Key Takeaways
- Fill weight, not fill type, is what most buyers get wrong first. A soft-bodied 30 cm toy takes 14–22 g/L; anything that must sit upright takes 25–35 g/L.
- “3D cotton” means two different things in the supply chain: 3 denier fibre, or 3D crimp. They are not interchangeable, and quoting one gets you the other.
- For the wash-and-recover cycle, the silicone finish and the hollow cross-section matter more than the denier number on the label.
- PE pellets solve drape and bottom weight, then create two problems you cannot undo: the toy cannot be machine washed, and it fails small-parts rules for under-threes.
- Filling is 3–6% of unit cost and roughly 70% of perceived quality. It is the cheapest place in the whole BOM to win or lose a repeat order.
Contents
1. Why the Filling Decides the Product
A plush toy is a fabric shell wrapped around air held in place by fibre. The shell gets the design review, the artwork approval and the Pantone meeting. The filling gets a line on the quotation that says “100% PP cotton” and a number nobody checks. Yet that line controls four things a buyer is judged on after the container lands: gram weight, hand feel, service life, and how much of the shape survives six weeks of vacuum compression.
Gram weight is the first argument. Two factories can quote the same unit price for a 30 cm bear and deliver 55 g of fill versus 78 g. Both are “PP cotton”. The lighter one is not defective; it is simply a thinner toy, and at retail the customer picks it up, feels the slack in the belly, and puts it back. Nobody returns a plush toy for being underfilled. They just do not buy the second one.
Hand feel is the second. The same 70 g of fibre, blown into the same shell, feels like two different products depending on denier and finish. Coarse fibre reads as “stuffed animal”. Fine siliconized fibre reads as “collectible”. That difference costs cents and changes the shelf it belongs on.
Service life is the third, and it is where washing lives. A toy that clumps into three lumps after the first machine wash has a one-star review attached to it forever, and the review names the brand, not the factory.
The fourth is the one buyers discover last, usually in a warehouse in November. Nearly all plush ships vacuum compressed at 4:1 to 6:1 to avoid paying for air across an ocean. A 45-day container journey through the tropics means the bale sits at 55–70 °C under load for weeks. Fibre that has good instantaneous loft but poor compression-set resistance comes out of the bag permanently thin, with hard creases along every fold. The recovery figure you want is the one measured after 24 hours at 50% deflection, not the one measured when the bag is opened.
The 40-gram problem
On a US$6.80 landed 30 cm plush, moving from 55 g to 78 g of 7D hollow fibre adds about US$0.04 per unit. It also adds roughly 0.9 L of compressed volume per carton of 60, which is about US$0.03 in ocean freight. For seven cents you change the product from “carnival prize” to “gift shop”. This is the highest-return seven cents in the entire bill of materials, and it is the line most often negotiated away to hit a target price.
If you are still deciding which of the plush toy filling options fits your product, the useful framing is not “which is best” but “which is best at this fill weight, this wash instruction and this retail price”. The rest of this guide works through the three families that cover almost every order, then puts them in one table.
2. PP Cotton: What 3D, 7D and 15D Actually Mean
“PP cotton” is polyester staple fibre — polyethylene terephthalate, the same polymer as a drinks bottle, spun into filament, cut to 32–64 mm, crimped, optionally siliconized, and opened into a loose cloud before it is blown into a shell. It is not cotton. The name survives because the fibre was developed as a cotton substitute in the 1950s, and it is the single most common source of confusion in a first-time sourcing conversation.
The number before the “D” is denier: the mass in grams of 9,000 metres of fibre. It is a linear density, so lower means finer. Converting it to something you can picture, at a polyester density of 1.38 g/cm³ a solid round fibre measures roughly:
| Fibre | Linear density | Approx. diameter | What it feels like in a toy |
|---|---|---|---|
| 3D | 3 denier | 10 μm | Cashmere-adjacent, almost powdery. Holds shape poorly on its own. |
| 7D | 7 denier | 15.5 μm | The industry default. Soft with enough spring to keep a belly round. |
| 15D | 15 denier | 23 μm | Firm, slightly crisp, high loft. Reads “cushion” rather than “plush”. |
Diameter assumes a solid circular cross-section. Hollow fibre has a larger outside diameter at the same denier, because the polymer is distributed around a void.
