Quick answer
A foamed TPU midsole can feel soft and responsive in an initial test yet lose thickness, cushioning response, or energy return after repeated loading. The failure is rarely explained by one number. Static compression set, creep, stress relaxation, cyclic compression fatigue, heat build-up, cell damage, bond failure, and part geometry describe different mechanisms.
A useful qualification plan recreates the real load–temperature–moisture duty cycle, measures retained properties at planned checkpoints, allows a defined recovery period, and tests production-intent components as well as simple foam specimens. The goal is not to prove that a foam has a high day-one rebound value. It is to show that the finished midsole retains the required response through the intended use and environmental history.
Day-one softness is not a durability result
Foam approval often begins with density, hardness, rebound, and a short compression test. Those measurements are useful for screening, but they do not establish long-term performance.
A midsole repeatedly deforms under a changing load. The foam warms, recovers between steps, rests between wear sessions, absorbs moisture, and experiences local pressure concentrations. Its skin, internal cells, bond lines, and molded geometry may not age at the same rate. A material that performs well in a uniform laboratory block can therefore behave differently in a sculpted heel, a thin forefoot zone, or a region beside a stiff plate.
Buyers should define performance as retention:
- retained thickness after loading and recovery;
- retained stiffness or compliance over the application strain range;
- retained energy return or hysteresis response;
- temperature rise during repeated loading;
- permanent deformation after a controlled rest period;
- visible or internal cell damage;
- cracking, splitting, or bond-line failure; and
- variation among locations, cavities, lots, and finished components.
The correct targets are application-specific. GOAFLEX should not publish a universal cycle count, mileage claim, or acceptable percentage loss for every midsole.
Separate the mechanisms before choosing a test
Compression set
Compression set measures residual deformation after a specimen has been compressed for a specified time and temperature and then allowed to recover under defined conditions. It is valuable for comparing permanent deformation under a controlled static history. It does not reproduce the repeated loading, changing strain rate, or heat generation of running.
Creep and stress relaxation
Creep is increasing deformation under a sustained load. Stress relaxation is decreasing force while deformation is held constant. Both matter when a foam or surrounding construction remains compressed for an extended period, but neither is identical to fatigue under thousands of load–unload events.
Cyclic compression fatigue
Cyclic compression asks whether the load–displacement response changes as loading repeats. The foam may soften, harden, lose thickness, generate heat, accumulate residual strain, or develop localized damage. The test must define waveform, peak load or strain, frequency, rest intervals, temperature, specimen geometry, and checkpoint measurements.
Cell-wall collapse and local bottoming-out
Permanent local collapse can occur when cell walls buckle, fracture, fuse, or stop recovering. A thin or highly loaded zone may reach a compacted condition while the rest of the midsole still looks acceptable. Average thickness or average hardness can hide this failure.
Environmental aging
Heat, humidity, water, sweat, detergent, UV exposure, and storage history can alter the polymer, cell structure, adhesives, or surrounding materials. Aging may precede dynamic loading, occur during loading, or be followed by loading. The sequence should match the intended risk rather than combine every exposure without a reason.
Why density and hardness are not enough
Density is important because it reflects the amount of polymer in a foam volume, but equal-density foams can have different cell-size distributions, skin thicknesses, anisotropy, defects, and processing histories. Hardness also depends on the method, specimen thickness, temperature, time, and surface condition. It cannot describe the complete dynamic response.
A practical comparison should control or record:
- TPU chemistry and exact grade;
- foam route and blowing process;
- molded density by location, not only nominal density;
- open- or closed-cell character where relevant;
- cell-size distribution and local voids;
- skin condition and skin-to-core ratio;
- part thickness and local geometry;
- molding, cooling, conditioning, and post-treatment history;
- pigment, additive, and regrind content;
- bond layers, outsole, inserts, or plates; and
- sample age and storage conditions before testing.
If one of these variables changes, the cause of a better result may not be the resin alone.
Build the duty cycle from the real use case
The most defensible test begins with a load map, not a convenient machine setting.
Define the user and product envelope
Record the intended wearer mass range, activity, duration, cadence or loading rate, surface, temperature range, humidity, rest between uses, and expected service interval. Translate these into forces, strains, deformation limits, and cycle blocks. If that translation is uncertain, instrumented product testing can provide a better starting point than guessing.
