- Email: [email protected]
Ultrasonic characterization of heat-treatment effects on SAE-1040 and -4340 steels
Journal of Materials Processing Technology 216 (2015) 188–198
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Ultrasonic characterization of heat-treatment effects on SAE-1040 and -4340 steels Magdy M. El Rayes ∗,1 , Ehab A. El-Danaf 2 , Abdulhakim A. Almajid Mechanical Engineering Department, College of Engineering, King Saud University, P.O. Box 800, 11421, Riyadh, Saudi Arabia
a r t i c l e
i n f o
Article history: Received 13 May 2014 Received in revised form 4 September 2014 Accepted 5 September 2014 Available online 16 September 2014 Keywords: Ultrasonic characterization Sound velocity Attenuation Heat treatment of steel Microstructural phases Mechanical testing
a b s t r a c t In this work microstructural characterization and mechanical testing results were correlated with ultrasonic velocity and sound attenuation of steels SAE-1040 and SAE-4340. Both types were subjected to three types of heat treatment; the first was annealing at 850 ◦ C, the second was austenitizing at 1000 ◦ C followed by oil quenching and the third was similar austenitizing then water quenching. Treatments of SAE-1040 steel resulted in microstructures containing different ferrite and pearlite contents, different inter-lamellar spacing and also different grain size. Similar ferrite and pearlite content was obtained when annealing SAE-4340 whereas, oil and water quenching resulted into martensite. With SAE-1040, the sound velocity was reduced in the order of annealing–oil–water quenching due to the reduction of ferrite on the expense of pearlite. The same order in sound velocity reduction was also obtained with SAE-4340 due to the change in microstructural phases from pearlite to martensite. In comparison to pearlite, the martensite possessed higher crystal lattice distortion, higher dislocation density and lower elastic modulus all of which contribute in reducing sound velocity. Attenuation of SAE-1040 increased in the order of annealing–oil–water quenching because of higher pearlite content and the reduction in inter-lamellar spacing. Attenuation of SAE-4340 gave an opposite order due to the reduction of the extent of microstructural anisotropy. The mechanical properties and hardness were predominantly affected by the microstructural phases leading to the logical correlation with ultrasonic parameters. © 2014 Elsevier B.V. All rights reserved.
1. Introduction One of the prime objectives of non-destructive testing (NDT) is to certify that the component being examined is fit for the intended service. The most common way of doing so is by examining the component with NDT to detect flaws or discontinuities such as voids, inclusions, cracks, in materials or structures. Another parameter which is equally important to flaw detection is to assess the material properties. Among various NDT methods, ultrasonic testing (UT) was applied rather extensively in a variety of publications related to the correlation of ultrasonic measurements with microstructural phases of steel as conducted by Gür and Tuncer (2004), duplex stainless steel by de Albuquerque et al. (2010a,b), Ni-base super alloy by de Albuquerque et al. (2012), thermally aged Nibase alloy by Nunes et al. (2013) and grain size as by Bouda et al.
∗ Corresponding author. Tel.: +966 11 467 9906/50099 1132; fax: +966 4676652. E-mail addresses: [email protected], [email protected] (M.M. El Rayes). 1 On leave from Production Engineering Department-Faculty of EngineeringAlexandria University. 2 On leave from Mechanical Design and Production Department-Faculty of Engineering-Cairo University. http://dx.doi.org/10.1016/j.jmatprotec.2014.09.005 0924-0136/© 2014 Elsevier B.V. All rights reserved.
(2003). UT results were also correlated with mechanical properties as investigated by Vijayalakshmi et al. (2011) and de Albuquerque et al. (2010a,b), as well as residual stresses as by Chaki and Bourse (2009). Therefore, the utilization of ultrasonic techniques to indirectly determine the microstructural features and the mechanical properties can be useful for numerous industrial applications. Plain carbon steels SAE-1020 and -1050 were heat treated by varying the austenitization temperature between 860 ◦ C and 1060 ◦ C, as well as the cooling rate was varied between furnace cooling, air cooling and oil quenching. These materials were characterized by ultrasonic velocity and attenuation measurements as reported by Gür and Keles (2003). Metallographic studies revealed that the amount of proeutectoid ferrite, the softer phase, in SAE1020 was higher than that in SAE-1050. Ultrasonic results showed that sound velocities vary depending on the severity of cooling. The lowest sound wave velocity was found with the oil quenched SAE1050 consisting of martensite, which possessed high dislocation density, distortion of crystalline lattice and maximum hardness among other specimens. On the other hand, the highest sound velocity was obtained with furnace cooled SAE-1020 which had the softest structure. It was also found that prior austenite grain size; and not the transformation products within it, as manifested
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198 Table 1 Chemical composition of steels SAE-1040 and -4340 in wt.%. Steel grade SAE
C
Mn
P
S
Si
Cr
Mo
Ni
1040 4340
0.41 0.36
0.6 0.25
0.04 0.04
0.05 0.04
0.3 0.25
– 1.4
– 0.2
– 1.4
by the author, had the predominant effect on the attenuation values. The larger the prior austenite grain size the higher the sound attenuation. Slower cooling rates led to larger prior austenite grain size, larger amounts of soft ferrite and wider interlamellar spacing between cementite [Fe3 C] plates of pearlite. Prasad and Kumar (1994) studied the influence of varying the grain size of cast steel on ultrasonic velocity and attenuation. Various grain sizes were achieved via hot upsetting at different percentages of height reduction. These samples were further heat treated through annealing, normalizing and hardening using oil quench followed by tempering. They reported that increasing the degree of deformation decreases the ultrasonic velocity. However, numerous research work concluded the fact that the longitudinal ultrasonic velocity varied from grain to grain because of misorientation of grains, which was related to the variation in the elastic constant as well. The objectives of the present work were to correlate ultrasonic measurements namely; ultrasonic longitudinal sound wave velocity and ultrasonic attenuation with the microstructural and mechanical properties of two types of steel which were subjected to different heat treatments. These treatments were selected to obtain different microstructural phases as well as different grain size. 2. Materials and methods Two types of steels, namely SAE-1040 (C-Mn steel) and SAE4340 (Cr-Ni-Mo steel), supplied in the form of 50 mm diameters rods, from which both materials were cut into three equal pieces 100 mm long using hack saw with lubricant to avoid excessive heating. The selection of these steel types was based on their importance in various fields of industrial applications in which they are mainly characterized by durability. Steel 1040 is normally used in axels, crank shafts and gears, whereas, steel 4340 is used in air craft landing gears, power transmission gears and shafts. In addition, the types of heat treatment chosen were expected to result into different phases that were ought to significantly affect the microstructural, mechanical and ultrasonic characteristics. Table 1 presents the average chemical composition of these steels in wt.% after three runs for each type using spectroscopic chemical analyzer. Both materials were subjected simultaneously to the same heat treatment type in an electric resistance furnace. The first treatment was austenitizing at 850 ◦ C for 2 h followed by furnace cooling (full annealing). The second and third were austenitizing at 1000 ◦ C for 3 h followed by oil and water quenching, respectively. The heat treated samples were sectioned into disks using hack saw with coolant to extract ultrasonic measurements samples [50 mm long], microstructural, hardness and tensile test samples. The microstructure was examined by secondary electron imaging using scanning electron microscopy (SEM), and electron back scattered diffraction (EBSD). For SEM the samples were prepared according to standard metallographic sample preparation which includes grinding using SiC sand paper, polishing using diamond paste of 1.0 and 0.05 m, and etched with 5% Nital to reveal the samples’ microstructure. The microstructure for both materials was also studied by EBSD using Oxford HKL system incorporated on a field emission scanning electron microscope (FESEM) 7600 JEOL. These samples were polished with colloidal silica as a final step prior to imaging.
