Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Cryogenic study of the magnetic and thermal stability of retained austenite in nanostructured bainite

Cryogenic study of the magnetic and thermal stability of retained austenite in nanostructured... SCIENCE AND TECHNOLOGY OF ADVANCED MATERIALS 2019, VOL. 20, NO. 1, 673–687 https://doi.org/10.1080/14686996.2019.1625722 Cryogenic study of the magnetic and thermal stability of retained austenite in nanostructured bainite a a b c Arántzazu Argüelles , Florentina Barbés , Jose I. Espeso and Carlos Garcia-Mateo Dpto. de Ciencia de los Materiales e Ingeniería Metalúrgica, Edificio Departamental Este-Campus de Viesques-Universidad de Oviedo, Gijón, Spain; Dpto. CITIMAC, Universidad de Cantabria, Santander, Spain; MATERALIA Research Group, National Center for Metallurgical Research CENIM-CSIC, Madrid, Spain ABSTRACT ARTICLE HISTORY Received 27 March 2019 First magnetic characterization of a recently developed generation of carbide free bainitic Revised 27 May 2019 steels, known as Nanobain, has been performed. Stability of its retained austenite at cryo- Accepted 28 May 2019 genic temperatures has been studied by means of X-ray diffraction, microscopy, dilatometry and magnetic measurements. Two morphologies for this phase (blocky-type and film-type) KEYWORDS appear in a different proportion depending on the chemical composition and the applied Bainitic steels; austenite-to- thermal treatment. Inhibition of the martensitic transformation, when decreasing the tem- martensite phase perature down to −271°C, has been observed in those microstructures with higher proportion transformation; nanostructured metals; of film-type austenite. The paramagnetic state of austenite at room temperature seems to magnetic properties; lead to different magnetic behaviors (ferromagnetic, antiferromagnetic) at cryogenic tem- cryogenic temperature peratures (T or T being around −23°C in all the studied samples), depending on the C N proportion of such morphological features. Furthermore, irreversibility with temperature on CLASSIFICATION the evolution of such magnetic behaviors has been observed for all the studied bainitic 10 Engineering and structures and is proposed to be due to a magnetic proximity effect. Structural materials; 106 Metallic materials; 302 Crystallization / Heat treatment / Crystal growth; 500 Characterization; 503 TEM, STEM, SEM; 504 X-ray / Neutron diffraction and scattering; 600 Others 1. Introduction of mechanical properties that placed it among the most important industrial materials even at present Despite steel is a long standing material and object [1]. However, this Fe-based alloy has much more to of a plentiful research from ancient times, it offer beyond its mechanical features. Thus, it is well remains the object of cutting-edge studies nowa- known the main function of steel in rotating days. Its metallurgy has given rise to a wide range CONTACT Arántzazu Argüelles arguellesarantzazu@uniovi.es Dpto. de Ciencia de los Materiales e Ingeniería Metalúrgica, Edificio Departamental Este-Campus de Viesques-Universidad de Oviedo, Gijón 33203, Spain © 2019 The Author(s). Published by National Institute for Materials Science in partnership with Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Sci. Technol. Adv. Mater. 20 (2019) 674 A. ARGÜELLES et al. electrical machines [2] and in designing eddy- Table 1. Chemical composition of the analyzed alloys (wt.%). Material C Si Mn Cr Cu Ni Mo current and magnetic rail brakes for high-speed Steel1 0.66 1.45 1.35 1.02 - 0.1 0.24 trains [3,4] relies on its electric and magnetic prop- Steel2 0.98 2.90 0.77 0.45 0.21 0.16 - erties. Also recently, semimagnetic stainless steel Steel3 0.99 2.47 0.74 0.97 0.17 0.12 0.028 has been tested and proposed as a suitable option in the design of high speed rotor permanent mag- net machines [5] because it meets both the mag- cementite precipitation during the bainitic transfor- netic and the mechanical requirements, with a high mation [8]. Those alloys were selected for being able yield strength. In this sense, the magnetic charac- to transform into nanostructured bainite when iso- terization of new developed steels is interesting by thermally transformed at low temperatures, see for itself, as can provide useful information for poten- example refs [6,9–12]. tial applications. Through this work, different types of heat treatments In this work and for the first time, the new gen- were applied, whose purpose will become clear later eration of steels, known as Nanobain, has been stu- when presenting the experimental results. In order to died from a magnetic point of view [6]. Its produce the desired bainitic structure, isothermal treat- microstructure, consisting of bainitic ferrite plates, ments after full austenitization were applied in condi- with thickness around tenths of nanometers, alternat- tions adapted to the different chemical compositions ing with carbon enriched retained austenite films, (treatment B in Figure 1). Selected bainitic microstruc- achieves exceptional mechanical properties (strengths tures were cooled down from room temperature (RT) in the range of 1.6–2.5 GPa and toughness of around to ~ −123°C and back to RT (treatment BC). Finally, pffiffiffiffi 30 MPa m [7]) comparable to that of maraging with the purpose of generating a fully ferritic micro- steels, but with a lower production cost. The goal of structure, samples of two of the alloys were quenched, the present work has been to get the first magnetic creating a microstructure composed of martensite (α´) characterization of Nanobain steels, and more speci- and retained austenite (γ), followed by a tempering fically, to find correlations between magnetic and treatment that allowed for the full decomposition of microstructural features, which is hoped to help in austenite into ferrite and scarce precipitates, (treatment designing the metallurgical scheme oriented to attain QT in Figure 1). a particular magnetic behavior together with specific Specifics of the different parameters used in the mechanical properties. heat treatments are gathered in Table 2. It has to be highlighted that the austenitization and isothermal temperatures and times, as well as the cooling rates, 2. Materials and experimental methods were adapted to the different alloys according to previous experiences [6,9–12]. In order to easily iden- The chemical composition of the steels used in this tify the samples, the following labelling procedure has work is listed in Table 1. All the alloys are high been used: a number identifying the steel category (as C steels with enough amount of Si (≥1.5%) to avoid Figure 1. Heat treatment schemes; treatment B (Bainite) for obtaining the bainitic microstructures, treatment BC (Bainite +Cryogenic) cooling down after having obtained the bainitic microstructure, treatment QT (Quench+Temper) in order to obtain an austenite-free microstructure. Sci. Technol. Adv. Mater. 20 (2019) 675 A. ARGÜELLES et al. Table 2. Heat treatments parameters: T and t are the ratio. XRD data were collected during 2 hours over γ γ austenitization temperature and time respectively; T and iso a2θ range 35 − 135°, with a step size of 0.01°. In this t are the temperature and time respectively, for the forma- iso study, the Rietveld analysis program TOPAS 4.2 tion of bainite by an isothermal treatment; T and t are cryo cryo (Bruker AXS) was used for quantification and calcu- the temperature and time, respectively, for the cryogenic lation of the structural parameters of both, the treatment. retained austenite and the bainitic ferrite. Line broad- SAMPLE T (°C) t (min) T (°C) t (h) T (°C) t (min) γ γ iso iso cryo cryo ening effects due to the lattice microstrains were 1B220+ 900 5 220 168 - - 1B220 900 5 220 24 - - analyzed with the double-Voigt approach [14]. In 1B250 900 5 250 8 - - order to eliminate the instrumental contribution to 1B300 900 5 300 5 - - 1B350 900 5 350 4.5 - - peak broadening, instrument functions were empiri- 1QT 900 5 - - - - cally parameterized from the profile shape analysis of 1BC220+ 900 5 220 168 −123 5 1BC350 900 5 350 4.5 −123 5 a corundum sample measured under the same con- 2B200 950 15 200 40 - - ditions. The austenite carbon content was estimated 2B350 950 15 350 8 - - 3B220 1050 5 220 40 - - using the well-known expression in ref [15], that 3B250 1050 5 250 25 - - relates the influence of different alloying elements to 3B350 1050 5 350 7 - - the lattice parameter. 3QT 1050 5 - - - - 3BC350 1050 5 350 7 −123 5 XRD sample preparation was performed using standard metallographic procedures but introducing a final set of cycles of etching and polishing in order shown in Table 1), the acronym corresponding to the to remove the surface layer that has been plastically thermal treatment and the temperature of the iso- deformed during the grinding step. That surface layer thermal treatment; the plus symbol in one of the may contain traces of martensite formed by transfor- experiments indicates extra time at the isothermal mation-induced plasticity (TRIP) of austenite due to transformation temperature. sample preparation, which would underestimate its While microstructures obtained with a heat treatment real fraction. of thetypeBwere allproducedinthe frameworkof The microstructure was revealed after a standard industrial collaborations according to the procedures metallographic preparation, followed by a final etch- described in [13], heat treatments denoted as BC and ing with a 2% nital solution. A JEOL JSM-6500 field QT were performed in a Bahr 805D high-resolution emission gun scanning electron microscope (FEG- dilatometer (TA Instruments, USA) equipped with an SEM; JEOL, Japan) operating at 10 kV was used to induction heating coil. Helium was used as quenching observe the microstructure. gas and the temperature was controlled by a K type Bainitic ferrite plate thickness (t ) was determined thermocouple welded to the central part of the sample on SEM micrographs by measuring the shortest dis- surface. For the cryogenic treatment, He flowing through tance perpendicular to the longitudinal dimension of a cooling coil immersed in liquid N was used instead. the ferrite plate, and correcting for stereological Thelengthchangeassociatedwiththedifferent metal- effects as described in reference [16]. lurgical events taking place during the heat treatments The magnetic measurements were made with were recorded by a linear variable displacement transdu- a Quantum Design Physical Property Measurement cer (LVDT) with a resolution of 0.05 μm. The mentioned System (PPMS; Quantum Design, USA), equipped with treatments were performed using fused silica push-rods a 9T superconducting magnet and able to cover to measure longitudinal changes in length, given the a temperature range from −271 to 77°C (2 to 350 K, small expansion coefficient of quartz (push rods), approximately). Saturation magnetization values were −6 −1 0.5 × 10 °C , when compared with the expansion obtained from the hysteresis loop at RT of every sample −6 −1 coefficient of steel, approximately 10 × 10 °C ;itis resulting from treatments B and QT (m 1). Subsequently, safe to conclude that the contribution of the push rods a thermomagnetic analysis was performed between RT to the measured change in length is negligible. and −271°C (2 K) first cooling and then heating. After Quantitative X-ray diffraction (XRD) analysis was this analysis, a second value for the saturation magneti- used to determine retained austenite and bainitic zation (m 2) was obtained from the hysteresis loop at RT ferrite volume fractions. Samples were step-scanned measured for each sample. in a Bruker AXS D8 X-ray diffractometer (BRUKER AXS, USA) with a rotating Co anode X-ray tube as a radiation source, Göebel mirror optics and 3. Results and discussion a LynxEye Linear Position Sensitive Detector for 3.1. Nanostructured bainite: microstructural ultra-fast XRD measurements. A current of 30 mA characterization at room temperature and a voltage of 40 kV were employed as tube set- tings. Operational conditions were selected to obtain For all the steels under investigation, the microstruc- X-ray diffraction data with a high signal to noise ture after the B treatment consists of bainitic ferrite Sci. Technol. Adv. Mater. 20 (2019) 676 A. ARGÜELLES et al. (α ) and retained austenite (γ). As reported in pre- The bainite transformation is a displacive and difus- vious works, the microstructure is essentially carbide sionless reaction, in which the ferrite is initially super- free and higher magnification techniques, such as saturated with respect to carbon. The carbon excess in transmission electron microscopy and atom probe the bainitic ferrite is subsequently and rapidly parti- tomography, only revealed scarce quantities of tioned into the residual austenite, but substitutional cementite precipitates [17,18]. Examples of the elements do not partition during the bainite reaction microstructures at selected temperatures can be [8]. In the absence of carbide precipitation, prevented in found in Figure 2, corresponding to Steel1 after iso- the present steels by the use of silicon, the austenite thermal treatment at 350°C (sample 1B350) (Figure 2 carbon enrichment is such, at all transformation tem- (a)) and at 220°C (sample 1B220+) (Figure 2(b)), peratures, that during the quenching to room tempera- respectively, where both phases have been identified. ture no martensite forms, i.e. the M (martensite start The darker long slender features are the plates of temperature) of C-enriched retained austenite is well bainitic ferrite (α ), and the lighter phase found as below room temperature. Table 3 summarizes the films and more blocky type features correspond to results of the measured C content in austenite (C ) retained austenite (γ). XRD results, presented in where it is clear that the level of C enrichment is Table 3, corroborate that only these two phases are above of that of the bulk content. present, being the bainitic ferrite the predominant Measurements of the bainitic ferrite plate thick- and the austenite the minor one. The maximum ness, t (see Table 3), indicate a coarsening of the αb extent of transformation increases as transformation ferritic matrix as the transformation T increases (see temperature decreases, i.e. bainitic ferrite fraction Figure 2). These results are in agreement to what is (V ) decreases as isothermal temperature increases, expected and already extensively reported, see for αb which is expected from the incomplete reaction phe- instance refs [16,19,20]. In this kind of microstruc- nomena ruling the bainitic transformation [8]. Note tures, austenite exhibits a well known multi-scale that the time used in the isothermal treatment of the character associated with two distinguishable 3B220 sample did not allow a full bainitic transfor- morphologies, i.e., thin films (γ ) between platelets mation, so that more bainite could have been formed of bainitic ferrite and blocks between sheaves of bai- in a longer time, i.e. the final amount of bainite is nite (γ ), see for example ref [9,10,21–23]. Those smaller than that contained in the 3B250 sample. have identified in Figure 2, i.e. in the main Figure 2. SEM micrographs showing examples of the bainitic microstructures of (a) 1B350, and (b) 1B220+, where bainitic ferrite (α ) and the two retained austenite morphologies, blocky (γ ) type and thin films (γ ) have been identified. b B f Sci. Technol. Adv. Mater. 