2.1 The “3D” ambiguity that costs buyers money
Ask five suppliers for “3D cotton” and you will get two different materials. In some mills, 3D means 3 denier — the fine fibre in the table above. In most of the Chinese and Vietnamese supply chain, 3D means three-dimensional crimp: a spiral, spring-like crimp set into the fibre so it resists packing, usually applied to a 7D hollow fibre and sold as “3D hollow cotton” or 三维中空棉.
These are not the same product, they do not feel the same, and they do not cost the same. A specification that says “3D PP cotton” can be satisfied by either, which means the factory's cheapest compliant interpretation wins. Write the denier and the crimp separately and the ambiguity disappears:
Write this instead
Filling: 100% virgin polyester staple fibre, 7 denier, hollow, siliconized, 3D spiral crimp, 51 mm staple, no recycled content.
Every clause closes a substitution route. “Virgin” and “no recycled content” matter for compliance, not just for feel — see section 7.
2.2 Hollow, siliconized, and why those two words outrank denier
Hollow fibre carries a void down its length. At the same denier it traps more still air, so it lofts higher per gram — typically 10–20% more volume than solid fibre — and it recovers from compression faster because the wall acts like a small spring rather than a solid rod. Almost all toy-grade fibre above 5D is hollow.
Silicone finishing is a surface treatment, usually a polysiloxane spin finish at 0.2–0.5% by weight. It lowers inter-fibre friction, which does three useful things: the fibre opens evenly instead of forming nits during carding; it slides rather than knots when the toy is squeezed; and it releases from the shell during washing instead of felting into a single lump. An unfinished 7D and a siliconized 7D feel almost identical in the showroom and behave completely differently after three wash cycles.
This is the point where cheap filling becomes expensive. Unsiliconeized fibre saves roughly US$0.10–0.15/kg. On a 30 cm toy using 70 g, that is one cent. One cent buys you a filling that clumps at the first wash and a review that says so.
2.3 Denier in practice: where each one belongs
3D fine. Use it where the toy is touched more than it is looked at: baby comfort items, small anime plush sold at a high unit price, the face and paws of a large character. It is too soft to carry a body on its own — a 40 cm bear filled only with 3D fibre slumps at the shoulders within a month of shelf life. The professional answer is a blend, not a substitution.
7D hollow siliconized. The workhorse. It carries a body, holds a belly round, survives vacuum compression and machine washing, and costs the least per litre of loft. If a buyer specifies nothing, this is what a good factory quotes anyway.
15D. For volume and firmness: floor cushions, oversized carnival plush, the arms and legs of mascot-style bodies, backrests. It has the best compression recovery of the three (96%+ at 24 h) and the worst hand feel against a cheek. Never use it in a product photographed in a child's arms.
A layered build gets the best of all three, and costs barely more than a single grade: 15D in the legs and base for structure, 7D through the torso for spring, 3D in the head, paws and any surface inside 50 mm of the fabric. Blowing three grades through three ports takes about ninety extra seconds per unit. It is the single most effective construction upgrade available to a mid-priced plush programme.
2.4 The three failure modes to inspect for
- Clumping. Fibre felts into discrete lumps, usually after washing or after a humid container. Caused by missing silicone finish, over-compaction during blowing, or washing above 30 °C. Visible as hard spots under the fabric; not repairable once set.
- Migration. Fill travels between compartments because the pattern has no internal baffles. A toy with a single continuous cavity will always end up bottom-heavy. Add one stitch line across the hips on anything above 35 cm.
