Decomposition of γ-Fe in 0.4C–1.8Si-2.8Mn-0.5Al steel during a continuous cooling process: A comparative study using in-situ HT-LSCM, DSC and dilatometry

Original Article Decomposition of g-Fe in 0.4Ce1.8Si-2.8Mn-0.5Al steel during a continuous cooling process: A comparative study using in-situ HT-LSCM, DSC and dilatometry Man Liu a,c , Michael Bernhard b,c,* , Monika Kawulokov a d , Josef Walek e , Maximilian Kern b , Simona Zl a d , Peter Presoly b , Bedrich Smetana d , Marketa Tkadle ckov a e , Guang Xu a , Youn-Bae Kang b , Christian Bernhard b a The State Key Laboratory of Refractories and Metallurgy, The Collaborative Innovation Center for Advanced Steels of Ministry of Education, The Key Laboratory for Ferrous Metallurgy and Resources Utilization of Ministry of Education, Wuhan University of Science and Technology, Wuhan, 430081, China b Graduate Institute of Ferrous and Energy Materials Technology, Pohang University of Science and Technology, 77, Cheongam-ro, Nam-gu, Pohang-si, Gyeongsangbuk-do, Republic of Korea c Chair of Ferrous Metallurgy, Montanuniversitaet Leoben, Franz-Josef-Strasse 18, 8700, Leoben, Austria d Department of Chemistry and Physico-Chemical Processes, Faculty of Materials Science and Technology, VSBTechnical University of Ostrava, 17. Listopadu 2172/15, 70800, Ostrava, Czech Republic e Department of Metallurgical Technologies, Faculty of Materials Science and Technology, VSB-Technical University of Ostrava, 17. Listopadu 2172/15, 70800, Ostrava, Czech Republic article info Article history: Received 6 March 2023 Accepted 1 April 2023 Available online 7 April 2023 Keywords: Thermal analysis HT-LSCM Dilatometry DSC CCT Heat treatment abstract Continuous cooling transformation (CCT) diagrams represent roadmaps for producing all heat-treatable steels. CCT curves provide valuable information on the solid-state phase transformation sequence, depending on the defined cooling strategies, the alloying concept of the steel and previous processing steps. The experimental characterization of CCT diagrams is usually done on a laboratory scale applying thermal analysis of dilatometry. In current research studies, however, also other in-situ methods such as high-temperature laser scanning confocal microscopy (HT-LSCM) or differential scanning calorimetry (DSC) are frequently used to investigate phase transformations during thermal cycling. In the present study, HT-LSCM observations and DSC analysis are critically compared with dilatometry results by investigating the CCT diagram of a 0.4%C-1.8%Si-2.8%Mn-0.5%Al (in mass pct.) advanced steel grade. Furthermore, classical examinations by optical microscopy and hardness measurements were performed to support the analysis. In general, very good consistencies between all experimental techniques were identified in determining the transformation start temperature for pearlite, bainite and martensite. The optical microscopy confirmed the observed phase transformations and the results *Corresponding author. E-mail address: [email protected]c.at (M. Bernhard). Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/jmrt journal of materials research and technology 2023;24:3534e3547 https://doi.org/10.1016/j.jmrt.2023.04.009 2238-7854/©2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http:// creativecommons.org/licenses/by/4.0/). correlated with the measured hardness response. Based on the results, coupling of HTLSCM and DSC is considered as a valuable novel approach to plot CCT diagrams in future research. ©2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 1. Introduction In continuous cooling transformation (CCT) diagrams, fundamental processing parameters, e.g., cooling rate, are graphically linked to the final microstructure and corresponding hardness of quenched steels, fully annealed steels or the heataffected zone in steel welding [1,2]. During the supercooling of austenite, phase transformations depend on the chemical composition of the steel [3,4], the extent of austenite homogenizing [1], the austenite grain size (AGS), and the austenitizing temperature and time [5,6]. Currently, CCT diagrams are obtained either from directly plotting the experimental