ASTM E23
Notched Bar Charpy Impact Testing of Metals

Downloads: Standard Test Methods for Notched Bar Impact Testing of Metallic Materials - PDF

Written by Chris Thomas, Universal Grip Company · Reviewed by Frank Bertini · Updated October 17, 2026

ASTM E23 is the standard test method for notched bar impact testing of metallic materials. It covers the Charpy (simple-beam) and Izod (cantilever-beam) tests, in which a swinging pendulum breaks a small notched bar and the energy absorbed in fracture is read from the machine. The test is fast and cheap, and it is one of the most common acceptance tests for structural steel, pressure vessel plate, pipeline steel, forgings, castings and welds, because it shows how a material behaves when a crack meets a sudden load, especially at low temperature.

This guide covers the specimens and their tolerances, the Charpy and Izod machines, temperature conditioning, the capacity rule that decides whether a result is valid, machine verification, how lateral expansion and percent shear are measured, worked examples, and the common errors that produce bad data. It also explains how E23 relates to ASTM A370 and ISO 148-1, which share the same specimen.

In this guide: What ASTM E23 Measures · Charpy and Izod Specimens · The Impact Machine · Choosing Machine Capacity: The 80 % Rule · Temperature Conditioning · Test Procedure Step by Step · Lateral Expansion and Percent Shear · Worked Examples · Machine Verification · Common Errors and How to Avoid Them · ASTM E23 vs ASTM A370 vs ISO 148-1 · Reporting Checklist · Equipment for ASTM E23 · FAQ

What ASTM E23 Measures

E23 reports the absorbed energy in joules (or foot-pounds): the difference between the pendulum's energy before and after it breaks the specimen. Absorbed energy is a measure of notch toughness, the material's resistance to fracture when a stress concentration and an impact load are present together. It is not a design property like yield strength, but it is a sensitive indicator of brittleness, and it is the basis of many material specifications that require a minimum energy at a given temperature, for example 27 J at −20 °C.

Two other quantities are measured on the broken halves and are often required by purchase specifications: lateral expansion, the amount the specimen bulged sideways opposite the notch, which is a measure of ductility; and percent shear fracture appearance, the fraction of the fracture surface that failed by ductile shear rather than brittle cleavage. Running the test over a range of temperatures gives the ductile-to-brittle transition curve, which tells you the lowest temperature at which a steel can be used safely.

E23 covers metals only. Plastics are tested by ASTM D256 (Izod) and ASTM D6110 (Charpy), which use different specimens and lighter machines. Instrumented impact testing, which records the load-time trace during the break, is covered by ASTM E2298.

Charpy and Izod Specimens

The specimen is the most tightly toleranced part of the test. A notch that is too shallow, too blunt or off-center changes the energy more than most machine errors.

Charpy V-notch (Type A)

• Size: 55 mm long (+0/−2.5 mm), 10 × 10 mm square cross-section (±0.075 mm).

• Notch: 45° ± 1° V-notch, 2 mm deep, leaving an 8 mm ± 0.025 mm ligament. Root radius 0.25 ± 0.025 mm.

• Position: the notch is centered along the length within ± 1 mm and cut perpendicular to the long axis.

• Finish: 2 µm Ra on the notched face and the face opposite, 4 µm Ra on the other two faces. Adjacent faces square within 10 minutes of arc.

• Marking: identify specimens on the ends or on the face that sits against the anvils, never within 10 mm of the notch.

Charpy U-notch and keyhole (Types B and C)

The U-notch has a 1 mm radius and a 5 mm ligament; the keyhole notch is a 1 mm radius hole with a slot no wider than 1.6 mm, leaving the same 5 mm ligament. Both are less severe than the V-notch and are mostly specified for older materials and legacy specifications. Results from different notch types are not comparable.

Izod (Type D)

The Izod specimen is 75 mm long (+0/−2.5 mm), 10 × 10 mm, with the same 45° V-notch and 0.25 mm root radius located 28 mm from the clamped end. It is held vertically in a vise and struck 22 mm above the notch. Because the vise cannot be chilled or heated with the specimen, Izod testing of metals is done at room temperature; Charpy is the test for low-temperature work.

