top of page

ISO 12135
Quasistatic Fracture Toughness of Metallic Materials

Universal Grip ISO 12135 fracture toughness bend fixture (senb)

Downloads: ISO 12135 Fracture Toughness Bend Fixture (SENB) – Universal Grip Data Sheet (PDF)

Written by Chris Thomas, Universal Grip Company · Reviewed by Frank Bertini · Updated October 24, 2026

ISO 12135 is the international standard for measuring the quasistatic fracture toughness of metallic materials. It is the unified method: one document gives the plane-strain stress intensity factor K (and the KIc qualification), the J-integral, the crack tip opening displacement δ (CTOD) and the resistance curves that describe how toughness rises as a crack tears. A specimen with a machined notch is fatigue-precracked to a sharp crack, loaded slowly in three-point bending or as a compact tension coupon while force and crack-mouth opening are recorded, then broken open so the crack can be measured. The result is the number a fracture-mechanics engineer needs to decide whether a flaw in a pressure vessel, pipeline girth weld, bridge girder or offshore jacket is acceptable under BS 7910 or API 579.

This guide covers the current edition, ISO 12135:2021 (corrected 2022), and what changed from 2016; the single-edge notched bend (SENB) and compact tension (CT) specimen proportions with their crack-length windows; fatigue precracking and the nine-point crack measurement; the loading rate and temperature control; the K, J and CTOD formulas with a checked worked example on a 100 mm wide bend specimen; the validity rules; the multi-specimen and single-specimen resistance-curve methods; a numeric comparison with ASTM E399, ASTM E1820 and the withdrawn BS 7448-1; and the bend fixture, clip gauge, precracking machine and load frames Universal Grip supplies. The ISO standards hub lists every ISO method we equip and the metals testing page gives the lab overview.

In this guide: What ISO 12135 Measures · Editions: 2002, 2016, 2021 and the 2022 Corrigendum · SENB and CT Specimens: Proportions, Crack Length and Notching · Fatigue Precracking and Crack Measurement · Apparatus: Bend Fixture, Clip Gauge, Rate and Temperature · Step-by-Step Procedure · Formulas: K, J and CTOD with a Worked Example · Validity, Pop-ins and the Size Criteria · Resistance Curves: Multi-Specimen and Single-Specimen Methods · Common Errors and How Big They Are · ISO 12135 vs ASTM E1820, ASTM E399 and BS 7448 · Equipment for ISO 12135 from Universal Grip · FAQ

What ISO 12135 Measures

The standard applies to homogeneous metallic materials loaded quasistatically, that is slowly enough that inertia and strain-rate effects can be ignored. Weldments, where the crack sits in weld metal or a heat-affected zone, are covered by the complementary ISO 15653, which uses the same specimens and formulas but adds rules for residual stress and notch placement. Crack arrest and dynamic (impact) fracture are outside the scope. Three families of result are defined:

  • K (stress intensity factor, MPa√m): the linear-elastic parameter. A result qualifies as the plane-strain fracture toughness KIc only if the specimen is thick enough and the force-displacement record is nearly linear, which in practice limits KIc to high-strength, low-toughness alloys. Otherwise the provisional value KQ is reported with the specimen size in brackets.

  • J (J-integral, kJ/m² or N/mm): the elastic-plastic energy parameter, split into an elastic part computed from K and a plastic part computed from the area under the force-displacement curve. Reported as Jc (unstable fracture with little tearing), Ju (unstable fracture after tearing), Jm (maximum force plateau) or as an initiation value from an R-curve.

  • δ (CTOD, mm): the crack tip opening displacement, the opening at the original crack tip when the specimen fails, calculated from the clip-gauge record on a rigid-rotation model. Reported as δc, δu, δm or as the engineering initiation δ0.2BL. CTOD is the parameter most weld and pipeline specifications call for, with acceptance values of the order of 0.10 to 0.30 mm.

