- Email: [email protected]
A study on aging carbide precipitation behavior of hadfield steel by dynamic elastic modulus
Author’s Accepted Manuscript A Study on Aging Carbide Precipitation Behavior of Hadfield Steel by Dynamic Elastic Modulus C. Chen, X.Y. Feng, B. Lv, Z.N. Yang, F.C. Zhang www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(16)30996-0 http://dx.doi.org/10.1016/j.msea.2016.08.084 MSA34042
To appear in: Materials Science & Engineering A Received date: 25 May 2016 Revised date: 18 August 2016 Accepted date: 19 August 2016 Cite this article as: C. Chen, X.Y. Feng, B. Lv, Z.N. Yang and F.C. Zhang, A Study on Aging Carbide Precipitation Behavior of Hadfield Steel by Dynamic Elastic Modulus, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2016.08.084 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A Study on Aging Carbide Precipitation Behavior of Hadfield Steel by Dynamic Elastic Modulus C. Chen a, X.Y. Feng a, B. Lv b,*, Z.N. Yang c, F.C. Zhang a, c a
State Key Laboratory of Metastable Materials Science and Technology, Yanshan University, Qinhuangdao 066004, China
b c
College of Environmental and Chemical Engineering , Yanshan University, Qinhuangdao 066004, China
National Engineering Research Center for Equipment and Technology of Cold Strip Rolling, Yanshan University, Qinhuangdao 066004, China
Abstract: In the present study, the carbon atom distribution and carbide precipitation behavior of Hadfield steel under three distinct states, namely, the quenched Hadfield steel, the deformed Hadfield steel (coarse grain), and the nanocrystallized Hadfield steel, were studied using dynamic thermomechanical analysis, transmission electron microscopy, three-dimensional atom probe, X-ray diffraction, and optical microscopy. Results showed apparent differences in elastic modulus change in the Hadfield steel under three distinct states in the dynamic thermomechanical analysis. Those results proved that the dynamic elastic modulus could be used to determine carbide precipitation behavior in steels during heat treatment. The carbide precipitation temperatures of the nanocrystallized Hadfield steel and the deformed Hadfield steel were higher than that of the quenched Hadfield steel. This phenomenon was attributed to the promoted diffusion of carbon atoms to the stress field of defects during plastic deformation and nanocrystallization process. This process induced a more homogeneous distribution of carbon atoms in the Hadfield steel. As a result, the precipitated carbides were much finer and more dispersive, especially in the nanocrystallized Hadfield steel. Key words: Hadfield steel; Carbides; Elastic modulus; Nanocrystalline microstructures 1 Introduction Severe plastic deformation (SPD), such as equal channel angular pressing, accumulative roll bonding, surface mechanical polishing, and shot peening, is considered the main method to prepare dense, nonpolluting, and ultrafine-grained microstructure materials [1–4]. However, these materials are in a thermodynamically metastable state because of wide grain boundary and lattice distortion induced by SPD [5–8]. Except for the evolution in grain size, twin density, and dislocation density, SPD could also result in the redistribution of alloying elements, and influence phase transformation [9–12]. Therefore, the microstructure evolution, including the kinetics and thermodynamics of phase transition, the microstructures, and the phase structures of the ultrafine-grained and nanocrystallized metal material prepared by SPD would change during heat treatment. Hadfield steel is a type of stable single-phase austenitic steel. Compared with other steels,
*
Corresponding author. Tel.: +86 335 8063949; e-mail addresses: [email protected] (B. Lv) 1
carbides easily precipitate in the Hadfield steel especially at the grain boundaries when the temperature is above 300 °C [13]. However, precipitated carbides can severely worsen mechanical properties. For this reason, some researchers investigated the delays and suppression in carbide precipitation during heat treatment [14]. For 130 years, material researchers have focused their attention on the work hardening property and microstructure reactions in Hadfield steel, some of which have not been determined yet. For instance, the excellent work hardening property is widely attributed to the block of dislocation movement by the dynamic strain aging effect resulting from the interaction between C–Mn couples and dislocations [15, 16]. Determining where the C atoms of the C–Mn couples in Hadfield steel are present, whether in the crystal lattice or in the strain field induced by vacancies and dislocations, is crucial but is rarely given attention. This process affects the deformation and the work hardening behaviors of Hadfield steel, and it influences atom diffusion and carbide precipitation behaviors in the heating process. Elastic modulus is an important property parameter that is a reflection of bond strength among atoms, ions, and molecules. Any factor that influences bond strength, such as crystal structure, residual stress, and temperature, can influence the elastic modulus [17, 18]. For the past few years, a number of reports have shown that the change in elastic modulus could be used to analyze residual stress [19], precipitation [20], and texture [21] during heat treatment or plastic deformation. The elastic modulus of a metal material is usually tested through tensile, nano-indentation, bending, ultrasonic, and vibration processes. The first four methods are known as static methods, and the last one is a dynamic method. Dynamic thermomechanical analysis is used to measure the elastic modulus of cement, concrete, and composite materials during heating [22-25]. For dynamic thermomechanical analysis method, deformation occurs in samples when mechanical stress with periodic variation acted on materials under different external conditions. When the stress acts on an elastomer, the resulting strain is proportional to the stress. The proportionality constant is the elastic modulus according to Hooke's law [26]. However, reports about using dynamic thermomechanical analysis to measure the elastic modulus of metal materials are rare [27], especially studies on the precipitation behaviors in steels. In the present work, the deformed and nanocrystallized Hadfield steel samples prepared by a new SPD process, as well as the quenched Hadfield steel sample, were studied. Dynamic thermomechanical analysis was used to measure the change in the elastic modulus of the Hadfield steel under three distinct states in the aging process. Methods of optical microscopy, X-Ray diffraction (XRD), transmission electron microscopy (TEM), and atom probe tomography (APT) were used to analyze the influence of deformation and nanocrystalline on carbon distribution, microstructure, and carbide precipitation behavior. Moreover, the carbide precipitation behavior was successfully predicted through the method and principle of dynamic elastic modulus. 2 Material and methods 2
The chemical compositions (in wt.%) of the forged Hadfield steel were 1.20 C, 12.30 Mn, 0.60 Si, 0.016 S, and 0.022 P with Fe as the balance. A uniform austenite microstructure with an average grain size of 150 μm was obtained after being austenitized at 1050 °C for 1 h with water quenching. The surface of the Hadfield steel was subjected to SPD using a new high-speed pounding (HSP) process [3, 28]. Compared to other SPD methods, HSP process is a convenient method with simple device which could obtain a uniform hardened layer on the surface of metal materials easily. Fig. 1 illustrates the sample preparation by HSP process. A high-strength metal pin pounded repeatedly on the surface of the Hadfield steel. When the sample was moving regularly, the surface of the sample underwent intense plastic deformation and a hardened surface layer is produced. The working pressure of the pin and the pounding times are the key parameters in HSP process to control the deformation degree of the surface of the Hadfield steel. A working pressure of 1.6 GPa was selected in the present study. After pounding for 2×104 times, the surface of the Hadfield steel was deformed, and with increasing pounding times to 8×104 times, a nanocrystalline layer was produced on the surface of the Hadfield steel. The nanocrystallized layer and the deformed layer were taken out from the surface of the Hadfield steel after HSP process under 1.6GPa for 8×104 and 2×104 times, respectively. The surface of the Hadfield steel sample during pounding was kept at room temperature using circulating water. The quenched Hadfield steel served as a comparison. The objects of the present study were the Hadfield steel samples under the three states described above.
