Combination of In Situ Confocal Microscopy and Calorimetry to Investigate Solidification of Super‐ and Hyper‐Duplex Stainless Steels

The solidification processes of super‐ and hyper‐duplex stainless steels (i.e., UNS S32750 and S 33 207) are investigated in situ by a high‐temperature confocal laser scanning microscope (HT‐CLSM) and differential scanning calorimetry (DSC). The variations of δ‐ferrite phase fraction during solidification are measured quantitatively. The results show that liquid L→δ‐ferrite transformation first occurs at a certain degree of supercooling during the solidification process of steel. UNS S3DSS 3207 with a higher Cr content can result in a higher nucleation temperature and faster growth of δ‐ferrite compared to those of UNS S332750 steel. Moreover, both the liquidus (TL) and solidus temperatures (TS) are increased with the increasing Cr content, while TL increases greater than TS. Electron microscopies are used to quantify the fraction and composition of each phase. Scheil equation is employed to predict the distribution behavior of the main alloying elements in the solidification process, and the predicted results are consistent with the experimental findings. This study aims to provide real‐time experimental insights into the solidification kinetics of state‐of‐the‐art high‐alloy‐grade duplex steels and benefits for controlling the casting process in the real production of stainless steels.


Introduction
Duplex stainless steels (DSSs) are considered as a category of stainless steels with an equal amount of dual phases, that is, δ (ferrite) and γ (austenite) microstructures.The optimum mechanical property of DSSs can be achieved at approximately equal amounts of both phases. [1]Currently, DSSs have been widely applied in harsh environments, for example, marine, construction, chemical engineering plants, and nuclear engineering due to their significant beneficial features, such as corrosion resistance, high tensile strength, and low cost compared to austenitic stainless steels. [2,3]Due to the strong influence of alloying content on the properties, there are a wide range of duplex grades available, ranging from low-alloyed lean grades, typically designed for structural applications to high-alloyed super-and hyper grades for use in severely corrosive environments.Among different grades of DSSs, that is, UNS S32750 and S33207 (named as 2507 and 3207 for short from herein) are super-and hypercategories including higher-alloying contents (mainly Cr, Ni, N, etc.) and they have become an attractive alternative to other high-performance materials such as superaustenitic stainless steels and Ni-based alloys. [4]he optimization of the mechanical properties and corrosion resistance of DSSs depends on the precise control of microstructure evolution. [5,6]One of the aspects that play an important role is the solidification process control.It is reported that the hot-crack susceptibility is related closely to the ferrite/austenite ratio during the solidification process of the steel. [7]Previously, the solidification process of steel was mainly investigated through postmortem analysis which will inevitably lose some important information about the solidification process. [8]The recent advancements in high-temperature confocal laser scanning microscope (HT-CLSM) have offered a great capability for real-time and continuous observation of phase transformation at high temperatures during the solidification process.The pioneering work was made by Yin and Shibata et al. [9][10][11] .They successfully applied the HT-CLSM technique to investigate the dynamic behavior of the δ-ferrite/γ-austenite interphase boundary during δ/γ transformations and peritectic reactions in Fe-C binary alloys and low-carbon steels.[16] In situ characterization of solidification of high-alloy stainless steels has rarely been reported.Huang et al. [17] observed the solidification processes of AISI 304 stainless steel in situ using HT-CLSM at a cooling rate of 3 °C min À1 , and found that δ-ferrite phase first appeared in liquid and γ-phase precipitated at the δ-grain boundary when the solid fraction was about 0.6 at 1452. 2 °C during solidification.Li et al. [18] studied the volume fraction variation of as-cast δ-ferrite at a cooling rate from 5 to 6000 °C min À1 (100 °C s À1 ) of 316H austenitic stainless steel and found that the solidification mode of 316 H was a ferrite-austenite (FA) type regardless of the cooling rate.The effects of the cooling rate on the solidification and microstructure evolution in the SAF 2205 DSS were studied using the differential scanning calorimetry (DSC) method and it was reported that the content of δ-ferrite increased with an increase in cooling rate.In addition, the solidification process occurred with a liquid-to-solid phase transformation, and segregation of main alloying elements such as Cr, Ni, N, and Mo in DSSs could lead to defects.Zhu et al. [19] reported that the formation of nitrogen-rich phases was beneficial to eliminate N segregation and suppress gas pore formation.It was also reported that austenite was enriched in Ni while ferrite had more Cr and Mo in DSSs, [20][21][22] and the partition of Mo between austenite and ferrite was very minor compared with Ni and Cr partitions. [23]So far, no research regarding the in situ observation of the solidification process of DSSs and the elemental segregation behavior utilizing HT-CSLM has been founded.
Chemical compositions play an important role in the solidification behavior.With increased Cr content from 25% in DSS 2507 to 32% in DSS 3207, it also resulted in different physical properties of steel, which could have effects on the solidification behavior of steel.Regarding the high-temperature phase equilibria research, Pettersson et al. [24] investigated the austenite-toferrite transformation at temperatures approaching 1400 °C using a novel in situ neutron scattering approach, and they found that higher austenite dissolution temperatures were measured for the higher-alloyed grades in DSS 3207 compared to that of DSS 2507.However, the effect of Cr content on solidification features in these high-alloyed DSS grades was not reported on.For instance, the phase formation kinetics (i.e., phase fraction evolution) as well as starting temperatures of δ-ferrite precipitation have not been reported.Therefore, it is of great interest to perform an in situ characterization of the solidification process of DSSs 2507 and 3207 utilizing the robustness of HT-CLSM and other related techniques.