  • Resistance curves (δ-Δa and J-Δa): toughness plotted against stable crack extension, from which the initiation values δ0.2BL and J0.2BL at the 0.2 mm offset blunting line, or the true initiation δi and Ji from the stretch-zone width measured in a scanning electron microscope, are read.

The subscript tells the reader how the specimen behaved: c means the crack went unstable (cleavage or fast fracture) after less than 0.2 mm of stable tearing, u means it went unstable after 0.2 mm or more of tearing, and m means the test reached a maximum-force plateau without instability. A pop-in, a small burst of crack extension that arrests, is evaluated under its own annex and can end the test early if it is significant. The result is always reported with the specimen thickness, for example δu(50) = 0.21 mm, because toughness depends on constraint and a 25 mm specimen is not the same test as a 50 mm one.

Editions: 2002, 2016, 2021 and the 2022 Corrigendum

ISO 12135 was first published in 2002 as the international successor to BS 7448-1:1991, which it followed closely, including the rigid-rotation CTOD formula with a plastic rotation factor of 0.4 for bend specimens. The 2016 second edition switched the single-point CTOD to a J-based form and added single-specimen R-curve guidance. The 2021 third edition, the current one, reversed part of that and made the following changes, as listed in its foreword:

  • CTOD back on a rigid-rotation basis throughout. The formulas that derived CTOD from J in the R-curve clauses were removed, and the bend-specimen CTOD now includes the yield-to-tensile ratio Rp0.2/Rm, so the plastic rotation factor is no longer a fixed 0.4 but a function of the material's strain hardening.

  • J from crack-mouth opening. J may now be determined directly from the clip-gauge (CMOD) record as well as from load-line displacement, which removes the comparator bar or second transducer that load-line measurement in three-point bending needs.

  • Modification techniques for welds. Local compression and similar treatments are permitted where residual stress stops the fatigue precrack front meeting the straightness requirement.

  • A new rotation-correction formula for compact specimens.

  • Corrected version 2022-08: the tip angle of the pop-in envelope in Figure 6a changed from 60° to 30°, and formulas (35), (38), (43), (D.7) and (H.13) and Table C.3 were fixed. Labs working from an uncorrected 2021 print should check those five equations.

BS 7448-1:1991 was withdrawn in August 2021 in favour of BS ISO 12135:2021, and BS 7448-4 (R-curves) went with it; BS 7448-2 (welds) maps to ISO 15653 and BS 7448-3 (rates above 3 MPa√m/s) remains the high-rate method. ISO 12737, the older KIc-only standard, is in practice superseded by the K clauses of ISO 12135. Several widely read pages still present the 1991 BS 7448 CTOD formula as current and describe ASTM E1820 as equivalent; the three documents give three different CTOD values from the same force-displacement record, which is why the edition must be stated on every report.

SENB and CT Specimens: Proportions, Crack Length and Notching

Two geometries are allowed. The single-edge notched bend specimen (SENB, also written SE(B)) is a rectangular bar of thickness B and width W loaded in three-point bending over a span S of four times the width. The compact tension specimen (CT or C(T)) is a pin-loaded plate with the crack growing in from the loaded edge. Bend specimens are cheaper to machine, suit plate and weld testing and are the usual choice for CTOD; compact specimens use less material per unit of crack-front length and are the usual choice for R-curves in plate and forgings. The proportions in the standard's figures are:

  • Width to thickness: 1.0 ≤ W/B ≤ 4.0, with W/B = 2 preferred. The two classic SENB shapes are the square B × B section and the preferred B × 2B section; for a 50 mm plate the preferred specimen is 50 mm thick, 100 mm wide and at least 460 mm long.

  • Length: at least 4.6W for bend specimens, so that the 4W span has rollers well inside the ends; the Universal Grip fixture's example specimen of 500 × 100 × 50 mm satisfies 500 ≥ 460 mm.

  • Initial crack length a0: 0.45 ≤ a0/W ≤ 0.70 for J, CTOD and R-curves; 0.45 ≤ a0/W ≤ 0.55 when a KIc value is sought. Shallow cracks (a0/W down to 0.10) are only permitted under ISO 15653 for welds.

