The Effect of Isothermal Multi-Directional Forging on the Grain Structure , Superplasticity , and Mechanical Properties of the Conventional Al – Mg-Based Alloy

The current study observed a grain structure evolution in the central part and periphery of the sample of an Al–Mg–Mn-based alloy during isothermal multidirectional forging (IMF) at 350 ◦C with a cumulative strain of 2.1–6.3 and a strain per pass of 0.7. A bimodal grain size distribution with areas of fine and coarse grains was observed after IMF and subsequent annealing. The grain structure, mechanical properties, and superplastic behavior of the samples subjected to IMF with a cumulative strain of 6.3 and the samples exposed to IMF with subsequent cold rolling were compared to the samples exposed to a simple thermo-mechanical treatment. The micro-shear bands were formed inside original grains after the first three passes. The fraction of recrystallized grains increased and the mean size decreased with an increasing cumulative strain from 2.1 to 6.3. Significant improvements of mechanical properties and superplasticity were observed due to the formation of a homogenous fine grain structure 4.8 μm in size after treatment including IMF and subsequent cold rolling.


Introduction
The effects of severe plastic deformation (SPD) to the microstructure and properties of aluminum alloys have hardly been studied during the last decade.Many SPD methods have been analyzed, and their advantages and disadvantages are known [1][2][3][4][5][6].Grain fragmentation and continuous dynamic/post-dynamic recrystallization [7,8] occur at SPD and provide an ultra-fine grain structure consisting of grains with high angle grain boundary misorientation [1,2,4,[6][7][8][9].As a result, ultra-fine grain aluminum alloys exhibit increased mechanical properties at room temperature [1,4,7,10] and superplasticity at elevated temperatures [11][12][13][14][15].The main disadvantage of the SPD technique is an increase in the cost of semi-finished products, which can be critical for many applications of aluminum-based alloys.Cost inefficiency delays the industrial application of many SPD methods.Under this point of view, the most cheap, simplest, and promising SPD technique is multi-directional, or multi-axial forging [16,17].This method provides significant grain refinement in various aluminum-based alloys and decreases the mean grain size to ~0.5-1 µm [18][19][20].The main multi-directional forging (MDF) problem is a strain inhomogeneity.As a result, an inhomogeneous grain structure forms in bulk material [21][22][23][24].The grain structure consists of very fine grains and coarse grains in pure Al subjected to multi-directional forging at room temperature [25,26].Similar grain bimodalities are observed in many aluminum-based alloys [9,21,[27][28][29][30][31].The final grain structure and its homogeneity are significantly affected by the deformation temperature, cumulative strain and alloy composition [2,[7][8][9][10]20].Montazeri-Pour et al. [25] concluded that a high imposed strain per pass by more shear planes with different orientations results in more homogeneous grain fragmentation.Wang et al. [32] found that grains were uniformly distributed after MDF at 350 • C with a cumulative strain of 3.6 in an Al-Cu-based alloy, due to the grain structure being stabilized by fine precipitates.Sidtikov et al. [20] observed grain structure homogeneity with a grain size of 1.2 µm after multi-directional forging under isothermal conditions (IMF) at 325 • C and cumulative strain of 8.3 in an Al-Mg-Sc-Zr alloy containing the fine precipitates of the Al 3 (Sc,Zr) phase.
The effect of MDF on the room temperature mechanical properties has been widely studied [19,21,33].It was found that the strength can be increased by several times [19] mostly due to grain boundary strengthening and the Hall-Petch relationship [1][2][3][4][5]19,25].The MDF-processed ultra-fine grain structure promises a decreased temperature and an increased strain rate of superplasticity for conventional aluminum-based alloys.Meanwhile, only several documents [34,35] have reported the superplastic deformation behavior of aluminum-based alloys subjected to multi-directional forging.Noda et al. [34] found low-temperature superplasticity in AA5083 type alloys subjected to MDF at 270 • C with cumulative strain of 6.The mean grain sizes were 0.8 and 10 µm in the samples annealed at 200 and 400 • C, respectively.Approximately 350% of elongation was achieved at 200 • C and 2.8 × 10 −4 s −1 , and the elongation exceeded 400% at 400 • C at the same strain rate.Alloy exhibited high strain rate superplasticity with an m-value of 0.39 and an elongation to failure of ~270% at a strain rate of 2.8 × 10 −2 s −1 and a temperature of 400 • C.Many studies have confirmed the effectiveness of MDF to improve of the superplastic properties of Mg-based alloys [36][37][38][39][40].There are insufficient data on the influence of MDF on the superplastic characteristics of other aluminum-based alloys exposed to MDF or MDF with subsequent rolling.This investigation focuses on the effect of IMF, with a large strain per pass of 0.7, on the grain structure evolution, the superplastic deformation behavior and the room temperature mechanical properties of an Al-Mg-Mn alloy (AA5000 type) subjected to IMF alone and to IMF with subsequent cold rolling.The samples of the same alloy exposed to a simple thermo-mechanical treatment, which included hot and cold rolling, were used as a reference.