Figure 1. Schematic illustration of the HSP process
The dynamic thermomechanical analyzer DMA861e (made by Mettler Toledo in Switzerland) was used to measure the elastic modulus as functions of temperature and time at different temperatures. The samples were 50 mm × 5 mm × 0.5 mm in size, and the vibration frequency was 1 Hz. Two results of each state were measured namely relationships between the elastic modulus and temperature from −100 °C to 400 °C with a heating rate of 3 °C/min, and the relationships between the elastic modulus and time heating at 300 °C, 350 °C, and 400 °C for 1 h. The metallurgical microscope Axiover200MAT was used to observe the optical micrographs, and TEM microstructures were observed under a JEM-2010 transmission electron microscope. Thin foils were prepared through a precision ion polishing system (Gatan). The phase composition in the Hadfield steel was analyzed by XRD using Cu target and Kα radiation. The characteristic wavelength λ was 1.540598 Å. The scanning scale was 20° - 120° and the step length was 0.02°. 3
The work voltage was 40 kV, and the current was 100 mA. The classical Williamson–Hall method [29] was used to calculate microstrain <ε2>1/2 and average subgrain size D. The formula for this analysis is as follows: βcosθ 1 2 sinθ 2 ε 2 1 / 2 ( ), λ D λ
(1)
where λ is the wavelength of diffraction wave, and β is the width of diffraction peak. The following formula was used to estimate the dislocation density of the Hadfield steels under the three states before and after aging at 400 °C for 1 h, where ρ is the dislocation density, and b is the magnitude of the Burgers vector. ρ 2 3 ε 2 1 / 2 /(Db) ,
(2)
A three-dimensional atom probe (3DAP) equipment was used to analyze the carbon atom distribution in Hadfield steel. Slender samples with a specification of 0.5 mm × 0.5 mm × 15mm were cut using a wire electrical discharge machine. The surfaces of the samples were polished using silicon carbide paper, and the blunt needle samples used in the test were prepared by chemical polishing. The chemical polishing process was completed in two steps. The samples were etched by mixing 75% acetic acid and 25% perchloric acid, followed by mixing 2% perchloric acid in glycol butyl ether solution [30]. A local electrode atom probe (LEAP) was employed in the atom probe instrument, Imago LEAP 3000X HR. In the analysis of the blunt needle samples, the sample temperature was -223 °C, the pulse frequency was 5 kHz, and the pulse fraction was 20%. The collected data was processed using IVAS software. The three-dimensional distribution of atoms in the Hadfield steel was rebuilt, and the data was analyzed. 3 Results
Figure 2. Elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel as a function of temperature
Fig. 2 shows the elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel in the heating process from −100 °C to 400 °C. Below 300 °C, the elastic modulus of the Hadfield steel under the three states decreased from 200GPa to about 100GPa with increasing temperature. Above 300 °C, the turning points in the elastic modulus of the three samples appeared, after which the elastic modulus increased rapidly. At a heating rate of 4
3 °C/min, the turning temperatures for the nanocrystallized, deformed, and quenched Hadfield steel samples were 371 °C, 377 °C, and 348 °C, respectively. Differences in the elastic modulus of the three samples at different temperatures were observed. On the one hand, the elastic modulus of the nanocrystallized Hadfield steel was the lowest, and that of the quenched Hadfield was the highest when the temperature was below 100 °C. On the other hand, the elastic modulus in increasing order was as follows: the quenched Hadfield steel, the nanocrystallized Hadfield steel, and the deformed Hadfield steel when it was above 100 °C.
Figure 3. Elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel aged at 300 °C (a), 350 °C (b), and 400 °C (c) for 1 h as a function of time
Fig. 3 shows the elastic modulus curves of the three samples aged at 300 °C, 350 °C, and 400 °C as a function of time. The elastic modulus showed no significant change when aged at 300 °C. When the temperature increased to 350 °C (Fig. 3b), the elastic modulus of the quenched Hadfield steel increased gradually at the preliminary stage, but hardly changed about 15 min later. The elastic modulus of the nanocrystallized Hadfield steel and the deformed Hadfield steel did not present the same trend. At 400 °C aging (Fig. 3c), the elastic modulus of all the three samples increased with extended isothermal time. This observation indicated that the bond force among atoms of the three samples did not change at 300 °C. At 350 °C aging, the bond force among atoms changed in the quenched Hadfield steel but not in the others. At 400 °C, continuous changes in the bond force between atoms were reported in all three samples. In addition, the continuous increase in elastic modulus proved the continuous microstructure transition in the samples. Fig. 3 also shows large differences among the elastic modulus of the three samples at different isothermal temperatures. At the aging temperature of 300 °C, the elastic modulus of the three samples in increasing order was as follows: the quenched Hadfield steel, nanocrystallized Hadfield steel, and deformed Hadfield steel. However, at 350 °C and 400 °C, the elastic modulus of the samples in increasing order was as follows: the nanocrystallized Hadfield steel, quenched Hadfield steel, and deformed Hadfield steel. These findings were consistent with the results in Fig. 2. Thus, the changes in elastic modulus were different in the Hadfield steels under distinct states at different temperatures. The differences in the bond force among atoms and the microscopic residual stress of the Hadfield steel under distinct states resulted in differences in the elastic modulus. 5
Figure 4. Optical micrographs of the Hadfield steel before and after aging at 400 °C for 1 h a, quenched Hadfield steel–not aged; b, deformed Hadfield steel–not aged; c, nanocrystallized Hadfield steel–not aged; d, quenched Hadfield steel–aged; e, deformed Hadfield steel–aged; f, nanocrystallized Hadfield steel–aged
Fig. 4 shows the optical micrographs of the Hadfield steel under three states before and after aging at 400 °C for 1 h. The microstructures of the three samples were single-phase austenite, deformed austenite, and microstructures with unclear details. After aging at 400 °C for 1 h, a precipitation reaction occurred in the Hadfield steel. The precipitates in the quenched Hadfield steel were coarse and needle shaped. The precipitates in the deformed Hadfield steel were also clearly observed. In addition, the precipitates were mainly observed at the prior austenite grain boundaries. However, this behavior was not observed for the precipitates in the nanocrystallized Hadfield steel. The reason might be that carbides mainly precipitated at the grain boundaries and the microstructures were nanocrystalline so that the precipitates in the nanocrystallized Hadfield steel were also very fine.
Figure 5. XRD patterns of the nanocrystallized sample (NS), deformed sample (DS), and quenched sample (QS) after aging at different temperatures for 1 h
To ensure the structure and chemical composition of the precipitates formed at the turning temperatures described in Fig. 2, XRD phase composition analysis of the three samples aged at 300 °C, 350 °C and 400 °C for 1 h was conducted. The results are shown in Fig. 5. The microstructure of the three samples was made of single-phase austenite after aging at 300 °C. When 6
the aging temperature was 400 °C, diffraction peaks of fcc austenite and a few Fe3C were obtained. These findings proved that the turning points of the elastic modulus curves shown in Fig. 2 were caused by carbide precipitation, and the precipitates observed in the optical micrographs were carbides in the Fe3C type. The intensity of the diffraction peaks of Fe3C showed that the amount of carbides in the nanocrystallized Hadfield steel was less than that in the quenched Hadfield steel.
Figure 6. TEM microstructures of the Hadfield steel and their corresponding diffraction patterns a, the quenched Hadfield steel–not aged; b, the deformed Hadfield steel–not aged; c, the nanocrystallized Hadfield steel–not aged; d, the quenched Hadfield steel–aged; e, the deformed Hadfield steel–aged; f, the nanocrystallized Hadfield steel–aged
TEM microstructures of the three samples before and after aging at 400 °C for 1 h are presented in Fig. 6. Before the aging process, the microstructure in the quenched Hadfield steel was prior austenite with some dislocations. A severely deformed microstructure with high density dislocations and twins was observed in the deformed Hadfield steel. The nanocrystallized Hadfield steel was reported to be nanocrystalline, and the average size of the grains was about 60 nm. After aging at 400 °C for 1 h, a large number of carbides were observed in all of the Hadfield steel samples, but the carbide morphologies and sizes were different. The average size of the precipitated carbides in the quenched Hadfield steel was as large as 500 nm, and they appeared in a block shape. The block-shaped carbides in the deformed Hadfield steel were about 300 nm in size, which was smaller than that in the quenched Hadfield steel. In the nanocrystallized Hadfield steel, the carbides were much smaller, more distributed, granular, and almost below 100 nm in size. These findings proved that plastic deformation could refine the carbides precipitated in the Hadfield steel, and that the size of the carbides in the nanocrystallized Hadfield steel reached nano-scale.