  • Notch: a machined slot (milled or wire-EDM) whose tip is centred within 0.005W of the specimen mid-width, followed by the fatigue precrack. The notch root is made narrow so that the fatigue crack starts cleanly and the final crack length can be met.

  • Knife edges: integral knife edges machined into the notch mouth, or attachable knife edges screwed to the face, locate the clip gauge; their height above the crack-mouth plane, z, enters the CTOD calculation (z = 0 for integral edges).

  • Side grooves: optional V-grooves cut after precracking to promote a straight crack front and raise constraint; the net thickness BN replaces B in the J and CTOD plasticity terms and the effective thickness √(B·BN) in K.

  • Compact specimens: W = 2B with the same a0/W windows, loaded through pins in a clevis, with the clip gauge at the load line or at the crack mouth.

Crack-plane orientation is coded per ISO 3785 (Annex B of the standard): L-T, T-L, L-S and so on, where the first letter is the loading direction and the second the crack-growth direction relative to the rolling or forging axis. Toughness in the T-L orientation of a rolled plate is typically well below L-T, so the code is part of the result, not a footnote. Specimens are machined with the notch plane square to the faces, the faces flat and parallel, and the loading surfaces smooth so the rollers seat evenly.

Fatigue Precracking and Crack Measurement

A machined notch is too blunt to represent a real crack, so every specimen is fatigue-cycled until a sharp crack has grown from the notch root, typically at a force ratio R of 0 to 0.1 on a servo-hydraulic or resonant machine fitted with the same three-point bend fixture or with fatigue-rated clevis grips for compact specimens. The standard limits the maximum fatigue force so that the plastic zone left by precracking is small compared with the one the test will create. For bend specimens the maximum precracking force is calculated from the thickness, the uncracked ligament b0 = W − a0, the span and the average of the yield and tensile strengths; for the 50 × 100 mm specimen at a0/W = 0.5 and a 450/550 MPa steel it works out to about 78 kN (the arithmetic is in the worked example below). The final increment of the precrack is grown at a reduced stress intensity so that the crack tip is as sharp as a service crack, and the precrack must extend far enough from the notch, of the order of 1.3 mm or 2.5 % of W, whichever governs, that the notch shape no longer influences the crack tip.

Welds and residual-stress bearing specimens often grow a bowed or asymmetric crack front. The 2021 edition explicitly permits modification techniques, most commonly local compression, in which two indenters squeeze the ligament region on both faces before precracking to relieve the residual stress, so that the straightness requirement can be met. After the test the specimen is heat-tinted or post-fatigued to mark the crack front, broken open, and the initial crack length a0 and final crack length af are each measured at nine equally spaced points across the thickness. The nine-point average weights the two surface readings at half each: a = [(a1 + a9)/2 + a2 + … + a8]/8. None of the nine readings may differ from the average by more than a stated fraction of a0, and a crack front that fails the check invalidates the test, which is why the fatigue parameters, local compression and side grooves all exist.

Apparatus: Bend Fixture, Clip Gauge, Rate and Temperature

The force train must be verified to ISO 7500-1 and the clip gauge calibrated to ISO 9513, and the fixture must do two things that an ordinary three-point bend fixture does not: it must keep the span at exactly 4W and it must let the support rollers roll. In fracture testing the specimen rotates through a large angle as the ligament yields; if the supports cannot roll, friction adds a horizontal force that both inflates the recorded force and corrupts the plastic work Up that the J and CTOD formulas use. The standard's fixture therefore has cylindrical support rollers that are held against stops by springs so that the span is correct at the start of the test, but which are free to roll outward on the base as the specimen deflects. The loading roller is likewise free to rotate, centred over the notch, and all three rollers are hardened. Roller diameter is chosen in proportion to the specimen width so that contact stresses stay elastic in high-strength steels.

