ASNT NDT Handbook Volume 7 ultrasonic Testing

Poor correlations or indeterminate results Modulus Degradation
were found in other time partition zones
or over the entire time window of the Another alternative for defining the stress
waveform. wave factor is given with respect to
measuring changes in modulus (stiffness)
Adhesive Bond Strength associated with cyclic fatiguing and
associated microcracking (see Fig. 26).196
An alternative approach to defining the In this case, the stress wave factor was
stress wave factor was found useful in defined as the root mean square value of
establishing a correlation with the shear the power spectrum of the
strength of adhesively bonded steel acoustoultrasonic waveform.190
plates.61 For the results given in Fig. 25,
the stress wave factor was expressed as a Note that the stress wave factor is
voltage weighted ringdown count: about ten times more sensitive to the
effects of fatigue damage than the secant
p modulus measurement. Moreover, the
∑ ( )(55) SWF = slopes of the two curves in Fig. 26 differ
Vi Ci − Ci +1 in detail because, although the secant
modulus refers to the length of the entire
i test object, the stress wave factor
measurements represent only part of it.

where Ci is the number of counts at the Limitations of
ith level, Vi is the threshold voltage at the Acoustoultrasonic
ith level and Vp is the peak voltage of the Techniques
waveform.
Acoustoultrasonic testing represents a
In this case, the entire raw waveform generalized approach for materials
characterization and carries both the
above a preselected minimum voltage capabilities and the limitations found in a
variety of kindred techniques. Beyond the
threshold was used to quantify the stress limitations common to all discontinuity
detection methods, the ultrasonic
wave factor. However, conditions may characterization of subtle discontinuities
and material properties is also subject to
dictate that the best correlation depends circumstances that affect sensitivity and
signal reproducibility.197
on first filtering the raw waveform and
Both the acoustoultrasonic and pulse
dealing only with that portion of the echo technique are vulnerable to
transducers misalignment and couplant
signal within a preselected bandwidth or variations. With pulse echo ultrasonic
frequency zone.195 testing, it is common to misread
attenuation by a factor of ten or more.
FIGURE 25. Stress wave factor versus shear
strength of adhesively bonded steel test Normalized stress wave factor
objects for range of test temperatures. Stress (relative unit)
wave factor was calculated as voltage
weighted ringdown count with correlation Normalized secant modulus
factor of 0.964.61 Normalized stress wave factor
FIGURE 26. Covariation of stress wave factor and secant
1.2
modulus with fatigue degradation in graphite epoxy fiber
1.0
composite laminate. Stress wave factor is the root mean
0.8 square of the power spectrum.190, 196

0.6 1.0 Loading ratio R = 0.1 1.0
0.96 [0, 90, ±45]S laminate 0.8
0.4 0.92 0.6
0.88 Stiffness 0.4
0.2 0.84 (left scale) 0.2

0 Stress
0.2 0.4 0.6 0.8 1.0 1.2 wave
factor
Normalized average shear strength
0.8 0
Legend 0 10 20 30 40 50 60 70
= 30 °C (85 °F) Number of fatigue cycles (thousands)
= 60 °C (140 °F)
= 90 °C (190 °F)
= 120 °C (250 °F)
= 150 °C (300 °F)

340 Ultrasonic Testing

This magnitude of error can occur if the Closing
transducer is not literally ground onto the
test object surface.28 This error is Ultrasonic nondestructive material
particularly common at high frequencies property characterization can be divided
(over 20 MHz) where couplant thickness into several categories.
variations, bubbles, surface porosity and
reflection coefficient anomalies can have 1. Elastic moduli are determined through
serious effects.29,33 measurements of ultrasonic speed or
dynamic vibration.
Factors Affecting Test Results
2. Some methods depend on speed and
Acoustoultrasonic measurements are attenuation measurements to
affected by several factors associated with characterize residual stress, grain size,
the attachment of the transducer to the porosity, texture and other
test object: (1) applied pressure, (2) type microstructural factors of material
and amount of couplant, (3) object behavior. Sometimes these
surface roughness, (4) transducer microstructural factors can be used to
alignment, (5) spacing between predict values or variations of
transducers and (6) exact location of mechanical properties such as strength
transducers on the object. or toughness.

Even if these are optimized, a further 3. Ultrasonic measurements can be
problem remains. In practice, correlated with mechanical properties:
acoustoultrasonic measurements must be speed with hardness or attenuation
repeatable over the test object surface. with toughness. This category
This may require lifting and recoupling generally includes empirical
the transducers or inventing transducers correlations that have been found to
that can scan while remaining in contact apply only to specific materials,
with the surface. usually in the form of laboratory
specimens.
The coupling problems associated with
scanning may be avoided by using Precise ultrasonic measurements must
noncontact laser ultrasonics. However, be made on test objects with specific size,
laser ultrasonics introduces other shape, thickness and surface condition.
problems that can limit signal control and Or, for the measurements to have at least
readout. Such problems arise from surface relative significance, geometric properties
roughness, reflectivity and other factors. of the test objects must be held constant,
as in damping measurements to obtain
The selection of sending and receiving relative modulus changes. This leads to an
transducers, their bandwidth, their alternative approach to materials
resonance frequencies and their internal characterization in which greater
damping all have an effect on test results, emphasis is placed on signal analysis to
primarily because ringdown in an extract information on relative changes in
undamped transducer can be confused material properties. In this approach,
with reverberations. In testing composite ultrasonic wave propagation, energy
panels, it is useful to select transducer transfer and signal modulation properties
frequencies that introduce wavelengths of a material are used to assess relative
less than the panel thickness. Or, in the variations in mechanical properties:
case of continuous fiber reinforced ultimate strength or bond strength.
composites, it is helpful to cover the
frequency range likely to be transmitted Empirical correlations between
by both the composite and fibers acting as ultrasonic measurements and mechanical
waveguides. properties have important roles in
industry. There are likely to be
These two considerations dictate ambiguities concerning the exact nature
transducer spacing, which must be small and influence of underlying
enough to avoid losing the reception of microstructural factors being measured.
high frequency signal components. This is true in the case of complex,
However, general guidelines for selecting heterogeneous, anisotropic, textured and
transducer frequency, bandwidth and composite materials where several
instrumentation parameters cannot be variables can simultaneously affect wave
prescribed for all cases. The best approach propagation. These factors introduce
is to use the successful examples cited complex relations among microstructure,
here and to experimentally seek the mechanical properties and load response
optimum conditions for particular (deformation and fracture modes).
applications. This approach can be
facilitated by waveform (or spectrum) The inferring of material properties
partitioning and by regression analysis to often depends on the use of two or more
identify the portion of the signal that best corroborative and complementary test
correlates with the material property of methods, some of which are neither
interest. widely applied nor widely accepted in
industry. The biggest challenge in

Ultrasonic Characterization of Material Properties 341

ultrasonic materials characterization has
been to apply the techniques to practical
field work.

Wider use of ultrasonic methods has
expanded through significant
improvements in computers,
instrumentation and materials
themselves, particularly with respect to
electronic materials and nanotechnology.
The literature describing these
advancements has continued to
grow.198-207

342 Ultrasonic Testing

References

1. Vary, A. Section 12, “Material 11. Nowich, A. and B. Berry. Anelastic
Property Characterization.” Relaxation in Crystalline Solids.
Nondestructive Testing Handbook, London, United Kingdom: Academic
second edition: Vol. 7, Ultrasonic Press (1972).
Testing. Columbus, OH: American
Society for Nondestructive Testing 12. Tittmann, B. “Apparatus for
(1991): p 383-431. Measuring High Internal Friction
Q-Factors.” Review of Scientific
2. Green, R.E. Ultrasonic Investigation of Instruments. Vol. 47. Melville, NY:
Mechanical Properties. Treatise on American Institute of Physics (1976):
Materials Science and Technology. p 1516.
Vol. 3. Saint Louis, MO:
Elsevier/Academic Press (1973). 13. Smith, C., ed. Internal Friction and
Ultrasonic Attenuation in Solids:
3. Green, R.E. “Ultrasonic Attenuation Proceedings of the European Conference
Detection of Fatigue Damage.” on Internal Friction and Ultrasonic
Ultrasonics International 1973. Attenuation [July 1980]. Oxford,
Guildford, United Kingdom: Elsevier United Kingdom: Pergamon Press
(1973): p 187-193. (1980).

4. Green, R.E. “Ultrasonic 14. Cawley, P. and R. Adams. “Vibration
Nondestructive Materials Techniques.” Non-Destructive Testing
Characterization.” Materials Analysis of Fibre-Reinforced Plastic Composites.
by Ultrasonics. Park Ridge, NJ: Noyes Vol. 1. London, United Kingdom:
Data Corporation (1987): p 1-29. Elsevier Applied Science (1987):
p 151-200.
5. Vary, A. “Ultrasonic Measurement of
Material Properties.” Research 15. Deka, M. and N. Eberhardt. “Internal
Techniques in Nondestructive Testing. Friction of Fe-Based Binary Alloys at
Vol. 4. London, United Kingdom: High Frequency.” Nondestructive
Academic Press (1980): p 159-204. Methods for Material Property
Determination [Hershey, PA,
6. Manderscheid, J.M. and April 1983]. New York, NY: Plenum
J.P. Gyekenyesi. “Fracture Mechanics (1984): p 135-148.
Concepts in Reliability Analysis of
Monolithic Ceramics.” Nondestructive 16. DiCarlo, J. and J. Maisel.
Testing of High-Performance Ceramics “Measurement of the
[Boston, MA, August 1987]. Time-Temperature Dependent
Columbus, OH: American Ceramics Dynamic Mechanical Properties of
Society (1987): p 59-72. Boron/Aluminum Composites.”
Composite Materials: Testing and
7. Kanninen, M. and C. Popelar. Design. Special Technical Publication
Advanced Fracture Mechanics. Oxford, 674. West Conshohocken, PA: ASTM
United Kingdom: Oxford University International (1979): p 201-227.
Press (1985): p 392-432.
17. Papadakis, E.P. and B. Kovacs.
8. Cawley, P. and R. Adams. “The “Theoretical Model for Comparison
Mechanics of the Coin-Tap Method.” of Sonic-Resonance and Ultrasonic
Journal of Sound and Vibration. Velocity Techniques for Assuring
Vol. 122, No. 2. Amsterdam, Quality in Nodular Iron Parts.”
Netherlands: Elsevier (1988): Materials Evaluation. Vol. 38, No. 6.
p 299-316. Columbus, OH: American Society for
Nondestructive Testing (May 1980):
9. Adams, R.D. and P. Cawley. p 25-30.
“Vibration Techniques in
Nondestructive Testing.” Research 18. Shiori, J., O. Furuta and K. Satoh.
Techniques in Nondestructive Testing. “Analysis of Elevated Property of
Vol. 8. London, United Kingdom: Heat Resistant Materials by Internal
Academic Press (1985): p 303-360. Friction.” Nondestructive
Characterization of Materials II
10. Uygur, E. “Nondestructive Dynamic [Montreal, Canada, July 1986]. New
Testing.” Research Techniques in York, NY: Plenum (1987): p 325-333.
Nondestructive Testing. Vol. 4.
London, United Kingdom: Academic 19. Matthews, J.R., ed. Acoustic Emission.
Press (1980): p 205-244. New York, NY: Gordon and Breach
Science Publishers (1983).

Ultrasonic Characterization of Material Properties 343

20. Kline, R.A. “Acoustic Emission Signal 32. Vary, A. “Concepts and Techniques
Characterization.” Acoustic Emission. for Ultrasonic Evaluation of Material
New York, NY: Gordon and Breach Mechanical Properties.” Mechanics of
Science Publishers (1983): p 105-138. Nondestructive Testing. New York, NY:
Plenum (1980): p 123-141.
21. Arrington, M. “Acoustic Emission.”
Non-Destructive Testing of 33. Generazio, E.R. “The Role of the
Fibre-Reinforced Plastic Composites. Reflection Coefficient in Precision
Vol. 1. London, United Kingdom: Measurement of Ultrasonic
Elsevier Applied Science (1987): Attenuation.” Materials Evaluation.
p 25-63. Vol. 45, No. 8. Columbus, OH:
American Society for Nondestructive
22. Liptai, R.G., D.O. Harris and Testing (July 1985): p 995-1004.
C.A. Tatro. Acoustic Emission. Special
Technical Publication 505. West 34. Vary, A. and H. Kautz. “Transfer
Conshohocken, PA: ASTM Function Concept for Ultrasonic
International (1972). Characterization of Material
Microstructures.” International
23. Spanner, J. Acoustic Emission: Advances in Nondestructive Testing.
Techniques and Applications. Evanston, Vol. 13. New York, NY: Gordon and
IL: Intex Publishing (1974). Breach (1988): p 193-249.

24. Wadley, H.N.G. and R. Mehrabian. 35. Hull, D., H. Kautz and A. Vary.
“Acoustic Emission for In-Process “Measurement of Ultrasonic Velocity
Monitoring and Microstructure Using Phase-Slope and
Control?” Nondestructive Methods for Cross-Correlation Methods.”
Material Property Determination Materials Evaluation. Vol. 43, No. 11.
[Hershey, PA, April 1983]. New York, Columbus, OH: American Society for
NY: Plenum (1984): p 207-236. Nondestructive Testing
(October 1985): p 1455-1460.
25. Nondestructive Testing Handbook,
second edition: Vol. 5, Acoustic 36. Sachse, W. and Y. Pao. “On the
Emission Testing. Columbus, OH: Determination of Phase and Velocity
American Society for Nondestructive of Dispersive Waves in Solids.”
Testing (1986). Journal of Applied Physics. Vol. 39.
Melville, NY: American Institute of
26. Scruby, C. “Quantitative Acoustic Physics (1978): p 4320.
Emission Techniques.” Research
Techniques in Nondestructive Testing. 37. Tittmann, B., H. Nadler, V. Clark,
Vol. 8. London, United Kingdom: L. Ahlberg and T. Spencer.
Academic Press (1985): p 141-210. “Frequency Dependence of Seismic
Dissipation in Rocks.” Geophysical
27. Green, R.E. “Basic Wave Analysis of Research Letters. Vol. 8. Washington,
Acoustic Emission.” Mechanics of DC: American Geophysical Union
Nondestructive Testing. New York, NY: (1981): p 36-88.
Springer (1980): p 55-76.
38. Winkler, K. “Frequency Dependent
28. Truell, R., C. Elbaum and B. Chick. Ultrasonic Properties of High
Ultrasonic Methods in Solid State Porosity Sandstones.” Journal of
Physics. Saint Louis, MO: Geophysical Research. Vol. 88.
Elsevier/Academic Press (1969). Washington, DC: American
Geophysical Union (1983):
29. Vary, A. “Simulation of p 9493-9499.
Transducer-Couplant Effects on
Broadband Ultrasonic Signals.” 39. Krautkrämer, J. and H. Krautkrämer.
International Advances in Ultrasonic Testing of Materials, fourth
Nondestructive Testing. Vol. 8. New edition. New York, NY:
York, NY: Gordon and Breach (1980): Springer-Verlag (1990).
p 167-200.
40. Kinsler, L.E., A.R. Frey, A.B. Coppens
30. Papadakis, E.P. “Ultrasonic Velocity and J.V. Sanders. Fundamentals of
and Attenuation: Measurement Acoustics, fourth edition. New York,
Methods with Scientific and NY: Wiley (1999).
Industrial Applications.” Physical
Acoustics: Principles and Methods. 41. Papadakis, E.P. “Buffer Rod System
Vol. 12. Saint Louis, MO: for Ultrasonic Attenuation
Elsevier/Academic Press (1976): Measurements.” Journal of the
p 277-374. Acoustical Society of America. Vol. 44,
No. 5. Melville, NY: American
31. Vary, A. “Computer Signal Processing Institute of Physics, for the
for Ultrasonic Attenuation and Acoustical Society of America (1968):
Velocity Measurement for Material p 1437-1441.
Property Characterization.”
Proceedings of the Twelfth Symposium
on Nondestructive Evaluation [San
Antonio, TX, April 1979]. San
Antonio, TX: Southwest Research
Institute (1979): p 33-46.

344 Ultrasonic Testing

42. Goebbels, K. “Structural Analysis by 53. Vary, A. “The Acousto-Ultrasonic
Scattered Ultrasonic Radiation.” Approach.” Acousto-Ultrasonics:
Research Techniques in Nondestructive Theory and Applications [Blacksburg,
Testing. Vol. 4. London, United VA, July 1987]. New York, NY:
Kingdom: Academic Press (1980): Plenum (1988): p 1-21.
p 87-157.
54. Scruby, C., R. Smith and B. Moss.
43. Theiner, W.A. and H.H. Willems. “Microstructural Monitoring by Laser
“Determination of Microstructural Ultrasonic Attenuation and Forward
Parameters by Ultrasonic NDE.” Scattering.” NDT International.
Nondestructive Methods for Material Vol. 19, No. 5. Guildford, Surrey,
Property Determination [Hershey, PA, United Kingdom: Butterworth
April 1983]. New York, NY: Plenum Scientific Limited (1986): p 307-313.
(1984): p 249-258.
55. Rosen, M. “Analytical Ultrasonics for
44. Willems, H. and K. Goebbels. Characterization of Metallurgical
“Characterization of Microstructure Microstructures and
by Backscatter Ultrasonic Waves.” Transformations.” Materials Analysis
Metal Science. London, United by Ultrasonics. Park Ridge, NJ: Noyes
Kingdom: Metals Society (1981): Data Corporation (1987): p 79-98.
p 549-553.
56. Tittmann, B., L. Ahlberg, J.
45. Fitting, D.W. and L. Adler. Ultrasonic Richardson and R. Thompson.
Spectral Analysis for Nondestructive “Determination of Physical Property
Evaluation. New York, NY: Plenum Gradients from Measured Surface
(1981). Wave Dispersion.” Transactions on
Sonics and Ultrasonics. Vol. 34, No. 5.
46. Good, M.S. and J.L. Rose. New York, NY: Institute of Electrical
“Measurement of Thin Case Depth in and Electronics Engineers (1987):
Hardened Steel by Ultrasonic p 500-507.
Pulse-Echo Technique.”
Nondestructive Methods for Material 57. Schreiber, E., O. Anderson and
Property Determination [Hershey, PA, N. Soga. Elastic Constants and Their
April 1983]. New York, NY: Plenum Measurement. New York, NY:
(1984): p 189-203. McGraw-Hill (1973).

47. Segal, E. and J. Rose. “Nondestructive 58. J.C. Duke, Jr., and E.G. Henneke, II.
Testing Techniques for Adhesive “Analytical Ultrasonics for
Bond Joints.” Research Techniques in Evaluation of Composite Material
Nondestructive Testing. Vol. 7. Response.” Materials Analysis by
London, United Kingdom: Academic Ultrasonics. Park Ridge, NJ: Noyes
Press (1980): p 275-316. Data Corporation (1987): p 148-163.

48. Telschow, K.L. and J.E. Flinn. 59. Kautz, H. Acousto-Ultrasonic
“Ultrasonic Characterization of Verification of the Strength of Filament
Consolidated Rapidly Solidified Wound Composite Material. NASA
Powders.” Nondestructive Technical Memorandum 88827.
Characterization of Materials II Washington, DC: National
[Montreal, Canada, July 1986]. New Aeronautics and Space
York, NY: Plenum (1987): p 149-157. Administration (1986).

49. Qu, J. and J. Achenbach. “Analytical 60. Dos Reis, H. and H. Kautz.
Treatment of Polar Backscatter from “Nondestructive Evaluation of
Porous Composites.” Review of Adhesive Bond Strength Using the
Progress in Quantitative Nondestructive Stress Wave Factor Technique.”
Evaluation. Vol. 6B. New York, NY: Journal of Acoustic Emission. Vol. 5,
Plenum (1987): p 1137-1146. No. 4. Los Angeles, CA: Acoustic
Emission Group (1986): p 144-147.
50. Tittmann, B., L. Ahlberg and
K. Fertig. “Ultrasonic 61. Fahr, A., S. Lee, S. Tanary and Y.
Characterization of Microstructures Haddad. “Estimation of Strength in
in Powder Metal Alloy.” Materials Adhesively Bonded Steel Specimens
Analysis by Ultrasonics. Park Ridge, by Acousto-Ultrasonic Technique.”
NJ: Noyes Data Corporation (1987): Materials Evaluation. Vol. 45, No. 2.
p 30-46. Columbus, OH: American Society for
Nondestructive Testing
51. Bar-Cohen, Y. “Ultrasonic NDE of (February 1989): p 233-239.
Composites — A Review.” Solid
Mechanics Research for Quantitative 62. Rokhlin, S.I. “Characterization of
Non-Destructive Evaluation [Evanston, Composites and Adhesive Cure by
IL, September 1985]. Dordrecht, Ultrasonic Waves.” Nondestructive
Netherlands: Martinus Nijhoff Characterization of Materials II
(1987): p 187-201. [Montreal, Canada, July 1986]. New
York, NY: Plenum (1987): p 105-113.
52. Duke, J.C., Jr., ed. Acousto-Ultrasonics:
Theory and Applications [Blacksburg,
VA, July 1987]. New York, NY:
Plenum (1988).

