How Non-Woven Bags Are Made: From Polypropylene Pellet to Packed Carton

A buyer approves a sample, signs off a Pantone reference, then opens the bulk shipment to find a bag half a shade off and stiffer in the hand. Nine times in ten, that colour and that hand-feel were fixed at a fabric mill weeks before anyone printed anything.

This is how a non-woven bag is made, from resin to marked carton, and what becomes permanent at each stage.

It starts as a pellet, not as cloth

Spunbond non-woven begins as polypropylene homopolymer pellets. Homopolymer, not copolymer: the regular chain structure holds the molecular orientation the drawing stage imposes, and that orientation is the fabric’s strength.

Grade selection turns on melt flow index (ISO 1133). Spunbond grades sit around 25–40 g/10 min. Higher MFI lowers melt viscosity, so the polymer clears spinneret capillaries a fraction of a millimetre across and spins finer filaments; too high and filaments snap during drawing, too low and the fabric comes out coarse and boardy at the same weight.

Colour is added here, which is what buyers most often misread. Pigment arrives as masterbatch — colourant concentrated in a polypropylene carrier — dosed into the extruder feed, so the colour lives inside the filament. Non-woven is body-coloured, not surface-dyed: the shade runs through the cross-section, it does not crock onto a white shirt, cut edges match the face, nothing washes off.

The constraint is the other half of that fact. You choose from a shade card rather than commissioning a colour, because an arbitrary Pantone in the fabric body needs its own masterbatch and a spinning run for your order alone. The usual answer is the nearest stock shade, with the brand colour in the print.

The spunbond line, from melt to wound roll

It all happens inside one continuous machine: no yarn, no loom, melt to wound roll in seconds. Pellets and masterbatch feed a single-screw extruder, where the polymer melts at roughly 220–250 °C, passes a screen changer and is metered by a gear pump into a spin beam, which forces it through a spinneret carrying thousands of capillaries. Quench air cools the filament curtain, then aerodynamic drawing attenuates the filaments in high-velocity air to around 18–25 µm and orients the polymer chains along the fibre axis. An under-drawn spunbond is weak however heavy it is.

The filaments land at random on a moving perforated belt with suction beneath, and at that instant the basis weight is fixed — throughput divided by belt speed and width. Nothing downstream adds it back.

The web becomes fabric at the calender: two heated steel rolls, one engraved with a raised point pattern, one smooth, pressed together near the softening point of polypropylene. Where a point presses, the filaments fuse into a weld; elsewhere they stay loose, so only a tenth to a quarter of the surface is bonded.

That pattern is a specification variable, not decoration. Larger bond area raises tensile and abrasion resistance but gives a stiffer, papery hand and worse tear, because welded rows act like perforation lines; finer dots give softer drape and better tear at the cost of fuzzing. Under-bond and fibres pull out under a fingernail; over-bond and the welds go glassy, so the fabric splits along the pattern. The web is then trimmed, slit and wound.

What “non-woven” changes about the finished bag

The name is literal: no warp, no weft, no interlacing, no yarn. Strength comes from a random filament network anchored at thermal welds, and three things follow for bag design.

It is directional despite looking random. Airflow and belt travel bias the filaments, so machine-direction tensile strength typically runs about 1.3 to 2 times the cross-direction figure. Panels are nested with the machine direction vertical, along the line of the hanging load.

Tear propagates differently. In woven cloth a tear is arrested by the next yarn; here there is none, so a crack runs between weld points with little resistance. A nick in a raw edge travels further than buyers expect, so a handle root needs a reinforcement patch or a bar-tack.

Cut edges do not fray. With no yarn ends to unravel, a hot-knife or ultrasonic cut ends sealed by its own melt, so no hemming is needed. That is much of why conversion is cheap, and why a D-cut bag, whose handle is punched out of the body, can exist at all.

Two things the roll fixes: weight and width

Basis weight in grams per square metre governs stiffness, opacity, how the bag stands and how much load the panels and welds carry. It is checked by cutting a known area and weighing it: a roll running a few grams light is invisible to the eye and obvious on a balance. Our guide to fabric weight covers choosing the figure.

Roll width is the variable buyers never think about, and it quietly moves price. The flat blank — body panels plus gusset, hem and seam allowances — has to nest into the slit width; if every repeat leaves a broad strip of offcut, you are buying fabric for the bin. A centimetre or two of change in bag width often lands the layout on a cleaner division of the roll, so if your dimensions carry tolerance, say so at enquiry.

The print decision: on the fabric, or under the film

This fork sets cost structure, artwork limits and durability at once. Our article on print methods and what drives their cost ranges wider; here is how it lands on spunbond.

Route A: print onto plain fabric. Screen printing is the workhorse — one mesh screen per colour, water-based or polyurethane ink pulled through onto the panel, then cured in a drying tunnel. It suits solid spot-colour blocks, logos and line art over few colours; cost scales with colour count rather than area, and registration is mechanical rather than optical, so tight multi-colour detail is the wrong brief. Heavy coverage stiffens the panel and can crack on a fold. Heat transfer handles fine detail over a small area but leaves a patch with a defined edge; flexography prints water-based ink on the roll before conversion. All three meet the same limit: the substrate is porous, ink sinks between filaments and reads duller than on coated paper, so approve a printed fabric drawdown, never a paper chip.