Experimental Methods
Chemical compositions of the proposed DSSs are listed in Table 1.HT-CLSM (Lasertec, 1LM21H) with a He-Ne laser beam with a power of 1.5 mW and wavelength of 632.8 nm was used to analyze the solidification of DSSs.Details of the instrumentation could be found in previous works. [25,26]DSS specimen was machined to 4 Â 4 Â 1.5 mm and subsequently one flat side was polished using abrasive papers and finally by alumina suspension (1 μm).Thereafter, the specimen was placed in the bottom of high-purity alumina crucibles (Ф 5.5 mm O.D., Ф 4.5 mm I.D., 5.0 mm height) and located at the focal point of a goldcoated ellipsoidal chamber.The chamber was cleaned thoroughly by a cycle of vacuum (less than 4 Â 10 À5 torr) and purging with high-purity Ar (purity>99.9999%)passing a 300 °C Mg column.To further remove the oxygen in the Ar, a Ti foil was wrapped to surround the top surface of the alumina crucible during observation.A Type B (PtRh30-PtRh6%) thermocouple was used to measure the temperature at the bottom of the crucible.Furthermore, pure Ni and Fe foils were used to be melted in the current HT-CLSM to calibrate the surface temperature of liquid metal.It was found that an average difference within 10 °C was found in the interested temperature range between 1450 and 1500 °C.
Besides HT-CLSM, DSC was used to detect the solidification temperature of DSSs before the in situ observation measurement.Details of the instrumentation can be found in Ref. [27].In order to identify the phases (ferrite and austenite) and quantify the fractions, electron backscatter diffraction (EBSD) was used to characterize the sample surface after the HT-CLSM experiments.It was reported that the specimen after HT-CLSM could be used for EBSD analysis directly without polishing due to good surface quality, [28] and the aim was to correlate the observation images by HT-CLSM and phase identification directly.Field-emission SEM (JEOL JSM-7800 F) equipped with a Nordlys Nano (Oxford Instruments) detector was used.The microscope was operated at an acceleration voltage of 20 kV and with a step size of 1.5 μm.HKL Channel 5 software (Oxford Instruments) was used for the data processing.
Wavelength-dispersive spectrometry (WDS) implemented in an electron probe microanalyzer (EPMA) (JEOL JXA-8200 WD/ED) was used for the quantitation of the chemical analysis.Before the measurement, standard pure metallic samples were used for calibration.