189
The ultrasonic measurement samples were machined using end milling then surface ground on both sides in order to ensure complete parallelism between faces. In order to eliminate roughness, visible irregularities and any oxides which might affect the ultrasonic measurements, the samples were further processed similar to the metallographic preparation steps. For ultrasonic measurements, the pulse-echo technique and direct contact method were applied to obtain ultrasonic velocities and attenuation coefficients using ultrasonic pulse-receiver equipment-Karl Deutsch (Model: Echograph-1085). Ultrasonic measurements for all samples were obtained by using commercial NDT ultrasoniclongitudinal wave transducers of 4 MHz. The coupling material Karl Deutsch-Ecotrace gel was used for the longitudinal wave measurements. During measurements a constant load was applied to the probe against the specimen surface so as to have a constant thickness of couplant layer at the interface between the specimen surface and the probe. Ultrasonic velocity was determined by dividing twice the specimen thickness by time of flight (TOF) obtained between zero crossing of the first and second back-wall echoes using Eq. (1) as applied by Vijayalakshmi et al. (2011). In order to check results repeatability, seven ultrasonic readings of each specimen were averaged to represent the data obtained, which gave an error of around ±0.5% with both types of steel. Velocity (m/s) = 2 ×
thickness (m) time (s)
(1)
During the same measurement, ultrasonic velocity was measured again using a function in the ultrasonic equipment which automatically calculates the ultrasonic velocity when locating two corresponding points on two successive echoes (first and second echoes) at the extreme sides of the screen with an error of less than the above mentioned one. The ultrasonic attenuation values were calculated according to Eq. (2) which is based on the reduction of the amplitude of an ultrasound pulse, measured in decibels per millimeter (dB/mm). This equation appeared in several publications such that of Stella et al. (2009), Vijayalakshmi et al. (2011) and Freitas et al. (2011) and given in as: ˛=
A0 20 log 2x A1
(2)
where ˛ is the attenuation coefficient [dB/mm], x is the thickness of the sample measured in the test [mm], A0 is the amplitude of the first echo in dB and A1 is the amplitude of the second echo. The constant 2 is because the pulse-echo technique is used. For each sample, the measurements were repeated seven times with maximum error of ±0.1% with both steels. Vickers macro-hardness tests were conducted using 10 kg force and were performed five times for each heat treated specimen and the average was taken. In order to evaluate the mechanical behavior of heat treated steels, tensile specimens were extracted from the center of the disk along its axis and were cut according to ASTM E-08. Tensile tests were conducted at room temperature and a cross-head speed velocity of 2 mm/min using Instron machine model 3385 H. The machine was equipped with a computer having software through which the load-elongation data were recorded. The tensile test for each type of steel and treatment was repeated three times and the average value was taken and presented hereafter. 3. Results and discussion 3.1. Microstructure Fig. 1a–c shows the SEM microstructures obtained with the three different treatments applied on SAE-1040 steel. Fig. 1a shows
190
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 1. Microstructures obtained by SEM of steel SAE-1040 in (a) annealed, (b) oil quenched and (c) water quenched conditions.
the annealed microstructure which contained pearlite colonies (light) in a matrix of ferrite (dark). The pearlite contained alternate lamellas of eutectoid ferrite (Fe) and cementite (Fe3 C) with random orientation. The pearlite constituted 46% volume-fraction, whereas ferrite was 54%. The average grain size of this structure was around 18 m. The oil quenched microstructure, shown in Fig. 1b, was also composed of pearlite and ferrite which transformed from austenite upon quenching. This microstructure conforms to the continuous cooling transformation (CCT) diagram of SAE-1040 steel. Measurements of phases yielded about 80% pearlite and 20% ferrite and mean grain size of 31.1 m. It can be also noted from Fig. 1b that the pearlite increased on the expense of ferrite and also the interlamellar spacing were much reduced due to the fast cooling rate compared to the annealed treatment. The water-quenched microstructure was similar to that obtained with oil-quench with respect to its pearlite and ferrite constituents. However, the volume fraction of pearlite increased to 92% still on the ferrite expense 8% as well as more dense interlamellar spacing between Fe3 C and Fe took place. The mean grain size corresponding to this treatment was around 54.5 m. Hence, it can be stated that the main difference in the microstructure resulting from the three treatments was the content and size of pearlite and ferrite phases as well as the interlamellar spacing. Fig. 2a–c shows, at higher magnifications, the microstructure corresponding to annealed, oil and water quenched respectively, with emphasis on pearlite structure. In the annealed condition the interlamellar spacing was relatively wider than that occurring with oil and water quenched structure where these spacing became much less, i.e. denser lamellas. Measurements of spacing in the annealed sample gave an average value of 0.362 m, whereas with the water quenched sample was much narrower around 0.0178 m. Fig. 3 shows an example of these measurements with SAE-1040 in the annealed condition using a software available in SEM. Fig. 4a–c shows the orientation imaging grain boundary maps representing the extent of sub-grain boundaries in red color, defined with misorientation less than 2◦ for SAE-1040 steel in the annealed, oil and water quenched conditions, respectively. Also shown in the same figure, the true grain boundaries, with misorientation larger than 15◦ across them, displayed in thick black color. Beside each grain boundary map the respective histogram for the misorientation angle distribution is presented. It is evident that the annealed structure had a relatively large average misorientation angle of about 35◦ with a relatively low percentage of low angle grain boundaries (LAGBs) of 12%, which is a typical finding for well recrystallized annealed microstructures. The oil quenched sample revealed a relatively good amount of LAGBs with a percentage of about 67%, and an average misorientation angle of 12◦ . The water quenched sample revealed, even more LAGBs with a percentage of about 73% and an average misorientation angle of 16◦ . Fig. 5a–c shows the annealed, oil and water quenched microstructures respectively of SAE-4340 steel taken by SEM, respectively. The annealed structure was composed of grains of pearlite (light) in a matrix of ferrite (dark). Measurements showed that this microstructure is composed of 59% pearlite and 41% ferrite. The grain size measured is around 11.6 m. The oil quenched structure, Fig. 5b, is composed of long martensite laths (light) coexisting with a substructure of few plate martensite (regions marked P-dark), which normally consists of fine internal twins as referred in ASM Metals Handbook. The water quenched microstructure is composed of 100% martensite with relatively short laths as in Fig. 5c. It should be noted that due to the presence of martensite laths, it was not possible to see the prior-austenite grain boundaries clearly in the quenched samples. Due to difficulties in indexing the Kikuchi patterns of martensite obtained in the oil and water quenched 4340, the EBSD study was
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
191
Fig. 3. Measurements of pearlite interlamellar spacing in SAE-1040 in the annealed condition.
4340 displayed a relatively higher amount of LAGBs almost double that displayed in the annealed 1040. 3.2. Hardness testing Vickers hardness measurements were conducted to indicate the hardness of phases obtained from different heat treatments. The results of the metallographic examinations are verified by hardness measurements. Fig. 7a and b shows that hardness increased when the heat treatment of both steel types changed from annealing to oil and to water quenching. With SAE-1040 steel, Fig. 7a, hardness increased due to the increase in pearlite content (hard phase) with respect to the ferrite (soft phase), as shown earlier in microstructural results. Similarly, with steel SAE-4340, the hardness increased due to the change in microstructural phases being pearlite + ferrite obtained from annealing, martensite transformed at slow and fast cooling rates obtained from oil and water quench, respectively, as depicted in Fig. 7b. The softness of oil-quenched martensite compared to the water-quenched one is due to that oil quenching leads to less extent of lattice distortion, residual stresses and dislocation density as reported by Gür and Tuncer (2004). 3.3. Tensile testing Table 2 summarizes the results obtained from tensile tests conducted on both types of steels subjected to the different heat treatments. The results of tensile tests are in line with hardness results, which in turn can be related to the phases present in the microstructure. With both types of steel, the ultimate tensile strength (UTS) and yield strength (YS) increased when the treatment was changed from annealing to oil and water quenching, whereas the elongation was reduced in the same order. The phases of steel obtained, as expected, were responsible for the increase in UTS and YS and the reduction in elongation which is similar to that found by Gür and Tuncer (2005). In SAE-1040, increasing Fig. 2. Variation of interlamellar spacing with heat treatment of SAE-1040 (a) annealed, (b) oil quenched and (c) water quenched.
confined to the annealed 4340. The structure is presented in Fig. 6. The average misorientation angle exhibited a value of about 33◦ with a relative low percentage of LAGBs of 25%. Worth noting, that both SAE-1040 and -4340 steels in annealed condition displayed almost similar average misorientation angle; however the annealed
Table 2 Mechanical properties of SAE-1040 and -4340 steels. Sample
UTS, MPa
YS, MPa
Elongation %
1040-annealed 1040-oil quench 1040-water quench 4340-annealed 4340-oil quench 4340-water quench
625 798 854 693 1721 1947
326 464 690 406 1425 1633
37.4 26.6 22.4 36.6 23.1 21.6
192
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 4. Orientation imaging grain boundary mapping (left) and misorientation angle distribution histogram (right) of SAE-1040 in the (a) annealed, (b) oil quenched and (c) water quenched structures.
