• ultrasonic testing;
  • prior austenitic grains;
  • Kohn etching;
  • Marshall etching;
  • martensite


  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

This study is focused on the sensitivity of ultra-sonic test (UT) performed on special carbon steel as a function of their microstructure. UTs was carried out on several samples divided into two families and featured by different process-histories. The estimation of the attenuation of the ultrasonic signals as a function of the microstructural features was performed. The samples underwent different thermal treatments to modify their parent austenite grain size and the associated martensite islands by re-austenitization process and by the following grain growth.

The difference of UT results mainly depends upon the wave scattering and the energy absorption due to the difference in the grain size and in the crystal lattice orientation.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

The inspection procedure and the non-destructive testing (NDT) during the manufacturing, the construction and the operation steps cover a fundamental role for granting the reliability of the metal products. In this scenario, ultrasonic tests represent an interesting tool to determine the quality of the material.[1]

The intensity of the ultrasonic wave decreases as the distance from the source increases during the propagation through the medium and this is due to the wave energy loss. These losses are due to diffraction, scattering, and absorption mechanisms, which take place in the medium.[2] The attenuation is caused by the sound energy loss as the ultrasonic beam crosses the investigated material. The intensity of an ultrasonic beam received by a transducer is significantly lower than the intensity of the initial transmission. The factors that are primarily responsible for the beam intensity loss is attributed to transmission losses, interference effects, and beam scattering.[3] In this study, the ultrasonic signal attenuation has been estimated as a function of the microstructural features.

The study aims at stating a relation between the ultrasonic attenuation and the lattice features of a special carbon steel. One of the most important factors affecting the ultrasonic measurement is the mean grain size of the metal and in this study the effect of the prior austenitic grain (PAG) size has been taken into account.[4]

Revealing the PAG is the first fundamental experimental input to obtain the austenitic grain size and it helps to correlate it with the ultra-sonic test (UT) response.[5] The measurement of the PAG size is needed because it determines the resulting martensitic grain size after quenching and tempering.

The Kohn metallographic etching[6] gives the correct image of the PAG and allows a simple count of the grain size that is useful to correlate the microstructural features with the UT response.

The Kohn method involves preferential transfer of material away from grain boundaries when the steel is exposed to high temperature in inert atmosphere. Thus, during austenitization of a pre-polished sample, grooves are formed at the austenite grain boundaries that emerge in the polished surface. These grooves remain intact after cooling and are clearly visible at room temperature outlining the austenite grain boundaries.

The obtained metallographic results have been discussed in order to understand the phenomena involved in the observed loss of transparency featuring the material in certain conditions.

2 Experimental Procedure

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

2.1 Specimens Preparation and Heat Treatments

The investigated samples are 26NiCrMoV11.5 steel (Table 1). This steel is applied after quenching and tempering and it is often used for the production of forged components.

Table 1. Chemical composition of 26NiCrMoV11.5

The specimens are two square bars (40 mm × 40 mm) 750 mm in length. The C bar and the D bar were produced by forging in the temperature range 1250–870°C by three re-heating steps, imposing a reduction ratio of 5, then they were oil quenched. The C bar has previously undergone a chemical homogenization thermal treatment at 1250°C for 48 h, to remove bandings and segregations.

In order to perform different thermal treatments on each sample, the C bar was cut into five pieces and the D bar into four pieces. The samples underwent different austenitization treatments (Table 2) to induce different microstructural features in term of PAG size.

Table 2. Austenitization parameters applied on each sample
NomenclatureTemperature [°C]Time [h]Cooling
C19000.5Water cooling
C29001Water cooling
C39004Water cooling
C49008Water cooling
C590016Water cooling
D112500.5Water cooling
D212501Water cooling
D312504Water cooling
D412508Water cooling

Argon atmosphere was employed during the austenitization process and only in the last 10 min the samples were exposed to oxidizing atmosphere. The sample oxidation during the last 10 min of treatment and the following water cooling is needed to perform the Kohn etching, used to detect and measure the PAG size (as described in the paragraphs 2.3).

