ASTM D903

What the test does

Two adherends are bonded over a defined length, one rigid and one flexible. The unbonded end of the flexible member is folded back on itself through 180° and clamped in the moving grip; the rigid member is held in the opposing grip or carried on a support plate. The crosshead separates at a constant rate and strips the flexible member off the bond while force is recorded continuously against travel.

What it measures, and why it matters

The result is peel strength — force per unit width of bond, reported from the running portion of the trace rather than a single peak, alongside the observed failure mode. Adhesive suppliers use it for lot release and shelf-life monitoring, because a drifting average peel value catches a bad batch before it reaches a customer. Formulators use it to rank candidate adhesives against a named adherend pair. In failure investigations, the split between adhesive and cohesive failure tells you whether to fix surface preparation or the adhesive itself.

Specimen

The bonded assembly is built as a strip nominally 25 mm wide with a bonded length of about 152 mm, leaving enough unbonded flexible material at one end to double back and reach the grip. The flexible adherend must tolerate bending through 180° without cracking, so the pairing of materials, not just the adhesive, decides whether the geometry is usable. Specimens are cut cleanly to width; ragged or nicked edges start a tear that reads as low peel strength. Condition and test in the standard laboratory atmosphere, nominally 23 °C and 50 % relative humidity — a conditioning room or desiccator is part of the required apparatus. Discard and replace any specimen that tears through the flexible adherend, delaminates at a cut edge, or peels outside the intended bond line.

What the machine must be capable of

Forces are low. A 25 mm strip carrying a weak pressure-sensitive adhesive may peel at around a newton, while a tough rubber-to-metal bond can demand a few hundred, so a 50 N to 1 kN load cell on a single-column frame covers essentially all of this work. Resolution at the bottom of that range matters more than capacity: a cell sized for the frame rather than the specimen buries a soft peel trace in noise. The method calls for a power-driven machine of the constant-rate-of-jaw-separation type, and separation is fixed at 152.4 mm/min (6 in./min); peel is rate-sensitive, so a machine that cannot hold that rate under a fluctuating load shifts the number.

Stroke is the constraint people miss. A 180° peel advances the peel front at roughly half the crosshead rate, so stripping a 152 mm bond needs on the order of 300 mm of usable travel. No extensometer is used — peel is a force-per-width measurement, not a strain measurement.

The fixture must hold the peel angle at 180° as the front travels down the bond. A fixed pair of grips lets the rigid adherend swing as the free length changes, so the angle drifts and the recorded force drifts with it; a support plate or roller arrangement that translates with the peel front holds it. Jaw faces must grip a 25 mm strip without slipping or cutting into it.

What goes wrong in practice

Peel angle drift is the commonest error: the rigid adherend rotates during the run and the trace slopes for geometric reasons that look like a real change in the bond. Stick-slip peel produces a saw-toothed trace whose peaks are not the peel strength — average the running region, and report that the trace was unstable. Confusing adhesive-interface failure with a cohesive split misdirects the corrective action entirely, so inspect both peeled faces before recording a number.

Related and equivalent standards

ISO 8510-2 is the nearest counterpart, covering 180° peel of a flexible-to-rigid assembly; specimen width, rate and reporting conventions differ, so results do not transfer without qualification. ASTM D3330/D3330M covers peel adhesion of pressure-sensitive tape, a related geometry with its own specimen and rate conventions, and is the usual choice for tape rather than a structural adhesive bond. ISO 8510-1 covers the 90° variant, which loads the bond differently and gives different figures.