the pearlite content on the expense of ferrite increased the UTS and YS and also reduced elongation as seen in Table 2. A similar trend was noted with SAE-4340 where the UTS and YS increased with simultaneous reduction in elongation accompanying the treatment. The results in Table 2 indicated that for both types of steel the mechanical properties are predominantly affected by the microstructural phases resulting from different heat treatments, which in turn affects the ultrasonic velocity and attenuation as will be shown in the following sections. 3.4. Ultrasonic testing 3.4.1. Sound velocity The ultrasonic velocity measurements of longitudinal waves were sensitive to microstructural variations resulting from different heat treatments applied on SAE-1040 steel. This can be noted when plotting sound velocity versus different fraction percent of pearlite and ferrite phases as in Fig. 8. The reason to this was related to the percentage of pearlite and ferrite phases within the whole structure as well as the interlamellar spacing between ferrite and Fe3 C phases within the pearlite grain. Gür and Tuncer (2004), Freitas et al. (2010) and others have agreed upon that sound velocity becomes lower when the ferrite content is reduced as well as when the interlamellar spacing is narrower compared to coarse pearlite. This is due to the fact that the ferrite phase has the least resistance to ultrasonic waves hence allowing the highest velocity. Simultaneously, pearlite grains with dense/narrow interlamellar spacing possess high resistance to ultrasonic waves and consequently, have low sound velocity as discussed by Freitas et al. (2010). The variation of interlamellar spacing with different treatments was noted in Fig. 2. Therefore, it can be stated that an inverse relation exists between pearlite content and sound velocity. The evolution of sub-grains and sub-grain boundaries contribute in the reduction of sound velocity. Gür and Keles (2003)
reported that the structures consisting of sub-grains and their boundaries have an effect on lattice straining and interrupts the matrix continuity thus lowers the elastic modulus and consequently, lowers the propagation rate of sound velocity. This can be inferred from Fig. 4, which shows a tremendous amount of sub-grain boundaries accompanying the water quench treatment (LAGBs = 73%) which is much higher than that occurring with the annealing treatment (LAGBs = 12%). Furthermore, the variation of the elastic constant from grain to grain in the direction of sound wave propagation may be another reason for the reduction in sound velocity as found by Prasad and Kumar (1994). Papadakis (1964) reported that there is a linear proportion between sound velocity and the elastic modulus of the structure within which the sound wave propagates. The elastic moduli decreases in the order of ferrite–pearlite–martensite. Hence, lower percentage of ferrite with respect to pearlite content corresponds to lower elastic modulus and consequently, lower sound velocity. A similar trend is also found with sound velocity for SAE-4340 when changing the heat treatment from annealing to oil and water quenching causing a reduction in sound velocity as seen in Fig. 9. The reason to this reduction is due to the change in microstructure from pearlite to martensite similar to that concluded by Prasad and Kumar (1994). When quenching, the steel is cooled rapidly from the austenitizing range to room temperature, in which each austenite grain (FCC) suddenly transforms into laths of martensite (BCT) by diffusionless lattice shear as confirmed by Gür and C¸am (2006). This transformation leads to high crystal lattice distortions due to the increase in volume during the austenite–martensite transformation resulting into great amount of internal tension/residual stresses. Therefore it can be stated that martensite is the phase with very high dislocation density and maximum randomness as reported by Papadakis (1964) and Gür and C¸am (2007). In the same context, Gür and Tuncer (2005) reported that the ultrasonic velocity in martensite is essentially affected by changes in the modulus of elasticity of individual grains, in the crystal
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
193
Fig. 6. Orientation imaging grain boundary mapping (left) and misorientation angle distribution histogram (right) of SAE-4340 in the annealed condition.
lattice distortion level and in the orientation of primary austenite grains. The increase in lattice distortion and the subsequent increase in dislocation density reduce sound velocity as reported by Gür and C¸am (2007). This was confirmed earlier by Papadakis (1964) who reported that the elastic moduli decrease in the order of pearlite–bainite–martensite and based on the proportional relation between the elastic modulus and sound velocity it can be stated that martensite has lower sound velocity when compared to pearlite. In addition, martensite presents high resistance to ultrasound waves, because of its compact and fine structure, hence possessing low sound wave propagation velocity as reported by Freitas et al. (2010). These results are in line with those obtained by Papadakis (1964), Gür and C¸am (2006) and Gür and Tuncer (2005).
Fig. 5. Microstructures obtained by SEM of steel SAE-4340 in (a) annealed, (b) oil quenched and (c) water quenched conditions.
3.4.2. Attenuation The attenuation of ultrasonic waves in a polycrystalline material is mainly due to either scattering within the grains and their structural boundaries or absorption due to dislocation damping, thermoelastic losses, and magnetic hysteresis loss or both as reported in various publications such as Papadakis (1964) and Kumar et al. (2002). When calculating attenuation coefficient (˛), most literature such that reported by Liu et al. (2007) have
194
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 7. Hardness variation with different heat treatments with: (a) SAE-1040 and (b) SAE-4340.
applied Eq. (3), which indicates that (˛) is equal to the sum of the absorption coefficient (˛a ) and the scattering coefficient (˛s ) as follows: ˛ = ˛a + ˛s
(3)
Scattering is usually divided into three different regimes depending on the ratio of grain size (d) to ultrasonic wavelength () as summarized by Khafri et al. (2012). These regimes are the Rayleigh region where d, the stochastic region where ≈ d and the geometric region where < d. In the present work the Rayleigh regime is applicable because ranges between 1.353 and 1.366 mm whereas the maximum grain size is about 54 m (depending on the sound velocities measured for each specific sample). In polycrystalline materials, it is the scattering of ultrasound from grains and interfaces that is the main cause of attenuation. In these materials,
grain scattering losses are large compared to absorption losses as agreed by Papadakis (1964) and Khafri et al. (2012). Fig. 10 shows that the attenuation coefficient for SAE-1040 was lowest with the annealed microstructure and gradually increases with oil then water quenched ones. This is due to the increase in pearlite content (i.e. decrease in proeutectoid ferrite), as inferred from Fig. 1, and the decrease in interlamellar spacing in pearlite as shown in Fig. 2, all of which contribute in increasing scattering, hence increasing attenuation. The reason to this is that scattering increases with samples having larger pearlite content and consequently, larger areas of interface [phase boundaries] between hard Fe3 C and soft ferrite, as reported by Gür and Tuncer (2004) and Gür and Keles (2003). In the same sense, ultrasonic wave attenuation due to scattering in pearlite microstructure is greater than that in ferrite because pearlite is more elastically anisotropic than ferrite. This result is confirmed by Ahn and Lee (2000). Since the pearlite content increased when the treatment changed from annealing to oil then water quenching, therefore the interface area increased as well. It was also reported by Gür and Keles (2003) that in pearlite containing steels, the attenuation of ultrasonic beams passing through grains are affected by pearlite dispersion, which conforms to that shown in Fig. 2. Fig. 11 for SAE-4340 shows that attenuation decreased as the steel phases changed from ferrite + pearlite, oil quenchedmartensite and water quenched-martensite. This result is in agreement with that obtained in various earlier works such as that of Kumar et al. (2002, 2003), which was applied on carbon and alloy steels. They reported that the attenuation of pearlite is higher than that of martensite. Gür and C¸am (2007) indicated that upon quenching, the sudden transformation of austenite into martensite resulted into high amount of lattice distortion, high dislocation density and maximum randomness of the structure. From the point of view of scattering theory, the randomness causes an increase in elastic isotropy of the grain volume, thus decreases scattering power as reported by Papadakis (1964) and Kumar et al. (2002). In addition, fine laths of martensite made the structure more isotropic than pearlite and ferrite hence decreasing its scattering power (Gür and Keles, 2003). 3.5. Correlation between ultrasonic and mechanical properties Fig. 12a and b shows the relation between hardness and sound velocity and attenuation coefficient of SAE-1040 steel at different pearlite–ferrite contents, respectively. Generally, it can be stated
Fig. 8. The influence of different pearlite (P) and ferrite (F) contents on sound velocity in SAE-1040.
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 9. The influence of different phases in SAE-4340 on sound velocity.
Fig. 10. The influence of different pearlite (P) and ferrite (F) contents on attenuation coefficient in SAE-1040.
Fig. 11. The influence of different phases in SAE-4340 on attenuation coefficient.
195
196
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 12. Relationship between hardness Vickers and (a) sound velocity and (b) attenuation coefficient at different pearlite–ferrite contents in SAE-1040. Fig. 13. Relationship between hardness Vickers and a) sound velocity and b) attenuation coefficient at different pearlite-ferrite contents in SAE-4340.
that an inverse relationship exists between ultrasonic velocity and hardness, which is in agreement with that found by Gür and C¸am (2007). Since the hardness of the phases increased in the order of increasing pearlite content, a corresponding reduction in the sound velocity was noted. The attenuation coefficient can also be correlated with hardness which in turn is related to the volume percent of pearlite and ferrite and the corresponding extent of elastic anisotropy. In addition, Prasad and Kumar (1994) indicated that hardening process introduces internal stresses into the lattice resulting into lattice distortion/deformation; which is accompanied by an increase in dislocation density thus increasing the attenuation coefficient. The same inverse relation between hardness and ultrasonic velocity was obtained with SAE-4340 steel as shown in Fig. 13a. This is attributed to the difference in microstructures where martensite possesses higher resistance to sound waves as well as higher lattice distortion and dislocation density all of which reduce the sound velocity more than pearlite and ferrite structure does. On the other hand, Fig. 13b shows that increasing hardness markedly reduced the attenuation coefficient. This result is confirmed with that obtained by Seok and Kim (2005). Again, the difference in microstructure was responsible to this reduction. This is due to the fact that laths of martensite
make the structure more isotropic than pearlite + ferrite structure does, hence decreasing its scattering power. This result is in line with that reported by Gür and Keles (2003) and Papadakis (1964) which stated that attenuation in pearlite is higher than that in martensite, which possesses least attenuation even among other steel phases such as ferrite and bainite as reported by Kumar et al. (2002). Fig. 14a shows that the increase of UTS and YS of SAE1040 steel, was accompanied by a reduction in sound velocity. A similar inverse relation, as that of hardness, is obtained due to the proportional relation between pearlite content and hardness on one hand and the mechanical properties on the other hand. Similarly, Fig. 14b plots the strength versus the attenuation coefficient. The strength is related to the increase in pearlite content which in turn causes the structure to be more anisotropic thus increases the attenuation [scattering power] of this structure compared to the less anisotropic one [lower pearlite content]. With SAE-4340 steel, increasing strength was accompanied by a reduction in ultrasonic velocity as in Fig. 15a, which is a trend similar that obtained with SAE-1040 steel. The reason to this reduction
Table 3 Comparison between SAE-1040 and -4340 in annealed condition. Sample
Sound velocity, m/s
Attenuation coefficient, dB/mm
Pearlie %, ferrite %
LAGBs %
1040-annealed 4340-annealed
5913 5906
0.0072 0.0074
46% P, 54% F 59% P, 41% F
12 25
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
197
Fig. 14. Relation between UTS and YS with ultrasonic parameters of SAE-1040 steel: (a) sound velocity and (b) attenuation coefficient.