Materials and Methods
A non-heat treatable conventional Al-Mg-based alloy [41] was studied in the present research.The alloy chemical composition is shown in Table 1.An ingot with a size of 250 × 120 × 30 mm 3 was cast using the semi-continuous casting method.Pure aluminum of 99.70%, pure magnesium of 99.95%, and pure zinc of 99.95%, along with master alloys Al-10% Mn, Al-4.3% Zr, Al-10% Cr, and Al-3.5% Ti were used for alloy preparation.All concentrations are given in wt.%.Melting was performed in a graphite-fireclay crucible using a laboratory electric resistance furnace (Naberterm S3, Nabertherm GmbH, Lilienthal, Germany).The temperature before casting was 740 • C. The casting cooling rate was ~3 K/min.
The grain structure, superplastic deformation behavior, and mechanical properties of the samples subjected to different thermo-mechanical treatments were compared.The sample after IMF (mode A), the sample subjected to IMF and cold rolling (mode B) and the sample subjected to hot rolling and cold rolling (mode C) were studied.The steps of the applied treatment modes are shown in Figure 1.Heat treatment was conducted in a Nabertherm N30/65A (Nabertherm GmbH, Lilienthal, Germany) furnace with an air atmosphere.Homogenization annealing was carried out in two stages at 430 °C for 5 h and at 480 °C for 3 h.
The isothermal multi-directional forging (IMF) was performed in a hydraulic press machine (Nordberg N3650E, Changshu Tongrun Auto Acessory Co., LTD, Changshu, China) with a maximum load of 50 tons.The scheme of the IMF deformation process is shown in Figure 2. The special dye was used to maintain a constant size in a one direction during forging.A constant sample geometry with an aspect ratio of 1 × 1 × 2 mm 3 was provided after each IMF pass.The sample size was 56 × 28 × 28 mm 3 .IMF was performed at a temperature of 350 °C and an initial strain rate of 6 × 10 −3 s −1 .The samples were maintained at 350 °C for 20 min before starting each compression.A strain of 0.7 per pass and a cumulative strain of ∑e = 2.1 per cycle were applied [42].A graphite-based lubricant (Liqui Moly, Ulm, Germany) was used to decrease the friction between the die elements and the sample.Hot and cold rolling were carried out on a laboratory rolling mill V-3P (GMT, Saint-Petersburg, Russia) with a roll diameter of 200 mm.The hot rolling reduction was 2.14 (88%) and the cold rolling reduction was 1.1 (67%).
An Axiovert 200 MMAT (Carl Zeiss, Oberkochen, Germany) optical light microscope (LM) in a polarized light regime was used for the grain structure analysis.Specimens for microstructure study were prepared by mechanical grinding on SiC papers and polishing on a colloidal silica suspension using a Struers LaboPol-5 (Struers APS, Ballerup, Denmark) machine.Subsequent anode oxidizing was performed in a Baker solution at 20 V. The grain structure was controlled after various IMF passes and thermo-mechanical treatment including IMF and hot and cold rolling.The microstructures of the IMF processed samples were analyzed in the near-surface layer (see Figure 2, Z-X plane after the first pass, Y-Z after the second pass and X-Y plane after the third pass) to study a Heat treatment was conducted in a Nabertherm N30/65A (Nabertherm GmbH, Lilienthal, Germany) furnace with an air atmosphere.Homogenization annealing was carried out in two stages at 430 • C for 5 h and at 480 • C for 3 h.
The isothermal multi-directional forging (IMF) was performed in a hydraulic press machine (Nordberg N3650E, Changshu Tongrun Auto Acessory Co., LTD, Changshu, China) with a maximum load of 50 tons.The scheme of the IMF deformation process is shown in Figure 2. The special dye was used to maintain a constant size in a one direction during forging.A constant sample geometry with an aspect ratio of 1 × 1 × 2 mm 3 was provided after each IMF pass.The sample size was 56 × 28 × 28 mm 3 .IMF was performed at a temperature of 350 • C and an initial strain rate of 6 × 10 −3 s −1 .The samples were maintained at 350 • C for 20 min before starting each compression.A strain of 0.7 per pass and a cumulative strain of ∑e = 2.1 per cycle were applied [42].A graphite-based lubricant (Liqui Moly, Ulm, Germany) was used to decrease the friction between the die elements and the sample.Heat treatment was conducted in a Nabertherm N30/65A (Nabertherm GmbH, Lilienthal, Germany) furnace with an air atmosphere.Homogenization annealing was carried out in two stages at 430 °C for 5 h and at 480 °C for 3 h.
The isothermal multi-directional forging (IMF) was performed in a hydraulic press machine (Nordberg N3650E, Changshu Tongrun Auto Acessory Co., LTD, Changshu, China) with a maximum load of 50 tons.The scheme of the IMF deformation process is shown in Figure 2. The special dye was used to maintain a constant size in a one direction during forging.A constant sample geometry with an aspect ratio of 1 × 1 × 2 mm 3 was provided after each IMF pass.The sample size was 56 × 28 × 28 mm 3 .IMF was performed at a temperature of 350 °C and an initial strain rate of 6 × 10 −3 s −1 .The samples were maintained at 350 °C for 20 min before starting each compression.A strain of 0.7 per pass and a cumulative strain of ∑e = 2.1 per cycle were applied [42].A graphite-based lubricant (Liqui Moly, Ulm, Germany) was used to decrease the friction between the die elements and the sample.Hot and cold rolling were carried out on a laboratory rolling mill V-3P (GMT, Saint-Petersburg, Russia) with a roll diameter of 200 mm.The hot rolling reduction was 2.14 (88%) and the cold rolling reduction was 1.1 (67%).
An Axiovert 200 MMAT (Carl Zeiss, Oberkochen, Germany) optical light microscope (LM) in a polarized light regime was used for the grain structure analysis.Specimens for microstructure study were prepared by mechanical grinding on SiC papers and polishing on a colloidal silica suspension using a Struers LaboPol-5 (Struers APS, Ballerup, Denmark) machine.Subsequent anode oxidizing was performed in a Baker solution at 20 V. The grain structure was controlled after various IMF passes and thermo-mechanical treatment including IMF and hot and cold rolling.The microstructures of the IMF processed samples were analyzed in the near-surface layer (see Figure 2, Z-X plane after the first pass, Y-Z after the second pass and X-Y plane after the third pass) to study a Hot and cold rolling were carried out on a laboratory rolling mill V-3P (GMT, Saint-Petersburg, Russia) with a roll diameter of 200 mm.The hot rolling reduction was 2.14 (88%) and the cold rolling reduction was 1.1 (67%).
An Axiovert 200 MMAT (Carl Zeiss, Oberkochen, Germany) optical light microscope (LM) in a polarized light regime was used for the grain structure analysis.Specimens for microstructure study were prepared by mechanical grinding on SiC papers and polishing on a colloidal silica suspension using a Struers LaboPol-5 (Struers APS, Ballerup, Denmark) machine.Subsequent anode oxidizing was performed in a Baker solution at 20 V. The grain structure was controlled after various IMF passes and thermo-mechanical treatment including IMF and hot and cold rolling.The microstructures of the IMF processed samples were analyzed in the near-surface layer (see Figure 2, Z-X plane after the first pass, Y-Z after the second pass and X-Y plane after the third pass) to study a grain structure in the periphery of the samples.The middle cross section (Z/2, X-Y plane in Figure 2 after the third pass) was analyzed to study the grain structure in the central part of the samples.
Tensile tests were performed at elevated temperatures of 500 and 540 • C on a Walter + Bay LFM100 test machine (Walter + Bai AG, Löhningen, Switzerland).Two test types were used: (1) step-by-step increase of the strain rate in a range of 8 × 10 −5 to 1 × 10 −1 s −1 and (2) constant strain rate tests at 1 × 10 −3 , 5 × 10 −3 and 1 × 10 −2 s −1 .The samples had a gauge size of F 0 = 6 × 1 mm 2 and a length of l 0 = 14 mm (l 0 = 5.65 √ F 0 ).The mechanical properties at room temperature were analyzed using a uniaxial tensile test on a Zwick Z250 (Zwick Roell Group, Ulm, Germany) test machine.A yield strength (YS), an ultimate tensile strength (UTS) and an elongation to failure were determined.The "dog bone" type samples had a gauge size of F 0 = 3 × 1 mm 2 and a length of L 0 = 10 mm (according to the ISO 6892-1 standard).The tensile tests were performed at traverse rate of 4 mm/min.Three samples per state were studied for room and elevated temperature tests.
Melting, heat and deformation treatments, and mechanical tests were performed in air atmosphere.The liner intercept method was used to determine a mean grain size.Three samples were studied for each state.The areas of coarse and fine grains were measured separately.The error bars were calculated as E = S × t/ √ N, were S is a standard deviation, t is a Student coefficient for confidence probability 0.95, and N is an amount of measurements (N ≈ 100 for coarse grained areas and N varied in a range of 300-500 for fine grained areas.