7
Figure 7. Atom distribution maps of quenched Hadfield steel (a), deformed Hadfield steel (b), and nanocrystallized Hadfield steel (c)
Fig. 7 shows the 3DAP results of the Hadfield steel under three states, presenting the atom distribution in an area of 40 nm × 25 nm × 0.3 nm. The thickness of the area was 0.3 nm, which was similar to the lattice thickness of the Hadfield steel (the thickness of a single lattice of f.c.c Hadfield steel was 0.36 nm). The red ball in Fig. 7 represents a gathering of three carbon atoms, the blue ball represents a gathering of two carbon atoms, the black red ball represents one carbon atom, and the green ball represents one manganese atom. The distribution of the manganese atoms was relatively uniform, whereas the distribution of the carbon atoms differed largely. 4 Discussion The elastic properties of solids came from the interaction among atoms, and the reaction of the solids to the interaction resulted from the interaction potential energy among atoms. As the temperature increased, the distance between the atoms increased, the reaction weakened, and the elastic modulus declined. Thus, the decrease in elastic modulus curves of the Hadfield steel as a function of temperature at the initial stage (below 350 °C) was caused by the increasing temperature as shown in Fig. 2. When the temperature was above 350 °C, the elastic modulus curves turned because the carbides precipitated in the Hadfield steel under the three distinct states with increasing temperature, shown in Fig. 4, 5, and 6. The elastic modulus of a metal material is a mechanical performance index, which is not sensitive to microstructures. Only factors that influence the bond strength can influence the elastic modulus. Carbides, as a precipitation, are greatly different from the matrix microstructures of the Hadfield steel. The bond force among atoms in the carbides was stronger than that in the austenitic lattice of the Hadfield steel, and it resulted in the elastic modulus of the austenitic Hadfield steel and carbide of Fe3C of 208 GPa and 216 GPa, respectively [31]. When carbides precipitated, the volume fraction of the interface in Hadfield steel increased. The increased interface volume limited the dislocation movement, resulting in the dislocation pile-up. In the aging process, the solute atoms would migrate to the center of the dislocations, and this migration would lead to a further increase in the elastic modulus [32]. Therefore, the elastic modulus of the Hadfield steel increased rapidly after the carbide precipitation. However, the carbide precipitation temperatures of the Hadfield steel under three distinct states were different. The 8
precipitation temperatures of the nanocrystallized Hadfield steel, the deformed Hadfield steel, and the quenched Hadfield steel were 371 °C, 377 °C, and 348 °C, respectively. The following section explains the reason behind these findings. After plastic deformation, the distortion of crystal lattice became engendered, and defect density increased in the Hadfield steel. Except for a small part transited to thermal energy, most of the mechanical energy during the HSP was absorbed as distortion energy, resulting in high internal energy. Compared with the quenched Hadfield steel in the low-temperature isothermal process, the distortion energy in the deformed Hadfield steel was released during heat treatment, and served as the energy for carbide precipitation. Thus, the carbide precipitation temperature relatively decreased. However, this theory was inconsistent with our results, suggesting that other factors could also influence carbide precipitation behavior. Table 1. Statistics of carbon atom distribution in a random area of 3000 nm3 Condition Three-carbon atom cluster Two-carbon atom cluster One carbon atom
The quenched Hadfield
The deformed Hadfield
The nanocrystallized Hadfield
steel
steel
steel
45.4%
42.5%
41.8%
39.8%
38.8%
37.9%
14.8%
18.7%
20.3%
To further analyze the carbon and manganese atom distribution and the interaction between them, the APT results were cut into pieces. Moreover, to avoid the overlap of congeneric atoms at the same position of different planes that could result in a misunderstanding of the atom distribution, each piece was cut with a thickness of 0.3 nm. Carbon atoms are generally considered to be located in the octahedral interstice. The distance between the opposite atoms at vertexes of the octahedron was 0.36 nm. Atoms could not overlap at the same positions on different planes. A rotating video of the APT result showing the atom distribution in an area of 30 nm × 10 nm × 0.6 nm is provided in Video Data to support the analysis. Based on that, 10 pieces of the APT results, each with an area of 40 nm × 25 nm × 0.3 nm, were statistically analyzed, and the distribution of the carbon atoms in the Hadfield steel was determined as shown in Table 1. The data in Table 1 were random with a volume of 3000 nm3. Compared with that in the quenched Hadfield steel, the fraction of three-carbon atom clusters and two-carbon atom clusters in the deformed Hadfield steel decreased by 2.9% and 1.0%, respectively, and the one-carbon atom increased by 3.9%. For the nanocrystallized Hadfield steel, the fractions of the three-carbon, two-carbon, and one-carbon atom clusters were 3.6%, 1.9%, and 5.5%, respectively. These findings showed that a redistribution of carbon atoms in the Hadfield steel occurred in the plastic deformation process. With increasing deformation extent, the number of three-carbon atom clusters and two-carbon atom clusters decreased, thus indicating that the carbon 9
atoms in the Hadfield steel were more uniform after deformation. The correlation degree between carbon and manganese atoms was not apparent, which indicated that the short-range order C-Mn couples could not be determined based on the APT results. However, according to previous reports and results obtained by other methods, Hadfield steels had a short-range order of C-Mn couples [15, 16]. Thus, the APT results demonstrated that plastic deformation enhanced the short-range order of the C-Mn couples in the Hadfield steel. Generally, carbon atoms were present in the octahedral interstice of the fcc γ-Fe in the Hadfield steel, and partial iron atoms were replaced by manganese atoms to form a Fe-Mn-C solid solution. Table 1 presents numerous three-carbon atom clusters and two-carbon atom clusters in the Hadfield steels under three distinct states. However, the octahedral interstice of γ-Fe in Hadfield steels could only accommodate one carbon atom according to the calculation results of exchange energy [33]. Thus, we inferred that the two-carbon and three-carbon atom clusters occurred at the vacancies, dislocations, or the grain boundaries. As shown in Table 1, after water quenching treatment, about 85% of carbon atoms existed in the dislocations and vacancies produced during quenching, as well as at the grain boundaries. After the HSP, a number of dislocations and vacancies were induced as shown in Table 2. The dislocation density in the quenched Hadfield steel was only 0.3 × 1014 m2, whereas those in the deformed Hadfield steel and nanocrystallized Hadfield steel were high at 9.2 × 1014 m2 and 2.2 × 1014 m2, respectively. Dislocations of such high density provided diffusion channels and existence spaces for carbon atoms. On the effect of thermal energy and deformation energy, the three-carbon and two-carbon atom clusters decomposed and diffused to newly generated dislocations and vacancies, thereby decreasing the distortion energy in the Hadfield steel. As shown in Fig. 6, Fig. 7, and Table 1, the diffusion process of carbon atoms could be explained as follows. In the quenched Hadfield steel, the three-carbon and two-carbon atom clusters were mainly observed at the grain boundaries, quenching vacancies, or in a few dislocations. When the Hadfield steel was subjected to plastic deformation, a number of vacancies, dislocations, and grain boundaries were introduced. Carbon atoms were dragged by the dislocations. Parts of the three-carbon and two-carbon atom clusters were decomposed to single carbon atoms. Therefore, the carbon atoms distributed more uniformly. Simply but, plastic deformation promoted the reaction between the carbon atoms and defects. The dislocation density decreased after aging treatment, but the dislocation density in the deformed Hadfield steel and the nanocrystallized Hadfield steel was much higher than that in the quenched Hadfield steel. According to binding energy theory of carbon atoms and dislocations [29, 34], carbide precipitation is inhibited when the binding enthalpy of carbon atom and dislocation is higher than the formation enthalpy of the carbides. When the Hadfield steel was subjected to plastic deformation, an extremely large number of dislocations and vacancies promoted the diffusion of carbon atoms to the dislocation strain field. This process promoted the homogenization of carbon atoms and delayed the precipitation of carbides in the 10