Calculation Methods
Theoretical solidus and liquidus temperatures were calculated using Thermo-Calc 2022a [29] with a TCFE12 [30] database, to compare with the on-set and peak of the melting temperatures measured by HT-CLSM and DSC.Besides, the amounts and formation temperature of different phases and their compositions were calculated using Equilibrium and Scheil simulation modules.

In Situ Characterization of Solidification of DSSs 2507 and 3207
Solidification generally proceeds by the nucleation and subsequent growth of the solid phase in the liquid matrix.
The solidification sequence of stainless steels is usually predicted using the Fe-Cr-Ni phase diagram. [31]The sequence is related to the ratio of Ni equivalent and Cr equivalent (Cr eq =Ni eq ), which are used to simplify a multicomposition system for the Fe-Cr-Ni ternary system.Here, Ni eq and Cr eq can be estimated by Equation ( 1) and ( 2), where contents of elements are given in wt%.
According to the compositions given in Table 1 and Equation ( 1) and ( 2), the Cr eq =Ni eq ratios are 1.74 for DSS2507 and 1.58 for DSS3207, respectively.Thus, the solidification mode of these steels falls into the ferrite-austenite mode (FA): [32] In other words, the δ-ferrite first precipitates from the liquid followed by further cooling, then the three-phase reaction (L þ δ þ γ) at the terminal solidification stage, and the ferrite transforms to austenite in the solid state until it is completely dissolved below the solidus line.
DSC measurement results of solidification of DSSs 2507 and 3207 are presented in Figure 1a,b.The heating and cooling rates of the specimen between 1300 and 1500 °C were set as 4 °C min À1 .It is seen that the DSS3207 has a higher on-set and peak temperature than DSS 2507 due to the different alloying compositions, that is, 1469.9 and 1471.2 °C for DSS 2507 and 1479.2 and 1481.3 °C for DSS 3207.In general, the on-set temperature of solidification should be higher than the peak temperature.In the current case, the peak temperature is observed to be higher due to the supercooling effect on the temperature profile, a bump can be clearly observed.The on-set temperature represents the formation of δ-ferrite from the melt.The aim to perform DSC analysis is used to design the HT-CLSM observation experiments for studying phase formation kinetics during solidification.
Figure 2 illustrates the representative images of DSS 2507 during the solidification process.The average cooling rate of the solidification process was set as an average value of 4 °C min À1 , which was the same as the DSC measurement.As shown in Figure 2a, the fully liquid phase remained until around 1468.5 °C, which was very close to the temperature obtained by DSC measurement (i.e., 1469.9 °C).Thereafter, δ-ferrite cell with a circular or elliptical morphology started to form in the liquid when critical undercooling occurred, and it grew with decreasing temperature (Figure 2b).In the early stage of solidification, the shape of the δ-ferrite cell was not affected by other grains during growth, which was mainly related to the shape of the cell itself and the surrounding temperature gradient.The formation and growth of δ-ferrite cell crystals during the solidification of liquid steel was the growth and coarsening process of dendrite tips.While growing, new δ-ferrite nucleation would form with time and areas of the liquid pool became smaller (Figure 2c).As the remained liquid phase became less, the growth of the formed δ-cell crystals was affected by other surrounding grains, presenting an irregular shape.As the temperature continued to decrease, some of the larger ferrite phases which were close to each other would form a bridge and congregate with each other (Figure 2d).The merging phenomenon between the cellular dendrites has a certain randomness, which will lead to the irregular shape of the merged cellular dendrites.The possible reason for this phenomenon might be that the growth conditions are not the same for each dendrite during the solidification process.Some cellular dendrites grow fast and become coarser, part of the solutes expelled during growth will be arranged to the top of the slow-growing and smaller cell dendrites, so that its growth will be hindered, and finally it will be submerged and swallowed by other cell dendrites.It should be noted that the merged cells may belong to one single δ-ferrite grain or different grains.Simultaneously, the remaining liquid was divided into many irregular blocks.Thereafter, the interphase boundary that usually separated these phases was not observed (Figure 2e).