- Strike-through. Not enough fill in a narrow limb, so the fabric goes slack and the seam puckers. The fix is gram weight per limb, not more fill overall — specify 8–12 g for a 22 cm arm on a 30 cm toy and check it with a scale, not by eye.
3. PE Pellets and Beans: Weight, Drape and the Rattle Problem
Polyethylene pellets are small moulded beads, typically 3–5 mm in diameter, with a solid density of 0.92–0.96 g/cm³ and a bulk density around 0.55–0.65 kg/L once poured. They are the standard answer to a specific problem: a plush toy that will not sit up, will not drape, or needs to feel heavier than its size suggests.
There are three legitimate uses, and it is worth being precise about which one you are buying, because they have different specifications.
Use 01
Bottom weighting
80–150 g in a sealed base pad so the toy sits upright on a shelf. The default for shelf-displayed retail plush.
Use 02
Hand and foot fill
20–40 g per limb for a beanbag-like feel. This is what makes a squishmallow-style toy read as “squishy” rather than “stuffed”.
Use 03
Full bean fill
Beanbag chairs and doorstops. 400–900 g/L of volume. Rarely a toy application above 3 years.
Pellet fill is always bagged, never loose in the shell. The inner bag should be a double-stitched, overlocked polyester bag with a minimum 10 mm seam allowance, sized so the pellets are held in a single layer rather than piled — a piled bag has a hard lump in the middle that no amount of outer fill hides. On a 30 cm toy, a 90 × 60 mm base pad holding 110 g is typical.
Three costs come attached, and they should be priced before the design is approved rather than after:
- It cannot be machine washed. A pellet bag survives a wash cycle maybe half the time. When it fails, the beads distribute through the shell and the toy is finished. If the care label says machine washable, the pellet bag contradicts it.
- It fails small-parts rules for under-threes. A 4 mm bead is a choking hazard if released, and EN 71-1 and ASTM F963 both treat it as a small part. A plush with loose pellet fill cannot be labelled for 0–36 months, whatever the outer shell is made of.
- It is dead weight in freight. At 0.6 kg/L, 150 g of pellets adds 0.25 L of non-compressible volume — and unlike fibre, it does not yield to a vacuum bag. On air freight this alone can cost more than the pellets.
There is also a sensory point that never appears in a spec sheet: pellets rattle. In a quiet retail environment, a 150 g pellet pad in a hard-shelled character is audible from a metre away. If the product is sold on tactile appeal in a quiet store, use a denser, smaller bead or a sewn-in weighted pad of steel-shot-filled fabric instead.
Two alternatives are worth knowing. EPS beads (expanded polystyrene) are 90% lighter and much cheaper, but they crunch, generate static, and crush permanently — they are a filler for shipping voids, not for a product. Glass microbeads at 60–120 μm give a much heavier, silkier drape per gram and are the right choice for weighted comfort plush, at four to six times the cost per kilogram and with a bag specification that must be airtight.
4. Memory Foam and Polyurethane: Slow Recovery, Narrow Use
Memory foam is polyurethane with a viscoelastic cell structure — usually TDI- or MDI-based, blown and cured into blocks, then cut. Its defining behaviour is not softness but time: press a thumb in and the dent returns over 3–5 seconds rather than instantly. That delay is what the market pays for, and it is the only reason to specify it.
Grades used in plush and comfort products sit at 40–80 kg/m³, most commonly 50–60. Below 40 the foam fatigues within a year of nightly use; above 80 it stops feeling soft and starts feeling heavy. Support is measured as IFD (indentation force deflection): 8–14 N at 25% compression for a soft pillow insert, 18–25 N for a seat cushion.
