Thermally sprayed ceramic coatings are widely used in industrial applications where the coated
component is subject to, e.g., high thermal loads or mechanical wear. Several non-destructive
testing techniques are available for the detection of defects in ceramic materials or for the evaluation of density and density variations. In addition to this, ultrasonic techniques can be used for quantitative evaluation of elastic properties of materials. This evaluation is based on the measurement of sound velocities of different wave modes in the material. Acoustic microscopy
operating at very high (> 100 MHz) frequencies has been used to measure the sound velocities in
homogeneous and thin coatings. With this type of equipment, reliable and accurate results have been achieved in laboratory measurements. However, less attention has been paid on the development of techniques for industrial applications on-site. The present work was focused on the development of measurement techniques for industrial applications.
Large-aperture low-frequency transducers were designed and constructed for the measurement of
sound velocities in thermally sprayed ceramic coatings. In the first type of transducers the
piezoelectric element was a plastic (PVDF) -film that was shaped to create the required focus. The practical measurement of the sound velocity is based on a modification of the V(z)-technique known from acoustic microscopy. The measurements carried out with specimens having different types of ceramic coatings and variable coating thickness show that the accuracy of the technique is sufficient for most industrial applications.
For industrial measurements the technique can be simplified and the measurement performed
The first large-aperture low-frequency transducers designed by VTT were used in immersion which is not always practical in industrial applications. Recently also contact transducers of similar type have been designed and constructed.
Figure 1 shows the incident pulse as originating at a highly focused lens which is used both as
transmitter and receiver. The pulse is incident perpendicular to the surface. A ray emerging from the lens at the critical angle for surface wave generation, qR , and hitting the surface at point  will generate a leaky surface wave. As this surface wave propagates, it will generate a compression wave in the fluid which emerges at the critical angle.
Fig 1: Transducer-lens configuration for acoustic microscope (Gilmore et al. 1989)|
A common approach to surface wave velocity measurement with acoustic microscopy is to monitor
the difference in arrival times of pulses propagating along ray paths - - - andthe pulse directlyreflected from the surface back to the transducer [I] in Fig. 1. This time difference is calculated by a simple consideration of difference in path lengths and wave velocities (Gilmore et al. 1989).
Based on the differences in round-trip travel time for the direct reflection (Dt1 ) and the surface wave (Dt2 ) at the two defocus distances ZAand ZBfollowing equation for calculation of velocity of Rayleigh waves (VR) has been presented by Gilmore et al. (1989):
Formula (1) permits the Rayleigh velocity to be determined from the time measurements at two
water path settings.
Measurement of elastic anisotropy
If the measurement of sound velocities is performed by using a spherical lens which is circularly
focused into a point leaky surface waves propagating in all directions are excited and the acoustic properties are measured as a mean value around the beam axis. This means that the sound velocities measured with this type of transducer can not be used for assessment of elastic anisotropy. For this purpose line-focussed acoustical transducers have been developed (Chubachi et al. 1991).
For on-site applications line-focused low-frequency wide-aperture ultrasonic transducers have been constructed by VTT. The structure of these probes is schematically shown in Fig. 2.
Fig 2: The construction of the line-focused low-frequency transducer.|
The ultrasonic beam is focused to the required depth by using, as an active element, a rectangular shaped piece of polyvinylidene fluoride (PVDF) film that has been glued on the cylindrical convex surface of the probe body. The flexibility of the PVDF film allows bending to the curvature needed for required focusing. In the probes described here the radius of the curvature was 20 mm and the aperture angle of the probe 120°. In this transducer the narrow-line-focus was created by bending the rectangular piece of film also in its width direction (Fig. 2). The curvature was adjusted to focus the beam in this direction at a distance 5 mm shorter than in the other direction. Thus the focus of the probe was a narrow line after defocusing the probe by 5 mm. The probe body was manufactured from bulk PVDF material that was machined to the shape of the final probe.
The application of the same approach to surface wave velocity measurement as presented by
Gilmore et al. (1989) and calculation of the time difference (Dt) between the sound beams following the two paths shown in Fig.1 yields following equation for the time difference:
Here the velocity of longitudinal waves propagating from transducer to the test material through the contact liquid is V 1 and the propagation velocity in the test material V2 .
A practical method for performing the measurement in a short time and with a reasonable accuracy
for industrial applications is following:
First the probe is adjusted at a distance from the specimen giving the maximum signal from
the directly reflected beam (the focus of the probe on the surface of the specimen). The Z
coordinate of the scanning system is reset to 0.
- The probe is moved in the Z direction, closer to the surface of the specimen (defocused). The
probe is stopped at the position where the distance (time difference t) between the signal from
the directly reflected beam and from the mode converted beam travelling on the surface of the
specimen can be accurately measured on the CRT screen.
- The displacement (z) of the scanner in the Z direction is recorded.
