Strain Evolution in Cold-Warm Forged Steel Components Studied by Means of EBSD Technique

Electron BackScatter Diffraction (EBSD) in conjunction with Field-Emission Environmental Scanning Electron Microscopy (FEG-ESEM) has been used to evaluate the microstructural and local plastic strain evolution in different alloys (AISI 1005, AISI 304L and Duplex 2205) deformed by a single-stage cold and warm forging process. The present work is aimed to describe the different behavior of the austenite and ferrite during plastic deformation as a function of different forging temperatures. Several topological EBSD maps have been measured on the deformed and undeformed states. Then, image quality factor, distributions of the grain size and misorientation have been analyzed in detail. In the austenitic stainless steel, the γ-phase has been found to harden more easily, then α-phase and γ-phase in AISI 1005 and in duplex stainless steel, sequentially. Compared to the high fraction of continuous dynamic recrystallized austenitic zones observed in stainless steels samples forged at low temperatures, the austenitic microstructure of samples forged at higher temperatures, 600–700 °C, has been found to be mainly characterized by large and elongated grains with some colonies of fine nearly-equiaxed grains attributed to discontinuous dynamic recrystallization.

In this scenario, stainless steels are an important class of alloys. Their importance is manifested in the plenitude of applications that rely on their use. The application of austenitic stainless steels in food, petrochemical and nuclear industries is due to their combination of mechanical and corrosion resistance. In particular, AISI 304L steel is widely used, not only for its high corrosion resistance but also for its excellent formability and mechanical behavior. Many researchers have studied the changes in 304L stainless steel. Its static plastic deformation and corresponding microstructural evolution was found different from dynamic loading conditions at high strain rate [5][6][7][8][9]. Another steel grade of great interest for forging industry is the Duplex Stainless Steel (DSS). DSS is a two-phase alloy (ferrite/austenite) which combines the properties of austenitic and ferritic stainless steels. The good combination of its mechanical properties and corrosion resistance makes it of great interest for a wide range of applications especially in the oil, chemical and power industry [10]. During the last years, in view of the great interest of forging industries on these materials, several studies on their formability were conducted. It is noted that its properties strongly depend on the microstructure and substructural changes of αand γ-phase during deformation under low and high strain rate conditions [11,12].
Grain boundary character plays a key role in the plastic deformation of polycrystalline materials and a beneficial combination of mechanical properties can be achieved by grain refinement. In particular, the mechanical properties of carbon and stainless steels can be improved by fine-grained structures [13][14][15][16]. Such materials do not undergo phase transformations within a wide temperature range, and small grain sizes can be produced by dynamic recrystallization (DRX) under warm or cold forging conditions [17,18]. Since size of the dynamically recrystallized grain sensitively depends on processing temperature, the fine-grained microstructures can be developed under warm deformation conditions, i.e., during plastic working at relatively low temperature (T = 0.5-0.7 T m with T m the melting temperature) [19]. Recently, two main DRX mechanisms have been found to operate in metallic materials with low stacking fault energy (SFE): discontinuous DRX (DDRX) and continuous DRX (CDRX). In the DDRX mechanism, the formation of a new grain structure results from the operation of a grain boundary bulging, namely grain boundary serration and migration consuming the strain hardened substructures [20]. The recrystallized structure can be achieved by using conventional metal working techniques consisting in recrystallized and work hardened component [15,21,22].
The other type is the continuous DRX, which operates mainly under conditions of warm working [23]. The new grains develop as a result of the gradual increase in the misorientations between the subgrains that are caused by the plastic deformation; thus, fine-grained materials cannot be produced by standard thermomechanical processing [17,19,20].
The present work is aimed to describe qualitatively and quantitatively the differences in the plastic behavior of ferrite and austenite during one-stage cold forging process to form a hex-head plug fitting used in thermo-hydraulic applications. The strain heterogeneities and microstructural evolution of γ-phase in AISI 304L and Duplex 2205 stainless steel during warm forging process at different temperatures (i.e., 400, 500, 600 and 700 • C) are also investigated. Finally, the strain hardening behavior of the steels at cold and warm working conditions is fully analyzed.