Grain Structure Evolution at IMF
The initial grain structure after homogenization annealing was equiaxed with a mean grain size of 143 ± 20 µm (Figure 3a).The grain structure evolution after IMF at 350 • C in the near-surface layer is shown at Figure 3b-f.The first IMF pass with a strain of 0.7 made the grains slightly elongated (Figure 3b), and the grain structure was transformed to the banded structure during the second and third IMF passes (Figure 3).The micro-shear bands (MSBs) [7,20] formed after the second IMF pass and their density increased after the third pass due an increase in the cumulative strain to 2.1.There was no recrystallization effect on the periphery of the sample at a cumulative strain of 2.1 (one cycle).An increase in the cumulative strain to 4.2 (two cycles) provided the dynamic/post-dynamic recrystallization [7,43].New equiaxed grains formed on the initial grain boundaries in the near-surface layer.The fraction of new recrystallized grains increased from 20 to 43%, and the grain size slightly decreased from 3.5 ± 0.1 to 2.9 ± 0.1 µm, with an increase in the cumulative strain from 4.2 to 6.3.The grain structure on the periphery of the samples remained non-uniform at ∑e = 6.3.
The microstructure of the central section of the samples was studied after cumulative strains of 2.1 (one cycle), 4.2 (two cycles), and 6.3 (three cycles) (Figure 4).The fine recrystallized grains were formed only on the initial grain boundaries at ∑e = 2.1 (Figure 4a).A recrystallized fraction occupied the larger areas at cumulative strains of 4.2 and 6.3 (Figure 4b,c).The volume fraction of recrystallized grains increased to 31% at ∑e = 4.2 and to 57% at ∑e = 6.3.As a result, a bimodal grain size distribution was observed in the central cross-section after IMF at 350 • C and ∑e = 6.3.Similar grain bimodality was observed after multi-directional forging of various aluminum alloys [9,[27][28][29][30][31] and after IMF with a lower strain per pass of 0.4 and a cumulative strain of ∑e = 6.1 in the studied Al-Mg-Mn alloy [43].
The mean size of the new recrystallized grains insignificantly changed with an increase in the cumulative strain.The mean grain size in the recrystallized area decreased from 1.6 ± 0.1 µm at ∑e = 2.1 to 1.2 ± 0.1 µm at ∑e = 6.3 in the central section.It is notable that grains in the central section were 2.2-2.4 times finer than those in the periphery of the sample.
In comparison, the grain structure of the sample hot-rolled at 350 the non-equiaxed recrystallized grains in the rolling direction varied within a range of 11 to 138 µm (Figure 4d). is shown at Figure 3b-f.The first IMF pass with a strain of 0.7 made the grains slightly elongated (Figure 3b), and the grain structure was transformed to the banded structure during the second and third IMF passes (Figure 3).The micro-shear bands (MSBs) [7,20] formed after the second IMF pass and their density increased after the third pass due an increase in the cumulative strain to 2.1.There was no recrystallization effect on the periphery of the sample at a cumulative strain of 2.1 (one cycle).An increase in the cumulative strain to 4.2 (two cycles) provided the dynamic/post-dynamic recrystallization [7,43].New equiaxed grains formed on the initial grain boundaries in the nearsurface layer.The fraction of new recrystallized grains increased from 20 to 43%, and the grain size slightly decreased from 3.5 ± 0.1 to 2.9 ± 0.1 µm, with an increase in the cumulative strain from 4.2 to 6.3.The grain structure on the periphery of the samples remained non-uniform at ∑e = 6.3.The microstructure of the central section of the samples was studied after cumulative strains of 2.1 (one cycle), 4.2 (two cycles), and 6.3 (three cycles) (Figure 4).The fine recrystallized grains were formed only on the initial grain boundaries at ∑e = 2.1 (Figure 4a).A recrystallized fraction occupied the larger areas at cumulative strains of 4.2 and 6.3 (Figure 4b,c).The volume fraction of recrystallized grains increased to 31% at ∑e = 4.2 and to 57% at ∑e = 6.3.As a result, a bimodal grain size distribution was observed in the central cross-section after IMF at 350 °C and ∑e = 6.3.Similar grain bimodality was observed after multi-directional forging of various aluminum alloys [9,[27][28][29][30][31] and after IMF with a lower strain per pass of 0.4 and a cumulative strain of ∑e = 6.1 in the studied Al-Mg-Mn alloy [43].
The mean size of the new recrystallized grains insignificantly changed with an increase in the cumulative strain.The mean grain size in the recrystallized area decreased from 1.6 ± 0.1 µm at ∑e = 2.1 to 1.2 ± 0.1 µm at ∑e = 6.3 in the central section.It is notable that grains in the central section were 2.2-2.4 times finer than those in the periphery of the sample.
In comparison, the grain structure of the sample hot-rolled at 350 °C (e = 2.14) consisted of nonrecrystallized elongated grains and several non-equiaxed recrystallized grains.The liner size of the non-equiaxed recrystallized grains in the rolling direction varied within a range of 11 to 138 µm (Figure 4d).