20 (2019) 677 A. ARGÜELLES et al. micrograph the blocky-type, and the film-type, due to The largest fraction of film type retained austenite their much smaller size, are identified in the higher is obtained for the sample 1B220+, while samples magnification inset. The two different morphologies treated at the highest austempering temperature are well known for also having very different carbon have the largest fraction of blocky type austenite contents in solid solution, i.e. the thin films are far (1B350, 2B350, 3B350). These results are consistent richer in carbon than the blocks [23–29]. with what is observed in the micrographs shown in There are two important effects on the microstruc- Figure 2, and they are expected from the incomplete ture as a consequence of decreasing the transformation reaction phenomena [8]. temperature. On the one hand, as the fraction of bainitic The displacive character of the bainitic transforma- ferrite increases, the fraction of thin films of austenite tion implies that there is a shape change along the (V ) increases at the expense of that of blocky austenite transformation. The plastic relaxation of this shape γf (V ). An estimate of the ratio between both morphol- change commonly takes place via generation of both, γB ogies can be done following the procedure described in dislocations in the austenite/bainitic ferrite interface, ref [30], where it is assumed that about 15% of the and also micro/nano-twins in the austenite, in contact volume contained within the boundaries of a bainite with bainitic ferrite plates [31–33]. As the transforma- sheaf consists of retained austenite films interspersed tion temperature decreases, there is an increase in the with bainite plates. Results obtained in this way are dislocation density caused by the yielding of the auste- summarized in Figure 3, presented as a function of V . nite. If we consider that microstrain is directly related αb to the dislocation density [34–36](see Table 3)it is clear that indeed, the lower is the transformation tem- perature, the higher is the dislocation density intro- duced in the microstructure. Note that the differences in microstrain between 1B220+ and 1B220 correspond to a restoration process due to the extended heat treat- ment [37]. 3.2. Nanostructured bainite: a preliminary magnetic characterization at room temperature (M versus H) In order to get an initial insight into the magnetic beha- vior of these bainitic microstructures, their hysteresis loops at room temperature (RT) have been measured. An example of the curves thus obtained is pre- sented in Figure 4 for the 1B220 and 3B220 micro- structures. From such curves, the value of the Figure 3. Relationship between the V =V ratio (thin film/ γf γB coercive field (H ) is obtained, and gathered in blocky type austenite) and the amount of bainitic ferrite (V ) αb Table 4. The obtained values range between 29–67 present in the studied microstructures. The line is just a guide Oe, close to the coercitivities corresponding to other for the eye. Table 3. Chemical and microstructural parameters of the studied samples. V ,V ,V are the volume fractions obtained by γXRD αb α’ XRD for austenite, bainitic ferrite and martensite, respectively. V is the volume fraction of austenite obtained by magnetic γmagn measurements. ε is the microstrain and t is the bainitic ferrite plate thickness. Those parameters for samples obtained after γ αb treatment B have been previously reported [6,9–11]. Sample V (%)±3% V (%)±3% C (wt.%) ±0.05 V (%)±3% V (%) ε (%) t (nm) γXRD αb γ α´ γmagn γ αb 1B220+ 14 86 0.89 - 11 0.0041 38 ± 6 1B220 17 83 0.71 - 17 0.0036 42 ± 7 1B250 25 75 0.81 - 19 0.0032 42 ± 11 1B300 27 72 1.25 - 33 0.0027 61 ± 10 1B350 47 53 1.12 - 44 0.0018 79 ± 14 1QT 0 100 1BC220+ 12 88 - 0.0029 36 ± 4 1BC350 29 62 9 0.0020 - 2B200 29 71 1.19 - 28 0.0033 40 ± 11 2B350 57 43 1.34 - 54 0.0017 155 ± 30 3B220 41 59 1.21 - 31 0.0029 34 ± 8 3B250 23 77 1.43 - 26 0.0027 43 ± 10 3B350 45 54 1.32 - 55 0.0016 74 ± 13 3QT 0 96a 3BC350 36 53 10.7 0.0023 - The 3QT sample contains 4% volume fraction of cementite. Sci. Technol. Adv. Mater. 20 (2019) 678 A. ARGÜELLES et al. steel grades [38–40], and closer to those values of soft a) magnetic materials than to those of hard magnetic ones, as is well known Hc is typically less than 10 Oe for the first ones and more than 100 for the second ones [41]. The initial magnetization curves, at RT, are pre- sented in Figure 5 for all the studied samples. It is well known that ferromagnetic materials follow the so- called law of approach to saturation given by Equation 1QT (2), where m is the saturation magnetization and the 1B220+ 1B220 parameters a (in Oe) and b (in Oe ) are positive and 1B250 1B300 related to crystallographic anisotropy and internal 1B350 strain [42]. As previously reported by other authors, Fits the saturation magnetization value in steel grades can be obtained from a fit of the experimental data to b) Equation (2) [43,44]. Such these fits, for an applied field H in the range 7000 Oe < H < 20,000 Oe, are 1B220 3QT 3B220 2B200 2B350 3B220 3B250 3B350 Fits 0 5 10 15 20 25 magnetic field (kOe) -50 Figure 5. Measured magnetization as a function of the applied magnetic field, at RT, for (a) samples 1B, (b) samples -100 -2 Hc 2B and 3B. The lines correspond to fits with Equation (2). -4 -150 -0.2 -0.1 0 0.1 0.2 presented in Figure 5.Theyresult in the m values -200 -25 -20 01 5 0 15 20 25 -15 -10 -5 denoted as m 1in Table 4. magnetic field (kOe) a b m ¼ m 1   (2) Figure 4. Measured magnetization as a function of the applied magnetic field (hysteresis loops) of the 1B220 and All the values obtained for the saturation magnetiza- 3B220 samples at RT. The inset shows the detail around 0 Oe tion are lower than that of pure Fe (217 emu/g) [45] in order to better appreciate the coercive field, H c. and close to values reported for other steel grades [43,46–49]. Differences between them can be attribu- Table 4. Magnetic parameters obtained for the studied sam- ted to the chemical composition and microstructural ples: coercitivity (H ), saturation magnetization for the baini- differences. It can be seen, as expected from references tic microstructures (m 1), and that of the same microstructures after undergoing a thermal cycle cooling [43,44,46,48,50,51], that for the same composition down to −271°C and back to room temperature (m 2). s there is a strong correlation between the amount of Sample H (Oe) m 1(emu/g) m 2(emu/g) c s s ferritic phase and the m values, i.e. the highest values 1B220+ 29 179.6 179.9 corresponding to the QT samples, which are almost 1B220 37 166.5 167.9 100% ferritic (Tables 3 and 4), are consistent with the 1B250 42 162.4 164.4 1B300 60 135.6 141.3 absence of austenite (paramagnetic at RT) in the 1B350 40 113.2 146.8 microstructure. Thus, a continuous decrease of m 1QT 191.1 2B200 57 137.8 139.8 value with increasing austempering temperature 2B350 38 86.9 141.2 (T ) is observed for the three analyzed steel grades. iso 3B220 67 131.9 135.7 3B250 67 140.9 144.4 The mentioned strong correlation between the m 3B350 65 86.1 143.2 values and the fraction of ferritic (ferromagnetic) phase 3QT 186.7 (189.13a) a has been exploited as a way for obtaining an estimate of Value corrected after considering the contribution of the cementite present in the microstructure, as explained in the text. the paramagnetic phase (austenite) present in the magnetization (emu/g) magnetization (emu/g) magnetization (emu/g) Sci. Technol. Adv. Mater. 20 (2019) 679 A. ARGÜELLES et al. microstructure of steels (see, for example, a behavior in the M(T) curve such as that of 1B220+ [43,44,46,48,50,51]).Thus, thevolumefractionofthe (see Figure 8) with a sharp increase in their magnetiza- retained austenite can be determined from the saturation tion when decreasing T down from ~ −23°C. This magnetization value according to Equation (3) [43,44], increase constitutes around 3% of the total magnetization value at this temperature range, what means a small variation respect to the total magnetization value of the V ¼ 100 1  (3) γmagn m ðÞ i s ferromagnetic iron phase (see inset in Figure 8). The significance of this transition will be clear later in the text. where m is the saturation magnetization of each sample On the contrary, magnetization decreases abruptly and m (i) is the intrinsic saturation magnetization, cor- when decreasing T to values lower than that value of ~ responding to an austenite-free microstructure with the −23°C in samples with a bigger amount of retained same chemical composition. The sample 3QT contains, austenite, mainly of blocky-type, such as 1B350 and besides ferrite, also a 4% volume fraction of cementite, so 3B350 (see Figure 9(a,b),respectively).Finally,athird in order to correct the experimental m value (~187 emu/ behavior, intermediate between those both already g) a lever rule was applied considering that cementite mentioned, is exhibited by other samples, such as saturation magnetization is ~129 emu/g [52,53]. The 3B250 (see Figure 9(c)) or 1B250, with a ‘cusp’ centered resulting corrected value for m is ~190 emu/g, which is around the same temperature (~ −23°C). Such samples close to that obtained in sample 1QT. contain a similar amount of both film-type and blocky- Furthermore, by plotting the measured volume frac- type austenite (see Figure 3). tion of bainitic ferrite, V , presented in Table 3, against αb Furthermore, it is worth to highlight that, for all the the specific saturation magnetization, m 1, reported in studied samples, irreversibility is observed for the mag- Table 4, it is possible to perform a linear fit, forcing it to netization evolution when returning to RT after cooling pass through zero, as m is supposed to be zero for (see Figures 8 and 9), which will be discussed later. a fully austenitic microstructure. From these fits, the intrinsic m (i) yields a value of ~201 emu/g for Steel1 (see Figure 6(a)), and a value of ~191 emu/g for Steel2 3.4. Stability of retained austenite in the and Steel3 (see Figure 6(b)). Such obtained values show cryogenic regime a reasonable level of agreement with those for the cor- responding QT samples (see Table 4). Despite small As thermomagnetic curves suggest changes of either chemical composition differences between Steel2 and structural or magnetic nature, a second hysteresis loop Steel3, both steel grades have been treated together as it has been measured at RT, for every sample, after cool- was checked that independent fits lead to similar values ing down to −271°C. This second measurement of the for m (i). For the sake of clarity, a comparison of the hysteresis loops let us check if there are differences in austenite content obtained from both methods, XRD the new saturation magnetization value (m 2) respect to and magnetic measurements (considering m (i) values that obtained from the virgin microstructure (m 1). from fits), is also presented in Figure 7. It can be seen These results are presented in Table 4 and, for discus- that the values obtained by both methods are in good sion purposes, are plotted in Figure 10. agreement, differing for most of the samples in no more The higher value of m 2 respect to m 1 for the s s than 6%, according with results reported by other samples 1B350, 2B350 and 3B350 is indicative of authors in different steel grades [44,50]. a higher fraction of a ferromagnetic phase, i.e. partial transformation from austenite (paramagnetic at RT) to martensite (ferromagnetic at RT), which is very 3.3. Thermomagnetic measurements (M versus T) likely to occur when cooling down to −271°C [46,54– 56]. On the contrary, in those samples where the m 1 In order to obtain further information on the auste- s and m 2 values are similar (see for example 1B220+), nite present in these samples, thermomagnetic curves s the occurrence of such transformation is unlikely and have been measured for each one of them, decreasing the microstructure is expected to be stable during the the temperature down to −271°C and increasing it up whole thermal cycle. A more direct way to visualize again to 27°C, while a small and constant magnetic this is by means of Equation (3), from which V field of 50 Oe was applied. Taking into account that γmagn (1) was calculated. Using this expression, new V ferrite gets ferromagnetically ordered below ~770°C, γmagn (2) values have been obtained from m 2 (see Table 4) and cementite below ~207°C, it can be assumed that s and they are plotted versus the previous V (1) any anomaly at lower temperatures has to be related γmagn ones in Figure 11. According to these results, the to either the austenite magnetic behavior or to struc- microstructures with the lowest V /V ratio, i.e. tural transformations from such phase. γf γB with higher fractions of blocky austenite (see Figure Some of the studied samples, with a high fraction of 3), as those transformed at 350°C, exhibit the highest bainitic ferrite, and therefore a high fraction of thin films degree of austenite transformation during the of austeniteoverthe blocky-type (Figure 3), show Sci. Technol. Adv. Mater. 20 (2019) 680 A. ARGÜELLES et al. Figure 7. Austenite volume fraction obtained from two dif- ferent methods: magnetic measurements (V ) and XRD γmagn (V ). The error associated to V and V is ±3% and γXRD γXRD γmagn ± 0.2% respectively. 1.45 1B220+ 1.44 H=50Oe 1.43 cooling 1.42 warming 2.0 1.41 1.5 1.40 1.0 -23ºC Figure 6. Saturation magnetization values, m 1, versus ferrite s 1.39 fraction (V ) for (a) Steel1; (b) Steel2 and Steel3. The lines 0.5 αb correspond to linear fits passing through the origin. The error 1.38 associated to m 1 is about 0.15%, and for V is ±3%. s αb -250 -150 -50 0 50 T(ºC) 1.37 -300 -250 -200 -150 -100 -50 0 50 T(ºC) cryogenic treatment. Note that these samples are the ones that deviate from the perfect correlation line. Figure 8. Thermomagnetic curve for 1B220+, measured dur- Before going deeper in the discussion, it is essen- ing cooling from RT down to −271°C and heating back to RT. The inset shows the same curve at a different scale to realize tial to comment on the parameters that may affect the about the relative change in the magnetization. thermal stability of austenite against martensitic transformation. composition, and elements such as C, Mn, Si and Al Martensitic transformation is displacive, involving [22,58,59] significantly enhance the austenite stability; the coordinated movement of atoms; and since the among them C is the element that exhibits the stron- glissile transformation interface has a dislocation gest influence. As already mentioned, the two differ- structure that has to move through any obstacles ent morphologies of retained austenite are strongly that exist in the austenite, martensitic transformation linked to different carbon contents, i.e. the thin films cannot be sustained against strong defects such as are far richer in carbon than the blocks. Since carbon grain boundaries. Less drastic defects, such as dislo- is an interstitial solute, such differences also imply big cations, also hinder the progress of such transforma- differences in their thermal stability [60]. tions, but can often be incorporated into the The differences in size of the austenitic features also martensite lattice [57]. have an important effect per-se on its stability against The resistance of austenite against the martensitic martensitic transformation [61–64]. This is in part due transformation partially depends on its chemical magnetization (emu/g) magnetization (emu/g) Sci. Technol. Adv. Mater. 