2.2 Attenuation Measurements

Ultrasonic testing is one of the preferred NDT technique for the material characterization. UT values are significantly affected by the microstructural features. A traditional experimental set-up was used for the ultrasonic measurements (Figure 1).


Figure 1. Sketch of the experimental set-up for ultrasonic measurements.

Download figure to PowerPoint

Ultrasonic tests were performed according to the contact pulse-echo method by Krautkramer USN 60 unit using 2 MHz longitudinal wave probe at room temperature. Ultragel was used as couplant for the probe, because it eliminates the air gap between the transducer and test specimen and it provides efficient transmission of waves from the transducer to the material.[7]

Attenuation can be determined by evaluating the multiple backwall reflections pointed out by typical A-scan display. The decibel number between two adjacent signals is measured and this value is divided by the time interval between them. This calculation determines an attenuation coefficient in decibels per unit time.[8] The ultrasound attenuation measurements have been expressed in term of minimum detectable defect (MDD), because this parameter is directly proportional to the signal attenuation. Moreover this parameter is widely used in the UT unit of forging plant and in many cases the transducer give simply this value instead the attenuation in dB.

In this work, the MDD value is considered acceptable if it is lower than 1 mm. This value usually complies with the strictest specifications required by the engineering companies and by the users.

To better understand the attenuation behavior the average grain size is related to the wavelength. The ξ factor (1) was calculated for all specimens and test conditions as

  • display math(1)

where λ is the wavelength, V the sound speed in the medium and ν the sound frequency. If the average grain size is lower than ξ, scattering is negligible and attenuation depends on absorption phenomena.[3] On the contrary, if the average grain size is >ξ, scattering conditions are satisfied.

2.3 Etching and Grain Size Measurements

Several procedures can be used to easy detect the PAGs, if the steel shows austenitic transformations during reheating. In this study the Kohn method was chosen, because it is simple and it reveals the PAG boundaries very well.

During the samples austenitization, the argon atmosphere prevents the surface oxidation and the decarburizing phenomenon. In the last 10 min of the treatments, the atmosphere is featured by oxygen presence in order to oxidize the sample surface. The oxidation is concentrated on the grain boundaries, because of their high reactivity. After the thermal treatment the samples were fast cooled in water to avoid the nucleation and the growth of ferrite–perlite pattern able to hide the austenite boundaries.

After the cooling and the ultrasonic tests, the surface was grinded and cleaned with 5% picral solution (5% picric acid, 95% ethanol), in order to polish the surface through the removing of the oxidized layer and for revealing only the prior austenite grain (PAG) boundaries.

The Kohn method accuracy was checked by the application of the Marshall etching[9] to compare the gain boundaries pointed out by this second etching. Marshall was performed after the sample cutting, on a central section. It was performed at room temperature and the microstructure was compared with the results obtained by the Kohn method on the surface.

The average prior austenitic grains size was measured using the “Mean Line Intercept Method” as reported in ASTM standard E112. Using this standard the grain size id defined thought the G index:

  • display math(2)

where NL is the number of intercepts per unit length of test line.

For the specimens coming from the D bar, Nital (2%) etching was also performed after grinding and polishing to reveal possible segregation bands or other non-homogeneous areas due to the chemical composition variations.

2.4 EBSD Misorientation Measurements

The influence of grain misorientation[10] on the wave attenuation was estimated by electron back scatter diffraction (EBSD) analysis. This analysis was performed by scanning electrons microscope (SEM) using 20 kV and 1 nA as power source. The samples surfaces were grinded and polished by colloidal-silica (0.20 μm) for 1 h at 50 rpm to allow the signal detection enough reliable to obtain the texture analysis realized by EBSD technique.

The measurements take into account a statistically reliable grains number for each sample.