data or from semi-empirical equations using the chemical composition of the steel and its austenitizing treatments as input values. Many attempts have been undertaken at modeling austenite transformations in steel during cooling [7e10]. The calculation methods provide a valuable alternative to experimental measurement in predicting the material data required for heat treatment simulation. However, the calculation results strongly depend on the quality and amount of the experimental data used in adjusting the model parameters. Thermal analysis by dilatometry is widely utilized to determine transformation temperatures in CCT diagrams. The measured dilatation is plotted over temperature and the tangent method is applied to characterize the inflection points of dilation curves, which represent the transformation temperatures due to the different expansion coefficients of gand a phases [11,12]. The cooling rate region of 0.08e44 Cs 1 was designed to obtain static CCT curves of bainitic forging steels [13]. The dynamic CCT curves of three CrMoV steels with cooling rates between 5 and 100 Cs 1 were plotted by Siwecki et al. [14]. The CCT diagrams of medium carbon steels with different Cr, V, Ti and Mn contents [15], low-carbon boron steels [16], linepipe steel [17] and additive manufacturing deposited steels [18] could be measured successfully and drawn through dilatometry. Differential scanning calorimetry (DSC) analysis represents another effective method to measure the transformation temperatures according to the change in heat flow [19]. Ganesh et al. [20] designed the cooling rate range of 0.017e0.05 Cs 1 to measure the martensite starting temperature of low-carbon 9Cr steels with the help of DSC. Additionally, high temperature-laser scanning confocal microscopy (HT-LSCM) was utilized to directly and dynamically observe the nucleation of ferrite, bainite and martensite at different cooling rates [21]. The CCT diagram with a cooling rate from 0.50 to 8 Cs 1 of a 0.14 mass pct. C microalloyed steel was obtained by HT-LSCM [22]. Hence, this technique provides another valuable option to draw the CCT diagram based on the thermally etched microstructure. Recently, the coupling of two selected in-situ methods mentioned above has been reported in the literature. Dilatometry and DSC analyses were used by Xu et al. [23] to plot the CCT diagram of medium-carbon microalloying steel. Kawulokova et al. [24] employed the two cooling rates (0.0833 and 0.333 Cs 1 ) to determine transformation temperature by DSC and dilatometry. DSC and dilatometry were also linked to detect the precipitation hardening of X5CrNiCuNb16-4 steel [25]. HT-LSCM and DSC are complementary techniques to characterize solid-state phase transformations in the g-loop of the binary FeeP system [26]. The inter-/intragranular ferrite transformation of 0.22 mass pct. C Ti 2 O 3 /TiN engineered steels was observed by HT-LSCM and measured by DSC [27]. Another combination of dilatometry and HT-LSCM was adopted to observe the intragranular ferrite transformation [28,29]. The purpose of the present study is to critically compare the accuracy of in-situ HT-LSCM and DSC with dilatometry in measuring solid-state phase transformations during austenite decomposition. Furthermore, it should be demonstrated that coupling DSC measurements and HT-LSCM observations can be used as a novel approach to establish CCT diagrams of advanced steel grades. In the first part of this work, dilatometry, DSC and HT-LSCM were applied to characterize the microstructure evolution in 0.4%C-1.8%Si-2.8%Mn-0.5%Al (in mass pct.) steel during constant cooling. The chemical composition represents a common alloying concept in the design of high-strength steels. Cooling rates (CR) were applied within the limited range for each piece of equipment used. Optical microscopy and hardness measurements of the samples were conducted ex-situ. The second part discusses the accuracy in determining ferrite, pearlite, bainite and martensite transformation temperatures. Finally, a CCT diagram for the steel grade investigated is proposed based on overall results by considering the applied heat treatment parameters. 