Test local geometry
A heel wall, rocker, cutout, plate edge, flex groove, and thin forefoot zone do not share the same stress state. Use pressure mapping, section measurements, or finite-element analysis as a guide, then confirm the high-risk locations physically. A full midsole or shoe test should preserve the support and boundary conditions that create local loading.
Control temperature and rest
Viscoelastic response changes with temperature and loading rate. Repeated compression can also generate heat, so the specimen may be tested at a different condition from the chamber setpoint. Measure surface or internal temperature where practical. Define recovery intervals because a result taken immediately after cycling may differ from the result after several hours or a day.
Use staged checkpoints
Do not wait for one final endpoint. Stop at planned intervals to track trends in thickness, load–displacement curves, stiffness, hysteresis, rebound, temperature, and damage. A gradual trend, early stabilization, and sudden collapse require different engineering decisions.
What to measure at each checkpoint
A useful checkpoint record combines numbers, curves, and physical inspection.
| Measurement | What it can reveal | Important controls |
|---|---|---|
| Thickness and residual set | Uniform thinning or local permanent deformation | Location, preload, measurement time, recovery period |
| Load–displacement curve | Softening, hardening, bottoming-out, and geometry change | Rate, temperature, fixture alignment, displacement range |
| Secant stiffness or compliance | Response within the intended operating region | Use the same curve segment and preconditioning method |
| Hysteresis or returned-energy ratio | Change in energy loss and elastic response | Cycle number, rate, temperature, integration method |
| Temperature rise | Viscoelastic heat generation or insufficient recovery | Sensor location, ambient condition, airflow |
| Rebound or impact response | Change in a defined dynamic screening response | Apparatus, specimen geometry, conditioning |
| Mass and dimensions | Moisture uptake, loss, or broader dimensional drift | Conditioning and measurement resolution |
| Visual and cross-section inspection | Cracks, voids, cell collapse, skin separation, bond failure | Lighting, magnification, section location, retained controls |
Report both absolute values and retention relative to a clearly defined baseline. The baseline may be an unaged specimen, a preconditioned specimen, or the first stable cycles—not automatically the first machine stroke.
Coupon, foam block, midsole, and whole shoe answer different questions
Standardized foam specimens
Uniform specimens are efficient for material and process screening. They help compare formulations or foam conditions while controlling geometry. Standards such as ASTM D3574 provide a recognized menu of flexible-cellular-material tests, but the relevant procedures, specimen preparation, and limits still need to be selected for the project.
Molded blocks or plaques
Production-intent foam blocks can reveal density gradients, skin effects, and process repeatability. They remain simplified compared with the final sole and may miss thin sections, plate interfaces, and local stress concentrations.
Finished midsoles
Component testing captures actual geometry, molded skin, local density, and bond features. Support fixtures and load locations must be realistic and repeatable. Measure several locations rather than relying on one center-point thickness.
Whole-shoe tests
Whole-shoe testing captures the interaction among midsole, outsole, upper, plate, sockliner, and assembly process. It is closest to product function but is less useful for isolating the exact cause of a failure. Use it after material and component screening, not instead of them.
The best program forms a ladder: material screen → process-intent block → finished midsole → whole shoe → controlled wear or field validation.
A failure map for faster diagnosis
| Observed change | Plausible mechanisms | First checks |
|---|---|---|
| Uniform thickness loss | Creep, compression set, broad cell damage, insufficient recovery | Recovery-time curve, density, temperature, static versus cyclic controls |
| Local collapse | Geometry concentration, density gradient, cell defect, plate edge, bottoming-out | Section the exact zone; map pressure, thickness, and cell structure |
| Progressive softening | Cell-wall fatigue, cracking, bond loss, polymer damage | Load–displacement trend, microscopy, tear/crack inspection, aged controls |
| Progressive hardening | Cell compaction, morphology change, temperature or conditioning effect | Density and thickness by zone, thermal history, recovery behavior |
| Rising hysteresis or heat | Increased internal friction, damage, unsuitable frequency or strain | Temperature trace, test rate, curve integration, specimen alignment |
| Rebound loss with limited thickness change | Viscoelastic or polymer-level response change | Dynamic curve, temperature, aging history, formulation comparison |
| Edge split or surface crack | Tensile strain concentration, skin defect, notch, poor geometry transition | Crack origin, radius, skin condition, flex direction, tear resistance |
| Midsole–outsole or plate separation | Interface preparation, adhesive, stiffness mismatch, heat/humidity aging | Failure surface, peel/flex after aging, process timing, contamination |
This table supports investigation; it does not prove a cause. Confirm the mechanism through controlled comparisons and physical evidence.