is related to the type of microstructure, where pearlite offers less resistance to ultrasonic waves thus allowing faster sound velocity to take place compared to martensite which possess higher resistance and consequently lowest speed as discussed by Freitas et al. (2010). The relation between strength and attenuation is plotted in Fig. 15b. Increasing strength was accompanied by a reduction in attenuation. The strength increased because of the change in microstructure from pearlite + ferrite to martensite which directly influenced the attenuation based on the fact that martensite structure is more isotropic thus possessing lower scattering power than pearlite does. This result agrees well with that reported by Papadakis (1964), and Kumar et al. (2002). Although the present work was not intended to compare the two types of steel, however, there were two reasons behind their comparison. First, was that both types of steel were annealed at the same temperature and second was that similar microstructural components were obtained. In light to the earlier discussion, comparison was done as a trial to further verify the influence of phase percent and the sub-grain boundaries on sound velocity and attenuation when steel type is changed. Table 3 summarizes the microstructural phases and grain size and their corresponding ultrasonic measurements conducted with both types of steels annealed at the same conditions. It can be noted that sound waves propagated faster in SAE1040 steel than in SAE-4340. This is due to that the former steel possessed higher percent of ferrite (54%; least resistant to ultrasonic waves) than the latter (41%). Attenuation of SAE-1040, on the other hand, was lower than that of SAE-4340 as in Table 3. This was again due to the lower percentage of pearlite in the former (46%) than in the latter (59%). Higher pearlite content made the structure more anisotropic thus leading to higher attenuation capability than lower pearlite content. Also, the higher percentage
Fig. 15. Relation between UTS and YS with ultrasonic parameters of SAE-4340 steel: (a) sound velocity and (b) attenuation coefficient.
of sub-grain boundaries occurring with SAE-4340 (LAGBs = 25%) compared to that with SAE-1040 (LAGBs = 12%) may be another reason why the former possesses higher attenuation than the latter.
4. Conclusions • In SAE-1040 steel, ultrasonic velocity is reduced in the order of annealing–oil–water quenching. • In SAE-4340, ultrasonic velocity is also reduced in the order of annealing–oil–water quenching as well as attenuation also decreases in the same order. • The main microstructural parameters affecting the nondestructive measurements are the volume fraction of phases and percentage of sub-grain boundaries. • The attenuation and ultrasonic velocity can be well correlated to evaluate the microstructural phases and the mechanical properties such as ultimate and yield strength as well as hardness.
Acknowledgment This work was supported by National Science, Technology and Innovation Plan (NSTIP) strategic technologies program, within the project number (08-ADV-209-02) in the Kingdom of Saudi Arabia.
198
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
References Ahn, B., Lee, S.S., 2000. Effect of microstructure of low carbon steels on ultrasonic attenuation. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 47 (May(3)). American Society of Metals, Metals Handbook, vol. 8, Metallography, Structures and Phase Diagrams. de Albuquerque, V.H.C., Silva, E.deM., Leite, J.P., de Moura, E.P., Freitas, V.L.deA., Tavares, J.M.R.S., 2010a. Spinodal decomposition mechanism study on the duplex stainless steel UNS S31803 using ultrasonic speed measurements. Mater. Des. 31, 2147–2150. de Albuquerque, V.H.C., Melo, T.A.deA., de Oliveira, D.F., Gomes, R.M., Tavares, J.M.R.S., 2010b. Evaluation of grain refiners influence on the mechanical properties in a CuAlBe shape memory alloy by ultrasonic and mechanical tensile testing. Mater. Des. 31, 3275–3281. de Albuquerque, V.H.C., Silva, C.C., Normando, P.G., Moura, E.P., Tavares, J.M.R.S., 2012. Thermal aging effects on the microstructure of Nb-bearing nickel based superalloy weld overlays using ultrasound techniques. Mater. Des. 36, 337–347. Bouda, A.B., Lebaili, S., Benchaala, A., 2003. Grain size influence on ultrasonic velocities and attenuation. NDT E Int. 36, 1–5. Chaki, S., Bourse, G., 2009. Guided ultrasonic waves for non-destructive monitoring of the stress levels in pre-stressed steel strands. Ultrasonics 49, 162–171. Freitas, V.L.deA., Normando, P.G., de Albuquerque, V.H.C., Silva, E.deM., Silva, A.A., Tavares, J.M.R.S., 2011. Nondestructive characterization and evaluation of embrittlement kinetics and elastic constants of duplex stainless steel SAF 2205 for different aging times at 425 ◦ C and 475 ◦ C. J. Nondestruct. Eval. 30, 130–136. Freitas, V.L.deA., de Albuquerque, V.H.C., Silva, E.deM., Silva, A.A., Tavares, J.M.R.S., 2010. Nondestructive characterization of microstructures and determination of elastic properties in plain carbon steel using ultrasonic measurement. Mater. Sci. Eng. A 527, 4431–4437. Gür, C.H., Tuncer, B.O., 2004. Investigating the microstructure-ultrasonic property relationships in steels. In: 16th WCNDT 2004 – World Conference on NDT, CDROM Proceedings, Internet Version of ∼600 Papers, August 30–September 3, Montreal, Canada. Gür, C.H., Keles, Y., 2003. Ultrasonic characterization of hot rolled heat treated steels. Mater. Charact. 45 (September(9)).
Gür, C.H., Tuncer, B.O., 2005. Characterization of microstructural phases of steels by sound velocity measurement. Mater. Charact. 55, 160–166. Gür, C.H., C¸am, I˙ ., 2006. Comparison of magnetic Barkhausen noise and sound velocity measurements for characterization of steel microstructures. In: ECNDT Proceedings Mo.2.2.4. Gür, C.H., C¸am, I˙ ., 2007. Comparison of magnetic Barkhausen noise and ultrasonic velocity measurements for microstructure evaluation of SAE 1040 and SAE 4140 steels. Mater. Charact. 58 (May(5)), 447–454. Khafri, M.A., Honarvar, F., Zanganeh, S., 2012. Characterization of grain size and yield strength in AISI 301 stainless steel using ultrasonic attenuation measurements. J. Nondestruct. Eval. 31 (September(3)), 191–196. Kumar, A., Laha, K., Jayakumar, T., Rao, K.B.S., Raj, B., 2002. Comprehensive microstructural characterization in modified 9Cr-1Mo ferritic steel by ultrasonic measurements. Metall. Mater. Trans. A 33A (June), 1617–1626. Kumar, A., Choudhary, B.K., Laha, K., Jayakumar, T., Rao, K.B.S., Raj, B., 2003. Characterization of microstructure in 9% chromium ferritic steels using ultrasonic measurements. Trans. Indian Inst. Met. 56 (October(5)), 483–497. Liu, X., Takamori, S., Osawa, Y., 2007. Effect of matrix structure on ultrasonic attenuation of ductile cast iron. J. Mater. Sci. 42, 179–184. Nunes, T.M., de Albuquerque, V.H.C., Papa, J.P., Silva, C.C., Normando, P.G., Moura, E.P., Tavares, J.M.R.S., 2013. Automatic microstructural characterization and classification using artificial intelligence techniques on ultrasound signals. Expert Syst. Appl. 40, 3096–3105. Papadakis, E., 1964. Ultrasonic attenuation and velocity in three transformation products in steel. J. Appl. Phys. 35 (May(5)). Prasad, R., Kumar, S., 1994. Study of the influence of deformation and thermal treatment on the ultrasonic behavior of steel. J. Mater. Process. Technol. 42, 51–59. Stella, J., Cerezo, J., Rodriguez, E., 2009. Characterization of the sensitization degree in the AISI 304 stainless steel using spectral analysis and conventional ultrasonic techniques. NDT E Int. 42, 267–274. Seok, C.S., Kim, J.P., 2005. Studies on the correlation between mechanical properties and ultrasonic parameters of aging I Cr-lMo-0.25V steel. J. Mech. Sci. Technol. 19 (2), 487–495. Vijayalakshmi, K., Muthupandi, V., Jayachitra, R., 2011. Influence of heat treatment on the microstructure, ultrasonic attenuation and hardness of SAF 2205 duplex stainless steel. Mater. Sci. Eng. A 529, 447–451.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Ultrasonic characterization of heat-treatment effects on SAE-1040 and -4340 steels Magdy M. El Rayes ∗,1 , Ehab A. El-Danaf 2 , Abdulhakim A. Almajid Mechanical Engineering Department, College of Engineering, King Saud University, P.O. Box 800, 11421, Riyadh, Saudi Arabia
a r t i c l e
i n f o
Article history: Received 13 May 2014 Received in revised form 4 September 2014 Accepted 5 September 2014 Available online 16 September 2014 Keywords: Ultrasonic characterization Sound velocity Attenuation Heat treatment of steel Microstructural phases Mechanical testing
a b s t r a c t In this work microstructural characterization and mechanical testing results were correlated with ultrasonic velocity and sound attenuation of steels SAE-1040 and SAE-4340. Both types were subjected to three types of heat treatment; the first was annealing at 850 ◦ C, the second was austenitizing at 1000 ◦ C followed by oil quenching and the third was similar austenitizing then water quenching. Treatments of SAE-1040 steel resulted in microstructures containing different ferrite and pearlite contents, different inter-lamellar spacing and also different grain size. Similar ferrite and pearlite content was obtained when annealing SAE-4340 whereas, oil and water quenching resulted into martensite. With SAE-1040, the sound velocity was reduced in the order of annealing–oil–water quenching due to the reduction of ferrite on the expense of pearlite. The same order in sound velocity reduction was also obtained with SAE-4340 due to the change in microstructural phases from pearlite to martensite. In comparison to pearlite, the martensite possessed higher crystal lattice distortion, higher dislocation density and lower elastic modulus all of which contribute in reducing sound velocity. Attenuation of SAE-1040 increased in the order of annealing–oil–water quenching because of higher pearlite content and the reduction in inter-lamellar spacing. Attenuation of SAE-4340 gave an opposite order due to the reduction of the extent of microstructural anisotropy. The mechanical properties and hardness were predominantly affected by the microstructural phases leading to the logical correlation with ultrasonic parameters. © 2014 Elsevier B.V. All rights reserved.