2. Materials and methods 2.1. Sample preparation and chemical analysis Samples of 50 g were produced in the high-frequency remelting (HFR) furnace Lifumat-Met-3.3-Vac from Linn High Therm GmbH (Eschenfelden, Germany). As the starting materials, technically pure iron cylinders (99.9 mass pct. Fe), metallic manganese flakes (99.89 mass pct. Mn), silicon lumps (Alfa Aesar, 99.9999 mass pct. Si, LOT# 61,600,760), high purity aluminum wire (Alfa Aesar, 99.999 mass pct. Al, LOT# 36,076) and a previously prepared Fe-4.4 mass pct. C alloy were used. The eutectic FeeC alloy was melted in a 20 kg induction furnace using synthetic desulfurized graphite powder (Alfa journal of materials research and technology 2023;24:3534e3547 3535 Aesar, LOT# BCBB5882). The melting was carried out in alumina crucibles under Ar 5.0 (purity 99.999%) over-pressure atmosphere. After 2 min of chemical homogenization by inductive bath movement, the melt was centrifugally cast into a copper mold [26,30]. The chemical analysis was determined using an optical emission spectrometer (OES), type Spectromax Version F (SPECTRO Analytical Instruments &Co.KG., Germany). The measured composition of the samples is listed as the average value in Table 1. 2.2. Differential scanning calorimetry (DSC) Experiments were carried out in a DSC 404F1 Pegasus from NETZSCH Geraetebau GmbH (Selb, Germany) with a rhodium (Rh) furnace (T max ¼1650 C) and a platinum DSC sensor instrumented with type S thermocouples. During the measurement, the protective tube of the Rh furnace was flushed with Ar 5.0 (70 ml min 1 ). A zirconium (Zr) getter was positioned below the DSC sensor to reduce the oxygen level at elevated temperatures (T >350 C) to a minimum. The DSC setup was calibrated by measuring the melting points and melting enthalpies of NETZSCH's standards of pure metals In, Bi, Al, Ag, Au, Ni and Co. Al 2 O 3 crucibles (85 ml) with lids were used; the reference was an empty crucible. A detailed description of the current DSC setup can be found in references [30,31]. It is well known that the DSC signal is sensitive to the heat exchange during the phase transformation (DH) per scanning time Dt(Dt¼DTSR 1 , SR is the scanning ratein Cs 1 )[26]. It is a common practice to increase the scanning rate or change the sample mass to intensify the DSC signal for a phase transformation. M. Bernhard et al. [26] proposed that a sample mass of 250 mg (3.5 3.5 2.6mm)issuitableinthecurrentDSC setup to characterize solid-state transformations in steels accurately. This approach was followed in the recent study for evaluating the cooling curves defined in Section 2.6. For detailed information on the interpretation of DSC signals to determine phase transformation temperatures in alloys (e.g., the definition of onset and peak temperatures), the authors refer to the NIST Recommended Practice Guide [32]. The onset temperature represents the first deviation from the stable baseline and is associated with the beginning of a phase transformation. A peak temperature is defined as a local maximum of the DSC signal and generally represents the end of the respective phase transformation. However, it has to be noted that each phase transformation may show characteristic curves of the DSC charts, depending on the steel's alloying system. 2.3. High-temperature laser scanning confocal microscopy (HT-LSCM) In-situ HT-LSCM has become an established tool to observe microstructural changes in steels up to the liquid state [26,30,33e35]. The experiments were carried out in a confocal laser scanning microscope VL2000DX manufactured by Lasertec (Yokohama, Japan) and a hightemperature furnace SVF17-SP from Yonekura (Yokohama, Japan). Small samples with 5 51.5 mm geometry were cut from the high-frequency remelted samples. The specimens were ground and polished. Within the HT-LSCM trials, the samples were put into alumina crucible and placed on the sample holder in the gold-coated, elliptical HT furnace. The furnace chamber was evacuated and flushed with Ar 5.0 to minimize the oxygen content. The temperature in the furnace was controlled using a type S thermocouple located at the bottom of the sample surface. A systematic referencing of the temperature was executed before every experimental campaign to determine the sample surface's temperature accurately. For this purpose, an external thermocouple was welded to the test-sample surface [36]. Throughout the experiment, a video was recorded with a maximum frame rate of 60 s 1 , enabling detailed studies of microstructure changes occurring post-experimentally due to the thermal etching effect [37]. 