Ultrasonic Characterization of Material Properties 345

63. Rokhlin, S.I. “Adhesive Joint 74. Williams, J.H., S.S. Lee and
Evaluation by Ultrasonic Interface H. Karagülle. “Input-Output
and Lamb Waves.” Materials Characterization of an Ultrasonic
Evaluation by Ultrasonics. Park Ridge, Testing System by Digital Signal
NJ: Noyes Data Corporation (1987): Analysis.” Materials Analysis by
p 299-310. Ultrasonics. Park Ridge, NJ: Noyes
Data Corporation (1987): p 302-330.
64. Bracewell, R.N. The Fourier Transform
and Its Applications, third edition. 75. Williams, J.H. and S.S. Lee. “Pattern
New York, NY: McGraw-Hill (1999). Recognition Characterizations of
Micromechanical and Morphological
65. Kline, R.A. “Measurement of Material States via Analytical
Attenuation and Dispersion Using an Ultrasonics.” Materials Analysis by
Ultrasonic Spectroscopy Technique.” Ultrasonics. Park Ridge, NJ: Noyes
Journal of the Acoustical Society of Data Corporation (1987): p 193-206.
America. Vol. 76, No. 2. Melville, NY:
American Institute of Physics, for the 76. Weaver, R. “Diffuse Field Decay Rates
Acoustical Society of America (1984): for Material Characterization.” Solid
p 498-504. Mechanics Research for Quantitative
Non-Destructive Evaluation [Evanston,
66. Kline, R. and D. Egle. “Applications IL, September 1985]. Dordrecht,
of Digital Methods to Ultrasonic Netherlands: Martinus Nijhoff
Materials Characterization.” NDT (1987): p 426-434.
International. Vol. 19, No. 5.
Guildford, Surrey, United Kingdom: 77. Weaver, R. “Diffuse Waves for
Butterworth Scientific (1986): Materials NDE.” Acousto-Ultrasonics:
p 341-347. Theory and Applications [Blacksburg,
VA, July 1987]. New York, NY:
67. Roberts, R. “Porosity Characterization Plenum (1988): p 35-43.
in Fiber-Reinforced Composites by
Backscatter.” Review of Progress in 78. Gericke, O. “Ultrasonic
Quantitative Nondestructive Evaluation. Spectroscopy.” Research Techniques in
Vol. 6B. New York, NY: Plenum Nondestructive Testing. London,
(1987): p 1147-1156. United Kingdom: Academic Press
(1970): p 31-62.
68. Henneke, E.G., II, J.C. Duke, Jr. and
R.C. Stiffler. “Characterization of the 79. Serabian, S. “Ultrasonic Material
Damage State of Composite Property Determinations.” Materials
Laminates via the Acousto-Ultrasonic Analysis by Ultrasonics. Park Ridge,
Technique.” Solid Mechanics Research NJ: Noyes Data Corporation (1987):
for Quantitative Non-Destructive p 211-224.
Evaluation [Evanston, IL,
September 1985]. Dordrecht, 80. Smith, R. “Ultrasonic Materials
Netherlands: Martinus Nijhoff Characterization.” NDT International.
(1987): p 217-235. Vol. 20, No. 1. Guildford, Surrey,
United Kingdom: Butterworth
69. Andrews, H.C. Introduction to Scientific Limited (1987): p 23-28.
Mathematical Techniques in Pattern
Recognition. New York, NY: 81. Tittmann, B., B. Hosten and M.
Wiley-Interscience (1972). Abdel-Gawad. “Ultrasonic
Attenuation in Carbon-Carbon
70. Johnson, R.A. and D.W. Wichern. Composites and the Determination
Applied Multivariate Statistical of Porosity.” Proceedings of the 1986
Analysis, fifth edition. Upper Saddle IEEE Ultrasonics Symposium. New
River, NJ: Prentice Hall (2002). York, NY: Institute of Electrical and
Electronics Engineers (1986).
71. Oppenheim, A.V., A.S. Willsky and
S.H. Nawab. Signals and Systems, 82. Vary, A. and R. Lark. “Correlation of
second edition. Upper Saddle River, Fiber Composite Tensile Strength
NJ: Prentice Hall (1996). with the Ultrasonic Stress Wave
Factor.” Journal of Testing and
72. Gammell, P. “Coherent Processing of Evaluation. Vol. 7, No. 4.
the Full Analytical Signal Philadelphia, PA: American Society
Information of Ultrasonic for Testing and Materials (1979):
Waveforms.” International Advances in p 185-191.
Nondestructive Testing. Vol. 10. New
York, NY: Gordon and Breach (1984): 83. Vary, A. and K. Bowles. “Ultrasonic
p 183-266. Evaluation of the Strength of
Unidirectional Graphite/Polyimide
73. Karagülle, H., J.H. Williams and S.S. Composites.” Proceedings of the
Lee. “Application of Homomorphic Eleventh Symposium on Nondestructive
Signal Processing to Stress Wave Evaluation. San Antonio, Texas:
Factor Analysis.” Materials Evaluation. Southwest Research Institute (1977):
Vol. 43, No. 11. Columbus, OH: p 242-258.
American Society for Nondestructive
Testing (October 1985): p 1446-1454.

346 Ultrasonic Testing

84. Vary, A. “Correlations among 94. Gruber, J., J. Smith and R.
Ultrasonic Propagation Factors and Brockelman. “Ultrasonic Velocity C-
Fracture Toughness Properties of Scans for Ceramics and Composites.”
Metallic Materials.” Materials Materials Evaluation. Vol. 46, No. 1.
Evaluation. Vol. 36, No. 7. Columbus, Columbus, OH: American Society for
OH: American Society for Nondestructive Testing (January
Nondestructive Testing (June 1978): 1988): p 90-96.
p 55-64.
95. Generazio, E.R., D.J. Roth and G.Y.
85. Vary, A. and D. Hull. “Interrelation Baaklini. “Acoustic Imaging of Subtle
of Material Microstructure, Porosity Variations in Ceramics.”
Ultrasonic Factors and Fracture Materials Evaluation. Vol. 46, No. 10.
Toughness of a Two-Phase Titanium Columbus, OH: American Society for
Alloy.” Materials Evaluation. Vol. 41, Nondestructive Testing (September
No. 3. Columbus, OH: American 1988): p 1338-1343.
Society of Nondestructive Testing
(March 1982): p 309-314. 96. Briggs, G.A.D. and M.G. Smoekh.
“Acoustic Microscopy of Surface
86. Briggs, A. An Introduction to Scanning Cracks: Theory and Practice.” Solid
Acoustic Microscopy. London, United Mechanics Research for Quantitative
Kingdom: Oxford University Press, Non-Destructive Evaluation [Evanston,
for the Royal Microscopical Society IL, September 1985]. Dordrecht,
(1986). Netherlands: Martinus Nijhoff
(1987): p 155-169.
87. Nikoonahad, M. “Reflection Acoustic
Microscopy for Industrial NDT.” 97. Rice, R. “Microstructure Dependence
Research Techniques in Nondestructive of Mechanical Behavior of
Testing. Vol. 7. London, United Ceramics.” Treatise on Materials
Kingdom: Academic Press (1984): Science and Technology. Vol. 11. New
p 217-257. York, NY: Academic Press (1977):
p 199-381.
88. Birnbaum, G. and G. White. “Laser
Techniques in Nondestructive 98. Mason, W.P. Physical Acoustics and the
Evaluation.” Research Techniques in Properties of Solids. Princeton, NJ: D.
Nondestructive Testing. Vol. 7. Van Nostrand Company (1958).
London, United Kingdom: Academic
Press (1984): p 259-365. 99. Klima, S. and G.Y. Baaklini.
“Ultrasonic Characterization of
89. Yuhas, D. and M. Oravecz. Structural Ceramics.” Materials
“Microstructure Characterization of Analysis by Ultrasonics. Park Ridge,
Titanium by Acoustic Microscopy.” NJ: Noyes Data Corporation (1987):
Nondestructive Methods for Material p 112-121.
Property Determination [Hershey, PA,
April 1983]. New York, NY: Plenum 100. Varshini, Y.P. Physical Review. Vol. 2.
(1984): p 259-270. Melville, NY: American Institute of
Physics (1970): p 3952-3955.
90. Rosenscwaig, A. Photoacoustics and
Photoacoustic Microscopy. New York, 101. Ledbetter, H. and D. Read. “Low
NY: Wiley (1980). Temperature Elastic Properties of a
300-Grade Maraging Steel.”
91. Rosenscwaig, A. “Thermal-Wave Metallurgical Transactions. Vol. 8A.
Imaging in a Scanning Electron Materials Park, OH: ASM
Microscope.” International Advances in International (1977): p 1805-1808.
Nondestructive Testing. Vol. 10. New
York, NY: Gordon and Breach (1985): 102. Monchalin, J.-P., R. Héon,
p 105-174. J.F. Bussière and B. Farahbakhsh.
“Laser Ultrasonic Determination of
92. Ringermacher, H. and C. Kitteredge. Elastic Constants at Ambient and
“Photoacoustic Microscopy of Elevated Temperatures.”
Ceramics.” Review of Progress in Nondestructive Characterization of
Quantitative Nondestructive Evaluation. Materials II [Montreal, Canada, July
Vol. 68. New York, NY: Plenum 1986]. New York, NY: Plenum (1987):
(1987): p 1231-1240. p 717-723.

93. Thomas, R.L., L.D. Favro and P.K. 103. Allen, D., W. Cooper, C. Sayers and
Kuo. “Thermal Wave Imaging for M. Silk. “The Use of Ultrasonics to
Quantitative Non-Destructive Measure Residual Stress.” Research
Evaluation.” Solid Mechanics Research Techniques in Nondestructive Testing.
for Quantitative Non-Destructive Vol. 6. London, United Kingdom:
Evaluation [Evanston, IL, Academic Press (1982): p 151-209.
September 1985]. Dordrecht,
Netherlands: Martinus Nijhoff 104. Pao, Y., W. Sachse and H. Fukuoka.
(1987): p 239-253. “Acoustoelasticity and Ultrasonic
Measurement of Residual Stress.”
Physical Acoustics. Vol. 17. Saint
Louis, MO: Elsevier/Academic Press
(1984): p 61-143.

Ultrasonic Characterization of Material Properties 347

105. Pao, Y.-H. “Theory of 114. Ledbetter, H. “Stainless Steel Elastic
Acoustoelasticity and Constants at Low Temperatures.”
Acoustoplasticity.” Solid Mechanics Journal of Applied Physics. Vol. 52,
Research for Quantitative No. 3. Melville, NY: American
Non-Destructive Evaluation [Evanston, Institute of Physics (1981):
IL, September 1985]. Dordrecht, p 1587-1589.
Netherlands: Martinus Nijhoff
(1987): p 257-273. 115. Lynnworth, L., E.P. Papadakis and
K. Fowler. “Ultrasonic Propagation
106. Droney, B.E. “Use of Ultrasonic Measurements and Applications.”
Techniques to Assess the Mechanical International Advances in
Properties of Steels.” Nondestructive Nondestructive Testing. Vol. 5. New
Methods for Material Property York, NY: Gordon and Breach
Determination [Hershey, PA, (1977): p 71-115.
April 1983]. New York, NY: Plenum
(1984): p 237-248. 116. Tittmann, B., G. Alers and
L. Graham. “Use of the Impulse
107. Hsu, N. “Acoustic Birefringence and Technique for Rapid Texture
Use of Ultrasonic Waves for Evaluation in Commercial Tube and
Experimental Stress Analysis.” Plate Materials.” Metallurgical
Experimental Mechanics. Vol. 14, Transactions. Vol. 7A. Materials Park,
No. 5. Bethel, CT: Society for OH: ASM International (1976):
Experimental Mechanics (1974): p 229.
p 169-176.
117. Schneider, E., S. Chu and K. Salama.
108. Okada, K. “Stress-Acoustic Relations “Influence of Texture on the
for Stress Measurements by Variations of Temperature
Ultrasonic Techniques.” Journal of the Dependence of Ultrasonic Wave
Acoustical Society of Japan. Vol. 1, Velocities with Stress.” Proceedings of
No. 3. Tokyo, Japan: Acoustical Ultrasonics International. Guildford,
Society of Japan (1980): p 193-200. Surrey, United Kingdom: Butterworth
Scientific (1985): p 133-138.
109. Mignogna, R.B., A.V. Clark, B.B. Rath
and C.L. Vold. “Effects of Rolled 118. Kupperman, D.S. “Analytical
Plate Thickness on Anisotropy with Ultrasonics for Structural Materials.”
Applications to Acoustic Stress Materials Analysis by Ultrasonics.
Measurement.” Nondestructive Park Ridge, NJ: Noyes Data
Methods for Material Property Corporation (1987): p 99-111.
Determination [Hershey, PA,
April 1983]. New York, NY: 119. Chandrasekaran, N. and K. Salama.
Plenum (1984): p 339-351. “Relationship between Stress and
Temperature Dependence of
110. Salama, K. and C.K. Ling. “The Effect Ultrasonic Shear Velocity.”
of Stress on the Temperature Nondestructive Methods for Material
Dependence of Ultrasonic Velocity.” Property Determination [Hershey, PA,
Journal of Applied Physics. Vol. 51. April 1983]. New York, NY: Plenum
Melville, NY: American Institute of (1984): p 393-403.
Physics (1980): p 1505.
120. Thompson, R., J. Smith and S. Lee.
111. Kwun, H. “Measurement of Stress in “Inference of Stress and Texture from
Steels Using Magnetically Induced Angular Dependence of Ultrasonic
Velocity Changes for Ultrasonic Plate Mode Velocities.” Materials
Waves.” Nondestructive Analysis by Ultrasonics. Park Ridge,
Characterization of Materials II NJ: Noyes Data Corporation (1987):
[Montreal, Canada, July 1986]. p 164-175.
New York, NY: Plenum (1987):
p 633-642. 121. Heyman, J. and M. Namkung.
“Residual Stress Measurements in
112. Bell, J. and J. Chen. “Pulse-Echo Carbon Steel.” Materials Evaluation by
Method of Determining the Elastic Ultrasonics. Park Ridge, NJ: Noyes
Constants of Rectangular Strips and Data Corporation (1987): p 61-74.
Square Plates.” NDT International.
Vol. 14, No. 6. Guildford, Surrey, 122. Namkung, M., D. Utrata,
United Kingdom: Butterworth J.S. Heyman and S.G. Allison.
Scientific (1981): p 325-327. “Low-Field Magnetoacoustic Residual
Stress Measurements in Steel.” Solid
113. Blessing, G.V., N.N. Hsu and Mechanics Research for Quantitative
T.M. Proctor. “Ultrasonic Shear Wave Non-Destructive Evaluation [Evanston,
Measurements of Known Residual IL, September 1985]. Dordrecht,
Stress in Aluminum.” Nondestructive Netherlands: Martinus Nijhoff
Methods for Material Property (1987): p 301-318.
Determination [Hershey, PA,
April 1983]. New York, NY:
Plenum (1984): p 353-363.

348 Ultrasonic Testing

123. Husson, D., S.D. Bennett and 132. Vary, A. “Concepts for Interrelating
G.S. Kino. “Rayleigh Wave Ultrasonic Attenuation,
Measurement of Surface Stresses in Microstructure and Fracture
Stainless Steel Piping.” Nondestructive Toughness in Polycrystalline Solids.”
Methods for Material Property Materials Evaluation. Vol. 46, No. 5.
Determination [Hershey, PA, April Columbus, OH: American Society for
1983]. New York, NY: Plenum (1984): Nondestructive Testing (April 1988):
p 365-375. p 642-649. Erratum, Vol. 46, No. 8
(July 1988): p 1118.
124. Egle, D.M. and D. Bray.
“Measurement of Acousto-Elastic and 133. Mason, W. and H. McSkimmin.
Third Order Elastic Constants for Rail “Attenuation and Scattering of High
Steel.” Journal of the Acoustical Society Frequency Sound Waves in Metals
of America. Vol. 60, No. 3. Melville, and Glasses.” Journal of the Acoustical
NY: American Institute of Physics, for Society of America. Vol. 19, No. 3.
the Acoustical Society of America Melville, NY: American Institute of
(1976): p 741-744. Physics, for the Acoustical Society of
America (1947): p 464-473.
125. Fukuoka, H. “Ultrasonic
Measurement of Residual Stress.” 134. Mason, W. and H. McSkimmin.
Solid Mechanics Research for Journal of Applied Physics. Vol. 19,
Quantitative Non-Destructive No. 10. Melville, NY: American
Evaluation [Evanston, IL, September Institute of Physics (1948):
1985]. Dordrecht, Netherlands: p 940-946.
Martinus Nijhoff (1987): p 275-299.
135. Lifsitz, I. and G. Parkomovskii.
126. Allen, D. and W. Cooper. “A Fourier Zhurnal Eksperimental’noi i
Transform Technique that Measures Teoreticheskoi Fiziki. Vol. 20. Moscow,
Phase Delays of Ultrasonic Pulses Russia: Izdatel’stvo Nauka. Melville,
with Accuracy for Determining NY: American Institute of Physics
Residual Stress in Metals.” NDT (1950): p 175-182.
International. Vol. 16, No. 4.
Guildford, Surrey, United Kingdom: 136. Merkulov, L. “Investigation of
Butterworth Scientific (1983): Ultrasonic Scattering in Metals.”
p 205-217. Soviet Physics. Technical Physics.
Vol. 1, No. 1. Melville, NY: American
127. Mott, G. and M. Tsao. Institute of Physics (1957): p 59-69.
“Acoustoelastic Effects in Two
Structural Steels.” Nondestructive 137. Kolsky, H. Stress Waves in Solids. New
Methods for Material Property York, NY: Dover (1963).
Determination [Hershey, PA,
April 1983]. New York, NY: 138. Lücke, K. “Ultrasonic Attenuation
Plenum (1984): p 377-392. Caused by Thermoelastic Heat Flow.”
Journal of Applied Physics. Vol. 27.
128. Schuckler, F. “Grain Size.” Melville, NY: American Institute of
Quantitative Microscopy. New York, Physics (1956): p 1433-1438.
NY: McGraw-Hill (1968): p 201-265.
139. Auld, B. Acoustic Fields and Waves in
129. DeHoff, R. “The Statistical Solids. Vol. 1. New York, NY: John
Background of Qualitative Wiley and Sons (1973).
Metallography.” Quantitative
Microscopy. New York, NY: McGraw 140. Granato, A.V. and K. Lücke. “Theory
Hill (1968): p 11-44. of Mechanical Damping Due to
Dislocations.” Journal of Applied
130. Papadakis, E.P. “Ultrasonic Physics. Vol. 27, No. 6. Melville, NY:
Diffraction from Single Apertures American Institute of Physics (1956):
with Application to Pulse p 583-593.
Measurements and Crystal Physics.”
Physical Acoustics: Principles and 141. Seeger, A. and P. Schiller. “The
Methods. Vol. 11. Saint Louis, MO: Formation and Diffusion of Kinks As
Elsevier/Academic Press (1975): the Fundamental Process of
p 152-211. Dislocation Movement in Internal
Movement Friction Measurements.”
131. Green, R.E. “Effect of Metallic Acta Metallurgica. Vol. 10. Elmsford,
Microstructure on Ultrasonic NY: Pergamon Press (1962):
Attenuation.” Nondestructive p 348-357.
Evaluation: Microstructural
Characterization and Reliability 142. Langlois, P. and J.F. Bussière.
Strategies. Warrensdale, PA: “Magnetoelastic Contribution to
Metallurgical Society of the American Ultrasonic Attenuation in Structural
Institute of Mechanical Engineers Steels.” Nondestructive Characterization
(1981): p 115-132. of Materials II [Montreal, Canada,
July 1986]. New York, NY: Plenum
(1987): p 291-298.

Ultrasonic Characterization of Material Properties 349

143. Winkler, K. and W. Murphy. 153. Serabian, S. “Frequency and Grain
“Scattering in Glass Beads: Effects of Size Dependency of Ultrasonic
Frame and Pore Fluid Attenuation in Polycrystalline
Compressabilities.” Journal of the Materials.” British Journal of
Acoustical Society of America. Vol. 76, Non-Destructive Testing. Vol. 22,
No. 3. Melville, NY: American No. 2. Northampton, United
Institute of Physics, for the Kingdom: British Institute of
Acoustical Society of America (1984): Non-Destructive Testing (1980):
p 820-825. p 69-77.