Route B: print the film, then laminate it. For laminated non-woven bags the artwork never touches the fabric. It is gravure printed in reverse on BOPP film, an engraved cylinder per colour, four-colour process plus spot colours, then bonded to the spunbond by dry lamination — adhesive coated, solvent driven off, nipped — or by a solvent-free thermal process. The image ends up sandwiched under the film: photographic quality, gradients, tight registration, ink nothing can abrade.

The commercial logic follows the tooling. Every colour needs an engraved cylinder, paid once and spread across the run, and the press consumes a run-up length of film before colour stabilises whatever the order size. So the route rewards quantity and complex artwork, and is the wrong answer for a two-colour logo on a modest run. Laminated fabric is also not practically recyclable, since film and substrate cannot be separated.

Two routes through the forming shop

The automatic bag machine. The roll unwinds, folds longitudinally to form the body, and the seams close by ultrasonic welding: a horn vibrating at around 20 kHz drives the fabric against a patterned anvil wheel, and interfacial friction fuses the polypropylene. No thread, no needles. Handles are welded on the same pass, or on a D-cut line the aperture is punched out of the body, and a cross-cut releases the bag.

It is fast and consistent, and it is where standard shapes belong: D-cut, vest styles and loop-handle non-woven shopping bags. The constraint is geometry — sizes step in the increments the tooling allows, gusset range is limited by the folding board, handle position is where the machine puts it. Too little weld amplitude gives a cold weld that peels; too much and the horn cuts the fabric.

The semi-automatic and hand line. Fabric is laid up in plies, cut to pattern by die or hot knife, and the panels joined by sewing with polyester thread, by ultrasonic welding, or both. Handles go on with a box-and-cross stitch or a reinforced weld, base boards drop in, and the bag is turned by hand.

Tooling here is a paper pattern rather than a machine change, so this route absorbs anything irregular: box bags with a separate base panel, wine carriers with dividers, cooler bags with a foil liner, added reinforcement, any construction with a zip. It carries more labour per bag and more unit-to-unit variation than machine welding, a limitation worth accepting knowingly rather than discovering. Sewn seams perforate the fabric, so bar-tacking the handle root matters.

Finishing, and where it collides with the fabric

• Base insert — a cut card or PP sheet in a base pocket, loose or stitched in, so the bottom stays flat under load.
• Four-sided forming — pressing the vertical corners so the bag stands square instead of slumping.
• Window — a die-cut aperture with clear PET or PVC film welded behind it; the cut edge is cleaner on laminated fabric.
• Zips, hook-and-loop tabs, cord channels and cord locks — zips are sewn rather than welded, because welding a zip tape reliably is harder than it looks; a drawstring channel is a folded hem closed before the cord is threaded.

Hot foil stamping behaves differently on the two substrates. Foil needs a smooth, closed surface to key onto. Plain spunbond is fibrous and porous, so foil transfers only to the raised bond points, reads as mottled and survives folding badly. On a laminated face the BOPP film gives exactly the continuous surface the process wants, and the result is crisp. Spot UV follows the same rule, so a brief calling for metallic detail or selective gloss has already chosen laminated fabric.

Where the quality gates actually sit

Inspecting only at the end tells you how much you have already lost. The gates that matter sit along the line.

1. Incoming fabric. Basis weight by cut-and-weigh under ISO 9073-1, slit width, roll-to-roll colour against the approved swatch, and a visual pass for holes, streaks, unbonded patches and oil marks.
2. Print first article. Registration, coverage and position measured from the bag mouth and side seam. Colour is judged in a light booth under a D65 standard illuminant, cross-checked under store lighting for metamerism, and quantified with a spectrophotometer as a Delta E value against the signed swatch. Around Delta E 3 is realistic on a porous substrate.
3. Forming first-off. Width × height × gusset measured flat, handle length and centres checked, and a pull test to confirm the fabric fails before the weld does.
4. In-process patrol. Dimensions, weld quality and print position re-checked at intervals, so drift is caught early.
5. Final random inspection. Sampling to ISO 2859-1 general inspection level II, commonly AQL 2.5 major / 4.0 minor, drawn from packed cartons. What counts as major is agreed in writing beforehand.
6. Cartoning. Inner count verified, carton gross weight compared against the calculated figure so a short count shows as a light box, and dimensions checked against the packing list.

Our role sits across those gates rather than at the machines: production runs through an SGS-verified facility in Wenzhou, and we fix the specification, run sampling, hold the signed reference at the line, check each first-off and inspect before shipment. Our quality policy sets out what gets checked, and our notes on vetting a non-woven bag manufacturer cover what to ask first.