1. Introduction One of the prime objectives of non-destructive testing (NDT) is to certify that the component being examined is fit for the intended service. The most common way of doing so is by examining the component with NDT to detect flaws or discontinuities such as voids, inclusions, cracks, in materials or structures. Another parameter which is equally important to flaw detection is to assess the material properties. Among various NDT methods, ultrasonic testing (UT) was applied rather extensively in a variety of publications related to the correlation of ultrasonic measurements with microstructural phases of steel as conducted by Gür and Tuncer (2004), duplex stainless steel by de Albuquerque et al. (2010a,b), Ni-base super alloy by de Albuquerque et al. (2012), thermally aged Nibase alloy by Nunes et al. (2013) and grain size as by Bouda et al.
∗ Corresponding author. Tel.: +966 11 467 9906/50099 1132; fax: +966 4676652. E-mail addresses: [email protected], [email protected] (M.M. El Rayes). 1 On leave from Production Engineering Department-Faculty of EngineeringAlexandria University. 2 On leave from Mechanical Design and Production Department-Faculty of Engineering-Cairo University. http://dx.doi.org/10.1016/j.jmatprotec.2014.09.005 0924-0136/© 2014 Elsevier B.V. All rights reserved.
(2003). UT results were also correlated with mechanical properties as investigated by Vijayalakshmi et al. (2011) and de Albuquerque et al. (2010a,b), as well as residual stresses as by Chaki and Bourse (2009). Therefore, the utilization of ultrasonic techniques to indirectly determine the microstructural features and the mechanical properties can be useful for numerous industrial applications. Plain carbon steels SAE-1020 and -1050 were heat treated by varying the austenitization temperature between 860 ◦ C and 1060 ◦ C, as well as the cooling rate was varied between furnace cooling, air cooling and oil quenching. These materials were characterized by ultrasonic velocity and attenuation measurements as reported by Gür and Keles (2003). Metallographic studies revealed that the amount of proeutectoid ferrite, the softer phase, in SAE1020 was higher than that in SAE-1050. Ultrasonic results showed that sound velocities vary depending on the severity of cooling. The lowest sound wave velocity was found with the oil quenched SAE1050 consisting of martensite, which possessed high dislocation density, distortion of crystalline lattice and maximum hardness among other specimens. On the other hand, the highest sound velocity was obtained with furnace cooled SAE-1020 which had the softest structure. It was also found that prior austenite grain size; and not the transformation products within it, as manifested
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198 Table 1 Chemical composition of steels SAE-1040 and -4340 in wt.%. Steel grade SAE
C
Mn
P
S
Si
Cr
Mo
Ni
1040 4340
0.41 0.36
0.6 0.25
0.04 0.04
0.05 0.04
0.3 0.25
– 1.4
– 0.2
– 1.4
by the author, had the predominant effect on the attenuation values. The larger the prior austenite grain size the higher the sound attenuation. Slower cooling rates led to larger prior austenite grain size, larger amounts of soft ferrite and wider interlamellar spacing between cementite [Fe3 C] plates of pearlite. Prasad and Kumar (1994) studied the influence of varying the grain size of cast steel on ultrasonic velocity and attenuation. Various grain sizes were achieved via hot upsetting at different percentages of height reduction. These samples were further heat treated through annealing, normalizing and hardening using oil quench followed by tempering. They reported that increasing the degree of deformation decreases the ultrasonic velocity. However, numerous research work concluded the fact that the longitudinal ultrasonic velocity varied from grain to grain because of misorientation of grains, which was related to the variation in the elastic constant as well. The objectives of the present work were to correlate ultrasonic measurements namely; ultrasonic longitudinal sound wave velocity and ultrasonic attenuation with the microstructural and mechanical properties of two types of steel which were subjected to different heat treatments. These treatments were selected to obtain different microstructural phases as well as different grain size. 2. Materials and methods Two types of steels, namely SAE-1040 (C-Mn steel) and SAE4340 (Cr-Ni-Mo steel), supplied in the form of 50 mm diameters rods, from which both materials were cut into three equal pieces 100 mm long using hack saw with lubricant to avoid excessive heating. The selection of these steel types was based on their importance in various fields of industrial applications in which they are mainly characterized by durability. Steel 1040 is normally used in axels, crank shafts and gears, whereas, steel 4340 is used in air craft landing gears, power transmission gears and shafts. In addition, the types of heat treatment chosen were expected to result into different phases that were ought to significantly affect the microstructural, mechanical and ultrasonic characteristics. Table 1 presents the average chemical composition of these steels in wt.% after three runs for each type using spectroscopic chemical analyzer. Both materials were subjected simultaneously to the same heat treatment type in an electric resistance furnace. The first treatment was austenitizing at 850 ◦ C for 2 h followed by furnace cooling (full annealing). The second and third were austenitizing at 1000 ◦ C for 3 h followed by oil and water quenching, respectively. The heat treated samples were sectioned into disks using hack saw with coolant to extract ultrasonic measurements samples [50 mm long], microstructural, hardness and tensile test samples. The microstructure was examined by secondary electron imaging using scanning electron microscopy (SEM), and electron back scattered diffraction (EBSD). For SEM the samples were prepared according to standard metallographic sample preparation which includes grinding using SiC sand paper, polishing using diamond paste of 1.0 and 0.05 m, and etched with 5% Nital to reveal the samples’ microstructure. The microstructure for both materials was also studied by EBSD using Oxford HKL system incorporated on a field emission scanning electron microscope (FESEM) 7600 JEOL. These samples were polished with colloidal silica as a final step prior to imaging.