20 (2019) 681 A. ARGÜELLES et al. 1.52 a) 1B350 -23ºC 1.50 H=50Oe 1.48 cooling 1.46 warming 1.44 1.42 m 1 1.40 m 2 -23ºC b) 3B350 H=50Oe 2.3 -40ºC 2.2 cooling Figure 10. Comparison of the saturation magnetization warming values before (m 1) and after (m 2) performing the thermo- s s 2.1 magnetic measurement, i.e. cooling down to −271°C. The error associated to m is about 0.15%. 2.0 1.9 -23ºC c) 3B250 1.85 H=50Oe 1.84 1.83 cooling warming 1.82 1.81 1.80 1.79 -300 -250 -200 -150 -100 -50 0 50 T(ºC) Figure 9. Thermomagnetic curves for (a) 1B350 sample; (b) 3B350 sample; (c) 3B250, measured during cooling from RT down to −271°C and heating back to RT. Figure 11. Correlation between the volume fraction of aus- to the fact that small retained austenite islands contain tenite (V ) obtained from m 1, and m 2. The associated γmagn s s lower potential nucleation sites for the transformation error is about 0.2%. to martensite and, consequently, require a greater total driving force for the nucleation of martensite [65]. Therefore, the smallest austenitic features are more other words a stronger matrix may prevent the mar- stable than the largest ones because of both their size/ tensitic transformation. morphology and their carbon content. Thus, in this work it is proposed a microstructural Then there is the influence that the strength of the stability parameter (msp) that accounts for almost all surrounding matrix, bainitic ferrite, has on the stabi- the mentioned factors influencing the stability of aus- lity of the retained austenite. Theoretically, the refine- tenite,i.e.carbon content (C ), dislocation density ment of the bainitic ferrite plates with different (microstrain ε ), morphology (the ratio r = V /V ) γ γf γB crystallographic orientations increases the stability and strength of the matrix (plate thickness t ). Such αb of the retained austenite. If the latter is closely sur- parameter is therefore defined as being proportional to rounded by the relatively rigid and refined bainitic msp , C rε =t (4) γ γ α ferrite, the stability of the retained austenite also increases due to the geometrical restrictions imposed The exact values of the different parameters can be by the surrounding bainitic ferrite laths [66,67]; in found in Table 3 and Figure 3. The higher the value magnetization (emu/g) magnetization (emu/g) magnetization (emu/g) m (emu/g) 1B220+ 1B220 1B250 1B300 1B350 2B200 2B350 3B220 3B250 3B350 Sci. Technol. Adv. Mater. 20 (2019) 682 A. ARGÜELLES et al. of this product is, the more stable is the austenite sample upon cooling from RT to −123°C, as there is within the initial bainitic microstructure (before the no deviation from the expected linear contraction. cryogenic treatment) against martensitic transforma- On the other hand, 350°C bainitic microstructures tion. This fact is strongly correlated with the variation exhibit deviations from linearity on cooling, indicat- of m when cooling sample down to −271°C, as it is ing the existence of a phase transformation: the shown in Figure 12. decomposition of austenite into martensite leads to In this sense, and from the magnetic point of view, a net expansion [68,69]. From these curves, the an equivalent stability parameter could be defined as beginning of the martensitic transformation has Δm =|m 1-m 2|, since, as already discussed, a zero been determined to take place at T = −7°C (M ) and s s s s or very small value for this parameter is indicative of T= −40°C (M ) for 1BC350 and 3BC350, respec- the austenite stability (paramagnetic at RT) respect to tively. Note that expansion is larger for 3BC350 its transformation in martensite (ferromagnetic at than for 1BC350, what agrees with an also higher RT) when undergoing the cryogenic treatment. difference between V (m 1) and V (m 2), as observed γ s γ s To further corroborate these results and the pre- in Figure 11. sent discussion, and by means of dilatometry, The transformation detected from dilatometry is a thermal cycle in the cryogenic regime (BC treat- barely reflected in the M(T) curve of the 3B350 ment in Figure 1) has been applied in three different sample (see Figure 9(b)), where a small anomaly at microstructures: one where the austenite is consid- around −40°C might indicate the start of the mar- ered as very stable (1B220+ sample), and other two tensitic transformation. As the austenite decomposi- where, according to the data so far presented, some tion into martensite is expected to evolve degree of austenite transformation has occurred, i.e. progressively in a wide temperature range, it is not 1B350 and 3B350 microstructures. The new samples strange that such subtle and progressive transforma- were codified as 1BC220+, 1BC350 and 3BC350. Due tion might not leave more evident traces on the to the equipment technical specifications, the mini- mentioned magnetic curves (Figure 9(a,b)). Note mum temperature is −123°C (the exact parameters of that for the 1B350 sample, the martensitic transfor- the thermal cycle are shown in Table 2). This range is mation starts at higher temperatures (M = −7°C) considered suitable, as the M(T) anomalies that could than that of the anomaly observed in the M(T) curve be related with phase transformations appear at much (−23°C) (see Figure 8(a)). higher temperatures (−23°C), as reported in Figures 8 Finally, in order to provide further additional and 9. experimental evidence, XRD analysis was performed The dilatometry measurements are shown in on the three samples that were measured in the Figure 13. There is no evidence of martensitic trans- dilatometer, to directly check the degree of transfor- formation from retained austenite in the 1BC220+ mation. These results are gathered in Table 3 and they agree with the previous discussion; i.e., the microstructures of 1B220+ and 1BC220+ are almost identical in terms of V , which means that no mar- tensitic transformation took place, while in the case Figure 12. Microstructural and magnetic stability parameters, msp and Δm respectively, as defined in the main body of the text. Note that, for the sake of clarity, the microstructural Figure 13. Dilatometric measurements on cooling from RT parameter has been represented in log scale. The solid line is down to −123°C for samples 1BC-220+, 1BC-350 and 3BC- just a guide for the eye. 350. Δl/lo represents the relative change in length. Sci. Technol. Adv. Mater. 20 (2019) 683 A. ARGÜELLES et al. of 1BC350 and 3BC350, 9.5 and 10.7 of fresh mar- where an increase should be expected as tensite (V ), respectively, was formed on cooling. a consequence of the appearance of a ferromagnetic α’ phase (fresh martensite α´) and the FM order of the thin films of austenite (see Figure 9(a,b)). 3.5. Magnetic behavior versus γ→α’ The decrease in M(T) is rationalized in terms of transformation the antiferromagnetism of the untransformed blocky austenite, which overcomes the expected increase due Once it has been established that for 1B220+ there is to the presence of α´ and thin films. In such blocky not martensitic transformation down to −123°C, austenite, much poorer in C than the thin films [23– there must be another effect that can explain the 29] the AF order is energetically favored between detected M(T) increase, below −23°C, when cooling those Fe atoms of the austenite cell with less the sample down to the cryogenic regime (see C neighbors [72,75–77]. In the present case, the Figure 8). establishment of such AF order is also found to be It is known that the magnetic state of γ (fcc), in Fe around T = −23°C (Néel temperature). Therefore, and Fe-based alloys, is strongly frustrated at cryo- the decrease of the M(T) is the net result of the genic temperatures, which leads to the establishment positive FM contribution of α´ and thin films, and of different magnetic structures, such as ferromag- the negative contribution of the AF blocky austenite. netic (FM), antiferromagnetic (AF), double layer Such overlapping of counteracting effects is evident antiferromagnetic (AFMD), Spin Spiral (SS) . . ., all when considering the smoother decrease of M(T) in of them with close energy values [70–73]. 1B350, where M ~ −7°C > T , as compared to that of At T = −23°C, the ferromagnetic bainitic ferrite s N 3B350, where M ~ −40°C < T (Figure 9(a,b)). matrix is already ordered, as its Curie temperature is s N In these particular cases, 1B350 and 3B350, the quite far above, ~770°C. As the inset in Figure 8 shows, measured microstrain before and after the cryogenic the change associated with the proposed transition cycle (RT→ −123°C → RT) at XRD shows an below −23°C it only represents a small, yet detectable, increase from 0.0018 and 0.0016 to 0.0020 and variation of M(T) as compared to that of the initial 0.0023 (%), respectively. Such increase would be con- microstructure, composed of 86% well ordered ferro- sistent with the formation of martensite on cooling, magnetic ferrite. In other words, such change can only and the concomitant introduction of dislocations due be attributed to a secondary phase, i.e., austenite. to its displacive character. In these samples, the An interesting observation is that the measured microstrain does not seem to change because of the austenite microstrain, before and after the dilato- establishment of the magnetic order (AF). metric cryogenic cycle at ~ −123°C, decreases signifi- An intermediate scenario, where the calculated cantly, from 0.0041 down to 0.0029 (%) (see Table 3). ratio between thin films and blocks of austenite is This fact points out a microstrain relief associated around 1, and no martensitic transformation was with the FM ordering of the austenite, as theoretically detected on cooling down to −123°C, corresponds predicted by Okatov et al. [74]. On the other hand, to the sample denoted as 3B250 (see Table 3 and Boukhvalov et al. [72], by means of ab initio electro- Figure 3). The thermal variation of this sample, pre- nic structure calculations, have shown that C, as an sented in Figure 9(c), shows a pronounced cusp cen- interstitial present in an Fe fcc lattice (austenite), tered around −23°C. This scenario reflects again the stabilized a FM state for the Fe atoms at the low competition between the FM arising from the bainite temperature ground state (−273°C), making its and thin films of austenite (prevailing at T > −23°C), energy to be lower than that of the antiferromagnetic and the AF of the blocks (prevailing at T < −23°C). (AF) state. Therefore, for the particular case of the It is clear now that the irreversibility observed in 1B220+ microstructure, the described microstructural the M(T) curves for all the analyzed samples can not features defining the retained austenite (mainly as be attributed to a martensitic transformation of the thin C-enriched films) seem to lead to its thermal retained austenite, opposite to what occurs in a Cr- stability when going to temperatures such as low as alloyed high C steel cooled down to −200°C, as −123°C, at the same time that a FM state for the Fe reported by Tavares et al. [46]. Following, this feature atoms is established. will be explained in terms of magnetism. Thus, it is proposed that the mentioned anomaly at There is a well known magnetic proximity effect that −23°C, in Figure 8, is the result of a ferromagnetic transi- may propagate the spin polarization of a magnetic tion of austenite, i.e. Curie temperature, T ~ −23°C. metal into a nonmagnetic one, as first reported by On the other hand, microstructures such as 1B350 Hauser in 1969 [78]. Since then, several works have and 3B350, with a higher proportion of austenite with used this effect to explain induced magnetic behaviors. block morphology, and martensitic transformations Unguris et al. [79] reported a spin-density wave (SDW) taking place at −7°C and −40°C, respectively, due to antiferromagnetism in Cr films deposited on an Fe their lower thermal stability, exhibit a sharp M(T) substrate, at temperatures above the T of bulk Cr, decrease as the temperature decreases below −23°C, N Sci. Technol. Adv. Mater. 20 (2019) 684 A. ARGÜELLES et al. due to the coupling of Cr and Fe at some distance from those microstructures where the morphology of the interface. Maccherozzi et al. [80] reported retained austenite is predominantly thin-film, a magnetic coupling between Mn and Fe in (Ga, Mn) martensitic transformation did not occur when As/Fe interfaces that extends over a region as thick as cooling down to −123°C; other microstructural 2 nm. Kravets et al. [81] investigated the interlayer factors such as microstrain, scale of the matrix exchange coupling in a ferromagnetic Ni-Cu trilayer or austenite carbon content have been found to system (strong/weak/strong) and explained the para- also play an important role. A microstructural magnetic-to-ferromagnetic transition of the spacer stability parameter (msp) that accounts for material by the magnetic proximity effect. almost all the mentioned factors influencing Furthermore, Zhang et al. [82] reported the application the stability of austenite is proposed. of this effect to their ab initio simulations of structural ● Thermomagnetic measurements have brought transformations between austenite and ferrite in Fe-C significant differences on the magnetic behavior alloys. And such a strongly coupling between both of the studied samples to light. Thus, phases supported also the work of Razumov et al. [83] a ferromagnetic order is detected for microstruc- about phase transformations in steels. In the same line, tures with a higher thin film/blocky ratio (pre- the actual work proposes that this magnetic proximity valence of film type austenite), while an effect is at the origin of the low-T magnetic order hold antiferromagnetic order is exhibited by those in austenite when heating to RT and, thus, explains the with lower film/blocky ratio (prevalence of observed irreversibility in the thermal evolution of the blocky type austenite). In fact, a competition of magnetization. Zhang et al. [82]reported thatthe both magnetic behaviors has been found for all exchange parameter, J, of FM bcc Fe (~0.2eV) is much the studied samples, always showing the transi- larger than that of the paramagnetic (PM) fcc Fe (close tion temperature around −23°C, transition that to 0eV), so that the Fe atoms at the interface between has been attributed to the retained austenite. PM austenite and FM ferrite are more strongly coupled ● Furthermore, these magnetic transitions of aus- to the FM ferrite side, and thus more likely to follow the tenite are irreversible for all the samples, sug- FM ordering. In this sense, for the samples studied in gesting a proximity magnetic effect between the present work, as the bainitic ferrite surrounding the such phase and the ferromagnetic ones (T retained austenite (both film-type and blocky-type in the ~770°C) surrounding it. different samples) is FM ordered down from ~770°C a ferromagnetic coupling to Fe atoms of the austenite films at the interface between both phases can be Disclosure statement expected. Such Fe atoms will be AF or FM ordered No potential conflict of interest was reported by the depending on the microstructural and chemical factors authors. relative to each sample, as already discussed, with a T or T around −23°C, and above this temperature they are PM. However, the exchange parameter, J, for FM Funding bainitic ferrite is supposed to be larger than that for the This work was supported by the Universidad de Oviedo PM austenite and, thus, it can keep the magnetic order [PAPI-18-EMERG-26]. already established at low temperatures. References 4. Conclusions [1] Kwon O. What’s new in steel? Nat Mater. In the present study, bainitic microstructures consist- 2007;6:713. ing of ferrite plates of nano- or submicro-sized [2] Sumner A, Gerada C, Brown N, et al. Controlling dimensions interspersed with thin-films of DC permeability in cast steels. J Magn Magn Mater. C-enriched retained austenite, and blocky austenite, 2017;429:79–85. [3] Kitanov S, Podol’skii A. Analysis of eddy-current were produced by isothermal treatment at different and magnetic rail brakes for high-speed trains. temperatures of three high C and high Si steels. Open Transp J. 2008;2:19–28. Their magnetic properties have been characterized [4] Ma D-M, Shiau J-K. The design of eddy-current magnet and correlated with very particular and distinctive brakes. Trans Can Soc Mech Eng. 2011;35(1):19–37. microstructural features. The study has been per- [5] Arumugam P, Xu Z, La Rocca A, et al. High-speed solid rotor permanent magnet machines: concept and design. formed at both, room temperature and in the cryo- IEEE Trans Transp Electrific. 