Materials and Methods
The chemical composition of the alloys analyzed (AISI 1005 (Wr. N. 1.0303), AISI 304L (Wr. N. 1.4307), and DDS 2205 (Wr. N. 1.4462)) are listed in Table 1. In the as-received conditions, the materials were obtained by continuous casting and then hot rolled down to a final bar diameter of 22 mm. AISI 304L and DDS 2205 steel bars were solution heat-treated at 1150 • C and 1050 • C, respectively, and water-quenched to avoid precipitation of secondary phases. Figure 1 shows the step-sequence of the analyzed one-stage forging process at different strokes. The process consisted of two forging phases: first, a compression to create the hex-head (named "A-phase"); and, second, a deep backward extrusion operation to form the "neck" of the plug fitting (named "B-phase"). Bottom punch was fixed during the forming cycle. Top punch and die were driven by press mechanism. Moreover, bottom die was floating and driven by the contact forces.
Materials 2017, 10,1441 3 of 17 Figure 1 shows the step-sequence of the analyzed one-stage forging process at different strokes. The process consisted of two forging phases: first, a compression to create the hex-head (named "A-phase"); and, second, a deep backward extrusion operation to form the "neck" of the plug fitting (named "B-phase"). Bottom punch was fixed during the forming cycle. Top punch and die were driven by press mechanism. Moreover, bottom die was floating and driven by the contact forces. 3D solid modeling of the workpiece (i.e., cylindrical billet, 18.3 mm height and 55 g weight) and tools were carried out by Pro/E ® (Needham, MA, USA) software and then imported into FORGE2011 ® (Mougins, France) numerical code.
Details about the numerical models such as materials rheology and friction conditions can be found in previous works [24,25]. In total, 550 cylindrical billets (50 in AISI 1005, 250 in AISI 304L and 250 in Duplex 2205) were used for cold and warm forging experimental tests. Samples were forged by using a 2453 kN single-station general-purpose mechanical knuckle press with 50 spm (stroke per minute).
For a detailed understanding of the effects caused by cold and warm forging processes on the alloy, metallographic longitudinal sections parallel to the compression z-axis (CA) were drawn from the cylindrical billets and forged samples at different temperatures ( Figure 2). The focus was set on the microstructural analysis of three areas corresponding to different strain levels, named zone A (no deformation), zone B (intermediate level of strain) and zone C (high strain level). The boundary between zones C and B was chosen, according to the numerical simulation, equal to effective strain of 0.6 ( Figure 5). Height reductions h (Figure 2d), defined as the ratio between the height of the deformed part (highlighted in Figure 2d) and the height of the initial billet before forging operations (Figure 2a), for each alloy at different forging temperatures are reported in Table 2.  3D solid modeling of the workpiece (i.e., cylindrical billet, 18.3 mm height and 55 g weight) and tools were carried out by Pro/E ® (Needham, MA, USA) software and then imported into FORGE2011 ® (Mougins, France) numerical code.
Details about the numerical models such as materials rheology and friction conditions can be found in previous works [24,25]. In total, 550 cylindrical billets (50 in AISI 1005, 250 in AISI 304L and 250 in Duplex 2205) were used for cold and warm forging experimental tests. Samples were forged by using a 2453 kN single-station general-purpose mechanical knuckle press with 50 spm (stroke per minute).
For a detailed understanding of the effects caused by cold and warm forging processes on the alloy, metallographic longitudinal sections parallel to the compression z-axis (CA) were drawn from the cylindrical billets and forged samples at different temperatures ( Figure 2). The focus was set on the microstructural analysis of three areas corresponding to different strain levels, named zone A (no deformation), zone B (intermediate level of strain) and zone C (high strain level). The boundary between zones C and B was chosen, according to the numerical simulation, equal to effective strain of 0.6 ( Figure 5). Height reductions h (Figure 2d), defined as the ratio between the height of the deformed part (highlighted in Figure 2d) and the height of the initial billet before forging operations (Figure 2a), for each alloy at different forging temperatures are reported in Table 2.
Materials 2017, 10, 1441 3 of 17 Figure 1 shows the step-sequence of the analyzed one-stage forging process at different strokes. The process consisted of two forging phases: first, a compression to create the hex-head (named "A-phase"); and, second, a deep backward extrusion operation to form the "neck" of the plug fitting (named "B-phase"). Bottom punch was fixed during the forming cycle. Top punch and die were driven by press mechanism. Moreover, bottom die was floating and driven by the contact forces. 3D solid modeling of the workpiece (i.e., cylindrical billet, 18.3 mm height and 55 g weight) and tools were carried out by Pro/E ® (Needham, MA, USA) software and then imported into FORGE2011 ® (Mougins, France) numerical code.
Details about the numerical models such as materials rheology and friction conditions can be found in previous works [24,25]. In total, 550 cylindrical billets (50 in AISI 1005, 250 in AISI 304L and 250 in Duplex 2205) were used for cold and warm forging experimental tests. Samples were forged by using a 2453 kN single-station general-purpose mechanical knuckle press with 50 spm (stroke per minute).
For a detailed understanding of the effects caused by cold and warm forging processes on the alloy, metallographic longitudinal sections parallel to the compression z-axis (CA) were drawn from the cylindrical billets and forged samples at different temperatures ( Figure 2). The focus was set on the microstructural analysis of three areas corresponding to different strain levels, named zone A (no deformation), zone B (intermediate level of strain) and zone C (high strain level). The boundary between zones C and B was chosen, according to the numerical simulation, equal to effective strain of 0.6 ( Figure 5). Height reductions h (Figure 2d), defined as the ratio between the height of the deformed part (highlighted in Figure 2d) and the height of the initial billet before forging operations (Figure 2a), for each alloy at different forging temperatures are reported in Table 2.   For optical investigations and micro-hardness measurements, AISI 1005 specimens were etched with 4% HNO 3 in ethanolic solution; AISI 304L was etched with a reagent for electrolytic etching (a mixture of 60% HNO 3 and 40% distilled water); and Duplex 2205 samples were etched with Beraha etching solution (10 mL HCl, 40 mL distilled water, 1 g K 2 S 2 O 5 ). The micro-hardness tester Vickers Leitz Wetzlar D-35578 (Leica, Wetzlar, Germany) was used to perform three micro-hardness profiles, as shown in Figure 2c,d. Measurements were carried out according to Standards ASTM E92-82 using a load of 100 g. Microstructural investigation was also carried out by using a FEI Quanta 250 scanning electron microscope (FEI, Hillsboro, OR, USA) equipped with an electron back scattering diffraction (EBSD) analyzer incorporating an orientation imaging microscopy (OIM) system (EDAX TSL software, version 5). The surfaces of the undeformed and cold-warm forged specimens were prepared by using a polishing solution of 0.05 µm colloidal silica suspension and then electropolished in an electrolytic etching solution (60 mL HClO 4 , 40 mL distilled water) at 20 • C to ensure the highest surface quality. Samples were placed in FEG-ESEM microscope (FEI, Hillsboro, OR, USA) immediately after preparation. To compare the strain levels of zone A and B, step and area size used in the EBSD scans were 50 nm and 300 × 300 µm 2 respectively; on the other hand, because of the different strain levels between zone B and C and thus different quality of the scanning micrographs, the comparison between zone A and B was made by using a step and area size of 70 nm and 150 × 150 µm 2 , respectively. The OIM images were subjected to clean-up procedures by setting a minimal confident index of 0.1. For the EBSD analysis a working distance in the range of 13-21 mm, a voltage eqaul to 20 kV, a beam current of 220 µA, an fps (frame per second) of 30 and a number of maps per zone equal to 3 were chosen.