189
The ultrasonic measurement samples were machined using end milling then surface ground on both sides in order to ensure complete parallelism between faces. In order to eliminate roughness, visible irregularities and any oxides which might affect the ultrasonic measurements, the samples were further processed similar to the metallographic preparation steps. For ultrasonic measurements, the pulse-echo technique and direct contact method were applied to obtain ultrasonic velocities and attenuation coefficients using ultrasonic pulse-receiver equipment-Karl Deutsch (Model: Echograph-1085). Ultrasonic measurements for all samples were obtained by using commercial NDT ultrasoniclongitudinal wave transducers of 4 MHz. The coupling material Karl Deutsch-Ecotrace gel was used for the longitudinal wave measurements. During measurements a constant load was applied to the probe against the specimen surface so as to have a constant thickness of couplant layer at the interface between the specimen surface and the probe. Ultrasonic velocity was determined by dividing twice the specimen thickness by time of flight (TOF) obtained between zero crossing of the first and second back-wall echoes using Eq. (1) as applied by Vijayalakshmi et al. (2011). In order to check results repeatability, seven ultrasonic readings of each specimen were averaged to represent the data obtained, which gave an error of around ±0.5% with both types of steel. Velocity (m/s) = 2 ×
thickness (m) time (s)
(1)
During the same measurement, ultrasonic velocity was measured again using a function in the ultrasonic equipment which automatically calculates the ultrasonic velocity when locating two corresponding points on two successive echoes (first and second echoes) at the extreme sides of the screen with an error of less than the above mentioned one. The ultrasonic attenuation values were calculated according to Eq. (2) which is based on the reduction of the amplitude of an ultrasound pulse, measured in decibels per millimeter (dB/mm). This equation appeared in several publications such that of Stella et al. (2009), Vijayalakshmi et al. (2011) and Freitas et al. (2011) and given in as: ˛=
A0 20 log 2x A1
(2)
where ˛ is the attenuation coefficient [dB/mm], x is the thickness of the sample measured in the test [mm], A0 is the amplitude of the first echo in dB and A1 is the amplitude of the second echo. The constant 2 is because the pulse-echo technique is used. For each sample, the measurements were repeated seven times with maximum error of ±0.1% with both steels. Vickers macro-hardness tests were conducted using 10 kg force and were performed five times for each heat treated specimen and the average was taken. In order to evaluate the mechanical behavior of heat treated steels, tensile specimens were extracted from the center of the disk along its axis and were cut according to ASTM E-08. Tensile tests were conducted at room temperature and a cross-head speed velocity of 2 mm/min using Instron machine model 3385 H. The machine was equipped with a computer having software through which the load-elongation data were recorded. The tensile test for each type of steel and treatment was repeated three times and the average value was taken and presented hereafter. 3. Results and discussion 3.1. Microstructure Fig. 1a–c shows the SEM microstructures obtained with the three different treatments applied on SAE-1040 steel. Fig. 1a shows
190
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 1. Microstructures obtained by SEM of steel SAE-1040 in (a) annealed, (b) oil quenched and (c) water quenched conditions.
the annealed microstructure which contained pearlite colonies (light) in a matrix of ferrite (dark). The pearlite contained alternate lamellas of eutectoid ferrite (Fe) and cementite (Fe3 C) with random orientation. The pearlite constituted 46% volume-fraction, whereas ferrite was 54%. The average grain size of this structure was around 18 m. The oil quenched microstructure, shown in Fig. 1b, was also composed of pearlite and ferrite which transformed from austenite upon quenching. This microstructure conforms to the continuous cooling transformation (CCT) diagram of SAE-1040 steel. Measurements of phases yielded about 80% pearlite and 20% ferrite and mean grain size of 31.1 m. It can be also noted from Fig. 1b that the pearlite increased on the expense of ferrite and also the interlamellar spacing were much reduced due to the fast cooling rate compared to the annealed treatment. The water-quenched microstructure was similar to that obtained with oil-quench with respect to its pearlite and ferrite constituents. However, the volume fraction of pearlite increased to 92% still on the ferrite expense 8% as well as more dense interlamellar spacing between Fe3 C and Fe took place. The mean grain size corresponding to this treatment was around 54.5 m. Hence, it can be stated that the main difference in the microstructure resulting from the three treatments was the content and size of pearlite and ferrite phases as well as the interlamellar spacing. Fig. 2a–c shows, at higher magnifications, the microstructure corresponding to annealed, oil and water quenched respectively, with emphasis on pearlite structure. In the annealed condition the interlamellar spacing was relatively wider than that occurring with oil and water quenched structure where these spacing became much less, i.e. denser lamellas. Measurements of spacing in the annealed sample gave an average value of 0.362 m, whereas with the water quenched sample was much narrower around 0.0178 m. Fig. 3 shows an example of these measurements with SAE-1040 in the annealed condition using a software available in SEM. Fig. 4a–c shows the orientation imaging grain boundary maps representing the extent of sub-grain boundaries in red color, defined with misorientation less than 2◦ for SAE-1040 steel in the annealed, oil and water quenched conditions, respectively. Also shown in the same figure, the true grain boundaries, with misorientation larger than 15◦ across them, displayed in thick black color. Beside each grain boundary map the respective histogram for the misorientation angle distribution is presented. It is evident that the annealed structure had a relatively large average misorientation angle of about 35◦ with a relatively low percentage of low angle grain boundaries (LAGBs) of 12%, which is a typical finding for well recrystallized annealed microstructures. The oil quenched sample revealed a relatively good amount of LAGBs with a percentage of about 67%, and an average misorientation angle of 12◦ . The water quenched sample revealed, even more LAGBs with a percentage of about 73% and an average misorientation angle of 16◦ . Fig. 5a–c shows the annealed, oil and water quenched microstructures respectively of SAE-4340 steel taken by SEM, respectively. The annealed structure was composed of grains of pearlite (light) in a matrix of ferrite (dark). Measurements showed that this microstructure is composed of 59% pearlite and 41% ferrite. The grain size measured is around 11.6 m. The oil quenched structure, Fig. 5b, is composed of long martensite laths (light) coexisting with a substructure of few plate martensite (regions marked P-dark), which normally consists of fine internal twins as referred in ASM Metals Handbook. The water quenched microstructure is composed of 100% martensite with relatively short laths as in Fig. 5c. It should be noted that due to the presence of martensite laths, it was not possible to see the prior-austenite grain boundaries clearly in the quenched samples. Due to difficulties in indexing the Kikuchi patterns of martensite obtained in the oil and water quenched 4340, the EBSD study was
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
191
Fig. 3. Measurements of pearlite interlamellar spacing in SAE-1040 in the annealed condition.
4340 displayed a relatively higher amount of LAGBs almost double that displayed in the annealed 1040. 3.2. Hardness testing Vickers hardness measurements were conducted to indicate the hardness of phases obtained from different heat treatments. The results of the metallographic examinations are verified by hardness measurements. Fig. 7a and b shows that hardness increased when the heat treatment of both steel types changed from annealing to oil and to water quenching. With SAE-1040 steel, Fig. 7a, hardness increased due to the increase in pearlite content (hard phase) with respect to the ferrite (soft phase), as shown earlier in microstructural results. Similarly, with steel SAE-4340, the hardness increased due to the change in microstructural phases being pearlite + ferrite obtained from annealing, martensite transformed at slow and fast cooling rates obtained from oil and water quench, respectively, as depicted in Fig. 7b. The softness of oil-quenched martensite compared to the water-quenched one is due to that oil quenching leads to less extent of lattice distortion, residual stresses and dislocation density as reported by Gür and Tuncer (2004). 3.3. Tensile testing Table 2 summarizes the results obtained from tensile tests conducted on both types of steels subjected to the different heat treatments. The results of tensile tests are in line with hardness results, which in turn can be related to the phases present in the microstructure. With both types of steel, the ultimate tensile strength (UTS) and yield strength (YS) increased when the treatment was changed from annealing to oil and water quenching, whereas the elongation was reduced in the same order. The phases of steel obtained, as expected, were responsible for the increase in UTS and YS and the reduction in elongation which is similar to that found by Gür and Tuncer (2005). In SAE-1040, increasing Fig. 2. Variation of interlamellar spacing with heat treatment of SAE-1040 (a) annealed, (b) oil quenched and (c) water quenched.
confined to the annealed 4340. The structure is presented in Fig. 6. The average misorientation angle exhibited a value of about 33◦ with a relative low percentage of LAGBs of 25%. Worth noting, that both SAE-1040 and -4340 steels in annealed condition displayed almost similar average misorientation angle; however the annealed
Table 2 Mechanical properties of SAE-1040 and -4340 steels. Sample
UTS, MPa
YS, MPa
Elongation %
1040-annealed 1040-oil quench 1040-water quench 4340-annealed 4340-oil quench 4340-water quench
625 798 854 693 1721 1947
326 464 690 406 1425 1633
37.4 26.6 22.4 36.6 23.1 21.6
192
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 4. Orientation imaging grain boundary mapping (left) and misorientation angle distribution histogram (right) of SAE-1040 in the (a) annealed, (b) oil quenched and (c) water quenched structures.
