Testing standard
Standard Test Method for Floating Roller Peel Resistance of Adhesives
Written and technically reviewed by Dak System Inc. engineeringLast reviewed
ASTM D3167 measures the peel resistance of a bond between one rigid and one flexible adherend. The rigid member runs through a cradle of unloaded rollers while the flexible one is pulled away, so the peel angle stays constant as the front advances. The strip is peeled at a constant head speed of 152 mm/min, and the result is the running force divided by the bonded width. The method states that both adherend thicknesses change the value and must be specified.
Two adherends are bonded along their length, one rigid and one flexible. The rigid member is drawn through a cradle of rollers while the free end of the flexible member is pulled upwards in the opposite grip, so the flexible adherend bends around a small roller and strips off the bond. The rollers turn freely, carrying the rigid adherend along as the peel front advances, so the angle at which the flexible member leaves the bond stays constant all the way down the specimen. Force is recorded continuously against travel, and the running force divided by the bonded width is the result.
Peel resistance, expressed as force per unit width. Peel destroys bonded assemblies that were perfectly adequate in shear, because it concentrates the whole load on a moving line rather than spreading it over an area: a structural adhesive carrying tens of megapascals in a lap joint can be walked apart by hand at the edge. The method puts a reproducible number on that. It is used to rank adhesives and primers, to qualify surface preparation, and as a production check where a drifting peel figure is the first sign that a pre-treatment line has changed.
The standard is explicit that the thicknesses of both adherends change the value materially, and that they must therefore be stated in the material requirements. A peel figure quoted without the adherend build is not a comparable number at all. Where no thickness is specified, the method falls back to a flexible adherend of 0.63 mm and a rigid adherend of 1.63 mm.
One rigid adherend, one flexible adherend, bonded along their length.
The flexible adherend has to bend around the roller without yielding or cracking. If it yields, the number describes the adherend and not the adhesive.
P = F_avg / w
Force per unit width, not a stress. There is no area in the denominator because peel loads a line, not a plane — which is exactly why a bond that is strong in shear can be weak here.
Flat vice-action faces hold the thin flexible adherend across its full width at a constant, gentle pressure. The arm is the specimen here, so a jaw that bites or marks it moves the result. The floating roller cradle itself is made to the method.
SpecificationsLow forces, generous stroke, steady rate. A 12.7 mm strip peels at anywhere from a few newtons for a light laminating adhesive to a couple of hundred for a toughened structural film, so a load cell in the 100 N to 1 kN class resolves the trace; one sized for the frame instead of the specimen buries a soft peel curve in noise. The method asks for a frame to Practices E4, self-aligning grips, a continuous record of load against separation, and a constant head speed of 152 mm/min (6 in./min) — at which the bond separates at that same rate, unlike a T-peel, where it opens at half the head speed. Holding the rate under a fluctuating load matters, because peel is rate-sensitive and a machine that surges through a stick-slip event records the surge. Stroke, not capacity, is the practical constraint: the average is read over at least 76.2 mm (3 in.) of bond line once the first 25.4 mm (1 in.) has been discarded. No extensometer is involved: this is force per unit width, not strain.
Yielding of the flexible adherend, which turns a peel test into a bend test and gives a number governed by the metal rather than the adhesive. Stick-slip peel, whose saw-toothed trace has peaks that are not the peel resistance; average the running region and say that the trace was unstable. Reading the initial spike as the result, when the method says to discard the first 25.4 mm of peel before averaging. Quoting a value without the adherend thicknesses, which the method itself warns against. And loading the specimen crookedly, so the peel front runs diagonally and the effective width is not the width you measured.
| ASTM D3167 | ASTM D1781 | ASTM D903 | ASTM D1876 | |
|---|---|---|---|---|
| Geometry | Floating roller | Climbing drum | 180° strip | T-peel |
| Adherends | One rigid, one flexible | Facing off a core or panel | One rigid, one flexible | Both flexible |
| Peel angle | Held by the roller cradle | Held by the drum | 180°, if the fixture keeps it there | Set by the two arms |
| Severity | Greater than D1781 | Less severe | Depends on the flexible member | Depends on both arms |
D3167 names itself an alternative to D1781 and the more severe of the two, because its peel angle is greater. Two peel numbers from different geometries are not interchangeable and should never be compared directly.
