TPU Tear Strength and Flex-Fatigue: How to Prevent Cracks from Becoming Field Failures

Quick answer: High tensile strength does not guarantee resistance to a cut, notch, weld edge, sharp radius, or repeated bending. TPU tear tests and flex-fatigue tests answer different questions, and their results depend strongly on specimen geometry, thickness, direction, temperature, conditioning, and test method. Approve the real construction by connecting material screening to the actual crack starter, service strain, environment, and cycle count.

A field crack rarely starts in a perfect unnotched plaque. It starts at an edge nick, gate scar, knit line, perforation, thin-to-thick transition, weld boundary, coating interface, void, or high-strain feature. The validation plan should reproduce that weak location rather than rely on one attractive tensile number.

Mechanical test data is useful only when specimen geometry, thickness, direction, conditioning, and method are reported.

Tensile strength, tear resistance, and fatigue are different

A standard tensile test stretches a defined specimen until yield, break, or another endpoint. It is valuable for comparing stiffness, strength, and elongation under a controlled monotonic load. However, an unnotched tensile bar distributes stress differently from a product containing an existing cut or geometric stress concentration.

Tear testing introduces a defined shape or notch and measures resistance as the tear initiates or propagates. Flex-fatigue or crack-growth testing applies repeated deformation and asks whether a crack begins, grows, or reaches an unacceptable length over cycles.

These properties can rank materials differently. A TPU grade with high tensile strength may still be sensitive to a sharp notch. Another grade may show good one-time tear resistance but accumulate damage under repeated flexing at low temperature or after aging.

For this reason, do not substitute one test for another:

• tensile data does not predict crack growth from a defect;
• a single tear value does not predict millions of cycles;
• a flex test without a controlled starter crack may measure crack initiation rather than propagation;
• a material coupon does not automatically represent a welded, laminated, molded, reinforced, or printed construction.

What common tear specimen shapes reveal

ASTM D624 and ISO 34-1 include defined tear approaches and specimen geometries. Depending on the method, laboratories may use trouser, angle, crescent, or other standardized test pieces. These shapes create different stress fields, so their numeric results should not be treated as interchangeable.

Trouser tear

A slit divides the specimen into two legs that are pulled apart. This configuration can help study propagation along a defined path. Thickness, cutting quality, direction, speed, and calculation method must be controlled.

Angle or crescent tear

These geometries concentrate stress near a notch or shaped region. They can be useful for comparing resistance to tear initiation and propagation under the specified method, but they do not reproduce every part edge or service load.

Finished-part or component tear

A product-specific test may load a perforation, eyelet, weld edge, rib, gate, hole, or bonded interface. This can be more relevant than a standard plaque, but it requires a documented fixture and acceptance criterion so results are repeatable.

Never compare “tear strength” values without confirming method, specimen type, thickness, test direction, conditioning, speed, and units. Even when two reports cite the same standard family, different specimen geometries may answer different questions.

Why repeated flexing changes the failure mechanism

A flexible TPU component may survive one large deformation but fail after many smaller cycles. Repeated loading can create heat, hysteresis, stress softening, surface damage, and progressive crack growth. The cycle frequency can also change temperature and recovery time.

A useful fatigue test defines:

• strain or displacement amplitude;
• bend radius or fixture geometry;
• frequency and waveform;
• minimum and maximum temperature;
• whether a starter crack is present;
• direction relative to processing or reinforcement;
• number of cycles and inspection intervals;
• failure criterion: first crack, crack length, leakage, force loss, or complete rupture.

ASTM D813 provides one established crack-growth context for rubber under repeated flexing, but its scope and specimen should be reviewed before extending it to a specific TPU product. A footwear flex zone, coated-fabric seam, tubing bend, bellows, or 3D-printed lattice may require a different component-level method.

Generic rubber example, not TPU. Bending can reveal cracks that are difficult to see in an unloaded specimen.

Where cracks start in real TPU products

Molded parts

Common crack starters include gate scars, weld or knit lines, sharp corners, insufficient radii, sink or voids, residual stress, and damage during demolding. A visually small notch can dominate performance when it sits in the highest-strain region.

Moisture or excessive thermal history during processing can reduce molecular weight and weaken durability. Check drying, residence time, melt temperature, shear, contamination, and regrind history before blaming the nominal grade.

Extruded film, tubing, and profiles

Edge damage, die lines, gels, thickness variation, orientation, scratches, and cut quality can start a tear. For film, machine direction and transverse direction may behave differently. Test both when the product can be loaded in both directions.

Welded or laminated constructions

The crack may run through TPU, along the weld edge, inside an adhesive, at a textile interface, or through the substrate. Report the failure mode rather than only the peak force. A strong weld can shift failure into the adjacent film or fabric; that may or may not be acceptable.

Footwear and flex zones

Grooves, perforations, tread edges, bonded overlays, and upper-film transitions concentrate strain. The service test should include the actual layer stack and flex line. A plaque tear value does not capture adhesive, fabric, foam, or outsole geometry.

Lattice and printed structures

Thin struts, layer lines, node geometry, build direction, incomplete fusion, and surface defects can control fatigue. Validate the printed process and orientation, not only the feedstock resin.

Bellows concentrate repeated bending at defined folds. The pictured component is generic rubber, not a GOAFLEX TPU product.