Hadfield steel. Therefore, the carbide precipitation temperature of the deformed Hadfield steel and the nanocrystallized Hadfield steel was higher than that of the quenched Hadfield steel (see Fig. 2). Table 2. Dislocation density of the Hadfield steels under three distinct states before and after aging at 400 °C ( m2) Condition
Quenched Hadfield steel 14
Before aging
0.3 × 10
After aging
0.2 × 1014
Deformed Hadfield steel
Nanocrystallized Hadfield steel
14
2.2 × 1014
3.3 × 1014
1.8 × 1014
9.2 × 10
The lattice constant of Hadfield steel does not change when carbon atoms segregate to vacancies, dislocations, and grain boundaries [35, 36]. The elastic modulus is only influenced by temperature if no carbide precipitation occurs. Subsequently, the carbon atoms around the grain boundaries segregate to the grain boundaries because of interfacial energy. When the carbon atom concentration increases to a certain degree, the carbides precipitate at the grain boundaries [37]. Notably, the energy at the grain boundaries was high with numerous defects, resulting in a faster atom diffusion rate than that in the grains. As a result, the carbide precipitation rate was higher at the grain boundaries than in the grains. Compared with the deformed Hadfield steel, more grain boundaries and much higher interfacial energy in unit volume were found in the nanocrystallized Hadfield steel. This finding illustrated the many nucleation sites and high driving force for carbide precipitation in the aging treatment. Thus, the carbide precipitation temperature of the nanocrystallized Hadfield steel was lower than that of the deformed Hadfield steel (see Fig. 2). Furthermore, the smaller the nanocrystallization was, the finer the carbides (see Fig. 6). Carbide precipitation was associated with the diffusion of carbon and alloying element atoms. Temperature determined the precipitation possibility and the precipitation rate. At a relatively low temperature (300 °C), no carbide precipitation was observed because of low diffusion capacity of the solute atoms (see Fig. 3). Therefore, the elastic modulus of the Hadfield steel under three distinct states did not change in the aging process (see Fig. 3a). By increasing the aging temperature to 400 °C, the diffusion capacity of atoms was enhanced, and the carbides precipitated gradually. In addition, the amount of carbides increased by extending the isothermal aging time, thereby continuously increasing the elastic modulus of the Hadfield steel (see Fig. 3c). However, the increase amplitude in the Hadfield steel under three distinct states was different, and the change in elastic modulus was the smallest in the nanocrystallized Hadfield steel. This result was attributed to the minimum transformation of carbides (see Fig. 5), and it further indicated that the nanocrystallization process impeded the formation of carbides in Hadfield steel. Table 3. Hardness of the Hadfield steels under three distinct states before and after aging at 400 °C (GPa) Condition
Quenched Hadfield steel
Deformed Hadfield steel
Nanocrystallized Hadfield steel
Before aging
1.68
7.43
7.52
After aging
1.85
8.43
8.85
The change regulation in the elastic modulus of the Hadfield steel was analyzed under three 11
distinct states when the aging temperature was below 350 °C (Fig. 2). Residual stress decreases the elastic modulus of metal materials [18]. After the plastic deformation and the nanocrystallization process, the hardness of the Hadfield steels increased as shown in Table 3. However, the hardness remained high even after aging treatment. This result indicated that high residual stress was induced in the Hadfield steels by the HSP, which would inevitably decrease the elastic modulus of the Hadfield steel. This result could explain why the elastic modulus of the deformed and the nanocrystallized Hadfield steel samples was lower than that of the quenched Hadfield steel when the testing temperature was below 100 °C. By increasing the testing temperature, local fine-tuning of atoms in the deformed and nanocrystallized Hadfield steel samples occurred, thus resulting in internal stress relaxation and elimination. In addition, the elastic modulus increased continuously. The amounts of the three-carbon and the two-carbon atom clusters were reduced by plastic deformation, as shown in Table 1. This process promoted the uniform distribution of carbon atoms. In other words, the short-range order of the C-Mn couples in the Hadfield steel increased. Thus, when the aging temperature ranged from 100 °C to 350 °C, the elastic modulus of the deformed and nanocrystallized Hadfield steel samples was higher than that of the quenched Hadfield steel. 5 Conclusions The following conclusions were drawn from the present work: (1) Most of the carbon atoms in the Hadfield steel were present in the defects, such as dislocations, vacancies, and grain boundaries. Plastic deformation caused a uniform distribution of the carbon atoms that increased the short-range order of the C-Mn couples in the Hadfield steel. (2) The HSP and the nanocrystallization process impeded carbide precipitation and improved the precipitation temperature in the Hadfield steel. The carbide precipitation temperatures of the quenched Hadfield steel, the deformed Hadfield steel, and the nanocrystallized Hadfield steel were 348 °C, 377 °C, and 371 °C, respectively at a heating rate of 3 °C/min. The nanocrystalline process of the Hadfield steel effectively reduced the carbide size to nano-scale, increased the carbide dispersion, and decreased the amount of carbides. (3) Carbide precipitation in the Hadfield steel remarkably increased the elastic modulus. Thus, the carbide precipitation behavior of steels could be determined by measuring the change in the elastic modulus through dynamic thermomechanical analysis.
Acknowledgement The present work was supported by the National Natural Science Youth Fund (No. 51501161).
References [1] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103-189. [2] A.P. Zhilyaev, T.G. Langdon. Prog. Mater. Sci. 53 (2008) 893-979. [3] F.C. Zhang, X.Y. Feng, Z.N. Yang, J. Kang & T.S. Wang, Sci. Rep. 5 (2015). [4] R.Z. Valiev, I.V. Alexandrov, Y.T. Zhu, T.C. Lowe, J. Mater. Res. 17 (2002) 5-8. 12
[5] K.S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mate. 51 (2003) 5743-5774. [6] K. Lu, Mater. Sci. Eng. R 16 (1996) 161-221. [7] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K. Lu, Acta Mater. 50 (2002) 4603-4616. [8] N.R. Tao, M.L. Sui, J. Lu, K. Lu, Nanostruct. Mater. 11 (1999) 433-440. [9] J.Y. Huang, Y.T. Zhu, X.Z. Liao, R.Z. Valiev, Philos. Mag. Lett. 84 (2004) 183-190. [10] L.C. Zhang, M. Calin, F. Paturaud, C. Mickel, J. Eckert, Appl. Phys. Lett. 90 (2007) 201908. [11] Y. Ivanisenko, I. Maclaren, X. Sauvage, R.Z. Valiev, H.J. Fecht, Acta Mater. 54 (2006) 1659-1669. [12] S.M. Fatemi-Varzaneha, A. Zarei-Hanzakia, H. Paul, Mater. Charact. 87 (2014) 27-35. [13] R.A. Hadfield, Science 12 (1888) 284-286. [14] E. Bayraktar, F.A. Khalid, C. Levaillant, J. Mater. Process Technol. 147 (2004) 145-154. [15] Y.N. Dastur, W.C. Leslie, Metall. Trans. A 12 (1981) 749-759. [16] F.C. Zhang, Y.Z. Zheng, R.F. Zhu, T.Q. Lei, Scripta Met. Mater. 32 (1995) 1477-1481. [17] B.A. Latella, S.R. Humphries, Scripta Mater. 51 (2004) 635-639. [18] R. Rahemi, D.Y. Li, Scripta Mater. 99 (2015) 41-44. [19] Q. Chen, W.G. Mao, Y.C. Zhou, C.Lu, Appl. Surf. Sci. 256 (2010) 7311-7315. [20] T. Tayeh, J. Douin, S. Jouannigot, M. Zakhour, M. Nakhl, J.F. Silvain, J.L. Bobet, Mater. Sci. Eng. A 591 (2014) 1-8. [21] A.P. Stebner, D.W. Brown, L.C. Brinson, Acta Mater. 61 (2013) 1944-1956. [22] H. Zhang, B.S. Mitchell, Mater. Sci. Eng. A 270 (1999) 237-243. [23] J.M. Fiorani, F.A. Kuhnast, C. Cunat, E. Illekova, Mater. Sci. Eng. A 165 (1993) 143-148. [24] A. Yasmin, I.M.Daniel, Polymer 45 (2004) 8211-8219. [25] K.J. Green, D.R. Dean, U.K. Vaidya, E. Nyairo, Compos. Part A-Appl. S. 40 (2009) 1470-1475. [26] S. Deng, M. Hou, L. Ye, Polym. Test. 26 (2007) 803-813. [27] G. Laino, R. De Santis, A. Gloria, T. Russo, D.S. Quintanilla, A. Laino, R. Martina, L. Nicolais, L. Ambrosio, J. Biomater. Appl. 26 (2012) 1-16. [28] F.C. Zhang, Z.N. Yang, L.H. Qian, F.C. Liu, B. Lv, M. Zhang, Scripta Mater. 64 (2011) 560-563. [29] V. Gavriljuk, Scripta Mater. 46 (2002) 175-177. [30] M.K. Miller, Atom Probe Tomography: Analysis at the atomic level, Kluwer Academic/Plenum Publishers, New York, 2000. [31] L.A. Glikman, A.M. Kartashov, Z.M. Rubashkina, A.F. Lobov, Strength Mater. 7 (1975) 523-524. [32] V.N. Bastun, N.I. Chernyak, Strength Mater. 3 (1971) 1062-1065. [33] W.S. Owen and M. Grujicic, Acta mater. 47 (1999) 111-126. [34] V.G. Gavriljuk, Scripta Mater. 45 (2001) 1469-1472. [35] Y. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.J. Fecht, Acta Mater. 51 (2003) 5555-5570. [36] C. Borchers, T. Al-Kassab, S. Goto & R. Kirchheim, Mater. Sci. Eng. A 502 (2009) 131-138. [37] J. Languillaume, G. Kapclski & B. Baudelet, Acta Mater. 45 (1997) 1201-1212.