2.4. Dilatometry Dilatometry measurements at low cooling rates (Section 2.6) were performed by a dilatometer DIL Expedis SUPREME NanoEye from NETZSCH Geraetebau GmbH (Selb, Germany) with a corundum holder, push rod and S-type thermocouple. The inner space around the samples was evacuated and flushed with helium before each analysis. Temperature calibration (Ag, 99.999%) and calibration using standard expansion (corundum) were carried out before the experimental series. The samples were processed into cylinders with a diameter of 6.2 mm and a length of 9.7 mm. The mass of the cylinders was approximately 2250 mg. The dynamic highpurity atmosphere of He 6.0 (purity 99.9999%, flow rate 120 ml min 1 ) was maintained during the analysis. The static force applied to the sample was 0.2 N. Experiments with high cooling rates (Section 2.6) were performed on a dilatometer DIL 805A/D/T manufactured by TA Instruments (New Castle, USA). DIL 805 is a hardening quenching type dilatometer (A) with the possibility of deformation by pressure (D) and tension (T). Samples of 10 mm in length and 4 mm in diameter were inductively heated in a vacuum atmosphere to the austenitizing temperature and then continuously cooled at the defined cooling rate. The inductor consists of two coils, whereas the inner coil is used for gas quenching (N 5.0, purity 99.999%) and the outer coil is water-cooled. The measuring chamber consists of two parts, with the linear variable differential transformer (LVDT) monitoring the change in sample length on the one side and the sample mounted in the measuring system on the other. The measuring system for the quenching measuring head consists of three push rods and the thermocouple. The push rods are made of fused silica. The standard thermocouple PtePt/Rh 10% (type S) was used to measure the temperature of the sample. The thermocouple was fixed to the sample by spot-welding technique. Before starting each experiment, a vacuum was created in the measuring chamber using nitrogen 5.0. Table 1 eChemical composition of the investigated steel grade. C [mass pct.] Si [mass pct.] Mn [mass pct.] Al [mass pct.] 0.40 ±0.004 1.79 ±0.03 2.81 ±0.04 0.51 journal of materials research and technology 2023;24:3534e35473536 2.5. Microstructure characterization and hardness measurement Subsequent to the DSC and dilatometry experiments, samples were cut off, embedded and prepared by mechanical polishing and etching with 4% nital to study the microstructure. Specimens of HT-LSCM were not considered for OM as the microstructural changes were observed in-situ on the surface. A Zeiss optical microscope (OM, Oberkochen, Germany) was used to observe the microstructure of all samples. An HV1000A Vickers hardness tester measured the hardness of all samples (DSC, HT-LSCM and DIL) with a load of 50 g, and the average of five separate indentations was calculated as the final hardness. 2.6. Time-temperature profile and measurement of the initial austenite grain size (AGS) In the first step, the A 1 and A 3 temperatures had to be determined to define theaustenitizing temperature. TheDSC method was used, applying a slow heating rate of 0.17 Cs 1 (10Cmin 1 ) and considering a small sample mass of 50 mg (1.5 1.5 2.1 mm) to reduce the thermal lag during scanning. The DSC charts of three independent measurements are plotted in Fig. 1 (note the signals are corrected to an arbitrary offset; the unitisgivenin mV).TheonsetforA 1 canbeobservedasanabrupt change in the DSC signal at 736.6 ±0.3 C. The minor peak below A 1 is the Curie temperature (T C ) given by 726.8 ±1.6 C. The A 3 temperature is detected at 858.5 ±3C, represented as endset rather than forming a peak in the DSC signal. Based on the measured value of A 3 , the austenitizing temperature