3 Results

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

The comparison of the PAG structures obtained both by Kohn and Marshall etchings was performed and the grain size revealed by the two etchings on each samples have given the same values (Figure 2).


Figure 2. Example of the prior austenite grains obtained after the application of Kohn etching (a) and Marshall etching (b) on D4 sample.

Download figure to PowerPoint

The results obtained for the two different etching methodologies allow to assess that the average grains size measured from Marshall etching micrographies fit exactly the ones determined by Kohn etching. Thus, although the Kohn procedure implies a re-heating in the austenite field, it does not change the microstructure and it gives the correct information about the austenite grain size. The microstructure transformation due to the austenitization does not affect the Kohn results because it will outline the new austenitic grain boundaries.

Untreated C-bar shows martensitic structure. Figure 3 shows both the martensite microstructure and the PAG one. PAGs was measured on a central surface after the Khon procedure described above, the austenitization time was just 15 min at 900 °C to obtain the present PAG and avoid its growth.[6] PAGs appear as equiaxic grains, without a preferential development direction and some of them seem to be coarsened, actually the grain size distribution is significantly heterogeneous. Indeed, it is possible to observe the size difference among the coarsened grains and the small ones. The G index of the as-received structure of the small observed grains can be stated at value of 8 while the largest grains reach a G index of 5. Also the Marshal procedure confirms these results.


Figure 3. Microstructure of the C-bar in the as-received state before any austenitizing treatment: (a) martensite microstructure (the martensite is present even after normalization, because the studied steel is featured by a chemical composition that can induce martensite by air cooling) and (b) prior austenite grain (PAG).

Download figure to PowerPoint

The austenitization treatments were able to change the PAG structure by nucleation and growth of new grains. The activation of the re-austenitization phenomenon is ruled by the temperature and the time of the applied thermal treatments. As the heat treatment time increases the grain size increases as well because the growth is favored (Figure 4).


Figure 4. PAGs of C samples treated at different holding time: C1, C2, C3, C4, and C5.

Download figure to PowerPoint

The grain size increases (decreasing of G index) by the increasing of the treatment time. Re-austenitization implies the replacement of the formerly grains by new grains due to the phase transformation. The following grain growth takes place favoring some of the re-austenitizated grains at the expense of other ones, causing a heterogeneous microstructure. The growth of the formed new grains at the expense of the polygonized matrix is promoted by the migration of high-angle grain boundaries.[11] The growth of the new grains proceeds rapidly because of the high mobility of the new high-angle grain boundaries, which are featured by misorientation angles ranging between 30° and 40°.

At room temperature, the final microstructure of all the samples is completely constituted by martensite, because water quenching was performed starting from the same austenitization temperature.

The ultrasound attenuation measurement on C-bar samples is expressed in term of MDD, because the minimum measurable size of a defect depends on the signal attenuation. In the industrial practice a MDD equal or lower than 1 mm is considered the limit for good transparency. After the quenching and tempering the values of the MDD point out an increase (loss of transparency) as the PAG size increases (Figure 5).


Figure 5. Minimum size of the detectable defects revealed in C samples (heat treated by different parameters) as a function of the PAG size (expressed in term of G index).

Download figure to PowerPoint

The attenuation increases (and the minimum size of the detectable defects increase) as the austenitization time increases, because it causes a significant grains growth.

The ultrasonic wave scattering occurs at crystal discontinuities, such as grain boundaries, twin boundaries and non-metallic inclusions. These defects tend to deflect ultrasonic energy out of the main ultrasonic beam. Scattering is highly dependent on the crystalline size. When the grain size is <0.01 times the wavelength (ξ), then the scatter is negligible.[3] Table 3 recorded the mean PAG size of the C samples and the theoretical possibility that the scattering can take place.

Table 3. Mean PAGs size and theoretical condition for the wave scattering occurrence
SampleAttenuation [dB mm−1]Sound speed [V m s−1]PAGs mean diameter [μm]ξ [μm]Condition
C10.0953251625.77No scattering
C20.16528722.526.04No scattering

The scattering condition is not verified by C1 and C2: these samples are featured by good transparency. On the other hand, samples C3 and C seem to respect the scattering condition as well. They are interested by attenuation but the transparency reaches acceptable values.