Results and Discussion
3.1. Optical Microscope Observations (Forging Temperature, 20 • C) Figure 3 shows the alloys microstructures before and after the cold forging test. In the as-received state (Figure 3a-zone A), AISI 1005 is characterized by a ferritic microstructure with a low amount of pearlite and an average grain size of 21 µm; its hardness value was found to be equal to 128 ± 3 HV 0.1 . AISI 304L (Figure 3b-zone A) shows the typical austenitic microstructure with twin boundaries; initial values of average grain size and hardness were found to be 42 µm and 207 ± 4 HV 0.1 , respectively.
Finally, the grain size and hardness of the ferritic-austenitic stainless steel (DDS 2205, Figure 3c-zone A) were 9 µm and 245 ± 6 HV 0.1 , respectively. In the as-received state, a balanced amount of austenite-ferrite was observed.  For optical investigations and micro-hardness measurements, AISI 1005 specimens were etched with 4% HNO3 in ethanolic solution; AISI 304L was etched with a reagent for electrolytic etching (a mixture of 60% HNO3 and 40% distilled water); and Duplex 2205 samples were etched with Beraha etching solution (10 mL HCl, 40 mL distilled water, 1 g K2S2O5). The micro-hardness tester Vickers Leitz Wetzlar D-35578 (Leica, Wetzlar, Germany) was used to perform three micro-hardness profiles, as shown in Figure 2c,d. Measurements were carried out according to Standards ASTM E92-82 using a load of 100 g. Microstructural investigation was also carried out by using a FEI Quanta 250 scanning electron microscope (FEI, Hillsboro, OR, USA) equipped with an electron back scattering diffraction (EBSD) analyzer incorporating an orientation imaging microscopy (OIM) system (EDAX TSL software, version 5). The surfaces of the undeformed and cold-warm forged specimens were prepared by using a polishing solution of 0.05 μm colloidal silica suspension and then electropolished in an electrolytic etching solution (60 mL HClO4, 40 mL distilled water) at 20 °C to ensure the highest surface quality. Samples were placed in FEG-ESEM microscope (FEI, Hillsboro, OR, USA) immediately after preparation. To compare the strain levels of zone A and B, step and area size used in the EBSD scans were 50 nm and 300 × 300 μm 2 respectively; on the other hand, because of the different strain levels between zone B and C and thus different quality of the scanning micrographs, the comparison between zone A and B was made by using a step and area size of 70 nm and 150 × 150 μm 2 , respectively. The OIM images were subjected to clean-up procedures by setting a minimal confident index of 0.1. For the EBSD analysis a working distance in the range of 13-21 mm, a voltage eqaul to 20 kV, a beam current of 220 μA, an fps (frame per second) of 30 and a number of maps per zone equal to 3 were chosen Figure 3 shows the alloys microstructures before and after the cold forging test. In the as-received state (Figure 3a-zone A), AISI 1005 is characterized by a ferritic microstructure with a low amount of pearlite and an average grain size of 21 μm; its hardness value was found to be equal to 128 ± 3 HV0.1. AISI 304L (Figure 3b-zone A) shows the typical austenitic microstructure with twin boundaries; initial values of average grain size and hardness were found to be 42 μm and 207 ± 4 HV0.1, respectively.

Optical Microscope Observations (Forging Temperature, 20 °C)
Finally, the grain size and hardness of the ferritic-austenitic stainless steel (DDS 2205, Figure 3c-zone A) were 9 μm and 245 ± 6 HV0.1, respectively. In the as-received state, a balanced amount of austenite-ferrite was observed.

Micro-Hardness Evolution (Forging Temperature, 20 °C)
Micro-hardness profiles reveal different hardening intensities for cold forged tested steels ( Figure 4a). AISI 1005, due to the almost full presence of α-phase, shows a nearly homogeneous hardening behavior. The highest mean values of hardness can be observed on the area close to the contact surface between the top punch and the workpiece (Figure 4a). This can be correlated to the combination of the material elastic-plastic properties (low stain hardening coefficient) and the forging technique used. The cold forged stainless steel samples show an inhomogeneous hardening behavior with a hardness increase in zone C ( Figure 4a). For AISI 304L, the hardness values vary from 356 to 257 HV0.1; while, in the case of DSS, they are in the range of 399 to 282 HV0.1. The highest hardness properties of the ferritic-austenitic stainless steel is mainly associated to the higher mean values of hardness reached on the as-received state ((~245 ± 6) HV0.1). Furthermore, due to the higher strain hardening coefficients of stainless steels compared to low carbon steel, the deformation tends to localize in zone C forming a sort of barrier that prevents the material flow to extend into other zones of the mold ( Figure 5). This has been also confirmed by the distribution of the micro-hardness increase . % (Figure 4b) defined as: where . and .
are the mean values of micro-hardness derived from the three profiles reported in Figure 2c,d respectively, calculated at the same distance along the compression axis (CA). It is easy to show that stainless steels are characterized by a rapid hardness increase from zone B to zone C. Even if DSS is characterized by a higher hardness value in the as-received conditions, which makes it more difficult to forge, it has a lower tendency to harden than

Micro-Hardness Evolution (Forging Temperature, 20 • C)
Micro-hardness profiles reveal different hardening intensities for cold forged tested steels ( Figure 4a). AISI 1005, due to the almost full presence of α-phase, shows a nearly homogeneous hardening behavior. The highest mean values of hardness can be observed on the area close to the contact surface between the top punch and the workpiece (Figure 4a). This can be correlated to the combination of the material elastic-plastic properties (low stain hardening coefficient) and the forging technique used.

