Radiographic sensitivity expresses the smallest detectable feature as a percentage of the penetrated thickness: sensitivity % = 100 x d / T, where d is the diameter of the finest wire resolved or the thickness of the hole-type IQI whose essential hole is visible. A 2 percent sensitivity on 25 mm of steel means a 0.5 mm feature was demonstrably visible.

Lower numbers are better, and 2 percent is the conventional benchmark that the ASME 2-2T quality level is built around. Achieving it depends on three independent things: contrast, set by beam energy, film class and processing; definition, set by geometric unsharpness and detector resolution; and noise, set by film grain or detector signal-to-noise. Any one of them can be the limiting factor, and improving the wrong one changes nothing.

For hole-type IQIs, ASTM E1025 defines the equivalent penetrameter sensitivity, EPS = 100 x sqrt(T_iqi x h / 2) / x, where h is the hole diameter and x the specimen thickness. It lets a 1T, 2T or 4T hole reading be put on a common scale. For the standard 2T hole the square root collapses to the IQI thickness and EPS reduces to the simple thickness ratio.

Sensitivity is a property of the radiograph, not of the weld. It is recorded on the report as evidence that the technique was capable of finding what the code requires, and a radiograph that fails its IQI requirement is rejected without ever being interpreted.

Worked example

IQI type	wire
Finest wire resolved	0.25 mm
Hole IQI thickness	0.43 mm
Smallest hole visible	h2
Penetrated thickness	12 mm
Required sensitivity	2 %
Radiographic sensitivity	2.08 %
Hole diameter used	0.86 mm
Equivalent penetrameter sensitivity	3.58 %
Smallest demonstrated feature	0.25 mm

A 0.25 mm wire resolved through 12 mm gives 100 x 0.25 / 12 = 2.08 percent, just outside a 2 percent requirement. For reference, the 0.43 mm plaque IQI with its 2T hole (0.86 mm) gives EPS = 100 x sqrt(0.43 x 0.86 / 2) / 12 = 100 x 0.43 / 12 = 3.58 percent.