Sub-size Charpy specimens (Annex A3)

When the material is too thin for a 10 mm specimen (sheet, small-diameter tube, thin flanges), E23 allows reduced-width specimens: 10 mm deep but 7.5, 5 or 2.5 mm wide (other widths are listed in the annex), with the same notch. Energy does not scale linearly with width, so sub-size results are reported with the size and are not converted to full-size values without a specification that allows it. Sub-size specimens must sit on raised support blocks so the notch stays on the striker's centerline.

Specimen preparation

• Machine the notch after the final heat treatment, never before.

• Cut the notch with a dedicated broach or a profile grinder and check the root radius with an optical comparator; a worn broach gives a larger radius and a higher, misleading energy.

• Orient the specimen correctly (longitudinal or transverse to rolling direction, notch position) and record it using the E23 orientation code, because transverse energy in rolled plate can be half the longitudinal value.

• Deburr lightly. Do not round the edges, which changes the constraint at the notch.

The Impact Machine

A pendulum impact machine stores energy in a raised hammer, releases it through a fixed arc, and reads the energy left after the break. E23 Annex A1 sets the machine requirements, and most of them are about the anvils, the striker and alignment.

• Impact velocity: 3 to 6 m/s at the center of strike. ISO 148-1 is tighter at 5.0 to 5.5 m/s, and machines sold for both standards fall in that band.

• Charpy anvils: span 40.00 ± 0.05 mm, anvil radius 1.00 ± 0.05 mm, anvil taper 80° ± 2°. Worn anvils widen the span and lower the reading.

• Striker: 8 mm radius is the standard E23 striker; a 2 mm radius is optional and is the ISO 148-1 standard. The two give different energies on the same material, typically within a few joules below 200 J but up to about 10 % apart on very tough steels, so the striker radius is reported with every result.

• Alignment: the specimen notch must be within 0.25 mm of the midpoint between the anvils, and the striker must pass through that midpoint within 0.40 mm. Self-centering tongs set the notch automatically.

• Friction and windage: total energy loss on a free swing must be no more than 0.75 % of range, and a routine check must show 0.4 % or less; a rise of more than 10 % from the previous check means the bearings or the pointer need attention.

• Mounting: the machine is bolted to a foundation of at least 40 times the pendulum mass or a concrete floor at least 150 mm thick, and leveled to 3 in 1,000.

Universal Grip's pendulum impact testers are built to E23 and ISO 148-1 with interchangeable 2 mm and 8 mm strikers, self-centering tongs, and capacities from 150 J to 750 J; see the 150 J, 300 J, 450 J, 600 J and 750 J models.

Choosing Machine Capacity: The 80 % Rule

E23 says that an absorbed energy above 80 % of the machine's range is inaccurate and must be reported as approximate. The reason is physical: a pendulum that spends most of its energy on the specimen slows down during the fracture, and at 80 % of capacity the striker velocity falls by about 45 % while the crack is still running, so the test is no longer an impact at the specified speed. The practical rules:

• Pick a machine whose range is at least 1.25 times the highest energy you expect. A 300 J machine is valid up to 240 J; a 450 J machine to 360 J; a 750 J machine to 600 J.

• If a specimen does not break and the energy is below 80 % of range, the value may be averaged with the others.

• If the energy is above 80 % and the specimen passes between the anvils unbroken, report the value as approximate and do not average it.

• If the specimen stops the pendulum or does not pass through, report that it exceeded the machine's capacity. Never strike a specimen a second time.

• At the low end, very small energies are limited by the scale resolution (0.25 % of range) and by friction; for tough thin material use a smaller machine or sub-size specimens rather than reading 3 J on a 750 J scale.

A specimen is counted as broken if the halves separate or can be separated by hand. Most tough steels, aluminum alloys and austenitic stainless steels bend and jam rather than break at room temperature, which is why their room-temperature Charpy values are often reported as approximate on 300 J machines, and why 450 J or 750 J machines exist.