Grain Structure Evolution of IMF Proceeded Samples at Annealing in a Temperature Range of 450-540 °C (Mode A)
Samples subjected to the IMF with ∑e = 6.3 (mode A) were annealed at 450, 500, and 540 °C for 30 min.The solidus temperature of an alloy is 560 °C [41].The grain structure exhibited bimodality after annealing in a studied temperature range (Figure 5).The grain size distributions in a fine grain area are shown in Figure 5h.The mean grain size in a fine-grained area increased from 1.2 ± 0.1 after IMF to 1.9 ± 0.1, 5.0 ± 0.1, and 9.4 ± 0.1 µm after annealing at 450, 500, and 540 °C, respectively.Coarse grains varied in a range of 13-150 µm at all studied temperatures (Figure 5i).Samples subjected to the IMF with ∑e = 6.3 (mode A) were annealed at 450, 500, and 540 • C for 30 min.The solidus temperature of an alloy is 560 • C [41].The grain structure exhibited bimodality after annealing in a studied temperature range (Figure 5).The grain size distributions in a fine grain area are shown in Figure 5h.The mean grain size in a fine-grained area increased from 1.2 ± 0.1 after Metals 2019, 9, 33 6 of 12 IMF to 1.9 ± 0.1, 5.0 ± 0.1, and 9.4 ± 0.1 µm after annealing at 450, 500, and 540 • C, respectively.Coarse grains varied in a range of 13-150 µm at all studied temperatures (Figure 5i).The dynamic recrystallization partially occurred in the places where a critical strain was reached at IMF (Figure 4c).The new grains were formed on the original grain boundaries and in the MS bands.The fraction of recrystallized equiaxed grains insignificantly increased at annealing (Figure 5).Therefore, in most of the non-recrystallized grains, the strain at IMF did not exceed the critical strain value required to start the recrystallization at subsequent annealing.As a result, the grain structure in the IMF processed sample and the sample after IMF and subsequent annealing exhibited significant inhomogeneity.The grain growth was significant especially at the sub-solidus temperature of 540 °C.The mean grain size in the recrystallized volume increased by ~8 times.