144. Papadakis, E.P. “Ultrasonic 154. Kopec, B. and V. Hanak. “Use of
Attenuation Caused by Rayleigh Ultrasonic Attenuation to Investigate
Scattering by Graphite Nodules in Anomalies in the Structure of
Nodular Cast Iron.” Journal of the Railway Axles.” NDT International.
Acoustical Society of America. Vol. 70, Vol. 17, No. 5. Saint Louis, MO:
No. 3. Melville, NY: American Elsevier (1984): p 265-268.
Institute of Physics, for the
Acoustical Society of America (1981): 155. Bussière, J.F. “Application of
p 782-787. Nondestructive Evaluation to
Processing of Metals.” Review of
145. Sayers, C. “Scattering of Ultrasound Progress in Quantitative Nondestructive
by Minority Phases in Polycrystalline Evaluation. Vol. 6B. New York, NY:
Metals.” Wave Motion. Vol. 7. Plenum (1987): p 1377-1393.
Amsterdam, Netherlands: Elsevier
Science Publications (1985): 156. Klinman, R., G. Webster, F. Marsh
p 95-104. and E. Stephenson. “Ultrasonic
Prediction of Grain Size, Strength
146. Evans, A., B. Tittmann, L. Ahlberg, and Toughness in Plain Carbon
B. Khuri-Yakub and G. Kino. Steel.” Materials Evaluation. Vol. 38,
“Ultrasonic Attenuation in No. 10. Columbus, OH: American
Ceramics.” Journal of Applied Physics. Society for Nondestructive Testing
Vol. 49, No. 5. Melville, NY: (September 1980): p 26-32.
American Institute of Physics, for the
Acoustical Society of America (1978): 157. Yada, H. and K. Kawashima.
p 2669-2679. “Important Metallurgical Parameters
That Must be Determined to Control
147. Perkeris, C. “Note on the Scattering the Properties of Steels during
of Radiation in an Inhomogeneous Processing.” Nondestructive
Medium.” Physical Review. Vol. 71, Characterization of Materials II
No. 4. Melville, NY: American [Montreal, Canada, July 1986]. New
Institute of Physics (1947): York, NY: Plenum (1987): p 195-209.
p 268-270.
158. Klinman, R. and E. Stephenson.
148. Roney, R. The Influence of Metal Grain “Ultrasonic Prediction of Grain Size
Structure on the Attenuation of an and Mechanical Properties in Plain
Ultrasonic Acoustic Wave. Dissertation. Carbon Steel.” Materials Evaluation.
Pasadena, CA: California Institute of Vol. 39, No. 12. Columbus, OH:
Technology (1950). American Society for Nondestructive
Testing (November 1981):
149. Serabian, S. and R.S. Williams. p 1116-1120.
“Experimental Determination of
Ultrasonic Attenuation 159. Generazio, E.R. “Ultrasonic
Characteristics Using the Roney Verification of Microstructural
Generalized Theory.” Materials Changes Due to Heat Treatment.”
Evaluation. Vol. 36, No. 8. Columbus, Materials Analysis by Ultrasonics. Park
OH: American Society for Ridge, NJ: Noyes Data Corporation
Nondestructive Testing (July 1978): (1987): p 200-210.
p 55-62.
160. Goebbels, K. and S. Hirsekorn.
150. Smith, R. “The Effect of Grain Size “A New Method for Stress
Distribution on the Frequency Determination in Textured
Dependence of the Ultrasonic Materials.” NDT International.
Attenuation in Polycrystalline Vol. 17, No. 6. Saint Louis, MO:
Materials.” Ultrasonics. Vol. 22, No. 9. Elsevier (1984): p 337-341.
Saint Louis, MO: Elsevier (1982):
p 211-214. 161. Hosten, B., M. Deschamps and
B. Tittmann. “Inhomogeneous Wave
151. Bozorg-Grayeli, N. Acoustic Generation and Propagation in Lossy
Nondestructive Evaluation of Anisotropic Solids — Composites
Microstructure. Dissertation. Stanford, Characterization.” Journal of the
CA: Stanford University (1981). Acoustical Society of America. Melville,
NY: American Institute of Physics, for
152. Vary, A. and D. Hull. TM-83358, the Acoustical Society of America
Ultrasonic Ranking of Toughness of (1986-1987).
Tungsten Carbide. Washington, DC:
National Aeronautics and Space
Administration (1983).

350 Ultrasonic Testing

162. Sayers, C. “Ultrasonic Determination 171. Schaller, R., J.J. Ammann and
of Texture and Residual Stress in P. Millet. “Mechanical Properties of
Polycrystalline Metals.” Solid Composite Materials by Internal
Mechanics Research for Quantitative Friction.” Nondestructive
Non-Destructive Evaluation [Evanston, Characterization of Materials II
IL, September 1985]. Dordrecht, [Montreal, Canada, July 1986]. New
Netherlands: Martinus Nijhoff York, NY: Plenum (1987): p 345-353.
(1987): p 319-333.
172. Petch, N. “The Cleavage Strength of
163. Mignogna, R., P.P. Delsanto, Polycrystals.” Journal of the Iron and
A.V. Clark, B.B. Rath and C.L. Vold. Steel Institute. Vol. 174. Materials
“Ultrasonic Measurements on Park, OH: ASM International (1953):
Textured Materials.” Nondestructive p 25.
Characterization of Materials II
[Montreal, Canada, July 1986]. New 173. Hall, E. “Deformation and Aging of
York, NY: Plenum (1987): p 545-553. Mild Steel III: Discussion and
Results.” Proceedings of the Physical
164. Kupperman, D.S., K.J. Reimann and Society. Vol. 64. London, United
J. Abrego-Lopez. “Ultrasonic NDE of Kingdom: Physical Society (1951):
Cast Stainless Steel.” NDT p 747.
International. Vol. 20, No. 3. Saint
Louis, MO: Elsevier (June 1987): 174. Hall, E. Yield Point Phenomena in
p 145-152. Metals and Alloys. New York, NY:
Plenum (1970).
165. Generazio, E.R. “Ultrasonic
Determination of Recrystallization.” 175. Armstrong, R. “The Influence of
Review of Progress in Quantitative Polycrystalline Grain Size on Several
Nondestructive Evaluation. Vol. 6B. Mechanical Properties.” Metallurgical
New York, NY: Plenum (1987): Transactions. Materials Park, OH:
p 1465-1475. ASM International (1970):
p 1169-1176.
166. Friant, C.L. and M. Rosen.
“Ultrasonic Materials 176. Armstrong, R. “The Influence of
Characterization of Melt Spun Polycrystal Grain Size on Mechanical
Metallic Ribbons.” Nondestructive Properties.” Advances in Materials
Methods for Material Property Research. Vol. 4. New York, NY: Wiley
Determination [Hershey, PA, (1970): p 101-146.
April 1983]. New York, NY:
Plenum (1984): p 301-314. 177. Smith, R. and W. Reynolds. “The
Correlation of Ultrasonic
167. Piché, L. “Application of Ultrasonics Attenuation, Microstructure and
to the Characterization of Ductile to Brittle Transition
Composites: A Method for the Temperature in Low Carbon Steels.”
Determination of Polyethylene Journal of Materials Science. Vol. 17.
Density.” Nondestructive London, United Kingdom: Chapman
Characterization of Materials II and Hall (1982): p 1420-1426.
[Montreal, Canada, July 1986]. New
York, NY: Plenum (1987): p 79-87. 178. Smith, R.L., K.L. Rusbridge,
W.N. Reynolds and B. Hudson.
168. Vary, A., E.R. Generazio, D.J. Roth “Ultrasonic Attenuation,
and G.Y. Baaklini. “Ultrasonic NDE Microstructure and Ductile to Brittle
of Structural Ceramics for Power and Transition Temperature in Fe-C
Propulsion Systems.” Non-Destructive Alloys.” Materials Evaluation. Vol. 41,
Testing: Proceedings of the Fourth No. 2. Columbus, OH: American
European Conference [London, United Society for Nondestructive Testing
Kingdom, September 1987]. Vol. 2. (February 1983): p 219-222.
Oxford, United Kingdom: Pergamon
Press (1988): p 1299-1307. 179. Smith, R. “Materials Characterization
by Ultrasonic Attenuation Spectral
169. Adler, L. and S. Wang. “Ultrasonic Analysis.” Review of Progress in
Measurement of Porosity in Casts Quantitative Nondestructive Evaluation.
and Welds.” Materials Analysis by Vol. 6B. New York, NY: Plenum
Ultrasonics. Park Ridge, NJ: Noyes (1987): p 1475-1483.
Data Corporation (1987): p 72-78.
180. Alford, N., K. Kendall, W. Clegg and
170. Tittmann, B. “Ultrasonic J. Birchall. “Strength/Microstructure
Measurements for the Prediction of Relation in Alumina and Titania.”
Mechanical Strength.” NDT Advanced Ceramic Materials. Vol. 3,
International. Vol. 11. Guildford, No. 2. Westerville, OH: American
Surrey, United Kingdom: Butterworth Ceramic Society (1988): p 113-117.
Scientific Limited (1978): p 17.
181. Shabel, B.S. and R.F. Young. “A New
Procedure for Rapid Determination
of Yield and Tensile Strength from
Hardness Tests.” Nondestructive
Characterization of Materials II
[Montreal, Canada, July 1986]. New
York, NY: Plenum (1987): p 335-343.

Ultrasonic Characterization of Material Properties 351

182. Rivenez, J., A. Lambert and 192. Duke, J.C., Jr., E.G. Henneke, II, and
C. Flambard. “Nondestructive W. Stinchcomb. CR-3976, Ultrasonic
Determination of Hardening Depths Stress Wave Characterization of
with Ultrasonic Surface Waves.” Composite Materials. Washington, DC:
Nondestructive Characterization of National Aeronautics and Space
Materials II [Montreal, Canada, Administration (1986).
July 1986]. New York, NY: Plenum
(1987): p 373-380. 193. Williams, J.H. and N. Lampert.
“Ultrasonic Evaluation of
183. Vary, A. “Ultrasonic Nondestructive Impact-Damaged Graphite Fiber
Evaluation, Microstructure and Composite.” Materials Evaluation.
Fracture Toughness Interrelations.” Vol. 38, No. 12. Columbus, OH:
Solid Mechanics Research for American Society for Nondestructive
Quantitative Non-Destructive Testing (November 1980): p 68-72.
Evaluation [Evanston, IL,
September 1985]. Dordrecht, 194. Williams, J.H. and S.S. Lee. “Pattern
Netherlands: Martinus Nijhoff Recognition Characterization of
(1987): p 135-152. Micromechanical and Morphological
Materials States via Analytical
184. Vary, A. “Correlations between Quantitative Ultrasonics.” Materials
Ultrasonic and Fracture Toughness Analysis by Ultrasonics. Park Ridge,
Factors in Metallic Materials.” NJ: Noyes Data Corporation (1987):
Fracture Mechanics. Special Technical p 187-199.
Publication 677. West
Conshohocken, PA: ASTM 195. Tanary, S. Characterization of
International (1979): p 563-578. Adhesively Bonded Joints Using
Acousto-Ultrasonics. Thesis. Ottawa,
185. Fu, L. “Mechanical Aspects of NDE Province of Quebec, Canada:
by Sound and Ultra-Sound.” Applied University of Ottawa (1988).
Mechanics Reviews. Vol. 55, No. 8.
New York, NY” ASME International 196. Duke, J.C., Jr., E.G. Henneke, II,
(1982): p 1047-1057. W. Stinchcomb and K. Reifsnider.
“Characterization of Composite
186. Fu, L. “On Ultrasonic Factors and Materials by Means of the Ultrasonic
Fracture Toughness.” Engineering Stress Wave Factor.” Proceedings of the
Fracture Mechanics. Vol. 18, No. 1. Second International Conference on
New York, NY: Pergamon (1983): Composite Structures. London, United
p 59-67. Kingdom: Applied Science Publishers
(1984): p 53-60.
187. Fu, L. “Micromechanics and Its
Application to Fracture and NDE.” 197. Russell-Floyd, R. and M.G. Phillips.
Developments in Mechanics. Vol. 12. “A Critical Assessment of
Iowa City, IA: University of Iowa Acousto-Ultrasonics As a Method of
Press (1983): p 263-265. Nondestructive Examination.” NDT
International. Vol. 21, No. 4.
188. Hahn, G., M. Kanninen and Guildford, Surrey, United Kingdom:
A. Rosenfeld. Annual Reviews of Butterworth Scientific (1988):
Materials Science. Vol. 2. Palo Alto, p 247-257.
CA: Annual Reviews (1972):
p 381-404. 198. Thompson, R.B. “Determination of
Texture and Grain Size in Metals: An
189. Canella, G. and M. Taddei. Example of Materials
“Correlation Between Ultrasonic Characterization.” Topics on
Attenuation and Fracture Toughness Nondestructive Evaluation: Vol. 1,
of Steels.” Nondestructive Sensing for Materials Characterization,
Characterization of Materials II Processing, and Manufacturing.
[Montreal, Canada, July 1986]. New Columbus, OH: American Society for
York, NY: Plenum (1987): p 261-269. Nondestructive Testing (1998):
p 23-45.
190. Govada, A., J.C. Duke, Jr.,
E.G. Henneke, II, and 199. Rohklin, S.I. and A.I. Lavrentyev.
W. Stinchcomb. CR-174870, A Study “Ultrasonic Characterization of Thin
of the Stress Wave Factor Technique for Surface and Interphase Layers.”
the Characterization of Composite Topics on Nondestructive Evaluation:
Materials. Washington, DC: National Vol. 1, Sensing for Materials
Aeronautics and Space Characterization, Processing, and
Administration (1985). Manufacturing. Columbus, OH:
American Society for Nondestructive
191. Govada, A., E.G. Henneke, II, and Testing (1998): p 47-83.
R. Talreja. “Acousto-Ultrasonic
Measurements to Monitor Damage
during Fatigue of Composites.”
Advances in Aerospace Sciences and
Engineering [New Orleans, LA,
December 1984]. New York, NY:
ASME International (1984): p 55-60.

352 Ultrasonic Testing

200. Hsu, D.K. “Ultrasonic Sensors for Bibliography
Robotics and Field Operation.” Topics
on Nondestructive Evaluation: Vol. 4, Achenbach, J. and Y. Rajapakse. Solid
Automation, Miniature Robotics, and Mechanics Research for Quantitative
Sensors for Nondestructive Testing and Non-Destructive Evaluation [Evanston,
Evaluation. Columbus, OH: American IL, September 1985]. Dordrecht,
Society for Nondestructive Testing Netherlands: Martinus Nijhoff (1987).
(2000): p 165-173.
Allen, D.R. and C.M. Sayers. “The
201. Sherrit, S., Y. Bar-Cohen and X. Bao. Influence of Stress on the Principal
“Ultrasonic Materials, Actuators, and Polarisation Directions of Ultrasonic
Motors (USM).” Topics on Shear Waves in Textured Steel Plates.”
Nondestructive Evaluation: Vol. 4, Journal of Physics D: Applied Physics.
Automation, Miniature Robotics, and Vol. 17, No. 7. Melville, NY: American
Sensors for Nondestructive Testing and Institute of Physics (1984): p 215-222.
Evaluation. Columbus, OH: American
Society for Nondestructive Testing Allen, D.R. and C.M. Sayers. “Ultrasonic
(2000): p 215-231. SH Waves in Textured Aluminum
Plates.” Ultrasonics. Vol. 23. Guildford,
202. Rogers, W.P. “Elastic Property Surrey: IPC Science and Technology
Measurement Using Rayleigh-Lamb Press (1984): p 215-222.
Waves.” Research in Nondestructive
Evaluation. Vol. 6. Columbus, OH: AMD, Vol 234: ASME International
American Society for Nondestructive Mechanical Engineering Congress and
Testing (1995): p 185-208. Exposition [Nashville, TN, November
1999]. NDE, Vol. 17: On the Recent
203. Roth, D.J., J.D. Kiser, S.M. Swickard, Advances of the Ultrasonic Evaluation
S.A. Szatmary and D.P. Kerwin. and Composite Material Characterization.
“Quantitative Mapping of Pore New York, NY: ASME International
Fraction Variations in Silicon Nitride (1999).
Using an Ultrasonic Contact Scan
Technique.” Research in Nondestructive ASTM E 1736, Standard Practice for
Evaluation. Vol. 6. Columbus, OH: Acousto-Ultrasonic Assessment of
American Society for Nondestructive Filament-Wound Pressure Vessels. West
Testing (1995): p 125-168. Conshohocken, PA: ASTM
International (2005).
204. Kawamoto, S., J.H. Muehl and
R.S. Williams. “Use of Bar-Cohen, Y. and D.E. Chimenti.
Acousto-Ultrasonic Techniques to “Detection of Porosity in Composite
Determine Properties of Laminates by Leaky Lamb Waves.”
Remanufactured Particle Boards 11th World Conference on Nondestructive
Made Solely from Recycled Particles.” Testing. Vol. 3. Columbus, OH.
Third International Workshop on Green American Society for Nondestructive
Composites [Kyoto, Japan, March Testing (1985): p 1661-1668.
20005]. Tokyo, Japan: Japanese
Society for Nondestructive Inspection Betz, D.C., G. Thursby, B. Culshaw and
(2005): p 184-189. W.J. Staszewski. “Acousto-Ultrasonic
Sensing Using Fiber Bragg Gratings.”
205. Kautz, H.E. NASA TM-1998-208410, Smart Material Structures. Vol. 12, No. 1.
Noncontact Acousto-Ultrasonics for Melville, NY: American Institute of
Material Characterization. Physics (February 2003): p 122-128.
Washington, DC: National
Aeronautics and Space Bhatia, A.B. Ultrasonic Absorption: An
Administration (1998). Introduction to the Theory of Sound
Absorption and Dispersion in Gases,
206. Kautz, H.E. NASA CR-2002-211881, Liquids, and Solids. New York, NY:
Acousto-Ultrasonics to Assess Material Dover (1967, 1985).
and Structural Properties. Washington,
DC: National Aeronautics and Space Chang, M., R. Chang and C. Shu. “The
Administration (2002). Application of Ultrasonic Attenuation
Measurements to Estimate Corrosion
207. Finlayson, R.D., M. Friesel, M. Carlos, in Pipeline at Support Structures.”
P. Cole and J.C. Lenain. “Health Materials Evaluation. Vol. 60, No. 5.
Monitoring of Aerospace Structures Columbus, OH: American Society for
with Acoustic Emission and Nondestructive Testing (May 2002):
Acousto-Ultrasonics.” Insight. Vol. 43, p 631-634.
No. 3. Northampton, United
Kingdom: British Institute of Chen, C.-H. Nonlinear Maximum Entropy
Non-Destructive Testing Spectral Analysis for Signal Recognition.
(March 2001): p 155-158. New York, NY: Research Studies Press
(1982).

Ultrasonic Characterization of Material Properties 353

Dos Reis, H. “Acousto-Ultrasonic Hosten, B., L. Alberg, B. Tittmann and
Non-Destructive Evaluation of Wire B. Springarn. “Ultrasonic
Rope Using the Stress Wave Factor Characterization of Diffusion Bonds.”
Technique.” British Journal of Review of Progress in Quantitative
Non-Destructive Testing. Vol. 28, No. 3. Nondestructive Evaluation. Vol. 6B. New
Northampton, United Kingdom: York, NY: Plenum (1987): p 1701-1706.
British Institute of Non-Destructive
Testing (1986): p 155-156. Kline, R.A. and D. Hashemi. “Ultrasonic
Guided-Wave Monitoring of Fatigue
Dos Reis, H., L. Bergman and J. Bucksbee. Damage Development in Bonded
“Adhesive Bond Strength Quality Joints.” Materials Evaluation. Vol. 45,
Assurance Using the No. 9. Columbus, OH: American
Acousto-Ultrasonic Technique.” British Society for Nondestructive Testing
Journal of Non-Destructive Testing. (September 1987): p 1076-1082.
Vol. 28, No. 6. Northampton, United
Kingdom: British Institute of Kumar, A., T. Jayakumar, B. Raj and
Non-Destructive Testing (1986): K.K. Ray. “Ultrasonic Time Domain
p 357-358. Technique for Frequency Dependent
Attenuation Measurement for
Ganesan, V., P. Palanichamy and B. Raj. Microstructural Characterization of a
“Elastic Modulus Determination as a Titanium Alloy.” Materials Evaluation.
Potential Tool for Predicting Vol. 61, No. 12. Columbus, OH:
Embrittlement in a Ferritic Steel.” American Society for Nondestructive
Materials Evaluation. Vol. 62, No. 2. Testing (December 2003):
Columbus, OH: American Society for p 1321-1326.
Nondestructive Testing
(February 2004): p 137-142. Kumar, A., V. Shankar, T. Jayakumar,
G. Srinivasan and B. Raj.
Hahn, H.T. “Application of Ultrasonic “Nondestructive Measurement of
Technique to Cure Characterization of Coating Thickness Using Through
Epoxies.” Nondestructive Methods for Thickness Ultrasonic Velocity
Material Property Determination Measurements.” Materials Evaluation.
[Hershey, PA, April 1983]. New York, Vol. 60, No. 6. Columbus, OH:
NY: Plenum (1984): p 315-326. American Society for Nondestructive
Testing (June 2002): p 791-794.
Haines, N., J. Bell and P. McIntyre. “The
Application of Broadband Ultrasonic Kutty, T., K. Chandrasekharan,
Spectroscopy to the Study of Layered J. Panakkal, S. Ghosal and P. De. “Use
Media.” Journal of the Acoustical Society of Ultrasonic Velocity for
of America. Vol. 64, No. 6. Melville, Nondestructive Evaluation of Ferrite
NY: American Institute of Physics, for Content in Duplex Stainless Steel.”
the Acoustical Society of America NDT International. Vol. 20, No. 6.
(1978): p 1645-1651. Guildford, Surrey, United Kingdom:
Butterworth Scientific Limited (1987):
Hemann, J.H., P. Cavano, H.E. Kautz and p 359-361.
K. Bowles. “Trans-Ply Crack Density
Detection by Acousto-Ultrasonics.” Ledbetter, H. “Texture in Stainless Steel
Acousto-Ultrasonics: Theory and Welds: an Ultrasonic Study.” Journal of
Applications [Blacksburg, VA, July Materials Science. Vol. 20. London,
1987]. New York, NY: Plenum (1988): United Kingdom: Chapman and Hall
p 319-325. (1985): p 1720-1724.