- Based on the measured values of z [mm] and t [ns] the sound velocity can be determined from
Results of experimental measurements
Experimental measurements with different types of coatings were carried out by using the
methodology described above and the C-SAM ultrasonic inspection system manufactured by
Ultrasonic Sciences Ltd., UK. In all the measurements the samples were immersed in water
(Kauppinen, 1997). Table 1 shows examples of velocities of surface acoustic waves (VSAW )
measured in alumina coatings with different thicknesses.
||Thickness of coating[mm]
|Table 1: Velocities of surface acoustic waves in thermally sprayed alumina coatings. The velocities
are measured in the direction of spraying (0°) and perpendicular to the spraying direction (90°).|
The velocities were measured in two directions perpendicular to each other, in order to assess
possible directional anisotropy of the coating. However, the results show no remarkable differences in the velocities measured in different directions when taking into account the general scatter of sound velocity due to the inhomogeneous structure of the coating.
It is evident that the velocities measured in samples AL500-AL2000 are affected by the substrate material and do not purely represent the coating. With the frequency of the probe (nominally 10 MHz) used in the measurements, the wavelength is too large, and therefore the surface waves penetrate also into the substrate. In layered specimens the surface acoustic waves are dispersive and the velocity VSAW is a function of both the frequency and the layer's film thickness and elastic parameters (Weglein 1985). Because in most of the samples studied here the thickness of the coating is less than the effective wave length (l > 2 mm) used in the measurement, the substrate has affected the result. The velocity of the Rayleigh wave in steel is 2996 m/s (Briggs 1992), which is higher than the velocities measured in samples AL500-AL2000. In this case, acoustically slow layers are
deposited on fast surfaces. This was also verified by measuring the velocity of the bulk longitudinal waves in the coating. The velocities measured (3200-4100 m/s) are clearly lower than the velocity of the longitudinal waves in steel VL = 5960 m/s (Briggs 1992). This means that also other modes of surface waves than Rayleigh waves (e.g. Lamb waves) can be formed in the coating. Only in the sample AL4500 was the measured velocity higher than the velocity of the Rayleigh waves in the substrate and the thickness of the coating was clearly higher than the effective wavelength. Thus this velocity can be used to evaluate the elasticity. By using the following approximation formula (3) presented by Bergman (1954) Young's modulus can be calculated.
In the calculation the value n = 0.25 for the Poisson ratio and r =3200 kg/m 3 for density have been used. This gives Young's modulus E = 100 GPa.
The same transducer can also be used for detection of delaminations and cracks. Figure 3 shows a C-scan image of alumina coating containing a delamination and two cracks opening to the surface of the coating. This C-scan image has been measured by using the longitudinal wave that is directly reflected back from the interface between the coating and the substrate. The cracks opening to the surface are strongly scattering the ultrasonic beam and can therefore be seen on the image.
Fig 3: C-scan image of thermally sprayed alumina coating on a steel substrate|
Low-frequency wide-aperture contact transducer
The measurements described above have been performed in immersion which is not always practical in industrial applications. Local immersion can be arranged by constructing a water chamber around the transducer but it is still more practical when measurements can be performed in direct contact with the material studied. For this type of measurements a low-frequency wide-aperture contact transducer was designed and constructed. With this transducer the coating thickness, possible delaminations and cracks and the relative sound velocity in the coating can be easily measured in industrial conditions. Due to the fact that in this transducer the degree of defocussing is fixed the transducer has to be separately designed for each application based on the approximate sound velocity in the material to be studied.
Low-frequency wide-aperture transducers can be used for the assessment of elasticity and integrity of thermally sprayed ceramic coatings in industrial conditions. Due to the inhomogeneous structure of the coating the scatter in sound velocities is relatively large and consequently only rough estimation of elasticity is possible. By using a transducer that has been focussed in two directions to create a line focus also the elastic anisotropy can be assessed. Furthermore, the same transducer can be used for the detection of delaminations and cracks in the coating.
- Gilmore, R. S., Hewes, R. A., Thomas, L. J. and Young, J. D. 1989. Broadband acoustic microscopy:
scanned images with amplitude and velocity information. In: Shimizu, H., Chubachi, N. and Kushibiki,
J. (eds). Acoustical Imaging, Volume 17. New York, USA: Plenum Press. Pp. 97-110.
- Chubachi, N., Kushibiki J., Sannomiya, T., Naruge, I, Saito, K. and Watanabe, S. 1991. Acoustical
images observed by directional PFB microscope. In: Lee, H. & Wade, G. (eds.). Acoustical Imaging,
Volume 18. New York, USA: Plenum Press. Pp. 255-260.
- Kauppinen, P. 1997. The evaluation of integrity and elasticity of thermally sprayed ceramic coatings
by ultrasonics. VTT Publications 325. Espoo, Finland: Technical Research Centre of Finland (VTT).
- Weglein, R. D. 1985. Acoustic Micro-Metrology. IEEE Transactions on Sonics and Ultrasonics, Vol.
SU-32, No. 2, pp. 225-234.
- Briggs, A. 1992. Acoustic Microscopy. Oxford, UK: Clarendon Press. 325 p.
- Bergmann, L. 1954. Der Ultraschall. Zürich, Switzerland: S. Hirzel Verlag. 1114 p.