Crack-mouth opening displacement V is measured with a clip gauge, a double-cantilever strain-gauged transducer that springs onto the knife edges. Its travel must cover the expected plastic opening of a tough steel, several millimetres for a 50 mm specimen, with class 1 accuracy at the small displacements that define the elastic slope. For J by load-line displacement a second transducer or comparator bar measures the displacement of the loading roller relative to the supports; since 2021 J may instead be derived from the clip-gauge record, which is simpler and is how most modern software works. Loading is quasistatic, applied under displacement control at a rate that keeps the stress intensity rising within the standard's window, of the order of 0.5 to 3 MPa√m/s during the elastic part of the test, which for a 100 mm wide steel specimen means a crosshead speed of a fraction of a millimetre per minute. Tests at other than room temperature, and most CTOD tests on structural steel are run at −10 to −60 °C, use an environmental chamber or a liquid bath with a thermocouple on the specimen near the crack; the specimen is soaked until the temperature is uniform through the thickness and held within about ±2 °C throughout, and the clip gauge is compensated or calibrated at the test temperature.

Step-by-Step Procedure

  • Machine the specimen to the chosen W/B (2 preferred), length ≥ 4.6W, notch tip centred within 0.005W, with integral or attached knife edges; record the orientation code.

  • Measure B, W and the notch length; calculate the maximum fatigue force from B, b0, S and the yield and tensile strengths at the precracking temperature.

  • Fatigue-precrack to the target a0/W (0.45–0.70, or 0.45–0.55 for KIc), reducing the stress intensity for the final increment; apply local compression first if the material carries residual stress. Cut side grooves afterwards if used.

  • Mount the specimen on the fixture with the span set to 4W on the graduated beam, the rollers against their spring stops, and the loading roller aligned with the notch; fit the clip gauge to the knife edges and, if used, the load-line transducer.

  • Bring the specimen to test temperature, soak until uniform, and verify the thermocouple reading.

  • Load at the quasistatic rate while recording force, crack-mouth opening and load-line displacement; watch for pop-ins; continue to unstable fracture, or to the maximum-force plateau, or unload at a chosen displacement for a multi-specimen R-curve point.

  • Mark the final crack front (heat tint or post-test fatigue), break the specimen open, and measure a0 and af at nine points.

  • Check the crack-front straightness, the a0/W window, the precracking forces and the pop-in criteria; calculate K, J and δ; classify the result as c, u or m, or evaluate the R-curve; qualify against the size criteria.

  • Report the specimen geometry, orientation, temperature, rate, Rp0.2 and Rm at test temperature, the record type, any pop-ins and the result with the thickness in brackets.

Formulas: K, J and CTOD with a Worked Example

Stress intensity factor. For the bend specimen K = [F·S/(B·W^1.5)]·f(a/W), where f is the geometry function of the crack-length ratio α = a0/W; at α = 0.5 it equals 2.6625. For the compact specimen K = [F/(B·W^0.5)]·f(a/W) with f(0.5) = 9.659. With side grooves B is replaced by √(B·BN). F in newtons and dimensions in millimetres give K in MPa·mm^0.5, which is converted to MPa√m by multiplying by 0.03162. For a KIc determination the force FQ is found by the 5 % secant construction on the record and the result qualifies only if Fmax/FQ ≤ 1.10 and the thickness, crack length and ligament all exceed 2.5(KQ/Rp0.2)².

J-integral. J is the sum of an elastic term and a plastic term: the elastic part is K²(1 − ν²)/E with K evaluated at the point of interest, and the plastic part is η·Up/(BN·b0), where Up is the plastic area under the force-displacement curve and η is a geometry factor equal to 2 for the bend specimen on load-line displacement (for the compact specimen it depends on b0/W, and the 2021 edition gives separate factors for J evaluated from crack-mouth opening). CTOD. The rigid-rotation model treats the two halves of the specimen as rotating about a hinge in the ligament. δ is the sum of a small elastic term, K²(1 − ν²)/(2·Rp0.2·E), and a plastic term equal to the plastic clip-gauge opening Vp scaled by the ratio of the hinge-to-crack-tip distance to the hinge-to-knife-edge distance; the hinge position rp(W − a0) uses a rotation factor rp that in the 2021 edition depends on Rp0.2/Rm rather than the fixed 0.4 of BS 7448-1.