Grain Structure of the Sheets (Modes B and C)
The grain structures of the sample subjected to IMF and cold rolling and the sample subjected to hot and cold rolling (mode C, Figure 1) after recrystallization annealing at 500 °C (mode B, Figure 1) are presented in Figure 6.The as-annealed samples processed by modes B (Figure 6a) and C (Figure 4b) demonstrated high homogeneity.The grain size distributions are shown in Figure 6c and the grain parameters are presented in Table 2.The mean grain size was 4.8 ± 0.1 µm in the sample pretreated by IMF (mode B).The sample exposed to simple thermo-mechanical treatment exhibited less equiaxed grains (mode C).The mean grain size was 8.0 ± 0.3 µm.The grain size varied in a range of 2 to 13 µm in mode B and from 2 to 26 µm in mode C (Figure 6c).The fraction of fine grains was significantly higher in the sample processed by mode B. The dynamic recrystallization partially occurred in the places where a critical strain was reached at IMF (Figure 4c).The new grains were formed on the original grain boundaries and in the MS bands.The fraction of recrystallized equiaxed grains insignificantly increased at annealing (Figure 5).Therefore, in most of the non-recrystallized grains, the strain at IMF did not exceed the critical strain value required to start the recrystallization at subsequent annealing.As a result, the grain structure in the IMF processed sample and the sample after IMF and subsequent annealing exhibited significant inhomogeneity.The grain growth was significant especially at the sub-solidus temperature of 540 • C. The mean grain size in the recrystallized volume increased by ~8 times.