2016;2(3):391–400. genic range. The main findings and conclusions are: [6] Garcia-Mateo C, Sourmail T, Caballero FG, et al. Nanostructured steel industrialisation: plausible At cryogenic temperatures, retained austenite reality. Mater Sci Technol. 2014;30(9):1071–1078. thermal stability was characterized by the occur- [7] García Mateo C, Caballero FC. Ultra-high-strength rence, or not, of martensitic transformation. In bainitic steels. ISIJ Int. 2005;45(11):1736–1740. Sci. Technol. Adv. Mater. 20 (2019) 685 A. ARGÜELLES et al. [8] Bhadeshia HKDH. Bainite in steels: theory and prac- [26] Le Houillier R, Bégin G, Dubé A. A study of the tice. 3rd ed. Wakefield (UK): Maney Publishing; peculiarities of austenite during the formation of 2015. bainite. Metall Trans A. 1971;2(9):2645–2653. [9] Garcia-Mateo C, Caballero FG. Nanocrystalline baini- [27] Babu SS, Specht ED, David SA, et al. In-situ observa- tic steels for industrial applications. In: tions of lattice parameter fluctuations in austenite Nanotechnology for energy sustainability. and transformation to bainite. Metall Mater Trans Weinheim: Wiley-VCH Verlag; 2017.p. 707–724. B. 2005;36A(12):3281–3289. [10] Sourmail T, Garcia-Mateo C, Caballero FG, et al. [28] Stone HJ, Peet MJ, Bhadeshia HKDH, et al. Tensile ductility of nanostructured bainitic steels: Synchrotron X-ray studies of austenite and bainitic influence of retained austenite stability. Metals. ferrite. Proc R Soc A. 2008;464(2092):1009–1027. 2017;7(1):31. [29] Rementeria R, Jimenez JA, Allain SYP, et al. [11] Morales-Rivas L, Garcia-Mateo C, Kuntz M, et al. Quantitative assessment of carbon allocation anoma- Induced martensitic transformation during tensile lies in low temperature bainite. Acta Mater. test in nanostructured bainitic steels. Mater Sci Eng 2017;133:333–345. A. 2016;662:169–177. [30] Bhadeshia HKDH, Edmonds DV. Bainite in silicon [12] Rementeria R, Poplawsky JD, Aranda MM, et al. steels: new composition-property approach. Part 1. Carbon concentration measurements by atom Met Sci. 1983;17(9):411–419. probe tomography in the ferritic phase of [31] Bhadeshia HKDH, Edmonds DV. The bainite trans- high-silicon steels. Acta Mater. 2017;125:359–368. formation in a silicon steel. Metall Trans A. 1979;10 [13] Sourmail T, Smanio V, Ziegler C, et al. Novel nanos- (7):895–896. tructured bainitic steel grades to answer the need for [32] Chang LC, Bhadeshia HKDH. Metallographic obser- high-performance steel components (Nanbain). vations of bainite transformation mechanism. Mater Luxembourg: European Comission; 2013. Sci Technol. 1995;11(2). [14] Balzar D, Ledbetter H. Voigt-function modeling in [33] Caballero FG, Yen HW, Miller MK, et al. fourier analysis of size- and strain-broadened X-ray Complementary use of transmission electron diffraction peaks. J Appl Crystallogr. 1993;26(pt microscopy and atom probe tomography for the 1):97–103. examination of plastic accommodation in nano- [15] Dyson DJ, Holmes B. Effect of alloying additions on crystalline bainitic steels. Acta Mater. 2011;59 lattice parameter of austenite. J Iron Steel Inst. (15):6117–6123. 1970;208:469–474. [34] Williamson GK, Smallman RE III. Dislocation den- [16] Garcia-Mateo C, Jimenez JA, Lopez-Ezquerra B, sities in some annealed and cold-worked metals from et al. Analyzing the scale of the bainitic ferrite plates measurements on the X-ray Debye-Scherrer by XRD, SEM and TEM. Mater Charact. spectrum. Philos Mag. 1956;1(1):34–46. 2016;122:83–89. [35] Dini G, Ueji R, Najafizadeh A, et al. Flow stress [17] Garcia-Mateo C, Caballero F, Bhadeshia H. analysis of TWIP steel via the XRD measurement Development of hard bainite. ISIJ Int. 2003;43 of dislocation density. Mater Sci Eng A. 2010;527 (8):1238–1243. (10–11):2759–2763. [18] Caballero FG, Yen H-W, Miller MK, et al. [36] Smallman RE, Westmacott KH. Stacking faults in Complementary use of transmission electron micro- face-centred cubic metals and alloys. Philos Mag. scopy and atom probe tomography for the examina- 1957;2(17):669–683. tion of plastic accommodation in nanocrystalline [37] Morales-Rivas L, Yen H-W, Huang B-M, et al. bainitic steels. Acta Mater. 2011;59(15):6117–6123. Tensile response of two nanoscale bainite [19] Singh SB, Bhadeshia HKDH. Estimation of bainite composite-like structures. JOM. 2015;67:2223–2235. plate-thickness in low-alloy steels. Mater Sci Eng A. [38] Ajus C, Tavares SSM, Silva MR, et al. Magnetic 1998;245(1):72–79. properties and retained austenite quantification in [20] Cornide J, Garcia-Mateo C, Capdevila C, et al. An SAE 4340 steel. Revista Matéria. 2009;14(3):993–999. assessment of the contributing factors to the nano- [39] Renzetti RA, Sandim HRZ, Sandim MJR, et al. scale structural refinement of advanced bainitic Annealing effects on microstructure and coercive steels. J Alloys Compd. 2013;577:S43–S47. field of ferritic–martensitic ODS eurofer steel. [21] Bhadeshia HKDH. Bainite in steels. Mater Sci Eng A. 2011;528(3):1442–1447. Transformations, microstructure and properties. [40] Oliveira VB, Sandim MJR, Stamopoulos D, et al. 3rd ed. London: H. Bhadeshia; 2015. Annealing effects on the microstructure and coercive [22] Caballero FG, Garcia-Mateo C, Chao J, et al. Effects field of two ferritic–martensitic eurofer steels: of morphology and stability of retained austenite on a comparative study. J Nucl Mater. the ductility of TRIP-aided bainitic steels. ISIJ Int. 2013;435:189–195. 2008;48(9):1256–1262. [41] Jiles D. Introduction to magnetism and magnetic [23] Garcia-Mateo C, Caballero FG, Miller MK, et al. On materials. New York: Taylor & Francis Group; 2010. measurement of carbon content in retained austenite [42] Grady DE. Origin of the linear term in the expres- in a nanostructured bainitic steel. J Mater Sci. sion for the approach to saturation in ferromagnetic 2012;47(2):1004–1010. materials. Phys Rev B. 1971;4(11):3982–3989. [24] Caballero FG, Garcia-Mateo C, Santofimia MJ, et al. [43] Zhao L, van Dijk NH, Brück E, et al. Magnetic and New experimental evidence on the incomplete trans- X-ray diffraction measurements for the determina- formation phenomenon in steel. Acta Mater. 2009;57 tion of retained austenite in TRIP steels. Mater Sci (1):8–17. Eng A. 2001;313(1):145–152. [25] Rees Bhadeshia GI. Bainite transformation kinetics. [44] Sicupira FL, Sandim MJR, Sandim HRZ, et al. Part 2. Non-uniform distribution of carbon. Mater Quantification of retained austenite by X-ray diffrac- Sci Technol. 1992;8:994–996. tion and saturation magnetization in Sci. Technol. Adv. Mater. 20 (2019) 686 A. ARGÜELLES et al. a supermartensitic stainless steel. Mater Charact. [63] García-Junceda A, Capdevila C, Caballero FG, 2016;115:90–96. et al. Dependence of martensite start temperature [45] Swartzendruber LJ. Properties, units and constants on fine austenite grain size. Scr Mater. 2008;58 in magnetism. J Magn Magn Mater. 1991;100:573– (2):134–137. 575. [64] Kitahara H, Tsuji N, Minamino Y. Martensite trans- [46] Tavares SSM, Mello SR, Gomes AM, et al. X-ray formation from ultrafine grained austenite in Fe-28.5 diffraction and magnetic characterization of the at.% Ni. Mater Sci Eng A. 2006;438:233–236. retained austenite in a chromium alloyed high car- [65] Brandt ML, Olson GB. Bainitic stabilization of aus- bon steel. J Mater Sci. 2006;41(15):4732–4736. tenite in low alloy sheet steels. Iron Steelmaker. [47] Tavares SSM, Pedrosa PDS, Teodósio JR, et al. 1993;20(5):55–60. Magnetic properties of the UNS S39205 duplex [66] Miihkinen VTT, Edmonds DV. M icrostructural stainless steel. J Alloys Compd. 2003;351(1):283–288. examination of two experimental high-strength bai- [48] Tavares SSM, Pardal JM, Souza JAD, et al. Magnetic nitic low-alloy steels containing silicon. Mater Sci phase quantification of the UNSS32750 superduplex Technol. 1987;3(6):422–431. stainless steel. J Alloys Compd. 2006;416(1):179–182. [67] Jacques PJ, Ladrière J, Delannay F. On the influence [49] Pardal JM, Tavares SSM, Cindra Fonseca MP, et al. of interactions between phases on the mechanical Influence of temperature and aging time on hardness stability of retained austenite in and magnetic properties of the maraging steel grade transformation-induced plasticity multiphase steels. 300. J Mater Sci. 2007;42(7):2276–2281. Metall Mater Trans A. 2001;32(11):2759–2768. [50] Shen YF, Qiu LN, Sun X, et al. Effects of retained [68] Garcia-Mateo C, Jimenez JA, Yen HW, et al. Low austenite volume fraction, morphology, and carbon temperature bainitic ferrite: evidence of carbon content on strength and ductility of nanostructured super-saturation and tetragonality. Acta Mater. TRIP-assisted steels. Mater Sci Eng A. 2015;91:162–173. 2015;636:551–564. [69] Grajcar A, Zalecki W, Skrzypczyk P, et al. [51] Solano-Alvarez W, Abreu HFG, daSilva MR, et al. Dilatometric study of phase transformations in Phase quantification in nanobainite via magnetic advanced high-strength bainitic steel. J Therm Anal measurements and X-raydiffraction. J Magn Magn Calorim. 2014;118:739–748. Mater. 2015;378:200–205. [70] Marcus PM, Moruzzi VL, Qiu S-L. Tetragonal equi- [52] Bozorth RM. Ferromagnetism. New York: D. Van librium states of iron. Phys Rev B. 1999;60:369–371. Nostrand Company, Inc.; 1961. [71] Pinski FJ, Staunton J, Gyorffy BL, et al. Ferromagnetism [53] Schweer FC, Spangler CE, Kelly JFJ. Temperature versus antiferromagnetism in face-centered-cubic iron. dependence of the magnetic coercitivity of perlite. Phys Rev Lett. 1986;56:2096–2099. Acta Metall. 1978;26:579–589. [72] Boukhvalov DW, Gornostyrev YN, Katsnelson MI, [54] King HW, Larbalestier DC. Austenitic stainless steels et al. Magnetism and local distorsion near carbon at cryogenic temperatures: the compositional depen- impurity in γ-iron. Phys Rev Lett. 2007;99:247205. dence of the Ms. Cryog. 1981;21(9):521–524. [73] Medvedeva NI, Van Aken D, Medvedeva JE. [55] Luzginova N, Zhao L, Sietsma J. Evolution and ther- Magnetism in bcc and fcc Fe with carbon and mal stability of retained austenite in SAE 52100 manganese. J Phys Condens Matter. 2010;22:316002. bainitic steel. Mater Sci Eng A. 2007;448(1):104–110. [74] Okatov SV, Gornostyrev YN, Lichtenstein AI, et al. [56] Koneshlou M, Meshinchi Asl K, Khomamizadeh F. Magnetoelastic coupling in γ-iron investigated Effect of cryogenic treatment on microstructure, within an ab initio spin spiral approach. Phys Rew mechanical and wear behaviors of AISI H13 hot B. 2011;84(7):214422. work tool steel. Cryog. 2011;51:55–61. [75] Warnes LAA, King HW. The low temperature mag- [57] Bhadeshia HKDH, Honeycombe RWK. Steels: netic properties of austentic Fe-Cr-Ni alloys 1. The microstructure and properties. London: effect of nickel concentration in Fe-Ni-20 wt% Cr Butterworths-Heinemann (Elsevier): 2006. alloys. Cryog. 1976;16(8):473–481. [58] Jacques PJ, Girault E, Mertens A, et al. The develop- [76] Lu J, Walsh RP, Han K. Low temperature physical ments of cold-rolled TRIP-assisted multiphase steels. properties of a high Mn austenitic steel JK2LB. Al-alloyed TRIP-assisted multiphase steels. ISIJ Int. Cryog. 2009;49(3):133–137. 2001;41(9):1068–1074. [77] Klimenko IN. Anomaly of the yield stress and mag- [59] Nohara K, Ono Y, Ohashi N. Composition and netic state in gamma Fe 18Cr Ni alloys. Acta Metall grain size dependence of strain induced marten- Mater. 1990;38(5):799–803. sitic transformation in metastable austenitic stain- [78] Hauser JJ. Magnetic proximity effect. Phys Rev. less steels. Tetsu to hagane. J ISIJ. 1977;63 1969;187(2):580–583. (5):212–222. [79] Unguris J, Celotta RJ, Pierce DT. Magnetism in Cr [60] Sourmail T, Garcia-Mateo C. Critical assessment of thin films on Fe(100). Phys Rev Lett. 1992;69 models for predicting the M-s temperature of steels. (7):1125–1128. Comput Mater Sci. 2005;34(4):323–334. [80] Maccherozzi F, Sperl M, Panaccione G, et al. [61] Yang H-S, Bhadeshia HKDH. Austenite grain size Evidence for a magnetic proximity effect up to and the martensite-start temperature. Scr Mater. room remperature at Fe/(Ga;Mn)As interfaces. Phys 2009;60(7):493–495. Rev Lett. 2008;101:267201. [62] Sadeghpour S, Kermanpur A, Najafizadeh A. [81] Kravets AF, Timoshevskii AN, Yanchitsky BZ, et al. Investigation of the effect of grain size on the Temperature-controlled interlayer exchange cou- strain-induced martensitic transformation in a pling in strong/weak ferromagnetic multilayers: high-Mn stainless steel using nanoindentation. a thermomagnetic Curie switch. Phys Rev B. Mater Sci Eng A. 2014;612:214–216. 2012;86:214413. Sci. Technol. Adv. Mater. 20 (2019) 687 A. ARGÜELLES et al. [82] Zhang X, Hickel T, Rogal J, et al. Structural trans- [83] Razumov IK, Boukhvalov DV, Petrik MV, et al. Role formations among austenite, ferrite and cementite in of magnetic degrees of freedom in a scenario of Fe–C alloys: a unified theory based on ab initio phase transformations in steel. Phys Rev B. simulations. Acta Mater. 2015;99:281–289. 2014;90:094101. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Science and Technology of Advanced Materials Pubmed Central

Cryogenic study of the magnetic and thermal stability of retained austenite in nanostructured bainite

Science and Technology of Advanced Materials , Volume 20 (1) – Jun 27, 2019

Loading next page...