It can also be seen from Figure 2e that the γ-austenite appeared (L þ δ!γ) among δ-ferrite grain boundary.The formation temperature of the γ-austenite was around 1455.3 °C.As γ phase was the last to solidify, it should be formed because of the partitioning of Ni into the melt during solidification.As the solidification continued, the amount of γ phase precipitated at the δ-grain boundary gradually increased, as shown in Figure 2f.Generally, γ-austenite formed not only at grain boundaries but also within grains.However, the further δ!γ transformation process and the growth of γ phase were not the focus of the current study.It should be noted that some liquid still existed in the final solidification process during the HT-CLSM observation.
A similar phenomenon has also been reported in the IN718 superalloy, where the remained liquid areas were the most segregated regions. [14]he representative micrographs of DSS3207 during the solidification process at the same cooling rate of 4 °C min À1 in average are presented in Figure 3.The liquid steel was supercooled to 1482.7 °C when δ-ferrite phase started to form on the surface of the liquid steel and slightly grew (Figure 3a) and gradually began to grow with time (Figure 3b).The arrangement of the newly formed δ-ferrite cells showed a certain regularity, and the growth in a certain direction was affected by the surrounding new phases.The growth rate was significantly faster, which led to the irregular shape of the δ-phase cells.Compared with DSS2507, the growth of δ-ferrite was much faster.When the size of δ-phase cells grew to some extent, the merge of several cells into bigger ones occurred (Figure 3c).The merging of such cell dendrites has a certain randomness, which would lead to the irregular shape of the merged dendrites.The remaining liquid became less as the continuation of solidification (Figure 3d).The measurement was interrupted when the formation of δ-ferrite was completed.DSSs 2507 and 3207 have similar solidification paths, ferrite solidified in the form of cellular crystals due to the slow cooling rate of steel at the current conditions.Figure 4 shows the schematic diagram of δ-ferrite formation during the solidification process of DSSs.In the HT-CLSM experiment, the solid phase in the molten steel nucleates and grows from the bottom of the crucible, and the remaining liquid remains at the top surface.The observed round or elliptical δ-ferrites are dendrite tips, they continue to grow, and several smaller cellular crystals start to form larger crystals.However, according to the HT-CLSM observations, the formation and growth characteristics of δ-ferrite in these two steels are different.These can possibly due to the increase of Cr content influencing the physical parameters of the liquid steels; details will be discussed in the following section.
The area fraction of δ-ferrite on the free surface of each picture recorded in HT-CLSM observation was measured by ImageJ software, according to the difference of brightness and contrast between the δ-ferrite and the liquid.Figure 5 presents the relationship between the area fraction of δ-ferrite and time as well as temperature.The curve was fit using Avrami's Equation (3) [33] and the results are available in Figure 5a.Besides, the maximum and average equivalent diameters of δ-ferrite are also plotted against time.
where f δ is the area fraction of δ-ferrite, n is the Avrami coefficient, k is the overall growth rate constant, and t is time.
