium is more robust but is brittle; it cracks at the pan radius when the pan is crimped or when the cake expands, and the crack becomes a corrosion initiation site. Plastic godets eliminate corrosion but introduce a new failure: the godet itself can be plasticized by the formula, soften, and release its grip on the cake.
The adhesive layer — hot melt, pressure-sensitive, or a press-fit with no adhesive — is a fourth variable. Hot melts based on polyolefin waxes are softened by oils and can creep at 45 °C. Pressure-sensitive adhesives can be plasticized at the plasticizer's migration front and lose peel strength gradually. A press-fit relies entirely on dimensional interference, which changes when the shell swells. In all three cases the failure looks the same to the buyer: the pan lifts, usually at the edge opposite the hinge, and the cake cracks.
Environmental stress cracking is the brittle failure of a polymer under a sustained load that is well below its yield strength, in the presence of a fluid that accelerates crack growth. It is the single most common cause of catastrophic palette failure, and it is almost never predicted by a simple immersion test, because immersion tests do not apply stress.
The ingredients are always the same: a moulded-in stress, a stress concentrator, a chemical agent, and time. Moulded-in stress comes from packing pressure, from a gate that freezes before the part, from uneven wall sections, and from ejection. Stress concentrators are the hinge web, the snap bead, the boss for a magnet, the sharp internal corner of a pan well, and any weld line. The chemical agent is the formula's oil phase, its volatile carrier, or a mould release residue. Time and temperature do the rest.
The failure signature is diagnostic. Cracks initiate at the stress concentrator, run perpendicular to the principal stress, and often show a mirror-smooth initiation zone followed by a rougher propagation zone. They appear on the inside surface first, where the formula is, and propagate outward. A buyer who sees a crack on the outside of a lid should assume it started on the inside three weeks earlier.
The six core tests below are the minimum for a bulk release decision. They are ordered by the sequence in which they should be run, because each one generates information that informs the next: aging reveals which interfaces are at risk, contact staining isolates the mechanism, pan adhesion quantifies the bond, drop and handling test the assembled system, closure and hinge tests quantify the mechanisms, and decoration testing closes the loop on the visible surfaces. A buyer with limited time should not skip any of them, but should run them in parallel on the same sample population so that a single aged unit can be evaluated for all attributes.
Accelerated aging is the backbone of the programme. The objective is not to prove that the product survives a specific temperature; it is to establish the shape of the property-versus-time curve so that the buyer can estimate where the product will be at the end of its real shelf life. A single time point is an anecdote. Three time points are a trend.
The standard design uses at least two elevated temperatures, typically 40 °C and 50 °C, with a controlled humidity condition of 75 percent relative humidity for the 40 °C cell and ambient humidity for the 50 °C cell. Samples are stored upright and inverted, because a cream will contact the lid and the mirror in one orientation and the pan well in the other, and the two orientations produce different failure modes. A freeze-thaw cell cycling between −10 °C and 45 °C for three to five cycles is added for products that will ship through cold-chain or unheated transport. Where the product will be displayed under strong lighting, a xenon-arc or fluorescent light exposure cell is added to catch photo-oxidation and colour shift in the decoration.
Pull points are typically 2, 4, 8, and 12 weeks at 40 °C, and 1, 2, 4, and 8 weeks at 50 °C. The 50 °C cell is a screening tool, not a shelf-life predictor, because some mechanisms — crystallisation of waxes, migration of plasticizer back out of the polymer, annealing of moulded-in stress — run in the opposite direction at high temperature and can make a failing system look better than it is. The 40 °C cell is the workhorse, and the freeze-thaw cell is the reality check.
Measurements at each pull point should include: mass change of the empty shell and of the filled unit, to quantify absorption and loss; dimensional change at three defined datum points, measured with a micrometer or a vision system; Shore D hardness on a defined area of the shell; gloss at 60 degrees; CIELAB colour of the shell, the decoration, and the mirror; visual inspection under defined lighting at 10× magnification; and a functional check of hinge torque, closure force, and lid flushness. Photographing every unit at every pull point against a grey card is cheap and settles most later disputes.
This test isolates the formula-to-polymer interface and is the only reliable way to rank materials and colours before committing to a mould. It uses coupons rather than finished parts, so that each polymer and each decoration can be evaluated independently and so that the result is not confounded by geometry.