 
/lp/pubmed-central/cryogenic-study-of-the-magnetic-and-thermal-stability-of-retained-XVnSupN2LS

References (87)

Publisher
Pubmed Central
Copyright
© 2019 The Author(s). Published by National Institute for Materials Science in partnership with Taylor & Francis Group.
ISSN
1468-6996
eISSN
1878-5514
DOI
10.1080/14686996.2019.1625722
Publisher site
See Article on Publisher Site

Abstract

SCIENCE AND TECHNOLOGY OF ADVANCED MATERIALS 2019, VOL. 20, NO. 1, 673–687 https://doi.org/10.1080/14686996.2019.1625722 Cryogenic study of the magnetic and thermal stability of retained austenite in nanostructured bainite a a b c Arántzazu Argüelles , Florentina Barbés , Jose I. Espeso and Carlos Garcia-Mateo Dpto. de Ciencia de los Materiales e Ingeniería Metalúrgica, Edificio Departamental Este-Campus de Viesques-Universidad de Oviedo, Gijón, Spain; Dpto. CITIMAC, Universidad de Cantabria, Santander, Spain; MATERALIA Research Group, National Center for Metallurgical Research CENIM-CSIC, Madrid, Spain ABSTRACT ARTICLE HISTORY Received 27 March 2019 First magnetic characterization of a recently developed generation of carbide free bainitic Revised 27 May 2019 steels, known as Nanobain, has been performed. Stability of its retained austenite at cryo- Accepted 28 May 2019 genic temperatures has been studied by means of X-ray diffraction, microscopy, dilatometry and magnetic measurements. Two morphologies for this phase (blocky-type and film-type) KEYWORDS appear in a different proportion depending on the chemical composition and the applied Bainitic steels; austenite-to- thermal treatment. Inhibition of the martensitic transformation, when decreasing the tem- martensite phase perature down to −271°C, has been observed in those microstructures with higher proportion transformation; nanostructured metals; of film-type austenite. The paramagnetic state of austenite at room temperature seems to magnetic properties; lead to different magnetic behaviors (ferromagnetic, antiferromagnetic) at cryogenic tem- cryogenic temperature peratures (T or T being around −23°C in all the studied samples), depending on the C N proportion of such morphological features. Furthermore, irreversibility with temperature on CLASSIFICATION the evolution of such magnetic behaviors has been observed for all the studied bainitic 10 Engineering and structures and is proposed to be due to a magnetic proximity effect. Structural materials; 106 Metallic materials; 302 Crystallization / Heat treatment / Crystal growth; 500 Characterization; 503 TEM, STEM, SEM; 504 X-ray / Neutron diffraction and scattering; 600 Others 1. Introduction of mechanical properties that placed it among the most important industrial materials even at present Despite steel is a long standing material and object [1]. However, this Fe-based alloy has much more to of a plentiful research from ancient times, it offer beyond its mechanical features. Thus, it is well remains the object of cutting-edge studies nowa- known the main function of steel in rotating days. Its metallurgy has given rise to a wide range CONTACT Arántzazu Argüelles arguellesarantzazu@uniovi.es Dpto. de Ciencia de los Materiales e Ingeniería Metalúrgica, Edificio Departamental Este-Campus de Viesques-Universidad de Oviedo, Gijón 33203, Spain © 2019 The Author(s). Published by National Institute for Materials Science in partnership with Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Sci. Technol. Adv. Mater. 20 (2019) 674 A. ARGÜELLES et al. electrical machines [2] and in designing eddy- Table 1. Chemical composition of the analyzed alloys (wt.%). Material C Si Mn Cr Cu Ni Mo current and magnetic rail brakes for high-speed Steel1 0.66 1.45 1.35 1.02 - 0.1 0.24 trains [3,4] relies on its electric and magnetic prop- Steel2 0.98 2.90 0.77 0.45 0.21 0.16 - erties. Also recently, semimagnetic stainless steel Steel3 0.99 2.47 0.74 0.97 0.17 0.12 0.028 has been tested and proposed as a suitable option in the design of high speed rotor permanent mag- net machines [5] because it meets both the mag- cementite precipitation during the bainitic transfor- netic and the mechanical requirements, with a high mation [8]. Those alloys were selected for being able yield strength. In this sense, the magnetic charac- to transform into nanostructured bainite when iso- terization of new developed steels is interesting by thermally transformed at low temperatures, see for itself, as can provide useful information for poten- example refs [6,9–12]. tial applications. Through this work, different types of heat treatments In this work and for the first time, the new gen- were applied, whose purpose will become clear later eration of steels, known as Nanobain, has been stu- when presenting the experimental results. In order to died from a magnetic point of view [6]. Its produce the desired bainitic structure, isothermal treat- microstructure, consisting of bainitic ferrite plates, ments after full austenitization were applied in condi- with thickness around tenths of nanometers, alternat- tions adapted to the different chemical compositions ing with carbon enriched retained austenite films, (treatment B in Figure 1). Selected bainitic microstruc- achieves exceptional mechanical properties (strengths tures were cooled down from room temperature (RT) in the range of 1.6–2.5 GPa and toughness of around to ~ −123°C and back to RT (treatment BC). Finally, pffiffiffiffi 30 MPa m [7]) comparable to that of maraging with the purpose of generating a fully ferritic micro- steels, but with a lower production cost. The goal of structure, samples of two of the alloys were quenched, the present work has been to get the first magnetic creating a microstructure composed of martensite (α´) characterization of Nanobain steels, and more speci- and retained austenite (γ), followed by a tempering fically, to find correlations between magnetic and treatment that allowed for the full decomposition of microstructural features, which is hoped to help in austenite into ferrite and scarce precipitates, (treatment designing the metallurgical scheme oriented to attain QT in Figure 1). a particular magnetic behavior together with specific Specifics of the different parameters used in the mechanical properties. heat treatments are gathered in Table 2. It has to be highlighted that the austenitization and isothermal temperatures and times, as well as the cooling rates, 2. Materials and experimental methods were adapted to the different alloys according to previous experiences [6,9–12]. In order to easily iden- The chemical composition of the steels used in this tify the samples, the following labelling procedure has work is listed in Table 1. All the alloys are high been used: a number identifying the steel category (as C steels with enough amount of Si (≥1.5%) to avoid Figure 1. Heat treatment schemes; treatment B (Bainite) for obtaining the bainitic microstructures, treatment BC (Bainite +Cryogenic) cooling down after having obtained the bainitic microstructure, treatment QT (Quench+Temper) in order to obtain an austenite-free microstructure. Sci. Technol. Adv. Mater. 20 (2019) 675 A. ARGÜELLES et al. Table 2. Heat treatments parameters: T and t are the ratio. XRD data were collected during 2 hours over γ γ austenitization temperature and time respectively; T and iso a2θ range 35 − 135°, with a step size of 0.01°. In this t are the temperature and time respectively, for the forma- iso study, the Rietveld analysis program TOPAS 4.2 tion of bainite by an isothermal treatment; T and t are cryo cryo (Bruker AXS) was used for quantification and calcu- the temperature and time, respectively, for the cryogenic lation of the structural parameters of both, the treatment. retained austenite and the bainitic ferrite. Line broad- SAMPLE T (°C) t (min) T (°C) t (h) T (°C) t (min) γ γ iso iso cryo cryo ening effects due to the lattice microstrains were 1B220+ 900 5 220 168 - - 1B220 900 5 220 24 - - analyzed with the double-Voigt approach [14]. In 1B250 900 5 250 8 - - order to eliminate the instrumental contribution to 1B300 900 5 300 5 - - 1B350 900 5 350 4.5 - - peak broadening, instrument functions were empiri- 1QT 900 5 - - - - cally parameterized from the profile shape analysis of 1BC220+ 900 5 220 168 −123 5 1BC350 900 5 350 4.5 −123 5 a corundum sample measured under the same con- 2B200 950 15 200 40 - - ditions. The austenite carbon content was estimated 2B350 950 15 350 8 - - 3B220 1050 5 220 40 - - using the well-known expression in ref [15], that 3B250 1050 5 250 25 - - relates the influence of different alloying elements to 3B350 1050 5 350 7 - - the lattice parameter. 3QT 1050 5 - - - - 3BC350 1050 5 350 7 −123 5 XRD sample preparation was performed using standard metallographic procedures but introducing a final set of cycles of etching and polishing in order shown in Table 1), the acronym corresponding to the to remove the surface layer that has been plastically thermal treatment and the temperature of the iso- deformed during the grinding step. That surface layer thermal treatment; the plus symbol in one of the may contain traces of martensite formed by transfor- experiments indicates extra time at the isothermal mation-induced plasticity (TRIP) of austenite due to transformation temperature. sample preparation, which would underestimate its While microstructures obtained with a heat treatment real fraction. of thetypeBwere allproducedinthe frameworkof The microstructure was revealed after a standard industrial collaborations according to the procedures metallographic preparation, followed by a final etch- described in [13], heat treatments denoted as BC and ing with a 2% nital solution. A JEOL JSM-6500 field QT were performed in a Bahr 805D high-resolution emission gun scanning electron microscope (FEG- dilatometer (TA Instruments, USA) equipped with an SEM; JEOL, Japan) operating at 10 kV was used to induction heating coil. Helium was used as quenching observe the microstructure. gas and the temperature was controlled by a K type Bainitic ferrite plate thickness (t ) was determined thermocouple welded to the central part of the sample on SEM micrographs by measuring the shortest dis- surface. For the cryogenic treatment, He flowing through tance perpendicular to the longitudinal dimension of a cooling coil immersed in liquid N was used instead. the ferrite plate, and correcting for stereological Thelengthchangeassociatedwiththedifferent metal- effects as described in reference [16]. lurgical events taking place during the heat treatments The magnetic measurements were made with were recorded by a linear variable displacement transdu- a Quantum Design Physical Property Measurement cer (LVDT) with a resolution of 0.05 μm. The mentioned System (PPMS; Quantum Design, USA), equipped with treatments were performed using fused silica push-rods a 9T superconducting magnet and able to cover to measure longitudinal changes in length, given the a temperature range from −271 to 77°C (2 to 350 K, small expansion coefficient of quartz (push rods), approximately). Saturation magnetization values were −6 −1 0.5 × 10 °C , when compared with the expansion obtained from the hysteresis loop at RT of every sample −6 −1 coefficient of steel, approximately 10 × 10 °C ;itis resulting from treatments B and QT (m 1). Subsequently, safe to conclude that the contribution of the push rods a thermomagnetic analysis was performed between RT to the measured change in length is negligible. and −271°C (2 K) first cooling and then heating. After Quantitative X-ray diffraction (XRD) analysis was this analysis, a second value for the saturation magneti- used to determine retained austenite and bainitic zation (m 2) was obtained from the hysteresis loop at RT ferrite volume fractions. Samples were step-scanned measured for each sample. in a Bruker AXS D8 X-ray diffractometer (BRUKER AXS, USA) with a rotating Co anode X-ray tube as a radiation source, Göebel mirror optics and 3. Results and discussion a LynxEye Linear Position Sensitive Detector for 3.1. Nanostructured bainite: microstructural ultra-fast XRD measurements. A current of 30 mA characterization at room temperature and a voltage of 40 kV were employed as tube set- tings. Operational conditions were selected to obtain For all the steels under investigation, the microstruc- X-ray diffraction data with a high signal to noise ture after the B treatment consists of bainitic ferrite Sci. Technol. Adv. Mater. 20 (2019) 676 A. ARGÜELLES et al. (α ) and retained austenite (γ). As reported in pre- The bainite transformation is a displacive and difus- vious works, the microstructure is essentially carbide sionless reaction, in which the ferrite is initially super- free and higher magnification techniques, such as saturated with respect to carbon. The carbon excess in transmission electron microscopy and atom probe the bainitic ferrite is subsequently and rapidly parti- tomography, only revealed scarce quantities of tioned into the residual austenite, but substitutional cementite precipitates [17,18]. Examples of the elements do not partition during the bainite reaction microstructures at selected temperatures can be [8]. In the absence of carbide precipitation, prevented in found in Figure 2, corresponding to Steel1 after iso- the present steels by the use of silicon, the austenite thermal treatment at 350°C (sample 1B350) (Figure 2 carbon enrichment is such, at all transformation tem- (a)) and at 220°C (sample 1B220+) (Figure 2(b)), peratures, that during the quenching to room tempera- respectively, where both phases have been identified. ture no martensite forms, i.e. the M (martensite start The darker long slender features are the plates of temperature) of C-enriched retained austenite is well bainitic ferrite (α ), and the lighter phase found as below room temperature. Table 3 summarizes the films and more blocky type features correspond to results of the measured C content in austenite (C ) retained austenite (γ). XRD results, presented in where it is clear that the level of C enrichment is Table 3, corroborate that only these two phases are above of that of the bulk content. present, being the bainitic ferrite the predominant Measurements of the bainitic ferrite plate thick- and the austenite the minor one. The maximum ness, t (see Table 3), indicate a coarsening of the αb extent of transformation increases as transformation ferritic matrix as the transformation T increases (see temperature decreases, i.e. bainitic ferrite fraction Figure 2). These results are in agreement to what is (V ) decreases as isothermal temperature increases, expected and already extensively reported, see for αb which is expected from the incomplete reaction phe- instance refs [16,19,20]. In this kind of microstruc- nomena ruling the bainitic transformation [8]. Note tures, austenite exhibits a well known multi-scale that the time used in the isothermal treatment of the character associated with two distinguishable 3B220 sample did not allow a full bainitic transfor- morphologies, i.e., thin films (γ ) between platelets mation, so that more bainite could have been formed of bainitic ferrite and blocks between sheaves of bai- in a longer time, i.e. the final amount of bainite is nite (γ ), see for example ref [9,10,21–23]. Those smaller than that contained in the 3B250 sample. have identified in Figure 2, i.e. in the main Figure 2. SEM micrographs showing examples of the bainitic microstructures of (a) 1B350, and (b) 1B220+, where bainitic ferrite (α ) and the two retained austenite morphologies, blocky (γ ) type and thin films (γ ) have been identified. b B f Sci. Technol. Adv. Mater. 