Charpy Impact Testing of Tool Steel

The test in one paragraph

In the Charpy test a notched bar is supported at both ends and broken by a single blow from a pendulum that strikes the middle of the specimen on the side away from the notch. The specimen breaks at the notch, the two halves fly clear, and the pendulum passes between the two parts of the anvil. The difference between the height of fall and the height of rise gives the energy absorbed in deforming and breaking the specimen, and the machine is calibrated to read that energy directly, adding back the friction and windage losses. This is three point loading. The Izod test uses the same kind of pendulum and a similar notched bar, but the specimen is held as a cantilever and struck above the notch, and the two tests do not return the same numbers.

The Charpy V-notch test is the most used and most widely specified impact test in the industry. Its limitations are known and they are real: the notch is blunt, the specimen is small, and the single energy figure mixes the initiation and the propagation of the crack. It survives because it is cheap, simple and extremely sensitive to the things that make steel brittle, and because decades of correlation with service performance have made it a usable acceptance test even where it does not predict the behaviour of a full-size part.

The specimen and the notch decide the result

There are three standard Charpy specimens: the V-notch, the U-notch and the keyhole. The V-notch bar is the primary specimen and the one used for tool steel. The other two are used in some industries, the steel casting industry among them, but the keyhole and the U-notch were recognised as early as 1945 as giving inadequate transition temperatures because the notch is blunt, and the constraint and triaxiality they generate are considerably lower than those of a V-notch bar. A transition temperature measured on a keyhole specimen is not a V-notch transition temperature.

Machining tolerances on the specimen are not a formality. The bar has to be square, and grinding opposite faces parallel does not make it square. The notch is the most critical feature on the whole specimen: its shape and size have to be held to the standard exactly, its finish has to be smooth but not polished, and it is worth examining at some magnification with a stereoscopic microscope or an optical comparator, where a V-notch template can be laid over it. V-notches are cut on dedicated broaching machines or on a milling machine with a fly cutter. A keyhole hole is drilled at low speed so that heat and work hardening are not introduced, and the slot is then cut without letting the cutter strike the back of the hole. One rule overrides all of this: all notching is done after any heat treatment, because a notch cut before hardening will not have the geometry the test assumes.

Orientation is the second variable, and on forged or rolled bar it is the one that catches people out. If nothing is stated, the specimen is oriented along the rolling direction of the plate or the forming direction of the part, with the notch perpendicular to that surface, which is the orientation that gives the maximum impact values. Cutting the specimen so that the notch lies along the fibre direction, or cutting across it, changes the result substantially. For tool steel supplied as forged or rolled bar, the direction a Charpy specimen is taken from belongs on the test certificate next to the number.

Sub-size specimens do not scale directly

Sub-size bars exist for the case where there is not enough material for a full-size specimen or the shape of the part will not allow one to be cut. They should be used only for that reason. Correlation between full-size and sub-size results exists but it is not direct, which is why the standards specify separate acceptance values for each size.

Minimum impact values by specimen size. Full-size specimens should be used wherever possible; the reduced sizes apply where there is not enough material for a full-size bar.

Test temperature is part of the result

A Charpy energy figure without a temperature is not a result. Unless something else is specified, testing is done between 21 and 32 °C (70 and 90 °F), but most Charpy testing is done below that, because the interesting behaviour of steel sits in the ductile to brittle transition, which for most steels occurs between room temperature and -46 °C (-50 °F). Testing also runs down to -196 °C (-320 °F) for cryogenic service, and lower for research work.

Temperature control is where the procedure is easiest to get wrong. Down to about -59 °C (-75 °F) the usual medium is ethyl alcohol with dry ice, which sets at about -68 °C (-90 °F). The specimens sit in an insulated container on a screen grid raised at least 25 mm above the bottom so the cooling liquid circulates under them, covered by at least 25 mm of liquid, with a calibrated thermometer or thermocouple reading the temperature near the centre of the group of specimens and the bath agitated so the temperature is uniform. The tongs used to move specimens go into the same bath, or they warm the specimen on the way to the machine. Specimens are held within +0 and -1.5 °C (+0 and -3 °F) of the test temperature for at least five minutes, then removed one at a time and broken within five seconds. Below -59 °C a liquid that does not solidify is needed, isohexane or liquid nitrogen among them. High temperature testing uses an agitated high flash point oil, with the bath and specimens held at temperature in a furnace for at least ten minutes.

When results are plotted against temperature, the transition shows up as the drop from the high energy upper shelf to the low energy lower shelf. Where the transition is sharp the transition temperature is easy to pick off the curve; where it is gradual, which is the common case, an energy level is specified instead and the material is treated as brittle below it. A value of 20 J (15 ft · lbf) is often used as that level.

Reading a Charpy value

Impact results are reported in three ways, and the standards allow more than one of them on the same specimen because they answer different questions. Energy is the figure most often specified. Percent shear and lateral expansion are taken from the broken halves after the test and describe how the specimen failed rather than how much energy it took, which is why they are the measures used when a Charpy result has to be correlated with a toughness test that uses a different specimen and a different loading rate.

The three ways a Charpy test result is reported, with the minima most often specified for steels. Compiled from the article text.

Three specimens are tested at each temperature and the average is the result. Where a minimum is specified for acceptance, not more than one of the three may fall below it, and if one specimen sits about 6 J (5 ft · lbf) below the average, the material is rejected or retested, in which case three further specimens have to be tested and all of them have to meet the minimum. The two thresholds worth carrying in your head when a toughness figure is being argued over are that Charpy energies of about 14 J (10 ft · lbf) and below are likely to initiate a fracture, and that around 27 J (20 ft · lbf) is where a brittle fracture, once started, is likely to propagate. Values well above that are needed to arrest a running crack.

What the test does not do is produce a number for design. Charpy energy is a comparative measure, and the empirical correlations that convert it to fracture toughness, K_Ic, are valid only inside a limited class of steel and a limited range of data; they also mix a blunt notch and a high loading rate into a result that is being compared with a sharp crack loaded slowly. They are useful as guides and they underpin some design codes, but a Charpy value should not be presented as a fracture toughness.

What this means for a tool steel order

A published toughness figure for a tool steel grade is measured on a full-size V-notch specimen, in a stated orientation, in a stated heat treated condition, and at a stated temperature. Change any one of those and the number moves, so a Charpy result is only comparable with another result taken the same way. The most common mismatches on tool steel are a figure quoted for the optimum tempering temperature compared against a tool that was tempered somewhere else, and a figure measured on a full-size bar compared against a sub-size result from a customer who could not cut one. When toughness is being used to choose between two grades, the comparison belongs on the same specimen size, the same orientation and the same tempering condition. The impact energies published for the common grades at their optimum tempering temperature are collected in our chart of tool steel impact toughness, and the way a toughness shortfall shows up in service is covered in impact wear and spalling of tooling.

Source: ASM Handbook, Volume 8, Mechanical Testing and Evaluation, ASM International, 2000. Article Impact Toughness Testing (PDF p1352-1392). Sub-size minimum values are Table 1 of that article.