It is seen from Figure 5a that the area fraction of δ-ferrite gradually increases with time during the whole solidification process, and some liquid phase still exists during the late stage of solidification.The growth of δ-ferrite is slower at the beginning after δferrite precipitation due to the limited solute segregation.During this stage, the degree of undercooling is small, and the required nucleation energy (driving force) is high.Therefore, the nucleation rate of δ-ferrite is slow, and the volume of the δ-ferrite phase increases slowly.After a certain period, the growth rate of δ-ferrite greatly increases with time due to the higher  undercooling degree and higher nucleation rate, which is more obviously seen in the case of DSS 3207.Moreover, the growth rate of δ-ferrite in DSS 3207 is larger than that of DSS 2507.Finally, when the remaining liquid becomes less, the growth rate of δ-ferrite decreases.The maximum and average diameter changes of δ-ferrite show a similar tendency with that of the area fraction change of δ-ferrite in DSS 3207 (Figure 5b).In addition, the maximum grain size (diameter) of δ-ferrite during the final solidification stage is larger in DSS 3207 (max.460 μm) compared to those in DSS 2507 (max.310 μm).It can be seen that the maximum diameter of δ-ferrite significantly increases at about 60 s in DSS 3207 and 80 s in DSS 2507 after the formation of δ-ferrite, which corresponds to the point when some separate δ-ferrite phases merge into larger cells.It can be indicated that the faster growth of δ-ferrite can result in a larger size of δ-ferrite.
When it comes to the temperature curve, the area fraction of δ-ferrite significantly increases with decreasing temperature from around 1467 to 1455 °C in DSS 2507; then it slightly increases as the solidification proceeds (Figure 5c).A similar tendency can be applied for DSS 3207 in the temperature range from 1482 to 1475 °C.In addition, δ-ferrite starts to form much earlier in DSS 3207 than that in DSS 2507.Besides, DSS 3207 has a narrower temperature range for the δ-ferrite growth.It can be indicated that the increased Cr content can result in higher formation temperature and faster growth of δ-ferrite.This information can provide guidance for the actual casting process of different grades of DSSs.
The area fraction of δ-ferrite ( f δ ) as a function of time in DSSs 2507 and 3207 can be fit by Equation ( 4) and (5).
It is clear that the higher growth rate constant and Avrami coefficient can be obtained for DSS 3207, which corresponds to the higher growth rate of δ-ferrite compared to that in DSS 2507.According to Figure 5a, it is known that the transition temperature of the liquid to δ-ferrite is around 1467.8 °C in DSS 2507 and 1482.7 °C in DSS 2507, respectively.Moreover, the cooling rate during the solidification process is in average 4 °C min À1 , which means the area fraction of δ-ferrite depends on temperature in the following relationship by Equation ( 6) in DSS 2507 and (7) in DSS 3207.f δ,2507 ¼ 1 À exp À6.9 Â 10 À4 1467.8À T 0.07 1.5 (6)   f δ,3207 ¼ 1 À exp À7.4 Â 10 À8 1482.7 À T 0.07 3.4 (7)   In the casting processes, the quality of products is highly related to the grain structure during solidification, including grain size and morphology of microstructure. [34]Besides the kinetic analysis, the microstructure evolution in the solidified DSSs is investigated by EBSD, and the typical results of band contrast, phase map, and inverse pole figure (IPF) of DSS2507 after HT-CLSM observation are presented in Figure 6a-c.It is seen that the dominant microstructure is BCC phase (δ-ferrite).FCC (γ-austenite) with a much finer morphology forms either in grain boundary (GB) or intragranlarly.The fraction of FCC is much lower, that is 15%, which is a typical DSS microstructure after solidification without heat treatment.In general, the postheat treatment is needed to increase FCC fraction to achieve a phase balance between γ-austenite and δ-ferrite. [35]Since the grain size is very large after the solidification, the orientation of the GB austenite shows an almost the same growth direction.The distributions of each chemical element in the solidified grains of DSS are investigated; a typical EPMA analysis of DSS2507 is presented in Figure 7.It is seen that there is no obvious element segregation in each chemical map, and this could be due to the slow cooling rate.Cr and Fe seem to be enriched in BCC phase than in FCC, and Ni and N are enriched in FCC.DSS 3207 shows an exact same tendency as DSS 2507.In addition, there are few nitrides formation randomly; however, it is not the focus of the current work.In general, the element distribution in FCC and BCC seems to be normal  according to metallurgical considerations, since Fe and Cr stabilize BCC and Ni and N stabilize FCC.More theoretical analysis of element distribution could be provided in the next section.