13
Video caption: "Video data of the APT results showing the atom distribution in an area of 30 nm × 10 nm × 0.6 nm in the quenched Hadfield steel".
14
PII: DOI: Reference:
S0921-5093(16)30996-0 http://dx.doi.org/10.1016/j.msea.2016.08.084 MSA34042
To appear in: Materials Science & Engineering A Received date: 25 May 2016 Revised date: 18 August 2016 Accepted date: 19 August 2016 Cite this article as: C. Chen, X.Y. Feng, B. Lv, Z.N. Yang and F.C. Zhang, A Study on Aging Carbide Precipitation Behavior of Hadfield Steel by Dynamic Elastic Modulus, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2016.08.084 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A Study on Aging Carbide Precipitation Behavior of Hadfield Steel by Dynamic Elastic Modulus C. Chen a, X.Y. Feng a, B. Lv b,*, Z.N. Yang c, F.C. Zhang a, c a
State Key Laboratory of Metastable Materials Science and Technology, Yanshan University, Qinhuangdao 066004, China
b c
College of Environmental and Chemical Engineering , Yanshan University, Qinhuangdao 066004, China
National Engineering Research Center for Equipment and Technology of Cold Strip Rolling, Yanshan University, Qinhuangdao 066004, China
Abstract: In the present study, the carbon atom distribution and carbide precipitation behavior of Hadfield steel under three distinct states, namely, the quenched Hadfield steel, the deformed Hadfield steel (coarse grain), and the nanocrystallized Hadfield steel, were studied using dynamic thermomechanical analysis, transmission electron microscopy, three-dimensional atom probe, X-ray diffraction, and optical microscopy. Results showed apparent differences in elastic modulus change in the Hadfield steel under three distinct states in the dynamic thermomechanical analysis. Those results proved that the dynamic elastic modulus could be used to determine carbide precipitation behavior in steels during heat treatment. The carbide precipitation temperatures of the nanocrystallized Hadfield steel and the deformed Hadfield steel were higher than that of the quenched Hadfield steel. This phenomenon was attributed to the promoted diffusion of carbon atoms to the stress field of defects during plastic deformation and nanocrystallization process. This process induced a more homogeneous distribution of carbon atoms in the Hadfield steel. As a result, the precipitated carbides were much finer and more dispersive, especially in the nanocrystallized Hadfield steel. Key words: Hadfield steel; Carbides; Elastic modulus; Nanocrystalline microstructures 1 Introduction Severe plastic deformation (SPD), such as equal channel angular pressing, accumulative roll bonding, surface mechanical polishing, and shot peening, is considered the main method to prepare dense, nonpolluting, and ultrafine-grained microstructure materials [1–4]. However, these materials are in a thermodynamically metastable state because of wide grain boundary and lattice distortion induced by SPD [5–8]. Except for the evolution in grain size, twin density, and dislocation density, SPD could also result in the redistribution of alloying elements, and influence phase transformation [9–12]. Therefore, the microstructure evolution, including the kinetics and thermodynamics of phase transition, the microstructures, and the phase structures of the ultrafine-grained and nanocrystallized metal material prepared by SPD would change during heat treatment. Hadfield steel is a type of stable single-phase austenitic steel. Compared with other steels,
*
Corresponding author. Tel.: +86 335 8063949; e-mail addresses: [email protected] (B. Lv) 1
carbides easily precipitate in the Hadfield steel especially at the grain boundaries when the temperature is above 300 °C [13]. However, precipitated carbides can severely worsen mechanical properties. For this reason, some researchers investigated the delays and suppression in carbide precipitation during heat treatment [14]. For 130 years, material researchers have focused their attention on the work hardening property and microstructure reactions in Hadfield steel, some of which have not been determined yet. For instance, the excellent work hardening property is widely attributed to the block of dislocation movement by the dynamic strain aging effect resulting from the interaction between C–Mn couples and dislocations [15, 16]. Determining where the C atoms of the C–Mn couples in Hadfield steel are present, whether in the crystal lattice or in the strain field induced by vacancies and dislocations, is crucial but is rarely given attention. This process affects the deformation and the work hardening behaviors of Hadfield steel, and it influences atom diffusion and carbide precipitation behaviors in the heating process. Elastic modulus is an important property parameter that is a reflection of bond strength among atoms, ions, and molecules. Any factor that influences bond strength, such as crystal structure, residual stress, and temperature, can influence the elastic modulus [17, 18]. For the past few years, a number of reports have shown that the change in elastic modulus could be used to analyze residual stress [19], precipitation [20], and texture [21] during heat treatment or plastic deformation. The elastic modulus of a metal material is usually tested through tensile, nano-indentation, bending, ultrasonic, and vibration processes. The first four methods are known as static methods, and the last one is a dynamic method. Dynamic thermomechanical analysis is used to measure the elastic modulus of cement, concrete, and composite materials during heating [22-25]. For dynamic thermomechanical analysis method, deformation occurs in samples when mechanical stress with periodic variation acted on materials under different external conditions. When the stress acts on an elastomer, the resulting strain is proportional to the stress. The proportionality constant is the elastic modulus according to Hooke's law [26]. However, reports about using dynamic thermomechanical analysis to measure the elastic modulus of metal materials are rare [27], especially studies on the precipitation behaviors in steels. In the present work, the deformed and nanocrystallized Hadfield steel samples prepared by a new SPD process, as well as the quenched Hadfield steel sample, were studied. Dynamic thermomechanical analysis was used to measure the change in the elastic modulus of the Hadfield steel under three distinct states in the aging process. Methods of optical microscopy, X-Ray diffraction (XRD), transmission electron microscopy (TEM), and atom probe tomography (APT) were used to analyze the influence of deformation and nanocrystalline on carbon distribution, microstructure, and carbide precipitation behavior. Moreover, the carbide precipitation behavior was successfully predicted through the method and principle of dynamic elastic modulus. 2 Material and methods 2
The chemical compositions (in wt.%) of the forged Hadfield steel were 1.20 C, 12.30 Mn, 0.60 Si, 0.016 S, and 0.022 P with Fe as the balance. A uniform austenite microstructure with an average grain size of 150 μm was obtained after being austenitized at 1050 °C for 1 h with water quenching. The surface of the Hadfield steel was subjected to SPD using a new high-speed pounding (HSP) process [3, 28]. Compared to other SPD methods, HSP process is a convenient method with simple device which could obtain a uniform hardened layer on the surface of metal materials easily. Fig. 1 illustrates the sample preparation by HSP process. A high-strength metal pin pounded repeatedly on the surface of the Hadfield steel. When the sample was moving regularly, the surface of the sample underwent intense plastic deformation and a hardened surface layer is produced. The working pressure of the pin and the pounding times are the key parameters in HSP process to control the deformation degree of the surface of the Hadfield steel. A working pressure of 1.6 GPa was selected in the present study. After pounding for 2×104 times, the surface of the Hadfield steel was deformed, and with increasing pounding times to 8×104 times, a nanocrystalline layer was produced on the surface of the Hadfield steel. The nanocrystallized layer and the deformed layer were taken out from the surface of the Hadfield steel after HSP process under 1.6GPa for 8×104 and 2×104 times, respectively. The surface of the Hadfield steel sample during pounding was kept at room temperature using circulating water. The quenched Hadfield steel served as a comparison. The objects of the present study were the Hadfield steel samples under the three states described above.