the pearlite content on the expense of ferrite increased the UTS and YS and also reduced elongation as seen in Table 2. A similar trend was noted with SAE-4340 where the UTS and YS increased with simultaneous reduction in elongation accompanying the treatment. The results in Table 2 indicated that for both types of steel the mechanical properties are predominantly affected by the microstructural phases resulting from different heat treatments, which in turn affects the ultrasonic velocity and attenuation as will be shown in the following sections. 3.4. Ultrasonic testing 3.4.1. Sound velocity The ultrasonic velocity measurements of longitudinal waves were sensitive to microstructural variations resulting from different heat treatments applied on SAE-1040 steel. This can be noted when plotting sound velocity versus different fraction percent of pearlite and ferrite phases as in Fig. 8. The reason to this was related to the percentage of pearlite and ferrite phases within the whole structure as well as the interlamellar spacing between ferrite and Fe3 C phases within the pearlite grain. Gür and Tuncer (2004), Freitas et al. (2010) and others have agreed upon that sound velocity becomes lower when the ferrite content is reduced as well as when the interlamellar spacing is narrower compared to coarse pearlite. This is due to the fact that the ferrite phase has the least resistance to ultrasonic waves hence allowing the highest velocity. Simultaneously, pearlite grains with dense/narrow interlamellar spacing possess high resistance to ultrasonic waves and consequently, have low sound velocity as discussed by Freitas et al. (2010). The variation of interlamellar spacing with different treatments was noted in Fig. 2. Therefore, it can be stated that an inverse relation exists between pearlite content and sound velocity. The evolution of sub-grains and sub-grain boundaries contribute in the reduction of sound velocity. Gür and Keles (2003)
reported that the structures consisting of sub-grains and their boundaries have an effect on lattice straining and interrupts the matrix continuity thus lowers the elastic modulus and consequently, lowers the propagation rate of sound velocity. This can be inferred from Fig. 4, which shows a tremendous amount of sub-grain boundaries accompanying the water quench treatment (LAGBs = 73%) which is much higher than that occurring with the annealing treatment (LAGBs = 12%). Furthermore, the variation of the elastic constant from grain to grain in the direction of sound wave propagation may be another reason for the reduction in sound velocity as found by Prasad and Kumar (1994). Papadakis (1964) reported that there is a linear proportion between sound velocity and the elastic modulus of the structure within which the sound wave propagates. The elastic moduli decreases in the order of ferrite–pearlite–martensite. Hence, lower percentage of ferrite with respect to pearlite content corresponds to lower elastic modulus and consequently, lower sound velocity. A similar trend is also found with sound velocity for SAE-4340 when changing the heat treatment from annealing to oil and water quenching causing a reduction in sound velocity as seen in Fig. 9. The reason to this reduction is due to the change in microstructure from pearlite to martensite similar to that concluded by Prasad and Kumar (1994). When quenching, the steel is cooled rapidly from the austenitizing range to room temperature, in which each austenite grain (FCC) suddenly transforms into laths of martensite (BCT) by diffusionless lattice shear as confirmed by Gür and C¸am (2006). This transformation leads to high crystal lattice distortions due to the increase in volume during the austenite–martensite transformation resulting into great amount of internal tension/residual stresses. Therefore it can be stated that martensite is the phase with very high dislocation density and maximum randomness as reported by Papadakis (1964) and Gür and C¸am (2007). In the same context, Gür and Tuncer (2005) reported that the ultrasonic velocity in martensite is essentially affected by changes in the modulus of elasticity of individual grains, in the crystal
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
193
Fig. 6. Orientation imaging grain boundary mapping (left) and misorientation angle distribution histogram (right) of SAE-4340 in the annealed condition.
lattice distortion level and in the orientation of primary austenite grains. The increase in lattice distortion and the subsequent increase in dislocation density reduce sound velocity as reported by Gür and C¸am (2007). This was confirmed earlier by Papadakis (1964) who reported that the elastic moduli decrease in the order of pearlite–bainite–martensite and based on the proportional relation between the elastic modulus and sound velocity it can be stated that martensite has lower sound velocity when compared to pearlite. In addition, martensite presents high resistance to ultrasound waves, because of its compact and fine structure, hence possessing low sound wave propagation velocity as reported by Freitas et al. (2010). These results are in line with those obtained by Papadakis (1964), Gür and C¸am (2006) and Gür and Tuncer (2005).
Fig. 5. Microstructures obtained by SEM of steel SAE-4340 in (a) annealed, (b) oil quenched and (c) water quenched conditions.
3.4.2. Attenuation The attenuation of ultrasonic waves in a polycrystalline material is mainly due to either scattering within the grains and their structural boundaries or absorption due to dislocation damping, thermoelastic losses, and magnetic hysteresis loss or both as reported in various publications such as Papadakis (1964) and Kumar et al. (2002). When calculating attenuation coefficient (˛), most literature such that reported by Liu et al. (2007) have
194
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 7. Hardness variation with different heat treatments with: (a) SAE-1040 and (b) SAE-4340.
applied Eq. (3), which indicates that (˛) is equal to the sum of the absorption coefficient (˛a ) and the scattering coefficient (˛s ) as follows: ˛ = ˛a + ˛s
(3)
Scattering is usually divided into three different regimes depending on the ratio of grain size (d) to ultrasonic wavelength () as summarized by Khafri et al. (2012). These regimes are the Rayleigh region where d, the stochastic region where ≈ d and the geometric region where < d. In the present work the Rayleigh regime is applicable because ranges between 1.353 and 1.366 mm whereas the maximum grain size is about 54 m (depending on the sound velocities measured for each specific sample). In polycrystalline materials, it is the scattering of ultrasound from grains and interfaces that is the main cause of attenuation. In these materials,
grain scattering losses are large compared to absorption losses as agreed by Papadakis (1964) and Khafri et al. (2012). Fig. 10 shows that the attenuation coefficient for SAE-1040 was lowest with the annealed microstructure and gradually increases with oil then water quenched ones. This is due to the increase in pearlite content (i.e. decrease in proeutectoid ferrite), as inferred from Fig. 1, and the decrease in interlamellar spacing in pearlite as shown in Fig. 2, all of which contribute in increasing scattering, hence increasing attenuation. The reason to this is that scattering increases with samples having larger pearlite content and consequently, larger areas of interface [phase boundaries] between hard Fe3 C and soft ferrite, as reported by Gür and Tuncer (2004) and Gür and Keles (2003). In the same sense, ultrasonic wave attenuation due to scattering in pearlite microstructure is greater than that in ferrite because pearlite is more elastically anisotropic than ferrite. This result is confirmed by Ahn and Lee (2000). Since the pearlite content increased when the treatment changed from annealing to oil then water quenching, therefore the interface area increased as well. It was also reported by Gür and Keles (2003) that in pearlite containing steels, the attenuation of ultrasonic beams passing through grains are affected by pearlite dispersion, which conforms to that shown in Fig. 2. Fig. 11 for SAE-4340 shows that attenuation decreased as the steel phases changed from ferrite + pearlite, oil quenchedmartensite and water quenched-martensite. This result is in agreement with that obtained in various earlier works such as that of Kumar et al. (2002, 2003), which was applied on carbon and alloy steels. They reported that the attenuation of pearlite is higher than that of martensite. Gür and C¸am (2007) indicated that upon quenching, the sudden transformation of austenite into martensite resulted into high amount of lattice distortion, high dislocation density and maximum randomness of the structure. From the point of view of scattering theory, the randomness causes an increase in elastic isotropy of the grain volume, thus decreases scattering power as reported by Papadakis (1964) and Kumar et al. (2002). In addition, fine laths of martensite made the structure more isotropic than pearlite and ferrite hence decreasing its scattering power (Gür and Keles, 2003). 3.5. Correlation between ultrasonic and mechanical properties Fig. 12a and b shows the relation between hardness and sound velocity and attenuation coefficient of SAE-1040 steel at different pearlite–ferrite contents, respectively. Generally, it can be stated
Fig. 8. The influence of different pearlite (P) and ferrite (F) contents on sound velocity in SAE-1040.
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 9. The influence of different phases in SAE-4340 on sound velocity.
Fig. 10. The influence of different pearlite (P) and ferrite (F) contents on attenuation coefficient in SAE-1040.
Fig. 11. The influence of different phases in SAE-4340 on attenuation coefficient.
195
196
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
Fig. 12. Relationship between hardness Vickers and (a) sound velocity and (b) attenuation coefficient at different pearlite–ferrite contents in SAE-1040. Fig. 13. Relationship between hardness Vickers and a) sound velocity and b) attenuation coefficient at different pearlite-ferrite contents in SAE-4340.
that an inverse relationship exists between ultrasonic velocity and hardness, which is in agreement with that found by Gür and C¸am (2007). Since the hardness of the phases increased in the order of increasing pearlite content, a corresponding reduction in the sound velocity was noted. The attenuation coefficient can also be correlated with hardness which in turn is related to the volume percent of pearlite and ferrite and the corresponding extent of elastic anisotropy. In addition, Prasad and Kumar (1994) indicated that hardening process introduces internal stresses into the lattice resulting into lattice distortion/deformation; which is accompanied by an increase in dislocation density thus increasing the attenuation coefficient. The same inverse relation between hardness and ultrasonic velocity was obtained with SAE-4340 steel as shown in Fig. 13a. This is attributed to the difference in microstructures where martensite possesses higher resistance to sound waves as well as higher lattice distortion and dislocation density all of which reduce the sound velocity more than pearlite and ferrite structure does. On the other hand, Fig. 13b shows that increasing hardness markedly reduced the attenuation coefficient. This result is confirmed with that obtained by Seok and Kim (2005). Again, the difference in microstructure was responsible to this reduction. This is due to the fact that laths of martensite
make the structure more isotropic than pearlite + ferrite structure does, hence decreasing its scattering power. This result is in line with that reported by Gür and Keles (2003) and Papadakis (1964) which stated that attenuation in pearlite is higher than that in martensite, which possesses least attenuation even among other steel phases such as ferrite and bainite as reported by Kumar et al. (2002). Fig. 14a shows that the increase of UTS and YS of SAE1040 steel, was accompanied by a reduction in sound velocity. A similar inverse relation, as that of hardness, is obtained due to the proportional relation between pearlite content and hardness on one hand and the mechanical properties on the other hand. Similarly, Fig. 14b plots the strength versus the attenuation coefficient. The strength is related to the increase in pearlite content which in turn causes the structure to be more anisotropic thus increases the attenuation [scattering power] of this structure compared to the less anisotropic one [lower pearlite content]. With SAE-4340 steel, increasing strength was accompanied by a reduction in ultrasonic velocity as in Fig. 15a, which is a trend similar that obtained with SAE-1040 steel. The reason to this reduction
Table 3 Comparison between SAE-1040 and -4340 in annealed condition. Sample
Sound velocity, m/s
Attenuation coefficient, dB/mm
Pearlie %, ferrite %
LAGBs %
1040-annealed 4340-annealed
5913 5906
0.0072 0.0074
46% P, 54% F 59% P, 41% F
12 25
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
197
Fig. 14. Relation between UTS and YS with ultrasonic parameters of SAE-1040 steel: (a) sound velocity and (b) attenuation coefficient.