20 (2019) 677 A. ARGÜELLES et al. micrograph the blocky-type, and the film-type, due to The largest fraction of film type retained austenite their much smaller size, are identified in the higher is obtained for the sample 1B220+, while samples magnification inset. The two different morphologies treated at the highest austempering temperature are well known for also having very different carbon have the largest fraction of blocky type austenite contents in solid solution, i.e. the thin films are far (1B350, 2B350, 3B350). These results are consistent richer in carbon than the blocks [23–29]. with what is observed in the micrographs shown in There are two important effects on the microstruc- Figure 2, and they are expected from the incomplete ture as a consequence of decreasing the transformation reaction phenomena [8]. temperature. On the one hand, as the fraction of bainitic The displacive character of the bainitic transforma- ferrite increases, the fraction of thin films of austenite tion implies that there is a shape change along the (V ) increases at the expense of that of blocky austenite transformation. The plastic relaxation of this shape γf (V ). An estimate of the ratio between both morphol- change commonly takes place via generation of both, γB ogies can be done following the procedure described in dislocations in the austenite/bainitic ferrite interface, ref [30], where it is assumed that about 15% of the and also micro/nano-twins in the austenite, in contact volume contained within the boundaries of a bainite with bainitic ferrite plates [31–33]. As the transforma- sheaf consists of retained austenite films interspersed tion temperature decreases, there is an increase in the with bainite plates. Results obtained in this way are dislocation density caused by the yielding of the auste- summarized in Figure 3, presented as a function of V . nite. If we consider that microstrain is directly related αb to the dislocation density [34–36](see Table 3)it is clear that indeed, the lower is the transformation tem- perature, the higher is the dislocation density intro- duced in the microstructure. Note that the differences in microstrain between 1B220+ and 1B220 correspond to a restoration process due to the extended heat treat- ment [37]. 3.2. Nanostructured bainite: a preliminary magnetic characterization at room temperature (M versus H) In order to get an initial insight into the magnetic beha- vior of these bainitic microstructures, their hysteresis loops at room temperature (RT) have been measured. An example of the curves thus obtained is pre- sented in Figure 4 for the 1B220 and 3B220 micro- structures. From such curves, the value of the Figure 3. Relationship between the V =V ratio (thin film/ γf γB coercive field (H ) is obtained, and gathered in blocky type austenite) and the amount of bainitic ferrite (V ) αb Table 4. The obtained values range between 29–67 present in the studied microstructures. The line is just a guide Oe, close to the coercitivities corresponding to other for the eye. Table 3. Chemical and microstructural parameters of the studied samples. V ,V ,V are the volume fractions obtained by γXRD αb α’ XRD for austenite, bainitic ferrite and martensite, respectively. V is the volume fraction of austenite obtained by magnetic γmagn measurements. ε is the microstrain and t is the bainitic ferrite plate thickness. Those parameters for samples obtained after γ αb treatment B have been previously reported [6,9–11]. Sample V (%)±3% V (%)±3% C (wt.%) ±0.05 V (%)±3% V (%) ε (%) t (nm) γXRD αb γ α´ γmagn γ αb 1B220+ 14 86 0.89 - 11 0.0041 38 ± 6 1B220 17 83 0.71 - 17 0.0036 42 ± 7 1B250 25 75 0.81 - 19 0.0032 42 ± 11 1B300 27 72 1.25 - 33 0.0027 61 ± 10 1B350 47 53 1.12 - 44 0.0018 79 ± 14 1QT 0 100 1BC220+ 12 88 - 0.0029 36 ± 4 1BC350 29 62 9 0.0020 - 2B200 29 71 1.19 - 28 0.0033 40 ± 11 2B350 57 43 1.34 - 54 0.0017 155 ± 30 3B220 41 59 1.21 - 31 0.0029 34 ± 8 3B250 23 77 1.43 - 26 0.0027 43 ± 10 3B350 45 54 1.32 - 55 0.0016 74 ± 13 3QT 0 96a 3BC350 36 53 10.7 0.0023 - The 3QT sample contains 4% volume fraction of cementite. Sci. Technol. Adv. Mater. 20 (2019) 678 A. ARGÜELLES et al. steel grades [38–40], and closer to those values of soft a) magnetic materials than to those of hard magnetic ones, as is well known Hc is typically less than 10 Oe for the first ones and more than 100 for the second ones [41]. The initial magnetization curves, at RT, are pre- sented in Figure 5 for all the studied samples. It is well known that ferromagnetic materials follow the so- called law of approach to saturation given by Equation 1QT (2), where m is the saturation magnetization and the 1B220+ 1B220 parameters a (in Oe) and b (in Oe ) are positive and 1B250 1B300 related to crystallographic anisotropy and internal 1B350 strain [42]. As previously reported by other authors, Fits the saturation magnetization value in steel grades can be obtained from a fit of the experimental data to b) Equation (2) [43,44]. Such these fits, for an applied field H in the range 7000 Oe < H < 20,000 Oe, are 1B220 3QT 3B220 2B200 2B350 3B220 3B250 3B350 Fits 0 5 10 15 20 25 magnetic field (kOe) -50 Figure 5. Measured magnetization as a function of the applied magnetic field, at RT, for (a) samples 1B, (b) samples -100 -2 Hc 2B and 3B. The lines correspond to fits with Equation (2). -4 -150 -0.2 -0.1 0 0.1 0.2 presented in Figure 5.Theyresult in the m values -200 -25 -20 01 5 0 15 20 25 -15 -10 -5 denoted as m 1in Table 4. magnetic field (kOe) a b m ¼ m 1   (2) Figure 4. Measured magnetization as a function of the applied magnetic field (hysteresis loops) of the 1B220 and All the values obtained for the saturation magnetiza- 3B220 samples at RT. The inset shows the detail around 0 Oe tion are lower than that of pure Fe (217 emu/g) [45] in order to better appreciate the coercive field, H c. and close to values reported for other steel grades [43,46–49]. Differences between them can be attribu- Table 4. Magnetic parameters obtained for the studied sam- ted to the chemical composition and microstructural ples: coercitivity (H ), saturation magnetization for the baini- differences. It can be seen, as expected from references tic microstructures (m 1), and that of the same microstructures after undergoing a thermal cycle cooling [43,44,46,48,50,51], that for the same composition down to −271°C and back to room temperature (m 2). s there is a strong correlation between the amount of Sample H (Oe) m 1(emu/g) m 2(emu/g) c s s ferritic phase and the m values, i.e. the highest values 1B220+ 29 179.6 179.9 corresponding to the QT samples, which are almost 1B220 37 166.5 167.9 100% ferritic (Tables 3 and 4), are consistent with the 1B250 42 162.4 164.4 1B300 60 135.6 141.3 absence of austenite (paramagnetic at RT) in the 1B350 40 113.2 146.8 microstructure. Thus, a continuous decrease of m 1QT 191.1 2B200 57 137.8 139.8 value with increasing austempering temperature 2B350 38 86.9 141.2 (T ) is observed for the three analyzed steel grades. iso 3B220 67 131.9 135.7 3B250 67 140.9 144.4 The mentioned strong correlation between the m 3B350 65 86.1 143.2 values and the fraction of ferritic (ferromagnetic) phase 3QT 186.7 (189.13a) a has been exploited as a way for obtaining an estimate of Value corrected after considering the contribution of the cementite present in the microstructure, as explained in the text. the paramagnetic phase (austenite) present in the magnetization (emu/g) magnetization (emu/g) magnetization (emu/g) Sci. Technol. Adv. Mater. 20 (2019) 679 A. ARGÜELLES et al. microstructure of steels (see, for example, a behavior in the M(T) curve such as that of 1B220+ [43,44,46,48,50,51]).Thus, thevolumefractionofthe (see Figure 8) with a sharp increase in their magnetiza- retained austenite can be determined from the saturation tion when decreasing T down from ~ −23°C. This magnetization value according to Equation (3) [43,44], increase constitutes around 3% of the total magnetization value at this temperature range, what means a small variation respect to the total magnetization value of the V ¼ 100 1  (3) γmagn m ðÞ i s ferromagnetic iron phase (see inset in Figure 8). The significance of this transition will be clear later in the text. where m is the saturation magnetization of each sample On the contrary, magnetization decreases abruptly and m (i) is the intrinsic saturation magnetization, cor- when decreasing T to values lower than that value of ~ responding to an austenite-free microstructure with the −23°C in samples with a bigger amount of retained same chemical composition. The sample 3QT contains, austenite, mainly of blocky-type, such as 1B350 and besides ferrite, also a 4% volume fraction of cementite, so 3B350 (see Figure 9(a,b),respectively).Finally,athird in order to correct the experimental m value (~187 emu/ behavior, intermediate between those both already g) a lever rule was applied considering that cementite mentioned, is exhibited by other samples, such as saturation magnetization is ~129 emu/g [52,53]. The 3B250 (see Figure 9(c)) or 1B250, with a ‘cusp’ centered resulting corrected value for m is ~190 emu/g, which is around the same temperature (~ −23°C). Such samples close to that obtained in sample 1QT. contain a similar amount of both film-type and blocky- Furthermore, by plotting the measured volume frac- type austenite (see Figure 3). tion of bainitic ferrite, V , presented in Table 3, against αb Furthermore, it is worth to highlight that, for all the the specific saturation magnetization, m 1, reported in studied samples, irreversibility is observed for the mag- Table 4, it is possible to perform a linear fit, forcing it to netization evolution when returning to RT after cooling pass through zero, as m is supposed to be zero for (see Figures 8 and 9), which will be discussed later. a fully austenitic microstructure. From these fits, the intrinsic m (i) yields a value of ~201 emu/g for Steel1 (see Figure 6(a)), and a value of ~191 emu/g for Steel2 3.4. Stability of retained austenite in the and Steel3 (see Figure 6(b)). Such obtained values show cryogenic regime a reasonable level of agreement with those for the cor- responding QT samples (see Table 4). Despite small As thermomagnetic curves suggest changes of either chemical composition differences between Steel2 and structural or magnetic nature, a second hysteresis loop Steel3, both steel grades have been treated together as it has been measured at RT, for every sample, after cool- was checked that independent fits lead to similar values ing down to −271°C. This second measurement of the for m (i). For the sake of clarity, a comparison of the hysteresis loops let us check if there are differences in austenite content obtained from both methods, XRD the new saturation magnetization value (m 2) respect to and magnetic measurements (considering m (i) values that obtained from the virgin microstructure (m 1). from fits), is also presented in Figure 7. It can be seen These results are presented in Table 4 and, for discus- that the values obtained by both methods are in good sion purposes, are plotted in Figure 10. agreement, differing for most of the samples in no more The higher value of m 2 respect to m 1 for the s s than 6%, according with results reported by other samples 1B350, 2B350 and 3B350 is indicative of authors in different steel grades [44,50]. a higher fraction of a ferromagnetic phase, i.e. partial transformation from austenite (paramagnetic at RT) to martensite (ferromagnetic at RT), which is very 3.3. Thermomagnetic measurements (M versus T) likely to occur when cooling down to −271°C [46,54– 56]. On the contrary, in those samples where the m 1 In order to obtain further information on the auste- s and m 2 values are similar (see for example 1B220+), nite present in these samples, thermomagnetic curves s the occurrence of such transformation is unlikely and have been measured for each one of them, decreasing the microstructure is expected to be stable during the the temperature down to −271°C and increasing it up whole thermal cycle. A more direct way to visualize again to 27°C, while a small and constant magnetic this is by means of Equation (3), from which V field of 50 Oe was applied. Taking into account that γmagn (1) was calculated. Using this expression, new V ferrite gets ferromagnetically ordered below ~770°C, γmagn (2) values have been obtained from m 2 (see Table 4) and cementite below ~207°C, it can be assumed that s and they are plotted versus the previous V (1) any anomaly at lower temperatures has to be related γmagn ones in Figure 11. According to these results, the to either the austenite magnetic behavior or to struc- microstructures with the lowest V /V ratio, i.e. tural transformations from such phase. γf γB with higher fractions of blocky austenite (see Figure Some of the studied samples, with a high fraction of 3), as those transformed at 350°C, exhibit the highest bainitic ferrite, and therefore a high fraction of thin films degree of austenite transformation during the of austeniteoverthe blocky-type (Figure 3), show Sci. Technol. Adv. Mater. 20 (2019) 680 A. ARGÜELLES et al. Figure 7. Austenite volume fraction obtained from two dif- ferent methods: magnetic measurements (V ) and XRD γmagn (V ). The error associated to V and V is ±3% and γXRD γXRD γmagn ± 0.2% respectively. 1.45 1B220+ 1.44 H=50Oe 1.43 cooling 1.42 warming 2.0 1.41 1.5 1.40 1.0 -23ºC Figure 6. Saturation magnetization values, m 1, versus ferrite s 1.39 fraction (V ) for (a) Steel1; (b) Steel2 and Steel3. The lines 0.5 αb correspond to linear fits passing through the origin. The error 1.38 associated to m 1 is about 0.15%, and for V is ±3%. s αb -250 -150 -50 0 50 T(ºC) 1.37 -300 -250 -200 -150 -100 -50 0 50 T(ºC) cryogenic treatment. Note that these samples are the ones that deviate from the perfect correlation line. Figure 8. Thermomagnetic curve for 1B220+, measured dur- Before going deeper in the discussion, it is essen- ing cooling from RT down to −271°C and heating back to RT. The inset shows the same curve at a different scale to realize tial to comment on the parameters that may affect the about the relative change in the magnetization. thermal stability of austenite against martensitic transformation. composition, and elements such as C, Mn, Si and Al Martensitic transformation is displacive, involving [22,58,59] significantly enhance the austenite stability; the coordinated movement of atoms; and since the among them C is the element that exhibits the stron- glissile transformation interface has a dislocation gest influence. As already mentioned, the two differ- structure that has to move through any obstacles ent morphologies of retained austenite are strongly that exist in the austenite, martensitic transformation linked to different carbon contents, i.e. the thin films cannot be sustained against strong defects such as are far richer in carbon than the blocks. Since carbon grain boundaries. Less drastic defects, such as dislo- is an interstitial solute, such differences also imply big cations, also hinder the progress of such transforma- differences in their thermal stability [60]. tions, but can often be incorporated into the The differences in size of the austenitic features also martensite lattice [57]. have an important effect per-se on its stability against The resistance of austenite against the martensitic martensitic transformation [61–64]. This is in part due transformation partially depends on its chemical magnetization (emu/g) magnetization (emu/g) Sci. Technol. Adv. Mater. 