Why the sample and the bulk are never identical

A pre-production sample and a production bag are made by different means: the sample is often screen printed by hand and assembled by a sample maker, while the bulk may be gravure printed, machine welded and die cut. Hand and machine registration do not give the same edge quality, and a hand-sewn handle does not feel like a welded one.

Fabric adds a second gap: masterbatch dosing and spinning conditions drift between lots, so a nominally identical shade shifts from batch to batch. Dimensional tolerance is a third — the fabric is compressible and cutting is done through a stack, so roughly a centimetre of variation on a mid-sized bag is normal rather than defective.

What controls all three is a signed reference sample in duplicate, one copy in the order file and one at the line, plus a written specification with explicit tolerances for dimensions and colour. Both sides also need to agree what the sample approves: if it was built before the bulk fabric was bought, approve shape and construction from it, but approve colour separately against a swatch cut from the production lot.

Packing, marks and the container

Non-woven bags are light and bulky, which inverts the usual freight arithmetic. Air and courier shipments price on volumetric weight, and LCL sea freight on cubic metres, so carton cube discipline matters more than saving grams.

Inner packing is a count decision: bags bundled into a poly bag or banded, the count stated on the inner label and chosen to divide cleanly for the receiving warehouse rather than for convenience on the line. Bags compress, but over-compressing laminated fabric sets creases into the film that will not fall out on the shelf.

Shipping marks should be approved as artwork before packing starts, because re-marking cartons at the port is slow and expensive. They normally carry item and colour, size, quantity per carton, carton number out of total, gross and net weight, carton dimensions, the order reference, the destination and any retailer-mandated labelling such as a GS1 barcode. Carton size matters too: a box slightly too tall to stack another tier inside the container height wastes cube on every unit shipped.

What to send with an enquiry

Quotations vary largely because suppliers are quoting different processes. A specification carrying these points can be priced properly instead of guessed at.

• Material and fabric weight in gsm, or simply what the bag has to carry.
• Finished dimensions as width × height × gusset, stating whether height includes the handle.
• Bag type: D-cut, loop handle, vest, box bag, drawstring, wine carrier, cooler.
• Print method, colour count, vector artwork, Pantone references, print size and position on each face.
• Plain or laminated, and whether foil or spot UV is wanted.
• Handle type: welded self-fabric, sewn webbing, rope with knot or barrel lock, or die-cut.
• Finishing: base board, window, zip, hook-and-loop, gusset construction, reinforcement.
• Packing, marks and any retail labelling standard you must meet.
• Destination port or delivery address, and your Incoterm.
• Whether you need material documentation. Any certification is held by the mill or converter, arranged per order, with the certificate number confirmed in writing before a claim is printed.

If the construction is not in a standard catalogue, our custom and OEM service covers pattern development and bespoke builds. Send the specification through the quotation form and we usually reply within one business day.

Frequently asked questions

Why can I not get an exact Pantone match on non-woven fabric?

The colour is spun into the fibre, not applied to the fabric. Masterbatch is dosed into the extruder before the filaments form, so a body shade is picked from the mill’s card rather than mixed to order. Put the exact brand colour in the print, where ink can be mixed.

Are non-woven bags sewn or welded?

Both, depending on the route. Standard shapes such as D-cut and loop-handle totes run on an automatic machine that welds seams and handles ultrasonically, with no thread. Irregular shapes, box bags, zips and reinforced constructions go down a semi-automatic or hand line, where panels are cut to pattern and sewn, welded, or both.

What does GSM actually control, and how is it verified?

GSM is grams per square metre, fixed on the spunbond line by polymer throughput divided by forming belt speed. It drives stiffness, opacity, tear resistance and how much load the panels carry, and is verified by cutting a known area and weighing it under ISO 9073-1.

Do I need laminated fabric for full-colour printing?

For photographic images, gradients and tight four-colour work, in practice yes. That artwork is gravure printed on BOPP film which is then bonded to the spunbond, so the image sits under the film. For solid spot colours and line art, screen printing on plain fabric is simpler and avoids cylinder engraving.

Why do the cut edges of a non-woven bag not fray?

There are no yarns to unravel. The fabric is a random web of continuous filaments fused at thermal bond points, so a cut edge simply ends and no hemming is needed. The trade-off is that a nick in a raw edge propagates further than in woven cloth, so handle roots are reinforced.

Buyer Summary

The manufacturing route changes with bag style, printing and load requirement. Use this guide to identify which production steps, tolerances and inspection records belong in a supplier brief.

Related Sourcing Resources

• Review our quality-control checkpoints
• See the custom production process

Frequently Asked Questions

What is spunbond non-woven fabric made from?

Most reusable non-woven shopping bags use polypropylene resin formed into continuous fibres and bonded into sheet material.

Are all non-woven bags sewn?

No. Many standard bags use ultrasonic welding, while complex or heavier constructions may use stitching or a combination of methods.

Which production checks matter most?

Confirm material GSM and colour, dimensions, seam and handle strength, print position, finished quantity and packing.

Prepared by the Global Bag Supply sourcing team · Updated August 8, 2026