Micro-Hardness Evolution (Forging Temperature, 20 °C)
Micro-hardness profiles reveal different hardening intensities for cold forged tested steels ( Figure 4a). AISI 1005, due to the almost full presence of α-phase, shows a nearly homogeneous hardening behavior. The highest mean values of hardness can be observed on the area close to the contact surface between the top punch and the workpiece (Figure 4a). This can be correlated to the combination of the material elastic-plastic properties (low stain hardening coefficient) and the forging technique used. The cold forged stainless steel samples show an inhomogeneous hardening behavior with a hardness increase in zone C (Figure 4a). For AISI 304L, the hardness values vary from 356 to 257 HV0.1; while, in the case of DSS, they are in the range of 399 to 282 HV0.1. The highest hardness properties of the ferritic-austenitic stainless steel is mainly associated to the higher mean values of hardness reached on the as-received state ((~245 ± 6) HV0.1). Furthermore, due to the higher strain hardening coefficients of stainless steels compared to low carbon steel, the deformation tends to localize in zone C forming a sort of barrier that prevents the material flow to extend into other zones of the mold ( Figure 5). This has been also confirmed by the distribution of the micro-hardness increase . % (Figure 4b) defined as: where . and .
are the mean values of micro-hardness derived from the three profiles reported in Figure 2c,d respectively, calculated at the same distance along the compression axis (CA). It is easy to show that stainless steels are characterized by a rapid hardness increase from zone B to zone C. Even if DSS is characterized by a higher hardness value in the as-received conditions, which makes it more difficult to forge, it has a lower tendency to harden than The cold forged stainless steel samples show an inhomogeneous hardening behavior with a hardness increase in zone C (Figure 4a). For AISI 304L, the hardness values vary from 356 to 257 HV 0.1 ; while, in the case of DSS, they are in the range of 399 to 282 HV 0.1 . The highest hardness properties of the ferritic-austenitic stainless steel is mainly associated to the higher mean values of hardness reached on the as-received state ((~245 ± 6) HV 0.1 ). Furthermore, due to the higher strain hardening coefficients of stainless steels compared to low carbon steel, the deformation tends to localize in zone C forming a sort of barrier that prevents the material flow to extend into other zones of the mold ( Figure 5). This has been also confirmed by the distribution of the micro-hardness increase HV 0.1 (%) (Figure 4b) defined as: where HV 0.1 (de f ormed) and HV 0.1 (as−received) are the mean values of micro-hardness derived from the three profiles reported in Figure 2c,d respectively, calculated at the same distance along the compression axis (CA). It is easy to show that stainless steels are characterized by a rapid hardness increase from zone B to zone C. Even if DSS is characterized by a higher hardness value in the as-received conditions, which makes it more difficult to forge, it has a lower tendency to harden than AISI 304L steel ( Figure 4b). The highest strain hardening effect of the fully austenitic stainless steel is associated to the twin boundaries formation [26], crossing of slip planes [27], the increase of dislocation and stacking fault density in the deformed regions [28,29].
Materials 2017, 10, 1441 6 of 17 AISI 304L steel (Figure 4b). The highest strain hardening effect of the fully austenitic stainless steel is associated to the twin boundaries formation [26], crossing of slip planes [27], the increase of dislocation and stacking fault density in the deformed regions [28,29].

Electron Backscatter Diffraction Analysis (Forging Temperature, 20 °C)
Several statistical analyses have been performed on the EBSD data from each scanned area (zone A-C) to compare local plastic strain and grain evolution behavior of α-and γ-phase on the as-received and deformed steels considered.

Image Quality (IQ) Factor
At each measurement point in an OIM scan, a parameter quantifying the quality of the corresponding diffraction pattern is recorded. It is well known [30] that the image quality (IQ) is affected by residual strain in the diffracting volume. Thus, an indication of the distribution of strain in the material can be observed through an IQ map. For a large scanned area on a bulk sample, if the average IQ value is assumed to correspond to the overall strain measured mechanically, the local strain can be quantified by assuming a linear relationship between the IQ value and the local plastic strain. In this work, the quantitative evaluation method of the local plastic strain rate is based on the concept proposed by Tarasiuk et al. [31]. The idea is reported in Figure 6. In each graph, two normalized IQ distributions are plotted which correspond to the undeformed and deformed sample. The total area under each distribution curve is equal to one since it includes all the points within the area under investigation used to estimate deformed and undeformed material volume fractions. By superposing these two plots, two areas are detected: region X, which corresponds to all the points deformed without ambiguity, and region Y, which corresponds to the still undeformed points ( Figure 6). The area of region X is used to estimate the minimal deformed volume fraction (Vf min) according to the following equation (Equation (2)): where p(x) and q(x) are the normalized IQ distributions for deformed and undeformed samples, respectively.

Electron Backscatter Diffraction Analysis (Forging Temperature, 20 • C)
Several statistical analyses have been performed on the EBSD data from each scanned area (zone A-C) to compare local plastic strain and grain evolution behavior of αand γ-phase on the as-received and deformed steels considered.

Image Quality (IQ) Factor
At each measurement point in an OIM scan, a parameter quantifying the quality of the corresponding diffraction pattern is recorded. It is well known [30] that the image quality (IQ) is affected by residual strain in the diffracting volume. Thus, an indication of the distribution of strain in the material can be observed through an IQ map. For a large scanned area on a bulk sample, if the average IQ value is assumed to correspond to the overall strain measured mechanically, the local strain can be quantified by assuming a linear relationship between the IQ value and the local plastic strain. In this work, the quantitative evaluation method of the local plastic strain rate is based on the concept proposed by Tarasiuk et al. [31]. The idea is reported in Figure 6. In each graph, two normalized IQ distributions are plotted which correspond to the undeformed and deformed sample. The total area under each distribution curve is equal to one since it includes all the points within the area under investigation used to estimate deformed and undeformed material volume fractions. By superposing these two plots, two areas are detected: region X, which corresponds to all the points deformed without ambiguity, and region Y, which corresponds to the still undeformed points ( Figure 6). The area of region X is used to estimate the minimal deformed volume fraction (V f min ) according to the following equation (Equation (2)): where p(x) and q(x) are the normalized IQ distributions for deformed and undeformed samples, respectively.  Figure 6 shows the IQ normalized distributions as a function of phase (ferrite, austenite) corresponding to the as-received (zone A) and cold forged (i.e., zone C) materials (the IQ distributions measured in zone B are directly used in the subsequent calculations of Vf min fraction). For AISI 304L, the absolute IQ ranges were 250-1500 (zone C) and 250-750 (zone A). Similar value ranges have been obtained for AISI 1005 and Duplex 2205. In both phases, plastic strain leads to a shift of the normalized IQ distribution peak to lower values.
In Table 3, the Vf min fraction is estimated (Equation (2)) by assuming that the as-received state of steels is deformation free (Vf min = 0%). In Table 3, it can be easily noted that the highest values of Vf min fraction are reached on γ-phase. This is confirmed by the average misorientation angles determined by EBSD that has been observed (in the present work and in a previous study [32]) to increase faster in γ-phase than in α-phase.
Variations of 39% to 44% in α-phase for AISI 1005 and Duplex 2205 steel, and of 46% to 50% in γ-phase for AISI 304L and DDS 2205 are observed; the difference is attributed to the slightly higher  Figure 6 shows the IQ normalized distributions as a function of phase (ferrite, austenite) corresponding to the as-received (zone A) and cold forged (i.e., zone C) materials (the IQ distributions measured in zone B are directly used in the subsequent calculations of V f min fraction). For AISI 304L, the absolute IQ ranges were 250-1500 (zone C) and 250-750 (zone A). Similar value ranges have been obtained for AISI 1005 and Duplex 2205. In both phases, plastic strain leads to a shift of the normalized IQ distribution peak to lower values.
In Table 3, the V f min fraction is estimated (Equation (2)) by assuming that the as-received state of steels is deformation free (V f min = 0%). In Table 3, it can be easily noted that the highest values of V f min fraction are reached on γ-phase. This is confirmed by the average misorientation angles determined by EBSD that has been observed (in the present work and in a previous study [32]) to increase faster in γ-phase than in α-phase.
Variations of 39% to 44% in α-phase for AISI 1005 and Duplex 2205 steel, and of 46% to 50% in γ-phase for AISI 304L and DDS 2205 are observed; the difference is attributed to the slightly higher tendency to work-harden of single-phase steels (AISI 304L) as mentioned above. Moreover, DDS 2205 shows the highest values of V f min fraction for both phases in zone C. This may be correlated with the highest micro-hardness values previously measured on that area.