Figure 1. Schematic illustration of the HSP process
The dynamic thermomechanical analyzer DMA861e (made by Mettler Toledo in Switzerland) was used to measure the elastic modulus as functions of temperature and time at different temperatures. The samples were 50 mm × 5 mm × 0.5 mm in size, and the vibration frequency was 1 Hz. Two results of each state were measured namely relationships between the elastic modulus and temperature from −100 °C to 400 °C with a heating rate of 3 °C/min, and the relationships between the elastic modulus and time heating at 300 °C, 350 °C, and 400 °C for 1 h. The metallurgical microscope Axiover200MAT was used to observe the optical micrographs, and TEM microstructures were observed under a JEM-2010 transmission electron microscope. Thin foils were prepared through a precision ion polishing system (Gatan). The phase composition in the Hadfield steel was analyzed by XRD using Cu target and Kα radiation. The characteristic wavelength λ was 1.540598 Å. The scanning scale was 20° - 120° and the step length was 0.02°. 3
The work voltage was 40 kV, and the current was 100 mA. The classical Williamson–Hall method [29] was used to calculate microstrain <ε2>1/2 and average subgrain size D. The formula for this analysis is as follows: βcosθ 1 2 sinθ 2 ε 2 1 / 2 ( ), λ D λ
(1)
where λ is the wavelength of diffraction wave, and β is the width of diffraction peak. The following formula was used to estimate the dislocation density of the Hadfield steels under the three states before and after aging at 400 °C for 1 h, where ρ is the dislocation density, and b is the magnitude of the Burgers vector. ρ 2 3 ε 2 1 / 2 /(Db) ,
(2)
A three-dimensional atom probe (3DAP) equipment was used to analyze the carbon atom distribution in Hadfield steel. Slender samples with a specification of 0.5 mm × 0.5 mm × 15mm were cut using a wire electrical discharge machine. The surfaces of the samples were polished using silicon carbide paper, and the blunt needle samples used in the test were prepared by chemical polishing. The chemical polishing process was completed in two steps. The samples were etched by mixing 75% acetic acid and 25% perchloric acid, followed by mixing 2% perchloric acid in glycol butyl ether solution [30]. A local electrode atom probe (LEAP) was employed in the atom probe instrument, Imago LEAP 3000X HR. In the analysis of the blunt needle samples, the sample temperature was -223 °C, the pulse frequency was 5 kHz, and the pulse fraction was 20%. The collected data was processed using IVAS software. The three-dimensional distribution of atoms in the Hadfield steel was rebuilt, and the data was analyzed. 3 Results
Figure 2. Elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel as a function of temperature
Fig. 2 shows the elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel in the heating process from −100 °C to 400 °C. Below 300 °C, the elastic modulus of the Hadfield steel under the three states decreased from 200GPa to about 100GPa with increasing temperature. Above 300 °C, the turning points in the elastic modulus of the three samples appeared, after which the elastic modulus increased rapidly. At a heating rate of 4
3 °C/min, the turning temperatures for the nanocrystallized, deformed, and quenched Hadfield steel samples were 371 °C, 377 °C, and 348 °C, respectively. Differences in the elastic modulus of the three samples at different temperatures were observed. On the one hand, the elastic modulus of the nanocrystallized Hadfield steel was the lowest, and that of the quenched Hadfield was the highest when the temperature was below 100 °C. On the other hand, the elastic modulus in increasing order was as follows: the quenched Hadfield steel, the nanocrystallized Hadfield steel, and the deformed Hadfield steel when it was above 100 °C.
Figure 3. Elastic modulus curves of the nanocrystallized Hadfield steel, deformed Hadfield steel, and quenched Hadfield steel aged at 300 °C (a), 350 °C (b), and 400 °C (c) for 1 h as a function of time
Fig. 3 shows the elastic modulus curves of the three samples aged at 300 °C, 350 °C, and 400 °C as a function of time. The elastic modulus showed no significant change when aged at 300 °C. When the temperature increased to 350 °C (Fig. 3b), the elastic modulus of the quenched Hadfield steel increased gradually at the preliminary stage, but hardly changed about 15 min later. The elastic modulus of the nanocrystallized Hadfield steel and the deformed Hadfield steel did not present the same trend. At 400 °C aging (Fig. 3c), the elastic modulus of all the three samples increased with extended isothermal time. This observation indicated that the bond force among atoms of the three samples did not change at 300 °C. At 350 °C aging, the bond force among atoms changed in the quenched Hadfield steel but not in the others. At 400 °C, continuous changes in the bond force between atoms were reported in all three samples. In addition, the continuous increase in elastic modulus proved the continuous microstructure transition in the samples. Fig. 3 also shows large differences among the elastic modulus of the three samples at different isothermal temperatures. At the aging temperature of 300 °C, the elastic modulus of the three samples in increasing order was as follows: the quenched Hadfield steel, nanocrystallized Hadfield steel, and deformed Hadfield steel. However, at 350 °C and 400 °C, the elastic modulus of the samples in increasing order was as follows: the nanocrystallized Hadfield steel, quenched Hadfield steel, and deformed Hadfield steel. These findings were consistent with the results in Fig. 2. Thus, the changes in elastic modulus were different in the Hadfield steels under distinct states at different temperatures. The differences in the bond force among atoms and the microscopic residual stress of the Hadfield steel under distinct states resulted in differences in the elastic modulus. 5
Figure 4. Optical micrographs of the Hadfield steel before and after aging at 400 °C for 1 h a, quenched Hadfield steel–not aged; b, deformed Hadfield steel–not aged; c, nanocrystallized Hadfield steel–not aged; d, quenched Hadfield steel–aged; e, deformed Hadfield steel–aged; f, nanocrystallized Hadfield steel–aged
Fig. 4 shows the optical micrographs of the Hadfield steel under three states before and after aging at 400 °C for 1 h. The microstructures of the three samples were single-phase austenite, deformed austenite, and microstructures with unclear details. After aging at 400 °C for 1 h, a precipitation reaction occurred in the Hadfield steel. The precipitates in the quenched Hadfield steel were coarse and needle shaped. The precipitates in the deformed Hadfield steel were also clearly observed. In addition, the precipitates were mainly observed at the prior austenite grain boundaries. However, this behavior was not observed for the precipitates in the nanocrystallized Hadfield steel. The reason might be that carbides mainly precipitated at the grain boundaries and the microstructures were nanocrystalline so that the precipitates in the nanocrystallized Hadfield steel were also very fine.
Figure 5. XRD patterns of the nanocrystallized sample (NS), deformed sample (DS), and quenched sample (QS) after aging at different temperatures for 1 h
To ensure the structure and chemical composition of the precipitates formed at the turning temperatures described in Fig. 2, XRD phase composition analysis of the three samples aged at 300 °C, 350 °C and 400 °C for 1 h was conducted. The results are shown in Fig. 5. The microstructure of the three samples was made of single-phase austenite after aging at 300 °C. When 6
the aging temperature was 400 °C, diffraction peaks of fcc austenite and a few Fe3C were obtained. These findings proved that the turning points of the elastic modulus curves shown in Fig. 2 were caused by carbide precipitation, and the precipitates observed in the optical micrographs were carbides in the Fe3C type. The intensity of the diffraction peaks of Fe3C showed that the amount of carbides in the nanocrystallized Hadfield steel was less than that in the quenched Hadfield steel.
Figure 6. TEM microstructures of the Hadfield steel and their corresponding diffraction patterns a, the quenched Hadfield steel–not aged; b, the deformed Hadfield steel–not aged; c, the nanocrystallized Hadfield steel–not aged; d, the quenched Hadfield steel–aged; e, the deformed Hadfield steel–aged; f, the nanocrystallized Hadfield steel–aged
TEM microstructures of the three samples before and after aging at 400 °C for 1 h are presented in Fig. 6. Before the aging process, the microstructure in the quenched Hadfield steel was prior austenite with some dislocations. A severely deformed microstructure with high density dislocations and twins was observed in the deformed Hadfield steel. The nanocrystallized Hadfield steel was reported to be nanocrystalline, and the average size of the grains was about 60 nm. After aging at 400 °C for 1 h, a large number of carbides were observed in all of the Hadfield steel samples, but the carbide morphologies and sizes were different. The average size of the precipitated carbides in the quenched Hadfield steel was as large as 500 nm, and they appeared in a block shape. The block-shaped carbides in the deformed Hadfield steel were about 300 nm in size, which was smaller than that in the quenched Hadfield steel. In the nanocrystallized Hadfield steel, the carbides were much smaller, more distributed, granular, and almost below 100 nm in size. These findings proved that plastic deformation could refine the carbides precipitated in the Hadfield steel, and that the size of the carbides in the nanocrystallized Hadfield steel reached nano-scale.