Worked example. A 50 mm thick, 100 mm wide bend specimen (W/B = 2) of a 450 MPa yield, 550 MPa tensile structural steel, a0/W = 0.5 so b0 = 50 mm, on a 400 mm span. Maximum fatigue precracking force: B·b0²·(Rp0.2 + Rm)/(4S) = 50 × 50² × 1,000/(4 × 400) = 125,000,000/1,600 = 78,125 N ≈ 78 kN. The plastic limit load that the ligament can carry in bending is about 1.5·B·b0²·Rp0.2/S = 1.5 × 50 × 2,500 × 450/400 ≈ 211 kN, so a 300 kN fixture rating leaves roughly 40 % margin. If the test record reaches 200 kN: K = 200,000 × 400/(50 × 100^1.5) × 2.6625 = 80,000,000/50,000 × 2.6625 = 1,600 × 2.6625 = 4,260 MPa·mm^0.5 = 134.7 MPa√m. The elastic part of J at that point, with E = 207 GPa and ν = 0.3, is (134.7 × 10⁶)² × 0.91/207 × 10⁹ ≈ 80 kJ/m², before the plastic area is added. The KIc size check 2.5(134.7/450)² = 0.224 m = 224 mm exceeds the 50 mm thickness many times over, so this steel cannot give a valid KIc at this size; the test is reported as δ or J, which is the normal outcome for structural steels at service temperatures.

Validity, Pop-ins and the Size Criteria

Every result passes through a list of checks before it can carry the standard's name: the crack-length window, the crack-front straightness from the nine readings, the precracking force and stress-intensity limits, the loading rate, the temperature record, and, for KIc, the linearity and thickness conditions above. For J and CTOD the specimen must be large enough that the crack tip is under the constraint the formulas assume; the criteria scale the thickness and ligament against J/Rp0.2, and ISO's factor of 25 is stricter than ASTM E1820's factor of 10, so a J value can qualify under E1820 and fall outside the ISO limit on the same specimen. Pop-ins are assessed under the standard's annex: a small, arrested burst of crack extension is significant if the force drop and displacement step exceed the envelope drawn on the record (the 2022 corrigendum changed that envelope's tip angle from 60° to 30°), in which case the toughness is evaluated at the pop-in rather than at the later maximum. Roller slip, clip-gauge seating, ice cracking in cold baths and electrical noise all produce steps that look like pop-ins and must be ruled out before a specimen is downgraded.

Resistance Curves: Multi-Specimen and Single-Specimen Methods

A single test gives one point; a resistance curve needs toughness at several amounts of stable crack extension Δa. The multiple-specimen method loads at least six nominally identical specimens to different displacements, marks and measures each crack, and plots δ or J against Δa. The single-specimen methods, unloading compliance (periodic partial unloads whose slope gives the current crack length) and electrical potential drop, are permitted provided the lab demonstrates sufficient accuracy, and are what most automated fracture software uses. The points are fitted with the standard's curve δ or J = α + βΔa^γ, and the initiation value δ0.2BL or J0.2BL is read where the fit crosses the blunting line offset by 0.2 mm. ISO's blunting line for CTOD has a slope of 1.87(Rm/Rp0.2), so for the 450/550 MPa steel above it is 2.29, whereas ASTM E1820 uses a fixed 1.4; the data-exclusion rules also differ (ISO requires at least one point between 0.1 and 0.3 mm and at least two between 0.1 and 0.5 mm of crack extension, ASTM at least one between 0.15 and 0.5 mm and one between 0.5 and 1.5 mm), and ASTM fits a power law J = C1·Δa^C2 rather than the three-parameter ISO curve. The same specimens therefore give different initiation values under the two standards, and a report that mixes them is wrong even when the raw data are identical. True initiation δi or Ji, independent of the blunting-line construction, is found by measuring the stretch-zone width on the fracture surface in a scanning electron microscope under the standard's Annex A.