Theoretical Consideration of Steel Grades on the Solidification Process
Equilibrium calculation was first applied for the solidification process of these two steel grades by Thermo-Calc software.In this case, it assumes that enough diffusion occurs in the solid phase by way of retaining an extremely low cooling rate and a homogeneous composition of both the solid and liquid phases.Moreover, the cooling rate is not slow enough for the solid phase to easily diffuse in most cases.It is known that various types of nonequilibrium interactions, linear or nonlinear dynamic processes, can be involved in the solidification process of steels.Therefore, the Scheil-Gulliver model within Thermo-Calc was also applied for the solidification calculation, assuming infinitely fast diffusion in the liquid phase and zero or limited diffusion in the solid phases.Figure 8 presents the simulation results for the phase fraction of DSSs 2507 and 3207 under the equilibrium state and the solidification process using Scheil-Gulliver module in Thermo-Calc.
According to the equilibrium calculation of phase diagrams presented in Figure 8a,c, the liquidus temperatures of DSS 2507 are about 1463 °C and that of DSS 3207 is about 1467 °C.The solidus temperatures of DSSs 2507 and 3207 are about 1370 and 1360 °C, respectively.It can be seen that the calculated temperature intervals (ΔT) of solidification of DSSs 2507 and 3207 are 93 and 107 °C, respectively.Furthermore, both two steels solidify with the primary formation of δ-ferrite (BCC).As the temperature decrease, the primarily formed δ-ferrite transforms into an γ-austenite (FCC).
In terms of the Scheil-Gulliver model calculations, the calculated solidification sequence is: which is consistent with the solidification characteristics of HT-CLSM images.Moreover, δ-ferrite first formed in liquid at 1465 °C, and the γaustenite phase precipitated at the δ-grain boundary when the solid fraction was about 0.82 at 1382 °C in DSS 2507.Besides, the respective temperatures are 1462 and 1367 °C in DSS 3207.The calculated liquidus temperatures are smaller than those according to the HT-CLSM and DSC experiments; this might be due to the different cooling rates.
It is evident according to the HT-CLSM results mentioned above that Cr has a significant effect on the solidification features of steels.Considering the different Cr contents in these two steels, phase equilibria for the composition of Fe-(20-36)% Cr-7% Ni-4% Mo-0.8% Mn-0.3% Si-0.02%C-0.4% N system was calculated in the temperature range 400-1600 °C and the results are presented in Figure 9a.It is clear that the single ferritic is the main phase at high temperatures regardless of the levels of Cr content.During the process of solidification, as the fraction of Cr increases, the BCC forms earlier, in another word, at a higher temperature.Moreover, the phase transition from BCC to FCC starts later at a much lower temperature compared to Scheil solidification process.With decreasing temperature, and under equilibrium conditions, the precipitation of sigma phase (σ) will proceed.However, the precipitation formation is neglected in the current experimental observation at a much higher temperature range; also they are hard to form in the actual solidification condition compared with the equilibrium state.The temperature is one important factor in the nucleation and growth of BCC phase.Therefore, the liquidus (T L ) and solidus (T S ) temperatures of steel based on the composition of the same system with different Cr contents were calculated using FactSage 7.3; the results are shown in Figure 9b.It can be seen that T L increases while T S slightly decreases first and then increases with increasing Cr content, and the increase of T L is greater than that of T S .As a result, the increased Cr content causes an extension of the solidification temperature range (ΔT = T L À T S ).A larger value of ΔT can result in a higher driving force for the growth of BCC phase; thus, the BCC phase more easily grew to larger sizes.
During the solidification process, the main alloying elements in steel will distribute in the liquid phase and solid phase.The solute distribution coefficients, that is, K-factor, reflecting the segregation of Cr, Ni, and Mo elements were calculated using the Scheil-Gulliver model with TCFE12 thermodynamic database in Thermo-Calc software.Initially, in the Scheil module, C i,s and C i,1 was calculated as a function of solid and liquid composition changes with temperature.The partitioning coefficient of each element is expressed by Equation ( 7) and (8).
where C i,s and C i,1 are the concentrations of alloying elements in the solid phase and liquid phase, respectively.C 0 i is the initial content of the element.f s is the solid fraction.
The solute partitioning coefficient K is defined as the dissolution tendency of the element in both liquid and solid phases.When K is less than 1, the elements tend to dissolve in the residual liquid phase.Conversely, the alloying elements dissolve in the solid phase.The calculated compositions in the solid ferrite phase and distribution coefficient K of the main alloying elements in the solidification process are shown in Figure 10.
It is found that Cr and Mo show a decreasing tendency while Ni shows an increasing tendency with the proceeds of solidification in both steels (Figure 10a).The elemental changes between the solid and liquid phase can be described by the solute partitioning coefficient K.The larger the difference between the K values and 1, the more severe segregation can occur.It can be seen in Figure 10b that the K values of Cr and Mo in both steels are larger than 1, which indicates that they tend to diffuse into the solid phase with decreasing temperature.Furthermore, the K value of Mo is larger than that of Cr, suggesting that the The L!δ transformation in these two steels belongs to diffusive transformation, and the higher Cr content results in a wider solidification temperature range (ΔT = T L À T S ).Thus, the temperature drops slowly, which is more beneficial for the full diffusion of various elements in the liquid steel.Therefore, it is easy to meet the energy requirements and composition requirements required for the phase change, and the phase transition is easy to occur in DSS 3207. [36]