Heyman, J., S. Allison and K. Salama. “The Lin, L., X.M. Li and J.S. Zhang.
Effect of Carbon Concentration and “Nondestructive Differentiation of
Plastic Deformation on Ultrasonic Three Transformation Products in Low
Higher Order Elastic Constants.” Alloy Steel Using Two Ultrasonic
Ultrasonics International Conference Methods.” Materials Evaluation.
Proceedings. Surrey, United Kingdom: Vol. 61. No. 4. Columbus, OH:
Butterworth Scientific Limited (1985): American Society for Nondestructive
p 786-791. Testing (April 2003): p 512-516.

Hinrichs, R. and J. Thuen. Control System MacCrone, R., ed. Treatise on Materials
for Processing Composite Materials. Science and Technology. Vol. 11. Saint
United States Patent 4 455 268 (1984). Louis, MO: Elsevier/Academic Press
(1977).
Höller, P. “Nondestructive Analysis of
Structure and Stresses by Ultrasonic Massines, F. and L. Piché. “Ultrasonic
and Magnetic Methods.” Nondestructive Characterization of Polymers in Their
Characterization of Materials II Evolution from Solid to Liquid State.”
[Montreal, Canada, July 1986]. New Nondestructive Characterization of
York, NY: Plenum (1987): p 211-225. Materials II [Montreal, Canada,
July 1986]. New York, NY: Plenum
(1987): p 49-60.

354 Ultrasonic Testing

Monchalin, J.-P. and J.F. Bussiére. Phani, K.K. and N.R. Bose. “Application of
“Measurement of Near-Surface Acousto-Ultrasonics for Predicting
Ultrasonic Absorption by Hygrothermal Degradation of
Thermo-Emissivity.” Nondestructive Unidirectional Glass-Fiber
Methods for Material Property Composites.” Acousto-Ultrasonics:
Determination [Hershey, PA, Theory and Applications [Blacksburg,
April 1983]. New York, NY: Plenum VA, July 1987]. New York, NY: Plenum
(1984): p 289-297. (1988): p 327-336.

Nondestructive Testing Handbook, second Piché, L. and A. Hamel. “Characterization
edition, Volume 5: Acoustic Emission of Isotropic Composites Containing
Testing. Columbus, OH: American Inclusions of Specific Shapes by
Society for Nondestructive Testing Ultrasonics.” Nondestructive
(1986). Characterization of Materials II
[Montreal, Canada, July 1986]. New
Palanichamy, P. and M. Vasudevan. York, NY: Plenum (1987): p 95-103.
“Ultrasonic Testing of Annealing
Behavior and Texture and Razvi, S., P. Li, K. Salama, J. Cantrell and
Determination of Texture Coefficients W. Yost. “Nondestructive
in Stainless Steel.” Materials Evaluation. Characterization of Aluminum
Vol. 61, No. 9. Columbus, OH: Alloys.” Review of Progress in
American Society for Nondestructive Quantitative Nondestructive Evaluation.
Testing (September 2003): Vol. 6B. New York, NY: Plenum (1987):
p 1020-1025. p 1403-1408.

Papadakis, E.P. “From Micrograph to Rebello, C. and J.C. Duke, Jr. “Factors
Grain-Size Distribution with Influencing the Ultrasonic Stress Wave
Ultrasonic Applications.” Journal of Factor Evaluation of Composite
Applied Physics. Vol. 35. Melville, NY: Material Structures.” Journal of
American Institute of Physics (1964): Composites Technology and Research.
p 1586-1594. Vol. 8, No. 1. West Conshohocken, PA:
ASTM International (1986): p 18-23.
Papadakis, E.P. “Revised Grain-Scattering
Formulas and Tables.” Journal of the Reed, R.W. and A.L. Bertram. “Ultrasonic
Acoustical Society of America. Vol. 37, Measurements of Elastic Moduli of
No. 4. Melville, NY: American Institute Thermally Cycled Metal Matrix
of Physics, for the Acoustical Society Composite Precursor Wires.”
of America (1965): p 703-710. Nondestructive Methods for Material
Property Determination [Hershey, PA,
Papadakis, E.P. “Ultrasonic Attenuation in April 1983]. New York, NY: Plenum
SAE 3140 and 4150 Steel.” Journal of (1984): p 327-336.
the Acoustical Society of America.
Vol. 32, No. 12. Melville, NY: Reynolds, W. and R. Smith. “Ultrasonic
American Institute of Physics, for the Wave Attenuation Spectra in Steels.”
Acoustical Society of America (1960): Journal of Physics D: Applied Physics.
p 1628-1639. Vol. 17, No. 1. Melville, NY: American
Institute of Physics (1984): p 109-116.
Papadakis, E.P. “Ultrasonic Attenuation
and Velocity in SAE 52100 Steel Rokhlin, S.I., A. Baltazar, B. Xie, J. Chen
Quenched from Various and R. Reuven. “Method for
Temperatures.” Metallurgical Monitoring Environmental
Transactions. Vol. 1. Materials Park, Degradation of Adhesive Bonds.”
OH: ASM International (1970): p 1053- Materials Evaluation. Vol. 59, No. 6.
1057. Columbus, OH: American Society for
Nondestructive Testing (June 2002):
Papadakis, E.P. “Ultrasonic Attenuation p 795-801.
and Velocity in Three Transformation
Products in Steel.” Journal of Applied Ruud, C.O. and R.E. Green, eds.
Physics. Vol. 35. Melville, NY: Nondestructive Methods for Material
American Institute of Physics (1964): Property Determination [Hershey, PA,
p 1474-1482. April 1983]. New York, NY: Plenum
(1984).
Papadakis, E.P. “Ultrasonic Attenuation
Caused by Scattering in Polycrystalline Scott, W. and P. Gordon. “Ultrasonic
Metals.” Journal of the Acoustical Society Spectrum Analysis for Nondestructive
of America. Vol. 37, No. 4. Melville, Testing of Layered Composite
NY: American Institute of Physics, for Materials.” Journal of the Acoustical
the Acoustical Society of America Society of America. Vol. 62, No. 1.
(1965): p 711-717. Melville, NY: American Institute of
Physics, for the Acoustical Society of
America (1977): p 108-116.

Ultrasonic Characterization of Material Properties 355

Shukla, S. and S. Yun. ”Ultrasonic
Attenuation in GaAs.” Journal of the
Acoustical Society of America. Vol. 70,
No. 6. Melville, NY: American Institute
of Physics, for the Acoustical Society
of America (1981): p 1713-1716.

Srivastava, V. and R. Prakash. “Fatigue
Prediction of Glass Fiber Reinforced
Plastics Using the Acousto-Ultrasonic
Technique.” International Journal of
Fatigue. Vol. 9, No. 3. Amsterdam,
Netherlands: Elsevier (1987):
p 175-178.

Srivastava, V. and R. Prakash. “Prediction
of Material Property Parameter of FRP
Composites Using Ultrasonic and
Acousto-Ultrasonic Techniques.”
Composite Structures. Vol. 8. Barking,
Essex, United Kingdom: Applied
Science (1987): p 311-321.

Vary, A. “Acousto-Ultrasonic
Characterization of Fiber Reinforced
Composites.” Materials Evaluation.
Vol. 40, No. 6. Columbus, OH:
American Society for Nondestructive
Testing (May 1982): p 650-654.

Vary, A., ed. Materials Analysis by
Ultrasonics — Metals, Ceramics,
Composites. Park Ridge, NJ: Noyes Data
Corporation (1987).

Willems, H. “Investigation of Creep
Damage in Alloy 800H Using
Ultrasonic Velocity Measurements.”
Nondestructive Characterization of
Materials II [Montreal, Canada,
July 1986]. New York, NY: Plenum
(1987): p 471-479.

Willems, H., W. Bendick and H. Weber.
“Nondestructive Evaluation of Creep
Damage in Service Exposed 14 MoV
6 3 Steel.” Nondestructive
Characterization of Materials II
[Montreal, Canada, July 1986]. New
York, NY: Plenum (1987): p 451-459.

Williams, J.H., H. Yuce and S.S. Lee.
“Ultrasonic and Mechanical
Characterization of Fatigue States of
Graphite Epoxy Composite
Laminates.” Materials Evaluation.
Vol. 40, No. 5. Columbus, OH:
American Society for Nondestructive
Testing (June 1982): p 560-565.

356 Ultrasonic Testing

9

CHAPTER

Ultrasonic Testing of
Advanced Materials1

Yoseph Bar-Cohen, Jet Propulsion Laboratory,
Pasadena, California
Ajit K. Mal, University of California at Los Angeles, Los
Angeles, California
Donald J. Roth, National Aeronautics and Space
Administration, Cleveland, Ohio

PART 1. Ultrasonic Testing of Advanced
Structural Ceramics

The high temperature thermal, focus can be maintained. For these
mechanical and physical properties of reasons, testing ceramics requires new
structural ceramics suit them for techniques to optimize available energy
applications such as advanced engines, and permit detection and characterization
enabling higher combustion temperatures of discontinuities as small as 20 µm
and therefore higher thermodynamic (0.0008 in.).
efficiencies. The relatively low fracture
toughness of structural ceramics and the In addition to the important task of
corresponding critical discontinuity size detecting critical discontinuities in
require significantly higher sensitivities structural ceramics, there is also a need to
than those of nondestructive test develop predictive capability for some of
techniques commonly used for detection the mechanical properties, based on easily
of discontinuities in metals. measured material characteristics. This has
been achieved for some metals, where a
Ultrasonic techniques for testing correlation has been found between the
ceramics use high frequency elastic waves fracture toughness and attenuation of
to probe both green state and sintered elastic waves.2
ceramic bodies. Discontinuities smaller
than 25 µm (0.001 in.) have been Green State Ceramics
detected in monolithic and composite
ceramics at depths of about 3 mm (0.1 in.) An intermediate step in the processing of
in alumina, silicon carbide and silicon structural ceramics is the fabrication of
nitride and at depths of about 5 mm green state bodies. At this first stage of
(0.2 in.) in zirconia. Described below are component shaping (such as through
procedures for the rapid determination of powder compaction), uncontrolled
attenuation as a function of frequency for conditions can lead to less than optimum
studies of microstructural variation. Both material and component properties.
frequency dependent and frequency Discovery and elimination of these
independent corrections are incorporated conditions through process optimization,
to yield the true material attenuation. An rejection or repair of the green bodies
example of the detection of whisker could lead to significant cost savings.
clumping in composite ceramics by
variations in the attenuation characteristic Several characteristics of green state
is also described. bodies have been identified as important
to the production of reliable structural
Discontinuity Detection in ceramic components. These include
Ceramics (1) selection and composition of binders
and sintering aids, (2) techniques of form
Critical discontinuities can be detected in removal and interactions with powders,
most structural metals with ultrasonic (3) elemental composition, (4) mechanical
wave frequencies from 1 to 10 MHz. properties and (5) discontinuities such as
Consequently, most research and porosity and surface condition. For some
development in ultrasonics has of these characteristics, ultrasonic
concentrated on this frequency range, measurements and test procedures have
with little activity above 15 MHz. In proven useful. The green state body has
conventional monolithic ceramics, the much lower density than the final
critical discontinuity size is often 20 µm sintered product because of its distributed
(0.0008 in.) or less and frequencies of porosity. Variations in properties of the
50 MHz and higher are required. finished component may result from
variations in the spatial distribution and
In addition, detection of such small size distribution of pores and the
discontinuities at reasonable depths attendant effect on density uniformity in
requires focused ultrasound, and the the green state. Characterization of the
propagation of such energy through the size and spatial distribution of porosity
ceramic surface introduces severe beam and density variations is an important
aberrations. This condition is made worse part of process control.
by the very high index of refraction of
typical ceramics, thus limiting to a few To determine the feasibility of
millimeters the depth at which effective nondestructive discontinuity detection in
green ceramics, many laboratory studies

358 Ultrasonic Testing

have been conducted: radiographic, production quantities or with sampling
ultrasonic, nuclear magnetic resonance tests during process development.
and small angle neutron scattering
techniques. Density

Studies of Green State Sintering of preformed, green state
Applications ceramic bodies at high temperatures
reduces or eliminates voids and increases
Despite some difficulties, ultrasonic density. Localized variations in the
techniques have been studied for testing attained density of ceramic shapes after
green state ceramics.3 It was determined sintering are undesirable because of the
that the low density and porous nature of effects on structural performance.
the green state limits the applicable Radiographic and ultrasonic
frequencies to less than 10 MHz nondestructive tests offer the potential for
(depending on test object thickness) detecting and measuring density
because of attenuation. The wavelength variations.
associated with the range of frequencies
limits the test sensitivity to discrete Investigations have been made into the
discontinuities approaching 1 mm feasibility of using ultrasonic velocity
(0.04 in.). measurements5-7 to determine bulk
density in sintered alpha silicon carbide
Correlation was observed between with densities varying from 2.8 to
measurements of ultrasonic velocity and 3.2 mg·mm–3. Using 20 MHz frequencies
material density. Ultrasonic couplants and commercial ultrasound equipment,
such as water or glycerol are generally the nominal bulk density could be
absorbed by green state ceramics, estimated within 1 percent.
eliminating the coupling effect and
potentially affecting subsequent Bulk Material and Mechanical
fabrication. In some cases, applying Properties
pressure to the transducer without
supplementary couplant has allowed Common practice for determining
adequate transmission into the test object. sintered ceramics’ mechanical properties
Care must be taken to avoid damage to (such as strength or fracture toughness) is
the test object. In addition, velocity data to perform destructive tests on large
vary with transducer pressure. numbers of specially prepared reference
standards (three and four point bending
An alternate coupling technique in standards for strength determination and
which the green ceramic is placed in an short rod, notched beam standards for
evacuated film enclosure also has been fracture toughness). Performance proof
successful.4 Through-transmission, tests can also be made on objects that
immersion ultrasonic techniques using a represent actual components. However,
10 MHz transducer allowed (1) detection these destructive tests are expensive and
of isolated inclusions with diameters on time consuming when providing
the order of 500 µm (0.02 in.) and statistical data on materials under
(2) mapping of velocity differences development.
attributed to density differences correlated
with X-ray tomography. The acceptable performance of a
ceramic material in service may depend as
Sintered Ceramics much on the basic bulk material and
mechanical properties of the individual
Heating or sintering is typically the final component as on the presence of discrete
fabrication stage for many ceramic discontinuities. For that reason,
components (in some cases, a grinding nondestructive testing of these properties
operation may be needed for surface on the actual item is a very powerful tool
preparation or sizing). Because sintering for predicting a newly fabricated
often completes the fabrication process, component’s serviceability or for
proper discontinuity characterization is determining the residual life of a
important to ensure that the desired component after service.
integrity and quality is achieved and
maintained. Measuring Correlated Properties

A number of critical characteristics Ultrasonic technology for discontinuity
affect the service life of ceramic detection can be used for measuring
components. These include local density parameters that correlate with desired
variations, microstructure, mechanical material properties. The ultrasonic
properties, physical properties, surface velocity is a function of the elastic
properties (affected by machining), properties of the material, and ultrasonic
elemental composition and distribution. attenuation can be related to bulk
For some of these characteristics, microstructure and to structural
ultrasonic techniques are used with discontinuities (both microscopic and

Ultrasonic Testing of Advanced Materials 359

macroscopic). Laboratory studies by be advantageous in whisker reinforced
several investigators confirm the ceramic composites, as described below.
feasibility of techniques for making such
measurements. The transducer chosen for attenuation
studies was a 100 MHz center frequency,
Ultrasonic longitudinal and transverse plane wave unit having a 20 dB
wave velocities have been measured in bandwidth ranging from 30 to 120 MHz.
different directions in hot pressed silicon This transducer had a quartz delay line
nitride.8 Anisotropy was observed and and was coupled to the sample with a
velocities perpendicular to the hot very thin layer of water.
pressing direction were about 5 percent
higher than velocities parallel to it. For the measurements to reflect the
Ultrasonic attenuation measured with intrinsic attenuation, the data must be
both longitudinal and transverse waves in corrected for nonmaterial losses. Two
the 30 to 130 MHz frequency range effects are impedance mismatch losses at
exhibits a frequency dependent the sample interfaces and beam spreading
characteristic proportional to the square or diffraction losses. Both of these losses
of the frequency. are often much larger than the intrinsic
material losses and, if uncorrected, will
Work on siliconized silicon carbide corrupt the measurements. Correction for
tubes indicated that the velocity of sound beam spread losses is made and all
changed as a function of the volume corrections are then combined in a
fraction of silicon and this may offer a program that acquires the needed data,
technique for indicating silicon content.9 deconvolves the input and output data,
The feasibility of using ultrasonic determines and applies the appropriate
techniques to measure elastic moduli, corrections and plots the resulting transfer
microstructure, hardness, fracture curve.14
toughness and strength has been
demonstrated for a wide range of Initial results indicated sharply
materials, including metals, ceramics and increasing attenuation above 30 MHz in
fiber composites.10,11 Ultrasonic all samples, even those for which the
techniques are particularly useful because grain size was too small (1 µm) to produce
they involve mechanical elastic waves measurable scattering at these frequencies.
modulated by some of the morphological Comparison of the signals from the end
factors that govern mechanical strength of the delay line with the transducer
and dynamic failure processes. coupled to the test object and with the
transducer in air indicated that the
Ultrasonic Attenuation Studies reflection coefficient was much higher
than assumed from the acoustic
Ultrasonic attenuation has also been impedances of the coupled materials.
studied for testing microstructural
properties.12 To measure the attenuation This effect was shown to be caused by
rapidly over a wide frequency range, a the water coupling layer, whose effect is
broad band elastic wave was transmitted entirely negligible at frequencies below
through a sample and a material transfer about 20 MHz.15 The condition was
function was computed from digitized subsequently analyzed as a three-layer
records of the input and transmitted problem and an expression was obtained
signals. A commercial 150 MHz ultrasonic for the reflection and transmission
system with an integral high speed (2 ns) coefficient as a function of frequency, for
peak detector and four-bit digitizer was both normal and nonnormal incidence
assembled for the tests. This system used on the interface. These results were
an immersion tank that included an incorporated in the analysis, which then
integral three-axis scanning system determined the thickness of the coupling
controlled by a dedicated computer. layer (and thus the frequency dependent
reflection coefficient) from the digitized
Although pure backscattering delay line signals in air and coupled to
measurements might appear preferable for the test object.
investigating the microstructure of
ceramics, through-transmission A value for thickness is determined at
measurements (or pulse echo, using the each frequency within the pass band of
backwall reflection) were used because of the transducer, including computation of
the difficulty in obtaining repeatable gaussian summary statistics. Experience
results from pure backscatter. The with this program has indicated that
literature details problems attendant to slight (less than 1 µm [4 × 10–5 in.])
backscattering measurements and offers wedging of the transducer can easily be
some hope for improving the accuracy detected from the standard deviation of
and repeatability of this test.13 Although the measurements.
through-transmission measurements are
affected by damping losses and scattering A typical value for the water layer
losses in the test object, this turned out to thickness is 0.5 µm (2 × 10–5 in.) and
Fig. 1 shows the reflection coefficient as a
function of frequency for this thickness.
With the latest corrections, transfer curve
results are highly repeatable, are

360 Ultrasonic Testing

independent of sample thickness andReflection coefficientabout 1 µm (4.0 × 10–5 in.). For such a
appear to reflect frequency dependent small source, frequencies of more than
Attenuation (dB·mm–1) losses commensurate with those expected Attenuation (dB·mm–1)1 GHz are necessary before scattering
from scanning electron microscope losses become important. The dominant
observations of the microstructure. A least losses in the range shown in Fig. 2 appear
squares fit to the experimental data was to be damping losses, linear with
found to coincide with theory. frequency.

Figure 2 shows results obtained on a The constant term in Fig. 2,
monolithic alumina. The solid line is the term 1 under coefficient of fit, is
experimental data and the dashed line is a 0.345 dB·mm–1 and is generally considered
least squares fit to these data. The to be a measure of the viscous damping
coefficients of fit indicate that a linear fit loss in the material. Unfortunately, this
was optimum. This is not surprising, for term is also affected by several error
the average grain size in this material is sources, such as attenuator errors, that are
difficult to determine. The repeatability of
FIGURE 1. Reflection coefficient at quartz silicon carbide this term is not particularly good, even for
interface coupled by 0.5 µm (2 × 10–5 in.) of water. the same test object. However, the linear
term is extremely repeatable for a single
1.0 object or for different objects produced by
a common process and is useful for
0.9 characterization.

0.8 Figure 3 shows the transfer curve
obtained from partially stabilized zirconia.
0.7 For this material, the average grain size is
about 100 µm (0.004 in.) and scattering
0.6 becomes important even for frequencies
as low as 10 MHz. This is shown in the
0.5 transfer curve, where losses increase
approximately as the square of the
0.4 frequency. No data are shown for
frequencies greater than 60 MHz because
0.3 those losses exceed the dynamic range of
the measuring system.
0.2
Because the transfer curve is highly
0.1 repeatable for a given material and
indicative of its microstructure, it would
0 appear to be an excellent variable for
10 20 30 40 50 60 70 80 90 100 correlation with fracture properties.