It is the ASTM floating roller peel test for adhesives, used where one adherend is rigid and the other flexible. The rigid member passes through a cradle of unloaded rollers while the flexible one is pulled away, so the peel angle stays constant as the front travels. The result is the average running force divided by the bonded width.
Because the cradle is not clamped in place. It carries the rigid adherend and is free to move as the peel front advances, which is what keeps the angle at which the flexible member leaves the bond constant from one end of the specimen to the other. A fixed arrangement lets the angle drift, and the recorded force drifts with it.
The method names itself an alternative to D1781, the climbing drum peel test, and says it is the more severe of the two because the peel angle is greater. D1781 suits a facing peeled from a sandwich panel or honeycomb core; D3167 suits a flexible sheet bonded to a rigid one. Results from the two are not interchangeable.
Because the thickness of both adherends changes the measured value materially, and the method says so in its own scope. A thicker flexible member resists bending and raises the recorded force without the adhesive having changed at all. A peel figure quoted without the adherend build cannot be compared with another laboratory's.
Where the material requirement does not specify them, the method falls back to a flexible adherend of 0.63 mm (0.025 in.) and a rigid adherend of 1.63 mm (0.064 in.). Those are defaults for the case where nothing has been agreed, not a specification to be applied over one that has. The default adherend material, again where nothing else is specified, is clad aluminium alloy 2024-T3.
Because it reflects the energy needed to start separation at one point, not the resistance of the bond to being peeled. The method is specific about it: discard the first 25.4 mm (1 in.) of peel, then average the load over at least the next 76.2 mm (3 in.) of bond line. The reproducible quantity is the steady running force once the front is advancing.
Stick-slip peel: the crack front arrests, load builds, and the front runs forward again. The peaks are not the peel resistance and must not be reported as such. Average the running region and state in the report that the trace was unstable, because that behaviour is itself a finding about the adhesive.
A small one. A 12.7 mm strip peels at a few newtons for a light laminating adhesive and at a couple of hundred for a toughened structural film, so a load cell in the 100 N to 1 kN class is right — the method wants the breaking load between 15 % and 85 % of full scale. Stroke matters more than force: the fixture takes up travel, and better than 100 mm of peel has to be recorded after that.
Dak verifies against whichever standard the method names, and where a class applies our frames sit a class tighter than it asks.
| The method asks for | Dak supplies | |
|---|---|---|
| Capacity | A 12.7 mm strip peels at a few newtons for a light laminating adhesive and at a couple of hundred for a toughened structural film, so a 100 N to 1 kN load cell suits almost all work. The method wants the breaking load to fall between 15 % and 85 % of full-scale capacity, which rules out running it on a cell sized for the frame. | Load cells from 1 kg to 60 ton on the Series 7200, and 0.5 to 100 kN on the Series 9000 |
| Force accuracy | ASTM E4 | Verified to ASTM E4, and to ISO 7500-1 Class 0.5 |
| Gripping | Floating roller peel fixture — a cradle of freely turning rollers attached to one crosshead, carrying the rigid adherend, with the flexible adherend clamped in the opposing grip | Our peel and adhesion fixtures, built to the specimen |
| Environment | Not fixed by the method. Conditioning before test is one of the variables the adhesive maker states, and it is reported with the result | 3009 series chambers, −150 °C to +400 °C — temperature only |
This page describes the method as practised. The governing text is the current edition from the issuing body. Tell us what you are testing and we will answer with the machine, the fixture and a quotation.