Where it earns its price: neck pillows, seat cushions, lumbar supports, weighted lap pads, and the small but growing category of anxiety and decompression plush where the product is bought for how it feels in the hand at the end of a day. In these products the foam is the product. In an ordinary character plush it is an expensive way to make a heavy toy.
| Property | What it means commercially |
|---|---|
| Cost | US$7–12/kg for virgin block versus US$1.6–2.2/kg for 7D fibre. A neck pillow insert is 4–8× the filling cost of the same volume in fibre. |
| Water | Open-cell. It absorbs, then takes 48–72 h to dry at room temperature. Never specify machine washable with a foam core; specify a removable cover instead. |
| Odour | Amine and residual-isocyanate smell, strongest in the first 72 h. Needs 24–72 h of forced-air off-gassing before bagging, or the smell arrives at retail with the product. |
| Temperature | Glass transition near 10 °C: below that it goes stiff and loses the slow-recovery effect. A container crossing a cold route in February delivers a hard block. |
| Compression | Recovers to 95%+ but slowly — 24–72 h after a 3:1 vacuum pack — so the retail unboxing experience is worse than fibre, not better. |
| Certification | Ask for CertiPUR-US (US market) or a REACH SVHC plus TDI-residue declaration (EU). Bonded “recycled PU” foam is cheaper and sheds crumbs; it should not appear in a product for children. |
One hybrid worth mentioning, because it is under-specified and over-sold: shredded memory foam. Offcuts are torn into 10–25 mm crumbs and used as loose fill, usually in pillows and “adjustable loft” products. It drapes beautifully and is genuinely comfortable. It is also impossible to keep tidy — the crumbs migrate, they pack down over months, and a percentage of every batch ends up as dust inside the cover. Use it where the cover is removable and washable; never in a sewn-closed toy.
5. Technical Comparison: Resilience, Weight, Washability, Cost
The table below compares the grades Ferpar actually specifies, at the fill weights each needs to deliver a comparable 30 cm toy body. Comparing fillings at equal mass rather than equal volume is the most common analytical error in plush sourcing — a kilogram of 3D fibre occupies far less space than a kilogram of 15D, so equal-mass comparison flatters the fine fibre and tells you nothing about the toy.
Swipe the table sideways →
| Filling | Resilience | Weight | Washability | Cost | Verdict |
|---|---|---|---|---|---|
| PP cotton, 3D fine 3 denier, solid or hollow |
Good initial loft, poor retention. 88–93% after 50% compression for 24 h. | Lightest hand feel. 12–18 g/L, needs 10% more mass for the same volume. | Machine wash 30 °C. Clumps at seams unless siliconized; 2–3 cycles before visible. | US$2.0–2.8/kg | Faces, paws, small pieces |
| PP cotton, 7D hollow Siliconized, 3D crimp |
92–96% at 24 h. Best loft-to-mass ratio in the category. | Standard. 15–22 g/L; ~70 g for a 30 cm body. | Machine wash 30 °C, low tumble with dryer balls. Recovers 90%+ over 6 cycles. | US$1.6–2.2/kg | The default grade |
| PP cotton, 15D Coarse, hollow |
96–98%. The springiest fibre, and the most resistant to vacuum set. | Heaviest fibre fill. 18–26 g/L but holds volume at low mass. | Machine wash 30 °C. Maintains structure; the fibre itself feels crisp through fabric. | US$1.5–2.0/kg | Structural, not cuddly |
| PE pellets 3–5 mm beads |
Does not compress. Shape retention is 100% because there is nothing to set. | 0.92–0.96 g/cm³ solid; 0.55–0.65 kg/L poured. 10–15× the mass of fibre per litre. | Not machine washable. Bag failure releases beads; wipe-clean only. | US$1.4–1.9/kg | Base pads only; no under-3s |
| Memory foam Viscoelastic PU, 50–80 kg/m³ |
95%+ recovery but over 3–5 seconds. Slow, deliberate, and the whole selling point. | 40–80 kg/m³. A neck-pillow insert runs 120–200 g. | Not washable. Absorbs water, dries in 48–72 h. Cover must be removable. | US$7–12/kg | Comfort use only, never bodies |
| Shredded PU foam 10–25 mm crumbs |
Good drape, progressive packing. Loses 20–30% loft over 12 months of use. | 35–60 kg/m³. Heavier than fibre, lighter than pellets. | Cover washable, fill not. Sheds dust through the shell over time. | US$3–5/kg | Pillows with removable covers |
| Glass microbeads 60–120 μm |
None needed. Flows to shape and stays there. | 1.4–1.6 kg/L. The heaviest sensible option per unit volume. | Not washable. Bag must be airtight; a leak is unrecoverable. | US$6–11/kg | Weighted plush only |
Resilience measured to ISO 1856 method B, 50% deflection for 24 h at 23 °C, recovery read at 24 h. Prices are China ex-works for virgin material, Q4 2026, before freight and duty.