Aging and service environment can change crack growth

Mechanical screening at room temperature is only a baseline. Add service exposures that can alter stiffness, strength, or interfaces:

• low-temperature flexing;
• hot service or heat build-up at high frequency;
• humidity or water immersion;
• hydrolysis aging where relevant;
• oils, fuels, sweat, cleaners, or plasticizers;
• UV or outdoor weathering;
• abrasion before or during flexing;
• compression or internal pressure combined with bending.

Test both during exposure and after conditioning or recovery when appropriate. A specimen can become stiff while cold, soften while hot, swell in fluid, or lose properties after aging even if it later returns to room temperature.

Do not accelerate one factor so aggressively that the mechanism no longer resembles service. The goal is not the fastest failure; it is a useful comparison tied to the expected field mode.

Diagnostic matrix: from crack location to likely investigation

Crack evidence	First areas to investigate	Useful confirmation
Starts at gate or vestige	Gate geometry, degating damage, residual stress, orientation	Sectioning, radius measurement, controlled gate redesign trial
Follows weld/knit line	Flow-front temperature, venting, contamination, fill balance	Short-shot study, microscopy, relocated gate or adjusted process trial
Begins at cut edge	Blade condition, edge notch, thickness, direction	Magnified edge review, controlled die-cut specimens
Runs next to a weld	Heat/pressure window, seam edge, local thinning, substrate	Peel/tear failure mode, seam cross-section, pressure-retention test
Appears at repeated bend	Radius, strain amplitude, temperature, frequency, starter defect	Fixed-fixture cycle test with inspection intervals
Random internal initiation	Voids, contamination, degradation, inclusions	Cross-section, microscopy, lot/process comparison
Surface network after aging	UV, oxidation, chemical attack, coating damage	Aged control, spectroscopy where justified, retained-property testing

A crack surface can preserve evidence. Protect failed samples from cleaning, cutting, or further flexing until photographed and reviewed. Mark the service orientation and crack origin.

A coupon-to-finished-part validation plan

Use a staged plan so each test answers a defined question.

Stage 1: material-level screening

Compare candidates using the same tensile, tear, hardness, and conditioning methods. Confirm specimen thickness, direction, die condition, and lot traceability.

Stage 2: defect-sensitive coupons

Introduce the notch, edge, weld, perforation, bondline, or thickness transition relevant to the application. Use controlled geometry rather than hand cuts.

Stage 3: environmental conditioning

Repeat critical tests after cold, heat, humidity, water, chemical, or UV exposure as service requires. Measure retained properties and note whether the failure mode changes.

Stage 4: cyclic component test

Run the real part in a representative fixture. Record frequency, temperature, displacement, force, cycle count, and inspection interval. Monitor heat build-up when cycling quickly.

Stage 5: production validation

Test representative lots made with production tooling, drying, regrind policy, color/additives, cutting, welding, and assembly. Set acceptance criteria before the trial.

Repeated flexing in real products combines material properties with geometry, reinforcement, joints, load, and environment.

Common mistakes

• Selecting TPU from tensile strength alone.
• Comparing tear values from different specimen shapes or thicknesses.
• Reporting a value without test direction and conditioning.
• Assuming a high one-time tear result guarantees fatigue life.
• Hand-cutting notches without controlling radius or crack length.
• Ignoring edge quality, gate scars, weld lines, voids, and residual stress.
• Running flex tests so fast that uncontrolled heat changes the result.
• Testing only virgin resin when production uses regrind, pigment, or additives.
• Inspecting only at final failure and missing the crack-growth history.
• Testing plaques when the field failure occurs in a layered or welded part.

What to send GOAFLEX for a tear or flex-fatigue review

Provide:

• failed parts plus unaffected controls;
• photographs showing crack origin and service orientation;
• part drawing, thickness, radius, gate, weld, perforation, and layer stack;
• TPU grade, hardness, lot, color/additives, and regrind level;
• molding, extrusion, cutting, welding, or lamination conditions;
• temperature, humidity, fluid, UV, abrasion, load, and cycle history;
• existing tensile, tear, flex, and aging methods with raw results;
• test direction, specimen geometry, speed, frequency, and failure criterion;
• required lifetime and consequence of failure.

GOAFLEX can help match the material screen to the product’s crack mechanism, but final approval must be based on production-intent parts and a customer-defined acceptance plan. Contact hello@goaflex.com with the failure evidence and application conditions.

Image credits

• Universal tensile testing machine: Cdang, Wikimedia Commons, public domain.
• Cracked porous rubber band: Michael Bemmerl, Wikimedia Commons, CC BY 3.0 DE. Generic rubber, not TPU.
• Rubber bellows: Venca24, Wikimedia Commons, CC BY-SA 4.0. Generic rubber, not TPU.
• Industrial conveyor belt: Getsuhas08, Wikimedia Commons, CC BY-SA 3.0. Belt material is not identified as TPU.

These search-sourced images provide testing or application context. They do not prove GOAFLEX ownership, creator endorsement, exact TPU chemistry, grade identity, or product performance.

References

• ASTM D624 — Tear Strength of Conventional Vulcanized Rubber and Thermoplastic Elastomers
• ISO 34-1 — Rubber, vulcanized or thermoplastic: determination of tear strength
• ASTM D813 — Rubber Deterioration: Crack Growth
• Instron — Guide to ASTM D624
• GOAFLEX — TPU Air Retention and Leak Testing
• GOAFLEX — TPU for Footwear
• GOAFLEX — TPU Injection Molding Problems