Temperature Conditioning

Most Charpy testing is done below room temperature because that is where steels become brittle. E23 section 9 sets the conditioning rules, and the time between bath and strike is the one most often violated.

• Room temperature is 20 ± 5 °C. Other test temperatures must be held within ± 1 °C.

• Liquid baths (alcohol and dry ice, chilled brine, liquid nitrogen mixtures): specimens are fully immersed under at least 25 mm of liquid for at least 5 minutes before testing.

• Gas or air chambers: at least 30 minutes.

• Transfer time: the specimen must be struck within 5 seconds of leaving the bath. A specimen at −40 °C in room air warms several degrees in that time; at 100 °C it loses about 10 °C in 5 seconds, which is why hot tests are run at a slightly higher bath temperature.

• Tongs: use self-centering tongs that are conditioned in the bath with the specimens, so they do not warm the specimen on the way to the anvils.

• Thermometers: check the bath thermometer or thermocouple against a reference every 6 months.

For Izod, the vise cannot be conditioned, so testing other than at room temperature is not practical. Our Charpy sample chiller baths hold −70 °C with an agitated alcohol bath and accept the standard conditioning rack and tongs.

Test Procedure Step by Step

• Verify the machine is within its 365-day verification window and run the daily friction and windage check (free swing with no specimen).

• Measure the specimen: ligament under the notch, width and notch position. Record the orientation code.

• Condition the specimen at the test temperature for the required time with the centering tongs in the bath.

• Raise and latch the pendulum. Zero or note the dial or encoder reading.

• Pick the specimen up with the tongs, place it against the anvils with the notch facing away from the striker, centered within 0.25 mm, and release the pendulum within 5 seconds of removing the specimen from the bath.

• Read and record the absorbed energy. Note whether the specimen broke, passed through unbroken, or stopped the pendulum.

• Recover both halves. Measure lateral expansion and, if required, percent shear.

• Repeat for the number of specimens in the specification (often three per temperature, with averaging rules set by the product specification).

Lateral Expansion and Percent Shear

Lateral expansion

When a ductile specimen breaks, the material opposite the notch is squeezed outward and the broken halves are wider than 10 mm at the fracture. Lateral expansion is measured with a gage that references the undeformed sides: take the two readings on each half (A1 and A2 on one half, A3 and A4 on the other), and report the sum of the larger reading from each half. Remove burrs first, and do not measure a half that is damaged at the protrusion. Many pressure vessel and pipeline specifications require a minimum lateral expansion, commonly 0.38 mm (15 mils), independent of the energy value.

Percent shear fracture appearance

The fracture surface has two regions: a dull, fibrous shear region and a bright, flat cleavage region. Percent shear is the shear area as a fraction of the whole fracture. E23 Annex A4 allows five methods: measure the cleavage region's length and width to 0.5 mm and read the percentage from Table A4.1; compare the surface with the standard fracture appearance chart; the same comparison under magnification; photograph and measure with a planimeter; or image analysis. For the standard 8 × 10 mm ligament, the table amounts to 100 × (1 − A × B ÷ 80), where A and B are the cleavage dimensions in millimeters. Percent shear is what defines the fracture appearance transition temperature (FATT), usually taken at 50 % shear.

Worked Examples

Energy conversion. A reading of 42 J is 42 ÷ 1.3558 = 31.0 ft·lbf. A specification minimum of 20 ft·lbf is 27.1 J. Report in the units of the specification and keep the raw reading.

Averaging with an unbroken specimen on a 300 J machine. Three results at 0 °C read 185 J (broken), 212 J (broken) and 251 J (unbroken, passed between the anvils). The 80 % limit for the machine is 240 J, so the third value is reported as approximate and is not averaged. The reported average is (185 + 212) ÷ 2 = 198.5 J with a note that one specimen exceeded 80 % of capacity. If all three are needed, repeat the test on a 450 J or 750 J machine.