5.1 Fill weight: the number that actually sets the price
Fibre is bought by the kilogram and consumed by the litre, so the specification that matters on the line is grams per litre of finished volume, not grams per toy. Measure the toy's internal volume from the pattern (or fill a sample with rice and weigh it), then apply the band for the product type:
| Product type | Fill weight | What happens below the band |
|---|---|---|
| Claw machine and carnival plush | 11–15 g/L | Loose skin at the belly; the claw grabs fabric instead of body and the toy tears |
| Mass-market retail plush | 15–22 g/L | Reads thin on shelf; shoulders collapse within a month |
| Premium and collector plush | 22–30 g/L | Misses the “dense, expensive” hand feel the price point promises |
| Shelf-display and sit-up characters | 25–35 g/L + base pad | Will not stand; the whole display plan fails at the fixture |
| Cushions and pillows | 18–26 g/L | Flattens to the seam line under body weight |
Bands assume 7D hollow siliconized fibre. Reduce by 15–20% for 15D, increase by 10–15% for 3D fine.
6. Application Scenarios
Filling decisions are made by channel, not by taste. A claw machine toy and a museum gift shop plush can share a pattern and share nothing else, because the economics of each channel reward different properties.
| Scenario | Specify this | Why |
|---|---|---|
| Claw machine and amusement High volume, low unit price |
7D hollow at 11–14 g/L, no pellets, single cavity | Must survive being grabbed and dropped hundreds of times, and must recover its shape overnight in the machine. Dense fill makes the claw slip and the fabric tear at the grab point. |
| Licensed retail IP US$18–45 retail |
7D body, 3D face and paws, 15D legs, 90 g base pad | The licensed character has to look right in a photograph and feel right in a hand. Layered fill is the cheapest route to both, and the base pad keeps it standing in the display. |
| Baby and toddler comfort 0–36 months |
3D/7D blend, hollow, siliconized, no pellets, no beads, no foam | Anything loose and small is a small-parts failure. Soft fine fibre also survives the frequent washing this category actually gets. |
| Squishy and “mochi” style Slow-rise hand feel |
3D fine, overfilled to 26–30 g/L, plus 25 g pellets per hand | The category is defined by a slow, doughy squeeze. Overfilling fine fibre with a little bottom weight delivers it at a fraction of memory-foam cost. |
| Wellness and decompression Neck pillows, lap pads |
Memory foam 50–60 kg/m³ core, removable cover, or microbead fill | Here the pressure-relief claim is the product. Foam is the only filling that supports it, and the removable cover is what makes the care label honest. |
| E-commerce and subscription Vacuum packed, air freight |
7D hollow siliconized, 5:1 compression, 48 h of relaxed packing before bagging | The unboxing is the review. Fibre that was compressed hot and immediately will arrive creased; fibre rested before compression arrives round. |
| Promotional and corporate gifting One campaign, 500–5,000 pcs |
7D at 15–18 g/L, single grade, no pellets | Cost and lead time dominate; nobody handles the toy long enough to notice layering. Spend the budget on embroidery and packaging instead. |
Two of these deserve a note. Claw machine plush is the one category where under-filling is correct, and buyers who apply retail standards to it make the product worse: a firmly stuffed toy does not compress under the claw, so the claw cannot grip, and the machine's win rate collapses. The toy is designed to be squeezed, not to look premium.
