Interpreting impact and fracture tests

Different tests measure different outcomes

A pendulum strikes a notched bar, the bar bends and separates, and the pendulum rises to a lower height than it would without the specimen. How much energy did the test consume? That is a useful question, but it is different from asking what loading will extend a sharp crack already present in a component.

A Charpy-type pendulum impact test records an absorbed energy in joules for its specified supported, notched specimen and loading arrangement. The notch starts a localized deformation and fracture process; its root has a finite machined geometry. The record can reflect work involved in deformation, crack initiation and subsequent extension. It does not isolate the energy of breaking atomic bonds alone.

A fracture-mechanics test starts with a defined crack and interprets the response through a crack-driving parameter. Many metallic tests use a sharp fatigue precrack, grown by cyclic loading beyond a starter notch. A machined notch with a rounded root is not automatically equivalent to that sharp crack. The preceding lessons explain why local field, constraint and crack-tip process zone matter.

Follow the energy from one height datum

Use an idealized pendulum energy account first. Let m be the mass used in the model and g the acceleration of gravity. Measure its mass-centre release height h₁ and post-impact return height h₂ from the same lowest-position datum. The corresponding gravitational energies are E₁ = mgh₁ and E₂ = mgh₂. Their difference, E_spent = mg(h₁ − h₂), is the energy no longer available as returned gravitational energy.

A real machine also loses energy through bearing friction, air resistance and other effects. Its defined reduction uses the appropriate correction for the actual swing. If an applicable loss E_loss has been supplied, the simple corrected balance is E_abs = E_spent − E_loss. Subtracting loss prevents machine dissipation from being assigned wholly to the specimen. A whole free-swing loss cannot simply be substituted without its required path adjustment.

The calculation assumes consistent heights and machine parameters; actual machines use calibrated angular/energy systems and method-specific corrections. The educational balance below is not a complete measurement procedure. Likewise, an instrumented force record can reveal more detail about the event, but dynamics, friction and vibration complicate the comparison between different energy reductions.

A force–displacement area gives work

If force F acts through an additional specimen displacement Δδ, the incremental work is approximately F Δδ. For straight segments between recorded points, use average force over each interval: ΔW_i = [(F_i + F_(i+1))/2](δ_(i+1) − δ_i). Summing these trapezoids gives the work along the prescribed path. No calculus is needed to calculate this piecewise-linear example.

Force must be paired with the relevant displacement. A machine’s actuator or crosshead can move partly because the rig deforms; an impact system also needs an appropriate displacement reconstruction. The example stipulates an already corrected force–specimen-displacement record so that the arithmetic can focus on work. Integrating force against time instead gives impulse, with units N·s, rather than energy.

The conversion is simple once the axes are correct: 1 kN·mm = 1000 N × 0.001 m = 1 J. The peak force is one height on the graph; work is the accumulated area. The brittle/ductile lesson showed why equal peaks can have very different areas. Neither the peak nor the total area alone identifies a unique crack mechanism.

A resistance parameter describes a different experiment

In an admitted linear-elastic mode-I test, measured force and a defined specimen/crack geometry are used to calculate stress intensity K_I. A selected critical value is a resistance to crack extension for the stated regime. K has units MPa√m. The restricted K_IC label additionally requires the appropriate plane-strain test validity; a large specimen or a high critical force alone does not establish it.

With substantial plastic deformation, a suitable test can instead characterize resistance using J, an energy-like crack-driving quantity with units J/m², or crack-tip opening displacement, a length defined near the tip. Such tests use a specified analysis of force, deformation, geometry and crack extension. A total area of 35 J is not J merely because both involve energy. Dividing it by an arbitrary specimen cross-section does not create a valid J resistance measurement.

J and elastic energy-release descriptions concern how a cracked body’s energy or deformation state changes with crack advance under their assumptions. They are not simply the total work accumulated before the specimen breaks, nor only the energy of making new surfaces. A changing resistance curve may also distinguish initiation from further stable extension. Keep the reported parameter and its definition with the number.

The tests need not rank materials identically. A rounded notch can allow different initiation and plastic-deformation processes from a sharp crack; rate, constraint, geometry and temperature also change the response. A calibration relating impact energy to a particular fracture resistance can be useful within the material and test population that established it. It is an empirical relationship with uncertainty, not a universal conversion between joules and MPa√m.