D-bar sample in the as-received condition shows martensitic microstructure (Figure 6a) featured by hues associated to segregative bands as a consequence of the local chemical difference. This is due to the solidification process, as a consequence of dendrites segregation that does not allow the chemical composition homogenization.[12]


Figure 6. D-bar: (a) microstructure in the as received state and (b) microstructure after 8 h of austenitization.

Download figure to PowerPoint

Even after the austenitization thermal treatments, the segregation bands and the composition in-homogeneities do not disappear, pointing out that the thermal treatment at the austenitization temperature is not enough to promote chemical homogenization (Figure 6b). The PAGs reveled by Kohn etching is characterized by equiaxic grains interested also by the significant coarsening phenomena (Figure 7). Indeed, the difference size among the coarsened grains and the small ones are evident. After the quenching treatments, the measurement of the PAGs final size has been performed. The most little grains seem to be maintained in the region interested by significant segregation and this is due to the pinning effect performed by the solute on the grain boundaries.


Figure 7. PAGs of D samples treated at different holding time: D1, D2, D, D3, and D4.

Download figure to PowerPoint

The re-austenitization took place changing the PAG structure and promoting a significant grain growth where the holding time was enough to allow the development of the process. The increase of the PAG final size is associated to an increase of the austenitization temperature holding time (Figure 8).[11]


Figure 8. Minimum detectable defects size revealed in D samples (heat treated by different parameters) as a function of the PAG size (expressed in term of G index).

Download figure to PowerPoint

The D-bar samples in as-received condition (before the development of any re-heating treatment of austenitization), show a MDD higher than 1 mm, that is usually not acceptable for the most requiring industrial specifications. After the quenching and tempering, the new minimum size of the detectable defects shows an increase (loss of transparency) associated to the growth of the PAGs size (Figure 8).

Although sample D (corresponding to as-received condition without any application of austenitizing treatment) is also featured by fine PAG that induces an average value of G index that is not particularly high, this condition is not enough to grant a good transparency. Thus, the presence of large grains seems to be the ruling factor determining the steel transparency.

D1 sample does not satisfy the theoretical scattering condition and therefore it is not certainly interested by scattering (Table 4), although D2 sample should be affected by scattering of the ultrasound wave, it shows an acceptable attenuation value.

Table 4. Mean PAGs size and theoretical condition for the occurrence of wave scattering
SampleAttenuation [dB mm−1]Sound speed [V m s−1]PAGs mean diameter [μm]ξ [μm]Condition
D10.1253071925.715No scattering
D20.22526131.425.84No scattering

3.1 EBSD Analysis

The EBSD analysis performed on martensite grains has allowed to quantify the grain boundaries misorientation, that is the rotation angle describing the different orientation of the lattices belonging to two adjacent grains.

The misorientation angles detected in the C and D samples were detected (Figure 9 and 10). It is clear that the most transparent samples are featured by the highest misorientation angles whereas the largest minimum size of the detectable defects are associated to specimens characterized by the lowest misorientation angles (under 15°).


Figure 9. Misorientation angles determined by EBSD test for C specimen: (a) C5 and C4, (b) C2 and C1.

Download figure to PowerPoint


Figure 10. Misorientation angles determined by EBSD test for D samples: (a) D5 and D4, (b) D2 and D1.

