Ultrasonic Welding vs Stitching: How Non-Woven Bags Are Joined

A buyer sends over a photograph of a non-woven tote with one handle hanging off. The weld has peeled open along its whole length, like a label lifted, and the fabric either side is untouched. The bag was ordered as “heavy duty”, it passed visual inspection, and the purchase order said nothing about how the handle was to be attached. There is no claim to make, because there was no requirement to breach.

How panels are joined decides how long a bag survives, and it is the line most often left blank on a specification. Here is what ultrasonic welding does to spunbond polypropylene, what stitching does instead, where each fails, and how to write the requirement down.

Why spunbond polypropylene can be welded at all

Non-woven PP is not a fabric in the textile sense. Molten polypropylene is extruded through a spinneret into continuous filaments, laid randomly onto a belt as a web, then consolidated between a heated engraved calender roll and a smooth roll. Those calender marks are the small diamonds visible when you hold 80 gsm spunbond up to a window: the web is held by thousands of tiny thermal bonds, not by weaving.

The consequence that matters is that the whole material is thermoplastic. PP homopolymer melts around 160–165 °C, so filaments at an interface can be re-melted and left to re-solidify as one mass. That is not glue or adhesion: it is polymer chains from two layers interdiffusing and crystallising together on cooling. Done properly, the joint is not a joint but one continuous piece of plastic.

The four settings that decide whether a weld holds

A generator converts mains supply into a high-frequency signal, most commonly 20 kHz for bag work. A piezoelectric converter turns that into mechanical vibration, a booster scales it, and a horn, or sonotrode, presses it into the material against a patterned anvil. On a bag line the horn is a rotating wheel over an engraved anvil wheel, so fabric runs continuously through a nip.

The motion is vertical and very small, typically 20–60 µm peak to peak at the horn face. Friction between the fabric faces and hysteresis within the polymer generate heat in a fraction of a second, concentrated on the raised anvil points, which act as energy directors. Four variables control the result:

• Frequency. 20 kHz is the workhorse: high amplitude, high power, tolerant of heavier stacks. 35 or 40 kHz give smaller amplitude and a gentler weld, useful on thin webs and laminates.
• Amplitude. Set by booster and horn gain, this is the energy per cycle — the fastest route to a burnt-through panel, and to an unwelded one.
• Nip force. Pressure couples the horn to the fabric and holds the melt in compression while it solidifies. Too little and energy bounces off; too much and the horn crushes the fibre structure and thins the joint.
• Dwell time. On a rotary system dwell is contact arc divided by line speed, which is why a setting that worked yesterday fails once an operator raises the output rate. The weld window moved; nothing else did.

A fifth variable matters just as much: the anvil pattern. Discrete dots give a flexible seam with modest peel strength; a dense lattice or continuous embossed band gives higher shear strength but a stiffer, more brittle line. Unfused fibre between the weld points lets a seam flex and share load.

What ultrasonic welding is genuinely good at

There is no needle and no thread, so nothing unravels: a broken stitch propagates along a seam, a failed weld point stays local. The process runs in line with the bag machine, forming side seams, bottom and gusset in one automated pass at a rate no sewing room approaches — the main reason machine-made non-woven shopping bags cost what they do. The horn cuts and seals in one action, so edges do not fray, and because the fabric is never perforated there is no needle-hole leak path, which matters on laminated non-woven bags where the film is doing a barrier job.

Where ultrasonic welding lets you down

A weld is a brittle joint. Inside the welded zone the fibre architecture is destroyed and replaced by re-solidified bulk polymer with no fibre-to-fibre give. Loaded in shear, along the plane of the seam, that is fine and often stronger than the fabric next to it. Loaded in peel — two layers pulled apart at an angle — or at a concentrated point, it is markedly weaker than stitching. A handle root is exactly a peel-and-point load, which is why it is the classic failure site.

There is also a notch effect: the boundary between hard weld and soft fabric is a step change in stiffness, so tears often start at the edge of the weld rather than within it. Over-welding turns the seam glassy and hard, and the line cracks when creased. Under-welding leaves layers that separate under finger pressure while passing visual inspection. Neither is repairable — a stitched seam can be unpicked and resewn, but welded material has already been converted, so rework means scrapping the panel.

The weld window also moves with the fabric. Melt flow index, fabric weight, calender bond area, colour masterbatch and above all calcium carbonate filler all shift it. Filler does not melt; it absorbs energy and seeds weak spots, so heavily filled low-cost fabric welds inconsistently, and recycled content narrows the window further. Heavy screen ink at a weld line acts as a barrier and prevents interdiffusion, so keep artwork clear of the seam by 8–10 mm. On laminated stock the BOPP film shrinks and blisters at temperatures the substrate tolerates comfortably, and gloss film shows scorch that matt film hides.