was set to 1000 C. In all experiments, the samples were heated at 0.50 Cs 1 to 1000 C and kept stable for 5 min. Subsequently, the samples were cooled to room temperature with the CR defined explicitly for each piece of equipment. DSC and HT-LSCM: The current DSC setup is limited to a maximum CR of 0.50 Cs 1 . The minimum CR was considered with 0.05 Cs 1 as slower cooling would significantly increase the experimental time and reduce the method's efficiency. Furthermore, only minor changes in the measured phase transformation temperatures are expected by further decreasing the CR. The HT-LSCM equipment enables cooling up to 20 Cs 1 . However, deviations from the linear CR must be accepted during fast cooling. Therefore, the highest cooling rate was defined as 10 Cs 1 . Prolonged cooling may result in surface oxidation of the sample. Based on pre-trials, the lowest CR of 0.20 Cs 1 was reasonable to avoid any surface oxidation of the sample. By coupling the available DSC setup with the HT-LSCM method, a wide range of cooling cycles can be investigated (0.05e10 Cs 1 ). To critically compare the obtained phase transformations by DSC and HT-LSCM, two similar cooling rates of 0.20 Cs 1 and 0.50 Cs 1 were applied in the CCT evaluation. DIL: The dilatometers were conserved as benchmark equipment to reconstruct the CCT diagram. The quenching dilatometer covered the high cooling rates (1e20 Cs 1 ). The high-resolution dilatometer investigated the range of very slow cooling (0.05e0.50 Cs 1 ). During the isothermal holding step at 1000 C, the austenite grains grow to minimize the system's free energy. The exact knowledge of the grain size at the start of cooling represents an essential parameter for evaluating the phase transformation temperatures and must be given for an accurately measured CCT. Large grains will delay the transformation started in diffusion-controlled mechanisms. Besides recording phase transformations, the HT-LSCM enables the determination of the grain size distribution, observed in-situ on the sample surface [36,38]. A highresolution picture of the video was analyzed at the final stage of isothermal annealing at 1000 C. As an example, the micrograph of the experiment with 0.5 Cs 1 is shown in Fig. 2 (a). Due to the thermal etching effect, the grain boundaries are clearly visible. The statistical analysis is plotted in Fig. 2 (b). The log-normal distribution confirms normal grain growth (NGG) with a mean austenite grain size (AGS) of 60 mm. Identical values were obtained for all other experiments performed by HT-LSCM (0.2, 1, 5 and 10 Cs 1 ) and were considered as representative for all in-situ trials by DSC, HTLSCM and dilatometry. The calculation of the CCT diagram in Fig. 3 (a) and (b) was performed using JMatPro software [39]. According to the HTLSCM results, the initial AGS was set to 60 mm. The equilibrium temperatures A 3 and A 1 are predicted as 816.6 C and 733.5 C, respectively. A 1 is in reasonable agreement with the DSC measurements (A 1 , DSC ¼736.6 C), where more significant deviations are observed in the case of A 3 (DT~40C). The bainite start temperature (B S ) at an infinite time is given by 457.8 C and the martensite start temperature (M S ) is calculated to be 245 C. At cooling rates lower than 0.25 Cs 1 , proeutectoid ferrite formation is expected, characterized by the temperature F S . The transformation curve of ferrite is very close to the pearlite start temperature (P S ). Pearlite is stable up to a critical CR of about 5 Cs 1 . The calculation indicates that the alloying elements in the studied steel grade suppress the bainite formation and shift the start temperature (B S ) to longer Fig. 1 eDSC measurements of A 1 and A 3 temperature with a heating rate of 10 Cmin ¡1 (0.17 Cs ¡1 ). journal of materials research and technology 2023;24:3534e3547 3537 incubation times. The critical CR for a martensitic steel matrix is 5 Cs 1 . 3. Results and discussion In the first section, the results of each in-situ technique are analyzed individually, supported by the OM results and hardness measurements. In the second part, the CCT is reconstructed and the potential and limitations of the experimental methods used are discussed critically. 