Frequency (MHz) FIGURE 3. Transfer curve of magnesia
stabilized zirconia with length of 2.6 mm
FIGURE 2. Transfer curve of monolithic alumina with length (0.1 in.) and acoustic impedance of
of 9.4 mm (0.37 in.) and acoustic impedance of 39.88 g·cm–2s–1. Coefficient of fit at term 1
43.02 g·cm–2·s–1. Coefficient of fit at term 1 is is 2.971 dB·mm–1, at term 2 is
0.345 dB·mm–1 and at term 2 is 0.007 dB·mm–1·MHz–1. –0.109 dB·mm–1·MHz–1 and at term 3 is
0.003 dB·mm–1·MHz–1.
12
12
10
10
8
8
6
6
4
4
2
2
0
10 20 30 40 50 60 70 80 90 100 0
10 20 30 40 50 60 70
Frequency (MHz)
Frequency (MHz)
Legend
= experimental data Legend
= least squares fit = experimental data
= least squares fit

Ultrasonic Testing of Advanced Materials 361

Microstructure in WhiskerAttenuation (dB·mm–1) for the same material with no whiskers. A
Reinforced Ceramics number of whisker reinforced samples
were tested with a constant ratio of
Whisker reinforced ceramics consist of whiskers to volume but with various
typical ceramic matrices in which ceramic degrees of clumping as determined with
whiskers are dispersed. A whisker is scanning electron microscopy. One of the
typically 0.5 µm (2 × 10–5 in.) in diameter, samples had no demonstrable clumping
20 to 30 µm (0.0008 to 0.001 in.) long and this was confirmed by ultrasonic
and is too small to be detected by the testing. The transfer curve’s linear term
frequencies available in a commercial for this test object was about three times
ultrasonic test system. The major problem that of the monolithic material. Other
with these materials is whisker clumping objects’ curves fell between the clump free
rather than voids or inclusions in the component and the one with severe
matrix. Clumping causes the matrix clumping.
material to be very sparse in large regions,
on the order of 100 to 200 µm (0.004 to While this study was not definitive, it
0.008 in.) in diameter, thus forming an appears that, in whisker reinforced
effective void on the scale of the clump. alumina, there is a relationship between
the degree of whisker clumping and the
These clumps are easily detectable by damping losses, as determined by the
ultrasonic scanning but the approach is ultrasonic attenuation characteristic.
relatively time consuming. The transfer
curve can be obtained in just a few Detection of
seconds and, if sensitive to whisker Discontinuities
clumping, is a valuable preliminary test
that could obviate detailed scanning for Various Techniques
voids.
Several nondestructive testing methods
Figure 4 shows the test results obtained have been studied for detection of
from a whisker reinforced alumina known discontinuities in sintered ceramics. These
to contain severe clumping. The presence methods include ultrasonic, radiographic,
of clumps was established by ultrasonic small angle neutron scanning, acoustic
scanning and by scanning electron emission, infrared, liquid penetrant,
microscope analysis of a fracture surface. photoacoustic microscopic and microwave
testing. The minimum detectable size of
The transfer curve shows a linear discrete discontinuities or distributed
characteristic, which is reasonable in light discontinuities varies significantly among
of the fact that the density of whisker the various methods.
clumps is probably too low to measurably
increase the scattering losses for plane Ultrasonic techniques for discontinuity
wave insonification. However, the linear detection include (1) traditional pulse
term is about seven times that obtained echo techniques for detecting reflections
from discrete discontinuities and
FIGURE 4. Transfer curve of silicon carbon whisker reinforced (2) measurement of ultrasonic attenuation
alumina (numerous whisker clumps), with length of 2.9 mm and velocity as a function of frequency
(0.11 in.) and acoustic impedance of 40.49 g·cm–2·s–1. with correlation to distributed
Coefficient of fit at term 1 is 0.192 dB·mm–1·MHz–1 and at discontinuity size. The literature includes
term 2 is –0.047 dB·mm–1·MHz–1. reports on a 50 MHz C-scan imaging
system for discontinuity detection in
12 silicon nitride and a high frequency (150
to 450 MHz) A-scan system for
10 discontinuity characterization and matrix
testing.16-19 Signal processing schemes
8 include temporal and spatial averaging,
filtering and corrections for diffraction
6 and for attenuation.

4 Backscattering Techniques

2 Difficulties have been encountered in
discontinuity characterization with single
0 backscattering measurements. For high
10 20 30 40 50 60 70 80 90 100 frequency systems, special transducers and
filtering techniques have been developed,
Frequency (MHz) allowing comparison of time domain
backscattered signals from inclusions with
Legend calculations from theory. Good agreement
= experimental data
= least squares fit

362 Ultrasonic Testing

was observed for 100 µm (0.004 in.) discontinuities (as well as others) were
inclusions in silicon nitride. successfully detected. In this case, the
critical strength limiting discontinuities
Synthetic aperture imaging at 50 MHz were complex shaped, three-dimensional
has been studied to obtain voids 75 to 200 µm (0.003 to 0.008 in.) in
three-dimensional images of size and two-dimensional surface cracks
discontinuities. Computer simulations less than 150 µm (0.006 in.) deep.
based on theoretical discontinuity models
have been implemented. Longitudinal An acoustic microscopy technique has
and transverse wave ultrasonic pulse echo been developed using high frequency
techniques have been investigated in (50 MHz) focused ultrasonic transducers
silicon nitride and silicon carbide and fast signal processing to rapidly scan
ceramics at frequencies between 25 and silicon carbide heat exchanger tubes for
45 MHz for detection of pores and discontinuities and to display the results
inclusions in the range of 10 to 130 µm in real time.28 The system can acquire
(0.0004 to 0.005 in.). data as fast as 2800 data points per second
and has demonstrated the ability to detect
Transverse wave techniques are more and image seeded discontinuities as small
sensitive than longitudinal wave as 0.05 mm (0.002 in.).
techniques for tests of ceramics.
Sensitivity to discontinuities as small as Techniques for Very Small
25 µm (0.001 in.) has been demonstrated. Discontinuities
Objects that had subsequent failure at
detected discontinuities showed lower Both theoretical and experimental studies
strengths.20,21 have been used in developing techniques
for very small critical discontinuities.12
Scanning Laser Techniques Reliable detection of discontinuities as
small as 20 µm (0.0008 in.) cannot be
Several investigators have reported studies achieved simply by increasing the
with a scanning laser acoustic microscope. interrogating frequency until the
In one study, a frequency of 100 MHz was wavelength is comparable to the
used to detect and display pores and discontinuity size. The reason for this is
inclusions from 50 to 100 µm (0.002 to that the amount of energy intercepted by
0.004 in.) in silicon nitride disks.22,23 such a discontinuity is a negligible
Distributed porosity smaller than 50 µm fraction of the total transducer energy for
(0.002 in.) was detected because of its a typical plane wave transducer.
effect on ultrasonic attenuation rather
than by discrete indications. In addition, the intensity of scattering
from a small discontinuity illuminated by
Statistical studies have been conducted plane wave radiation decreases
to determine the probability of detecting approximately as (a·d–1)2, where a is the
surface and internal voids using 100 MHz discontinuity radius and d is the depth of
scanning laser acoustic microscopy in the discontinuity, assuming that the
sintered components of silicon carbide wavelength is small compared to the
and silicon nitride.24,25 Surface voids as discontinuity. Therefore, discontinuities
small as 100 µm (0.004 in.) in diameter can be detected much smaller than the
were reliably detected in polished test transducer size only at relatively shallow
objects. If close to the surface, internal depths for plane wave insonification. To
voids as small as 30 µm (0.0012 in.) in reliably detect 20 µm (0.0008 in.)
silicon nitride and 60 µm (0.0024 in.) in discontinuities at depths of several
silicon were detected. Larger voids were millimeters (0.1 in.), sharply focused
reliably detected at greater depths. transducers are typically required.
Ultrasonic techniques have been applied Focusing in turn dictates a small scan
up to 45 MHz and show potential for index with its attendant large scan time.
detecting clusters of subsurface Unfortunately, the propagation of energy
discontinuities near 25 µm (0.001 in.) in through the ceramic surface under these
hot pressed silicon nitride.26 conditions introduces into the beam
severe spherical aberration that partially
Similar studies in reaction bonded offsets the benefits of focusing.
silicon nitride show the potential for
detecting clusters in the 125 µm Spherical Void Measurement
(0.005 in.) range. Silicon rich inclusions
are difficult to detect when less than In monolithic ceramics, one of the
250 to 500 µm (0.01 to 0.02 in.). A common discontinuities is a
36 MHz pulse echo ultrasonic technique quasispherical void, so the typical
and scanning laser techniques have been scattering center is modeled as a spherical
applied to the detection of intentionally cavity. Although detection is not
seeded discontinuities in the range of 50 restricted to this shape, scattering data are
to 125 µm (0.002 to 0.005 in.) and 150 to interpreted in the light of this model,
250 µm (0.006 to 0.010 in.) in silicon unless there are valid reasons for
carbide disk thicknesses ranging from
2.5 to 130 mm (0.1 to 5 in.).27 The seeded

Ultrasonic Testing of Advanced Materials 363

suspecting otherwise. Models have been fine grained zirconia. The transducer and
developed for other discontinuity shapes system response have been removed from
such as planar cracks that commonly the data by deconvolution and the results
occur in ceramics.29 depend only on the discontinuity
characteristics. If the turning point in the
There is also an analytical model for response at about 89 MHz is equated with
the spherical cavity30 and Fig. 5 shows the the theoretical transition at ka = 1, a
theoretical response from such a cavity in discontinuity diameter of about 25 µm
silicon nitride. Here k is the wave vector (0.001 in.) is estimated. Smaller
of the ultrasound and a is the radius of discontinuities can be detected at higher
the cavity. The scattering is characterized frequencies.
by a rapid increase in cross section with
frequency (the rayleigh region), followed With these capabilities, discontinuities
by a transition to an oscillatory cross on the order of 20 µm (0.0008 in.) can be
section at about ka = 1. These oscillations detected at depths up to 3 mm (0.1 in.) in
are caused by interference between direct alumina or silicon nitride and at depths
and creeping waves and are often missing up to 5 mm (0.2 in.) in tetragonal zirconia
in real discontinuities because of their polycrystalline. In coarse grained
surface irregularities. In this case, the materials, severe scattering losses increase
scattering response is characterized by the the minimum discontinuity detection
rayleigh tail and a transition to a cross threshold but the critical discontinuity
section nearly constant with frequency. size is also larger in these materials.

Figure 6 shows experimental data Detection of Other Discontinuity
obtained from a natural discontinuity in Types
tetragonal zirconia polycrystalline (TZP), a
Figure 7 shows numerous indications
FIGURE 5. Theoretical scattering from spherical void. obtained in three modulus-of-rupture
(MOR) bars. Note the high density of
Rayleigh indications near the ends of two of the
region bars. These indications are equivalent to
100 µm (0.004 in.) and larger diameter
Differential cross section 0.4 High frequency region voids. The variations among the
0.35 indications are more visible in color
0.3 1 2 3 4 5 6 7 8 9 10 11 12 13 enhanced displays.
0.25 Wave vector k × radius a
0.2 Figure 8 shows the results obtained
0.15 from a 650 mm2 (1 in.2) piece of
0.1 monolithic silicon nitride. The smallest
0.05
0 FIGURE 7. Gray scale presentation of
discontinuities in zirconia
modulus-of-rupture bars.

FIGURE 6. Experimental scattering from a 25 µm (0.001 in.)Normal amplitude (V)
discontinuity in tetragonal polycrystalline zirconia.

1

0.8

0.6

0.4

0.2

0
10 20 30 40 50 60 70 80 90 100
Frequency (MHz)

364 Ultrasonic Testing

indications probably correspond to voids for accurate discontinuity detection and
with diameters of 20 µm (0.0008 in.) and microstructural characterization.
lie at a depth of 3 mm (0.1 in.). Mechanical and environmental loads
applied to ceramic matrix composites can
Figure 9 shows the results obtained cause degradation in the form of discrete
from whisker reinforced alumina known discontinuity nucleation and distributed
to contain clumping. The indications are microscopic damage that plays a
from clumps on the order of 100 to significant role in reduction of desirable
200 µm (0.004 to 0.008 in.) in diameter. physical properties. Categories of
Similar results were obtained from whisker microscopic damage include
reinforced silicon nitride. However, a fiber-to-matrix disbonding (interface
monolithic sample prepared from the failure), matrix microcracking, fiber
same powder was virtually free of fracture and buckling, oxidation and
indications. second phase formation. An ultrasonic
guided wave scan system was developed
Advanced Ceramics to characterize various microstructural
and discontinuity conditions in ceramic
Ceramic matrix composites developed for matrix composite samples: (1) silicon
advanced aerospace propulsion to save carbide fiber in silicon carbide matrix and
weight, to improve reuse capability and to (2) carbon fiber in silicon carbide matrix
increase performance provide a challenge (Figs. 10 to 12).31

FIGURE 8. Threshold presentation of Guided wave ultrasonic testing is
discontinuities in silicon nitride (frequency is generally thought to be an attractive
50 MHz). alternative to scanning because guided
waves can be excited at one location of a
structure by a single transducer or line of
transducers with returning or received
echoes indicating the presence of

FIGURE 10. Ultrasonic guided wave test system: (a) scanner
hardware; (b) typical waveform.

(a)

Stage Z stage
indexer pressure

gage

Computer X and
Y stages
FIGURE 9. Photomicrograph of whisker
clumps in silicon carbide whisker reinforced Load cell
alumina. Z stage actuator
Transducers

(b)

Ultrasonic Testing of Advanced Materials 365

discontinuities. This type of test has lowered to be in contact with the sample,
detected discontinuities and material and another measurement is made. This
degradation in many types of materials routine is repeated to perform raster
and components — in some applications scans. Scan increments vary between 1
over significant distances. The guided and 5 mm (0.05 and 0.25 in.). Images
wave signal in raw form is very complex constructed included those calculated
(many dispersive and interfering modes, from centroid mean time,
traveling at different velocities) with time-versus-distance skew factor, zeroth
significant coherent noise that cannot be moment and frequency centroid of the
averaged out. As a result, guided wave power spectrum, frequency dependent
techniques seem to be most successful for ultrasonic decay rate and frequency
discontinuity detection when tuning for a dependent energy density initial value.
minimally or nondispersive mode of
ultrasound, at a particular excitation Figure 11 shows an ultrasonic guided
frequency, in one direction, to control wave scan image showing delamination
coherent noise. One guided wave within a silicon carbide fiber composite in
technique takes a different approach by a silicon carbide matrix. The delamination
using the total, multiple-mode ultrasonic was most easily discriminated in the
response, using separate sending and centroid mean time image (whitish areas
receiving transducers, doing so in a in image). Centroid mean time can be
scanning configuration and using thought of as the time in the raw
specialized signal processing routines to waveform demarcating the location of
extract parameters of the time domain energy balance. Time domain waveforms
and frequency domain signals.31 These associated with the delaminated area and
parameters have proven sensitive to a nondelaminated area are shown in the
changes in microstructural conditions and figure. The shift in centroid mean time
to the presence of discontinuities and away from the origin is quite apparent for
appear promising at monitoring the delamination when viewing the
degradation in ceramic matrix waveforms.
composites. There may be several further
advantages of using guided wave scanning Figure 12 shows the ability of guided
over conventional ultrasonic techniques. wave scanning to detect cracking
perpendicular to wave travel in such a
1. Guided wave scanning can be fiber matrix.
performed directionally, allowing
correlations between ultrasonic FIGURE 11. Silicon carbide fiber in silicon carbide matrix:
parameters and directionally (a) centroid mean time image with delamination showing as
dependent material properties, such as white area; (b) time domain waveforms.
for unidirectional composites or to test
the premise of nondirectionality of (a)
properties.
(b)
2. The sample under test does not have
to be immersed in fluid as for most Delamination
conventional ultrasonic
characterization.

3. Guided wave scanning can be applied
to components with mildly curved
surfaces.

4. Guided wave scanning can be more
versatile in characterizing local
modulus changes than resonant
frequency techniques that require
nodal excitation and generation and
are thus not applicable for scanning.

Typical experimental setup parameters
when scanning ceramic composites
include broad band ultrasonic transducers
with center frequencies f ranging from 1
to 3.5 MHz (both sender and receiver of
the same frequency). Ultrasound is
coupled to the material via elastic
coupling pads. The distance between
sending and receiving transducers is
25 mm (1 in.). Analog-to-digital sampling
rates used are 10 MHz to 25 MHz. A
measurement is made (contact load = 35.6
± 2.3 N [8 ± 0.5 lb]), the sender receiver
pair is lifted, moved to the next location,

366 Ultrasonic Testing

Detection of Surface (0.004 in.) deep in silicon nitride and
Properties silicon carbide.32,33 The technique was
found to be sensitive to surface conditions
The integrity of a sintered ceramic surface such as grinding damage as well as to
is critical because of the small critical discontinuities. Flexural strength was
discontinuity size, the tensile stress correlated qualitatively with ultrasonic
concentrations possible on the object response to machining damage.
surface during service and the potential Sensitivity to discontinuities is limited by
for damage during fabrication (during the depth of machining damage and the
grinding or other machining, for focal spot size of the ultrasonic beam. For
example). For these reasons, surface the 5.8 mm (0.230 in.) focal spot, the
testing is useful in demonstrating ceramic smallest verified discontinuity was a
performance. semicircular crack with a depth of 30 µm
(0.0012 in.).
Several different methods have been
used to accomplish high resolution, high Surface acoustic waves in ceramics have
sensitivity tests, including optical also been studied for detecting individual
techniques, ultrasonics, penetrants and cracks with depths as small as 60 µm
others. Some of the technology used for (0.0024 in.).33,34 Detectability is affected
detection of discontinuities in the by the size distribution of adjacent
microstructure is also applicable for background microcracks (such as those
surface testing. caused by grinding).

High frequency ultrasonic surface A preliminary correlation has been
waves have been studied by several observed between attenuation of the
investigators and shown to be useful for surface wave and the extremes of the
the detection of small surface crack size distribution. In the studies
discontinuities. A 45 MHz ultrasonic particularly addressed to machining
surface wave technique has been damage (suspected of being a major cause
developed for detecting surface of test object failure), a measurement
discontinuities less than 100 µm technique was developed to find the
microcracks and a long wavelength

FIGURE 12. Value images of carbon fiber in silicon carbide matrix: (a) zeroth energy density
after 14 h; (b) frequency dependent energy initial density after 14 h; (c) zeroth energy
density after 16 h; (d) frequency dependent energy initial density after 16 h. Note the white
indication in both 16 h images at X = 100 to 110 mm, which was not apparent after 14 h of
testing and which was at the eventual failure location.

(a) (c)

(b) (d)

Ultrasonic Testing of Advanced Materials 367

scattering theory was proposed for
predicting size. The techniques are
affected by variations in the plastic zone
and by the crack closure at the surface
resulting from residual stresses. Fracture
stress prediction may be possible if the
long wavelength criterion is met and the
size of the plastic zone is correctly
estimated.