Lateral expansion. Readings of A1 = 0.58 mm, A2 = 0.52 mm on one half and A3 = 0.49 mm, A4 = 0.61 mm on the other. Lateral expansion = max(0.58, 0.52) + max(0.49, 0.61) = 0.58 + 0.61 = 1.19 mm.

Percent shear. The cleavage region measures A = 4.0 mm by B = 6.5 mm. Cleavage area = 26 mm² out of 80 mm², so percent shear = 100 × (1 − 26 ÷ 80) = 67.5 %, reported as 68 %.

Transition curve. Testing three specimens at each of −60, −40, −20, 0 and 20 °C and plotting average energy and percent shear against temperature gives the ductile-to-brittle transition. The temperature at 27 J (or the specification's energy) and the temperature at 50 % shear are the two numbers most often reported.

Machine Verification

Charpy results are only valid when the machine has passed both direct verification (dimensions, friction, velocity, alignment) and indirect verification within the previous 365 days, and after any repair that could affect the energy reading. Indirect verification uses certified reference specimens from NIST: sets of five at a low energy level (about 13 to 20 J), a high level (about 88 to 136 J) and a super-high level (about 176 to 244 J), with separate lots for 2 mm and 8 mm strikers. The machine passes a level if the average of five falls within ± 1.4 J or ± 5 % of the certified value, whichever is larger. Common causes of failure are worn or chipped anvils, off-center placement from worn tongs or a viscous bath, loose bolts, a bent pendulum, and bearing friction. Izod machines require direct verification annually.

E23's own precision data, from 29 laboratories testing five specimens each, gives a repeatability of about 2.4 J at a 16 J level and 8.3 J at a 96 J level. Keep that in mind when a specification minimum is close to your average: three specimens scattered by 8 J is normal, not a machine fault.

Common Errors and How to Avoid Them

• Slow transfer from the bath: more than 5 seconds warms the specimen and raises the energy. Practice the motion and keep the tongs in the bath.

• Notch off-center: more than 0.25 mm from the anvil midpoint produces a mixed-mode break and scattered energies. Use self-centering tongs and check them for wear.

• Notch toward the striker: the notch must face away from the striker, so the striker hits the back of the specimen opposite the notch.

• Blunt notch cutter: a root radius above 0.275 mm gives optimistic energies. Check radius on a comparator every few hundred specimens.

• Reporting values above 80 % of range as valid: they are approximate by definition.

• Mixing 2 mm and 8 mm striker data: the strikers give different energies on tough materials; report the radius every time.

• Notching before heat treatment: decarburization or scale at the notch root changes the result.

• Skipping the daily friction check: dirty bearings or a dragging pointer add energy loss that looks like a tougher material.

ASTM E23 vs ASTM A370 vs ISO 148-1

The three standards use the same 10 × 10 × 55 mm V-notch specimen, and a machine built for one is usually verified for the others, but there are differences worth knowing when a specification cites one and your lab runs another.

• ASTM A370 is the mechanical testing standard for steel products. Its impact section refers to E23 for the method and adds product-specific rules on sampling location, orientation and sub-size specimens for thin products. If a steel is bought to A370, run E23 with A370's sampling rules.

• ISO 148-1 is the international Charpy standard. It uses the 2 mm striker as standard (KV2) where E23 uses 8 mm (KV8), sets impact velocity at 5.0 to 5.5 m/s instead of 3 to 6, allows ± 2 °C on test temperature instead of ± 1 °C, and permits a 10-second transfer when the temperature difference to ambient is small. Results are labeled KV2 or KV8 so they are not mixed.

• ASTM E2298 adds instrumented striker load measurement to E23 for dynamic yield and crack initiation energy.

• Fracture mechanics standards such as ASTM E399 and E1820 measure fracture toughness as a material property rather than absorbed energy, using pre-cracked specimens on a universal testing machine.

Reporting Checklist

• Specimen type (Charpy V, U, keyhole, Izod) and size, including sub-size width.

• Orientation code and location in the product.