Microstructural Evolution (Forging Temperature, 20 • C)
A detailed analysis of the misorientation angle distributions by EBSD with the aim to estimate the amount of low-angle boundaries (LABs) (θ = 2 • -5 • ) [33] and high-angle boundaries (HABs) (θ > 15 • ) [34] has been carried out. Figures 7 and 8 show the histograms of LABs and HABs volume fractions (defined as the ratio between LABs length (or HABs) and the total grain boundary length, mm/mm * 100) as a function of the analyzed zone. A detailed analysis of the misorientation angle distributions by EBSD with the aim to estimate the amount of low-angle boundaries (LABs) (θ = 2°-5°) [33] and high-angle boundaries (HABs) (θ > 15°) [34] has been carried out. Figures 7 and 8 show the histograms of LABs and HABs volume fractions (defined as the ratio between LABs length (or HABs) and the total grain boundary length, mm/mm * 100) as a function of the analyzed zone.  It can be noted that the volume fraction of LABs increases and HABs decreases after a cold forging cycle in both phases. This can be attributed to the development of a sub-grains microstructure, characterized by dislocation walls, which forms during plastic deformation.
With the exception of AISI 1005, a higher amount of LABs has been detected in the α-phase compared to γ-phase in the as-received state. This upholds the hypothesis of incomplete recrystallization of α-phase. On the other hand, HABs prevail in γ-phase. They form through the fragmentation of elongated grains, as provided after complete recrystallization. The high difference among initial LABs and HABs volume fraction values of AISI 1005 was due to the supply conditions and the presence of a single-phase microstructure. tendency to work-harden of single-phase steels (AISI 304L) as mentioned above. Moreover, DDS 2205 shows the highest values of Vf min fraction for both phases in zone C. This may be correlated with the highest micro-hardness values previously measured on that area.