Download figure to PowerPoint

4 Discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

The samples featured by the largest PAGs are also characterized by the lowest misorientation angles among the different martensite blocks formed by quenching (by low relative rotation between two adjacent grains).[13] The boundaries featured by angles lower than 15° can be classified as low-angle grain boundaries. The obtained results clearly pointed out that microstructure composed by fine PAGs implies higher fraction of martensite featured by high misorientation angles. This trend is consistent with the thermo-mechanical processes the samples have undergone. After the re-austenitization has occurred, the grain growth is promoted by the good mobility of grain boundaries that is stimulated by high misorientation angles, featured by highest boundary energy. This involves the preferential boundary migration of small austenite grain boundaries and so small grains disappear by a coarsening phenomenon. Some small grains cannot activate the growth process because they are surrounded by the solute elements that pin their boundaries forbidding their migration. Thus, at the first stage of the growth process the variation of the chemical composition (caused by segregation phenomena) induces a very heterogeneous microstructure interested by the simultaneous presence of little and large grains. As the grain growth goes on the largest grains, featured by high boundary curvatures, incorporates the smallest grains that are featured by a less favorable ratio between the surface boundary energy and the bulk energy. Thus, the formerly grown grains incorporate the most little ones characterized by a high boundary curvature and by higher ratio between surface energy and bulk energy.[11]

This microstructural organization affects the transparency observed on each samples: as the grain size increases, the average misorientation decreases and loss in transparency is pointed out.

This is consistent with the wave behavior into the material: low misoriented grains could be considered as a very big grain sized structure. This kind of structure behaves as a big ordered one in which the wave propagation is hindered. In other word, the propagation is easier in a random structure featured by a big amount of small grains high misoriented than in a fewer number of grains featured by long range ordered crystals.

The ultrasonic wave intensity decreases with the distance from source during the propagation through the medium due to loss of energy. These losses are due to diffraction, scattering, and absorption mechanisms, which take place in the medium. The ultrasonic energy absorption by the medium may be due to loss caused by imperfection, phonon interaction with medium and thermo-elastic losses. Scattering energy loss in polycrystalline solids is due to grain boundaries, cracks, precipitates, inclusions, etc.

When the wave front crosses the edge of a reflecting surface, the wave front bends around the edge and produces an interference pattern in a zone immediately behind the reflector, because of the phase differences induced in the different portions of the forward-scattered beam. A sound beam striking an interface (such as a grain boundary) is reflected and refracted and so part of the beam energy is scattered. Beyond the interface, a coherent wave must re-form through phase reinforcement and cancellation interference; then the wave continues to propagate as a modified wave.

The diffraction losses are little or not concerned with the material properties.

A microstructure featured by small grain size is interested by a big amount of grain boundaries involving high beam energy scattering. However, the transparency loss due to such a scattering is lower than the one due to absorption produced by crystal energy dissipation. Thus, the energy absorption due to scattering phenomena is higher in a finer structure but not as high as the attenuation contribution caused by the energy dissipation by a crystal structure developed on a long range. Thus, an ordered and extended crystal structure can perform higher wave energy absorption due to relaxation process: this process is known as thermo-elastic attenuation.[2]

Actually, when the grain size is >0.01 the wavelength, the main energy dissipating phenomenon is the absorption that increase the attenuation.[3] The attenuation of the samples featured by average grain size larger than the boundary wave length, is approximately linearly proportional to the average grain size.[4, 14]

In the case of large PAGs (that imply also large martensite blocks) grown by migration of their boundaries two detrimental phenomena take place. Not only does absorption occur but also acoustic wave refraction unfolds. When ultrasonic wave is incident at particular angles on interface between two grains, transmission and reflection occur at the interface without any change in beam direction. At any other incidence angle, the phenomenon of mode conversion (a change in the nature of the wave motion) and refraction (a change in the direction of wave propagation) must be considered. These phenomena may affect the entire beam or only its portion, and the sum of all the changes that occur at the interface depends on the incidence angle. If the incidence angle is small, the sound waves propagating across the grains will undergo mode conversion at boundaries and this phenomenon causes the simultaneous propagation of longitudinal and transverse (shear) waves in the following grains. When grains misorientations are higher than a critical value, the refracted longitudinal wave will no longer propagate across the grains and only refracted (mode-converted) shear wave can propagate across the following grains beyond the boundary. The grown coarsened grains are featured even by significant fraction of high misoriented grain boundaries and this situation gives sits contribution to the wave attenuation.[14, 15]

5 Conclusions

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental Procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions

The effect of PAGs size on the attenuation of ultrasonic test have been studied on two different samples families. The two families are featured by the same chemical composition and the same geometry but characterized by two different thermo-mechanical treatments.