20 (2019) 681 A. ARGÜELLES et al. 1.52 a) 1B350 -23ºC 1.50 H=50Oe 1.48 cooling 1.46 warming 1.44 1.42 m 1 1.40 m 2 -23ºC b) 3B350 H=50Oe 2.3 -40ºC 2.2 cooling Figure 10. Comparison of the saturation magnetization warming values before (m 1) and after (m 2) performing the thermo- s s 2.1 magnetic measurement, i.e. cooling down to −271°C. The error associated to m is about 0.15%. 2.0 1.9 -23ºC c) 3B250 1.85 H=50Oe 1.84 1.83 cooling warming 1.82 1.81 1.80 1.79 -300 -250 -200 -150 -100 -50 0 50 T(ºC) Figure 9. Thermomagnetic curves for (a) 1B350 sample; (b) 3B350 sample; (c) 3B250, measured during cooling from RT down to −271°C and heating back to RT. Figure 11. Correlation between the volume fraction of aus- to the fact that small retained austenite islands contain tenite (V ) obtained from m 1, and m 2. The associated γmagn s s lower potential nucleation sites for the transformation error is about 0.2%. to martensite and, consequently, require a greater total driving force for the nucleation of martensite [65]. Therefore, the smallest austenitic features are more other words a stronger matrix may prevent the mar- stable than the largest ones because of both their size/ tensitic transformation. morphology and their carbon content. Thus, in this work it is proposed a microstructural Then there is the influence that the strength of the stability parameter (msp) that accounts for almost all surrounding matrix, bainitic ferrite, has on the stabi- the mentioned factors influencing the stability of aus- lity of the retained austenite. Theoretically, the refine- tenite,i.e.carbon content (C ), dislocation density ment of the bainitic ferrite plates with different (microstrain ε ), morphology (the ratio r = V /V ) γ γf γB crystallographic orientations increases the stability and strength of the matrix (plate thickness t ). Such αb of the retained austenite. If the latter is closely sur- parameter is therefore defined as being proportional to rounded by the relatively rigid and refined bainitic msp , C rε =t (4) γ γ α ferrite, the stability of the retained austenite also increases due to the geometrical restrictions imposed The exact values of the different parameters can be by the surrounding bainitic ferrite laths [66,67]; in found in Table 3 and Figure 3. The higher the value magnetization (emu/g) magnetization (emu/g) magnetization (emu/g) m (emu/g) 1B220+ 1B220 1B250 1B300 1B350 2B200 2B350 3B220 3B250 3B350 Sci. Technol. Adv. Mater. 20 (2019) 682 A. ARGÜELLES et al. of this product is, the more stable is the austenite sample upon cooling from RT to −123°C, as there is within the initial bainitic microstructure (before the no deviation from the expected linear contraction. cryogenic treatment) against martensitic transforma- On the other hand, 350°C bainitic microstructures tion. This fact is strongly correlated with the variation exhibit deviations from linearity on cooling, indicat- of m when cooling sample down to −271°C, as it is ing the existence of a phase transformation: the shown in Figure 12. decomposition of austenite into martensite leads to In this sense, and from the magnetic point of view, a net expansion [68,69]. From these curves, the an equivalent stability parameter could be defined as beginning of the martensitic transformation has Δm =|m 1-m 2|, since, as already discussed, a zero been determined to take place at T = −7°C (M ) and s s s s or very small value for this parameter is indicative of T= −40°C (M ) for 1BC350 and 3BC350, respec- the austenite stability (paramagnetic at RT) respect to tively. Note that expansion is larger for 3BC350 its transformation in martensite (ferromagnetic at than for 1BC350, what agrees with an also higher RT) when undergoing the cryogenic treatment. difference between V (m 1) and V (m 2), as observed γ s γ s To further corroborate these results and the pre- in Figure 11. sent discussion, and by means of dilatometry, The transformation detected from dilatometry is a thermal cycle in the cryogenic regime (BC treat- barely reflected in the M(T) curve of the 3B350 ment in Figure 1) has been applied in three different sample (see Figure 9(b)), where a small anomaly at microstructures: one where the austenite is consid- around −40°C might indicate the start of the mar- ered as very stable (1B220+ sample), and other two tensitic transformation. As the austenite decomposi- where, according to the data so far presented, some tion into martensite is expected to evolve degree of austenite transformation has occurred, i.e. progressively in a wide temperature range, it is not 1B350 and 3B350 microstructures. The new samples strange that such subtle and progressive transforma- were codified as 1BC220+, 1BC350 and 3BC350. Due tion might not leave more evident traces on the to the equipment technical specifications, the mini- mentioned magnetic curves (Figure 9(a,b)). Note mum temperature is −123°C (the exact parameters of that for the 1B350 sample, the martensitic transfor- the thermal cycle are shown in Table 2). This range is mation starts at higher temperatures (M = −7°C) considered suitable, as the M(T) anomalies that could than that of the anomaly observed in the M(T) curve be related with phase transformations appear at much (−23°C) (see Figure 8(a)). higher temperatures (−23°C), as reported in Figures 8 Finally, in order to provide further additional and 9. experimental evidence, XRD analysis was performed The dilatometry measurements are shown in on the three samples that were measured in the Figure 13. There is no evidence of martensitic trans- dilatometer, to directly check the degree of transfor- formation from retained austenite in the 1BC220+ mation. These results are gathered in Table 3 and they agree with the previous discussion; i.e., the microstructures of 1B220+ and 1BC220+ are almost identical in terms of V , which means that no mar- tensitic transformation took place, while in the case Figure 12. Microstructural and magnetic stability parameters, msp and Δm respectively, as defined in the main body of the text. Note that, for the sake of clarity, the microstructural Figure 13. Dilatometric measurements on cooling from RT parameter has been represented in log scale. The solid line is down to −123°C for samples 1BC-220+, 1BC-350 and 3BC- just a guide for the eye. 350. Δl/lo represents the relative change in length. Sci. Technol. Adv. Mater. 20 (2019) 683 A. ARGÜELLES et al. of 1BC350 and 3BC350, 9.5 and 10.7 of fresh mar- where an increase should be expected as tensite (V ), respectively, was formed on cooling. a consequence of the appearance of a ferromagnetic α’ phase (fresh martensite α´) and the FM order of the thin films of austenite (see Figure 9(a,b)). 3.5. Magnetic behavior versus γ→α’ The decrease in M(T) is rationalized in terms of transformation the antiferromagnetism of the untransformed blocky austenite, which overcomes the expected increase due Once it has been established that for 1B220+ there is to the presence of α´ and thin films. In such blocky not martensitic transformation down to −123°C, austenite, much poorer in C than the thin films [23– there must be another effect that can explain the 29] the AF order is energetically favored between detected M(T) increase, below −23°C, when cooling those Fe atoms of the austenite cell with less the sample down to the cryogenic regime (see C neighbors [72,75–77]. In the present case, the Figure 8). establishment of such AF order is also found to be It is known that the magnetic state of γ (fcc), in Fe around T = −23°C (Néel temperature). Therefore, and Fe-based alloys, is strongly frustrated at cryo- the decrease of the M(T) is the net result of the genic temperatures, which leads to the establishment positive FM contribution of α´ and thin films, and of different magnetic structures, such as ferromag- the negative contribution of the AF blocky austenite. netic (FM), antiferromagnetic (AF), double layer Such overlapping of counteracting effects is evident antiferromagnetic (AFMD), Spin Spiral (SS) . . ., all when considering the smoother decrease of M(T) in of them with close energy values [70–73]. 1B350, where M ~ −7°C > T , as compared to that of At T = −23°C, the ferromagnetic bainitic ferrite s N 3B350, where M ~ −40°C < T (Figure 9(a,b)). matrix is already ordered, as its Curie temperature is s N In these particular cases, 1B350 and 3B350, the quite far above, ~770°C. As the inset in Figure 8 shows, measured microstrain before and after the cryogenic the change associated with the proposed transition cycle (RT→ −123°C → RT) at XRD shows an below −23°C it only represents a small, yet detectable, increase from 0.0018 and 0.0016 to 0.0020 and variation of M(T) as compared to that of the initial 0.0023 (%), respectively. Such increase would be con- microstructure, composed of 86% well ordered ferro- sistent with the formation of martensite on cooling, magnetic ferrite. In other words, such change can only and the concomitant introduction of dislocations due be attributed to a secondary phase, i.e., austenite. to its displacive character. In these samples, the An interesting observation is that the measured microstrain does not seem to change because of the austenite microstrain, before and after the dilato- establishment of the magnetic order (AF). metric cryogenic cycle at ~ −123°C, decreases signifi- An intermediate scenario, where the calculated cantly, from 0.0041 down to 0.0029 (%) (see Table 3). ratio between thin films and blocks of austenite is This fact points out a microstrain relief associated around 1, and no martensitic transformation was with the FM ordering of the austenite, as theoretically detected on cooling down to −123°C, corresponds predicted by Okatov et al. [74]. On the other hand, to the sample denoted as 3B250 (see Table 3 and Boukhvalov et al. [72], by means of ab initio electro- Figure 3). The thermal variation of this sample, pre- nic structure calculations, have shown that C, as an sented in Figure 9(c), shows a pronounced cusp cen- interstitial present in an Fe fcc lattice (austenite), tered around −23°C. This scenario reflects again the stabilized a FM state for the Fe atoms at the low competition between the FM arising from the bainite temperature ground state (−273°C), making its and thin films of austenite (prevailing at T > −23°C), energy to be lower than that of the antiferromagnetic and the AF of the blocks (prevailing at T < −23°C). (AF) state. Therefore, for the particular case of the It is clear now that the irreversibility observed in 1B220+ microstructure, the described microstructural the M(T) curves for all the analyzed samples can not features defining the retained austenite (mainly as be attributed to a martensitic transformation of the thin C-enriched films) seem to lead to its thermal retained austenite, opposite to what occurs in a Cr- stability when going to temperatures such as low as alloyed high C steel cooled down to −200°C, as −123°C, at the same time that a FM state for the Fe reported by Tavares et al. [46]. Following, this feature atoms is established. will be explained in terms of magnetism. Thus, it is proposed that the mentioned anomaly at There is a well known magnetic proximity effect that −23°C, in Figure 8, is the result of a ferromagnetic transi- may propagate the spin polarization of a magnetic tion of austenite, i.e. Curie temperature, T ~ −23°C. metal into a nonmagnetic one, as first reported by On the other hand, microstructures such as 1B350 Hauser in 1969 [78]. Since then, several works have and 3B350, with a higher proportion of austenite with used this effect to explain induced magnetic behaviors. block morphology, and martensitic transformations Unguris et al. [79] reported a spin-density wave (SDW) taking place at −7°C and −40°C, respectively, due to antiferromagnetism in Cr films deposited on an Fe their lower thermal stability, exhibit a sharp M(T) substrate, at temperatures above the T of bulk Cr, decrease as the temperature decreases below −23°C, N Sci. Technol. Adv. Mater. 20 (2019) 684 A. ARGÜELLES et al. due to the coupling of Cr and Fe at some distance from those microstructures where the morphology of the interface. Maccherozzi et al. [80] reported retained austenite is predominantly thin-film, a magnetic coupling between Mn and Fe in (Ga, Mn) martensitic transformation did not occur when As/Fe interfaces that extends over a region as thick as cooling down to −123°C; other microstructural 2 nm. Kravets et al. [81] investigated the interlayer factors such as microstrain, scale of the matrix exchange coupling in a ferromagnetic Ni-Cu trilayer or austenite carbon content have been found to system (strong/weak/strong) and explained the para- also play an important role. A microstructural magnetic-to-ferromagnetic transition of the spacer stability parameter (msp) that accounts for material by the magnetic proximity effect. almost all the mentioned factors influencing Furthermore, Zhang et al. [82] reported the application the stability of austenite is proposed. of this effect to their ab initio simulations of structural ● Thermomagnetic measurements have brought transformations between austenite and ferrite in Fe-C significant differences on the magnetic behavior alloys. And such a strongly coupling between both of the studied samples to light. Thus, phases supported also the work of Razumov et al. [83] a ferromagnetic order is detected for microstruc- about phase transformations in steels. In the same line, tures with a higher thin film/blocky ratio (pre- the actual work proposes that this magnetic proximity valence of film type austenite), while an effect is at the origin of the low-T magnetic order hold antiferromagnetic order is exhibited by those in austenite when heating to RT and, thus, explains the with lower film/blocky ratio (prevalence of observed irreversibility in the thermal evolution of the blocky type austenite). In fact, a competition of magnetization. Zhang et al. [82]reported thatthe both magnetic behaviors has been found for all exchange parameter, J, of FM bcc Fe (~0.2eV) is much the studied samples, always showing the transi- larger than that of the paramagnetic (PM) fcc Fe (close tion temperature around −23°C, transition that to 0eV), so that the Fe atoms at the interface between has been attributed to the retained austenite. PM austenite and FM ferrite are more strongly coupled ● Furthermore, these magnetic transitions of aus- to the FM ferrite side, and thus more likely to follow the tenite are irreversible for all the samples, sug- FM ordering. In this sense, for the samples studied in gesting a proximity magnetic effect between the present work, as the bainitic ferrite surrounding the such phase and the ferromagnetic ones (T retained austenite (both film-type and blocky-type in the ~770°C) surrounding it. different samples) is FM ordered down from ~770°C a ferromagnetic coupling to Fe atoms of the austenite films at the interface between both phases can be Disclosure statement expected. Such Fe atoms will be AF or FM ordered No potential conflict of interest was reported by the depending on the microstructural and chemical factors authors. relative to each sample, as already discussed, with a T or T around −23°C, and above this temperature they are PM. However, the exchange parameter, J, for FM Funding bainitic ferrite is supposed to be larger than that for the This work was supported by the Universidad de Oviedo PM austenite and, thus, it can keep the magnetic order [PAPI-18-EMERG-26]. already established at low temperatures. References 4. Conclusions [1] Kwon O. What’s new in steel? Nat Mater. In the present study, bainitic microstructures consist- 2007;6:713. ing of ferrite plates of nano- or submicro-sized [2] Sumner A, Gerada C, Brown N, et al. Controlling dimensions interspersed with thin-films of DC permeability in cast steels. J Magn Magn Mater. C-enriched retained austenite, and blocky austenite, 2017;429:79–85. [3] Kitanov S, Podol’skii A. Analysis of eddy-current were produced by isothermal treatment at different and magnetic rail brakes for high-speed trains. temperatures of three high C and high Si steels. Open Transp J. 2008;2:19–28. Their magnetic properties have been characterized [4] Ma D-M, Shiau J-K. The design of eddy-current magnet and correlated with very particular and distinctive brakes. Trans Can Soc Mech Eng. 2011;35(1):19–37. microstructural features. The study has been per- [5] Arumugam P, Xu Z, La Rocca A, et al. High-speed solid rotor permanent magnet machines: concept and design. formed at both, room temperature and in the cryo- IEEE Trans Transp Electrific. 