Grain Structure of the Sheets (Modes B and C)
The grain structures of the sample subjected to IMF and cold rolling and the sample subjected to hot and cold rolling (mode C, Figure 1) after recrystallization annealing at 500 • C (mode B, Figure 1) are presented in Figure 6.The as-annealed samples processed by modes B (Figure 6a) and C (Figure 4b) demonstrated high homogeneity.The grain size distributions are shown in Figure 6c and the grain parameters are presented in Table 2.The mean grain size was 4.8 ± 0.1 µm in the sample pre-treated by IMF (mode B).The sample exposed to simple thermo-mechanical treatment exhibited less equiaxed grains (mode C).The mean grain size was 8.0 ± 0.3 µm.The grain size varied in a range of 2 to 13 µm in mode B and from 2 to 26 µm in mode C (Figure 6c).The fraction of fine grains was significantly higher in the sample processed by mode B.  For mode B-processed samples, the liner part of the stress-strain rate curves corresponded to the superplastic behavior with m > 0.3 being observed in a range of 1 × 10 −4 to 2 × 10 −2 s −1 at 500 °C and in a range of 1 × 10 −3 to 4 × 10 −2 s −1 at 540 °C (red curves in Figure 7).Mode C provided lower superplastic strain rates and m-values (blue curves in Figure 7).The samples exposed to IMF (mode A) exhibited m > 0.3 at strain rates lower than 2 × 10 −3 s −1 (black curves in Figure 7) due to significant grain structure inhomogeneity after IMF.For mode B-processed samples, the liner part of the stress-strain rate curves corresponded to the superplastic behavior with m > 0.3 being observed in a range of 1 × 10 −4 to 2 × 10 −2 s −1 at 500 °C and in a range of 1 × 10 −3 to 4 × 10 −2 s −1 at 540 °C (red curves in Figure 7).Mode C provided lower superplastic strain rates and m-values (blue curves in Figure 7).The samples exposed to IMF (mode A) exhibited m > 0.3 at strain rates lower than 2 × 10 −3 s −1 (black curves in Figure 7) due to significant grain structure inhomogeneity after IMF.For mode B-processed samples, the liner part of the stress-strain rate curves corresponded to the superplastic behavior with m > 0.3 being observed in a range of 1 × 10 −4 to 2 × 10 −2 s −1 at 500 • C and in a range of 1 × 10 −3 to 4 × 10 −2 s −1 at 540 • C (red curves in Figure 7).Mode C provided lower superplastic strain rates and m-values (blue curves in Figure 7).The samples exposed to IMF (mode A) exhibited m > 0.3 at strain rates lower than 2 × 10 −3 s −1 (black curves in Figure 7) due to significant grain structure inhomogeneity after IMF.Constant strain rate tests were carried out at strain rates of 1 × 10 −3 , 5 × 10 −3 and 1 × 10 −2 s −1 (Figure 8).The flow stress was lower at 540 • C as compared with at 500 • C for all studied modes and strain rates (Figure 8a-c).The samples exposed to IMF (mode A) demonstrated non-superplastic behavior with necking and an elongation to failure of less than 150% due to the presence of a non-uniform grain structure after IMF.
Metals 2019, 9 FOR PEER REVIEW 8 Constant strain rate tests were carried out at strain rates of 1 × 10 −3 , 5 × 10 −3 and 1 × 10 −2 s −1 (Figure 8).The flow stress was lower at 540 °C as compared with at 500 °C for all studied modes and strain rates (Figure 8a-c).The samples exposed to IMF (mode A) demonstrated non-superplastic behavior with necking and an elongation to failure of less than 150% due to the presence of a non-uniform grain structure after IMF.
Mode B provided lower stress values and a more stable steady stage as compared with mode C due to a finer grain structure.A maximum elongation of 500% and a stress value below 7 MPa were observed in the samples processed by IMF and cold rolling (Mode B) at 540 °C and 1 × 10 −3 s −1 .This mode provided high strain rate superplasticity in the studied conventional alloy.At 540 °C and 1 × 10 −2 s −1 , the mean elongation to failure exceeded 340% (Figure 8d) and m = 0.6 (Figure 7b).