7
Figure 7. Atom distribution maps of quenched Hadfield steel (a), deformed Hadfield steel (b), and nanocrystallized Hadfield steel (c)
Fig. 7 shows the 3DAP results of the Hadfield steel under three states, presenting the atom distribution in an area of 40 nm × 25 nm × 0.3 nm. The thickness of the area was 0.3 nm, which was similar to the lattice thickness of the Hadfield steel (the thickness of a single lattice of f.c.c Hadfield steel was 0.36 nm). The red ball in Fig. 7 represents a gathering of three carbon atoms, the blue ball represents a gathering of two carbon atoms, the black red ball represents one carbon atom, and the green ball represents one manganese atom. The distribution of the manganese atoms was relatively uniform, whereas the distribution of the carbon atoms differed largely. 4 Discussion The elastic properties of solids came from the interaction among atoms, and the reaction of the solids to the interaction resulted from the interaction potential energy among atoms. As the temperature increased, the distance between the atoms increased, the reaction weakened, and the elastic modulus declined. Thus, the decrease in elastic modulus curves of the Hadfield steel as a function of temperature at the initial stage (below 350 °C) was caused by the increasing temperature as shown in Fig. 2. When the temperature was above 350 °C, the elastic modulus curves turned because the carbides precipitated in the Hadfield steel under the three distinct states with increasing temperature, shown in Fig. 4, 5, and 6. The elastic modulus of a metal material is a mechanical performance index, which is not sensitive to microstructures. Only factors that influence the bond strength can influence the elastic modulus. Carbides, as a precipitation, are greatly different from the matrix microstructures of the Hadfield steel. The bond force among atoms in the carbides was stronger than that in the austenitic lattice of the Hadfield steel, and it resulted in the elastic modulus of the austenitic Hadfield steel and carbide of Fe3C of 208 GPa and 216 GPa, respectively [31]. When carbides precipitated, the volume fraction of the interface in Hadfield steel increased. The increased interface volume limited the dislocation movement, resulting in the dislocation pile-up. In the aging process, the solute atoms would migrate to the center of the dislocations, and this migration would lead to a further increase in the elastic modulus [32]. Therefore, the elastic modulus of the Hadfield steel increased rapidly after the carbide precipitation. However, the carbide precipitation temperatures of the Hadfield steel under three distinct states were different. The 8
precipitation temperatures of the nanocrystallized Hadfield steel, the deformed Hadfield steel, and the quenched Hadfield steel were 371 °C, 377 °C, and 348 °C, respectively. The following section explains the reason behind these findings. After plastic deformation, the distortion of crystal lattice became engendered, and defect density increased in the Hadfield steel. Except for a small part transited to thermal energy, most of the mechanical energy during the HSP was absorbed as distortion energy, resulting in high internal energy. Compared with the quenched Hadfield steel in the low-temperature isothermal process, the distortion energy in the deformed Hadfield steel was released during heat treatment, and served as the energy for carbide precipitation. Thus, the carbide precipitation temperature relatively decreased. However, this theory was inconsistent with our results, suggesting that other factors could also influence carbide precipitation behavior. Table 1. Statistics of carbon atom distribution in a random area of 3000 nm3 Condition Three-carbon atom cluster Two-carbon atom cluster One carbon atom
The quenched Hadfield
The deformed Hadfield
The nanocrystallized Hadfield
steel
steel
steel
45.4%
42.5%
41.8%
39.8%
38.8%
37.9%
14.8%
18.7%
20.3%
To further analyze the carbon and manganese atom distribution and the interaction between them, the APT results were cut into pieces. Moreover, to avoid the overlap of congeneric atoms at the same position of different planes that could result in a misunderstanding of the atom distribution, each piece was cut with a thickness of 0.3 nm. Carbon atoms are generally considered to be located in the octahedral interstice. The distance between the opposite atoms at vertexes of the octahedron was 0.36 nm. Atoms could not overlap at the same positions on different planes. A rotating video of the APT result showing the atom distribution in an area of 30 nm × 10 nm × 0.6 nm is provided in Video Data to support the analysis. Based on that, 10 pieces of the APT results, each with an area of 40 nm × 25 nm × 0.3 nm, were statistically analyzed, and the distribution of the carbon atoms in the Hadfield steel was determined as shown in Table 1. The data in Table 1 were random with a volume of 3000 nm3. Compared with that in the quenched Hadfield steel, the fraction of three-carbon atom clusters and two-carbon atom clusters in the deformed Hadfield steel decreased by 2.9% and 1.0%, respectively, and the one-carbon atom increased by 3.9%. For the nanocrystallized Hadfield steel, the fractions of the three-carbon, two-carbon, and one-carbon atom clusters were 3.6%, 1.9%, and 5.5%, respectively. These findings showed that a redistribution of carbon atoms in the Hadfield steel occurred in the plastic deformation process. With increasing deformation extent, the number of three-carbon atom clusters and two-carbon atom clusters decreased, thus indicating that the carbon 9
atoms in the Hadfield steel were more uniform after deformation. The correlation degree between carbon and manganese atoms was not apparent, which indicated that the short-range order C-Mn couples could not be determined based on the APT results. However, according to previous reports and results obtained by other methods, Hadfield steels had a short-range order of C-Mn couples [15, 16]. Thus, the APT results demonstrated that plastic deformation enhanced the short-range order of the C-Mn couples in the Hadfield steel. Generally, carbon atoms were present in the octahedral interstice of the fcc γ-Fe in the Hadfield steel, and partial iron atoms were replaced by manganese atoms to form a Fe-Mn-C solid solution. Table 1 presents numerous three-carbon atom clusters and two-carbon atom clusters in the Hadfield steels under three distinct states. However, the octahedral interstice of γ-Fe in Hadfield steels could only accommodate one carbon atom according to the calculation results of exchange energy [33]. Thus, we inferred that the two-carbon and three-carbon atom clusters occurred at the vacancies, dislocations, or the grain boundaries. As shown in Table 1, after water quenching treatment, about 85% of carbon atoms existed in the dislocations and vacancies produced during quenching, as well as at the grain boundaries. After the HSP, a number of dislocations and vacancies were induced as shown in Table 2. The dislocation density in the quenched Hadfield steel was only 0.3 × 1014 m2, whereas those in the deformed Hadfield steel and nanocrystallized Hadfield steel were high at 9.2 × 1014 m2 and 2.2 × 1014 m2, respectively. Dislocations of such high density provided diffusion channels and existence spaces for carbon atoms. On the effect of thermal energy and deformation energy, the three-carbon and two-carbon atom clusters decomposed and diffused to newly generated dislocations and vacancies, thereby decreasing the distortion energy in the Hadfield steel. As shown in Fig. 6, Fig. 7, and Table 1, the diffusion process of carbon atoms could be explained as follows. In the quenched Hadfield steel, the three-carbon and two-carbon atom clusters were mainly observed at the grain boundaries, quenching vacancies, or in a few dislocations. When the Hadfield steel was subjected to plastic deformation, a number of vacancies, dislocations, and grain boundaries were introduced. Carbon atoms were dragged by the dislocations. Parts of the three-carbon and two-carbon atom clusters were decomposed to single carbon atoms. Therefore, the carbon atoms distributed more uniformly. Simply but, plastic deformation promoted the reaction between the carbon atoms and defects. The dislocation density decreased after aging treatment, but the dislocation density in the deformed Hadfield steel and the nanocrystallized Hadfield steel was much higher than that in the quenched Hadfield steel. According to binding energy theory of carbon atoms and dislocations [29, 34], carbide precipitation is inhibited when the binding enthalpy of carbon atom and dislocation is higher than the formation enthalpy of the carbides. When the Hadfield steel was subjected to plastic deformation, an extremely large number of dislocations and vacancies promoted the diffusion of carbon atoms to the dislocation strain field. This process promoted the homogenization of carbon atoms and delayed the precipitation of carbides in the 10