2016;2(3):391–400. genic range. The main findings and conclusions are: [6] Garcia-Mateo C, Sourmail T, Caballero FG, et al. Nanostructured steel industrialisation: plausible At cryogenic temperatures, retained austenite reality. Mater Sci Technol. 2014;30(9):1071–1078. thermal stability was characterized by the occur- [7] García Mateo C, Caballero FC. Ultra-high-strength rence, or not, of martensitic transformation. In bainitic steels. ISIJ Int. 2005;45(11):1736–1740. Sci. Technol. Adv. Mater. 20 (2019) 685 A. ARGÜELLES et al. [8] Bhadeshia HKDH. Bainite in steels: theory and prac- [26] Le Houillier R, Bégin G, Dubé A. A study of the tice. 3rd ed. Wakefield (UK): Maney Publishing; peculiarities of austenite during the formation of 2015. bainite. Metall Trans A. 1971;2(9):2645–2653. [9] Garcia-Mateo C, Caballero FG. Nanocrystalline baini- [27] Babu SS, Specht ED, David SA, et al. In-situ observa- tic steels for industrial applications. In: tions of lattice parameter fluctuations in austenite Nanotechnology for energy sustainability. and transformation to bainite. Metall Mater Trans Weinheim: Wiley-VCH Verlag; 2017.p. 707–724. B. 2005;36A(12):3281–3289. [10] Sourmail T, Garcia-Mateo C, Caballero FG, et al. [28] Stone HJ, Peet MJ, Bhadeshia HKDH, et al. Tensile ductility of nanostructured bainitic steels: Synchrotron X-ray studies of austenite and bainitic influence of retained austenite stability. Metals. ferrite. Proc R Soc A. 2008;464(2092):1009–1027. 2017;7(1):31. [29] Rementeria R, Jimenez JA, Allain SYP, et al. [11] Morales-Rivas L, Garcia-Mateo C, Kuntz M, et al. Quantitative assessment of carbon allocation anoma- Induced martensitic transformation during tensile lies in low temperature bainite. Acta Mater. test in nanostructured bainitic steels. Mater Sci Eng 2017;133:333–345. A. 2016;662:169–177. [30] Bhadeshia HKDH, Edmonds DV. Bainite in silicon [12] Rementeria R, Poplawsky JD, Aranda MM, et al. steels: new composition-property approach. Part 1. Carbon concentration measurements by atom Met Sci. 1983;17(9):411–419. probe tomography in the ferritic phase of [31] Bhadeshia HKDH, Edmonds DV. The bainite trans- high-silicon steels. Acta Mater. 2017;125:359–368. formation in a silicon steel. Metall Trans A. 1979;10 [13] Sourmail T, Smanio V, Ziegler C, et al. Novel nanos- (7):895–896. tructured bainitic steel grades to answer the need for [32] Chang LC, Bhadeshia HKDH. Metallographic obser- high-performance steel components (Nanbain). vations of bainite transformation mechanism. Mater Luxembourg: European Comission; 2013. Sci Technol. 1995;11(2). [14] Balzar D, Ledbetter H. Voigt-function modeling in [33] Caballero FG, Yen HW, Miller MK, et al. fourier analysis of size- and strain-broadened X-ray Complementary use of transmission electron diffraction peaks. J Appl Crystallogr. 1993;26(pt microscopy and atom probe tomography for the 1):97–103. examination of plastic accommodation in nano- [15] Dyson DJ, Holmes B. Effect of alloying additions on crystalline bainitic steels. Acta Mater. 2011;59 lattice parameter of austenite. J Iron Steel Inst. (15):6117–6123. 1970;208:469–474. [34] Williamson GK, Smallman RE III. Dislocation den- [16] Garcia-Mateo C, Jimenez JA, Lopez-Ezquerra B, sities in some annealed and cold-worked metals from et al. Analyzing the scale of the bainitic ferrite plates measurements on the X-ray Debye-Scherrer by XRD, SEM and TEM. Mater Charact. spectrum. Philos Mag. 1956;1(1):34–46. 2016;122:83–89. [35] Dini G, Ueji R, Najafizadeh A, et al. Flow stress [17] Garcia-Mateo C, Caballero F, Bhadeshia H. analysis of TWIP steel via the XRD measurement Development of hard bainite. ISIJ Int. 2003;43 of dislocation density. Mater Sci Eng A. 2010;527 (8):1238–1243. (10–11):2759–2763. [18] Caballero FG, Yen H-W, Miller MK, et al. [36] Smallman RE, Westmacott KH. Stacking faults in Complementary use of transmission electron micro- face-centred cubic metals and alloys. Philos Mag. scopy and atom probe tomography for the examina- 1957;2(17):669–683. tion of plastic accommodation in nanocrystalline [37] Morales-Rivas L, Yen H-W, Huang B-M, et al. bainitic steels. Acta Mater. 2011;59(15):6117–6123. Tensile response of two nanoscale bainite [19] Singh SB, Bhadeshia HKDH. Estimation of bainite composite-like structures. JOM. 2015;67:2223–2235. plate-thickness in low-alloy steels. Mater Sci Eng A. [38] Ajus C, Tavares SSM, Silva MR, et al. Magnetic 1998;245(1):72–79. properties and retained austenite quantification in [20] Cornide J, Garcia-Mateo C, Capdevila C, et al. An SAE 4340 steel. Revista Matéria. 2009;14(3):993–999. assessment of the contributing factors to the nano- [39] Renzetti RA, Sandim HRZ, Sandim MJR, et al. scale structural refinement of advanced bainitic Annealing effects on microstructure and coercive steels. J Alloys Compd. 2013;577:S43–S47. field of ferritic–martensitic ODS eurofer steel. [21] Bhadeshia HKDH. Bainite in steels. Mater Sci Eng A. 2011;528(3):1442–1447. Transformations, microstructure and properties. [40] Oliveira VB, Sandim MJR, Stamopoulos D, et al. 3rd ed. London: H. Bhadeshia; 2015. Annealing effects on the microstructure and coercive [22] Caballero FG, Garcia-Mateo C, Chao J, et al. Effects field of two ferritic–martensitic eurofer steels: of morphology and stability of retained austenite on a comparative study. J Nucl Mater. the ductility of TRIP-aided bainitic steels. ISIJ Int. 2013;435:189–195. 2008;48(9):1256–1262. [41] Jiles D. Introduction to magnetism and magnetic [23] Garcia-Mateo C, Caballero FG, Miller MK, et al. On materials. New York: Taylor & Francis Group; 2010. measurement of carbon content in retained austenite [42] Grady DE. Origin of the linear term in the expres- in a nanostructured bainitic steel. J Mater Sci. sion for the approach to saturation in ferromagnetic 2012;47(2):1004–1010. materials. Phys Rev B. 1971;4(11):3982–3989. [24] Caballero FG, Garcia-Mateo C, Santofimia MJ, et al. [43] Zhao L, van Dijk NH, Brück E, et al. Magnetic and New experimental evidence on the incomplete trans- X-ray diffraction measurements for the determina- formation phenomenon in steel. Acta Mater. 2009;57 tion of retained austenite in TRIP steels. Mater Sci (1):8–17. Eng A. 2001;313(1):145–152. [25] Rees Bhadeshia GI. Bainite transformation kinetics. [44] Sicupira FL, Sandim MJR, Sandim HRZ, et al. Part 2. Non-uniform distribution of carbon. Mater Quantification of retained austenite by X-ray diffrac- Sci Technol. 1992;8:994–996. tion and saturation magnetization in Sci. Technol. Adv. Mater. 20 (2019) 686 A. ARGÜELLES et al. a supermartensitic stainless steel. Mater Charact. [63] García-Junceda A, Capdevila C, Caballero FG, 2016;115:90–96. et al. Dependence of martensite start temperature [45] Swartzendruber LJ. Properties, units and constants on fine austenite grain size. Scr Mater. 2008;58 in magnetism. J Magn Magn Mater. 1991;100:573– (2):134–137. 575. [64] Kitahara H, Tsuji N, Minamino Y. Martensite trans- [46] Tavares SSM, Mello SR, Gomes AM, et al. X-ray formation from ultrafine grained austenite in Fe-28.5 diffraction and magnetic characterization of the at.% Ni. Mater Sci Eng A. 2006;438:233–236. retained austenite in a chromium alloyed high car- [65] Brandt ML, Olson GB. Bainitic stabilization of aus- bon steel. J Mater Sci. 2006;41(15):4732–4736. tenite in low alloy sheet steels. Iron Steelmaker. [47] Tavares SSM, Pedrosa PDS, Teodósio JR, et al. 1993;20(5):55–60. Magnetic properties of the UNS S39205 duplex [66] Miihkinen VTT, Edmonds DV. M icrostructural stainless steel. J Alloys Compd. 2003;351(1):283–288. examination of two experimental high-strength bai- [48] Tavares SSM, Pardal JM, Souza JAD, et al. Magnetic nitic low-alloy steels containing silicon. Mater Sci phase quantification of the UNSS32750 superduplex Technol. 1987;3(6):422–431. stainless steel. J Alloys Compd. 2006;416(1):179–182. [67] Jacques PJ, Ladrière J, Delannay F. On the influence [49] Pardal JM, Tavares SSM, Cindra Fonseca MP, et al. of interactions between phases on the mechanical Influence of temperature and aging time on hardness stability of retained austenite in and magnetic properties of the maraging steel grade transformation-induced plasticity multiphase steels. 300. J Mater Sci. 2007;42(7):2276–2281. Metall Mater Trans A. 2001;32(11):2759–2768. [50] Shen YF, Qiu LN, Sun X, et al. Effects of retained [68] Garcia-Mateo C, Jimenez JA, Yen HW, et al. Low austenite volume fraction, morphology, and carbon temperature bainitic ferrite: evidence of carbon content on strength and ductility of nanostructured super-saturation and tetragonality. Acta Mater. TRIP-assisted steels. Mater Sci Eng A. 2015;91:162–173. 2015;636:551–564. [69] Grajcar A, Zalecki W, Skrzypczyk P, et al. [51] Solano-Alvarez W, Abreu HFG, daSilva MR, et al. Dilatometric study of phase transformations in Phase quantification in nanobainite via magnetic advanced high-strength bainitic steel. J Therm Anal measurements and X-raydiffraction. J Magn Magn Calorim. 2014;118:739–748. Mater. 2015;378:200–205. [70] Marcus PM, Moruzzi VL, Qiu S-L. Tetragonal equi- [52] Bozorth RM. Ferromagnetism. New York: D. Van librium states of iron. Phys Rev B. 1999;60:369–371. Nostrand Company, Inc.; 1961. [71] Pinski FJ, Staunton J, Gyorffy BL, et al. Ferromagnetism [53] Schweer FC, Spangler CE, Kelly JFJ. Temperature versus antiferromagnetism in face-centered-cubic iron. dependence of the magnetic coercitivity of perlite. Phys Rev Lett. 1986;56:2096–2099. Acta Metall. 1978;26:579–589. [72] Boukhvalov DW, Gornostyrev YN, Katsnelson MI, [54] King HW, Larbalestier DC. Austenitic stainless steels et al. Magnetism and local distorsion near carbon at cryogenic temperatures: the compositional depen- impurity in γ-iron. Phys Rev Lett. 2007;99:247205. dence of the Ms. Cryog. 1981;21(9):521–524. [73] Medvedeva NI, Van Aken D, Medvedeva JE. [55] Luzginova N, Zhao L, Sietsma J. Evolution and ther- Magnetism in bcc and fcc Fe with carbon and mal stability of retained austenite in SAE 52100 manganese. J Phys Condens Matter. 2010;22:316002. bainitic steel. Mater Sci Eng A. 2007;448(1):104–110. [74] Okatov SV, Gornostyrev YN, Lichtenstein AI, et al. [56] Koneshlou M, Meshinchi Asl K, Khomamizadeh F. Magnetoelastic coupling in γ-iron investigated Effect of cryogenic treatment on microstructure, within an ab initio spin spiral approach. Phys Rew mechanical and wear behaviors of AISI H13 hot B. 2011;84(7):214422. work tool steel. Cryog. 2011;51:55–61. [75] Warnes LAA, King HW. The low temperature mag- [57] Bhadeshia HKDH, Honeycombe RWK. Steels: netic properties of austentic Fe-Cr-Ni alloys 1. The microstructure and properties. London: effect of nickel concentration in Fe-Ni-20 wt% Cr Butterworths-Heinemann (Elsevier): 2006. alloys. Cryog. 1976;16(8):473–481. [58] Jacques PJ, Girault E, Mertens A, et al. The develop- [76] Lu J, Walsh RP, Han K. Low temperature physical ments of cold-rolled TRIP-assisted multiphase steels. properties of a high Mn austenitic steel JK2LB. Al-alloyed TRIP-assisted multiphase steels. ISIJ Int. Cryog. 2009;49(3):133–137. 2001;41(9):1068–1074. [77] Klimenko IN. Anomaly of the yield stress and mag- [59] Nohara K, Ono Y, Ohashi N. Composition and netic state in gamma Fe 18Cr Ni alloys. Acta Metall grain size dependence of strain induced marten- Mater. 1990;38(5):799–803. sitic transformation in metastable austenitic stain- [78] Hauser JJ. Magnetic proximity effect. Phys Rev. less steels. Tetsu to hagane. J ISIJ. 1977;63 1969;187(2):580–583. (5):212–222. [79] Unguris J, Celotta RJ, Pierce DT. Magnetism in Cr [60] Sourmail T, Garcia-Mateo C. Critical assessment of thin films on Fe(100). Phys Rev Lett. 1992;69 models for predicting the M-s temperature of steels. (7):1125–1128. Comput Mater Sci. 2005;34(4):323–334. [80] Maccherozzi F, Sperl M, Panaccione G, et al. [61] Yang H-S, Bhadeshia HKDH. Austenite grain size Evidence for a magnetic proximity effect up to and the martensite-start temperature. Scr Mater. room remperature at Fe/(Ga;Mn)As interfaces. Phys 2009;60(7):493–495. Rev Lett. 2008;101:267201. [62] Sadeghpour S, Kermanpur A, Najafizadeh A. [81] Kravets AF, Timoshevskii AN, Yanchitsky BZ, et al. Investigation of the effect of grain size on the Temperature-controlled interlayer exchange cou- strain-induced martensitic transformation in a pling in strong/weak ferromagnetic multilayers: high-Mn stainless steel using nanoindentation. a thermomagnetic Curie switch. Phys Rev B. Mater Sci Eng A. 2014;612:214–216. 2012;86:214413. Sci. Technol. Adv. Mater. 20 (2019) 687 A. ARGÜELLES et al. [82] Zhang X, Hickel T, Rogal J, et al. Structural trans- [83] Razumov IK, Boukhvalov DV, Petrik MV, et al. Role formations among austenite, ferrite and cementite in of magnetic degrees of freedom in a scenario of Fe–C alloys: a unified theory based on ab initio phase transformations in steel. Phys Rev B. simulations. Acta Mater. 2015;99:281–289. 2014;90:094101.

Journal

Science and Technology of Advanced MaterialsPubmed Central

Published: Jun 27, 2019

There are no references for this article.