Microstructural Evolution (Forging Temperature, 20 °C)
A detailed analysis of the misorientation angle distributions by EBSD with the aim to estimate the amount of low-angle boundaries (LABs) (θ = 2°-5°) [33] and high-angle boundaries (HABs) (θ > 15°) [34] has been carried out. Figures 7 and 8 show the histograms of LABs and HABs volume fractions (defined as the ratio between LABs length (or HABs) and the total grain boundary length, mm/mm * 100) as a function of the analyzed zone.  It can be noted that the volume fraction of LABs increases and HABs decreases after a cold forging cycle in both phases. This can be attributed to the development of a sub-grains microstructure, characterized by dislocation walls, which forms during plastic deformation.
With the exception of AISI 1005, a higher amount of LABs has been detected in the α-phase compared to γ-phase in the as-received state. This upholds the hypothesis of incomplete recrystallization of α-phase. On the other hand, HABs prevail in γ-phase. They form through the fragmentation of elongated grains, as provided after complete recrystallization. The high difference among initial LABs and HABs volume fraction values of AISI 1005 was due to the supply conditions and the presence of a single-phase microstructure. It can be noted that the volume fraction of LABs increases and HABs decreases after a cold forging cycle in both phases. This can be attributed to the development of a sub-grains microstructure, characterized by dislocation walls, which forms during plastic deformation.
With the exception of AISI 1005, a higher amount of LABs has been detected in the α-phase compared to γ-phase in the as-received state. This upholds the hypothesis of incomplete recrystallization of α-phase. On the other hand, HABs prevail in γ-phase. They form through the fragmentation of elongated grains, as provided after complete recrystallization. The high difference among initial LABs and HABs volume fraction values of AISI 1005 was due to the supply conditions and the presence of a single-phase microstructure.
The variation of LAB and HAB fraction within a phase may depend on different parameters. In Figure 9, the LABs increase and HABs decrease is calculated as difference between the associated deformed (zones B and C) and undeformed (i.e., zone A) values reported in Figures 7 and 8. The variation of LAB and HAB fraction within a phase may depend on different parameters. In Figure 9, the LABs increase and HABs decrease is calculated as difference between the associated deformed (zones B and C) and undeformed (i.e., zone A) values reported in Figures 7 and 8. In Figure 8a, it can be observed that the highest increase of LABs volume fractions has been found on the α-phase of low carbon steel (zone C). This is due to the higher formability properties of this material compared to the stainless steels analyzed. It is also in good agreement with the high Vf min increase estimated on that phase for AISI 1005 (Table 3). On the other hand, the LABs fraction increase in zone B is higher for DDS 2205 compared to AISI 1005 steel as a consequence of the higher Vf min fraction value reached in that zone.
The increase of LABs volume fractions in γ-phase, on both deformed zones (~20-25%), is higher in DSS 2205 than AISI 304L. This is due to the building up of higher amount of dislocation microstructure composed by sub-grains induced by the higher levels of strain (Vf min fraction) observed on γ-phase of DDS 2205 compared to AISI 304L. Moreover, the highest increase of LABs volume fractions observed on zone C (high strain area) for the stainless steels is directly related to their high strain-hardening behavior (Figure 4b). A very similar histogram is observed for the HABs volume fractions decrease (Figure 8b).
A detailed statistical analysis of the misorientation distribution angles across the so-called special γ-grain boundaries, i.e., those having dense Coincident Site Lattice (CSL), was also carried out. By using EBSD analysis, the CSL numbers (Σ) were measured by means of the following equation: Σ = ℎ of lattice points in a unit cell of the generating lattice In face centered cubic metals and alloys with low stacking fault energy (SFE), most of these special boundaries are Ʃ3 or Ʃ3 n CSL boundaries related to twin boundaries. On zone A, about 59.2% and 65.9% of the HABs on AISI 304L and DSS 2205, respectively, display the first-order twin CSL orientation relationship Ʃ3 (within a deviation of 2°) characterized by 60° rotation about <111> axis. About 3.2% of boundaries in AISI 304L and 3.0% of boundaries in DDS 2205 appear to correspond to the second-order twins represented by Ʃ9 (38.9°/<011>) CSL orientation relationship.
On zone B, the γ-phase regions become slightly more elongated and locally fragmented compared to zone A. They show a tendency to become preferentially aligned at determined angles. The originally sharp peak in the γ-phase misorientation distribution centered on the ideal Ʃ3 CSL orientation relationship becomes broader and the portion of first-order twin boundaries among the HABs decrease to about 55.8% and 64.7% on AISI 304L and Duplex 2205 steel, respectively. In Figure 8a, it can be observed that the highest increase of LABs volume fractions has been found on the α-phase of low carbon steel (zone C). This is due to the higher formability properties of this material compared to the stainless steels analyzed. It is also in good agreement with the high V f min increase estimated on that phase for AISI 1005 (Table 3). On the other hand, the LABs fraction increase in zone B is higher for DDS 2205 compared to AISI 1005 steel as a consequence of the higher V f min fraction value reached in that zone.
The increase of LABs volume fractions in γ-phase, on both deformed zones (~20-25%), is higher in DSS 2205 than AISI 304L. This is due to the building up of higher amount of dislocation microstructure composed by sub-grains induced by the higher levels of strain (V f min fraction) observed on γ-phase of DDS 2205 compared to AISI 304L. Moreover, the highest increase of LABs volume fractions observed on zone C (high strain area) for the stainless steels is directly related to their high strain-hardening behavior (Figure 4b). A very similar histogram is observed for the HABs volume fractions decrease (Figure 8b).
A detailed statistical analysis of the misorientation distribution angles across the so-called special γ-grain boundaries, i.e., those having dense Coincident Site Lattice (CSL), was also carried out. By using EBSD analysis, the CSL numbers (Σ) were measured by means of the following equation: In face centered cubic metals and alloys with low stacking fault energy (SFE), most of these special boundaries are Σ3 or Σ3 n CSL boundaries related to twin boundaries. On zone A, about 59.2% and 65.9% of the HABs on AISI 304L and DSS 2205, respectively, display the first-order twin CSL orientation relationship Σ3 (within a deviation of 2 • ) characterized by 60 • rotation about <111> axis. About 3.2% of boundaries in AISI 304L and 3.0% of boundaries in DDS 2205 appear to correspond to the second-order twins represented by Σ9 (38.9 • /<011>) CSL orientation relationship.
On zone B, the γ-phase regions become slightly more elongated and locally fragmented compared to zone A. They show a tendency to become preferentially aligned at determined angles. The originally sharp peak in the γ-phase misorientation distribution centered on the ideal Σ3 CSL orientation relationship becomes broader and the portion of first-order twin boundaries among the HABs decrease to about 55.8% and 64.7% on AISI 304L and Duplex 2205 steel, respectively.
As the plastic strain increases (i.e., zone C), the austenite areas become more elongated. The broadening of the original Σ3 peak in the γ-phase misorientation distribution becomes more pronounced; the portion of the first-order twin boundaries among the HABs further decreases to about 14.3% and 12.3% on fully austenitic stainless steel and DDS 2205, respectively. The observed presence of second-order twin boundaries in the misorientation spectra is about 1.1% for AISI 304L and 2.0% for the duplex stainless steel. Table 4 summarizes the main fractions of CSL boundaries examined in the γ-phase; α-phase is practically free of them.
The present results show that pre-existing annealing twin regions within the austenite display a tendency to progressively rotate away from the ideal CSL orientation relationship during straining.
Thus, the corresponding originally straight coherent twin boundaries become gradually converted to general HABs during the deformation process. Similar results showing that such rotations appear to occur very early in the deformation process and might reach values of several tens of degrees at large strains have been reported by Cizek et al. [35].  Figure 10a,b shows micro-hardness profiles on stainless steel samples forged at different temperatures. The effect of increasing temperature tends to continuously decrease the micro-hardness profiles on each type of steel. This behavior is mainly associated with the higher dislocation mobility and lower dislocation density at higher forging temperatures [22]. All micro-hardness profiles confirm the presence of the high strain-hardened area around zone C under the forging impact at different temperatures. This effect is less pronounced at higher forging temperatures tests. As the plastic strain increases (i.e., zone C), the austenite areas become more elongated. The broadening of the original Ʃ3 peak in the γ-phase misorientation distribution becomes more pronounced; the portion of the first-order twin boundaries among the HABs further decreases to about 14.3% and 12.3% on fully austenitic stainless steel and DDS 2205, respectively. The observed presence of second-order twin boundaries in the misorientation spectra is about 1.1% for AISI 304L and 2.0% for the duplex stainless steel. Table 4 summarizes the main fractions of CSL boundaries examined in the γ-phase; α-phase is practically free of them.

Micro-Hardness Evolution on Zones B and C at Different Warm Forging Temperatures
The present results show that pre-existing annealing twin regions within the austenite display a tendency to progressively rotate away from the ideal CSL orientation relationship during straining.
Thus, the corresponding originally straight coherent twin boundaries become gradually converted to general HABs during the deformation process. Similar results showing that such rotations appear to occur very early in the deformation process and might reach values of several tens of degrees at large strains have been reported by Cizek et al. [35].  Figure 10a,b shows micro-hardness profiles on stainless steel samples forged at different temperatures. The effect of increasing temperature tends to continuously decrease the microhardness profiles on each type of steel. This behavior is mainly associated with the higher dislocation mobility and lower dislocation density at higher forging temperatures [22]. All micro-hardness profiles confirm the presence of the high strain-hardened area around zone C under the forging impact at different temperatures. This effect is less pronounced at higher forging temperatures tests.  In Figure 11a,b, a comparison between micro-hardness profiles obtained at 20 • C and at different forging temperatures has been made on each stainless steel in terms of micro-hardness decrease rate measurements HV 0.1 [%] estimated by Equation (1).