Common Errors and How Big They Are

  • Wrong CTOD formula for the edition. The 1991 BS 7448 rigid-rotation formula with rp = 0.4, the 2016 J-based formula and the 2021 yield-ratio form give values that differ by tens of percent on the same record, and ASTM E1820's δ = J/(m·σY) differs again; literature comparisons show 15 to 60 % spreads and up to several-fold on high yield-ratio aluminium.

  • Fixed support rollers. Friction at supports that cannot roll adds to the measured force and to the plastic area, inflating J and δ by several percent in tough steels; spring-retained free rollers are the cure.

  • Span not 4W. A span error of 1 % changes K by 1 % directly and shifts the η and rotation-factor assumptions; set the span on a graduated base and check it with a rule before each test.

  • Crack outside 0.45–0.70 W or a bowed front. Either invalidates the test outright; most such failures trace to precracking at too high a force or to residual stress in welds without local compression.

  • Vp read at the wrong point. The plastic component of clip-gauge opening must be taken by drawing a line parallel to the elastic loading slope from the point of evaluation; using total opening overstates δ by the elastic part, which is small for tough steels but dominant for high-strength ones.

  • Unrecognised pop-in. Evaluating at maximum force when a significant pop-in occurred earlier reports a toughness the structure never had.

  • Clip gauge or thermocouple not at test temperature. Gauge sensitivity drifts with temperature; a gauge calibrated warm and used at −40 °C mis-reads the opening unless compensated.

  • Reporting δ without the thickness. δu = 0.21 mm means nothing to an assessor without the (50) that tells them the constraint; the standard's notation is δu(50) = 0.21 mm.

ISO 12135 vs ASTM E1820, ASTM E399 and BS 7448

The three families of fracture standards share specimens and the K function but differ in rates, size criteria, CTOD definition and R-curve construction. The figures below are the ones that cause reports to disagree; our ASTM E399 guide covers the ASTM KIc method in full.

  • Scope: ISO 12135 gives K, J, δ and R-curves in one method; ASTM E399 is KIc only; ASTM E1820 gives K, J, δ and R-curves (absorbing the withdrawn E1290); BS 7448-1 (K, J, δ) and -4 (R-curves) were withdrawn in 2021 in favour of ISO 12135.

  • Crack length: ISO a0/W 0.45–0.70 (0.45–0.55 for KIc); E399 0.45–0.55; E1820 0.45–0.70; welds under ISO 15653 down to 0.10.

  • Loading rate: ISO quasistatic within a dK/dt window of the order of 0.5–3 MPa√m/s; E399 0.55–2.75 MPa√m/s (30–150 ksi√in/min); E1820 quasistatic with a time-to-fracture guideline; BS 7448-3 covers rates above 3 MPa√m/s.

  • KIc qualification: both ISO and E399 require Fmax/FQ ≤ 1.10 and B, a0, b0 ≥ 2.5(KQ/σys)².

  • J size criterion: ISO b0 and B ≥ 25·J/Rp0.2; E1820 ≥ 10·J/σY, so ISO is 2.5 times stricter.

  • CTOD: ISO 2021 rigid rotation with rp a function of Rp0.2/Rm; E1820 δ = J/(m·σY) with m a cubic in σYS/σTS (coefficients 3.62, 4.21, 4.33, 2.00); BS 7448-1 rigid rotation with rp = 0.4.

  • Blunting line: ISO δ = 1.87(Rm/Rp0.2)·Δa; E1820 δ = 1.4·Δa. Exclusion lines ISO 0.1/0.3/0.5 mm, E1820 0.15/0.5/1.5 mm. Fit ISO α + βΔa^γ, E1820 power law.

  • Side grooves: permitted in both; ISO limits the total reduction more tightly than E1820's 25 % of B, so a specimen grooved to the ASTM limit may fail the ISO check.

  • Multi-specimen R-curve: ISO at least six specimens; E1820 is written around the single-specimen compliance method with multi-specimen as an alternative.