Use a comparison only within its evidence

First match what was tested. A result belongs to its specimen dimensions, notch or precrack, material condition and orientation, temperature, loading rate and measurement route. For example, comparing a cold, fast test of one treatment with a warm, slow test of another changes several possible causes at once. The difference cannot be assigned to heat treatment alone.

Temperature deserves particular care near a brittle-to-ductile transition. A small temperature difference can produce a substantial energy difference where the response curve is steep. A transition temperature also needs its stated energy or appearance criterion, as the preceding mechanisms lesson demonstrates. One isolated energy number does not define that curve.

Repeat tests reveal scatter under the repeated conditions. Machine bias, specimen-temperature error, geometry measurement and material heterogeneity can contribute differently. Calibration and verification address instrument behavior, while uncertainty describes the limits on a particular reported result. A verified machine does not remove specimen variability or establish that the result represents another geometry or service environment.

Report the actual quantity, units and conditions; then state which inference it supports. A repeated impact-energy trend can compare matched conditions or motivate a mechanism investigation. A valid sharp-crack resistance can support a suitable crack assessment. Neither result, by itself, supplies every condition needed for component lifetime or behavior under cyclic or sustained loading.

Worked examples

Try the reasoning

1. Match these quantities to their meanings: 35 J, 25 MPa√m and 100 kJ/m². Why does having an energy-related unit not make them interchangeable?

   Show the answer and reasoning

   35 J can be test work; 25 MPa√m can be a stress-intensity resistance; 100 kJ/m² can be a J-based resistance, if each is actually established by its method.

   Their dimensions and normalizations differ, and units do not prove validity. Total work, field intensity and an energy-like crack-driving parameter come from different defined analyses.

2. Two alloys absorb the same energy in matched pendulum impact specimens. Must their valid plane-strain fracture toughnesses be equal? Explain a physical reason for your answer.

   Show the answer and reasoning

   No equality follows.

   The notch-impact event can include different initiation and plastic-deformation work from the response to a sharp precrack. Constraint and rate can also differ from the fracture test. A valid empirical correlation would need its own matched evidence and uncertainty.

3. A student divides 154 J by a 100 mm² cross-section to obtain 1.54 J/mm² and labels it J_IC. Check the unit conversion, then identify the more important interpretive error.

   Show the answer and reasoning

   1.54 J/mm² equals 1.54 × 10⁶ J/m², but that quotient is not thereby a valid J_IC.

   The arithmetic area normalization is possible. It does not supply the defined crack, force/deformation analysis, crack-extension criterion or test validity that establish a J-based initiation resistance. Correct dimensions are necessary, not sufficient.

4. A constant force of 1000 N acts for 0.001 s. Another statement says it acts through 0.001 m. Which gives 1 J, and what does the other give?

   Show the answer and reasoning

   1000 N × 0.001 m = 1 J of work. 1000 N × 0.001 s = 1 N·s of impulse.

   Force–time area is not force–displacement area. To infer displacement from a dynamic time record, additional motion information and a suitable analysis are required.

5. A treatment gives 80 J in a test at 20 °C, while another gives 40 J at −40 °C with a different notch. Can the energy difference be assigned to the treatment alone? Propose a useful next comparison.

   Show the answer and reasoning

   No. Test both treatments with matched specimen/notch geometry, orientation, temperature and loading conditions, with repeat measurements.

   The existing comparison confounds treatment with temperature and notch response. The lower value may reflect either or several of these changes. Matching conditions helps isolate the treatment question while retaining scatter and uncertainty.

6. A verified impact machine gives a lower result for a colder specimen. Must verification have failed?

   Show the answer and reasoning

   No. The specimen response can change with temperature even when the instrument behaves correctly.

   Check the actual specimen temperature, repeated material response and measurement uncertainty. Verification concerns machine performance; it does not require every material and condition to absorb the same energy.

What to carry forward

• Track gravitational energy, test work and machine loss before interpreting an impact number.
• Use force with displacement to calculate work; force with time gives impulse.
• A machined-notch impact energy and a valid sharp-crack resistance have different physical meanings and analyses.
• Matched conditions, resistance labels and uncertainty determine which comparison a reported number supports.