Common qualification mistakes
- Using one rebound number as a durability claim.
- Comparing foams at different density, thickness, temperature, or recovery time.
- Calling a cyclic test “compression set” without defining the actual method.
- Reporting only the final value and losing the shape of the degradation curve.
- Testing at a convenient force that never reaches the product’s critical strain—or exceeds it unrealistically.
- Ignoring specimen self-heating during high-frequency testing.
- Measuring only the center of a part when failure is local.
- Changing resin, foam density, geometry, and process together, then assigning the result to one variable.
- Approving a laboratory block without testing the production-intent midsole.
- Making consumer mileage, injury-prevention, comfort, or running-economy claims from material tests.
A practical validation sequence
- Define the product duty cycle, critical zones, failure criteria, and allowed property change.
- Establish controlled baselines for density, dimensions, cell structure, and dynamic response.
- Screen candidate grades and foam conditions with standardized specimens.
- Compare static compression, sustained-load behavior, and cyclic compression as separate tests.
- Add heat, humidity, water, sweat, or other aging only where the use case requires it.
- Measure retained properties at planned cycle checkpoints and after defined recovery periods.
- Inspect cross-sections and failure surfaces from high-risk locations.
- Test production-intent midsoles with representative geometry, skin, inserts, plates, and bond layers.
- Confirm whole-shoe response and controlled wear performance.
- Repeat across cavities, lots, process shifts, and meaningful material changes before release.
What to send GOAFLEX for a midsole durability review
To structure a useful material and sample review, send:
- intended product, wearer/load range, and duty cycle;
- current TPU grade, foam route, density, hardness, and processing history;
- midsole drawings, local thicknesses, high-load zones, and sectioned samples;
- outsole, plate, insert, adhesive, and skin-layer details;
- baseline and aged load–displacement curves where available;
- compression-set, cyclic-compression, rebound, hysteresis, and temperature data with full methods;
- checkpoint values and recovery times, not only a final pass/fail;
- photos or cross-sections of thinning, cracks, collapse, or delamination;
- heat, humidity, water, sweat, detergent, UV, and storage exposures;
- cavity, lot, and process records; and
- the exact retention targets and acceptance criteria.
GOAFLEX can help compare candidate TPU systems and build a staged validation plan around the actual foam process and finished geometry. Final approval should remain tied to the exact formulation, molded construction, production route, and application-specific tests. Contact hello@goaflex.com with samples and test records.
Sources and further reading
- ASTM D3574 — Standard Test Methods for Flexible Cellular Materials. Use the applicable edition and procedures selected with the laboratory; the standard is a test menu, not a universal footwear specification.
- ASTM D395 — Standard Test Methods for Rubber Property—Compression Set. Useful for explaining controlled compression-set concepts; suitability and specimen details must be defined for the TPU foam project.
- Bianchi et al. — Mechanical behaviour of conventional and negative Poisson’s ratio polyurethane foams under compressive cyclic loading, International Journal of Fatigue. Supports the principle that foam architecture and repeated compression must be evaluated together; study results should not be transferred directly to a commercial midsole.
- GOAFLEX — TPU Compression Set, Creep, and Stress Relaxation.
- GOAFLEX — High-Rebound TPU and Elastic Recovery.
- GOAFLEX — Supercritical Foaming TPU for Footwear.
- GOAFLEX — TPU Tear Strength and Flex-Fatigue.
Image credits and contextual caution
- Image 1 — “Black sneaker on sandy ground top view,” A S M Jobaer, Wikimedia Commons, CC BY-SA 4.0. Cropped for the featured image.
- Image 2 — “Solid polymeric foam,” Eva Santini, Giovanna Canu, Michela E. Pedretti, CNR-ICMATE, Wikimedia Commons, CC BY 4.0.
- Image 3 — “Compression test—Universal testing machine,” Cjp24, Wikimedia Commons, CC BY-SA 3.0.
- Image 4 — “Close-up of a determined athlete running on a treadmill,” Shixart1985, Wikimedia Commons, CC BY 2.0.
These images are contextual. They do not prove GOAFLEX ownership, supplier relationships, endorsement, exact TPU chemistry, foam process, grade identity, or midsole performance.