The long wavelength (low frequency)
showed good agreement between
predicted and actual fracture stresses in
silicon nitride test objects containing
semicircular surface cracks with radii
(depths) ranging from 50 to 275 µm
(0.002 to 0.011 in.).35 The scattered
ultrasonic radiation patterns from surface
cracks show that modulation of the
ultrasonic frequency spectrum is related
to crack length and aspect ratio
(geometric crack parameters important for
failure prediction).36-39

368 Ultrasonic Testing

PART 2. Ultrasonic Testing of Adhesive Bonds

Adhesives are used to maintain structural when the stress exceeds the material
integrity and to transfer loads between strength in its weakest spot, not when a
the components of an assembly. For stress exceeds the average bond strength,
adhesives in discontinuity sensitive which may be substantially higher.
structures with complex configurations,
there is a need for highly reliable Bond Strength
nondestructive testing techniques for
evaluating bond performance. The mechanical strength of an adhesive
bond depends on two factors, cohesion
The capability of ultrasonic tests of and adhesion. Cohesion is the attraction
bond quality in structural components is between the molecules of the adhesive
described below. Also discussed is the layer. Cohesive strength is determined by
limitation of other nondestructive tests the type of adhesive, its elastic properties
for providing quantitative results, along and its thickness. Limited information can
with the details of a joint theoretical and be obtained about these parameters when
experimental research program using using nondestructive test methods.
leaky lamb waves on laboratory test
objects. The leaky lamb wave technique is Adhesion is the molecular attraction
shown to have advantages over other between dissimilar bodies in physical
nondestructive tests. contact or the bond between an adhesive
and the adherends. Adhesion quality is
Characteristics of Adhesive critical to the performance of a bond
Bonds between components of an assembly.
Because a bond interface layer is often a
The condition of a bond can be evaluated fraction of a micrometer thick, it is very
best only with a test that detects difficult to characterize it
unbonds, characterizes them and nondestructively. Weakness of this layer
determines the properties of the adhesive. can be caused by poor surface preparation
Nondestructive determination of the and is not detectable by contemporary
strength of a bonded joint provides the nondestructive test methods. In cases
most meaningful information about bond such as diffusion bonding and
quality. To this end, material parameters steel-to-rubber bonding, weak adhesion
that might be sensitive to strength have can also result from many minute
been sought by many investigators.40 separations over a certain area. This type
Limited success was achieved for cases in of unbond can be measured
which material variables were carefully nondestructively for the average degree of
controlled and where there was a direct separation rather than for the weakness of
relationship between the average strength the bond.
and a single property of the adhesive
material, such as thickness. Because of these limitations,
nondestructive testing is used to detect
In practice, it is difficult to correlate and characterize unbonds rather than to
nondestructive test data with the strength determine bond strength. Acoustics and
of a bond. Bond strength is not a physical ultrasonics are the primary means for
property but a structural parameter, an nondestructive testing of unbonds.
indication of the highest stress that a Resonance, pulse echo,
specific structure’s weakest spot can bear. through-transmission and ultrasonic
Generally, there is no nondestructive test spectroscopy are the most widely used
method that can search systematically techniques and are reviewed below. The
through a structure to identify all the leaky lamb wave technique reportedly has
weak points and then determine which is potential for nondestructive testing of
the weakest. bonds and is discussed later in this
section.
Furthermore, bonding is particularly
sensitive to interface characteristics, Tap Testing
difficult to determine nondestructively.
Therefore, even though it may be possible The simplest and most common adhesive
to estimate the average strength of a bond test is called tap testing. The
bonded system, the usefulness of this inspector strikes the test object, typically
measure is questionable. A bond fails

Ultrasonic Testing of Advanced Materials 369

with a metal implement such as a phase and ρ is the material density (gram
hammer or coin, and listens to the per cubic millimeter).
characteristics of the resulting sound. This
is a qualitative, sonic technique rather The change in the acoustic impedance
than an ultrasonic technique and, because of unbond also changes the
although fast and simple, it can only electric impedance of the transducer
detect gross unbonds. loaded by the material. The change in the
load depends on the distance traveled by
A systematic study has been made of the acoustic energy in the material (the
the tap testing method by examining the effective thickness) causing an appropriate
input force.41 The study found that the change in z. A proper selection of the test
characteristics of the force input by a tap frequency allows an increase in the
are changed by an unbond. Furthermore, response to unbond and this is done
the impact duration increases and the during instrument calibration.
peak input force of the impact decreases
as the discontinuity becomes larger. Applications of Resonance Tests

Even though this study resulted in a Resonance techniques for
substantial improvement in the reliability nondestructively detecting unbonds are
and detectability of tapping, the method based on the difference in characteristic
is still limited to thin layers less than response between bonded and unbonded
1 mm (0.04 in.) thick. The smallest areas. Commercial instruments developed
unbond detectable with tapping is 10 mm for this application are typically calibrated
(0.4 in.) in diameter in typical test objects. by using the response from a bonded or
an unbonded area as a reference.
Resonance Tests
One of these instruments displays the
The resonance method can be used to test transducer’s electrical impedance at a
adhesive bonding by accessing the test selected frequency as a point on the
object from one side using broad band or complex plane. Figure 13 shows typical
swept frequency excitation. Resonance responses to transducer loading by
testing is based on establishing a standing bonded and unbonded metal-to-metal
wave in the material when the effective areas of a reference standard. In this
thickness of the material is equal to an figure, three impedance points are
integral number of half wavelengths.42 displayed: (1) air loading, where the
The effective thickness of the test material transducer is not loaded, (2) an unbonded
is inversely related to the resonant layer of aluminum and (3) a bonded
frequency. An unbonded material gives system of two aluminum plates. The
rise to a higher resonant frequency than transducer’s impedance changes with
the resonance frequency of the bonded loading.
structure.
The other type of bond tester displays
The shift in frequency can be the impedance amplitude and phase as a
determined by measuring its loading function of frequency on two different
effect on piezoelectric transducers. This indicators, thereby providing additional
effect can be analyzed using transmission data. Using this system, it has been shown
line theory, where the test material serves that a change in each of these parameters
as a termination or load. The transducer’s is related to a change in adhesive layer
electrical impedance and its resonant thickness.44,45 Because the thickness of the
frequency are affected by the load. Their adhesive can be correlated to the bond
values for bonded materials can be used as strength, the instrument can be used as
a reference when searching for unbonds. an indirect means of bond strength
measurement. This application was widely
To determine the effect of the unbond used during the 1960s but is not reliable
on the induced ultrasonic wave, the wave for general adhesives.
behavior must be analyzed. At low
frequencies (in the kilohertz range), the FIGURE 13. Display of transducer impedance
attenuation in the test material can be and changes caused by changes in loading.
neglected. Assuming an incident plane
wave, the acoustic pressure can be related Electronic impedance Bonded
to particle velocity through acoustic (relative scale)
impedance z:43 Unbonded

(1) z = ρv tan h ⎣⎡ α + i(β + kd) ⎦⎤ Air loaded
transducer
where d is the distance (millimeter), k is
the wave number, v is acoustical velocity Resistance (relative scale)
(millimeter per second), α is the
reflectivity constant, β is the change in

370 Ultrasonic Testing

Ultrasonic Pulse Echo and transducer. Thus, a single transducer is
Through-Transmission used and the effect of attenuation is
Tests increased by doubling the wave path.

An elastic wave traveling from one Using the through-transmission
material to another is partially reflected technique in field conditions is sometimes
and partially transmitted through the difficult because of the required access to
interface. The relative amount of energy both sides, the alignment of the two
of each of the two generated wave transducers during testing and the need to
components is determined by the degree maintain water columns for coupling. The
of mismatch in the acoustic impedance pulse echo technique, on the other hand,
between the two materials. provides information about the bonded
areas without these difficulties. Pulse echo
The reflection coefficient R and tests can be used to determine whether
transmission coefficient T for an incident the unbond is above or below the
wave normal to an interface may be adhesive layer or, in the case of
determined: composites, if it is an unbond or
delamination. When the delamination is
⎛ z1 − z2 ⎞ 2 detected, its depth can be determined
⎝⎜ z1 + z2 ⎠⎟ with an accuracy of ±0.1 mm (±0.004 in.)
(2) R = in typical test objects.46

and Typical pulse echo data for a relatively
thick aluminum plate and the results of
(3) T = 1 − R theoretical calculations are shown in
Fig. 14. The agreement between data and
where z1 and z2 are the acoustic theory is excellent.
impedances of the two materials, defined
through the relation: Examples of response when using pulse
echo and through-transmission
(4) zj = pj vj techniques in more complex test objects

where j = 1 or 2. The larger the impedance FIGURE 14. Comparison of typical pulse echo
mismatch between the two materials, the data with theory for unbonded aluminum:
higher the reflection coefficient. As an (a) pulse echo response for wrought
example, if aluminum is bonded to an aluminum alloy at 5 MHz; (b) theoretical
adhesive layer of epoxy, then R = 0.48. If response for same test object. The first pulse
the aluminum is not bonded, then at the represents reflection from the front surface
interface of aluminum and air, R ഡ 1 and of the aluminum plate immersed in water;
total reflection occurs. all other pulses are from the back of the
plate.
On the other hand, if water penetrates
the unbonded area during a standard (a)
ultrasonic test, then R < 0.71 and the
sensitivity to the unbond decreases. This Response (relative scale 0.5
analysis can be applied to 0.4
through-transmission by using Eq. 2. In 0.3
the case of an unbond, the discontinuity 0.2
causes a complete blockage of the wave 0.1
transmission, because T = 0 for an 0
interface of aluminum and air. –0.1
–0.2
These two ways of identifying an –0.3
unbond are the nondestructive techniques –0.4
known as pulse echo and –0.5
through-transmission. For such tests,
short ultrasonic pulses are used at 1 2 3 4 5 6 7 8 9 10
frequencies higher than the resonance Time (µs)
technique (0.5 to 10 MHz). For pulse echo
tests, a single transducer is used and (b)
access to only one side of the test object is
required. For through-transmission tests, Response (relative scale 0.5
two transducers are required (one on each 0.4
side of the test material) and must be 0.3
maintained along the wave path. To 0.2
improve on the detectability of 0.1
discontinuities in a through-transmission 0
test, a reflector plate can be used to reflect –0.1
the transmitted signal back to the –0.2
–0.3
–0.4
–0.5

1 2 3 4 5 6 7 8 9 10
Time (µs)

Ultrasonic Testing of Advanced Materials 371

are shown in Figs. 15 and 16. For (5) T = 2m
through-transmission testing (Fig. 15), the
difference between bonded and unbonded ( )4m2 + 1 − m2 2 sin2 kd
areas is related to the amplitude of the
received signal after traveling through the where is d is the thickness of the gap
test object. This technique is widely used (millimeter), k is the wave number in the
because it inspects the entire volume of gap and m is the degree of impedance
an object in one test. mismatch (z1·z2–1).

Detected discontinuities are accepted The transmission coefficient T of the
or rejected by comparing the response air gap between two aluminum half
from a discontinuity to the response from spaces is shown in Fig. 17. The ultrasonic
a reference standard defined by the frequency times the gap thickness is on
material specifications. It is common to the logarithmic abscissa. As shown for
define levels of quality that depend on 1 MHz, a gap of 5 µm (0.0002 in.) allows
the discontinuity size with letters A, B and very little transmission. However, if the
C, where A identifies the highest quality frequency is reduced to 10 kHz, then
used for primary structures and C about 25 percent transmission occurs and
identifies the lowest quality.45 the detectability of the discontinuity is
substantially reduced. This shows that
The detectability of unbonds using
either of the two methods is critically AmplitudeFIGURE 16. Pulse echo from
dependent on the unbond gap. If the gap (relative scale)graphite-to-epoxy [0,90]2S, glass-to-epoxy
is much smaller than the wavelength of layer bonded with adhesive to aluminum
the ultrasound, then both bonded and honeycomb: (a) bonded sample;
unbonded areas have the same response. (b) delamination between fourth and fifth
Assuming that a layer of foreign material layers of the graphite-to-epoxy skin;
is sandwiched between two half spaces of (c) unbond between glass-to-epoxy and
the same material. The transmission adhesive layers.
coefficient T for normal incidences is
given by:47 (a)

FIGURE 15. Through-transmission response 1
of sandwich structure made of
graphite-to-epoxy skin and metallic 2
honeycomb: (a) bonded sandwich;
(b) unbonded sandwich. 3

(a) Time (relative scale)
(b)
Amplitude (relative scale)
1

Amplitude
(relative scale)
4
Amplitude (relative scale) Time (relative scale) Amplitude
(b) Time (relative scale)(relative scale)
(c)
Time (relative scale)
372 Ultrasonic Testing 1

4

Time (relative scale)
Legend
1. Front surface reflection.
2. Reflection from interface of graphite epoxy to glass

epoxy.
3. Reflection from adhesive honeycomb interface.
4. Disbond or delamination.

unbonds can best be detected using account by assuming that the wave
frequencies with wavelengths sufficiently number and the acoustic impedance are
smaller than the suspected unbond gap. complex. Efficient computer codes have
been developed for the calculation of the
Ultrasonic Spectroscopy Reflection amplitudefrequency dependent reflection
coefficientcoefficient.50
Data signals measured by pulse echo or
through-transmission are usually Reflection amplitudeThis analysis can be useful when
examined in the time domain using a coefficientexamining the effect of the bonded and
broad band signal. The ultrasonic unbonded materials on the reflection
spectroscopy technique is based on coefficient. As an example, consider a test
analyzing the received signal in the object consisting of two identical
frequency domain. For this purpose, the aluminum plates bonded by a layer of
signal is transformed to the frequency epoxy. Assume that the aluminum layers
domain using a fast fourier are 1 mm (0.04 in.) thick and the epoxy
transformation algorithm. layer is 0.1 mm (0.004 in.) thick and with
attenuation of 10 dB·mm–1·MHz–2. The
The advantage of spectroscopy is its calculated reflection coefficients as a
ability to reveal frequency dependent function of frequency for several possible
features that cannot be easily identified in cases are shown in Fig. 18.
time domain signals. Furthermore, the
signal can be processed and enhanced to FIGURE 18. Amplitude spectra of calculated reflection at
improve its detection of discontinuities. 0 rad: (a) perfectly bonded aluminum plate; (b) plate with
Examples of such processes include complete disbonding at lower epoxy-to-aluminum interface;
filtering, convolution and correlation. (c) plate with complete disbonding at upper interface.
This technique is called ultrasonic
spectroscopy and was studied as a potential 1.2
nondestructive testing tool during the
1970s.48 1.0

The technique is based on the analysis 0.8
of the spectral response of the bonded test
object compared to the response from an 0.6
unbonded reference standard. The
frequency dependent reflection and 0.4
transmission coefficients can be
determined for any given number of 0.2
elastic, isotropic and homogeneous layers,
by means of several well established 0
techniques.49 A plane longitudinal wave is
assumed to propagate normal to these 1 2 3 4 5 6 7 8 9 10
layers.
Frequency (MHz)
The reflection coefficient has a 1.2
frequency dependence related to the
thickness and the elastic properties of 1.0
each layer of the bonded medium. The
attenuation in each layer and specifically 0.8
in the adhesive layer can be taken into
0.6
FIGURE 17. Transmission response for air gap between two
aluminum half spaces. (X axis is logarithmic scale.) 0.4

Transmission coefficient T 1.2 10 kHz × 5 µm 0.2Reflection amplitude
1.0 1 MHz × 5 µm coefficient
0.8 0
0.6 –7 –6 –5 –4 –3 –2 1 2 3 4 5 6 7 8 9 10
0.4 Gap thickness × frequency (mm·MHz)
0.2 Frequency (MHz)
0 1.2

–8 1.0

0.8

0.6

0.4

0.2

0

1 2 3 4 5 6 7 8 9 10

Frequency (MHz)

Ultrasonic Testing of Advanced Materials 373

The predicted spectra can be compared the wave is refracted and mode converted
with measured responses to determine if to induce plate waves. When excited,
the aluminum layers are bonded. The these waves propagate along the plate and
bonded system has six minima within the are strongly affected by the properties of
frequency range of 1 to 10 MHz. The the bond.50,51
unbonded aluminum layer has only three
minima. The minima are the result of The leaky lamb wave technique uses
destructive interference of the waves two transducers in a pitch catch
within the test object. In the time arrangement. The test object is typically
domain, a large number of minima are immersed in a water tank or a water
associated with fewer reflections in a column is maintained between the
given time. Multiple reflections with a transducers and the object surface. For a
short time of flight identify unbonded fixed angle of insonification, the acoustic
aluminum plate and are used in a test waves are mode converted to lamb waves
called the ringing technique. at specific frequencies, resulting in leakage
of acoustic radiation into the fluid.
The spectral representation of the pulse
echo results in Fig. 19a are shown in When a leaky wave is introduced, the
Fig. 19b. There is general agreement field of the specularly reflected wave
between the theoretical and measured (reflection from a half space) is distorted.
spectra but a number of detailed features The specular component of the reflected
of the theoretical predictions are not wave and the leaky wave interfere, a
reproduced in the measured spectra. phase cancellation occurs and two
components are generated with a null
Leaky Lamb Waves between them.52 A schematic diagram of
the leaky lamb wave technique using a
In the techniques discussed above, the plate immersed in fluid is shown in
ultrasonic waves were incident normal to Fig. 20. A typical pulsed schlieren image
the test object. The leaky lamb wave of the leaky lamb wave response using a
technique is based on insonification of glass epoxy laminate is shown in Fig. 21.
the test object at an oblique angle, where
Figures 22 and 23 show the spectral
FIGURE 19. Measured data from joint in thin bonded response of a uniform aluminum plate
aluminum plates immersed in water. Epoxy layer is 0.1 mm and a bonded aluminum epoxy plate
(0.004 in.) thick and plate is 0.8 mm (0.032 in.) thick: immersed in water and insonified at
(a) time domain data; (b) spectral data. 0.34 rad (19.5 deg). The minima often are
associated with the excitation of leaky
(a) lamb wave modes in the test object. The
agreement between theory and
1.0 experiment is excellent for the unbonded
plate and reasonably good for the bonded
Response 0.5 plate, indicating a need for further study.
(relative units)
0 The possible lamb modes at various
angles of incidence are shown by
–0.5 dispersion curves. Figure 24 shows the
calculated dispersion curves for the three
–1.0 1 2 3 4 5 6 7 8 9 10 possible cases of a bonded aluminum
Time (µs) plate. The dispersion curves in Fig. 24c
(b) show good agreement between theory and
experiment for unbonded plate. The
0.15 agreement is good for the bonded plate,
except at high frequencies and high phase
velocities.

FIGURE 20. Schematic diagram of leaky lamb
wave field.

Amplitude (V) Transmitter Receiver

0.1

0.05 Fluid
0 Plate

1 2 3 4 5 6 7 8 9 10 Fluid
Frequency (MHz)
Leaky wave
Null zone

374 Ultrasonic Testing

Generally, a bonded structure can give Amplitude (V)FIGURE 22. Leaky lamb wave spectra from unbonded
rise to many possible modes that are aluminum plate (0.8 mm [0.032 in.]) immersed in water
modifications of those for a single layer. with incidence angle of 0.34 rad (19.5 deg): (a) measured
For a bonded system, the excitation of the spectrum; (b) calculated reflection spectrum.
spectral response predicted for a single
layer is an indication of unbonding in the (a)
tested area. Leaky lamb waves can be used
for nondestructive testing of bonds by 9
scanning an assembly and detecting areas 8
at which the leaky lamb wave modes 7
appear. These modes are different from 6
those of the bonded assembly. 5
4
Leaky Lamb Wave Characteristics 3
2
Leaky lamb wave phenomena have two 1
characteristics that make them useful for 0
nondestructive tests of bonds. First, phase
cancellation in the null zone of the leaky 1 2 3 4 5 6 7 8 9 10
lamb wave field is sensitive to changes in
interface conditions. The presence or (b) Amplitude (V) Frequency (MHz)
absence of bonding as well as the change 1 2 3 4 5 6 7 8 9 10
in the properties of the adhesive 9
significantly alter the leaky lamb wave 8
response. 7
6
Furthermore, two types of stress 5
(compression and shear, corresponding to 4
longitudinal and transverse waves) are 3
encountered simultaneously when a lamb 2
wave travels in a plate. Only one type of 1
0
FIGURE 21. Pulsed schlieren image of leaky
lamb wave mode for tone burst signal Frequency (MHz)
before and after impinging on
glass-to-epoxy sample: (a) incident beam; FIGURE 23. Leaky lamb wave reflection spectra from bondedAmplitude (V)
(b) reflected and transmitted beams. aluminum plate [0.8 mm (0.032 in.)] immersed in water
with incidence angle of 0.34 rad (19.5 deg). Epoxy layer
(a) 0.1 mm (0.004 in.): (a) measured spectrum; (b) calculated
spectrum.
(b)
(a) 9
Amplitude (V)
8
7
6
5
4
3
2
1
0

1 2 3 4 5 6 7 8 9 10

Frequency (MHz)

(b) 9

8
7
6
5
4
3
2
1
0

1 2 3 4 5 6 7 8 9 10

Frequency (MHz)

Ultrasonic Testing of Advanced Materials 375

stress is involved in other ultrasonic tests. quality on the location of the minima in
Because the two types of stress are affected the spectra is shown. In Fig. 25d, the
differently by different material and dispersion curves for the same bonded
discontinuity parameters, the lamb wave system are shown.
technique can potentially provide better
diagnostics of interfacial bonds. The influence of the bonding layer’s
elastic properties on the lamb wave phase
An example of the influence of velocity is significant over a specific
bonding layer properties on leaky lamb frequency range. An inversion technique
waves is given in Fig. 25. In Figs. 25a to has been developed to extract the elastic
25c, the reflected amplitude spectra from properties of the adhesive from the
a single layered half space are shown for bonded joint dispersion curve.53
three possible cases: perfect bonding, a
weak bond and complete disbond at the FIGURE 25. Influence of bond properties on leaky waves in
interface. The strong influence of bond single-layered medium (titanium bonded to beryllium
substrate): (a) perfect bonding; (b) thin low velocity
FIGURE 24. Calculated dispersion curves for interfacial layer; (c) complete disbonding at interface;
lamb waves: (a) in bonded plate of 1 mm (d) frequency versus wave speed.
(0.04 in.) aluminum, 0.1 mm (0.004 in.)
epoxy and 1 mm (0.04 in.) aluminum; (a) 0.35 rad
(b) with disbonding at lower interface of (20 deg)
plate with 1 mm (0.04 in.) aluminum and 1.5
0.1 mm (0.004 in.) epoxy; (c) with 1.0
disbonding at upper interface of 1 mm Amplitude
(0.04 in.) aluminum plate. reflection 0.5 Water
coefficient h Titanium
(a) 0
0 0.5 1.0 1.5 2.0 2.5 Beryllium
1
0Phase velocity (km·s–1) Normalized frequency (fh·β–1)
9
8 (b) 0.35 rad
7 (20 deg)
6 1.5
5 Amplitude
4 reflection 1.0 h Titanium 0.01 h
3 coefficient Beryllium
2 0.5
1
0 0
0 0.5 1.0 1.5 2.0 2.5
1 2 3 4 5 6 7 8 9 10
Normalized frequency (fh·β–1)
Frequency × thickness (MHz·mm)
Phase velocity (km·s–1) (c) 0.35 rad
(b) (20 deg)
1.5
1 Amplitude
0 reflection 1.0
9 coefficient
8 0.5 h Titanium
7 Vacuum
6 0
5 0 0.5 1.0 1.5 2.0 2.5
4
3 Normalized frequency (fh·β–1)
2
1Phase velocity (km·s–1) (d)Normalized
0 wave speed
3.0
1 2 3 4 5 6 7 8 9 10 2.5(c·β–1)
2.0
Frequency × thickness (MHz·mm) 1.5
1.0
(c) 0.5