Once the initial attenuation value was measured, the samples underwent different heat treatments. The austenitization time was changed in order to promote different distribution of the austenite grain size.

  • The finer are the prior austenitic grains the less is the steel attenuation;
  • the PAG size influences the martensite block size and the consequent energy wave dissipation;
  • the scattering is not the main responsible for the attenuation of the acoustic wave, on the contrary the phenomenon is mainly ruled by absorption due to ordered crystal lattice;
  • the segregated and banded microstructure (i.e., pointed out by D samples of this studies) can cause a very heterogeneous distribution in the grain size due to the pinning effect performed by the increase of the solute alloying elements: the large coarsened grains cause the strong wave attenuation, so in a heterogeneous microstructure featured by large and little grains the maximum grain size is more significant than the average one;
  • low misoriented martensite blocks are induced by large PAGs (usually associated to abnormal grain growth of the grains) and they decrease the ultrasound wave propagation, actually the wave energy is dissipated during the crossing of an extended ordered crystal;
  • the higher the PAG size is, the lower the misorientation of the martensite blocks, the higher the attenuation of the ultrasound wave;
  • for detecting defects in steel components by ultrasonic testing it is important to assure homogeneous chemical composition (without significant traces of segregation) and to control re-austenitization phenomenon by correct heat treatment, because fine grain size distribution grants good transparency. Thus, the heat treatment parameters do not have to promote PAG growth.
  • 1
    VV AA., in Le prove non distruttive, Associazione Italiana Metallurgia, Milano 1999, p. 165.
  • 2
    D. K. Pandey, S. Pandey, in Ultrasonics: A Technique of Material Characterization, Acoustic Waves, Don Dissanayake 2010, p. 414.
  • 3
    VV AA., in Asm Metals Handbook Volume 17, Nondestructive Evaluation and Quality Control, ASM International, Materials Park, Ohio 1992, p. 345.
  • 4
    B. Ahn, S. Seok Lee, IEEE T-UFFC 2000, 47, 620.
  • 5
    C. G. de Andres, M. J. Bartolome, C. Capdevila, D. San Martin, F. G. Caballero, V. Lopez, Mater. Charact. 2001, 46, 389.
  • 6
    G. F. Vander Voort, Ind. Heat. 2010, 78, 48.
  • 7
    ASTM International, E 1158-09, Standard Guide for Material Selection and Fabrication of Reference Blocks for the Pulsed Longitudinal Wave Ultrasonic Testing of Metal and Metal Alloy Production Material.
  • 8
    J. Krautkramer, H. Krautkramer, in Ultrasonic Testing of Materials, Springer-Verlag, Berlin 1990.
  • 9
    A. O. Benscoter, M. J. Perricone, Microsc. Microanal. 2005, 11, 76.
  • 10
    U. F. Kocks, C. N. Tomè, H. R. Wenk, in Textures and Anisotropy, Cambridge University Press, Cambridge 1998, p. 205.
  • 11
    F. J. Humphres, M. Hatherly, in Recrystallization and Related Annealing Phenomena, Elsevier, Oxford 2004, p. 154.
  • 12
    W. Nicodemi, in Metallurgia, Zanichelli, Milano 2002, p. 45.
  • 13
    J. W. Morris, ISIJ Int. 2011, 51, 1569.
  • 14
    N. Grayeli, J. C. Shyne, in Review of Progress in Quantitative Nondestructive Evaluation, Plenum Press, New York 1985, 927.
  • 15
    W. N. Reynolds, R. L. Smith, J. Phys. D: Appl. Phys. 1984, 17, 106.