Prepare coupons of every candidate substrate — PP, SAN, ABS, PETG, acrylic, and any regrind blend — at the production wall thickness, and coupons of every decoration — spray-coated, metallized, hot-stamped, UV-varnished, and printed — on their production substrate. Apply the formula in two configurations. In the direct-contact configuration, apply a 2 mm layer of the bulk to a 20 mm diameter area and cover it with a watch glass to prevent drying. In the extraction configuration, prepare the oil phase alone, or a 50:50 dilution of the bulk in the carrier, and apply it as a film. The second configuration is important because it separates the effect of the oil phase from the effect of the pigment, and because it lets a buyer test a formula that has not yet been finalised.
Age the coupons at 50 °C for four weeks and at 40 °C for eight weeks, with a control set at 23 °C. At each pull point, remove the formula with a defined wipe protocol — a dry lint-free wipe followed by three wipes with isopropyl alcohol, then a 24-hour recovery — and measure ΔE against the unexposed control using a spectrophotometer with a 10-degree observer and D65 illuminant. A ΔE above 2.0 is visible to a trained eye; above 3.0 it is visible to a consumer; above 5.0 the part is scrap. For mirror surfaces, measure haze and distinctness of image in addition to colour, because a mirror can fail by fogging without changing colour.
Run the same coupons with a simulated perspiration and sebum solution, because the consumer's skin contact adds a second chemical load that the formula alone does not represent. A shell that survives the formula but fails after ten hand contacts will fail in the field within a week of launch.
Pan adhesion must be measured after the pan has been in contact with the formula and after thermal aging, not on a clean, freshly assembled sample. The reason is that the bond is changed by the formula: hot melt softens, pressure-sensitive adhesive plasticizes, and press-fit interference relaxes as the shell swells.
The measurement is a pull-out force using a force gauge with a suitable fixture — a vacuum cup for a flat cake, a slotted plate for a domed cake, or a custom shoe that engages the pan rim without touching the cake. Apply the load perpendicular to the pan plane at a constant rate of 10 mm per minute and record the peak force and the failure location: adhesive to pan, adhesive to shell, cohesive failure within the adhesive, or cake failure. The failure location is more informative than the number, because it tells the buyer which interface to change.
A typical acceptance target for a small eyeshadow pan is 15 to 30 N, with a lower 5th-percentile limit written into the specification and a requirement that no sample fails by clean interfacial separation. For larger blush or bronzer pans the target scales with pan area and should be set by the buyer's own application, not copied from a supplier's catalogue. Measure at 23 °C and at 45 °C, because a bond that passes at room temperature can creep below the limit in a warm warehouse.
A complementary test is the twist test, in which a defined torque is applied about the pan axis. It catches pans that are held only by friction and that will spin in the consumer's hand when the brush is pressed into the cake. A pan that spins is functionally as bad as a pan that lifts, and it is invisible to a pull test.
Drop testing an empty palette is a moulding test. Drop testing a filled palette is a compatibility test, because the mass of the cake changes the load path and the formula changes the stiffness of the assembly. The filled unit should be dropped from 1.2 m onto a rigid concrete surface, on each of six faces and on the hinge edge, using a guided-drop apparatus that controls attitude. A production-relevant sequence is five drops per orientation on three units per cell, with inspection after each drop.
Inspection criteria should be defined before the test, not after. Acceptable: scuffing of the decoration that does not expose substrate, a hinge that still holds the lid at its designed angle, and a cake that remains intact. Unacceptable: any crack visible at 10× magnification, a lid that separates from the hinge, a pan that lifts, a cake that cracks through, a mirror that shatters or separates, and a closure that no longer latches. Photograph the failure and record the drop number and orientation.
Add a handling test that simulates the retail environment: 200 open-close cycles by hand or by actuator, fifty deliberate knocks against a hard edge, and a pocket test in which the closed unit is tumbled with keys and coins for 30 minutes. The pocket test is the most realistic predictor of decoration wear and is the one most often omitted.
Closure performance is measured on filled, aged units. The relevant metrics are opening force, closing force, lid-to-base gap or flushness, and, where a gasket or a seal is present, leak tightness. Measure opening force with a force gauge at a defined point on the lid edge and a defined rate; measure flushness with a feeler gauge at four quadrants; measure leak tightness by vacuum bubble emission or pressure decay if the product is a cream that must not dry out.
