Application: Ultrasonically identifying presence or absence of bonding between an outer layer of high impedance material such as metal or ceramic and an inner layer of low impedance material such as plastic, composite, or rubber. This technique also applies to many metal-to-metal glue joints.

Background: When materials of similar acoustic impedance (density multiplied by sound velocity) are joined to each other, such as metal welded or brazed to metal, or plastic fused to plastic, the presence or absence of bonding can usually be determined from the amplitude of the first reflection from the boundary. Typically there will be a significant difference in reflected signal amplitude between bonded and disbonded conditions. However, if there is a large difference in the acoustic impedance of the two materials, as in the case of bonds between metal and polymers, then due to that acoustic impedance mismatch there will be a significant reflection from the boundary even if the materials are mechanically bonded. The change in echo amplitude between bonded and disbonded conditions may be small and hard to detect, especially in situations where transducer coupling conditions are not uniform. This same situation often occurs in cases where metal is bonded to metal by means of epoxy or other low impedance adhesive. Unless the adhesive layer is extremely thin, it will constitute a low impedance bond line with a significant reflection even from a good joint.

Note that in cases where the low impedance material is on the outside, or part geometry otherwise permits coupling to the low impedance side of the joint, then the recommended procedure is a phase shift test. This is described in the Panametrics-NDT application note on Phase Shift Test for Bond Integrity.

In cases where the test must be performed from the high impedance side of the joint, the ringdown technique described in this application note is often the best approach. This involves observing a series of multiple reflections from the boundary and looking for a difference in the rate at which the amplitude of successive echoes, or the ringdown envelope, changes between bond and disbond conditions. As the sound wave reverberates in the metal or other high impedance material, the amplitude change is multiplied at each bounce, so for example a hard-to-see 5% amplitude drop between disbond and bond conditions at the first echo becomes a 25% drop at the fifth echo and a 50% drop at the tenth echo. While the change in a single echo may be small, the collective change across a series of multiple echoes will be much larger and easier to identify. In a typical joint, bond and disbond conditions will each produce their own distinctive ringdown pattern.

Equipment: This test can be performed with any standard ultrasonic flaw detector such as the Epoch XT, Epoch LTC, Epoch 600, or Epoch 1000. Ultrasonic pulser/receivers can also be employed. The transducer will usually be a common delay line transducer in the frequency range between 2.25 MHz and 20 MHz, such as a V207-RB, V206-RB, V202-RM, or V208-RM. The specific transducer for a given test will be selected based on material thickness, geometry, and acoustic properties.

Procedure: As with any other ultrasonic flaw detection procedure, setting up this test requires reference standards representing the conditions to be detected, in this case samples of the specific materials and geometry in question with known good and band bonds.

A delay line (or immersion) transducer is normally used for this type of test because it will typically generate more multiple echoes from a high impedance test piece than a contact transducer whose wearplate is closely matched to the target. Select a delay line transducer that generates at least five multiple backwall echoes, preferably more, from the bond line at the backwall of the outer material. Set the instrument time base to a large enough range to display a series of at least five multiple echoes on the screen. Couple the transducer to a reference standard representing unbonded material and adjust damping, filtering, and rectification as necessary to generate the cleanest and sharpest peaks.

Set gain such that the first backwall echo is at full screen height. In the waveform below, the first peak is the delay line interface echo and the subsequent peaks are multiple echoes from a 0.1 in. (2.5 mm) unbonded steel standard. Because the metal is air-backed, the multiple echoes diminish at a relatively slow rate.

Slow ringdown pattern from unbonded metal:



Then couple the transducer to the bonded reference standard and observe the change in the rate at which the height of the peaks diminishes. The waveform below represents a 0.1 in. (2.5 mm) bonded steel standard in which the steel is bonded to a polymer liner. The rate at which the peaks diminish is faster than in the first waveform.

Fast ringdown pattern from bonded metal:



Notice that the amplitude of the multiple echoes decreases rapidly towards the right side of the screen. The bonded polymer on the back side of the steel acts as a mechanical damping mechanism and causes the multiple echoes to decay more quickly. The exact variance in the ringdown pattern between bonded and unbonded conditions will depend on the relative acoustic impedance of the two materials as well as attenuation factors, but in principle this test works for any situation involving a high impedance material bonded to one of low impedance. By recording the two characteristic patterns observed in the reference standards and comparing them to the waveforms from a test piece, the presence or absence of bonding in the test piece can be determined.

When using instruments that incorporate Distance Amplitude Correction (DAC) software, it will often be helpful to use the DAC function to draw the ringdown profile of a properly bonded part on the screen. That DAC curve may then be used to quickly identify absence of bonding, which will be indicated by a series of peaks that rise above the line as seen below. Additionally, instrument alarms may be set to trigger on echoes that rise above the curve to further identify suspect bond conditions.

DAC curve marking ringdown profile of good bond, with waveform from bad bond:



It is important to remember that this type of test qualifies the presence or absence of bonding only, not the strength of bonding. Also, the materials must be acoustically bonded by an adhesive or similar bonding agent. Pressure alone is rarely sufficient to create an acoustic bond.