What stitching buys you, and what it costs

Stitching anchors thread through the fabric at many discrete points, so load distributes across all of them, the fabric between them shares it, and failure is progressive — a stitch at a time, with visible warning — rather than one sudden separation. That is why peel and tear resistance are so much better. It is also adjustable: thread ticket, stitch type (301 lockstitch, 401 chainstitch, 504 overlock, 516 safety stitch), stitch density, needle size and seam allowance are independent levers, and stitching permits construction welding cannot produce — bound edges, turned hems, boxed corners, a patch behind a handle, webbing down the side wall. Above all, a sewn handle root can be worked as a box-X or bar-tack, spreading load across a rectangle rather than a line, which is why heavier non-woven tote bags almost always contain some sewing.

The cost is labour: every seam is operator time, and the sewing room becomes the constraint on output. A needle hole is also a defect you choose to create — each penetration weakens the fabric locally, and pushed too far the seam perforates like a postage stamp and tears along its own stitch line. Every hole is a water path, so sewing a laminated bag partly undoes the reason for laminating it, and a blunt or overheated needle can fuse spunbond into a run of melted holes. Add skipped stitches, thread breakage, loose bobbin tension, seam pucker, alignment that depends on the operator, and thread with its own abrasion and UV behaviour.

The hybrid build, and when it earns its cost

Most well-built machine-made bags are not purely one thing. The standard hybrid welds the body on the automatic line, then reinforces the handle roots with a sewn box-X or bar-tack through handle and panel, a welded patch behind the root, or both. Some builds add a strip running under the base and up both side walls, so handle load transfers into the base rather than into two small welds.

It earns its cost when the rated load is real rather than nominal, when the bag carries bottles or glassware, when it is meant to be kept and reused, when it will hang from a shoulder (dynamic loading repeatedly peels the root), or when the retailer runs a documented pull protocol at goods-in. It does not earn its cost on a lightweight giveaway, and we would rather say so than upsell it.

Bag type decides the method, and canvas decides it for you

D-cut bags have no handle joint at all: the handle is a die-cut aperture, so strength is the tear resistance of the fabric around the cut-out, not a seam — the answer to a weak D-cut bag is heavier fabric, not better welding. T-shirt and vest bags are similarly welded throughout, limited by the fabric at the base of the loop. The classic machine-made handle bag, where a separate loop is welded onto the mouth, is the cheapest construction and likeliest to fail at the root. Box bags carry more, and above roughly 90–100 gsm they commonly move to a hybrid or fully sewn build with a bound top edge; our fabric weight guide covers how weight and construction interact.

Cotton canvas bags settle the question outright. Canvas is woven cellulose: cotton does not melt, it chars, so ultrasonic energy only abrades and scorches without forming a bond. Every seam on a genuine canvas tote is sewn, and the strength conversation moves to fabric weight in ounces, thread ticket, stitch density and webbing. A useful test: if something is offered as canvas with welded seams, it is not cotton.

Testing the joint: pull, peel and drop

Three tests cover almost everything, and one rule of interpretation matters more than the numbers.

Handle pull. The workshop version loads the bag to rated weight, lifts it clear by both handles, holds for 60 seconds, then repeats the lift-and-lower cycle; 100 cycles is a reasonable benchmark for a reusable bag. The instrumented version clamps handle and body in opposing jaws of a tensile tester, pulls at a constant crosshead speed such as 100 mm/min, and records peak load in newtons. Draw samples across the run, not off one carton.

Seam peel. Cut a 25 mm strip across the seam, clamp each layer in opposing jaws, pull them apart and record average force per unit width in N/25 mm. Peel is the measurement most sensitive to weld setting, and the best early warning that the line has drifted. A grab test across the same seam gives the in-plane figure.

Drop. Fill to rated load and drop flat and on each bottom corner from a defined height such as 750 mm.

Now the rule: where it broke is more diagnostic than what it read. Fabric tearing with the seam intact means the joint is stronger than the material; to go higher, change fabric weight, not machine settings. A clean separation with a polished, glassy face means over-welding. A fuzzy peel with only part of the anvil pattern transferred means under-welding or poor coupling. A tear running along the weld boundary means the notch effect dominates, so reduce amplitude or widen the pattern. Handle webbing snapping while the bag is undamaged means the handle material, not the joint, is the limit. “Handle pull, both handles, 100 mm/min, n = 5, all failures in fabric” tells you something; “passed” does not.