The critical variable is the change, not the absolute value. A palette whose opening force drifts from 8 N to 3 N over eight weeks at 40 °C has a closure that will feel loose in the consumer's hand even though 3 N is still a positive retention. A palette whose opening force rises from 8 N to 20 N has a shell that is swelling into its own interference fit and will eventually crack at the snap bead. Both are failures, and both are only visible if the baseline was measured on the same fixture before aging.
Hinge fatigue is tested separately on an automated cycling rig at 5,000, 10,000, and 20,000 cycles, at 23 °C and at 40 °C, with torque measured every 1,000 cycles. The pass criteria are: no crack at the hinge web, torque within ±30 percent of the initial value, and the lid holding its open angle without droop. A hinge that passes 20,000 cycles at 23 °C but fails at 6,000 at 40 °C has told you that the formula is plasticizing the hinge, and the fix is a material or geometry change rather than a longer cycle specification.
Decoration is part of the package's function, not a cosmetic afterthought. A logo that blurs, a metallization that blisters, or a label that lifts is a compatibility failure, and it is usually the first one the consumer sees.
Test adhesion by cross-hatch according to ASTM D3359, before and after 40 °C aging and after a defined rub with the formula. Test solvent resistance by MEK double rubs according to ASTM D5402 on a non-visible area. Test abrasion by Taber according to ASTM D4060 with a defined wheel and load, and report the haze change rather than the weight loss alone. Test print legibility after the pocket test and after a wipe with the formula and with a 70 percent alcohol wipe, because consumers clean their compacts.
For metallized and mirror surfaces, add a humidity test at 40 °C and 90 percent relative humidity for 96 hours, followed by a tape pull. Silver and aluminium coatings fail by hazing, by pitting, and by delamination at the edge, and the failure is accelerated by sulphur compounds in the formula and by residual moulding volatiles in the shell.
The table below is a working template. The sample sizes and pass criteria are starting points calibrated to a typical eyeshadow or blush compact; a buyer should tighten them for a premium product and loosen them only with a written rationale. Every row should appear in the purchase specification with the test method, the lab, the pull points, and the disposition rule for a failure.
| Test | Method outline | Sample size | Pass criteria | Typical failure mode |
|---|---|---|---|---|
| Accelerated aging | 40 °C / 75% RH for 12 weeks; 50 °C for 8 weeks; freeze-thaw −10 to 45 °C, 5 cycles; upright and inverted | 30 filled units per cell; 10 empty shells as control | ΔE ≤ 2.0; dimensional change ≤ 0.3%; mass change ≤ 1.0%; no crack at 10×; hinge torque within ±30% | Swelling, lid warp, hinge stiffening, mirror haze |
| Contact staining | Formula or oil phase on coupons, 50 °C / 4 weeks and 40 °C / 8 weeks, defined wipe, spectrophotometer | 5 coupons per substrate-decoration pair; minimum 12 pairs | ΔE ≤ 2.0 after wipe; no residual colour at 10×; grey scale ≥ 4–5 | Permanent halo, dye migration, yellowing |
| Pan adhesion | Pull-out at 10 mm/min, 23 °C and 45 °C, after 4-week aging; twist test at defined torque | 12 pans per cell from 3 production lots | Peak force 15–30 N; 5th percentile ≥ 12 N; no interfacial separation; no rotation at 0.5 N·m | Pan pop, spinning pan, cake cracking |
| Drop and handling | Guided drop 1.2 m, 6 faces + hinge edge, 5 drops per orientation; 200 open-close; pocket tumble | 9 filled units per cell | No crack, no hinge separation, no pan lift, closure still latches | Hinge snap fracture, lid split, mirror shatter |
| Closure performance | Opening force at defined point and rate; feeler-gauge flushness at 4 quadrants; vacuum leak test | 12 filled units per cell, measured before and after aging | Opening force 5–15 N; flushness ≤ 0.15 mm; no bubble in 60 s at −40 kPa | Loose lid, gap, drying of cream, leakage |
| Hinge fatigue | Automated cycling at 23 °C and 40 °C, torque logged every 1,000 cycles | 6 units per condition | 20,000 cycles without crack; torque within ±30%; no droop | Craze at hinge web, snap failure, lid droop |
| Decoration adhesion | ASTM D3359 cross-hatch before and after aging; tape pull | 5 areas per decorated part; 10 parts | Classification 4B or better; no lifting at edges | Print blur, coating blister, foil lift |
| Solvent and abrasion | ASTM D5402 MEK double rubs; ASTM D4060 Taber, 500 cycles | 3 specimens per decoration | ≥ 50 double rubs without breakthrough; haze change ≤ 10% | Gloss loss, colour shift, exposed substrate |