Micro-Hardness Evolution on Zones B and C at Different Warm Forging Temperatures
Materials 2017, 10, 1441 11 of 17 In Figure 11a,b, a comparison between micro-hardness profiles obtained at 20 °C and at different forging temperatures has been made on each stainless steel in terms of micro-hardness decrease rate measurements . % estimated by Equation (1). As shown in Figure 11b, Duplex 2205 has a lower tendency to decrease micro-hardness values at different forging temperatures than AISI 304L; moreover, the increase of temperature seems to drastically decrease the strain hardening effect in zone C. This behavior is not clear on AISI 304L samples forged at 400 and 500 °C, respectively, due to lower temperatures and similar micro-hardness profiles as compared to cold forged material. Figure 12 shows the trends of Vf min fractions for AISI 304L and Duplex 2205 stainless steel at different forging temperatures.  As shown in Figure 11b, Duplex 2205 has a lower tendency to decrease micro-hardness values at different forging temperatures than AISI 304L; moreover, the increase of temperature seems to drastically decrease the strain hardening effect in zone C. This behavior is not clear on AISI 304L samples forged at 400 and 500 • C, respectively, due to lower temperatures and similar micro-hardness profiles as compared to cold forged material. In Figure 11a,b, a comparison between micro-hardness profiles obtained at 20 °C and at different forging temperatures has been made on each stainless steel in terms of micro-hardness decrease rate measurements . % estimated by Equation (1). As shown in Figure 11b, Duplex 2205 has a lower tendency to decrease micro-hardness values at different forging temperatures than AISI 304L; moreover, the increase of temperature seems to drastically decrease the strain hardening effect in zone C. This behavior is not clear on AISI 304L samples forged at 400 and 500 °C, respectively, due to lower temperatures and similar micro-hardness profiles as compared to cold forged material. Figure 12 shows the trends of Vf min fractions for AISI 304L and Duplex 2205 stainless steel at different forging temperatures.  As can be seen in Figure 12, the V f min fractions are almost constant with an increase in forging temperature from 20 to 400 • C on zone C in both steels analyzed; they slightly decrease on zone B. The warm-working temperature of 400 • C gives not enough relevant alteration and evolution on γ-grains and sub-grains structures at different levels of strain (i.e., zones B and C). In zone B, with an increase in temperature from 400 to 700 • C, the V f min fractions are in the range of 38-52% and 10-49% in AISI 304L and DSS 2205, respectively. On the other hand, in the same range of temperatures, the V f min fractions vary from 47% to 68% and from 50% to 70% in the austenitic and dual-phase stainless steel, respectively. If this range of temperatures is considered, variations of 14% and 39% in V f min fractions are observed in zone B for AISI 304L and Duplex 2205, respectively. On the other hand, the raising of V f min fractions are almost constant and set at about 20% on zone C for both steels; it means that the γ-phase presents the same formability properties at different temperatures in both steels.

Minimum Deformed Volume Fractions
In zone B, the different increase of V f min fractions on the two steels analyzed is a direct consequence of the forging process used. Zone B is the last zone of the workpiece to be deformed and the lowest formability properties DDS 2205 steel involve lower V f min fractions at low temperatures as compared to AISI 304L steel. Due to the temperature increase effect, a reduction of V f min fractions gap between the steels analyzed is revealed and the formability properties of Duplex 2205 are greatly improved. In fact, in Figure 12, it is noted that γ-phase reaches almost similar V f min fractions at 700 • C in both steels considered.

Microstructural Evolution of γ-phase on Zones B and C at Different Warm Forging Temperatures
The deformation microstructures obtained after one-stage forging process at temperatures of 20 • C, 400 • C, 500 • C, 600 • C and 700 • C are shown in Figure 13 (black regions in the maps of Duplex 2205 are used for highlighting only the of γ-phase).
The single-stage warm forging process at all studied zones (i.e., Zones B and C) and temperatures results in slight γ-grain refinement from 20 to 500 • C for each stainless steel considered. A non-uniform fine grained structure evolves in the samples processed at these temperatures in zone B and C, as reported in Figure 14. The γ-grains size of AISI 304L (forging temperature from 20 to 500 • C) are in the range of 3671-2285 µm 2 and 267-77 µm 2 in zones B and C, respectively. In the same temperature range, the average γ-grains size of DDS 2205 changes from 116 to 58 µm 2 in zone B and from 25 to 12 µm 2 in zone C. The microstructure at these temperatures is characterized by some heterogeneities. In addition to the equiaxed fine dynamic recrystallized (DRX) grains, these microstructures contain a low amount of large-elongated γ-grains with irregular boundaries, which are the remainders of the original grains.
At 600 and 700 • C the microstructures are almost fully composed of fine nearly-equiaxed and large-elongated γ-grains. No significant DRX took place at those temperatures. Since γ-phase is characterized by a low value of stacking fault energy, the partially recrystallized microstructure at 600 and 700 • C is attributed to DDRX [36,37]. Figure 15 shows an examination of the boundary and sub-boundary misorientation distribution during deformation process at different forging temperatures. The fractions of LABs (i.e., 2 ≤ θ ≤ 5 • ) and HABs (i.e., 15 • ≤ θ ≤ 90 • ) tend to become almost similar at higher temperatures (i.e., 600 and 700 • C) for both deformed zones and stainless steels considered. The originally large differences between fractions of LABs on both deformed zones at 20 • C (i.e.,~12% for AISI 304L and 15% for Duplex 2205), decrease remarkably to about 6% and 2% on single-phase and dual-phase steel, respectively, at 700 • C. At this temperature, the fractions of LABs vary slightly from 66% to 72% (Figure 15a). On the other hand, in Figure 15b, a very similar behavior of HABs fractions are also revealed at different forging temperatures. At 20 • C the differences between fractions of HABs on both deformed zones are 30% for AISI 304L steel and 26% for Duplex 2205 steel, which decrease drastically to 1% and 5% on single-phase and dual-phase steel, respectively, at 700 • C.    As regards the specific misorientations that might be present within the stainless steels, the fraction of CSL boundaries has also been examined in the γ-phase. Resulting values are graphically shown in Figure 16. At lower temperatures (i.e., from 20 to 500 °C), the fraction of CSL boundaries decreases in both deformed zones of stainless steels analyzed. On zone B, Ʃ3 boundaries fraction decreases from 56% to 45% for AISI 304L and from 68% to 59% for DDS 2205. Moreover, it decreases from 14% to 9% and from 25% to 21% for austenitic and dual-phase stainless steel, respectively, on zone C. At the same time the percentage of total CSL boundaries decreases simultaneously on both deformed zones of stainless steels. On the other hand, the fractions of CSL boundaries at 600 and 700 °C increase as a consequence of the new grains nucleated as a result of local bulging of grain boundaries during the dynamic recrystallization (DRX) mechanism [38,39].   As regards the specific misorientations that might be present within the stainless steels, the fraction of CSL boundaries has also been examined in the γ-phase. Resulting values are graphically shown in Figure 16. At lower temperatures (i.e., from 20 to 500 °C), the fraction of CSL boundaries decreases in both deformed zones of stainless steels analyzed. On zone B, Ʃ3 boundaries fraction decreases from 56% to 45% for AISI 304L and from 68% to 59% for DDS 2205. Moreover, it decreases from 14% to 9% and from 25% to 21% for austenitic and dual-phase stainless steel, respectively, on zone C. At the same time the percentage of total CSL boundaries decreases simultaneously on both deformed zones of stainless steels. On the other hand, the fractions of CSL boundaries at 600 and 700 °C increase as a consequence of the new grains nucleated as a result of local bulging of grain boundaries during the dynamic recrystallization (DRX) mechanism [38,39]. As regards the specific misorientations that might be present within the stainless steels, the fraction of CSL boundaries has also been examined in the γ-phase. Resulting values are graphically shown in Figure 16. At lower temperatures (i.e., from 20 to 500 • C), the fraction of CSL boundaries decreases in both deformed zones of stainless steels analyzed. On zone B, Σ3 boundaries fraction decreases from 56% to 45% for AISI 304L and from 68% to 59% for DDS 2205. Moreover, it decreases from 14% to 9% and from 25% to 21% for austenitic and dual-phase stainless steel, respectively, on zone C. At the same time the percentage of total CSL boundaries decreases simultaneously on both deformed zones of stainless steels. On the other hand, the fractions of CSL boundaries at 600 and 700 • C increase as a consequence of the new grains nucleated as a result of local bulging of grain boundaries during the dynamic recrystallization (DRX) mechanism [38,39].