is related to the type of microstructure, where pearlite offers less resistance to ultrasonic waves thus allowing faster sound velocity to take place compared to martensite which possess higher resistance and consequently lowest speed as discussed by Freitas et al. (2010). The relation between strength and attenuation is plotted in Fig. 15b. Increasing strength was accompanied by a reduction in attenuation. The strength increased because of the change in microstructure from pearlite + ferrite to martensite which directly influenced the attenuation based on the fact that martensite structure is more isotropic thus possessing lower scattering power than pearlite does. This result agrees well with that reported by Papadakis (1964), and Kumar et al. (2002). Although the present work was not intended to compare the two types of steel, however, there were two reasons behind their comparison. First, was that both types of steel were annealed at the same temperature and second was that similar microstructural components were obtained. In light to the earlier discussion, comparison was done as a trial to further verify the influence of phase percent and the sub-grain boundaries on sound velocity and attenuation when steel type is changed. Table 3 summarizes the microstructural phases and grain size and their corresponding ultrasonic measurements conducted with both types of steels annealed at the same conditions. It can be noted that sound waves propagated faster in SAE1040 steel than in SAE-4340. This is due to that the former steel possessed higher percent of ferrite (54%; least resistant to ultrasonic waves) than the latter (41%). Attenuation of SAE-1040, on the other hand, was lower than that of SAE-4340 as in Table 3. This was again due to the lower percentage of pearlite in the former (46%) than in the latter (59%). Higher pearlite content made the structure more anisotropic thus leading to higher attenuation capability than lower pearlite content. Also, the higher percentage
Fig. 15. Relation between UTS and YS with ultrasonic parameters of SAE-4340 steel: (a) sound velocity and (b) attenuation coefficient.
of sub-grain boundaries occurring with SAE-4340 (LAGBs = 25%) compared to that with SAE-1040 (LAGBs = 12%) may be another reason why the former possesses higher attenuation than the latter.
4. Conclusions • In SAE-1040 steel, ultrasonic velocity is reduced in the order of annealing–oil–water quenching. • In SAE-4340, ultrasonic velocity is also reduced in the order of annealing–oil–water quenching as well as attenuation also decreases in the same order. • The main microstructural parameters affecting the nondestructive measurements are the volume fraction of phases and percentage of sub-grain boundaries. • The attenuation and ultrasonic velocity can be well correlated to evaluate the microstructural phases and the mechanical properties such as ultimate and yield strength as well as hardness.
Acknowledgment This work was supported by National Science, Technology and Innovation Plan (NSTIP) strategic technologies program, within the project number (08-ADV-209-02) in the Kingdom of Saudi Arabia.
198
M.M. El Rayes et al. / Journal of Materials Processing Technology 216 (2015) 188–198
References Ahn, B., Lee, S.S., 2000. Effect of microstructure of low carbon steels on ultrasonic attenuation. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 47 (May(3)). American Society of Metals, Metals Handbook, vol. 8, Metallography, Structures and Phase Diagrams. de Albuquerque, V.H.C., Silva, E.deM., Leite, J.P., de Moura, E.P., Freitas, V.L.deA., Tavares, J.M.R.S., 2010a. Spinodal decomposition mechanism study on the duplex stainless steel UNS S31803 using ultrasonic speed measurements. Mater. Des. 31, 2147–2150. de Albuquerque, V.H.C., Melo, T.A.deA., de Oliveira, D.F., Gomes, R.M., Tavares, J.M.R.S., 2010b. Evaluation of grain refiners influence on the mechanical properties in a CuAlBe shape memory alloy by ultrasonic and mechanical tensile testing. Mater. Des. 31, 3275–3281. de Albuquerque, V.H.C., Silva, C.C., Normando, P.G., Moura, E.P., Tavares, J.M.R.S., 2012. Thermal aging effects on the microstructure of Nb-bearing nickel based superalloy weld overlays using ultrasound techniques. Mater. Des. 36, 337–347. Bouda, A.B., Lebaili, S., Benchaala, A., 2003. Grain size influence on ultrasonic velocities and attenuation. NDT E Int. 36, 1–5. Chaki, S., Bourse, G., 2009. Guided ultrasonic waves for non-destructive monitoring of the stress levels in pre-stressed steel strands. Ultrasonics 49, 162–171. Freitas, V.L.deA., Normando, P.G., de Albuquerque, V.H.C., Silva, E.deM., Silva, A.A., Tavares, J.M.R.S., 2011. Nondestructive characterization and evaluation of embrittlement kinetics and elastic constants of duplex stainless steel SAF 2205 for different aging times at 425 ◦ C and 475 ◦ C. J. Nondestruct. Eval. 30, 130–136. Freitas, V.L.deA., de Albuquerque, V.H.C., Silva, E.deM., Silva, A.A., Tavares, J.M.R.S., 2010. Nondestructive characterization of microstructures and determination of elastic properties in plain carbon steel using ultrasonic measurement. Mater. Sci. Eng. A 527, 4431–4437. Gür, C.H., Tuncer, B.O., 2004. Investigating the microstructure-ultrasonic property relationships in steels. In: 16th WCNDT 2004 – World Conference on NDT, CDROM Proceedings, Internet Version of ∼600 Papers, August 30–September 3, Montreal, Canada. Gür, C.H., Keles, Y., 2003. Ultrasonic characterization of hot rolled heat treated steels. Mater. Charact. 45 (September(9)).
Gür, C.H., Tuncer, B.O., 2005. Characterization of microstructural phases of steels by sound velocity measurement. Mater. Charact. 55, 160–166. Gür, C.H., C¸am, I˙ ., 2006. Comparison of magnetic Barkhausen noise and sound velocity measurements for characterization of steel microstructures. In: ECNDT Proceedings Mo.2.2.4. Gür, C.H., C¸am, I˙ ., 2007. Comparison of magnetic Barkhausen noise and ultrasonic velocity measurements for microstructure evaluation of SAE 1040 and SAE 4140 steels. Mater. Charact. 58 (May(5)), 447–454. Khafri, M.A., Honarvar, F., Zanganeh, S., 2012. Characterization of grain size and yield strength in AISI 301 stainless steel using ultrasonic attenuation measurements. J. Nondestruct. Eval. 31 (September(3)), 191–196. Kumar, A., Laha, K., Jayakumar, T., Rao, K.B.S., Raj, B., 2002. Comprehensive microstructural characterization in modified 9Cr-1Mo ferritic steel by ultrasonic measurements. Metall. Mater. Trans. A 33A (June), 1617–1626. Kumar, A., Choudhary, B.K., Laha, K., Jayakumar, T., Rao, K.B.S., Raj, B., 2003. Characterization of microstructure in 9% chromium ferritic steels using ultrasonic measurements. Trans. Indian Inst. Met. 56 (October(5)), 483–497. Liu, X., Takamori, S., Osawa, Y., 2007. Effect of matrix structure on ultrasonic attenuation of ductile cast iron. J. Mater. Sci. 42, 179–184. Nunes, T.M., de Albuquerque, V.H.C., Papa, J.P., Silva, C.C., Normando, P.G., Moura, E.P., Tavares, J.M.R.S., 2013. Automatic microstructural characterization and classification using artificial intelligence techniques on ultrasound signals. Expert Syst. Appl. 40, 3096–3105. Papadakis, E., 1964. Ultrasonic attenuation and velocity in three transformation products in steel. J. Appl. Phys. 35 (May(5)). Prasad, R., Kumar, S., 1994. Study of the influence of deformation and thermal treatment on the ultrasonic behavior of steel. J. Mater. Process. Technol. 42, 51–59. Stella, J., Cerezo, J., Rodriguez, E., 2009. Characterization of the sensitization degree in the AISI 304 stainless steel using spectral analysis and conventional ultrasonic techniques. NDT E Int. 42, 267–274. Seok, C.S., Kim, J.P., 2005. Studies on the correlation between mechanical properties and ultrasonic parameters of aging I Cr-lMo-0.25V steel. J. Mech. Sci. Technol. 19 (2), 487–495. Vijayalakshmi, K., Muthupandi, V., Jayachitra, R., 2011. Influence of heat treatment on the microstructure, ultrasonic attenuation and hardness of SAF 2205 duplex stainless steel. Mater. Sci. Eng. A 529, 447–451.