| Humidity and mirror | 40 °C / 90% RH, 96 h, tape pull and haze measurement | 10 mirrors per coating type | No haze increase > 2%; no delamination; no pitting | Mirror fog, edge corrosion, silver tarnish |
| Compatibility of gasket or liner | Compression set at 40 °C for 72 h; swell in formula; hardness change | 10 specimens per material | Compression set ≤ 25%; swell ≤ 10%; hardness change ≤ 10 Shore A | Gasket swell, loss of seal, odour |
| Analytical root cause | FTIR-ATR, DSC, GC-MS of migrants, gravimetric sorption | 3 aged and 3 control samples | No unaccounted spectral change; Tg shift ≤ 5 °C | Plasticization, oxidation, unknown migrant |
Field failures are more instructive than laboratory passes because they show which mechanisms actually reach the consumer. The list below is drawn from the recurring defects that buyers report after launch, organised by symptom so that an investigation can start from what is visible.
The lid will not sit flush after two months. This is almost always differential swelling: the base has absorbed more oil than the lid because the formula sits in the base, and the resulting dimensional mismatch opens the seam at the front edge. The fix is a symmetric wall section, a more resistant polymer in the base, or a closure geometry that tolerates 0.3 mm of growth. Testing empty shells will never reveal it.
A hairline crack runs from the hinge boss. This is environmental stress cracking, initiated at the highest-stress point and propagated by the volatile carrier. It is more common in hybrids and in palettes with a thin hinge web. The fix is to relieve moulded-in stress, increase the fillet radius, or change the resin; a thicker hinge does not help if the stress is tensile on the inside surface.
The mirror is fogged or has a milky edge. Volatile carriers and fragrance components condense on the mirror and attack the coating or its adhesive. The edge haze is the first sign, because the coating is thinnest and the adhesive is exposed there. A mirror recess that is open to the formula chamber makes this worse; a physical barrier or a change of mirror adhesive solves it.
Pans lift at the edge opposite the hinge. This is adhesive creep under thermal load, often combined with shell swelling that relieves the press-fit. It appears first on the pans furthest from the gate, where moulded-in stress is highest. A pull test at 45 °C after four weeks of aging will predict it.
A pink or grey halo surrounds the pan well. Pigment migration into a light shell, driven by humidity and by direct contact with loose powder. It is irreversible, and it is more severe on matt or textured surfaces because the surface area is higher. Testing with the actual shade on the actual shell colour is the only defence.
The decoration blisters or the logo rubs off. Either the coating was under-cured, or the formula's solvent has penetrated it, or the moulding released agent was incompletely removed before coating. A cross-hatch test after aging distinguishes the three, and an FTIR of the blister surface confirms solvent attack.
The case smells of solvent after storage. Absorption followed by slow desorption. The polymer acts as a reservoir, and the odour is released over weeks. It is a claim-level failure for fragrance-free products and is best screened by a sensory panel on aged units, not by instrumental analysis alone.
The gasket swells and the lid no longer seals. Elastomer swell in the oil phase, usually accompanied by a hardness drop and a permanent set. Compression-set testing at 40 °C catches it before launch, and a change of elastomer compound is usually the only fix.
A filling trial is the bridge between a laboratory coupon and a production line, and it is where most compatibility programmes either become credible or collapse. The trial must use production tooling, production resin, production decoration, and production formula, on the line that will actually run the order. A trial run on a different machine with a lab-mixed batch answers a different question.
Select the trial cells deliberately. At minimum, run the highest oil-load shade, the most strongly pigmented shade, the lightest shell colour, and the most heavily decorated lid, because those four carry the highest risk and they are rarely the same SKU. If the range includes a hybrid or a balm, it gets its own cell. If the range includes a refillable or magnetic format, that gets a separate cell because the magnet pocket and the removable pan introduce a different stress path.