3.1. Phase transformation temperatures and microstructure analysis 3.1.1. DSC analysis Fig. 4 (a) graphically summarizes all DSC signals obtained during cooling at 0.05e0.50 Cs 1 . Note that the DSC charts were corrected by an arbitrary offset; the unit is mV. The first onset in the signal may be assigned to the pearlite start temperature (P S ), with a significant heat exchange during slow cooling (0.05e0.25 Cs 1 ). The intensity of the DSC signal decreases with increasing cooling rate due to a less forming fraction of pearlite. As expected, P S is shifted to lower values at higher CR. No pearlite was identified at the CR of 0.35 and 0.50 Cs 1 . Austenite transformation to martensite is detected in the low-temperature range (T <300 C). Martensite forms already at a prolonged cooling rate of 0.10 Cs 1 . The phase fraction of martensite, and as a consequence, the DSC signal's intensity, continuously increases if the CR is increased. The M S seems to be slightly dependent on the CR. M S first increases and becomes constant at CR 0.25 Cs 1 with the average value of M S ¼270 C. M S is strongly affected by preliminary phase transformations. Due to the larger consumption of austenite to form pearlite at a CR of 0.10 Cs 1 , less austenite is available Fig. 2 eAustenite grains at the final stage of isothermal annealing at 1000 C (a) and analyzed AGS statistics (b). Fig. 3 eCooling rates defined in HT-LSCM and DSC trials (a) and DIL (b) along with the CCT calculated using JMatPro software [39]. DIL NE …DIL Expedis SUPREME NanoEye. DIL Q …DIL 805A/D/T. journal of materials research and technology 2023;24:3534e35473538 for martensite, resulting in a lower M S of 249 C. The decreased intensity of the DSC signal of P S with higher CR indicates less decomposition of austenite into pearlite. Hence, more untransformed austenite is available for the bainite and martensite transformations. At the same time, the DSC signal of bainite transformation is very weak. Thus, martensite transformation happens at a higher temperature, indicated by the increase of M S at the CR of 0.15e0.25 Cs 1 . At 0.50 Cs 1 , martensite is the primary phase, and M S becomes constant. Between 500 and 300 C, a change in the DSC signal is evident in Fig. 4 (a). Depending on the alloying concept of the steel grade, the bainite formation usually starts in this intermediate temperature range. However, an accurate prediction of the B S temperature is challenging in the original charts. The first derivative by time is plotted in Fig. 4 (b), representing the transformation rate in mVs 1 . From Fig. 4 (b), the B S temperature becomes visible as a deviation from the actual signal trend line at CR of 0.15e0.35 Cs 1 . The enlarged section of the bainite formation is plotted in Fig. 5. Bainite growth starts at a temperature close to that of martensite. The critical cooling rate of bainite formation is obtained by 0.50 Cs 1 , above which no bainite forms at a slower CR. However, as expected from the CCT calculations in Fig. 3, a clear dependence on the CR is not established. Fig. 6 shows the micrographs of all DSC samples analyzed by OM (4% nital etchant). The obtained microstructure constituents and corresponding hardness values are summarized along with the DSC results in Table 2. At CR of 0.05e0.20 Cs 1 , small fractions of pro-eutectoid ferrite are observed along the prior austenite grains. The ferrite formation could not be obtained in the DSC signal, as the ferrite fraction formed was too low and the resulting heat change was below the resolution limit of the DSC setup. Furthermore, the CCT calculations in Fig. 3 indicate that ferrite starts to grow nearly simultaneously along with pearlite, which may lead to less clear phase separation in the DSC chart. Pearlite is the dominant microstructure constituent in samples cooled at 0.05 and 0.10 Cs 1 ; the amount continuously decreases with a higher CR. At 0.35 Cs 1 and 0.50 Cs 1 , Fig. 4 eOffset of DSC signals during cooling with 0.05e0.50 Cs ¡1 , corrected to the baseline of 0 mV (a) and first derivative by time for obtaining the B S temperature (b). Fig. 5 eDSC transformation rate in the