1 0 0.2 0.4 0.6 0.8
0
9 Normalized frequency (fh·β–1)
8
7 Legend
6 f = frequency (MHz)
5 h = thickness (mm)
4 β = transverse wave speed in titanium (mm·s–1)
3
2
1
0

1 2 3 4 5 6 7 8 9 10

Frequency × thickness (MHz·mm)

376 Ultrasonic Testing

Applications of Leaky Lamb Wave Fig. 27. The pulse echo technique shows a
Tests relatively small difference, at the level of
the material variations across the bonded
Various applications of the leaky lamb area. On the other hand, the unbond is
wave phenomenon have been clearly indicated when using leaky lamb
investigated for nondestructive tests of waves.
bonds. For studies of composites, a
precured sandwich was prepared, Assessment of Ultrasonic
containing [0,90]2S carbon epoxy skins Tests of Bonds
with a 13 mm (0.5 in.) high, 3.2 mm
(0.13 in.) polyamide paper, phenolic Technology does not provide a physical
honeycomb cell and simulated unbonds parameter that can be directly correlated
made of synthetic fluorine wafers of 25, with bond strength for a practical test
19, 13 and 6.4 mm (1, 0.75, 0.5 and method. Nondestructive test techniques
0.25 in.) diameter.54 The reference are more capable of detecting unbonds
standard was insonified at 0.26 rad either at the adhesive layer or at its
(15 deg) and the leaky lamb wave modes interface with the adherend. Because
were measured. adhesive bond strength strongly depends
on surface preparation, a nondestructive
A C-scan system was connected to the means for determining surface quality
leaky lamb wave setup and the amplitude (the presence of contamination, for
was recorded as a function of location. instance) is essential.56
Initially, a frequency sweep was made and
the minima associated with the leaky FIGURE 27. C-scan images of steel-to-rubber bonds: (a) leaky
lamb wave modes were recorded. The test
was conducted at 5.31 MHz, which lamb wave image from 4 to 2 V in 15-color scale; (b) pulse
represents one of the leaky lamb wave echo image from 0.55 to 0.25 V in 15−color scale, showing
modes in the unbonded skin. The test 13 mm (0.5 in.) unbond.54
results are shown in Fig. 26, where the
unbonds are clearly identified — the (a)
generation of a leaky lamb wave mode
creates a null detected by the receiver. 229 (9.02)

Detection of unbonds between metals Y dimension, mm (in.) 210 (8.27)
and rubber is another difficult
nondestructive test application. The 191 (7.52) Unbond
problems result from the low acoustic 171 (6.76)
impedance of rubber and the large
mismatch in acoustic impedance between 152 (6.01)
rubber and metals. This makes the
difference in the reflected signal from 133 (5.26)
bonded and unbonded rubber relatively
small. Because the leaky lamb wave 114 (4.5)
technique is based on measurement of the
amplitude of the null due to a phase 95 (3.75)
cancellation, the technique is very
sensitive to changes in boundary 75 (3.0)
conditions.55 A 6.4 mm (0.25 in.) thick
steel plate bonded to a 3.2 mm (0.13 in.) 0 475 500 525 550 575 600 625
thick rubber mat has been tested with the (0) (19) (20) (21) (22) (23) (24) (25)
pulse echo technique at 10 MHz and with
the leaky lamb wave technique at X dimension, mm (in.)
4.63 MHz. The results are shown in
(b) 163 (6.41)
FIGURE 26. Leaky lamb wave C-scan
showing disbonds in substrate of graphite 145 (5.69)
epoxy sandwich.54
Y dimension, mm (in.) 126 (4.98)
Above
adhesive 108 (4.27) Unbond
layer 90 (3.56)

Below 72 (2.85)
adhesive
layer 54 (2.14)

36 (1.42)

18 (0.71) 0 22 44 66 88 110 130
(0) (0.85)(1.7) (2.55)(3.4)(4.25) (5.1)

X dimension, mm (in.)

Ultrasonic Testing of Advanced Materials 377

Ultrasonic techniques can provide Assessment of Ultrasonic
information about adhesive bond Tests of Coatings
properties but some relevant parameters
in the detected signals may be unused. An Coatings require ultrasonic testing for
increase in the signal acquisition speed, verification of integrity in high
an improvement in signal processing temperature thermal barrier coatings and
techniques and an increase in the size and in some cases, evaluation of the process
speed of access to computer memory are itself.64 Coating means include plasma
expected to improve the capability of the spray, chemical vapor deposition and
technique. Improved techniques should physical vapor deposition. For instance,
allow several parameters to be captured the chemical vapor deposition
while a bonded area is being scanned and environment for silicon nitride reacts
this should enable an accurate assessment with carbon-to-carbon surfaces at 1400 °C
of bond quality. (2550 °F). Oxidation barriers require
noncontact measurement of coating
One approach assesses the durability of thickness and modulus while undergoing
commercial epoxy adhesive bonds by deposition in real time. A laser ultrasonics
measuring ultrasonic reflection from an system was successfully implemented for
interphase region between the adhesive monitoring this process. This
and the adherend.57 The technique uses a methodology for determining reaction
specimen geometry that overcomes the parameters in situ also provides the
drawbacks of the conventional adhesive opportunity to develop a system model of
sandwich. The interphase region is the process so that production runs are
modeled with spring boundary always controlled in the same way for
conditions. The normal and tangential repeatable coatings.
spring constants are determined as a
function of epoxy degradation, from Rayleigh wave velocities have been
normal incidence longitudinal and from measured in metallic coatings
transverse wave measurements. Obliquely electrodeposited on steel and an
incident transverse waves are also experimental correlation was found
measured with a dual-sensor between hardness and velocity.65 This
thermoplastic (polyvinyllidene difluoride) result suggests that surface wave velocity
transducer. An efficient angular spectrum measurements can be used to evaluate
approach was used to model the oblique coating hardness over regions inaccessible
incidence measurements, and the to conventional hardness tests. A laser
predictions of the model are compared ultrasonic system was set up with a
with the measurements for various levels chromium coating on a steel right circular
of degradation. cylinder. Laser generation of the surface
wave and detection was accomplished
Pulse echo and through-transmission using a plastic wedge with a 5 MHz broad
have traditionally been the most widely band transducer. Dispersion curves were
used production ultrasonic techniques for calculated using a wavelet decomposition
tests of adhesive bonding. Resonance and for time versus frequency. The correlation
pulse echo are used in field conditions. was made using calibration data from
However, leaky lamb waves have been these measurements.
used because of superiority in cases such
as steel-to-rubber and composite bonds. The elastic constants of zirconia air
Related techniques have been used for plasma sprayed thermal barrier coatings
interphase characterization.58 were studied using two 15 MHz
transducers in a water immersion tank.66
For composites, the leaky lamb wave Data were taken at various incident angles
technique has been found useful for and fitted to a calculated theoretical
laminates with limited types of fiber transfer function for an unlimited
orientation. Development of theoretical isotropic plate with plane waves. The
analysis of wave propagation in average longitudinal and transverse wave
anisotropic, multilayered media would velocities are 2560 m·s–1 and 1710 m·s–1
lead to a better understanding of wave respectively with variation lower than
behavior.59-63 Such a development has led 3 percent over a wide refraction angle
to an increased use of leaky lamb waves range, 0 to 1.1 rad (0 to 60 deg). Young’s
for nondestructive testing of bonded modulus was measured to be 33 GPa in
composites (the methods are presently too comparison to compact zirconia at
complex to interpret for multilayered 241 GPa, indicating a high void count of
laminates). Another area that benefits 15 percent in the presence of a microcrack
from ultrasonic testing is analysis of the network. In another application, the
bond between steel and rubber. elastic constants obtained from bulk wave
measurements were used to assess the
quality of thermal barrier coatings, which
may develop a substantial amount of

378 Ultrasonic Testing

porosity if the process is not controlled
properly.67

Short pulse scanning acoustic
microscopy has been developed to
investigate the structure, properties and
geometry of highly absorptive
multilayered polymer media.68 The
evaluation and visualization of internal
layers having three percent of the total
thickness have been demonstrated. The
approach also included a time domain
digital algorithm to provide the precision
needed. The polymer sample was an
advanced material for gas storage tanks
consisting of high density polyethylene,
an upper layer of the same high density
polyethylene with a black pigment filler, a
barrier layer of ethylene vinyl alcohol
copolymer and an adhesive layer of
polyethylene based modified polyolefin
adhesive resins.

Testing can be implemented remotely
to detect discontinuities of large metallic
pipes, tubes and plates with a surface
coating added for corrosion protection or
insulation.69 Because the coatings are
usually viscoelastic, the guided wave
ranges may be severely reduced unless a
proper mode and an adequate frequency
range are selected. To overcome this
limitation, a hybrid finite element
boundary element method which
explicitly includes the attenuating
properties of the coating was used to
determine the lamb and transverse
horizontal mode conversion factors at the
corrosion discontinuities under the
coating. Monotonic variations of the
primary mode conversion factors with
discontinuity depth enabled weakly
attenuated modes to be inspected.

Ultrasonic Testing of Advanced Materials 379

PART 3. Ultrasonic Tests of Composite Laminates

Composites are useful structural materials stacking is determined by design
because of their high ratios of strength to requirements and can be done in manual
weight and moduli to weight. Composites layup or by automated filament winding.
are multilayered, heterogeneous and Different types of discontinuities can be
anisotropic on both macroscopic and introduced during the production of
microscopic levels. Most discontinuities in preimpregnated fibers: (1) not enough
composites are different from those in resin, (2) inclusions and contaminations,
metals, and the fracture mechanisms are (3) excessive variability in fiber or resin
much more complex. properties, (4) nonuniform hardener
content and (5) fiber misalignment.
Various discontinuities or
combinations of discontinuities can have To make a composite, the
specific degradation effects on the preimpregnated layup or the filament
performance of a given composite and wound structure is cured by exposure to
these effects are determined by several elevated temperature and pressure in a
factors: discontinuity characteristics (such predetermined procedure. A vacuum is
as dimensions and location), geometry, maintained in the curing environment to
composition and other properties of the eliminate porosity in the resin. Resin is
host composite, the type and magnitude cured in three stages: (1) the fluid stage
of the applied stress and the environment (the resin is liquid and its molecules
to which the structure is exposed during combine to form a reactive polymerizable
service. material), (2) the polymerization stage
(polymers of long chains are formed) and
For many discontinuities, the specific (3) the hardening stage (polymeric chains
mechanisms of degradation and the cross link to produce a three-dimensional
effects of the above factors are not well network).
understood. The anticipated durability of
composite structural components can be The progressive physical condition of
confirmed by efficient, reliable and cost the curing resin determines the duration
effective ultrasonic tests. These techniques of each stage and the times when various
should allow the determination of temperatures and pressures are applied.
performance levels and serviceability at an The resin condition must therefore be
acceptable probability of detection. within specifications to ensure the final
quality.
Described below are (1) the life cycle of
a composite, (2) the type of Once curing is completed, the
discontinuities that can be induced at composite is postcured to relieve stresses.
each stage of the life cycle and (3) the These stresses are induced mainly by a
ultrasonic techniques in use or under mismatch of thermal expansion
development for tests of composites. coefficients between the various layers of
Although the tests described here are the composite and between the fibers and
applicable to a wide variety of composite the matrix. The laminate or filament
materials, including metal matrix wound structure might contain several
composites, the following discussion types of discontinuities, depending on the
concentrates on fiber reinforced plastic production process. Each of these
composites, particularly those used for discontinuities degrades the performance
aircraft structures. The fiber materials of the host composite structure and its
include graphite, glass and boron. Epoxy durability in service. Possible
resin is the usual matrix material. Boron discontinuities induced by fabrication
epoxy was the first material to be used for include (1) delamination, broken fibers
primary composite structures but was and matrix cracking, (2) fiber
replaced by graphite epoxy because of misalignment, (3) inclusions and
cost. contaminations, (4) inadequate volume
ratio of fiber to resin, (5) wrong layup
Sources of Discontinuities order, (6) overlap or the gap between the
fiber bundles in a layer, (7) insufficient
Fiber reinforced plastic components are curing or overcuring, (8) excessive
commonly made by curing porosity or voids and (9) knots or missing
preimpregnated fibers stacked in layers of roving of the winding fibers.
a certain orientation. The sequence of
Aircraft composite structures are
assembled using adhesive bonds to join

380 Ultrasonic Testing

combinations of composite laminates to variations affect the intensity, velocity,
metallic or other composite structures scattering, mode conversion and
such as honeycomb. Two basic fabrication reflection characteristics of an ultrasonic
concepts, precure and cocure, are used to beam that insonifies the test object.
build up a composite assembly. In
precuring, skins made of laminates with Different types of information can be
the required layup order are cured first obtained when the incident acoustic wave
and then bonded to an adherend. In is normal or at an angle to the material
cocure, the assembly is prepared by surface. To understand wave behavior and
stacking the adhesive layers over the to predict its characteristics in composite
adherend, laying up laminate plies at the material, a theory for the analysis of the
required sequence and then curing the wave behavior in anisotropic layered
complete structure. media is needed. A general theory of wave
propagation in multilayered composite
These processes can induce laminates has been presented in the
discontinuities such as unbonds, porosity literature.60 The fundamental features of
and voids as well as contamination of the the theory are briefly described below.
adhesive joint. Other means of assembly
include rivets or bolts, but these can Theory of Wave Propagation in
initiate cracking and delaminations near Composites
the fastener hole.
In a general test configuration for
In service, structural composites are composite materials, the composite
exposed to conditions that can induce laminate contains N layers (called
discontinuities different from those laminae) and is of total thickness H. Each
produced during fabrication. These lamina is a unidirectional fiber reinforced
include environmental degradation, layer and may have different orientations,
erosion and damage from impact, weather depending on the design requirements of
and fatigue. In the case of aircraft the laminate. The laminae are assumed to
components, impact damage is typically be perfectly bonded at their interfaces.
caused by dropped tools, birds or debris
encountered during taxiing or landing, In many multiple-orientation
military action or weather conditions laminates, the interfacial zones are matrix
during flight. Erosion is caused by rain, rich and therefore have material
hail, dust or sand. These damage sources properties that may be significantly
degrade the performance of the composite different from those of the adjacent
skins, the adhesive and the adherend. laminae. In such cases, the interfacial
zone should be represented by an
In some cases, the mechanism of additional layer of material with certain
degradation is not well understood or assumed properties and the solution
predictable. The response of composite procedure can still be applied to the
materials to fatigue conditions depends resulting problem. The laminate is
on the layer’s orientation, on the stacking assumed to be immersed in water and
order and on the nature of the applied insonified by a plane acoustic wave at an
loads. Fatigue can cause matrix cracking incidence angle θ. The incident wave can
and crazing, fiber failure, delamination be either time harmonic or pulsed. The
and disruption of the bond between fibers theory of wave propagation in
and matrix. multilayered composite laminates has
been presented in the literature.1,57 The
Impact damage is a primary concern to equations are well behaved at all
users of composite materials because the frequencies and can be solved by means
damage can appear at any location over of standard techniques.70
the structure and at any time. This is in
contrast to fatigue damage, which can be This general treatment of wave
induced only at high stress concentration propagation in composites can be applied
areas after exposure to a sufficiently large for any angle of incidence, including
number of mechanical loading cycles. normal incidence at which composites
Even low levels of impact damage can be behave as a layered isotropic medium. The
serious over time because of growth main interest of ultrasonic testing is the
resulting from stress concentration. reflection coefficient R as a function of
the frequency and incident angle. For this
Role of Ultrasonic Testing purpose, the order of the system of
equations that need to be solved can be
Ultrasonic techniques play a major role in somewhat reduced. Also, by treating k0 as
the nondestructive testing of composites, the unknown wave number, the
both in research and in practical dispersion equation for guided wave
applications. Ultrasound provides many propagation in the medium can be
parameters that can be used to detect and derived.
characterize discontinuities and to
determine the elastic properties of the
composite. Anomalies and property

Ultrasonic Testing of Advanced Materials 381

Testing Composites at 100 ns pulses, this technique has detected
Normal Incidence 1 mm (0.04 in.) diameter delaminations in
graphite epoxy laminates with ±0.2 mm
In most ultrasonic tests, a longitudinal (±0.008 in.) depth accuracy.72
beam is incident normal to the composite
surface. Under these conditions, the effect The accuracy of assessing depth
of material anisotropy can be neglected depends on (1) the consistency of the
and the approach is similar to tests of ultrasonic velocity across the material and
isotropic media. Even though the material (2) the similarity of velocity in the test
behaves as if it were isotropic, its layers material to velocity in a reference
cause extraneous reflections and increase material. In composites, this accuracy is
attenuation. limited because of elastic property
variations within the materials and
Two basic testing modes of normal between various components. These
incidence are commonly performed. In property variations are caused by
through-transmission tests, the nonuniform volume ratios of resin to
attenuation is determined from the fiber, by differences in polymerization
amplitude of the wave after it has traveled levels and by variations in material
through the test object. In pulse echo content between batches of the same
tests, several parameters can be evaluated: composite.73
(1) changes in the back reflection
amplitude, (2) amplitude of extraneous Velocity Measurements
reflections and (3) variations in time of
flight measured from the front reflection Using time-of-flight measurements, the
to the back reflection or the extraneous ultrasonic velocity of a given mode along
reflection. the material can be determined. Studies of
both longitudinal and transverse wave
Attenuation velocities can be used to determine some
of the elastic constants of a composite.
Attenuation is relatively high in The anisotropic nature of composites is
composites, primarily because of evident when examining the ultrasonic
scattering by the fibers and isothermal velocities of various modes parallel and
absorption in the resin. High attenuation normal to the fibers. Use of these
can be reduced by using the measurements with lamb waves has also
through-transmission mode, passing been shown to provide the capability to
through the test object only once. For monitor fatigue and thermal damage in
thick composites, the attenuation needs aerospace composite materials.73,74
to be reduced further by using lower
frequencies (0.5 to 2.25 MHz) and Theoretical velocity curves for models
sacrificing resolution or by using high of graphite epoxy and glass epoxy
power signals. The through-transmission
technique and C-scan systems are widely FIGURE 28. Pulse echo response from
used for detection of delaminations, 24-layer unidirectional graphite-to-epoxy
voids, resin rich and resin starved areas laminate: (a) without discontinuities;
and other anomalies that significantly (b) with discontinuities.
affect attenuation.71
1 23
The use of attenuation as a
characterization parameter is hampered (a)
by variables such as surface roughness or
wave coupling and their effects are 2
difficult to deconvolve from the data. In
addition, the transducer and instrument (b)
characteristics are difficult to control in a
reproducible manner. Therefore, Legend
attenuation measurements are applied 1. Front surface reflection.
mainly to identify a significant deviation 2. Delamination.
of material response from an average 3. Reflection from graphite-to-epoxy layer.
attenuation value for the tested laminate.

Time of Flight

The depth of discontinuities in composite
laminates can be determined with
relatively high precision by measuring
time of flight. For this purpose, short
duration pulses are used in the pulse echo
mode (Fig. 28) and the test procedure is
similar to those for metallic objects. Using

382 Ultrasonic Testing

composites are shown in Fig. 29. As the Resonance
figure shows, graphite epoxy is much
more anisotropic than glass epoxy. Studies Resonance conditions are established in a
indicate that velocity calculations can be composite plate when the thickness is a
applied for fiber reinforced plastic numeric multiple of half the wavelength.
materials containing matrix voids.75 The For this purpose, low frequencies are used
same studies also indicate the feasibility of to reduce the effect of attenuation on the
predicting fiber volume fraction, assuming measurements. Generally the resonance
low frequencies or small fiber diameter. technique is used to test bonded
Deviation from these conditions gives rise structures or to detect delaminations.
to velocity dispersion. This dispersion is
most noticeable for boron fiber because it The resonance technique can also be
has ten times greater diameter than glass used to measure the depth of
or graphite fibers.77 discontinuities but is not as practical as
other techniques. Depth testing requires
FIGURE 29. Theoretical velocity curves for fabrication of a large set of calibration
three basic modes of propagation standards with controlled discontinuity
(longitudinal horizontal transverse, vertical size and depth. In addition, test results are
transverse) in: (a) unbounded significantly affected by pressure on the
graphite-to-epoxy composite at 0.62 fiber transducer, variations in the material
volume fraction; (b) unbounded surface roughness and variations in the
glass-to-epoxy composite at 0.4 fiber elastic properties of the test object.
volume fraction.70
Spectroscopy
(a)
Conventional ultrasonic tests are based on
9.7 studies of the time domain acoustic
intensity integrated over the transducer
Normal to fiber axis (km·s–1) 7.76 area.78 If these data are analyzed in the
frequency domain using signal processing
5.82 techniques, frequency dependent features
can be determined. The interaction
3.88 between ultrasonic waves and
discontinuities — for example, scattering,
1.94 absorption and, in particular, interference
of wavelets scattered from various parts of
0 1.94 3.88 5.82 7.76 9.7 discontinuities — depends on the wave
0 Fiber axis (km·s–1) frequency. The frequency dependence of
the signal, detected by a wideband
(b) transducer, contains useful information
for characterizing discontinuities. Using
5 spectral analysis, discontinuities such as
delaminations have been studied by
Normal to fiber axis (km·s–1) 4 several investigators.79

3 Spectral analysis is finding limited
application for discontinuity testing
2 because of the large number of factors
(coupling and surface roughness) affecting
1 the spectrum and the fact that their
effects cannot be predetermined.
0 5 Deconvolution methods have been used
0123 4 to extract the relevant signal from the
characteristic signal of the transducer,80
Fiber axis (km·s–1) thus reducing the number of factors
contaminating the signal scattered from
Legend the discontinuity. It remains difficult,
= horizontal transverse however, to evaluate the scattered field
= longitudinal from discontinuities in composites as a
= vertical transverse function of frequency.