Run a minimum of three consecutive production lots of shells if the schedule allows, because lot-to-lot variation in regrind content, colour masterbatch, and moulding conditions is a real source of compatibility scatter. A single lot that passes tells you the lot passed; three lots that pass tell you the process is in control.
A practical trial size is 300 to 500 filled units per cell: enough for 30 units per pull point across four pull points, plus drop and hinge samples, plus retain samples for the brand and the factory, plus a margin for handling loss. Retain at least 20 filled units from each cell in the original shipper, unopened, at 23 °C, for the life of the programme; they are the reference against which every later complaint is judged.
In-process checks should be simple and fast: fill weight within tolerance, pan height relative to the shell, lid flushness, closure force on a sample every 30 minutes, and a visual check under a defined light. Record the line speed, the fill temperature, the cooling time, and the torque of any capping station, because these variables change the stress state of the finished unit and therefore change its compatibility outcome.
After filling, condition the units for 24 hours at 23 °C before packing, and pack them in the production shipper, not in laboratory trays. The shipper is part of the stress environment: a nested stack applies a sustained load to the lid, and a carton that is over-packed can pre-load the hinge. Compatibility failures that appear only in shipper-packed units are common and are missed by lab testing on loose samples.
The trial is not finished when the units pass. It is finished when the process window is documented and the change-control rules are written. Record the approved resin grade and supplier, the masterbatch percentage, the regrind allowance, the moulding conditions, the decoration supplier and cure schedule, the pan coating lot, the adhesive lot, and the formula version. Any change to any of these should trigger a partial re-test — at minimum a four-week 50 °C aging and a pull test — because compatibility is a system property, and a system change invalidates the previous result.
The most expensive compatibility failures in the industry are not caused by a bad first design. They are caused by an unannounced resin substitution, a regrind percentage increase, or a decoration supplier change made to protect a delivery date, six months after the original approval. A change-control clause with a defined re-test protocol is worth more than an extra pull point.
Compatibility decisions interact with commercial ones, and buyers should understand the trade-offs before they fall in love with a custom shell. Custom injection tooling for a multi-cavity palette mould is a significant fixed cost, and it is usually amortised into the unit price across the first order or charged separately. Stock moulds avoid that cost but limit the geometry, the wall thickness, and often the polymer choice, which in turn limits the compatibility options. A stock mould in a general-purpose styrenic may be perfectly adequate for a pressed powder and marginal for a high-oil cream, and no amount of testing will change the resin if the tool is cut for it.
MOQ is driven by three independent numbers: the moulding MOQ per colour, the decoration MOQ per artwork, and the pan or godet MOQ per size. The effective MOQ for a SKU is the highest of the three, which is why a 3,000-unit palette order can become a 10,000-unit commitment once a custom spray colour and a hot-stamp foil are added. Buyers should ask for the three numbers separately and negotiate them separately; bundling them hides the real constraint.
Lead time has a compatibility dimension that is often ignored. A 30-day moulding lead time plus a 15-day decoration lead time plus a 20-day filling lead time leaves the formula in contact with the shell for a shorter or longer period before the consumer sees it, depending on the order sequence. If the palettes are decorated and then stored for two months before filling, the decoration sees no formula and the shell sees no oil; if they are filled immediately, the clock starts earlier. The compatibility protocol should define t=0 as the date of filling, not the date of moulding, and the shelf-life claim should be referenced to that date.
Logistics is where laboratory aging meets reality. A marine container on a tropical route can reach 60 to 70 °C in the headspace for days at a time, with a diurnal cycle that pumps vapour in and out of the package. That condition is more severe than a 50 °C oven because it is cyclic and because the package is under the mechanical load of stacking. Buyers shipping to hot climates should add a container simulation cell — 60 °C for 72 hours, followed by 25 °C for 24 hours, repeated three times, with the units in the production shipper under a 20 kg stack load — and should specify a desiccant and a liner board that can survive the humidity.
Packaging of the empty palettes themselves matters for the same reason. Palettes shipped nested without individual protection will abrade each other's decoration; palettes shipped in a rigid tray with a lid will resist compression but add freight volume. The choice affects the decoration durability result, so it should be fixed before the decoration test is run, not after.