Room Temperature Tensile Tests
The room temperature mechanical properties of the samples exposed to modes A, B, and C and the as-annealed samples are presented in Table 3. IMF provided a YS of 300 MPa and a UTS of 405 MPa.However, a typical YS value is 170-180 MPa and a typical UTS value is 300-320 MPa for conventionally proceeded samples [44], and mode C-processed samples exhibited similar values.Despite the inhomogeneous grain structure, samples exposed to IMF exhibited 17% elongation, which is close to the values of conventionally processed sheets.The increased strength is explained by the presence of an ultrafine grain structure with ~50% of fine grains being 1.2-2.9µm in size.The rolling at room temperature increased the YS to 500 MPa, Mode B provided lower stress values and a more stable steady stage as compared with mode C due to a finer grain structure.A maximum elongation of 500% and a stress value below 7 MPa were observed in the samples processed by IMF and cold rolling (Mode B) at 540 • C and 1 × 10 −3 s −1 .This mode provided high strain rate superplasticity in the studied conventional alloy.At 540 • C and 1 × 10 −2 s −1 , the mean elongation to failure exceeded 340% (Figure 8d) and m = 0.6 (Figure 7b).

Room Temperature Tensile Tests
The room temperature mechanical properties of the samples exposed to modes A, B, and C and the as-annealed samples are presented in Table 3. IMF provided a YS of 300 MPa and a UTS of 405 MPa.However, a typical YS value is 170-180 MPa and a typical UTS value is 300-320 MPa for conventionally proceeded samples [44], and mode C-processed samples exhibited similar values.Despite the inhomogeneous grain structure, samples Metals 2019, 9, 33 9 of 12 exposed to IMF exhibited 17% elongation, which is close to the values of conventionally processed sheets.The increased strength is explained by the presence of an ultrafine grain structure with ~50% of fine grains being 1.2-2.9µm in size.The rolling at room temperature increased the YS to 500 MPa, which is higher than that of many high strength Al-based alloys of the AA7000 series after solid solution treatment and peak aging.The cold-rolled samples processed by IMF-free Mode C exhibited a YS of 235 MPa and a UTS of 475 MPa and 10% elongation.The significant increase in YS and the decrease in room temperature elongation were observed, due to the inclusion of the IMF treatment in the sheet processing.The elevated temperature annealing at 500 • C for mode B-processed samples provided slightly higher strength characteristics (YS = 195 MPa and UTS = 335 MPa) and significantly higher elongation at room temperature (27%) compare to simple treatment by Mode C.
In summary, samples exposed to IMF exhibited increased strength, due to their high fraction of ultrafine grains, and failed superplasticity, due to their grain structure inhomogeneity.Further increasing the cumulative strain at MDF should increase grain structure homogeneity and should improve the superplastic and mechanical properties.Despite the fact that increasing the number of passes makes the production process more expensive and complex, further investigations on the microstructure and properties of light aluminum alloys exposed to multi-axial deformation are important, and those materials are promising for various industrial applications.Applying the IMF as an intermediate operation in the sheet processing helped to form fine homogeneous grain structures after cold rolling and subsequent recrystallization annealing.As a result, samples exposed to treatment with IMF exhibited increased strength at room temperature and high strain rate superplasticity at elevated temperatures.