Hadfield steel. Therefore, the carbide precipitation temperature of the deformed Hadfield steel and the nanocrystallized Hadfield steel was higher than that of the quenched Hadfield steel (see Fig. 2). Table 2. Dislocation density of the Hadfield steels under three distinct states before and after aging at 400 °C ( m2) Condition
Quenched Hadfield steel 14
Before aging
0.3 × 10
After aging
0.2 × 1014
Deformed Hadfield steel
Nanocrystallized Hadfield steel
14
2.2 × 1014
3.3 × 1014
1.8 × 1014
9.2 × 10
The lattice constant of Hadfield steel does not change when carbon atoms segregate to vacancies, dislocations, and grain boundaries [35, 36]. The elastic modulus is only influenced by temperature if no carbide precipitation occurs. Subsequently, the carbon atoms around the grain boundaries segregate to the grain boundaries because of interfacial energy. When the carbon atom concentration increases to a certain degree, the carbides precipitate at the grain boundaries [37]. Notably, the energy at the grain boundaries was high with numerous defects, resulting in a faster atom diffusion rate than that in the grains. As a result, the carbide precipitation rate was higher at the grain boundaries than in the grains. Compared with the deformed Hadfield steel, more grain boundaries and much higher interfacial energy in unit volume were found in the nanocrystallized Hadfield steel. This finding illustrated the many nucleation sites and high driving force for carbide precipitation in the aging treatment. Thus, the carbide precipitation temperature of the nanocrystallized Hadfield steel was lower than that of the deformed Hadfield steel (see Fig. 2). Furthermore, the smaller the nanocrystallization was, the finer the carbides (see Fig. 6). Carbide precipitation was associated with the diffusion of carbon and alloying element atoms. Temperature determined the precipitation possibility and the precipitation rate. At a relatively low temperature (300 °C), no carbide precipitation was observed because of low diffusion capacity of the solute atoms (see Fig. 3). Therefore, the elastic modulus of the Hadfield steel under three distinct states did not change in the aging process (see Fig. 3a). By increasing the aging temperature to 400 °C, the diffusion capacity of atoms was enhanced, and the carbides precipitated gradually. In addition, the amount of carbides increased by extending the isothermal aging time, thereby continuously increasing the elastic modulus of the Hadfield steel (see Fig. 3c). However, the increase amplitude in the Hadfield steel under three distinct states was different, and the change in elastic modulus was the smallest in the nanocrystallized Hadfield steel. This result was attributed to the minimum transformation of carbides (see Fig. 5), and it further indicated that the nanocrystallization process impeded the formation of carbides in Hadfield steel. Table 3. Hardness of the Hadfield steels under three distinct states before and after aging at 400 °C (GPa) Condition
Quenched Hadfield steel
Deformed Hadfield steel
Nanocrystallized Hadfield steel
Before aging
1.68
7.43
7.52
After aging
1.85
8.43
8.85
The change regulation in the elastic modulus of the Hadfield steel was analyzed under three 11
distinct states when the aging temperature was below 350 °C (Fig. 2). Residual stress decreases the elastic modulus of metal materials [18]. After the plastic deformation and the nanocrystallization process, the hardness of the Hadfield steels increased as shown in Table 3. However, the hardness remained high even after aging treatment. This result indicated that high residual stress was induced in the Hadfield steels by the HSP, which would inevitably decrease the elastic modulus of the Hadfield steel. This result could explain why the elastic modulus of the deformed and the nanocrystallized Hadfield steel samples was lower than that of the quenched Hadfield steel when the testing temperature was below 100 °C. By increasing the testing temperature, local fine-tuning of atoms in the deformed and nanocrystallized Hadfield steel samples occurred, thus resulting in internal stress relaxation and elimination. In addition, the elastic modulus increased continuously. The amounts of the three-carbon and the two-carbon atom clusters were reduced by plastic deformation, as shown in Table 1. This process promoted the uniform distribution of carbon atoms. In other words, the short-range order of the C-Mn couples in the Hadfield steel increased. Thus, when the aging temperature ranged from 100 °C to 350 °C, the elastic modulus of the deformed and nanocrystallized Hadfield steel samples was higher than that of the quenched Hadfield steel. 5 Conclusions The following conclusions were drawn from the present work: (1) Most of the carbon atoms in the Hadfield steel were present in the defects, such as dislocations, vacancies, and grain boundaries. Plastic deformation caused a uniform distribution of the carbon atoms that increased the short-range order of the C-Mn couples in the Hadfield steel. (2) The HSP and the nanocrystallization process impeded carbide precipitation and improved the precipitation temperature in the Hadfield steel. The carbide precipitation temperatures of the quenched Hadfield steel, the deformed Hadfield steel, and the nanocrystallized Hadfield steel were 348 °C, 377 °C, and 371 °C, respectively at a heating rate of 3 °C/min. The nanocrystalline process of the Hadfield steel effectively reduced the carbide size to nano-scale, increased the carbide dispersion, and decreased the amount of carbides. (3) Carbide precipitation in the Hadfield steel remarkably increased the elastic modulus. Thus, the carbide precipitation behavior of steels could be determined by measuring the change in the elastic modulus through dynamic thermomechanical analysis.
Acknowledgement The present work was supported by the National Natural Science Youth Fund (No. 51501161).
References [1] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103-189. [2] A.P. Zhilyaev, T.G. Langdon. Prog. Mater. Sci. 53 (2008) 893-979. [3] F.C. Zhang, X.Y. Feng, Z.N. Yang, J. Kang & T.S. Wang, Sci. Rep. 5 (2015). [4] R.Z. Valiev, I.V. Alexandrov, Y.T. Zhu, T.C. Lowe, J. Mater. Res. 17 (2002) 5-8. 12
[5] K.S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mate. 51 (2003) 5743-5774. [6] K. Lu, Mater. Sci. Eng. R 16 (1996) 161-221. [7] N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K. Lu, Acta Mater. 50 (2002) 4603-4616. [8] N.R. Tao, M.L. Sui, J. Lu, K. Lu, Nanostruct. Mater. 11 (1999) 433-440. [9] J.Y. Huang, Y.T. Zhu, X.Z. Liao, R.Z. Valiev, Philos. Mag. Lett. 84 (2004) 183-190. [10] L.C. Zhang, M. Calin, F. Paturaud, C. Mickel, J. Eckert, Appl. Phys. Lett. 90 (2007) 201908. [11] Y. Ivanisenko, I. Maclaren, X. Sauvage, R.Z. Valiev, H.J. Fecht, Acta Mater. 54 (2006) 1659-1669. [12] S.M. Fatemi-Varzaneha, A. Zarei-Hanzakia, H. Paul, Mater. Charact. 87 (2014) 27-35. [13] R.A. Hadfield, Science 12 (1888) 284-286. [14] E. Bayraktar, F.A. Khalid, C. Levaillant, J. Mater. Process Technol. 147 (2004) 145-154. [15] Y.N. Dastur, W.C. Leslie, Metall. Trans. A 12 (1981) 749-759. [16] F.C. Zhang, Y.Z. Zheng, R.F. Zhu, T.Q. Lei, Scripta Met. Mater. 32 (1995) 1477-1481. [17] B.A. Latella, S.R. Humphries, Scripta Mater. 51 (2004) 635-639. [18] R. Rahemi, D.Y. Li, Scripta Mater. 99 (2015) 41-44. [19] Q. Chen, W.G. Mao, Y.C. Zhou, C.Lu, Appl. Surf. Sci. 256 (2010) 7311-7315. [20] T. Tayeh, J. Douin, S. Jouannigot, M. Zakhour, M. Nakhl, J.F. Silvain, J.L. Bobet, Mater. Sci. Eng. A 591 (2014) 1-8. [21] A.P. Stebner, D.W. Brown, L.C. Brinson, Acta Mater. 61 (2013) 1944-1956. [22] H. Zhang, B.S. Mitchell, Mater. Sci. Eng. A 270 (1999) 237-243. [23] J.M. Fiorani, F.A. Kuhnast, C. Cunat, E. Illekova, Mater. Sci. Eng. A 165 (1993) 143-148. [24] A. Yasmin, I.M.Daniel, Polymer 45 (2004) 8211-8219. [25] K.J. Green, D.R. Dean, U.K. Vaidya, E. Nyairo, Compos. Part A-Appl. S. 40 (2009) 1470-1475. [26] S. Deng, M. Hou, L. Ye, Polym. Test. 26 (2007) 803-813. [27] G. Laino, R. De Santis, A. Gloria, T. Russo, D.S. Quintanilla, A. Laino, R. Martina, L. Nicolais, L. Ambrosio, J. Biomater. Appl. 26 (2012) 1-16. [28] F.C. Zhang, Z.N. Yang, L.H. Qian, F.C. Liu, B. Lv, M. Zhang, Scripta Mater. 64 (2011) 560-563. [29] V. Gavriljuk, Scripta Mater. 46 (2002) 175-177. [30] M.K. Miller, Atom Probe Tomography: Analysis at the atomic level, Kluwer Academic/Plenum Publishers, New York, 2000. [31] L.A. Glikman, A.M. Kartashov, Z.M. Rubashkina, A.F. Lobov, Strength Mater. 7 (1975) 523-524. [32] V.N. Bastun, N.I. Chernyak, Strength Mater. 3 (1971) 1062-1065. [33] W.S. Owen and M. Grujicic, Acta mater. 47 (1999) 111-126. [34] V.G. Gavriljuk, Scripta Mater. 45 (2001) 1469-1472. [35] Y. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.J. Fecht, Acta Mater. 51 (2003) 5555-5570. [36] C. Borchers, T. Al-Kassab, S. Goto & R. Kirchheim, Mater. Sci. Eng. A 502 (2009) 131-138. [37] J. Languillaume, G. Kapclski & B. Baudelet, Acta Mater. 45 (1997) 1201-1212.
13
Video caption: "Video data of the APT results showing the atom distribution in an area of 30 nm × 10 nm × 0.6 nm in the quenched Hadfield steel".
14