A defensible programme takes 12 to 16 weeks from receipt of the production-intent formula and the production mould. The critical path is usually the 40 °C aging cell, which needs 8 to 12 weeks to show a trend, plus four weeks for the contact staining coupons and two weeks for the filling trial and the drop, hinge, and closure work. The 50 °C screening cell can run in parallel and gives an early warning at two weeks, but it cannot replace the 40 °C data because some mechanisms reverse at high temperature. Buyers who are offered a two-week "compatibility report" are being offered a screening result, not a shelf-life basis.
Only if the supplier's data was generated on the same formula, the same resin grade, the same pan coating, and the same decoration, at the same wall thickness. In practice that is almost never true. Supplier data is usually generated on a reference formula — often a standard oil blend or a generic cream — and on a reference colour. It is useful for screening out obviously incompatible materials and for setting expectations, but it cannot substitute for a test on the buyer's own system, because the two highest-risk variables, the pigment system and the fragrance, are exactly the ones the reference formula omits.
No. Wall thickness, gate location, hinge geometry, pan depth, closure interference, and the presence of a mirror all change the stress state and the diffusion path. A cream that is stable in a thick-walled compact can crack a thin-walled palette at the hinge within a month, because the thinner section has less material to absorb the same migrant and a higher moulded-in stress. Every new mould requires at least a partial re-test, even when the formula and the resin are unchanged.
Thirty filled units per aging cell, twelve pans per pull test cell, nine units per drop cell, and five coupons per substrate-decoration pair is a workable minimum for a single-SKU launch. For a range of six shades on two shell colours, the matrix multiplies quickly, so the practical approach is to test the worst-case shade on the worst-case shell for every attribute, and to test the remaining shades only for staining and pan adhesion. Document the rationale for the reduction; an auditor or a retailer will ask.
Test the oil phase and the pigment dispersion separately, then test the finished formula when it is locked. The oil phase alone is a conservative proxy for plasticization and stress cracking, because it omits the pigment that can absorb some of the oil and slow migration. The pigment dispersion alone is a conservative proxy for staining. Running both on coupons takes two to three weeks and lets the buyer eliminate materials before the formula is frozen, which is when a change is still cheap.
Yes, and the differences are significant. A removable pan introduces a repeated insertion and extraction cycle that stresses the pan wall and the cavity, so the pull test must be run after 50 insertion cycles, not once. The magnet pocket is a stress concentrator and a trap for loose powder, and the magnet itself can corrode if the formula reaches it. The gap between the refill pan and the cavity is a capillary path for oil, so the wicking test becomes essential. A refillable format should be treated as a new system even when the shell is unchanged.
Compatibility is not a certificate and it is not a datasheet. It is a measured property of one formula inside one package, and the only way to know it is to put the two together under conditions that are harsher than the ones the consumer will meet, for longer than the consumer will keep the product, and to measure what happens at intervals rather than once. The six tests described here — accelerated aging, contact staining, pan adhesion after contact, filled drop and handling, closure and hinge performance, and decoration durability — cover the mechanisms that actually cause returns: oil migration and plasticization, environmental stress cracking, irreversible staining, pan pop, hinge fatigue, and coating breakdown.
The commercial argument for running them is straightforward. A failed bulk order costs the moulding, the decoration, the formula, the freight, the launch window, and the retail relationship. A compatibility programme costs a few hundred sample units, twelve to sixteen weeks, and a laboratory invoice. The ratio is not close. Buyers who treat compatibility as a line item in the sampling budget rather than as a gate before the purchase order consistently discover the difference at the worst possible moment: after the container has been loaded, in a market where the launch date cannot move.
The discipline that makes the programme work is not sophisticated instrumentation. It is the habit of defining the pass criterion before the test, sampling the worst case rather than the nominal, ageing the filled assembly rather than the empty component, measuring at more than one time point, and writing the change-control rule that forces a re-test when anything in the system moves. Those four habits, applied consistently, will catch more failures than any single expensive instrument, and they will make the difference between a palette that survives its first summer and one that comes back.
Jinhua Xingqiao Plastic Industry Co., Ltd.
Mobile phone: +86-13868982062
Telephone: +86-579-82196667
Fax: +86-579-82196656
E-mail: sales@zjxqpackaging.com
Address: 377 Yongxue Street, Wucheng District, Jinhua City, Zhejiang Province, China
Copyright © 2022 Jinhua Xingqiao Plastic Industry Co., Ltd
This website uses cookies to ensure you get the best experience on our website.
Phone
Comment
（0）