ASTM D1876

What the test does

Two flexible strips bonded together over most of their length are pulled apart. The short unbonded ends are bent away from each other and clamped in opposing grips, so the assembly forms a T. The crosshead separates the grips at a constant rate and the bond splits progressively along the joint, while force and head travel are recorded continuously through the separation.

What it measures, and why it matters

The reported quantity is peel resistance: the average force per unit specimen width sustained while the peel front advances steadily, taken over a fixed length of peel and excluding the initial load spike. It is a lot-release number for tapes, films and laminates, and a screening number during adhesive formulation. It also grades surface preparation, since the same adhesive on a degreased and on an untreated adherend separates by a wide margin. It ranks bonds; it is not a design allowable for a joint of different geometry.

The T-peel assembly

Adherends

Both flexibleThin sheet metal, film or laminate. If one is rigid, this geometry cannot be built and ASTM D903 applies instead.

Common width

25 mmMeasured on each specimen rather than assumed from the die — it is the divisor in the result.

Bonded length

Most of the stripWith a short unbonded length free at one end to form the arms.

Arm stiffness

Flexible enough to bend without permanent setA stiff arm turns this into a bending test and the number stops being peel.

Conditioning

About 23 °C and 50 % RH, at least 24 h

Replicate count

From the standard's own text

Record the failure mode

Every specimenDakSame reasoning as any bond test: adhesive and cohesive failure call for different corrective action.

Test speed

Head speed

254 mm/min (10 in./min), constantThe bond itself separates at 127 mm/min (5 in./min), because both arms move away from the crack front.

Excluded from the average

The initial load spike

Averaged over

A fixed length of peelNot over the whole trace, and not over a length chosen after seeing the data.

Calculations

How the test runs

1. 01Prepare both adherends with the specified surface treatment.
2. 02Bond over most of the length, leaving a short unbonded run at one end.
3. 03Cure to the adhesive's schedule and condition at least 24 hours.
4. 04Cut to width and measure the width of each specimen.
5. 05Bend the free ends away from each other to form the T.
6. 06Clamp one arm in each grip, square and without pre-tension.
7. 07Separate the grips at a constant 254 mm/min (10 in./min), recording force against head travel.
8. 08Let the peel front reach steady advance.
9. 09Average over the fixed peel length, excluding the initial spike.
10. 10Record the failure mode.
11. 11Discard any specimen where an arm took a permanent set or failed itself.

The two arms must stay flexible throughout. Once an arm yields and holds a bend, part of the measured force is going into deforming metal rather than separating adhesive — and the trace still looks like a peel curve.

The fixture this method needs

Standard 25 mm

Pneumatic Vice Action Grip

Flat vice-action faces hold each thin arm across its full width at a constant, gentle pressure — the arms are the specimen here, so a grip that marks or cuts them changes the result.

Specifications

What the report has to contain

• Reference to ASTM D1876 and the edition
• Adhesive identification, batch and cure schedule
• Both adherends and their surface preparation
• Specimen width as measured on each piece
• Conditioning and test atmosphere
• Rate of grip separation
• The peel length over which the average was taken
• Peel resistance as force per unit width
• Failure mode
• Number of specimens, mean and standard deviation
• Any specimen discarded, with the reason

What the machine must be capable of

Forces are modest. A film laminate peels at tens of newtons on a 25 mm strip and a toughened structural adhesive on thin metal arms at a few hundred, so a 100 N to 1 kN load cell covers almost all of this work, with 5 kN reserved for outliers. Capacity matching is the binding constraint rather than frame size: a 20 N peel plateau read on a 5 kN cell sits in the noise, and the fine structure of the trace is lost. The force-accuracy class required is set in the standard's own text.

The method fixes head speed at 254 mm/min (10 in./min), which separates the bond at 127 mm/min because both arms move. Peel is rate-sensitive, so the speed must be held constant, not merely averaged. Stroke matters more than it does in tension: recording the required 127 mm of peel after the initial peak needs roughly 250 to 300 mm of usable travel, and a short-stroke bench runs out mid-test.

No extensometer is used — the record is force against head travel. No dedicated peel fixture is called for either; the two bent arms go into plain opposing tension grips, mechanical vice-action or pneumatic side-action, with non-slip faces. Alignment is the critical detail: the arms must sit symmetric about the load axis. Testing is done in the same 23 °C, 50 % RH atmosphere used for conditioning.

What goes wrong in practice

Arm yield is the commonest invalidation — a metal adherend that bends permanently absorbs work that is then reported as adhesion. Asymmetric clamping runs the peel front crooked across the width, and the trace drifts downward as the effective bonded width shrinks. Stick-slip peel gives a sawtooth trace as the crack alternately arrests and runs; the mean is meaningful, single points are not. Including the initial peak in the average inflates the result, sometimes by a factor of two.

D1876 or D903

Both rank bonds and detect process drift; neither is a design allowable. Peel results depend on geometry, peel angle and adherend stiffness, so they do not transfer to a joint of different shape.

Questions we are asked about this test