Licensed IP is the mirror image. Here the plush is a photograph before it is a toy, and the filling has to hold a silhouette through a 40-image lookbook and then a year on a shelf. Layering is not a premium flourish; it is what makes the ears stand up in the pack shot.
7. What to Write on the PO
“100% PP cotton” is not a specification. It has no denier, no finish, no fill weight and no compression requirement, which means any fibre at any weight satisfies it. Replace it with the lines below; they are written to be pasted into an RFQ unchanged.
| Spec line | Write this | What it closes |
|---|---|---|
| Fibre grade | 100% virgin polyester staple, 7 denier, hollow, siliconized, 3D spiral crimp, 51 mm | The 3D ambiguity, recycled content, and the missing silicone finish |
| Fill weight | 18 g/L ±5%, verified by weighing three units per lot | Silent under-filling on a repeat order |
| Layering | 15D in limbs, 7D in torso, 3D in head and paws; one baffle stitch at the hips | Migration and the bottom-heavy toy |
| Pellet bag, if used | PE pellets 4 mm, 110 g ±5%, double-stitched overlocked inner bag, 10 mm seam allowance | Bead release, and the washable claim that cannot be honoured |
| Compression | ISO 1856, 50% deflection, 24 h, 23 °C; recovery ≥92% at 24 h | “Soft cotton” that arrives flat after 45 days at sea |
| Hygiene | New material only, no reclaimed or shredded post-consumer fill, stored sealed before filling | The most common EN 71 and GPSR documentation failure |
| Care claim | Machine wash 30 °C only if no pellet, foam or electronic component is present | A care label the construction cannot survive |
Add one sentence to the sample approval clause: “The approved sample is the reference for filling grade, fill weight and layering; any substitution requires written approval.” On a 3,000-piece order, that line is worth more than any price negotiation you will win in the same meeting.
Frequently Asked Questions
What is the best stuffing material for custom plush?
For most orders, 7 denier hollow siliconized polyester at 15–22 g/L. It gives the best loft per kilogram, survives vacuum compression and machine washing, and costs US$1.6–2.2/kg. Move to 3 denier for faces and paws, to 15 denier for limbs and structural volume, and to PE pellets only for base weighting. There is no single best material; there is a best specification per part of the toy.
Is 3D cotton better than 7D cotton?
They do different jobs. If “3D” means 3 denier, it is finer and softer but holds shape poorly and clumps faster; it belongs in faces and small pieces. If “3D” means 3D crimp, it is usually a 7D hollow fibre with a spiral crimp, which is the better general-purpose fill. Always ask the supplier which definition they mean, and write both the denier and the crimp into the specification.
Can a plush toy with PE pellets be machine washed?
In practice, no. A pellet bag survives perhaps half of all wash cycles, and when the seam fails the beads distribute through the shell and the toy is finished. If the product must carry a machine-wash claim, keep pellets out and use a sewn-in weighted pad or a denser fill instead. Pellets also make the toy a small-parts hazard, so it cannot be labelled for children under three.
Why does my plush arrive flat after vacuum packing?
Either the fibre has poor compression-set resistance, or it was compressed while still warm and immediately after filling. A 45-day container at 60 °C under 5:1 load will set creases into any fibre that has not been rested. Specify ISO 1856 recovery above 92% at 24 h, and require 48 hours of relaxed packing at ambient before bagging. Most “flat arrival” complaints are a process problem, not a material one.
Can you send filling samples before we specify?
Yes. The material sample kit contains 3D, 7D and 15D hollow fibre, PE pellets, a memory-foam block and a microbead pouch, each labelled with density, denier and the ISO 1856 recovery figure, plus three 100 g pre-weighed fill bags so your team can feel the difference between 15 and 22 g/L in the hand. It ships free, freight collect, within three working days.