Conclusions
The grain structure evolution, mechanical properties and superplastic behavior of the samples of an Al-Mg-Mn-based alloy subjected to isothermal multidirectional forging (IMF) at 350 • C with a cumulative strain of 1.4 to 6.3 and a strain per pass of 0.7 were analyzed.Microstructure and properties of Samples exposed to IMF with subsequent cold rolling, and samples after simple hot and cold rolling were compared.The conclusions are summarized as follows: (1) The micro-shear bands were formed inside the original grains after the second (∑e = 1.4) and the third (∑e = 2.1) IMF passes, and the dynamic recrystallization started during the third IMF pass.The fraction of recrystallized grains increased, and their size decreased in both the periphery and central part of the sample as the cumulative strain increased from 2.1 to 6.3.(2) The IMF processed samples exhibited a non-homogeneous grain structure consisting of fine and coarse grain areas.The sizes of original coarse grains varied in a range of 30-140 µm.Fifty-seven percent of recrystallized grains with a mean size of 1.2 µm were observed in the central part and 43% of recrystallized grains with a mean size of 2.9 µm were observed in the periphery of the samples.The recrystallized fraction was unchanged, and the grain growth occurred in the fine-grained area during annealing in a temperature range of 450-540 • C, leading to an increase in the mean grain size of 1.9-9.4µm.The IMF processed samples exhibited an ultimate tensile strength (UTS) of 405 MPa, a yield strength (YS) of 300 MPa and an elongation to failure of 17%.The non-homogeneous grain structure resulted in non-superplastic behavior of the alloy in a strain rate range of 10 −3 to 10 −2 s −1 at 500 and 540 • C. (3) The sheets exposed to IMF with subsequent cold rolling and recrystallization annealing at 500 • C demonstrated a homogeneous fine grain structure with a mean size of 4.8 µm and high strain rate superplasticity with elongation of 350-500%, a strain rate sensitivity index of m = 0.5-0.6 in a strain rate range of 10 −3 to 10 −2 s −1 at 500 and 540 • C. In comparison, the elongation to failure was 250-350%, and the mean grain size was 8.0 µm in the sheets processed by simple thermo-mechanical treatment.Cold rolling of IMF-processed samples led to an increase in YS to 500 MPa: two times higher than that of samples subjected to a simple thermo-mechanical treatment which included hot and cold rolling.

Figure 1 .
Figure 1.The thermo-mechanical treatment regimes (modes A, B and C).

Figure 2 .
Figure 2. The scheme of sample rotation during isothermal multidirectional forging.

Figure 1 .
Figure 1.The thermo-mechanical treatment regimes (modes A, B and C).

Metals 2019, 9 FOR PEER REVIEW 3 Figure 1 .
Figure 1.The thermo-mechanical treatment regimes (modes A, B and C).

Figure 2 .
Figure 2. The scheme of sample rotation during isothermal multidirectional forging.

Figure 2 .
Figure 2. The scheme of sample rotation during isothermal multidirectional forging.

Figure 3 . 5 Figure 3 .
Figure 3. Evolution of the grain structure in the near-surface layer during isothermal multidirectional forging at 350 • C: (a) as-cast structure; (b) the first pass along the z-axis; (c) the second pass along the y-axis; (d) the third pass along the x-axis; (e) two cycles with ∑e = 4.2; (f) three cycles with ∑e = 6.3; inserts in the (e) and (f) grain structures of the local places at a higher magnification; polarized light, optical microscope.

Figure 4 .
Figure 4.The grain structure in the center part of the sample (a-c) isothermal multidirectional forging at 350 °C at various cumulative strains: (a) 2.1, (b) 4.2, and (c) 6.3;(d) grain structure after hot rolling with a strain of 2.14; polarized light, optical microscope, (e) grain size distribution.

Figure 4 .
Figure 4.The grain structure in the center part of the sample (a-c) isothermal multidirectional forging at 350 • C at various cumulative strains: (a) 2.1, (b) 4.2, and (c) 6.3;(d) grain structure after hot rolling with a strain of 2.14; polarized light, optical microscope, (e) grain size distribution.

3. 2 .
Grain Structure Evolution of IMF Proceeded Samples at Annealing in a Temperature Range of 450-540 • C (Mode A)

Metals 2019, 9 FOR PEER REVIEW 6 Figure 5 .
Figure 5.The grain structure in the center part of the sample after isothermal multidirectional forging at 350 °C with a cumulative strain of 6.3 and subsequent annealing at (a,d,g) 450, (b,e) 500 and (c,f) 540 °C for 30 min; (h) and (i) the grain size distribution in fine grain and coarse grain areas.

Figure 5 .
Figure 5.The grain structure in the center part of the sample after isothermal multidirectional forging at 350 • C with a cumulative strain of 6.3 and subsequent annealing at (a,d,g) 450, (b,e) 500 and (c,f) 540 • C for 30 min; (h) and (i) the grain size distribution in fine grain and coarse grain areas.

Figure 7 .
Figure 7.The flow stress (a,c) and m-index (b,d) vs. the strain rate at 500 (a,b) and 540 °C (c,d).

Figure 7 .
Figure 7.The flow stress (a,c) and m-index (b,d) vs. the strain rate at 500 (a,b) and 540 °C (c,d).

Figure 7 .
Figure 7.The flow stress (a,c) and m-index (b,d) vs. the strain rate at 500 (a,b) and 540 • C (c,d).

Table 3 .
Mechanical properties at room temperature.

Table 3 .
Mechanical properties at room temperature.