Equipment for ISO 12135 from Universal Grip

Fracture toughness bend fixture (part UGC-ISO12135-BF). A heavy-duty three-point bend fixture for single-edge notched bend specimens: support rollers held against stops by springs sit on a graduated base beam set to four times the specimen width, and a central roller loads the fatigue-precracked specimen. The spring-retained support rollers are free to roll, limiting friction effects; the graduated base beam gives accurate span setting; the fixture is shown with a 400 mm span for 100 mm wide SENB specimens; it is rated 300 kN per drawing with an M64×2 machine thread and is supplied with adapters to suit your load frame (clevis or threaded). Construction is a high-strength steel base with hardened cylindrical loading and support rollers, for electromechanical and servo-hydraulic universal test machines, with other capacities on request. The worked example above shows why 300 kN is the right rating for a 50 × 100 mm steel specimen: about 78 kN for precracking, 211 kN at the ligament's limit load, 200 kN at 135 MPa√m. When ordering, specify specimen width and thickness, span, load capacity, clip gauge requirements and the test machine connection. Our general bend testing fixtures and the bend testing application page cover the standard flexure fixtures for ISO 7438 and ISO 178, which are not substitutes: their supports do not roll and their spans are not graduated for 4W.

Clip gauge and displacement. A crack-mouth opening clip gauge from our clip-on extensometer range, specified by knife-edge spacing and the travel the expected Vp needs, with low-temperature versions for chamber and bath work; a video extensometer can add a second, non-contact measurement of load-line displacement or rotation. Precracking. Fatigue precracking runs on a servo-hydraulic or resonant frame using the same bend fixture or, for compact specimens, fatigue-rated clevis grips with pins sized to the specimen holes. Load frames. The 211 kN limit load of the example specimen needs a frame with margin: the 200 kN floor-standing UTM for specimens up to about 40 mm thick in structural steel, the 600 kN floor-standing UTM for 50 mm and thicker, and our hydraulic universal testing machines for heavy plate and high-strength alloys. Temperature. Environmental chambers to −70 °C fit around the fixture on these frames; for −100 °C and below a liquid bath around the specimen is used. The companion tensile data, Rp0.2 and Rm at the test temperature, come from an ISO 6892-1 test, so plan the tensile and fracture programmes together.

ISO 12135 FAQ

What is ISO 12135?

ISO 12135 is the international unified method for measuring the quasistatic fracture toughness of metallic materials: the stress intensity factor K (and KIc), the J-integral, the crack tip opening displacement δ (CTOD) and δ-Δa or J-Δa resistance curves, using fatigue-precracked single-edge notched bend or compact tension specimens. The current edition is ISO 12135:2021, corrected in 2022; it replaced the 2016 edition and BS 7448-1 was withdrawn in its favour in August 2021.

What specimens does ISO 12135 use?

Single-edge notched bend (SENB) bars loaded in three-point bending over a span of four times the width, and pin-loaded compact tension (CT) plates. Width to thickness W/B is between 1 and 4 with 2 preferred, bend specimens are at least 4.6W long, the notch tip is centred within 0.005W, and the fatigue crack is grown to a0/W between 0.45 and 0.70, or 0.45 to 0.55 for a KIc test.

What is the difference between K, J and CTOD?

K is the linear-elastic stress intensity factor in MPa√m, valid only when the specimen is large enough for plane-strain conditions; J is the elastic-plastic energy parameter in kJ/m², the sum of K²(1 − ν²)/E and a plastic term from the area under the force-displacement curve; CTOD δ is the crack tip opening in millimetres at failure, calculated from the clip-gauge opening on a rigid-rotation model. Structural steels at service temperatures almost always give J or δ rather than a valid KIc.

What changed in ISO 12135:2021?

CTOD returned to a rigid-rotation basis throughout, with the bend-specimen rotation factor now a function of the yield-to-tensile ratio Rp0.2/Rm instead of a fixed 0.4; J may be calculated from crack-mouth opening as well as load-line displacement; local compression and similar modification techniques are permitted for welds; and a new rotation correction was added for compact specimens. The 2022 corrected version fixed the pop-in envelope angle (60° to 30°) and five formulas.