Testing Composites at
Oblique Incidence

The behavior of an acoustic wave
impinging at an angle on a composite is
significantly affected by the layered,
heterogeneous and anisotropic nature of a
composite. Oblique incidence is used to

Ultrasonic Testing of Advanced Materials 383

test composites in two modes. First is the placed so that the fibers were parallel to
backscattering mode, where a single the axis of rotation and its front surface
transducer is used. Scattering sources in matched this axis. The receiver was placed
the composite are identified as a function in a stationary position outside the
of fiber orientations. turntable to allow measuring the scattered
wave amplitude as a function of the
Second is the leaky lamb wave mode, angle.
where two transducers are used in a pitch
catch arrangement. The transducers can The measured field consists of side
be on the same side or on opposite sides lobes that strongly depend on the
of the laminate. The leaky lamb wave frequency but do not present any
receiver is placed at the null zone caused consistent trend. The side lobes are
by the interference of the specular dominated by local variations and
reflection and the leaky wave irregularities in the order of the fibers
components. As detailed below, many within the laminate. Even when the fiber
discontinuities are detectable by these two volume fraction is high, as in the case of
oblique incidence modes. graphite epoxy, the diffracted field is
strongly perturbed by local variations.
Backscattering
Although it is difficult to obtain a
When a composite half space or a plate is typical or characteristic scattered field,
insonified at an angle, a specular backscattering measurements using pulses
reflection occurs. The characteristics of and spatial averaging have been highly
reflection depend on the material and the successful. Because a single transducer is
fluid loading. While an unperturbed used, the number of variables associated
specular reflection is observed along the with the test setup is limited, making
fiber direction, strong scattering takes such a setup more practical. Another
place as the wave propagates normal to advantage of the backscattering method is
the fibers. A schlieren image of this
behavior is shown in Fig. 30.81,82 Many FIGURE 31. Acoustic backscattering from [0,±45,90]s
investigators have tried to develop graphite-to-epoxy laminate: (a) setup; (b) as function of
theories to predict this scattered field
from the fibers but none of the proposed rotation angle β for 0.5 rad (30 deg) incident angle.71
models has been corroborated by
experiment with composites. (a) Incident and backscattered
acoustic beam
The scattered field has been studied Transducer
extensively at different frequencies.81,82 Z
An experimental setup was prepared with α
a transmitter set in a fixed location
normal to the sample and rotating with it βY
on a turntable. To eliminate the side
lobes, transducers with a gaussian X
directivity were used. The laminate was
Fibers
FIGURE 30. Schlieren image of incident and
reflected waves from unidirectional Resin
glass-to-epoxy laminate: (a) along fiber
direction; (b) normal to fiber direction.70 (b)

(a) 2
0
Amplitude (dB)
–2

–4

–6

(b) 0 0.5 1.0 1.5 2.0 2.5 3.0

(30) (60) (90) (120) (150) (180)

Rotation angle β, rad (deg)

Legend
α = angle of incidence
β = angle between Y axis and transmitter beam trajectory on layer
plane

384 Ultrasonic Testing

that the scattered wave is physically from the 1.57 rad (90 deg) fibers in a
separated from the specular reflection. [0,±45,90]S laminate. This fact allows the
detection and imaging of transverse
Using the setup shown in Fig. 31a,83 it cracks, using a C-scan system, by
was observed that backscattering occurs discriminating the scattering amplitude of
only when the angle of insonification is the cracks from that of the fibers. An
normal to the fiber axis.81 As shown in example image of transverse cracks in
Fig. 31b, a laminate with a [0,±45,90]S graphite epoxy is shown in Fig. 33.
layup (0 indicates symmetric) gives rise to
a maximum backscatter each time the Porosity is another discontinuity for
beam is normal to the fibers. The finite which backscattering provides unique
width of the angular spectrum is information. Generally, porosity tends to
determined by the transducer directivity. accumulate between the layers of

In another study, the signals were FIGURE 32. Backscattering characterization
spatially averaged to reduce the effects of for the orientation of graphite-to-epoxy
local variations and small layers:71 (a) [0]8; (b) [0,±45,90]S;
discontinuities.83 Tests were conducted in (c) [0,90]2S.
the frequency range from 1 to 25 MHz.
Using pulses and a boxcar averager, the (a)
peak amplitude of the backscattering was
measured. (b)

Backscattering at the plane normal to (c)
the fibers, measured as a function of
incident angle, showed two maxima at FIGURE 33. Acoustic backscattering images
the critical angles. The mechanism of of transverse cracking in graphite-to-epoxy
generating a maximum at the critical laminate: (a) fatigued sample; (b) statically
angles is not clear and is not explained by loaded sample.70
documented theories. Note that the wave
number length ka (k is the wave number (a)
and a is the fiber diameter) is relatively
small and reaches only about 0.2 mm (b) Surface Transverse
(0.008 in.) at 25 MHz for graphite epoxy.
roughness cracks
Fiber orientations and stacking order
determine the final properties of
structural composites. Fiber layup is
identified as follows: fiber orientations are
identified in square brackets, an index
indicates the number of times these
orientations are repeated and an S
subscript denotes that the layup is
symmetric. Using backscattering
measurements, fiber orientation can be
determined ultrasonically by measuring
the spatial distribution of intensity for a
constant angle of incidence.83 This has
been implemented with the aid of a polar
C-scan system. As shown in Fig. 32, the
fiber orientations of graphite epoxy
laminates in several layups are easily
resolved. It was also shown that
backscattering can be used to determine
fiber misalignment and to detect ply gaps.

Matrix cracking is common in
composite materials. It is induced by
mechanical and thermoelastic stresses and
serves as an initiation site for
delaminations under interlaminar shear
stresses. The cracks typically extend
through the total thickness of a lamina (a
stack of layers of a given orientation
within a laminate) and are at least an
order of magnitude larger than the fiber
diameter.

For example, the thickness of a
graphite ply is about 125 µm (0.005 in.)
and the diameter of a graphite fiber is
about 5 µm (0.0002 in.). The
backscattering from transverse cracks is
more than 30 dB higher than scattering

Ultrasonic Testing of Advanced Materials 385

composites. To simulate porosity, a because of (1) insufficient understanding
[0,90]2S laminate was prepared using of this subtle interference phenomenon
microscopic balloons 40 µm (0.0015 in.) and (2) the complexity of its associated
in diameter and a 2 µm (8 × 10–5 in.) shell acoustic field. In particular, it is difficult
thickness, spread in a thin coat between to interpret the received wave for
the first and second layers. Because laminates with multiple orientations.
porosity as randomly spread spheres does Extensive research has been conducted to
not have any preferred orientation, it corroborate a theoretical model for
generates backscattering of equal determining wave behavior.87
amplitude for all angles of insonification.
Leaky Lamb Wave Theory
This behavior is shown in Fig. 34
where porosity causes an increase in For the analysis of leaky lamb wave
scattering for angles not normal to the behavior in composite materials, a matrix
fiber axes. This behavior generates a technique can be used.1 A pitch catch
spatial window in the backscattering field setup has been used to verify the theory: a
through which discontinuities of different receiver was placed at the null zone of the
scattering directivity can be detected and leaky wave field for testing at various
characterized. With limited success, angles of incidence. Pairs of flat broad
attempts have been made to develop a band transducers examined the leaky
theoretical model for the relation between lamb wave field in the frequency range of
porosity and backscattering.84,85 0.1 to 15 MHz.

Leaky Lamb Wave Tone burst signals, with durations
Applications sufficiently long to establish a steady state
condition in the test laminates, were
Leaky lamb waves have been applied to induced with the aid of a function
the nondestructive testing of generator. The received signals were
composites.82,86 The technique can amplified and two aspects of the leaky
provide an excellent nondestructive test lamb wave phenomenon were examined:
for detecting and characterizing various the reflection coefficient and the
discontinuities, including delamination, dispersion curve.
porosity, ply gaps and variations in resin
content.81 At specific angles of incidence, the
amplitude is acquired as a function of
The phenomena of leaky lamb waves frequency. These experimental results
are described above, with reference to were then compared with the theoretical
adhesive bonding. The leaky lamb wave reflection coefficient for a graphite epoxy
field for a typical composite laminate is AS4/3501-6 [O]8 laminate tested at 0, 0.79
shown schematically in Fig. 35. The and 1.57 rad (0, 45 and 90 deg). In testing
technique is applicable even with along the fibers, at 0.26 rad (15 deg) angle
relatively thick laminates. Leaky lamb of incidence, the theoretical data have
waves are not widely used, however, been convolved with the transducer
characteristic response. A relatively good
FIGURE 34. Backscattering from [0,90]2S glass-to-epoxy agreement of theory and experiment was
laminate with and without porosity.70 observed.

0 The phase velocity of the leaky lamb
wave modes was recorded as a function of
the frequency for incidence angles in the
range of 0.2 to 1.1 rad (10 to 60 deg) at
35 mrad (2 deg) increments. These data
allowed the preparation of characteristic

Relative amplitude (dB) –2 FIGURE 35. Schematic diagram of leaky lamb wave acoustic
field for immersed laminate.
–4
Transmitter θ Receiver
–6
Intralaminar porosity
–8
0 Normal composite

0.5 1.0 1.5 2.0 2.5 3.0 Fluid φ X1
(30) (60) (180) Fluid X2
(90) (120) (150)
X3
Rotation angle, rad (deg)

386 Ultrasonic Testing

dispersion curves for lamb waves thickness, making it difficult to
propagating at 0, 0.79 and 1.57 rad (0, 45 discriminate between insignificant
and 90 deg) to the fiber orientation. The variations and the presence of
results were then compared to theoretical discontinuities.
predictions for a [O]8 graphite epoxy
laminate. The agreement of the theory To reduce the sensitivity of the leaky
and the experiment is excellent at lower lamb wave phenomenon to insignificant
frequencies and acceptable at higher changes and to take advantage of the
frequencies. information available in a broad
frequency range, a forward fast fourier
This theory has also been successfully transform is used.91 In this process, a
applied to multiple-layered periodic minimum in the frequency
multiple-orientation laminates, including domain is characterized by a single peak
interfacial matrix rich layers.88,89 The value representing the inverse of the
agreement between theoretical and period of leaky lamb wave excitation
experimental dispersion curves for a modes.
[0,90]2S graphite epoxy laminate is
excellent at lower frequencies and Using this procedure, a substantial
acceptable at higher frequencies. increase in the signal-to-noise ratio (more
than 20 dB) has been achieved and
Leaky Lamb Wave Portable discontinuity detection is significantly
Immersion Fixture improved. To perform such a test, a signal
analyzer can be programmed to perform
The leaky lamb wave method requires a the fast fourier transform of the leaky
fluid medium between the transducers lamb wave spectral response at the null
and the test object. With this constraint, zone. Relatively high noise contaminated
leaky lamb wave tests cannot be used in the signal (1) after the process of
the field without a device that maintains obtaining the spectrum and (2) after the
a water column between the transducers fast fourier transform of the spectrum. To
and the laminate.90 As is evident in the reduce this noise, a cross correlation was
dispersion curves of composites, leaky applied between the resulting fast fourier
lamb wave modes are relatively constant transform and the discontinuity free
for angles of incidence below 0.35 rad reference signal.
(20 deg). Therefore, using a device (known
as a bubbler) with an angle of incidence The location of the peak depends on
lower than 0.35 rad (20 deg) can reduce the depth of the delamination. The closer
sensitivity to surface curvature of the test a delamination is to the upper surface of a
object. laminate, the smaller the value of the
transformed frequency (abscissa of the fast
The bubbler can be provided with a fourier transform graph) of the associated
means of positioning the receiver at the peak.
null zone by simultaneously adjusting the
ultrasonic frequency and the height of the Leaky Lamb Wave Time Domain
transducers until the measured amplitude Analysis
is at minimum. The height adjustment
replaces changing the distance between Although a forward fast fourier transform
transducers and needs to be done only of the leaky lamb wave spectra allows a
once at calibration. Generally, responses significant improvement in sensitivity to
obtained with a bubbler are similar to discontinuities, it is time consuming and
those obtained with the test object requires relatively complex equipment.
immersed in water. The small changes Generally, a fourier transform of a
observed in the measured values are periodic function either forward or
attributed to the difference in boundary backward leads to a similar functional
conditions, namely the presence of a result. This means that a transform of the
stress free back surface when using the leaky lamb wave spectrum to the time
bubbler. domain (backward) or to the transformed
frequency domain (forward) effectively
Signal Processing of Leaky Lamb produces the same data. Time domain
Wave Response response is the commonest form of signal
presentation on commercial ultrasonic
Leaky lamb wave tests for discontinuities instruments.
can be performed in two ways: (1) by
evaluating amplitude changes at a tone Time domain leaky lamb wave
burst frequency that induces a leaky lamb techniques require placing the receiver in
wave mode in a discontinuity free sample the null zone of the reflected wave using a
or (2) by evaluating changes in the tone burst transmitter. Once this
spectral response while sweeping through positioning is properly established, the
a given frequency range. Both of these pulsing transmitter is substituted to
techniques are very sensitive to small induce short pulses instead of tone bursts.
variations in elastic properties or plate The reflected signal can be processed very
rapidly for C-scan or computer analysis
and is associated with a very high

Ultrasonic Testing of Advanced Materials 387

signal-to-noise ratio when compared to In addition, changes in the resin-to-fiber
the forward fast fourier transform process. ratio significantly affected the amplitude.

The advantages of the Using time The variations of the time-of-flight
domain approach, a C-scan image can be values for the [O]24 sample can also be
made for improved sensitivity over recorded on a C-scan image (Fig. 36b).
previously documented results.90 While all the embedded discontinuities
have been detected, they are observed
Time domain leaky lamb wave tests in with a sensitivity superior to the one
the radiofrequency form provide obtained with tone burst leaky lamb wave
significant information about composite tests. The dark lines along the C-scan
laminates. Several unidirectional image are parallel to the fiber orientation
laminates were examined along the fibers and are a result of the migration of the
at 0, 0.79 and 1.57 rad (0, 45 and 90 deg) porosity 40 µm (0.0016 in.) from the
propagation and at 0.26 and 0.35 rad center of the object during the cure of the
(15 and 20 deg) incidence. The responses laminate. A close look at the image shows
along the fibers and normal to the fibers that the porosity also migrated through
show repetitive reflections with a constant the plies during cure.
time of flight between them. The time
duration between the reflections is Different colors were assigned to the
determined by the thickness of the various time-of-flight ranges, thus
laminate. This result conforms with the presenting the depth of the
results from the forward fast fourier discontinuities in the sample. The depth
transform process, where the location of distribution of the trapped clusters of
the peak correlates well with the porosity can be identified by the color of
thickness. pixels around the porosity. The leaky
lamb wave theory has been applied to
When testing the sample normal to the corroborate time domain results. Using
fiber, the reflections pattern is convolution of the transducer impulse
significantly different from the one response in the frequency domain and a
obtained along the fibers. The time fast fourier transform, the frequency
duration between the reflections is no dependent, complex valued reflection
longer constant and the second reflection coefficient was transformed to the time
appears closer than expected in the case domain. The theoretical pulse responses
of the internal specular reflection of a
longitudinal wave. If this were a FIGURE 36. C-scan image of [0]24 graphite-to-epoxy laminate
longitudinal mode, it would have been tested with pulsers at 0.75 rad (45 deg) to fiber orientation:
three times slower and would have
appeared much later. (a) amplitude; (b) time of flight.

When testing a unidirectional laminate (a) 150 (6)Y dimension
at 0.79 rad (45 deg) from the fiber mm (in.)
orientation, reflections beyond the first 125 (5)
one are weak and complex. On the other 100 (4)
hand, at this direction of wave
propagation, discontinuities produce 75 (3)
significant reflection amplitude with the 50 (2)
time of flight correlated to discontinuity 25 (1)
depth.
0
Generally, it is determined that the
leaky lamb wave behavior, as a function 30 60 95 125 155 190 220 250 280
of the fiber orientation, is responsible for (1.3) (2.5) (3.8) (5) (6.3) (7.5) (8.8) (10) (11.3)
the above phenomena. In a pulsed form,
these phenomena are sensitive to changes X dimension, mm (in.)
in boundary conditions. For example, a
delamination (a stress free surface) (b) Delamination Porosity Delaminations
completely changed the response at
0.79 rad (45 deg). 150 (6)
Y dimension, mm (in.)
This technique has been used with a 125 (5) Porosity
C-scan setup and a time gate placed over 100 (4) Migration
the range beyond the position of the
specular reflection to produce a C-scan 75 (3)
image. The host computer acquired the
peak-to-peak value of the maximum 50 (2)
reflection as well as the time-of-flight
value. Figure 36a shows a C-scan image 25 (1)
presenting the amplitude variations
obtained within the time gate. As can be 0
seen, all three types of imbedded 30 60 95 125 155 190 220 250 280
discontinuities (delaminations, ply gap (1.3) (2.5) (3.8) (5) (6.3) (7.5) (8.8) (10) (11.3)
and porosity) are clearly distinguishable.
X dimension, mm (in.)

388 Ultrasonic Testing

for various propagation orientations with Only a leaky wave whose critical angle θR
the fibers agree with experimental data. is less than the half-aperture angle θm of
the transducer, θR < θm, can be excited
A technique for determining elastic and detected in this scheme. The time
constants of laminar composite materials
using line focused acoustic microscopic delay Δz in the responses to the ray R is
experiments is based on time domain
response studies.92 The microscopy directly related to the velocity CR of the
response is complicated by multiple leaky surface lamb waves:95,96
reflections in the layers and also by the
anisotropic nature of the material. The (6) CR = 1
model used is based on a stable, recursive,
stiffness matrix algorithm that can be Δt − 1 ⋅ ⎛ Δt ⎞
applied to the interpretation of the time C ⋅ Δz 4 ⎝⎜ Δz ⎟⎠
resolved acoustic microscopy signature. It
has been shown that the fluid load has a The maximum value of Δz is limited by
significant effect on the leaky surface
waves in these materials, increasing the the focal distance of the lens and the half
wave speed above that for the slow aperture angle θm. Thus Δz < F·cos(θm).
transverse wave. This results in its absence The maximum value of F is usually
from the microscopy signature of the
surface wave. Time resolved acoustic limited by the sound attenuation in the
microscopy has been applied to the
determination of elastic constants of a liquid. It is also possible to obtain better
unidirectional composite or of one lamina accuracy by decreasing θm, but this is not
of a cross ply composite. The lateral waves desirable because of the reduction in the
and multiple reflections of bulk waves
appearing in the microscopy signature are critical angle range. However, the
used for the elastic properties accuracy does increase with increasing Δz.
reconstruction. The reconstruction results
can then be compared to the data Extensions of the V(z) technique have
obtained by the self-reference,
double-through-transmission bulk wave been described for another configuration
technique.
in which the received voltage V(x,t) is
Other advances in high resolution
acoustical imaging and quantitative acquired during relative translation along
acoustical microscopy for advanced the specimen surface.96 Angular resolution
material evaluation based on leaky surface
lamb waves (sometimes called simply leaky of the measurements in the V(x,t)
surface acoustic waves) have been
demonstrated. The most popular technique is better than the V(z)
quantitative technique in acoustic
microscopy is the V(z) technique, in technique for small incident angles.
which the acoustic velocity and
attenuation of the leaky surface acoustic Computerized Testing
waves can be determined from the output
signal V of the transducer, which is Nondestructive testing of composites and
acquired as a function of specimen other advanced materials requires
displacement relative to the ultrasonic gathering many data, statistically
sources and receiver.93,94 The output processing them using numerical
voltage is recorded as a function of the techniques and comparing them with a
distance between the focus and the reference data bank. These tasks are time
surface of the specimen. The phase consuming and can include human error
velocity and propagation attenuation of if done manually, so they are ideally
the leaky surface acoustic waves, as well as
the reflectance function for the FIGURE 37. Ray model of the V(z) technique for high
specimen-to-liquid coupling interface, can resolution acoustic imaging.
be obtained from the recorded amplitude
data. The experimental configuration as D
shown in Fig. 37 is called the V(z) system. R
In this arrangement, the leaky surface
lamb waves are generated by a ray θR
incident on the liquid-to-solid interface at
the critical angle θR. A surface wave AB
propagates along the interface reradiating
back to the liquid at the same angle θR. As θm Δz
the transducer is moved toward the
specimen, one of the reradiated rays is Legend
effectively received by the transducer. A = ray of reflected wave
B = ray of reflected wave
D = ray of normal incidence

θm = half-aperture angle
θR = critical angle
Δz = time delay related to test object thickness

Ultrasonic Testing of Advanced Materials 389