Conclusions
Compared to previous works, both cold and warm single-phase forging processes of different steels (AISI 1005, AISI 304L and DDS 2205) were investigated. To the best of authors knowledge, α and γ phases metallurgical behaviors at different temperatures were compared for the first time by means of the same investigation techniques in alloys characterized by the presence of both or one of them. Because of the different microstructure that characterized the alloys analyzed, a different behavior was observed between ferrite and austenite during the cold and warm forging process. Duplex stainless steels are known to be difficult to cold forging. The obtained results give new insights about the cold and warm forging process of the analyzed alloys and above all they show the possibility to cold and warm forging duplex stainless steels.
The main results can be summarized as follows:  The α-phase in AISI 1005 steel has a lower tendency to harden compared to γ-phase of AISI 304L steel. On the other hand, γ-phase tends to harden easier on austenitic than on duplex stainless steel. The highest strain hardening effect of γ-phase is associated to the crossing of slip planes, twin boundaries formation, the increase of dislocation and stacking fault density in the deformed regions.  During the cold forging process, the estimated deformed volume fraction is higher in the γphase compared to the α-phase. Furthermore, the γ-phase grains deform more homogeneously than the initially large α-phase grains.  Samples forged at 20 °C result in the development of fine grained microstructures. Low-angle boundaries (LABs) increase and high-angle boundaries (HABs) decrease, as a direct consequence of the dislocation microstructures formation.  The γ-phase microstructure which develops during single-stage forging from 400 to 500 °C is characterized by fine grained microstructure at different strain levels. The fraction of special boundaries decreases rapidly from 400 to 500 °C for both stainless steel analyzed. On the other hand, the microstructures of γ-phase detected at higher forging temperatures (i.e., 600 and 700 °C) are almost fully composed of large-elongated and fine nearly-equiaxed grains, which are considered to be discontinuous dynamic recrystallized (DRX). Similar values of LABs and HABs fractions and annealing twins formation are observed on the stainless steels investigated.

Conclusions
Compared to previous works, both cold and warm single-phase forging processes of different steels (AISI 1005, AISI 304L and DDS 2205) were investigated. To the best of authors knowledge, α and γ phases metallurgical behaviors at different temperatures were compared for the first time by means of the same investigation techniques in alloys characterized by the presence of both or one of them. Because of the different microstructure that characterized the alloys analyzed, a different behavior was observed between ferrite and austenite during the cold and warm forging process. Duplex stainless steels are known to be difficult to cold forging. The obtained results give new insights about the cold and warm forging process of the analyzed alloys and above all they show the possibility to cold and warm forging duplex stainless steels.
The main results can be summarized as follows: • The α-phase in AISI 1005 steel has a lower tendency to harden compared to γ-phase of AISI 304L steel. On the other hand, γ-phase tends to harden easier on austenitic than on duplex stainless steel. The highest strain hardening effect of γ-phase is associated to the crossing of slip planes, twin boundaries formation, the increase of dislocation and stacking fault density in the deformed regions.

•
During the cold forging process, the estimated deformed volume fraction is higher in the γ-phase compared to the α-phase. Furthermore, the γ-phase grains deform more homogeneously than the initially large α-phase grains. • Samples forged at 20 • C result in the development of fine grained microstructures. Low-angle boundaries (LABs) increase and high-angle boundaries (HABs) decrease, as a direct consequence of the dislocation microstructures formation.

•
The γ-phase microstructure which develops during single-stage forging from 400 to 500 • C is characterized by fine grained microstructure at different strain levels. The fraction of special boundaries decreases rapidly from 400 to 500 • C for both stainless steel analyzed. On the other hand, the microstructures of γ-phase detected at higher forging temperatures (i.e., 600 and 700 • C) are almost fully composed of large-elongated and fine nearly-equiaxed grains, which are considered to be discontinuous dynamic recrystallized (DRX). Similar values of LABs and HABs fractions and annealing twins formation are observed on the stainless steels investigated.