How fast is an ISO 12135 test run?

Quasistatically, under displacement control at a rate that keeps the stress intensity rising within the standard's window, of the order of 0.5 to 3 MPa√m/s during elastic loading; for a 100 mm wide steel specimen that is a crosshead speed of a fraction of a millimetre per minute. ASTM E399 specifies 0.55 to 2.75 MPa√m/s, and rates above 3 MPa√m/s belong to the high-rate method BS 7448-3.

What is the K formula for a bend specimen?

K = [F·S/(B·W^1.5)]·f(a/W), with the geometry function f equal to 2.6625 at a/W = 0.5; force in newtons and dimensions in millimetres give MPa·mm^0.5, multiplied by 0.03162 for MPa√m. For a 50 × 100 mm specimen on a 400 mm span at 200 kN, K = 1,600 × 2.6625 = 4,260 MPa·mm^0.5 ≈ 135 MPa√m. The compact-specimen function is 9.659 at a/W = 0.5.

How much force does an ISO 12135 bend test need?

The ligament's plastic limit load is about 1.5·B·b0²·Rp0.2/S: for a 50 mm thick, 100 mm wide, a0/W = 0.5 specimen of 450 MPa steel on a 400 mm span that is about 211 kN, and precracking needs up to about 78 kN. A 300 kN fixture and a 600 kN frame suit 50 mm plate; 25 mm specimens of the same steel need about 26 kN at the limit load and run comfortably on a 200 kN frame.

Why must the support rollers roll?

As the ligament yields the specimen rotates through a large angle and slides over the supports; fixed supports generate friction that adds to the measured force and to the plastic work used in J and CTOD, inflating the result by several percent in tough steels. ISO 12135 fixtures hold the rollers against stops with springs so the 4W span is right at the start and the rollers are free to roll outward during the test.

How is the crack length measured?

After the test the crack front is marked by heat tinting or post-test fatigue, the specimen is broken open, and the initial and final crack lengths are each measured at nine equally spaced points across the thickness; the average weights the two surface readings at half: a = [(a1 + a9)/2 + a2 + … + a8]/8. A crack front whose individual readings differ from the average by more than the allowed fraction invalidates the test.

How many specimens does an R-curve need?

At least six for the multiple-specimen method, each loaded to a different displacement and broken open; single-specimen methods, unloading compliance and electrical potential drop, are permitted when the lab demonstrates sufficient accuracy. The initiation values δ0.2BL and J0.2BL are read at the 0.2 mm offset from a blunting line whose CTOD slope is 1.87(Rm/Rp0.2) in ISO and 1.4 in ASTM E1820.

What is the difference between ISO 12135 and ASTM E1820?

Both give K, J, CTOD and R-curves from the same specimens, but ISO's J size criterion is 25·J/Rp0.2 against E1820's 10·J/σY, ISO calculates CTOD by rigid rotation while E1820 uses δ = J/(m·σY), the blunting lines are 1.87(Rm/Rp0.2)·Δa versus 1.4·Δa, the exclusion lines are 0.1/0.3/0.5 mm versus 0.15/0.5/1.5 mm, and ISO fits α + βΔa^γ where ASTM fits a power law. The same data give different initiation values, so the standard must be stated.

What are typical CTOD acceptance values?

Weld and pipeline specifications commonly require δ of 0.10 to 0.30 mm at the minimum design temperature, for example 0.15 mm at −10 °C for many offshore steels, reported with the thickness, such as δu(50) = 0.21 mm. The value depends on constraint, so a 25 mm specimen and a 50 mm specimen are different tests even from the same plate.

Need a fracture toughness bend fixture, clip gauge, precracking setup or a complete system for ISO 12135? Contact Universal Grip with your specimen width and thickness, span, expected forces and machine connection and we will quote the right setup.

bottom of page