• Test temperature and conditioning medium.

• Absorbed energy for each specimen in J (and ft·lbf if required), with unbroken, approximate or exceeded-capacity flags.

• Lateral expansion in mm and percent shear, when required.

• Striker radius (2 mm or 8 mm) and machine capacity.

• Date of the last machine verification.

Equipment for ASTM E23

• Pendulum impact testers from 150 J to 750 J with E23 and ISO 148-1 anvils, 2 mm and 8 mm strikers, digital encoder readout and safety enclosure.

• Specimen chiller baths to −70 °C, conditioning racks and self-centering tongs.

• Charpy notch broaches and V-notch cutters, plus a lateral expansion gage.

• NIST reference specimens and verification service through our partners.

• Universal testing machines and wedge grips for the tensile tests that usually accompany impact testing under ASTM E8 and A370.

The videos on this page show a Charpy test on one of our pendulum machines and a walk-through of the full E23 procedure.

ASTM E23 FAQ

What is ASTM E23?

ASTM E23 is the standard test method for notched bar impact testing of metals. A pendulum breaks a small notched specimen, and the energy absorbed in fracture is reported as a measure of notch toughness. It covers both the Charpy and Izod tests.

What are the Charpy V-notch specimen dimensions?

55 mm long, 10 × 10 mm square, with a 45° V-notch 2 mm deep and a 0.25 mm root radius centered on the length. Tolerances are tight: ± 0.075 mm on the cross-section and ± 0.025 mm on the ligament and root radius.

What is the difference between Charpy and Izod?

In the Charpy test the specimen rests horizontally on two anvils 40 mm apart and is struck behind the notch; in the Izod test the specimen is clamped vertically in a vise and struck 22 mm above the notch. Charpy is used for almost all metals work because the specimen can be chilled or heated; Izod is limited to room temperature.

Why must the specimen be struck within 5 seconds?

A small steel specimen changes temperature quickly in air. A 100 °C specimen loses about 10 °C in 5 seconds, and a cold specimen warms just as fast, so a longer delay means the test is run at the wrong temperature and the energy is wrong.

What machine capacity do I need for Charpy testing?

The absorbed energy must be no more than 80 % of the machine's range to be valid. A 300 J machine covers results up to 240 J, which suits most carbon and low-alloy steels; tough steels, aluminum and stainless at room temperature often need 450 J or 750 J.

What is lateral expansion?

The amount the specimen bulges sideways at the fracture, measured on the broken halves opposite the notch. It is a ductility measure, and specifications often require a minimum of 0.38 mm (15 mils).

What does percent shear mean?

The fraction of the fracture surface that failed by ductile shear rather than brittle cleavage. It is measured from the size of the flat cleavage area and is used to find the fracture appearance transition temperature.

What is the difference between a 2 mm and an 8 mm striker?

E23 uses an 8 mm radius striker as standard; ISO 148-1 uses 2 mm. Both are allowed under E23. They give slightly different energies on the same material, so the radius is always reported and results from the two are not mixed.

How often must an impact machine be verified?

Direct and indirect verification annually, and after any repair that affects the energy reading. Results are valid only within 365 days of the last passed indirect verification with NIST reference specimens.

What are sub-size Charpy specimens?

Reduced-width specimens (for example 7.5, 5 or 2.5 mm wide, still 10 mm deep) for material too thin for a full-size bar. The width is reported with the result, and sub-size energies are not directly comparable with full-size values.

How many Charpy specimens are tested?

E23 does not fix the number; product specifications usually require three per temperature and set their own averaging rules. Transition curves use three or more at each of several temperatures.

Is ASTM E23 the same as ISO 148-1?

They use the same specimen and the same principle, but differ in standard striker radius, impact velocity range, temperature tolerance and transfer time. A machine can be verified to both; results are labeled KV8 (E23) or KV2 (ISO).

Need a fixture, grips or a complete system for ASTM E23? Contact Universal Grip with your specimen details, capacity and machine connection and we will quote the right setup.