enlarged section of the bainite formation for cooling rates of 0.15e0.35 Cs ¡1 . journal of materials research and technology 2023;24:3534e3547 3539 no pearlite can be found in the micrographs, which is in excellent agreement with the DSC analysis. The highest fractions of bainite can be observed in Fig. 6 (c) and (d), corresponding to cooling rates of 0.15 and 0.20 Cs 1 . In this case, the most significant change in the DSC signals was also detected. However, the amount of bainite in Fig. 6 (b) (CR ¼0.10 Cs 1 ) was too low to observe accurately in the DSC measurements. Martensite is observed from 0.10 to 0.50 Cs 1 in Fig. 6 (b)e(g). Based on the DSC analysis and the micrographs, the critical CR, above which only martensite is obtained, is higher than 0.50 Cs 1 . 3.1.2. HT-LSCM observations Fig. 7 shows the occurrence of phase transformations during different cooling. White ellipses point out the positions where the phase transformation happens. The transformation type can be determined based on the characteristics of phase transformation products in conjunction with the temperature. Pearlite nucleated at prior austenite grain boundaries, growing into the austenite grains. The morphology of pearlite was an irregular shape. Bainite and martensite nucleated at the grain or phase boundaries and developed in a lath morphology. The main difference between bainite and martensite is the growth rate. Lath bainite occurred gradually, while lath martensite happened instantaneously and widely. The formation of ferrite could not be observed at any cooling rate. However, it must be noted that the range of adjustable magnification is limited during the investigation of CCT by HT-LSCM. A high magnification, which would be required to detect minor phase fractions forming from a single austenite grain, shows the disadvantage that the observation may not represent the whole sample surface. At a higher cooling rate, the sample does not stay long in the higher temperature region, indicating that a larger driving force is required for pearlite transformation. Hence, P S gradually decreases from 0.20 to 1 Cs 1 from 604 C to 553 C. The critical cooling rate at which pearlite does not form anymore is 5 Cs 1 .B S first increases and then decreases, while M S changes differently when increasing the cooling rate from 0.2 to 10 Cs 1 . Lower P S means lower consumption of austenite. In this case, the remaining austenite is easily transformed to bainite, showing an increase in B S within 0.20e1Cs 1 . As the bainite transformation is accompanied by carbon diffusion, less austenite with high carbon concentration is retained and M S decreases. When increasing the cooling rate, a more considerable undercooling degree for bainite transformation is needed, resulting in a decrease in B S . If austenite decomposes in less bainite fraction, more untransformed austenite is available for the martensite transformation; consequently, M S increases. The determined phase transformations and the observed microstructure constitution are listed in Table 3 along with the measured hardness. 3.1.3. Dilatometry Fig. 8 depicts the measured dilation curves for several cooling rates, represented versus temperature and given by an Fig. 6 eOptical microstructure of DSC samples at different cooling rates. 0.05 Cs ¡1 (a), 0.10 Cs ¡1 (b), 0.15 Cs ¡1 (c), 0.20 Cs ¡1 (d), 0.25 Cs ¡1 (e), 0.35 Cs ¡1 (f) and 0.50 Cs ¡1 (g). (F: ferrite; P: pearlite; B: bainite; M: martensite). Table 2 eDSC results of transformation temperatures compared with microstructure constituents obtained in the optical microscope. Cooling rate [Cs 1 ]Ps[ C] Bs [C] Ms [C] Phase (in-situ) Phase (OM) Hardness [HV 0.05 ] 0.05 644 ee PPþF 316 ±4 0.10 622 e249 P þMPþFþBþM 325 ±17 0.15 619 363 262 P þBþMPþFþBþM 494 ±33 0.20 610 360 265 P þBþMPþFþBþM 512 ±21 0.25 615 343 268 P þBþMPþBþM 608 ±16 0.35 e370 272 B þMBþM 612 ±18 0.50 ee 271 M B þM 614 ±16 journal of materials research and technology 2023;24:3534e35473540 arbitrary offset. Depending on the cooling rate, the shape of the curve and the transformation temperatures vary. P S is lowered as the cooling rate increases, whereas the M S points of samples at higher