Conclusion
Based on the in situ investigation using HT-CLSM and DSC, in combination with thermodynamic calculations and electron microscopy characterizations, the solidification of DSSs 2507 and 3207 was quantitatively studied.The obtained conclusions are summarized as follows.1) The δ-ferrite phase first appeared in liquid during the solidification, and the cellular δ-ferrite phase formed at the tip of the dendrite and gradually grew and coarsened, and part of the small δ-ferrite phase merged to form a large cellular crystal during the growth process.Then it is followed by the precipitation of γ-austenite phase at the δ-ferrite grain boundary and also inside the grain.2) The area fraction of δ-ferrite during solidification of 2507 and DSS 3207s can be expressed as a unction of time.The specific functions are f δ,2507 ¼ 1 À exp À6.9 Â 10 À4 Â t 1.5 ð Þ for DSS 2507 and f δ,3207 ¼ 1 À exp À7.4 Â 10 À8 Â t 3.4 ð Þ for DSS 3207.3) As the Cr content increased, the δ-ferrite can form earlier and at a higher temperature, the growth rate and size of δ-ferrite also increased.The increased Cr content can result in the increase of both the liquidus temperature (T L ) and solidus temperature (T S ), and T L increases greater than that of T S .The alloying distribution behavior of the solidified HT-CLSM sample was simulated by the Scheil-Gulliver module, which was in good agreement with the experimental result.

Figure 1 .
Figure 1.DSC curves of the solidification of a) DSS 2507 and b) 3207 samples.

Figure 4 .
Figure 4. Schematic diagram of cellular δ-ferrite phase formation during solidification of DSSs: a) cellular δ-ferrite phase growth in HT-CLSM, b) δ-ferrite phase precipitates from liquid and gradually grows up, c) δ-ferrite phase continues to grow and merges together.

Figure 5 .
Figure 5. a,b) Relationships of area fraction and diameter of δ-ferrite with time and c) relationships of area fraction of δ-ferrite with temperature.

Figure 6 .
Figure 6.EBSD analysis of a typical DSS2507 sample after solidification: a) band contrast image, b) phase map, and c) IPF.

Figure 7 .
Figure 7. EPMA analysis of a typical DSS2507 sample after solidification.

Figure 8 .
Figure 8. Equilibrium and Scheil-Gulliver module simulation result for solidification process of a,b) 2507 and c,d) DSS 3207.