A defect atlas, read backwards to the machine

• Hard, whitened weld points that crack when creased — over-weld: excess amplitude or nip force, line speed too low, or a worn anvil.
• Seam that opens under thumb pressure — under-weld: low amplitude, line too fast, insufficient force, or ink, dust or silicone contamination.
• Pinholes or burn-through along the weld — amplitude and pressure both too high, or a hard inclusion in the web.
• Weld wandering off the edge, pattern half missing — feed and web tension: unwinder imbalance, worn guides, panel misalignment.
• Puckered or blistered film on a laminated bag — BOPP shrinkage; the setting suits the substrate but not the film.
• Skipped stitches, loops underneath, broken thread — needle and thread selection, machine timing, tension, or a burr on the throat plate.
• Tearing along the stitch line — stitch density too high or needle too coarse for the fabric weight.
• Asymmetric handles — a cutting or alignment error; the shorter side takes disproportionate load and fails early, so it is a strength defect, not a cosmetic one.

Writing the joining method into your specification

“Good quality, strong handles” cannot be inspected against. Five lines make it enforceable:

• Body joining method — ultrasonic weld, anvil pattern, weld band width (6 mm, say) and its position relative to the cut edge; or sewn, with stitch type and seam allowance.
• Handle joining method — welded, sewn, or welded plus a sewn box-X bar-tack at each root, with thread and stitch density stated (Tex 40 spun polyester at 10 stitches per inch is a sensible default).
• Reinforcement — whether a patch is required, its material, size (40 mm × 50 mm, say) and placement, and whether the handle passes through it or sits on top.
• Rated load and acceptance by failure mode — the load, the number of lift cycles, and an explicit statement that fabric tear is acceptable while seam separation or handle detachment is a reject.
• Inspection plan — first article approval against a signed retained sample, in-process patrol with a hand-peel check on each machine, and final random inspection to ISO 2859-1, general inspection level II, AQL 2.5 major / 4.0 minor.

That last point deserves emphasis: most seam disputes are not about whether a defect exists but about whether it was major or minor, so classify handle detachment and seam separation in writing before production starts; our quality and inspection approach sets out how first article, patrol and final inspection fit together. Four things let us answer a seam question properly: what the bag will carry and how heavy, whether it is a handout or meant for repeated use, whether the retailer runs its own pull protocol, and whether artwork comes near the seam lines. Send those through our quotation form and we usually reply within one business day.

Frequently asked questions

Is an ultrasonic weld as strong as a sewn seam?

It depends on how the load arrives. In shear, along the plane of the seam, a correctly set weld is usually stronger than the fabric beside it. In peel, and at point loads such as a handle root, stitching is clearly better: it anchors through the fabric and fails progressively rather than all at once.

Can canvas bags be ultrasonically welded?

No. Cotton canvas is woven cellulose: it does not melt, so there is nothing to fuse, and ultrasonic energy only abrades and scorches it. Every seam on a cotton canvas bag is sewn. If a bag is offered as canvas with welded seams, the fabric is a synthetic.

Why do non-woven bag handles fail at the root?

The root sees peel and point loading, the weakest direction for a weld, and the boundary between hard weld and soft fabric is a stiffness step that concentrates stress. A box-X bar-tack or a reinforcement patch spreads that load over an area instead of a line.

What does a hard, glassy weld point tell me?

It is over-welded: too much amplitude, too much nip pressure or too slow a line has melted more polymer than the joint needs and destroyed the fibre structure, so the seam is brittle and cracks when flexed. Under-welding is the opposite — layers peel apart fuzzy under finger pressure.

How should the joining method be written into a specification?

State the body joining method and weld pattern, the handle method including any reinforcement patch or bar-tack, thread and stitch density if sewn, the rated load, and the acceptance criterion by failure mode — fabric tear acceptable, seam separation a reject. Add the inspection plan and the AQL.

Welded vs Stitched Construction

Choose the seam method after the bag structure and target load are defined. Neither method is automatically stronger without a production-equivalent finished-bag test.

Decision point	Ultrasonic welding	Stitching
Typical use	Efficient standard non-woven seams and handles designed for welding	Complex shapes, bindings, reinforced details and mixed components
Appearance	Clean fused seam without thread	Visible thread and stitch pattern
Key controls	Energy, pressure, speed, overlap and weld consistency	Thread, stitch density, seam allowance, tension and back-tacking
Approval test	Loaded finished bag using production material and settings	Loaded finished bag using production material and stitch standard

Compare custom non-woven bag constructions

Buyer Summary

Ultrasonic welding is efficient for many standard non-woven constructions, while stitching can support complex shapes, bindings and reinforced details. The decision should follow the bag design and validated loaded performance.

Related Sourcing Resources

• See the complete manufacturing process
• Review seam and handle quality checks
• Discuss the right construction

Frequently Asked Questions

Is stitching always stronger than ultrasonic welding?

No. Strength depends on material, seam geometry, settings, thread or weld quality and the complete bag construction.

When is ultrasonic welding commonly used?

It is common for efficient production of standard polypropylene non-woven seams and handles designed for welding.

How should seam construction be approved?

Test a production-equivalent finished bag at the intended load and inspect seams, handles, gussets and repeated-use performance.

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