cooling rates seem stable within 271e282 C; see Fig. 8 (b). The transformed fraction curve exhibits several slope changes in the particular cases of 0.20 Cs 1 and 0.50 Cs 1 as well as 0.10 Cs 1 and 1 Cs 1 . For other cooling rates, and even if the optical micrographs show that the microstructures are composed of a mixture of different phases, only one phase transformation was recorded by expansion. It should be mentioned that the critical cooling rate for pure pearlite transformation is 0.05 Cs 1 , above Fig. 7 eThe occurrence of phase transformations at different CRs and their corresponding optical microstructures: 0.2 Cs ¡1 (a)e(c), 0.5 Cs ¡1 (d)e(f), 1 Cs ¡1 (g)e(i), 5 Cs ¡1 (j)e(k) and 10 Cs ¡1 (l)e(m). journal of materials research and technology 2023;24:3534e3547 3541 which other phase transformations happened during cooling. The required cooling rate for pure martensite transformation is 5 Cs 1 . Additional information can be extracted if the slope of the dilatation is given by the first derivative by time (dD/dt). The analysis to recognize the transformation types based on this approach is demonstrated in Fig. 9, showing the transformation rates as a function of the temperature. For the 0.20 and 0.50 Cs 1 cooling rates, three different peaks correspond to the formation of the pearlite, bainite and martensite mixture. At cooling rates of 1 Cs 1 , the transformation rate curve has a sharp and narrow peak related to the formation of pearlite. Similar to the DSC analysis, the onset of the peak is considered the start of the transformation. For the dilation curves to be interpreted, a careful and systematic analysis of resulting data in conjunction with optical micrographs in Fig. 10 was done (4% nital etchant). Though dilatation curves do not recognize ferrite transformation, it is observed by optical micrographs at the CR below 0.05 Cs 1 , as pointed out by arrows in Fig. 10 (a)e(d). The proportion of ferrite and pearlite fractions decreases with an increasing CR and disappears when the CR is larger than 0.5 Cs 1 . The martensite phase starts at 0.1 Cs 1 and increases when the CR is raised to 0.5 Cs 1 . Martensite consequently becomes the main phase in the microstructure of samples at the CR of 1e20 Cs 1 . The bainite structure begins to form at the CR of 0.2 Cs 1 and disappears when the CR is above 0.5 Cs 1 . The corresponding hardness response of samples at different CRs is listed in Table 4. 3.2. CCT diagrams: critical evaluation of techniques used in-situ and future perspectives Fig. 11 (a) and Fig. 11 (b) show the experimental CCT diagrams determined by HT-LSCM and DSC and dilatometry, respectively. Calculations performed with JMatPro software [39] are plotted for comparison. Though pro-eutectoid ferrite was identified in the OM analysis of a sample cooled at 0.05e0.20 Cs 1 , its formation could not be measured by any in-situ technique. The ferrite transformation curve is close to the pearlite transformation curve, as presented by the dashed lines calculated by JMatPro software. Hence, it is reasonable to speculate that the ferrite transformation happens first and then the pearlite transformation occurs quickly. The ferrite and pearlite transformations may overlap, which could not be distinguished from dilation curves. In addition, it can be concluded that all methods show a limited resolution in this case. The P S excellently agrees between HT-LSCM/DSC and DIL. The HT-LSCM performs better than the DSC in measuring P S at CR >0.35 Cs 1 . DIL enables the determination of P S up to aCRof1Cs 1 , as can be seen in Fig. 11 (b). The data sets obtained are in good accordance with the calculations. In Table 3 eHT-LSCM results of transformation temperatures compared with microstructure constituents obtained in the optical microscope. Cooling rate [ Cs 1 ]Ps[ C] Bs [C] Ms [C] Phase (in-situ) Hardness [HV 0.05 ] 0.2 604 351 260 P þBþM 510 ±26 0.5 604 393 255 P þBþM 603 ±24 1 553 424 233 P þBþM 610 ±22 5e403 222 B þM 598 ±21 10 e393 265 B þM 600 ±13 Fig. 8 eDilatation curves of high-resolution dilatometer at CR of 0.05e0.50 Cs ¡1 (a) and CR of 1e20 Cs ¡1 (b). journal of materials research and technology 2023;24:3534e35473542